Syddansk Universitet Characterization of Cellular and Molecular Heterogeneity of Bone Marrow Stromal Cells Elsafadi, Mona; Manikandan, Muthurangan ; Atteya, Muhammad ; Hashmi, Jamil Amjad ; Iqbal, Zafar ; Aldahmash, Abdullah ; Alfayez, Musaad ; Kassem, Moustapha; Mahmood, Amer Published in: Stem Cells International DOI: 10.1155/2016/9378081 Publication date: 2016 Document Version Publisher's PDF, also known as Version of record Link to publication Citation for pulished version (APA): Elsafadi, M., Manikandan, M., Atteya, M., Hashmi, J. A., Iqbal, Z., Aldahmash, A., ... Mahmood, A. (2016). Characterization of Cellular and Molecular Heterogeneity of Bone Marrow Stromal Cells. Stem Cells International, 2016, [9378081]. DOI: 10.1155/2016/9378081 General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal ? Take down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Download date: 09. Jan. 2017
20
Embed
Syddansk Universitet Characterization of Cellular and ...Research Article Characterization of Cellular and Molecular Heterogeneity of Bone Marrow Stromal Cells MonaElsafadi, 1,2 MuthuranganManikandan,
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Syddansk Universitet
Characterization of Cellular and Molecular Heterogeneity of Bone Marrow Stromal Cells
Elsafadi Mona Manikandan Muthurangan Atteya Muhammad Hashmi Jamil Amjad Iqbal Zafar Aldahmash Abdullah Alfayez Musaad Kassem Moustapha Mahmood AmerPublished inStem Cells International
DOI10115520169378081
Publication date2016
Document VersionPublishers PDF also known as Version of record
Link to publication
Citation for pulished version (APA)Elsafadi M Manikandan M Atteya M Hashmi J A Iqbal Z Aldahmash A Mahmood A (2016)Characterization of Cellular and Molecular Heterogeneity of Bone Marrow Stromal Cells Stem CellsInternational 2016 [9378081] DOI 10115520169378081
General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors andor other copyright ownersand it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights
bull Users may download and print one copy of any publication from the public portal for the purpose of private study or research bull You may not further distribute the material or use it for any profit-making activity or commercial gain bull You may freely distribute the URL identifying the publication in the public portal
Take down policyIf you believe that this document breaches copyright please contact us providing details and we will remove access to the work immediatelyand investigate your claim
Download date 09 Jan 2017
Research ArticleCharacterization of Cellular and Molecular Heterogeneity ofBone Marrow Stromal Cells
Mona Elsafadi12 Muthurangan Manikandan1 Muhammad Atteya13
Jamil Amjad Hashmi4 Zafar Iqbal5 Abdullah Aldahmash16 Musaad Alfayez1
Moustapha Kassem12 and Amer Mahmood12
1Stem Cell Unit Department of Anatomy College of Medicine King Saud University Riyadh Saudi Arabia2Molecular Endocrinology Laboratory (KMEB) Department of Endocrinology University of Southern Denmark Odense Denmark3Department of Histology Faculty of Medicine Cairo University Egypt4Center for Genetics and Inherited Diseases Taibah University Al-Madina Al-Munawara Saudi Arabia5College of Applied Medical Sciences King Saud Bin Abdulaziz University for Health Sciences (KSAU-HS)National Guards Health Affairs Riyadh Saudi Arabia6Prince Naıf Health and Research Center College of Medicine King Saud University Riyadh Saudi Arabia
Correspondence should be addressed to Amer Mahmood ammahmoodksuedusa
Received 14 April 2016 Accepted 26 May 2016
Academic Editor Giorgio Mori
Copyright copy 2016 Mona Elsafadi et al This is an open access article distributed under the Creative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited
Human bone marrow-derived stromal stem cells (hBMSC) exhibit multiple functions including differentiation into skeletal cells(progenitor function) hematopoiesis support and immune regulation (nonprogenitor function)We have previously demonstratedthe presence of morphological and functional heterogeneity of hBMSC cultures In the present study we characterized in detailtwo hTERT-BMSC clonal cell populations termed here CL1 and CL2 that represent an opposing phenotype with respect tomorphology markers expression alkaline phosphatase (ALP) and CD146 and ex vivo differentiation potential CL1 differentiatedreadily to osteoblasts adipocytes and chondrocytes as shown by expression of lineage specific genes and proteins Whole genometranscriptome profiling of CL1 versus CL2 revealed enrichment in CL1 of bone- mineralization- and skeletal muscle-relatedgenes for example ALP POSTN IGFBP5 BMP4 and CXCL12 On the other hand CL2 transcriptome was enriched in immunemodulatory genes for exampleCD14 CD99NOTCH3 CXCL6 CFB andCFI Furthermore gene expressionmicroarray analysis ofosteoblast differentiated CL1 versus CL2 showed significant upregulation in CL1 of bone development and osteoblast differentiationgenes which included several homeobox genes TBX15 HOXA2 andHOXA10 and IGF1 FGFR3 BMP6 MCAM ITGA10 IGFBP5and ALP siRNA-based downregulation of the ALP gene in CL1 impaired osteoblastic and adipocytic differentiation Our studiesdemonstrate the existence of molecular and functional heterogeneity in cultured hBMSC ALP can be employed to identifyosteoblastic and adipocytic progenitor cells in the heterogeneous hBMSC cultures
1 Introduction
Human bonemarrow stromal (also known as skeletal ormes-enchymal) stem cells (hBMSC) are increasingly employedin clinical trials for enhancing tissue regeneration followinginjury [1] Typically hBMSC are isolated by their abilityto adhere to the plastic surfaces of in vitro culture platesHowever the cultured hBMSC exhibit morphological hetero-geneity suggesting the presence of functional heterogeneity[2 3] It has also been suggested that the use of heterogeneous
cell populations in clinical trials of hBMSC-based therapiescaused variability in the observed treatment effects [4] Thusfor the efficient use of hBMSC in therapy better cellular andmolecular characterization of hBMSC is required [1 4]
There exist no specific markers that define the hBMSCphenotype The plastic-adherent hBMSC are defined by thepresence of surface expression of some CD surface markerswith variable sensitivity and specificity [1] Single cell clonalanalysis revealed that only 25 of the cells are true stemcells based on their ability to differentiate into osteoblasts
adipocytes and chondrocytes (trilineage differentiation) andto form heterotopic bone and bone marrow organ whenimplanted in vivo subcutaneously in immune deficient mice[5] The identity of the remaining cells is not clarified butthey may represent lineage-committed cells [3] Therefore itis plausible that functional heterogeneity exists in culturedhBMSC reflecting the in vivo functional and developmentalheterogeneity of hBMSC [6]
In addition to their ability to differentiate into skeletaltissue cells (known as progenitor function) hBMSC possessimmunomodulatory characteristics (known as nonprogen-itor functions) [7] It is not clear whether these differentfunctions are mediated by a number of independent sub-populations within the hBMSC [2] Only a few studies havetried to identify the subpopulation within cultured hBMSCbased on surface markers for example STRO1 and alkalinephosphatase (ALP) but limited molecular phenotyping hasbeen conducted [8]
We have previously demonstrated the presence of mor-phological and functional heterogeneity of clones isolatedfrom telomerized hMSC (hMSC-TERT) cell line [3]The aimof the present study was therefore to further study in detailthe heterogeneity of cultured hBMSC as demonstrated bytwo clonal cell lines with opposite cellular and functionalphenotypeWe also employed theDNAmicroarrays to definetheir molecular signature and signaling pathways associatedwith their functional phenotype
2 Experimental Procedures21 Cell Culture As a model for hBMSC we employedimmortalized hBMSC-TERT cell line that is created fromnormal human BMSC by overexpression of human telom-erase reverse transcriptase gene (hTERT) [9] The hBMSC-TERT cells have been extensively characterized and theyexhibit similar cellular and molecular phenotype to primaryMSC [10] CL1 and CL2 cells are clonal cell populationsof hBMSC-TERT identified in long term culture (passagenumbers 15ndash25) of hBMSC-TERT and were chosen based ontheir distinct and different morphologies Cells were culturedin Dulbeccorsquos Modified Eagle Medium (DMEM) supple-mented with D-glucose 4500mgL 4mM L-glutamine and110mgL sodium pyruvate 10 Fetal Bovine Serum (FBS) 1xpenicillin-streptomycin (Pen-strep) and nonessential aminoacids (all purchased from Gibco-Invitrogen USA) For somecontrol experiments primary bone marrow derived MSC(phBMSC) were employed Sixty milliliters of bone marrowwas aspirated from the iliac crest bone of consenting healthydonors This procedure was approved by the King KhalidUniversity Hospital-King Saud University ethics committeephBMSCwere isolated from bonemarrowmononuclear cellsby plastic adherence as described previously [9]
22 Cell Proliferation Cell proliferation rate was determinedby counting cell number and calculating population doubling(PD) rate The cells were cultured in 25 cm2 tissue culturePetri dish at cell density 05 times 106 cells (28000 cellscm2) Atconfluence the cells were trypsinized and counted manuallyby hemocytometer At each passage population doubling wasdetermined by the following formula log119873 log 2 where 119873
is the number of cells at confluence divided by the initialcell number Cumulative PD level is the sum of populationdoublings and PD rate is PDtime in culture
23 Flow Cytometry Cells were trypsinized to a singlecell suspension were recovered by centrifugation at 200 gfor 5min washed twice in ice-cold PBS supplementedwith 2 BSA and resuspended at a concentration of105 cellsantibody After incubation with the preconjugatedantibodies or matched isotype controls for 30min on icein the dark cells were washed with PBS resuspended in500120583L of PBS and analyzed in the BD FACSCalibur flowcytometer (BD Biosciences) Living cells were gated in adot plot of forward versus side scatter signals acquired ona linear scale At least 10000 gated events were acquiredon a log fluorescence scale Positive staining was distinct asthe emission of a fluorescence signal that surpassed levelsachieved by gt99 of control cell population stained withcorresponding isotype antibodies The ratios of fluorescencesignals versus scatter signals were calculated and histogramswere generated using the software Cell Quest Pro SoftwareVersion 33 (BD Biosciences) The following antibodies wereused all from BD Biosciences FITC-PE-APC-Mouse IgG1kisotype control APC-Mouse Anti-Human CD44 (559942)FITC-Mouse Anti-HumanCD63 (557305) PE-Mouse Anti-Human CD73 (550257) PE-Mouse Anti-Human CD105(560839) PE-Mouse Anti-Human CD146 (550315) PE-Mouse Anti-HumanCD166 (560903) and Alexa Fluor 488Mouse Anti-Human Alkaline Phosphatase (561495)
24 Electron Microscopy After trypsinizing the hBMSC cellsfrom the flasks or 6-well plates were collected the sampleswere washed with PBS and the pellets were resuspendeddirectly in 25 glutaraldehyde in 01M phosphate buffer (pH72) and kept at 4∘C for 4 hr First the cells were washed with01Mphosphate buffer (pH72) and transferred to 1osmiumtetroxide (OsO4) solution in 01M phosphate buffer (pH 72)for two hr The cells were dehydrated in ascending grades ofethanolThe cells were then resuspended in acetone and werealiquoted into BEEM embedding capsules and infiltratedwith acetone resin mixture followed by embedding in apure resin mixture for two hr Semithin sections (05 120583mthickness) and ultrathin sections (70 nm thickness) wereprepared examined and photographed under a transmissionelectron microscope (TEM) (Jeol 1010 Jeol Tokyo Japan)
25 In Vitro Osteoblast Differentiation Cells were grownin standard DMEM growth medium in 6-well plates at20000 cellcm2 When 70ndash80 confluence was achieved testcells were cultured inDMEM supplementedwith osteoblasticinduction mixture (referred to as OS) containing 10 FBS1 Pen-strep 50120583gmL L-ascorbic acid (Wako ChemicalsNeuss Germany) 10mM 120573-glycerophosphate (Sigma) and10 nM calcitriol (112057225-dihydroxyvitamin D3 Sigma) and10 nM dexamethasone (Sigma) noninduced cells (referredto as Cont) were cultured in normal growth media for thesame duration as induced The media were replaced threetimes per week Cells cultured in standard culture mediumwere considered as control At day 14 of differentiation
Stem Cells International 3
mineralized nodules became apparent and were stained withAlizarin Red S and ALP
26 In Vitro Adipocyte Differentiation Cells were grown instandard DMEM growth medium in 6-well plates at 03times 106 cellsmL At 90ndash100 confluence cells were culturedin DMEM supplemented with adipogenic (Adip) inductionmixture containing 10 FBS 10 Horse Serum (Sigma)1 Pen-strep 100 nM dexamethasone 045mM isobutylmethylxanthine (Sigma) 3120583gmL insulin (Sigma) and 1 120583MRosiglitazone (Novo Nordisk Bagsvaerd Denmark) Themedia were replaced three times per week Cells cultured instandard culture medium were considered as control Fromday 3 of differentiation small lipid droplets became visibleand at day 7 they were stained with Oil Red-O and Nile red
27 In Vitro Chondrogenic Differentiation Both CL1 and CL2cells were trypsinized and counted around 1 times 106 cellstaken in each 15mL conical tube centrifuged at 400timesg for5 minutes For chondrocyte differentiation pellet culturesystem used chondrocyte induction was done in mediacontaining advanced DMEMF12 supplemented with 1 ITSPremix Tissue Culture Supplement 100 nM dexamethasoneGlutamax and 10 ngmL transforming growth factor-beta-3(TGF120573-3) Cells were maintained in chondrocyte differentia-tion media for 21 days and changed every two days
28 Cytochemical Staining281 Alkaline Phosphatase (ALP) Staining CL1 and CL2 cellswere stained before OS differentiation for the basal ALPexpression and after OB differentiation at day 7 of inductionCells cultured in 6-well plates were washed in PBS minusminus(-Ca -Mg) andfixed in acetonecitrate buffer 10mMat pH42for 5min at room temperature The NaphtholFast Red stain[02mgmL Naphthol AS-TR phosphate substrate (Sigma)][0417mgmL of Fast Red (Sigma)] was added for one hourat room temperature
Histological tissue blocks were sectioned at 4 micronsImmunohistochemical staining was performed on CL1 andCL2 chondrocyte 3D pellets using DAKO EnVision andPowerVision according to the manufacturerrsquos instructions(DAKO Glostrup Denmark) Briefly paraffin sections wereincubated for 1 hour at room temperature with primaryantibodies diluted in ChemMate (DAKO) (HumanAnti-Col-10 and Human Anti-Col-2 ABI) Sections were washed sub-sequently in Tris-buffered saline (TBS 005M pH 74) incu-bated for 30 minutes with secondary anti-mouse IgHRP-conjugated polymers (K4001 En Visionthorn DAKO) and visu-alizedwith 330-diaminobenzidine tetrahydrochloride (DABS3000 DAKO) or with 3-amino-9-ethylcarbazole (AECDAKO) according to manufacturerrsquos instruction Controlswere performed with nonimmune immunoglobulins of thesame isotype as the primary antibodies (negative controls)and processed under identical conditions Alcian blue stain-ing was used to detect chondrocytes Sections of paraffin-embedded implants were stained with Alcian blue (Sigma)solution pH 25 at this pH all the glycoproteins (neutral andacidic) will be stained blue
29 Alizarin Red S Staining for Mineralized Matrix Seven-day-old OS differentiated cells in 6-well plates were used forAlizarin Red S staining The cell layer was washed with PBSand then fixed with 70 ice-cold ethanol for 1 hr at minus20∘CAfter removing the ethanol the cell layer was rinsed withdistilled water and stained with 40 nM AR-S (Sigma) pH 42for 10 minutes at room temperature Excess dye was washedoff with water followed by a wash with PBS for few minutesto minimize nonspecific AR-S stain
For quantifying the Alizarin Red S staining the air-driedplates the Alizarin Red S dye was eluted in 800 120583L of aceticacid incubated in each well for 30 minutes at room temper-ature as described [11] and measured in spectrophotometer(BioTek Epoch) at 405 nm
210 Quantitative ALP Activity To quantify ALP activity inCL1 and CL2 hBMSC before and after OS differentiationwe used the BioVision ALP activity colorimetric assay kit(BioVision Inc CA USA) with some modifications Cellswere cultured in 96-well plates under normal conditionsthen on day of analysis wells were rinsed once with PBSand were fixed using 37 formaldehyde in 90 ethanol for30 seconds at room temperature Subsequently fixative wasremoved and 50120583L of pNPP solution was added to each welland incubated for 1 hour in the dark at room temperatureThe reaction was subsequently stopped by adding 20 120583L stopsolution and gently shaking the plate ODwas thenmeasuredat 405 nm
211 Oil Red-O Staining for Lipid Droplets CL1 and CL2 cellsdifferentiated to adipocytes with Adip induction media atday 7 were used Accumulated cytoplasmic lipid dropletswere visualized by staining with Oil Red-O After washingcells grown in 6-well plates with PBS the cells were fixedin 4 formaldehyde for 10min at room temperature andthen rinsed once with 3 isopropanol and stained for 1 hr atroom temperature with filtered Oil Red-O staining solution(prepared by dissolving 05 g Oil Red-O powder in 60isopropanol) To quantify staining of fat droplets Oil Red-O was used as a stain Oil Red-O was eluted by adding 100isopropanol to each well and color changes were measuredby spectrophotometer at 510 nm (BioTek SpectrophotometerEpoch)
212 Nile Red Fluorescence Determination and Quantificationof Adipogenesis A stock solution of Nile red (1mgmL) inDMSO was prepared and stored at minus20∘C protected fromlight Staining was performed on unfixed cells Culturedundifferentiated and day 7 adipocyte differentiated cells weregrown in Corning polystyrene flat bottom 96-well TC-treated black microplates (Corning NY USA) were washedonce with PBS The dye was then added directly to the cells(5 120583gmL in PBS) and the preparation was incubated for10min at room temperature and then washed twice with PBSFluorescent signal was measured using SpectraMaxM5 fluo-rescence spectrophotometer plate reader (Molecular DevicesCo Sunnyvale CA USA) using bottom well-scan modewhere nine readings were taken per well using Ex (485 nm)and Em (572 nm) spectra
4 Stem Cells International
213 Quantitative Real-Time PCR (qRT-PCR) Analysis TotalRNA was extracted using MagNA pure compact RNA iso-lation kit (Roche Applied Science Germany Cat number04802993001) in an automatedMagNA pure compact system(Roche Germany) as recommended by the manufacturerThe total RNAwas quantified by Nanodrop spectrophotome-ter (Nanodrop 2000 Thermo Scientific USA) Complemen-tary DNA (cDNA) was synthesized from 1 120583g of the RNAsamples usingHighCapacity cDNAReverse Transcription kit(Applied Biosystems USA) using Labnet Multigene thermo-cycler according to the manufacturerrsquos instructions Relativelevels of mRNA were determined from cDNA by real-timePCR (Applied Biosystems-Real-Time PCRDetection System)with Power SYBR Green PCR kit (Applied BiosystemsUK) according to the manufacturerrsquos instructions Followingnormalization to the reference gene GAPDH quantificationof gene expression was carried out using a comparative Ctmethod where ΔCt is the difference between the CT valuesof the target and the reference gene and fold induction wasperformed from the control (Cont) for the same time pointPrimers (Supplementary Table 1 in Supplementary Materialavailable online at httpdxdoiorg10115520169378081)were obtained from Applied Biosystems (USA) as TAQMANprimers or previously published primers were used (seeSupplementary Table 1)
214 DNAMicroarray Global Gene Expression Analysis Fourhundred ng of total RNA was used as input for gener-ating biotin-labeled cRNA (Ambion Austin TX UnitedStates) cRNA samples were then hybridized onto Illuminahuman-8 BeadChips version 3 Hybridization washing Cy3-streptavidin staining and scanning were performed on theIllumina BeadStation 500 platform (Illumina San DiegoCA USA) according to the manufacturerrsquos instructions andeverything was done in triplicate Expression data analysiswas carried out using the Partek genomic suite softwareRaw data were background-subtracted normalized usingthe ldquorank invariantrdquo algorithm and filtered for significantexpression on the basis of negative control beads Geneswere considered significantly expressed with detection 119901values le 001 Differential expression analysis was performedwith the Illumina custom method using freshly isolatedprimary hBMSC (used at passage 3) as a reference controlThe following parameters were set to identify statisticalsignificance differential 119901 values le 001 fold change ratiogt 15 Pathway analysis was performed using DAVID Bioin-formatics Resources 67 (httpdavidabccncifcrfgov) andGeneSpring GX software (Agilent Technologies) Pathwayanalysis for CL1 OS D14 versus CL2 OS D14 was conductedusing the Single Experiment Pathway analysis feature inGeneSpring 120 (Agilent Technologies)
215 Small Interfering (si)RNATransfection For transfectionhBMSC in logarithmic growth phase were transfected withSilencer Select Predesigned ALP siRNA (25 nM) (Assay IDs1298 and Cat number 4390824) (Ambion The RNA Com-pany USA) using Lipofectamine RNAiMAX Reagent (Invit-rogen CA USA) plus serum-free Opti-MEMI mediumunder the conditions described by the manufacturer At
day 3 of transfection the cells were induced for osteogenicdifferentiation for an additional 7 days ALP staining wasused as a control for the siRNA transfection efficiency andtimeline
216 Statistical Analysis All of the results were presentedas the mean and standard deviation (SD) of at least 3independent experiments with 3ndash5 technical repeats in eachexperiment Studentrsquos 119905-test (two-tailed) was used for testingdifferences between groups 119901 value lt005 was consideredstatistically significant
3 Results
31 Comparison between CL1 and CL2 Differences in Mor-phology Proliferation and Marker Expression Profile Weisolated two distinct clonal cell populations of hBMSC-TERThBMSC-CL1 and hBMSC-CL2 (for easiness will be termedhereafter CL1 and CL2) based on differences in cell morphol-ogy (Figure 1(a)) CL1 cells had cuboidalmorphologywhereasCL2 cells have spindle-shaped fibroblast-like morphologyCL1 cells had higher proliferation rate compared to CL2(Figure 1(b)) mean PD rates of CL1 and CL2 were 0714 and0429 PDday respectively (Figure 1(b)) Both CL1 and CL2expressed surface marker profiles characteristics of hBMSC(gt90) CD44+ CD63+ CD73+ CD105+ and CD166+(Figure 1(c)) However CL1 cells showed higher expressionof CD146 (927 versus 12) and ALP (98 versus 0)compared to CL2 (Figure 1(d)) TEM revealed the presenceof abundant pseudopodia in CL1 indicating high motility(Figure 1(e)(A)) as well as well-developed mitochondriaand rough endoplasmic reticulum (rER) suggesting highmetabolic activity CL2 cells contained abundant phagocyticvacuole (pv)microvilli (mi) and lysosomes (ly) (Figure 1(e))
We performed quantitative real-time PCR (RT-PCR)for genes expressed in mesodermal progenitor cells [12]CL1 expressed higher levels of BMP4 MIXL1 WNT3a andTWIST compared to CL2 (Figure 1(f) 119901 lt 001) In contrastCL2 expressed higher levels of Kinase Insert Domain Recep-tor (Type III Receptor Tyrosine Kinase) (KDR) expressed inendothelial cells and smoothmusclemyosin heavy chain gene(smMHC) expressed in smooth muscle cells (Figure 1(f))
32 CL1 Cells Exhibit Enhanced Osteoblast DifferentiationFollowing osteoblast (OB) differentiation induction ALPstaining andALP enzymatic activity were significantly higherin CL1 compared to CL2 cells (Figure 2(a) 119901 lt 001)Similarly Alizarin Red staining and quantitation of formedmineralized matrix were more pronounced in CL1 cells (Fig-ure 2(b) 119901 lt 001) In addition CL1 cells expressed higherlevels of osteoblastic genes ALP RUNX2 and osteopontin(OPN) (Figure 2(c) upper panel) compared to CL2 cells
Global gene expression microarray analysis of OB dif-ferentiated cells at day 14 showed around 1060 genes signif-icantly upregulated more than 2-fold (119901 lt 001) in CL1Among the upregulated genes 80 genes were annotated tobone development and osteoblast differentiation (Table 1)The highest upregulated genes included paired-like home-odomain 2 (PITX2) Insulin-like growth factor 1 (IGF1) and
Stem Cells International 5
CL1
CL2
(a)
Cum
ulat
ive P
D
Cumulative days
Growth curve
0 10 20 30 40
CL1 CL2
CL1 CL20
5
10
15
20
25
0
01
02
03
04
05
06
07
08
(PD
day
)
(b)
CL2
103
102
101
100
103
102
101
100
103
102
101
100
103
102
101
100
103
102
101
100
103
101
100
103
102
102
101
100
103
102
101
100
103
102
101
100
103
102
101
100
CL1
CD44 CD63 CD73 CD166 CD105
T3 CD44 T3 CD63 T3 CD73 T3 CD166 T3 CD105
0
20
40
60
80
Cou
nt
0
20
40
60
80
100
120
Cou
ntC
ount
Cou
nt
Cou
nt
Cou
nt
Cou
nt0
20
40
60
80
Cou
nt
0
20
40
60
80
100
Cou
nt
0
20
40
60
80
Cou
nt
80
60
40
20
0
100
50
0
120
100
80
60
40
20
0
120
100
80
60
40
20
0
100
80
60
40
20
0
(c)
ISO CD146 ALP
ISO CD146 ALP
FL2-
H
FL2-
H
FL2-
HPEPEPECL
2CL
1
FL1-H FL1-H FL1-H
FITCFITCFITC
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
(d)
Figure 1 Continued
6 Stem Cells International
(A) (B)
(e)
Fold
indu
ctio
n
CL1CL2
lowast
lowastlowast lowast
BMP4 MIXL1 WNT3a TWIST
lowastlowast
KDR smMHCminus01
01
03
05
07
09
11
13
15
mdash
050
100
150
200
250
300
CL1CL2
(f)
Figure 1 Cellular phenotype of two clonal populations of humanbonemarrow stromal stem cells CL1 andCL2 (a) Cellmorphology CL1 cellsshowed large cuboidal morphology while CL2 cells had spindle-shaped fibroblast-like morphology (b) Growth curves showing populationdoubling (PD) rate during long term culture (c) Flow cytometry analysis (FACS) shows expression of CD44 CD63 CD73 CD105 andCD166in CL1 and CL2 cells Matched isotype control was used for gating (d) Flow cytometry analysis presented as dot blot of CD146 and alkalinephosphatase (ALP) cell surface proteins (e) Transmission electron microscope (TEM) (A) CL1 (1200x) (B) CL2 (1500x) n nucleus nunucleolus rER rough endoplasmic reticulum ly lysosomes pv phagocytic vacuole and rer reticular stalk of rER (f) Gene expressionanalysis using RT-PCR for a group of mesodermal and stromal genes Gene expression was normalized to GAPDH and presented as foldchange Data is shown as mean plusmn SD of three independent experiments lowast119901 lt 005 119901 lt 0001
collagen type V alpha 3 (COL5A3) osteomodulin (OMD)and T-box 15 (TBX15) (Table 1) Furthermore several knownosteoblast-related genes were upregulated in CL1 cells suchas bone morphogenetic protein 6 (BMP6) fibroblast growthfactor receptor 3 (FGFR3) insulin-like growth factor bindingprotein 5 (IGFBP5) and vitamin D (125-dihydroxyvitaminD3) receptor (VDR) (Table 1) On the other hand 1200 geneswere upregulated in CL2 cells 255 genes were annotated toimmunity and immune response and defense This categoryincluded genes from complement system chemokine (C-Cmotif) ligands interferon family chemokine (C-X-C motif)ligands and receptor major histocompatibility complex classII molecules interleukins and tumor necrosis factor receptorsuperfamily (Table 2 and Supplementary Table 3)
33 CL1 Cells Exhibit Enhanced Adipocyte DifferentiationWe observed significant differences between CL1 and CL2in their response to adipocytic differentiation inductionCL1 differentiated readily to adipocytes compared to CL2(Figure 2(d) lower panel) evidenced by higher levels ofadipocyticmarkers gene expression LPL (lipoprotein lipase)and adiponectin as well as formation of mature lipid filledadipocytes visualized by Oil Red-O staining and quantitativeNile red staining (Figure 2(d))
34 CL1 Cells Differentiate to Chondrocytic Lineage Inpellet cultures CL1 cells formed 3D pellets containingproteoglycan-secreting chondrocytes which stained positivewith Alcian blue Limited chondrocyte differentiation wasvisible in cell pellets of CL2 cells The differentiated chon-drocytes in CL1 pellets expressed higher levels of collagen Xand collagen II which was overlapping the Alcian blue stain(Figure 2(e))
35 Molecular Signature of CL1 and CL2 Cells To define themolecular signature and molecular differences between CL1
and CL2 we compared the basal gene expression patternof CL1 and CL2 cells using DNA microarrays The PCAanalysis showed a clear separation between CL1 and CL2(Supplementary Figure 1) Comparison between CL1 withCL2 showed that 915 geneswere differentially expressed in thetwo cell lines (gt2-fold119901 lt 001) 462 genes were upregulatedand 452 were downregulated in CL1 versus CL2 The mostrelevant genes that were upregulated in CL1 are listed inTable 3(a) Among these 35 highly expressed genes in CL1 thefollowing 11 genes were present in skeletal and muscular sys-tem development and function FOLR3 CCL3L1 SERPINB2POSTN IGFBP5 CCL3 NOV ALP TNFRSF11B ACTG2and CDH11 (Table 3(a)) Functional annotation of the upreg-ulated genes in CL1 using the Ingenuity Pathway Analysis(IPA) revealed enrichment in the following categories ldquotissuedevelopmentrdquo ldquoskeletal and muscular system developmentand functionrdquo and ldquoorganismal developmentrdquo (Table 3(b))Furthermore the DAVID annotation tool was employed toassess the functional relationships of the upregulated genesin CL1 showing enrichment in ontologies ldquoskeletal andmuscular system development and functionrdquo that includedbone size osteoblast differentiation bone mineralizationand bone mineral density (Table 3(c)) CL1 exhibited upreg-ulation of WNT pathway ligands WNT5B (2-fold) andLRP5 (2-fold) (Table 1) Also ALP was among the highlyexpressed genes together with POSTN IGFBP5 SPP1 IL-6 and DKK1 (Tables 3(a) and 1) These genes are knownto play an important role in osteoblast differentiation andbone formation For CL2 inhibitors of WNT pathway wereupregulated and included SFRP1 (11-fold) DKK2 (32-fold)FGF2 (31-fold) and GBP2 (24-fold) Functional annotationof the upregulated genes in CL2 revealed enrichment in thefollowing categories ldquodevelopmental processrdquo ldquomulticellularorganismal processrdquo ldquobiological adhesionrdquo and ldquoimmunesystem processrdquo (Supplementary Table 2A) In-depth analysis
Stem Cells International 7
CL1
CL2
ALPCont D 14
Cont OB I
ALP activity
mdash100
200
300
400
500
600
Fold
chan
ge
CL1CL2
(a)
CL1
CL2
ALZCont D 14
Fold
chan
ge
25
20
15
10
5
0
Cont d14 d21
CL1CL2
ALZ quantification
(b)
CL1CL2
CL1CL2
LPL ADIPOQ
Cont ContADIP D14 ADIP D14
000
100
200
300
400
500
600
700
0
1000
2000
3000
4000
5000
CL1CL2
CL1CL2
CL1CL2
RUNX2 ALP OPN
Cont Cont ContOS D7 OS D7 OS D7
00
05
10
15
20
25
30
00
05
10
15
20
25
30
35
00
05
10
15
20
25
30
35
(c)
CL1CL2
CL1
CL2
Control Adip d7 Adip d14
d7 d140
005
01
015
02
025
03
035
04
045
Nile
red
fluor
esce
nt in
tens
ity
(d)
CL1
CL2
Alcian blue Col X Col II
(e)
Figure 2 In vitro osteoblastic and adipocytic differentiation of two clonal populations of human bone marrow stromal stem cells CL1 andCL2 cells Both cell lines were induced for osteoblast differentiation using standard protocol described in the Methods (a) ALP staining atday 14 in control noninduced (Cont) and osteoblast induced cells (D14) Right panel shows ALP activity (119899 = 3 independent experiments119901 lt 0001) (b) Mineralized matrix formation visualized by Alizarin Red S staining Right panel shows Alizarin Red quantification at day 14and 21 after osteoblast differentiation (119899 = 3 independent experiments 119901 lt 0001) (c) Quantitative RT-PCR of osteoblastic and adipocytegene markers in CL1 and CL2 during osteoblast (upper panel) and adipocyte (lower panel) differentiation ALP = alkaline phosphataseOPN = osteopontin LPL = lipoprotein lipase and ADIPOQ = adiponectin Data are presented as fold change in expression of each targetgene normalized to GAPDH (119899 = 3 independent experiments 119901 lt 005
119901 lt 0001) (d) CL1 and CL2 lines were induced for adipocytedifferentiation using standard protocol described in the Methods Adipocyte formation was visualized at day 7 (Adip d7) and day 14 (Adipd14) by Oil Red-O staining Lower panel presents quantification of Nile red staining (119899 = 3 independent experiments 119901 lt 0001) (e) CL1and CL2 lines were induced for chondrocyte differentiation using 21-day pellet culture method as described in the Methods The pellets werestained with Alcian blue collagen 10 (Col X) and collagen 2 (Col II) (original magnification 5x)
8 Stem Cells International
Table 1 Microarray data analysis showing genes related to bone development and osteoblast differentiation upregulated in CL1 versus CL2cells
Probe ID Genbank accession Gene name Gene symbol FCA 23 P167367 NM 153426 Paired-like homeodomain 2 PITX2 30884
of the biological processes revealed several immune-relatedpathways ldquoMAPKKK cascaderdquo ldquoimmunity and defenserdquoldquosignal transductionrdquo ldquoextracellular matrix protein-mediatedsignalingrdquo and ldquointerferon-mediated immunityrdquo among oth-ers that were upregulated (Supplementary Table 2B) Also40 genes related to immune system related factors wereidentified as significantly enriched in CL2 compared to CL1cells (Supplementary Table 2C)We chose the following genesfor validation of the microarray results NOV IGFBP5 ALPTAGLN and CDH11 as they were highly expressed in CL1RT-PCR analysis confirmed themicroarray results (Figure 3)
Furthermore we compared the molecular phenotype ofCL1 and CL2 cells with that of phBMSCWe found that morethan 80 of the genes expressed in CL1 and 90 in CL2 cellswere commonwith primary phBMSC (Supplementary Figure2) suggesting that CL1 and CL2 molecular phenotype existwithin the heterogeneous population of phBMSC cultures
36 ALP Knockdown Impairs Differentiation of CL1 CellsSince ALP has been suggested as a marker for hBMSCprogenitor cell lineage commitment [13] and was highlyupregulated in CL1 cells we tested its biological role in
CL1 cells ALP siRNA transfection decreased ALP proteinlevel ALP activity and mRNA gene expression compared tocontrol cells transfected with control siRNA (119901 lt 001) andthis inhibition was detectable up to day 7 days after osteoblastdifferentiation induction (Figures 4(A) and 4(B)) At day 14 ofdifferentiation mineralization ability of CL1 was significantlyimpaired (Figure 4(C)) In addition we found that thenumber of mature adipocyte formations was significantlyreduced to more than 75 (119901 lt 001) (Figure 4(D))
To identify relevant adipocyte differentiation associatedgenes that were targeted by ALP deficiency we compared thedownregulated genes of ALP deficient CL1 with the upreg-ulated genes identified during adipocytic differentiation ofCL1 We identified 62 genes that were common (Figure 5(a)Table 4) and among these genes were genes related tometabolism (primarily lipid and carbohydrate) and transportincluding CYB5B CHST1 TAP1 ATP8A1 LRP8 PLCD1and FABP5 (Table 4) We further performed quantitativereal-time PCR of ALP deficient CL1 cells during adipocytedifferentiation The following adipocyte-associated geneswere downregulated PPAR1205742 LPL and aP2 (Figure 5(b))
12 Stem Cells International
Table 3 Whole genome microarray analysis of two clonal popu-lations of human bone marrow stromal stem cells CL1 and CL2cells (a) Top 35 highly upregulated genes in CL1 cells versus CL2cells (b) Ingenuity Pathway Analysis (IPA) showing the differentphysiological system development and function genes found in eachcategory and the corresponding 119901 value (c) Upper part IPA analysisshowing the categories for skeletal and muscular system developmentand function upregulated in CL1 cells and lower part showing genesupregulated in CL2 cells NOV = nephroblastoma overexpressedIGFBP5 = insulin-like growth factor binding protein 5 ALP TAGLN= transgelin and CDH11 = OB-cadherin (osteoblast)
Name 119901 value moleculesPhysiological system development and function
Organismal development 705119864 minus 10ndash143119864 minus 03 181Embryonic development 100119864 minus 09ndash143119864 minus 03 154Organ development 100119864 minus 09ndash143119864 minus 03 145
(b) Continued
Name 119901 value moleculesSkeletal and muscularsystem development andfunction
100119864 minus 09ndash102119864 minus 03 123
Tissue development 100119864 minus 09ndash143119864 minus 03 236
(c)
Functions annotation 119901 value moleculesSkeletal and muscular system development and
function upregulated in CL1 cellsSize of bone 143119864 minus 06 24Differentiation ofosteoblasts 384119864 minus 06 25
Mineralization ofbone 493119864 minus 06 19
Bone mineral density 365119864 minus 05 19Skeletal and muscular system development and
function upregulated in CL2 cellsDevelopment ofmuscle 100119864 minus 09 44
Proliferation ofmuscle cells 229119864 minus 06 35
Remodeling of bone 363119864 minus 06 21Resorption of bone 394119864 minus 06 19
confirming impairment of adipocytic differentiation of ALPdeficient CL1 cells
4 Discussion
We extensively studied two cell populations within culturedhBMSC that were identified based on differences in mor-phology Cellular and molecular studies revealed differencesin growth differentiation capacity and molecular signatureOur data support the notion of the presence of cellular andfunctional heterogeneity among cultured hBMSC
Cellular heterogeneity of cultured hBMSC is recognizedin an increasing number of reports Several extrinsic andintrinsic factors may contribute to the observed hBMSCheterogeneity Extrinsic factors include donor-to-donor vari-ations in the number and quantity of initiating cells whichresult in differences in cell growth rate and differentia-tion capacity [14 15] Intrinsic factors have been examinedemploying single cell clonal analysis and revealed variationsin differentiation potential among individual colonies withinhBMSC cultures ranging from the presence of cells withtrilineage (osteoblast adipocytes and chondrocyte) potencyto cells with null potency [16] Also variations in the abilityof clonal cells to form heterotopic bone when implanted invivo have been reported [5] Our study corroborates thesefindings and provides more detailed cellular and molecularphenotyping of two examples of cell populations that existwithin the heterogeneous hBMSC cultures [17]
Determining the molecular signature of CL1 and CL2using whole genomemicroarray analysis showed enrichmentof lineage-commitment associated genes inCL1 For exampleinsulin-like growth factor 5 (IGFBP5) and interleukin 6 (IL6)
Stem Cells International 13
NOV0040
0030
0020
0010
mdash
IGFBP50015
0010
0005
mdash
ALP0100
0080
0060
0040
0020
mdash
TAGLN0800
0600
0400
0200
CDH11
0060
0040
0020
mdashmdash
LRP30015
0010
0005
mdash
CL1 CL2 CL1 CL2 CL1 CL2
CL1 CL2CL1 CL2CL1 CL2
Figure 3 Validation of whole genome microarray analysis of two clonal populations of human bone marrow stromal stem cells CL1 andCL2 cells Quantitative real-time PCR for highly expressed genes in CL1 cells NOV = nephroblastoma overexpressed IGFBP5 = insulin-likegrowth factor binding protein 5 ALP TAGLN = transgelin and CDH11 = OB-cadherin (osteoblast) Data are presented as fold change inexpression of each target gene normalized to GAPDH (119899 = three independent experiments 119901 lt 0001) (see also Table 3)
were 147- and 33-fold upregulated in CL1 cells respectivelyBoth factors are expressed in osteoprogenitor cells andimportant for osteoblast maturation [18] We also observedthat periostin (POSTN) gene was highly upregulated inCL1 cells (156-fold) POSTN is a 90 kDa secreted proteinoriginally identified in murine osteoblast-like cells and isupregulated by PTH [19] Several studies employing murineand human cells have revealed important role of POSTNin osteoblast differentiation and during development inintramembranous ossification [20ndash23] Another factor iden-tified in CL1 cells is nephroblastoma overexpression (NOV)which is a member of the Cyr 61 connective tissue growthfactor (CNN) family The CCN family of proteins promotesosteoblast differentiation through interaction with integrinsWNT BMP and NOTCH signaling pathways [24ndash26] Inaddition a large number of signaling molecules known tobe regulators of hBMSC lineage specific differentiation forexample insulin-like growth factors [27ndash29] WNT [30ndash32]and MAPK [33ndash35] were enriched in CL1 cells In contrastCL2 expressed high levels of immune-related genes whichmay explain the poor differentiation response to osteoblastor adipocyte lineage In a recent study the authors usedtelomerized hBMSC and showed clearly a clonal populationthat had very low in vitro and in vivo differentiation abilityhowever they had enhanced immune-related features includ-ing high IL7 expression These nullipotent cells expressedCD317 which was associated with remarkably high basal levelexpression of factors with a proinflammatory and antiviralfunction [17] We observed that this molecular phenotypewas associated with distinct ultrastructural characteristicsof the cells In particular CL2 had abundant phagocyticvacuole microvilli and lysosomes features reminiscent of
ultrastructure of immune-regulatory cells Our data thussupport the increasingly recognized feature that hBMSCexhibit immune modulatory functions and a part of theinnate immune response [17]
We observed that ALP protein expression and enzymaticactivity were significantly different between CL1 and CL2cell lines and were thus a potential marker that distinguishesdifferent cell populations with progenitor functions (CL1)from cells with nonprogenitor functions (CL2) ALP isexpressed in a wide variety of tissues including kidneysbone and liver [36 37] but tissue-nonspecific ALP (ALPL)is considered a commitment marker for osteoblastic lineage[13 38] However in a recent study the authors examined thedifferentiation potential of a number of hMSC clones in vitroand in vivo and reported that the hMSC clones with highlevels of ALP expression were committed to trilineage differ-entiation [13] Our data corroborate and extend these findingsby reporting the effects of siRNA-mediated inhibition ofALP that resulted in an impaired hBMSC differentiation notonly to osteoblasts but also to adipocytes Also our resultscorroborate earlier studies that demonstrated in human bonebiopsies the presence of ALP expression in bone marrowadipocytic cells [13 17] All these data suggest that ALP isa ldquostemnessrdquo marker of hBMSC and not just an indicator ofosteoblastic lineage commitment
While CL1 and CL2 were isolated from telomerizedhMSC cell line they are relevant to normal human physi-ology We observed that the molecular phenotypes of CL1and CL2 were contained within the molecular signatureof primary hBMSC suggesting that CL1 and CL2 repre-sent cell populations within the heterogeneous cultures ofhBMSCWe have also previously reported that the molecular
14 Stem Cells International
ALZ
stai
ning
Control siRNA ALP siRNA
ALP
Beta
-act
in
Con
trol
ALP
siRN
A d3
ALP
siRN
A d7
Nile red quantificationlowast
NI ADIP I ALP siRNA0
50
100
150
200
Con
trol s
iRN
AA
LP si
RNA
Day 3 Day 5 OS induction Day 7 OS induction
Fold
indu
ctio
n lo
g sc
ale
10000
1000
100
10
1
Day 0 OS day 7
Control siRNAALP siRNA
(A) (B)
(C)
(D)
Figure 4 Effect of alkaline phosphatase (ALP) gene silencing by small interfering RNA (siRNA) on a clonal population of human bonemarrow stromal stem cell CL1 (A) ALP staining at day 3 day 5 or day 7 days of osteoblast differentiation (OB induction) (B) Quantitativereal-time PCR for ALP gene following ALP siRNA transfection at day 0 OB and day 7 of OB Data are presented as fold change in expressionof each target gene normalized to GAPDH (119899 = three independent experiments 119901 lt 001) Western blotting analysis of day 3 and day 7after siRNA ALP transfection of CL1 cells ALPL specific antibody and B-actin was used (C) Mineralized matrix formation as visualizedby Alizarin Red S staining in siRNA transfected CL1 cells after 14 days of OB induction (D) Nile red quantification of mature lipid filledadipocyte in control noninduced (Cont) adipocyte induced (Adip I) and ALP siRNA transfected cells that are adipocyte induced (ALPsiRNA) Adipocyte induction was carried out for 7 days lowast119901 lt 005
phenotype and cellular responses of hMSC-TERT are similarto those of primary hMSC [10]While we have identified ALPas a marker that can be used for a prospective identificationof differentiation committed population of hBMSC we iden-tified additional distinctive molecular markers of the cellsFor example IGF-1 IGF-2 and IGF binding protein 5 wereenriched in CL1 compared to CL2 IGFs and their bindingproteins are very well-studied factors that play a role inhBMSC proliferation and osteoblast differentiation [18] On
the other hand annexin A3 as well as several immune-relatedgenes was highly enriched in CL2 compared to CL1 Futurestudies are needed to determine the functional significanceof these molecules in relation to the functional identity ofvarious cell populations within the hBMSC cultures andtheir usefulness as biomarkers to dissect the heterogeneouspopulation of cultured hBMSC
Our finding of the presence of functional diversity withinhBMSC cultures that contain progenitor and nonprogenitor
Stem Cells International 15
Table 4 Microarray data analysis showing genes found to be upregulated during adipogenic differentiation and downregulated after ALPKO
62 common elements inldquoAD uprdquo and ldquoALP downrdquo Gene name FC (ALP siRNA versus
control siRNA)APOBEC3G Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3G minus15938025IFI44L Interferon-induced protein 44-like minus11520283PAQR5 Progestin and adipoQ receptor family member V minus6868241PNMA2 Paraneoplastic antigen MA2 minus59695344DUSP23 Dual specificity phosphatase 23 minus54786854CLDN23 Claudin 23 minus51885047ANKDD1A Ankyrin repeat and death domain containing 1A minus51646647IL8 Interleukin 8 minus4887188LRRC23 Leucine rich repeat containing 23 minus47611775IL6 Interleukin 6 (interferon beta 2) minus4693139LIFR Leukemia inhibitory factor receptor alpha minus46540866PTGFR Prostaglandin F receptor (FP) minus4457529FAM134B Family with sequence similarity 134 member B minus4403495CYFIP2 Cytoplasmic FMR1 interacting protein 2 minus4260462METTL7A Methyltransferase-like 7A minus40480843APOBEC3F Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3F minus39502614CA5B Carbonic anhydrase VB mitochondrial minus393889ITGA10 Integrin alpha 10 minus39143775FMO3 Flavin containing monooxygenase 3 minus3852087IMPA2 Inositol monophosphatase 2 (human) minus38374884CDO1 Cysteine dioxygenase type I minus38181455CCDC68 Coiled-coil domain containing 68 minus37292893
HEXDC Hexosaminidase (glycosyl hydrolase family 20 catalytic domain) containing minus2094836SARM1 Sterile alpha and TIR motif containing 1 minus20797038BBS1 Bardet-Biedl syndrome 1 minus20146718SERPING1 Serpin peptidase inhibitor clade G (C1 inhibitor) member 1 minus20102212FAM162A Family with sequence similarity 162 member A minus2005807TCTN1 Tectonic family member 1 minus20033443
AD up ALP down
1006 62 1145
(a)
140
120
100
080
060
040
020
mdash
ALPL PPARg aP2 LPL
lowastlowastlowastlowastlowastlowast
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
lowast
(b)
Figure 5 Adipocyte related genes downregulated in ALP knockdown CL1 cells (a) Venn diagram of whole gene expression analysis of ADupregulated genes compared with ALP KD downregulated genes (b) Quantitative real-time PCR of four selected common genes from Venndiagram including ALPL PPARg2 aP2 and LPL Data are presented as fold change in expression of each target gene normalized to GAPDH(119899 = three independent experiments lowast119901 lt 005 lowastlowast119901 lt 001 and lowastlowastlowast119901 lt 0001)
Stem Cells International 17
cell populations has a clinical relevance It demonstrates thatthe progenitor function and the immune modulatory rolesof hBMSC [39] are mediated by specific and distinguishablepopulations of hBMSC Thus future clinical studies employ-ing hBMSC should attempt to administer the relevant sub-population of hBMSC dependent on the experimental aim asa novel approach to improving the clinical efficiency insteadof the current use of heterogeneous hBMSC populations
Competing Interests
The authors declare that they have no competing interests
Authorsrsquo Contributions
Mona Elsafadi and Amer Mahmood contributed equally tothis work
Acknowledgments
These studies are supported by a grant from the College ofMedicine Research Centre (CMRC) Deanship of ScientificResearch King Saud University Riyadh Saudi Arabia Theauthors would like to thank Mr Ali Al-Roalle for all histechnical help They thank Natalie Twine for advice on DNAmicroarray analysis
References
[1] W Zaher L Harkness A Jafari and M Kassem ldquoAn update ofhuman mesenchymal stem cell biology and their clinical usesrdquoArchives of Toxicology vol 88 no 5 pp 1069ndash1082 2014
[2] S Post B M Abdallah J F Bentzon andM Kassem ldquoDemon-stration of the presence of independent pre-osteoblastic andpre-adipocytic cell populations in bone marrow-derived mes-enchymal stem cellsrdquo Bone vol 43 no 1 pp 32ndash39 2008
[3] K H Larsen C M Frederiksen J S Burns B M Abdallahand M Kassem ldquoIdentifying a molecular phenotype for bonemarrow stromal cells with in vivo bone-forming capacityrdquoJournal of Bone and Mineral Research vol 25 no 4 pp 796ndash808 2010
[4] A Rosenzweig ldquoCardiac cell therapymdashmixed results frommixed cellsrdquoThe New England Journal of Medicine vol 355 no12 pp 1274ndash1277 2006
[5] S A Kuznetsov P H Krebsbach K Satomura et al ldquoSingle-colony derived strains of human marrow stromal fibroblastsform bone after transplantation in vivordquo Journal of Bone andMineral Research vol 12 no 9 pp 1335ndash1347 1997
[6] M Kassem and P Bianco ldquoSkeletal stem cells in space and timerdquoCell vol 160 no 1-2 pp 17ndash19 2015
[7] A I Caplan and D Correa ldquoPDGF in bone formation andregeneration new insights into a novel mechanism involvingMSCsrdquo Journal of Orthopaedic Research vol 29 no 12 pp 1795ndash1803 2011
[8] K Stewart S Walsh J Screen et al ldquoFurther characterizationof cells expressing STRO-1 in cultures of adult human bonemarrow stromal cellsrdquo Journal of Bone and Mineral Researchvol 14 no 8 pp 1345ndash1356 1999
[9] J L Simonsen C Rosada N Serakinci et al ldquoTelomeraseexpression extends the proliferative life-span and maintains
the osteogenic potential of human bone marrow stromal cellsrdquoNature Biotechnology vol 20 no 6 pp 592ndash596 2002
[10] M Al-Nbaheen R Vishnubalaji D Ali et al ldquoHuman stromal(mesenchymal) stem cells from bone marrow adipose tissueand skin exhibit differences in molecular phenotype and differ-entiation potentialrdquo Stem Cell Reviews and Reports vol 9 no 1pp 32ndash43 2013
[11] C A Gregory W G Gunn A Peister and D J Prockop ldquoAnAlizarin red-based assay of mineralization by adherent cells inculture comparison with cetylpyridinium chloride extractionrdquoAnalytical Biochemistry vol 329 no 1 pp 77ndash84 2004
[12] AMahmood L Harkness H D Schroslashder B M Abdallah andM Kassem ldquoEnhanced differentiation of human embryonicstem cells to mesenchymal progenitors by inhibition of TGF-120573activinnodal signaling using SB-431542rdquo Journal of Bone andMineral Research vol 25 no 6 pp 1216ndash1233 2010
[13] H-J Prins A K Braat D Gawlitta et al ldquoIn vitro inductionof alkaline phosphatase levels predicts in vivo bone formingcapacity of human bone marrow stromal cellsrdquo Stem CellResearch vol 12 no 2 pp 428ndash440 2014
[14] K Stenderup J Justesen C Clausen and M Kassem ldquoAgingis associated with decreased maximal life span and acceleratedsenescence of bone marrow stromal cellsrdquo Bone vol 33 no 6pp 919ndash926 2003
[15] D G Phinney G Kopen W Righter S Webster N Tremainand D J Prockop ldquoDonor variation in the growth propertiesand osteogenic potential of human marrow stromal cellsrdquoJournal of Cellular Biochemistry vol 75 no 3 pp 424ndash436 1999
[16] C M Digirolamo D Stokes D Colter D G Phinney R Classand D J Prockop ldquoPropagation and senescence of humanmarrow stromal cells in culture a simple colony-forming assayidentifies samples with the greatest potential to propagate anddifferentiaterdquo British Journal of Haematology vol 107 no 2 pp275ndash281 1999
[17] S James J Fox F Afsari et al ldquoMultiparameter analy-sis of human bone marrow stromal cells identifies distinctimmunomodulatory and differentiation-competent subtypesrdquoStem Cell Reports vol 4 no 6 pp 1004ndash1015 2015
[18] B Peruzzi A Cappariello A Del Fattore N Rucci F DeBenedetti and A Teti ldquoc-Src and IL-6 inhibit osteoblastdifferentiation and integrate IGFBP5 signallingrdquo Nature Com-munications vol 3 article 630 2012
[19] D Fortunati S Reppe A-K Fjeldheim M Nielsen V TGautvik and K M Gautvik ldquoPeriostin is a collagen associatedbonematrix protein regulated by parathyroid hormonerdquoMatrixBiology vol 29 no 7 pp 594ndash601 2010
[20] A Oshima H Tanabe T Yan G N Lowe C A Glackin andA Kudo ldquoA novel mechanism for the regulation of osteoblastdifferentiation transcription of periostin a member of thefasciclin I family is regulated by the bHLH transcription factorTwistrdquo Journal of Cellular Biochemistry vol 86 no 4 pp 792ndash804 2002
[21] T G Kashima T Nishiyama K Shimazu et al ldquoPeriostina novel marker of intramembranous ossification is expressedin fibrous dysplasia and in c-Fos-overexpressing bone lesionsrdquoHuman Pathology vol 40 no 2 pp 226ndash237 2009
[22] CGalliM PiergianniM Piemontese et al ldquoPeriostin improvescell adhesion to implantable biomaterials and osteoblasticdifferentiation on implant titanium surfaces in a topography-dependent fashionrdquo Journal of Biomedical Materials Research -Part A vol 102 no 11 pp 3855ndash3861 2014
18 Stem Cells International
[23] D L Coutu H W Jian A Monette G-E Rivard M DBlostein and J Galipeau ldquoPeriostin amember of a novel familyof vitamin K-dependent proteins is expressed by mesenchymalstromal cellsrdquo The Journal of Biological Chemistry vol 283 no26 pp 17991ndash18001 2008
[24] T Minamizato K Sakamoto T Liu et al ldquoCCN3NOV inhibitsBMP-2-induced osteoblast differentiation by interacting withBMP andNotch signaling pathwaysrdquo Biochemical and Biophysi-cal Research Communications vol 354 no 2 pp 567ndash573 2007
[25] W Si Q Kang H H Luu et al ldquoCCN1Cyr61 is regulated bythe canonicalWnt signal and plays an important role inWnt3A-induced osteoblast differentiation of mesenchymal stem cellsrdquoMolecular and Cellular Biology vol 26 no 8 pp 2955ndash29642006
[26] Q Luo Q Kang W Si et al ldquoConnective tissue growth factor(CTGF) is regulated by Wnt and bone morphogenetic proteinssignaling in osteoblast differentiation of mesenchymal stemcellsrdquo The Journal of Biological Chemistry vol 279 no 53 pp55958ndash55968 2004
[27] Y-S Yoon N Lee and H Scadova ldquoMyocardial regenerationwith bone-marrow-derived stem cellsrdquo Biology of the Cell vol97 no 4 pp 253ndash263 2005
[28] Y S Yoon ldquoCardiac regeneration with novel bone marrow-derived multipotent stem cellsrdquo Discovery Medicine vol 5 pp204ndash208 2005
[29] Y-S Yoon A Wecker L Heyd et al ldquoClonally expanded novelmultipotent stem cells from human bone marrow regeneratemyocardium after myocardial infarctionrdquo Journal of ClinicalInvestigation vol 115 no 2 pp 326ndash338 2005
[30] GM BolandG Perkins D J Hall andR S Tuan ldquoWnt 3a pro-motes proliferation and suppresses osteogenic differentiationof adult human mesenchymal stem cellsrdquo Journal of CellularBiochemistry vol 93 no 6 pp 1210ndash1230 2004
[31] D A Cook S W Fellgett M E Pownall P J OrsquoShea and PG Genever ldquoWnt-dependent osteogenic commitment of bonemarrow stromal cells using a novel GSK3120573 inhibitorrdquo Stem CellResearch vol 12 no 2 pp 415ndash427 2014
[32] J De Boer R Siddappa C Gaspar A VanApeldoorn R Foddeand C Van Blitterswijk ldquoWnt signaling inhibits osteogenicdifferentiation of humanmesenchymal stem cellsrdquoBone vol 34no 5 pp 818ndash826 2004
[33] MAl-Toub A AlmusaM Almajed et al ldquoPleiotropic effects ofcancer cellsrsquo secreted factors on human stromal (mesenchymal)stem cellsrdquo Stem Cell Research and Therapy vol 4 article 1142013
[34] L Chen D Qanie A Jafari et al ldquoDelta-like 1fetal antigen-1(Dlk1FA1) is a novel regulator of chondrogenic cell differenti-ation via inhibition of the Akt kinase-dependent pathwayrdquoTheJournal of Biological Chemistry vol 286 no 37 pp 32140ndash321492011
[35] T Matsunobu K Torigoe M Ishikawa et al ldquoCritical roles ofthe TGF-120573 type I receptor ALK5 in perichondrial formationand function cartilage integrity and osteoblast differentiationduring growth plate developmentrdquo Developmental Biology vol332 no 2 pp 325ndash338 2009
[36] B C Nair D E Johnson R J Majeska J A Rodkey C DBennett and G A Rodan ldquoRat alkaline phosphatase II Struc-tural similarities between the osteosarcoma bone kidney andplacenta isoenzymesrdquo Archives of Biochemistry and Biophysicsvol 254 no 1 pp 28ndash34 1987
[37] P J Butterworth ldquoAlkaline phosphatase Biochemistry of mam-malian alkaline phosphatasesrdquo Cell Biochemistry and Functionvol 1 no 2 pp 66ndash70 1983
[38] A T Ali C B Penny J E Paiker et al ldquoAlkaline phosphataseis involved in the control of adipogenesis in the murinepreadipocyte cell line 3T3-L1rdquo Clinica Chimica Acta vol 354no 1-2 pp 101ndash109 2005
[39] G M Spaggiari and L Moretta ldquoCellular and molecularinteractions of mesenchymal stem cells in innate immunityrdquoImmunology and Cell Biology vol 91 no 1 pp 27ndash31 2012
Research ArticleCharacterization of Cellular and Molecular Heterogeneity ofBone Marrow Stromal Cells
Mona Elsafadi12 Muthurangan Manikandan1 Muhammad Atteya13
Jamil Amjad Hashmi4 Zafar Iqbal5 Abdullah Aldahmash16 Musaad Alfayez1
Moustapha Kassem12 and Amer Mahmood12
1Stem Cell Unit Department of Anatomy College of Medicine King Saud University Riyadh Saudi Arabia2Molecular Endocrinology Laboratory (KMEB) Department of Endocrinology University of Southern Denmark Odense Denmark3Department of Histology Faculty of Medicine Cairo University Egypt4Center for Genetics and Inherited Diseases Taibah University Al-Madina Al-Munawara Saudi Arabia5College of Applied Medical Sciences King Saud Bin Abdulaziz University for Health Sciences (KSAU-HS)National Guards Health Affairs Riyadh Saudi Arabia6Prince Naıf Health and Research Center College of Medicine King Saud University Riyadh Saudi Arabia
Correspondence should be addressed to Amer Mahmood ammahmoodksuedusa
Received 14 April 2016 Accepted 26 May 2016
Academic Editor Giorgio Mori
Copyright copy 2016 Mona Elsafadi et al This is an open access article distributed under the Creative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited
Human bone marrow-derived stromal stem cells (hBMSC) exhibit multiple functions including differentiation into skeletal cells(progenitor function) hematopoiesis support and immune regulation (nonprogenitor function)We have previously demonstratedthe presence of morphological and functional heterogeneity of hBMSC cultures In the present study we characterized in detailtwo hTERT-BMSC clonal cell populations termed here CL1 and CL2 that represent an opposing phenotype with respect tomorphology markers expression alkaline phosphatase (ALP) and CD146 and ex vivo differentiation potential CL1 differentiatedreadily to osteoblasts adipocytes and chondrocytes as shown by expression of lineage specific genes and proteins Whole genometranscriptome profiling of CL1 versus CL2 revealed enrichment in CL1 of bone- mineralization- and skeletal muscle-relatedgenes for example ALP POSTN IGFBP5 BMP4 and CXCL12 On the other hand CL2 transcriptome was enriched in immunemodulatory genes for exampleCD14 CD99NOTCH3 CXCL6 CFB andCFI Furthermore gene expressionmicroarray analysis ofosteoblast differentiated CL1 versus CL2 showed significant upregulation in CL1 of bone development and osteoblast differentiationgenes which included several homeobox genes TBX15 HOXA2 andHOXA10 and IGF1 FGFR3 BMP6 MCAM ITGA10 IGFBP5and ALP siRNA-based downregulation of the ALP gene in CL1 impaired osteoblastic and adipocytic differentiation Our studiesdemonstrate the existence of molecular and functional heterogeneity in cultured hBMSC ALP can be employed to identifyosteoblastic and adipocytic progenitor cells in the heterogeneous hBMSC cultures
1 Introduction
Human bonemarrow stromal (also known as skeletal ormes-enchymal) stem cells (hBMSC) are increasingly employedin clinical trials for enhancing tissue regeneration followinginjury [1] Typically hBMSC are isolated by their abilityto adhere to the plastic surfaces of in vitro culture platesHowever the cultured hBMSC exhibit morphological hetero-geneity suggesting the presence of functional heterogeneity[2 3] It has also been suggested that the use of heterogeneous
cell populations in clinical trials of hBMSC-based therapiescaused variability in the observed treatment effects [4] Thusfor the efficient use of hBMSC in therapy better cellular andmolecular characterization of hBMSC is required [1 4]
There exist no specific markers that define the hBMSCphenotype The plastic-adherent hBMSC are defined by thepresence of surface expression of some CD surface markerswith variable sensitivity and specificity [1] Single cell clonalanalysis revealed that only 25 of the cells are true stemcells based on their ability to differentiate into osteoblasts
adipocytes and chondrocytes (trilineage differentiation) andto form heterotopic bone and bone marrow organ whenimplanted in vivo subcutaneously in immune deficient mice[5] The identity of the remaining cells is not clarified butthey may represent lineage-committed cells [3] Therefore itis plausible that functional heterogeneity exists in culturedhBMSC reflecting the in vivo functional and developmentalheterogeneity of hBMSC [6]
In addition to their ability to differentiate into skeletaltissue cells (known as progenitor function) hBMSC possessimmunomodulatory characteristics (known as nonprogen-itor functions) [7] It is not clear whether these differentfunctions are mediated by a number of independent sub-populations within the hBMSC [2] Only a few studies havetried to identify the subpopulation within cultured hBMSCbased on surface markers for example STRO1 and alkalinephosphatase (ALP) but limited molecular phenotyping hasbeen conducted [8]
We have previously demonstrated the presence of mor-phological and functional heterogeneity of clones isolatedfrom telomerized hMSC (hMSC-TERT) cell line [3]The aimof the present study was therefore to further study in detailthe heterogeneity of cultured hBMSC as demonstrated bytwo clonal cell lines with opposite cellular and functionalphenotypeWe also employed theDNAmicroarrays to definetheir molecular signature and signaling pathways associatedwith their functional phenotype
2 Experimental Procedures21 Cell Culture As a model for hBMSC we employedimmortalized hBMSC-TERT cell line that is created fromnormal human BMSC by overexpression of human telom-erase reverse transcriptase gene (hTERT) [9] The hBMSC-TERT cells have been extensively characterized and theyexhibit similar cellular and molecular phenotype to primaryMSC [10] CL1 and CL2 cells are clonal cell populationsof hBMSC-TERT identified in long term culture (passagenumbers 15ndash25) of hBMSC-TERT and were chosen based ontheir distinct and different morphologies Cells were culturedin Dulbeccorsquos Modified Eagle Medium (DMEM) supple-mented with D-glucose 4500mgL 4mM L-glutamine and110mgL sodium pyruvate 10 Fetal Bovine Serum (FBS) 1xpenicillin-streptomycin (Pen-strep) and nonessential aminoacids (all purchased from Gibco-Invitrogen USA) For somecontrol experiments primary bone marrow derived MSC(phBMSC) were employed Sixty milliliters of bone marrowwas aspirated from the iliac crest bone of consenting healthydonors This procedure was approved by the King KhalidUniversity Hospital-King Saud University ethics committeephBMSCwere isolated from bonemarrowmononuclear cellsby plastic adherence as described previously [9]
22 Cell Proliferation Cell proliferation rate was determinedby counting cell number and calculating population doubling(PD) rate The cells were cultured in 25 cm2 tissue culturePetri dish at cell density 05 times 106 cells (28000 cellscm2) Atconfluence the cells were trypsinized and counted manuallyby hemocytometer At each passage population doubling wasdetermined by the following formula log119873 log 2 where 119873
is the number of cells at confluence divided by the initialcell number Cumulative PD level is the sum of populationdoublings and PD rate is PDtime in culture
23 Flow Cytometry Cells were trypsinized to a singlecell suspension were recovered by centrifugation at 200 gfor 5min washed twice in ice-cold PBS supplementedwith 2 BSA and resuspended at a concentration of105 cellsantibody After incubation with the preconjugatedantibodies or matched isotype controls for 30min on icein the dark cells were washed with PBS resuspended in500120583L of PBS and analyzed in the BD FACSCalibur flowcytometer (BD Biosciences) Living cells were gated in adot plot of forward versus side scatter signals acquired ona linear scale At least 10000 gated events were acquiredon a log fluorescence scale Positive staining was distinct asthe emission of a fluorescence signal that surpassed levelsachieved by gt99 of control cell population stained withcorresponding isotype antibodies The ratios of fluorescencesignals versus scatter signals were calculated and histogramswere generated using the software Cell Quest Pro SoftwareVersion 33 (BD Biosciences) The following antibodies wereused all from BD Biosciences FITC-PE-APC-Mouse IgG1kisotype control APC-Mouse Anti-Human CD44 (559942)FITC-Mouse Anti-HumanCD63 (557305) PE-Mouse Anti-Human CD73 (550257) PE-Mouse Anti-Human CD105(560839) PE-Mouse Anti-Human CD146 (550315) PE-Mouse Anti-HumanCD166 (560903) and Alexa Fluor 488Mouse Anti-Human Alkaline Phosphatase (561495)
24 Electron Microscopy After trypsinizing the hBMSC cellsfrom the flasks or 6-well plates were collected the sampleswere washed with PBS and the pellets were resuspendeddirectly in 25 glutaraldehyde in 01M phosphate buffer (pH72) and kept at 4∘C for 4 hr First the cells were washed with01Mphosphate buffer (pH72) and transferred to 1osmiumtetroxide (OsO4) solution in 01M phosphate buffer (pH 72)for two hr The cells were dehydrated in ascending grades ofethanolThe cells were then resuspended in acetone and werealiquoted into BEEM embedding capsules and infiltratedwith acetone resin mixture followed by embedding in apure resin mixture for two hr Semithin sections (05 120583mthickness) and ultrathin sections (70 nm thickness) wereprepared examined and photographed under a transmissionelectron microscope (TEM) (Jeol 1010 Jeol Tokyo Japan)
25 In Vitro Osteoblast Differentiation Cells were grownin standard DMEM growth medium in 6-well plates at20000 cellcm2 When 70ndash80 confluence was achieved testcells were cultured inDMEM supplementedwith osteoblasticinduction mixture (referred to as OS) containing 10 FBS1 Pen-strep 50120583gmL L-ascorbic acid (Wako ChemicalsNeuss Germany) 10mM 120573-glycerophosphate (Sigma) and10 nM calcitriol (112057225-dihydroxyvitamin D3 Sigma) and10 nM dexamethasone (Sigma) noninduced cells (referredto as Cont) were cultured in normal growth media for thesame duration as induced The media were replaced threetimes per week Cells cultured in standard culture mediumwere considered as control At day 14 of differentiation
Stem Cells International 3
mineralized nodules became apparent and were stained withAlizarin Red S and ALP
26 In Vitro Adipocyte Differentiation Cells were grown instandard DMEM growth medium in 6-well plates at 03times 106 cellsmL At 90ndash100 confluence cells were culturedin DMEM supplemented with adipogenic (Adip) inductionmixture containing 10 FBS 10 Horse Serum (Sigma)1 Pen-strep 100 nM dexamethasone 045mM isobutylmethylxanthine (Sigma) 3120583gmL insulin (Sigma) and 1 120583MRosiglitazone (Novo Nordisk Bagsvaerd Denmark) Themedia were replaced three times per week Cells cultured instandard culture medium were considered as control Fromday 3 of differentiation small lipid droplets became visibleand at day 7 they were stained with Oil Red-O and Nile red
27 In Vitro Chondrogenic Differentiation Both CL1 and CL2cells were trypsinized and counted around 1 times 106 cellstaken in each 15mL conical tube centrifuged at 400timesg for5 minutes For chondrocyte differentiation pellet culturesystem used chondrocyte induction was done in mediacontaining advanced DMEMF12 supplemented with 1 ITSPremix Tissue Culture Supplement 100 nM dexamethasoneGlutamax and 10 ngmL transforming growth factor-beta-3(TGF120573-3) Cells were maintained in chondrocyte differentia-tion media for 21 days and changed every two days
28 Cytochemical Staining281 Alkaline Phosphatase (ALP) Staining CL1 and CL2 cellswere stained before OS differentiation for the basal ALPexpression and after OB differentiation at day 7 of inductionCells cultured in 6-well plates were washed in PBS minusminus(-Ca -Mg) andfixed in acetonecitrate buffer 10mMat pH42for 5min at room temperature The NaphtholFast Red stain[02mgmL Naphthol AS-TR phosphate substrate (Sigma)][0417mgmL of Fast Red (Sigma)] was added for one hourat room temperature
Histological tissue blocks were sectioned at 4 micronsImmunohistochemical staining was performed on CL1 andCL2 chondrocyte 3D pellets using DAKO EnVision andPowerVision according to the manufacturerrsquos instructions(DAKO Glostrup Denmark) Briefly paraffin sections wereincubated for 1 hour at room temperature with primaryantibodies diluted in ChemMate (DAKO) (HumanAnti-Col-10 and Human Anti-Col-2 ABI) Sections were washed sub-sequently in Tris-buffered saline (TBS 005M pH 74) incu-bated for 30 minutes with secondary anti-mouse IgHRP-conjugated polymers (K4001 En Visionthorn DAKO) and visu-alizedwith 330-diaminobenzidine tetrahydrochloride (DABS3000 DAKO) or with 3-amino-9-ethylcarbazole (AECDAKO) according to manufacturerrsquos instruction Controlswere performed with nonimmune immunoglobulins of thesame isotype as the primary antibodies (negative controls)and processed under identical conditions Alcian blue stain-ing was used to detect chondrocytes Sections of paraffin-embedded implants were stained with Alcian blue (Sigma)solution pH 25 at this pH all the glycoproteins (neutral andacidic) will be stained blue
29 Alizarin Red S Staining for Mineralized Matrix Seven-day-old OS differentiated cells in 6-well plates were used forAlizarin Red S staining The cell layer was washed with PBSand then fixed with 70 ice-cold ethanol for 1 hr at minus20∘CAfter removing the ethanol the cell layer was rinsed withdistilled water and stained with 40 nM AR-S (Sigma) pH 42for 10 minutes at room temperature Excess dye was washedoff with water followed by a wash with PBS for few minutesto minimize nonspecific AR-S stain
For quantifying the Alizarin Red S staining the air-driedplates the Alizarin Red S dye was eluted in 800 120583L of aceticacid incubated in each well for 30 minutes at room temper-ature as described [11] and measured in spectrophotometer(BioTek Epoch) at 405 nm
210 Quantitative ALP Activity To quantify ALP activity inCL1 and CL2 hBMSC before and after OS differentiationwe used the BioVision ALP activity colorimetric assay kit(BioVision Inc CA USA) with some modifications Cellswere cultured in 96-well plates under normal conditionsthen on day of analysis wells were rinsed once with PBSand were fixed using 37 formaldehyde in 90 ethanol for30 seconds at room temperature Subsequently fixative wasremoved and 50120583L of pNPP solution was added to each welland incubated for 1 hour in the dark at room temperatureThe reaction was subsequently stopped by adding 20 120583L stopsolution and gently shaking the plate ODwas thenmeasuredat 405 nm
211 Oil Red-O Staining for Lipid Droplets CL1 and CL2 cellsdifferentiated to adipocytes with Adip induction media atday 7 were used Accumulated cytoplasmic lipid dropletswere visualized by staining with Oil Red-O After washingcells grown in 6-well plates with PBS the cells were fixedin 4 formaldehyde for 10min at room temperature andthen rinsed once with 3 isopropanol and stained for 1 hr atroom temperature with filtered Oil Red-O staining solution(prepared by dissolving 05 g Oil Red-O powder in 60isopropanol) To quantify staining of fat droplets Oil Red-O was used as a stain Oil Red-O was eluted by adding 100isopropanol to each well and color changes were measuredby spectrophotometer at 510 nm (BioTek SpectrophotometerEpoch)
212 Nile Red Fluorescence Determination and Quantificationof Adipogenesis A stock solution of Nile red (1mgmL) inDMSO was prepared and stored at minus20∘C protected fromlight Staining was performed on unfixed cells Culturedundifferentiated and day 7 adipocyte differentiated cells weregrown in Corning polystyrene flat bottom 96-well TC-treated black microplates (Corning NY USA) were washedonce with PBS The dye was then added directly to the cells(5 120583gmL in PBS) and the preparation was incubated for10min at room temperature and then washed twice with PBSFluorescent signal was measured using SpectraMaxM5 fluo-rescence spectrophotometer plate reader (Molecular DevicesCo Sunnyvale CA USA) using bottom well-scan modewhere nine readings were taken per well using Ex (485 nm)and Em (572 nm) spectra
4 Stem Cells International
213 Quantitative Real-Time PCR (qRT-PCR) Analysis TotalRNA was extracted using MagNA pure compact RNA iso-lation kit (Roche Applied Science Germany Cat number04802993001) in an automatedMagNA pure compact system(Roche Germany) as recommended by the manufacturerThe total RNAwas quantified by Nanodrop spectrophotome-ter (Nanodrop 2000 Thermo Scientific USA) Complemen-tary DNA (cDNA) was synthesized from 1 120583g of the RNAsamples usingHighCapacity cDNAReverse Transcription kit(Applied Biosystems USA) using Labnet Multigene thermo-cycler according to the manufacturerrsquos instructions Relativelevels of mRNA were determined from cDNA by real-timePCR (Applied Biosystems-Real-Time PCRDetection System)with Power SYBR Green PCR kit (Applied BiosystemsUK) according to the manufacturerrsquos instructions Followingnormalization to the reference gene GAPDH quantificationof gene expression was carried out using a comparative Ctmethod where ΔCt is the difference between the CT valuesof the target and the reference gene and fold induction wasperformed from the control (Cont) for the same time pointPrimers (Supplementary Table 1 in Supplementary Materialavailable online at httpdxdoiorg10115520169378081)were obtained from Applied Biosystems (USA) as TAQMANprimers or previously published primers were used (seeSupplementary Table 1)
214 DNAMicroarray Global Gene Expression Analysis Fourhundred ng of total RNA was used as input for gener-ating biotin-labeled cRNA (Ambion Austin TX UnitedStates) cRNA samples were then hybridized onto Illuminahuman-8 BeadChips version 3 Hybridization washing Cy3-streptavidin staining and scanning were performed on theIllumina BeadStation 500 platform (Illumina San DiegoCA USA) according to the manufacturerrsquos instructions andeverything was done in triplicate Expression data analysiswas carried out using the Partek genomic suite softwareRaw data were background-subtracted normalized usingthe ldquorank invariantrdquo algorithm and filtered for significantexpression on the basis of negative control beads Geneswere considered significantly expressed with detection 119901values le 001 Differential expression analysis was performedwith the Illumina custom method using freshly isolatedprimary hBMSC (used at passage 3) as a reference controlThe following parameters were set to identify statisticalsignificance differential 119901 values le 001 fold change ratiogt 15 Pathway analysis was performed using DAVID Bioin-formatics Resources 67 (httpdavidabccncifcrfgov) andGeneSpring GX software (Agilent Technologies) Pathwayanalysis for CL1 OS D14 versus CL2 OS D14 was conductedusing the Single Experiment Pathway analysis feature inGeneSpring 120 (Agilent Technologies)
215 Small Interfering (si)RNATransfection For transfectionhBMSC in logarithmic growth phase were transfected withSilencer Select Predesigned ALP siRNA (25 nM) (Assay IDs1298 and Cat number 4390824) (Ambion The RNA Com-pany USA) using Lipofectamine RNAiMAX Reagent (Invit-rogen CA USA) plus serum-free Opti-MEMI mediumunder the conditions described by the manufacturer At
day 3 of transfection the cells were induced for osteogenicdifferentiation for an additional 7 days ALP staining wasused as a control for the siRNA transfection efficiency andtimeline
216 Statistical Analysis All of the results were presentedas the mean and standard deviation (SD) of at least 3independent experiments with 3ndash5 technical repeats in eachexperiment Studentrsquos 119905-test (two-tailed) was used for testingdifferences between groups 119901 value lt005 was consideredstatistically significant
3 Results
31 Comparison between CL1 and CL2 Differences in Mor-phology Proliferation and Marker Expression Profile Weisolated two distinct clonal cell populations of hBMSC-TERThBMSC-CL1 and hBMSC-CL2 (for easiness will be termedhereafter CL1 and CL2) based on differences in cell morphol-ogy (Figure 1(a)) CL1 cells had cuboidalmorphologywhereasCL2 cells have spindle-shaped fibroblast-like morphologyCL1 cells had higher proliferation rate compared to CL2(Figure 1(b)) mean PD rates of CL1 and CL2 were 0714 and0429 PDday respectively (Figure 1(b)) Both CL1 and CL2expressed surface marker profiles characteristics of hBMSC(gt90) CD44+ CD63+ CD73+ CD105+ and CD166+(Figure 1(c)) However CL1 cells showed higher expressionof CD146 (927 versus 12) and ALP (98 versus 0)compared to CL2 (Figure 1(d)) TEM revealed the presenceof abundant pseudopodia in CL1 indicating high motility(Figure 1(e)(A)) as well as well-developed mitochondriaand rough endoplasmic reticulum (rER) suggesting highmetabolic activity CL2 cells contained abundant phagocyticvacuole (pv)microvilli (mi) and lysosomes (ly) (Figure 1(e))
We performed quantitative real-time PCR (RT-PCR)for genes expressed in mesodermal progenitor cells [12]CL1 expressed higher levels of BMP4 MIXL1 WNT3a andTWIST compared to CL2 (Figure 1(f) 119901 lt 001) In contrastCL2 expressed higher levels of Kinase Insert Domain Recep-tor (Type III Receptor Tyrosine Kinase) (KDR) expressed inendothelial cells and smoothmusclemyosin heavy chain gene(smMHC) expressed in smooth muscle cells (Figure 1(f))
32 CL1 Cells Exhibit Enhanced Osteoblast DifferentiationFollowing osteoblast (OB) differentiation induction ALPstaining andALP enzymatic activity were significantly higherin CL1 compared to CL2 cells (Figure 2(a) 119901 lt 001)Similarly Alizarin Red staining and quantitation of formedmineralized matrix were more pronounced in CL1 cells (Fig-ure 2(b) 119901 lt 001) In addition CL1 cells expressed higherlevels of osteoblastic genes ALP RUNX2 and osteopontin(OPN) (Figure 2(c) upper panel) compared to CL2 cells
Global gene expression microarray analysis of OB dif-ferentiated cells at day 14 showed around 1060 genes signif-icantly upregulated more than 2-fold (119901 lt 001) in CL1Among the upregulated genes 80 genes were annotated tobone development and osteoblast differentiation (Table 1)The highest upregulated genes included paired-like home-odomain 2 (PITX2) Insulin-like growth factor 1 (IGF1) and
Stem Cells International 5
CL1
CL2
(a)
Cum
ulat
ive P
D
Cumulative days
Growth curve
0 10 20 30 40
CL1 CL2
CL1 CL20
5
10
15
20
25
0
01
02
03
04
05
06
07
08
(PD
day
)
(b)
CL2
103
102
101
100
103
102
101
100
103
102
101
100
103
102
101
100
103
102
101
100
103
101
100
103
102
102
101
100
103
102
101
100
103
102
101
100
103
102
101
100
CL1
CD44 CD63 CD73 CD166 CD105
T3 CD44 T3 CD63 T3 CD73 T3 CD166 T3 CD105
0
20
40
60
80
Cou
nt
0
20
40
60
80
100
120
Cou
ntC
ount
Cou
nt
Cou
nt
Cou
nt
Cou
nt0
20
40
60
80
Cou
nt
0
20
40
60
80
100
Cou
nt
0
20
40
60
80
Cou
nt
80
60
40
20
0
100
50
0
120
100
80
60
40
20
0
120
100
80
60
40
20
0
100
80
60
40
20
0
(c)
ISO CD146 ALP
ISO CD146 ALP
FL2-
H
FL2-
H
FL2-
HPEPEPECL
2CL
1
FL1-H FL1-H FL1-H
FITCFITCFITC
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
(d)
Figure 1 Continued
6 Stem Cells International
(A) (B)
(e)
Fold
indu
ctio
n
CL1CL2
lowast
lowastlowast lowast
BMP4 MIXL1 WNT3a TWIST
lowastlowast
KDR smMHCminus01
01
03
05
07
09
11
13
15
mdash
050
100
150
200
250
300
CL1CL2
(f)
Figure 1 Cellular phenotype of two clonal populations of humanbonemarrow stromal stem cells CL1 andCL2 (a) Cellmorphology CL1 cellsshowed large cuboidal morphology while CL2 cells had spindle-shaped fibroblast-like morphology (b) Growth curves showing populationdoubling (PD) rate during long term culture (c) Flow cytometry analysis (FACS) shows expression of CD44 CD63 CD73 CD105 andCD166in CL1 and CL2 cells Matched isotype control was used for gating (d) Flow cytometry analysis presented as dot blot of CD146 and alkalinephosphatase (ALP) cell surface proteins (e) Transmission electron microscope (TEM) (A) CL1 (1200x) (B) CL2 (1500x) n nucleus nunucleolus rER rough endoplasmic reticulum ly lysosomes pv phagocytic vacuole and rer reticular stalk of rER (f) Gene expressionanalysis using RT-PCR for a group of mesodermal and stromal genes Gene expression was normalized to GAPDH and presented as foldchange Data is shown as mean plusmn SD of three independent experiments lowast119901 lt 005 119901 lt 0001
collagen type V alpha 3 (COL5A3) osteomodulin (OMD)and T-box 15 (TBX15) (Table 1) Furthermore several knownosteoblast-related genes were upregulated in CL1 cells suchas bone morphogenetic protein 6 (BMP6) fibroblast growthfactor receptor 3 (FGFR3) insulin-like growth factor bindingprotein 5 (IGFBP5) and vitamin D (125-dihydroxyvitaminD3) receptor (VDR) (Table 1) On the other hand 1200 geneswere upregulated in CL2 cells 255 genes were annotated toimmunity and immune response and defense This categoryincluded genes from complement system chemokine (C-Cmotif) ligands interferon family chemokine (C-X-C motif)ligands and receptor major histocompatibility complex classII molecules interleukins and tumor necrosis factor receptorsuperfamily (Table 2 and Supplementary Table 3)
33 CL1 Cells Exhibit Enhanced Adipocyte DifferentiationWe observed significant differences between CL1 and CL2in their response to adipocytic differentiation inductionCL1 differentiated readily to adipocytes compared to CL2(Figure 2(d) lower panel) evidenced by higher levels ofadipocyticmarkers gene expression LPL (lipoprotein lipase)and adiponectin as well as formation of mature lipid filledadipocytes visualized by Oil Red-O staining and quantitativeNile red staining (Figure 2(d))
34 CL1 Cells Differentiate to Chondrocytic Lineage Inpellet cultures CL1 cells formed 3D pellets containingproteoglycan-secreting chondrocytes which stained positivewith Alcian blue Limited chondrocyte differentiation wasvisible in cell pellets of CL2 cells The differentiated chon-drocytes in CL1 pellets expressed higher levels of collagen Xand collagen II which was overlapping the Alcian blue stain(Figure 2(e))
35 Molecular Signature of CL1 and CL2 Cells To define themolecular signature and molecular differences between CL1
and CL2 we compared the basal gene expression patternof CL1 and CL2 cells using DNA microarrays The PCAanalysis showed a clear separation between CL1 and CL2(Supplementary Figure 1) Comparison between CL1 withCL2 showed that 915 geneswere differentially expressed in thetwo cell lines (gt2-fold119901 lt 001) 462 genes were upregulatedand 452 were downregulated in CL1 versus CL2 The mostrelevant genes that were upregulated in CL1 are listed inTable 3(a) Among these 35 highly expressed genes in CL1 thefollowing 11 genes were present in skeletal and muscular sys-tem development and function FOLR3 CCL3L1 SERPINB2POSTN IGFBP5 CCL3 NOV ALP TNFRSF11B ACTG2and CDH11 (Table 3(a)) Functional annotation of the upreg-ulated genes in CL1 using the Ingenuity Pathway Analysis(IPA) revealed enrichment in the following categories ldquotissuedevelopmentrdquo ldquoskeletal and muscular system developmentand functionrdquo and ldquoorganismal developmentrdquo (Table 3(b))Furthermore the DAVID annotation tool was employed toassess the functional relationships of the upregulated genesin CL1 showing enrichment in ontologies ldquoskeletal andmuscular system development and functionrdquo that includedbone size osteoblast differentiation bone mineralizationand bone mineral density (Table 3(c)) CL1 exhibited upreg-ulation of WNT pathway ligands WNT5B (2-fold) andLRP5 (2-fold) (Table 1) Also ALP was among the highlyexpressed genes together with POSTN IGFBP5 SPP1 IL-6 and DKK1 (Tables 3(a) and 1) These genes are knownto play an important role in osteoblast differentiation andbone formation For CL2 inhibitors of WNT pathway wereupregulated and included SFRP1 (11-fold) DKK2 (32-fold)FGF2 (31-fold) and GBP2 (24-fold) Functional annotationof the upregulated genes in CL2 revealed enrichment in thefollowing categories ldquodevelopmental processrdquo ldquomulticellularorganismal processrdquo ldquobiological adhesionrdquo and ldquoimmunesystem processrdquo (Supplementary Table 2A) In-depth analysis
Stem Cells International 7
CL1
CL2
ALPCont D 14
Cont OB I
ALP activity
mdash100
200
300
400
500
600
Fold
chan
ge
CL1CL2
(a)
CL1
CL2
ALZCont D 14
Fold
chan
ge
25
20
15
10
5
0
Cont d14 d21
CL1CL2
ALZ quantification
(b)
CL1CL2
CL1CL2
LPL ADIPOQ
Cont ContADIP D14 ADIP D14
000
100
200
300
400
500
600
700
0
1000
2000
3000
4000
5000
CL1CL2
CL1CL2
CL1CL2
RUNX2 ALP OPN
Cont Cont ContOS D7 OS D7 OS D7
00
05
10
15
20
25
30
00
05
10
15
20
25
30
35
00
05
10
15
20
25
30
35
(c)
CL1CL2
CL1
CL2
Control Adip d7 Adip d14
d7 d140
005
01
015
02
025
03
035
04
045
Nile
red
fluor
esce
nt in
tens
ity
(d)
CL1
CL2
Alcian blue Col X Col II
(e)
Figure 2 In vitro osteoblastic and adipocytic differentiation of two clonal populations of human bone marrow stromal stem cells CL1 andCL2 cells Both cell lines were induced for osteoblast differentiation using standard protocol described in the Methods (a) ALP staining atday 14 in control noninduced (Cont) and osteoblast induced cells (D14) Right panel shows ALP activity (119899 = 3 independent experiments119901 lt 0001) (b) Mineralized matrix formation visualized by Alizarin Red S staining Right panel shows Alizarin Red quantification at day 14and 21 after osteoblast differentiation (119899 = 3 independent experiments 119901 lt 0001) (c) Quantitative RT-PCR of osteoblastic and adipocytegene markers in CL1 and CL2 during osteoblast (upper panel) and adipocyte (lower panel) differentiation ALP = alkaline phosphataseOPN = osteopontin LPL = lipoprotein lipase and ADIPOQ = adiponectin Data are presented as fold change in expression of each targetgene normalized to GAPDH (119899 = 3 independent experiments 119901 lt 005
119901 lt 0001) (d) CL1 and CL2 lines were induced for adipocytedifferentiation using standard protocol described in the Methods Adipocyte formation was visualized at day 7 (Adip d7) and day 14 (Adipd14) by Oil Red-O staining Lower panel presents quantification of Nile red staining (119899 = 3 independent experiments 119901 lt 0001) (e) CL1and CL2 lines were induced for chondrocyte differentiation using 21-day pellet culture method as described in the Methods The pellets werestained with Alcian blue collagen 10 (Col X) and collagen 2 (Col II) (original magnification 5x)
8 Stem Cells International
Table 1 Microarray data analysis showing genes related to bone development and osteoblast differentiation upregulated in CL1 versus CL2cells
Probe ID Genbank accession Gene name Gene symbol FCA 23 P167367 NM 153426 Paired-like homeodomain 2 PITX2 30884
of the biological processes revealed several immune-relatedpathways ldquoMAPKKK cascaderdquo ldquoimmunity and defenserdquoldquosignal transductionrdquo ldquoextracellular matrix protein-mediatedsignalingrdquo and ldquointerferon-mediated immunityrdquo among oth-ers that were upregulated (Supplementary Table 2B) Also40 genes related to immune system related factors wereidentified as significantly enriched in CL2 compared to CL1cells (Supplementary Table 2C)We chose the following genesfor validation of the microarray results NOV IGFBP5 ALPTAGLN and CDH11 as they were highly expressed in CL1RT-PCR analysis confirmed themicroarray results (Figure 3)
Furthermore we compared the molecular phenotype ofCL1 and CL2 cells with that of phBMSCWe found that morethan 80 of the genes expressed in CL1 and 90 in CL2 cellswere commonwith primary phBMSC (Supplementary Figure2) suggesting that CL1 and CL2 molecular phenotype existwithin the heterogeneous population of phBMSC cultures
36 ALP Knockdown Impairs Differentiation of CL1 CellsSince ALP has been suggested as a marker for hBMSCprogenitor cell lineage commitment [13] and was highlyupregulated in CL1 cells we tested its biological role in
CL1 cells ALP siRNA transfection decreased ALP proteinlevel ALP activity and mRNA gene expression compared tocontrol cells transfected with control siRNA (119901 lt 001) andthis inhibition was detectable up to day 7 days after osteoblastdifferentiation induction (Figures 4(A) and 4(B)) At day 14 ofdifferentiation mineralization ability of CL1 was significantlyimpaired (Figure 4(C)) In addition we found that thenumber of mature adipocyte formations was significantlyreduced to more than 75 (119901 lt 001) (Figure 4(D))
To identify relevant adipocyte differentiation associatedgenes that were targeted by ALP deficiency we compared thedownregulated genes of ALP deficient CL1 with the upreg-ulated genes identified during adipocytic differentiation ofCL1 We identified 62 genes that were common (Figure 5(a)Table 4) and among these genes were genes related tometabolism (primarily lipid and carbohydrate) and transportincluding CYB5B CHST1 TAP1 ATP8A1 LRP8 PLCD1and FABP5 (Table 4) We further performed quantitativereal-time PCR of ALP deficient CL1 cells during adipocytedifferentiation The following adipocyte-associated geneswere downregulated PPAR1205742 LPL and aP2 (Figure 5(b))
12 Stem Cells International
Table 3 Whole genome microarray analysis of two clonal popu-lations of human bone marrow stromal stem cells CL1 and CL2cells (a) Top 35 highly upregulated genes in CL1 cells versus CL2cells (b) Ingenuity Pathway Analysis (IPA) showing the differentphysiological system development and function genes found in eachcategory and the corresponding 119901 value (c) Upper part IPA analysisshowing the categories for skeletal and muscular system developmentand function upregulated in CL1 cells and lower part showing genesupregulated in CL2 cells NOV = nephroblastoma overexpressedIGFBP5 = insulin-like growth factor binding protein 5 ALP TAGLN= transgelin and CDH11 = OB-cadherin (osteoblast)
Name 119901 value moleculesPhysiological system development and function
Organismal development 705119864 minus 10ndash143119864 minus 03 181Embryonic development 100119864 minus 09ndash143119864 minus 03 154Organ development 100119864 minus 09ndash143119864 minus 03 145
(b) Continued
Name 119901 value moleculesSkeletal and muscularsystem development andfunction
100119864 minus 09ndash102119864 minus 03 123
Tissue development 100119864 minus 09ndash143119864 minus 03 236
(c)
Functions annotation 119901 value moleculesSkeletal and muscular system development and
function upregulated in CL1 cellsSize of bone 143119864 minus 06 24Differentiation ofosteoblasts 384119864 minus 06 25
Mineralization ofbone 493119864 minus 06 19
Bone mineral density 365119864 minus 05 19Skeletal and muscular system development and
function upregulated in CL2 cellsDevelopment ofmuscle 100119864 minus 09 44
Proliferation ofmuscle cells 229119864 minus 06 35
Remodeling of bone 363119864 minus 06 21Resorption of bone 394119864 minus 06 19
confirming impairment of adipocytic differentiation of ALPdeficient CL1 cells
4 Discussion
We extensively studied two cell populations within culturedhBMSC that were identified based on differences in mor-phology Cellular and molecular studies revealed differencesin growth differentiation capacity and molecular signatureOur data support the notion of the presence of cellular andfunctional heterogeneity among cultured hBMSC
Cellular heterogeneity of cultured hBMSC is recognizedin an increasing number of reports Several extrinsic andintrinsic factors may contribute to the observed hBMSCheterogeneity Extrinsic factors include donor-to-donor vari-ations in the number and quantity of initiating cells whichresult in differences in cell growth rate and differentia-tion capacity [14 15] Intrinsic factors have been examinedemploying single cell clonal analysis and revealed variationsin differentiation potential among individual colonies withinhBMSC cultures ranging from the presence of cells withtrilineage (osteoblast adipocytes and chondrocyte) potencyto cells with null potency [16] Also variations in the abilityof clonal cells to form heterotopic bone when implanted invivo have been reported [5] Our study corroborates thesefindings and provides more detailed cellular and molecularphenotyping of two examples of cell populations that existwithin the heterogeneous hBMSC cultures [17]
Determining the molecular signature of CL1 and CL2using whole genomemicroarray analysis showed enrichmentof lineage-commitment associated genes inCL1 For exampleinsulin-like growth factor 5 (IGFBP5) and interleukin 6 (IL6)
Stem Cells International 13
NOV0040
0030
0020
0010
mdash
IGFBP50015
0010
0005
mdash
ALP0100
0080
0060
0040
0020
mdash
TAGLN0800
0600
0400
0200
CDH11
0060
0040
0020
mdashmdash
LRP30015
0010
0005
mdash
CL1 CL2 CL1 CL2 CL1 CL2
CL1 CL2CL1 CL2CL1 CL2
Figure 3 Validation of whole genome microarray analysis of two clonal populations of human bone marrow stromal stem cells CL1 andCL2 cells Quantitative real-time PCR for highly expressed genes in CL1 cells NOV = nephroblastoma overexpressed IGFBP5 = insulin-likegrowth factor binding protein 5 ALP TAGLN = transgelin and CDH11 = OB-cadherin (osteoblast) Data are presented as fold change inexpression of each target gene normalized to GAPDH (119899 = three independent experiments 119901 lt 0001) (see also Table 3)
were 147- and 33-fold upregulated in CL1 cells respectivelyBoth factors are expressed in osteoprogenitor cells andimportant for osteoblast maturation [18] We also observedthat periostin (POSTN) gene was highly upregulated inCL1 cells (156-fold) POSTN is a 90 kDa secreted proteinoriginally identified in murine osteoblast-like cells and isupregulated by PTH [19] Several studies employing murineand human cells have revealed important role of POSTNin osteoblast differentiation and during development inintramembranous ossification [20ndash23] Another factor iden-tified in CL1 cells is nephroblastoma overexpression (NOV)which is a member of the Cyr 61 connective tissue growthfactor (CNN) family The CCN family of proteins promotesosteoblast differentiation through interaction with integrinsWNT BMP and NOTCH signaling pathways [24ndash26] Inaddition a large number of signaling molecules known tobe regulators of hBMSC lineage specific differentiation forexample insulin-like growth factors [27ndash29] WNT [30ndash32]and MAPK [33ndash35] were enriched in CL1 cells In contrastCL2 expressed high levels of immune-related genes whichmay explain the poor differentiation response to osteoblastor adipocyte lineage In a recent study the authors usedtelomerized hBMSC and showed clearly a clonal populationthat had very low in vitro and in vivo differentiation abilityhowever they had enhanced immune-related features includ-ing high IL7 expression These nullipotent cells expressedCD317 which was associated with remarkably high basal levelexpression of factors with a proinflammatory and antiviralfunction [17] We observed that this molecular phenotypewas associated with distinct ultrastructural characteristicsof the cells In particular CL2 had abundant phagocyticvacuole microvilli and lysosomes features reminiscent of
ultrastructure of immune-regulatory cells Our data thussupport the increasingly recognized feature that hBMSCexhibit immune modulatory functions and a part of theinnate immune response [17]
We observed that ALP protein expression and enzymaticactivity were significantly different between CL1 and CL2cell lines and were thus a potential marker that distinguishesdifferent cell populations with progenitor functions (CL1)from cells with nonprogenitor functions (CL2) ALP isexpressed in a wide variety of tissues including kidneysbone and liver [36 37] but tissue-nonspecific ALP (ALPL)is considered a commitment marker for osteoblastic lineage[13 38] However in a recent study the authors examined thedifferentiation potential of a number of hMSC clones in vitroand in vivo and reported that the hMSC clones with highlevels of ALP expression were committed to trilineage differ-entiation [13] Our data corroborate and extend these findingsby reporting the effects of siRNA-mediated inhibition ofALP that resulted in an impaired hBMSC differentiation notonly to osteoblasts but also to adipocytes Also our resultscorroborate earlier studies that demonstrated in human bonebiopsies the presence of ALP expression in bone marrowadipocytic cells [13 17] All these data suggest that ALP isa ldquostemnessrdquo marker of hBMSC and not just an indicator ofosteoblastic lineage commitment
While CL1 and CL2 were isolated from telomerizedhMSC cell line they are relevant to normal human physi-ology We observed that the molecular phenotypes of CL1and CL2 were contained within the molecular signatureof primary hBMSC suggesting that CL1 and CL2 repre-sent cell populations within the heterogeneous cultures ofhBMSCWe have also previously reported that the molecular
14 Stem Cells International
ALZ
stai
ning
Control siRNA ALP siRNA
ALP
Beta
-act
in
Con
trol
ALP
siRN
A d3
ALP
siRN
A d7
Nile red quantificationlowast
NI ADIP I ALP siRNA0
50
100
150
200
Con
trol s
iRN
AA
LP si
RNA
Day 3 Day 5 OS induction Day 7 OS induction
Fold
indu
ctio
n lo
g sc
ale
10000
1000
100
10
1
Day 0 OS day 7
Control siRNAALP siRNA
(A) (B)
(C)
(D)
Figure 4 Effect of alkaline phosphatase (ALP) gene silencing by small interfering RNA (siRNA) on a clonal population of human bonemarrow stromal stem cell CL1 (A) ALP staining at day 3 day 5 or day 7 days of osteoblast differentiation (OB induction) (B) Quantitativereal-time PCR for ALP gene following ALP siRNA transfection at day 0 OB and day 7 of OB Data are presented as fold change in expressionof each target gene normalized to GAPDH (119899 = three independent experiments 119901 lt 001) Western blotting analysis of day 3 and day 7after siRNA ALP transfection of CL1 cells ALPL specific antibody and B-actin was used (C) Mineralized matrix formation as visualizedby Alizarin Red S staining in siRNA transfected CL1 cells after 14 days of OB induction (D) Nile red quantification of mature lipid filledadipocyte in control noninduced (Cont) adipocyte induced (Adip I) and ALP siRNA transfected cells that are adipocyte induced (ALPsiRNA) Adipocyte induction was carried out for 7 days lowast119901 lt 005
phenotype and cellular responses of hMSC-TERT are similarto those of primary hMSC [10]While we have identified ALPas a marker that can be used for a prospective identificationof differentiation committed population of hBMSC we iden-tified additional distinctive molecular markers of the cellsFor example IGF-1 IGF-2 and IGF binding protein 5 wereenriched in CL1 compared to CL2 IGFs and their bindingproteins are very well-studied factors that play a role inhBMSC proliferation and osteoblast differentiation [18] On
the other hand annexin A3 as well as several immune-relatedgenes was highly enriched in CL2 compared to CL1 Futurestudies are needed to determine the functional significanceof these molecules in relation to the functional identity ofvarious cell populations within the hBMSC cultures andtheir usefulness as biomarkers to dissect the heterogeneouspopulation of cultured hBMSC
Our finding of the presence of functional diversity withinhBMSC cultures that contain progenitor and nonprogenitor
Stem Cells International 15
Table 4 Microarray data analysis showing genes found to be upregulated during adipogenic differentiation and downregulated after ALPKO
62 common elements inldquoAD uprdquo and ldquoALP downrdquo Gene name FC (ALP siRNA versus
control siRNA)APOBEC3G Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3G minus15938025IFI44L Interferon-induced protein 44-like minus11520283PAQR5 Progestin and adipoQ receptor family member V minus6868241PNMA2 Paraneoplastic antigen MA2 minus59695344DUSP23 Dual specificity phosphatase 23 minus54786854CLDN23 Claudin 23 minus51885047ANKDD1A Ankyrin repeat and death domain containing 1A minus51646647IL8 Interleukin 8 minus4887188LRRC23 Leucine rich repeat containing 23 minus47611775IL6 Interleukin 6 (interferon beta 2) minus4693139LIFR Leukemia inhibitory factor receptor alpha minus46540866PTGFR Prostaglandin F receptor (FP) minus4457529FAM134B Family with sequence similarity 134 member B minus4403495CYFIP2 Cytoplasmic FMR1 interacting protein 2 minus4260462METTL7A Methyltransferase-like 7A minus40480843APOBEC3F Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3F minus39502614CA5B Carbonic anhydrase VB mitochondrial minus393889ITGA10 Integrin alpha 10 minus39143775FMO3 Flavin containing monooxygenase 3 minus3852087IMPA2 Inositol monophosphatase 2 (human) minus38374884CDO1 Cysteine dioxygenase type I minus38181455CCDC68 Coiled-coil domain containing 68 minus37292893
HEXDC Hexosaminidase (glycosyl hydrolase family 20 catalytic domain) containing minus2094836SARM1 Sterile alpha and TIR motif containing 1 minus20797038BBS1 Bardet-Biedl syndrome 1 minus20146718SERPING1 Serpin peptidase inhibitor clade G (C1 inhibitor) member 1 minus20102212FAM162A Family with sequence similarity 162 member A minus2005807TCTN1 Tectonic family member 1 minus20033443
AD up ALP down
1006 62 1145
(a)
140
120
100
080
060
040
020
mdash
ALPL PPARg aP2 LPL
lowastlowastlowastlowastlowastlowast
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
lowast
(b)
Figure 5 Adipocyte related genes downregulated in ALP knockdown CL1 cells (a) Venn diagram of whole gene expression analysis of ADupregulated genes compared with ALP KD downregulated genes (b) Quantitative real-time PCR of four selected common genes from Venndiagram including ALPL PPARg2 aP2 and LPL Data are presented as fold change in expression of each target gene normalized to GAPDH(119899 = three independent experiments lowast119901 lt 005 lowastlowast119901 lt 001 and lowastlowastlowast119901 lt 0001)
Stem Cells International 17
cell populations has a clinical relevance It demonstrates thatthe progenitor function and the immune modulatory rolesof hBMSC [39] are mediated by specific and distinguishablepopulations of hBMSC Thus future clinical studies employ-ing hBMSC should attempt to administer the relevant sub-population of hBMSC dependent on the experimental aim asa novel approach to improving the clinical efficiency insteadof the current use of heterogeneous hBMSC populations
Competing Interests
The authors declare that they have no competing interests
Authorsrsquo Contributions
Mona Elsafadi and Amer Mahmood contributed equally tothis work
Acknowledgments
These studies are supported by a grant from the College ofMedicine Research Centre (CMRC) Deanship of ScientificResearch King Saud University Riyadh Saudi Arabia Theauthors would like to thank Mr Ali Al-Roalle for all histechnical help They thank Natalie Twine for advice on DNAmicroarray analysis
References
[1] W Zaher L Harkness A Jafari and M Kassem ldquoAn update ofhuman mesenchymal stem cell biology and their clinical usesrdquoArchives of Toxicology vol 88 no 5 pp 1069ndash1082 2014
[2] S Post B M Abdallah J F Bentzon andM Kassem ldquoDemon-stration of the presence of independent pre-osteoblastic andpre-adipocytic cell populations in bone marrow-derived mes-enchymal stem cellsrdquo Bone vol 43 no 1 pp 32ndash39 2008
[3] K H Larsen C M Frederiksen J S Burns B M Abdallahand M Kassem ldquoIdentifying a molecular phenotype for bonemarrow stromal cells with in vivo bone-forming capacityrdquoJournal of Bone and Mineral Research vol 25 no 4 pp 796ndash808 2010
[4] A Rosenzweig ldquoCardiac cell therapymdashmixed results frommixed cellsrdquoThe New England Journal of Medicine vol 355 no12 pp 1274ndash1277 2006
[5] S A Kuznetsov P H Krebsbach K Satomura et al ldquoSingle-colony derived strains of human marrow stromal fibroblastsform bone after transplantation in vivordquo Journal of Bone andMineral Research vol 12 no 9 pp 1335ndash1347 1997
[6] M Kassem and P Bianco ldquoSkeletal stem cells in space and timerdquoCell vol 160 no 1-2 pp 17ndash19 2015
[7] A I Caplan and D Correa ldquoPDGF in bone formation andregeneration new insights into a novel mechanism involvingMSCsrdquo Journal of Orthopaedic Research vol 29 no 12 pp 1795ndash1803 2011
[8] K Stewart S Walsh J Screen et al ldquoFurther characterizationof cells expressing STRO-1 in cultures of adult human bonemarrow stromal cellsrdquo Journal of Bone and Mineral Researchvol 14 no 8 pp 1345ndash1356 1999
[9] J L Simonsen C Rosada N Serakinci et al ldquoTelomeraseexpression extends the proliferative life-span and maintains
the osteogenic potential of human bone marrow stromal cellsrdquoNature Biotechnology vol 20 no 6 pp 592ndash596 2002
[10] M Al-Nbaheen R Vishnubalaji D Ali et al ldquoHuman stromal(mesenchymal) stem cells from bone marrow adipose tissueand skin exhibit differences in molecular phenotype and differ-entiation potentialrdquo Stem Cell Reviews and Reports vol 9 no 1pp 32ndash43 2013
[11] C A Gregory W G Gunn A Peister and D J Prockop ldquoAnAlizarin red-based assay of mineralization by adherent cells inculture comparison with cetylpyridinium chloride extractionrdquoAnalytical Biochemistry vol 329 no 1 pp 77ndash84 2004
[12] AMahmood L Harkness H D Schroslashder B M Abdallah andM Kassem ldquoEnhanced differentiation of human embryonicstem cells to mesenchymal progenitors by inhibition of TGF-120573activinnodal signaling using SB-431542rdquo Journal of Bone andMineral Research vol 25 no 6 pp 1216ndash1233 2010
[13] H-J Prins A K Braat D Gawlitta et al ldquoIn vitro inductionof alkaline phosphatase levels predicts in vivo bone formingcapacity of human bone marrow stromal cellsrdquo Stem CellResearch vol 12 no 2 pp 428ndash440 2014
[14] K Stenderup J Justesen C Clausen and M Kassem ldquoAgingis associated with decreased maximal life span and acceleratedsenescence of bone marrow stromal cellsrdquo Bone vol 33 no 6pp 919ndash926 2003
[15] D G Phinney G Kopen W Righter S Webster N Tremainand D J Prockop ldquoDonor variation in the growth propertiesand osteogenic potential of human marrow stromal cellsrdquoJournal of Cellular Biochemistry vol 75 no 3 pp 424ndash436 1999
[16] C M Digirolamo D Stokes D Colter D G Phinney R Classand D J Prockop ldquoPropagation and senescence of humanmarrow stromal cells in culture a simple colony-forming assayidentifies samples with the greatest potential to propagate anddifferentiaterdquo British Journal of Haematology vol 107 no 2 pp275ndash281 1999
[17] S James J Fox F Afsari et al ldquoMultiparameter analy-sis of human bone marrow stromal cells identifies distinctimmunomodulatory and differentiation-competent subtypesrdquoStem Cell Reports vol 4 no 6 pp 1004ndash1015 2015
[18] B Peruzzi A Cappariello A Del Fattore N Rucci F DeBenedetti and A Teti ldquoc-Src and IL-6 inhibit osteoblastdifferentiation and integrate IGFBP5 signallingrdquo Nature Com-munications vol 3 article 630 2012
[19] D Fortunati S Reppe A-K Fjeldheim M Nielsen V TGautvik and K M Gautvik ldquoPeriostin is a collagen associatedbonematrix protein regulated by parathyroid hormonerdquoMatrixBiology vol 29 no 7 pp 594ndash601 2010
[20] A Oshima H Tanabe T Yan G N Lowe C A Glackin andA Kudo ldquoA novel mechanism for the regulation of osteoblastdifferentiation transcription of periostin a member of thefasciclin I family is regulated by the bHLH transcription factorTwistrdquo Journal of Cellular Biochemistry vol 86 no 4 pp 792ndash804 2002
[21] T G Kashima T Nishiyama K Shimazu et al ldquoPeriostina novel marker of intramembranous ossification is expressedin fibrous dysplasia and in c-Fos-overexpressing bone lesionsrdquoHuman Pathology vol 40 no 2 pp 226ndash237 2009
[22] CGalliM PiergianniM Piemontese et al ldquoPeriostin improvescell adhesion to implantable biomaterials and osteoblasticdifferentiation on implant titanium surfaces in a topography-dependent fashionrdquo Journal of Biomedical Materials Research -Part A vol 102 no 11 pp 3855ndash3861 2014
18 Stem Cells International
[23] D L Coutu H W Jian A Monette G-E Rivard M DBlostein and J Galipeau ldquoPeriostin amember of a novel familyof vitamin K-dependent proteins is expressed by mesenchymalstromal cellsrdquo The Journal of Biological Chemistry vol 283 no26 pp 17991ndash18001 2008
[24] T Minamizato K Sakamoto T Liu et al ldquoCCN3NOV inhibitsBMP-2-induced osteoblast differentiation by interacting withBMP andNotch signaling pathwaysrdquo Biochemical and Biophysi-cal Research Communications vol 354 no 2 pp 567ndash573 2007
[25] W Si Q Kang H H Luu et al ldquoCCN1Cyr61 is regulated bythe canonicalWnt signal and plays an important role inWnt3A-induced osteoblast differentiation of mesenchymal stem cellsrdquoMolecular and Cellular Biology vol 26 no 8 pp 2955ndash29642006
[26] Q Luo Q Kang W Si et al ldquoConnective tissue growth factor(CTGF) is regulated by Wnt and bone morphogenetic proteinssignaling in osteoblast differentiation of mesenchymal stemcellsrdquo The Journal of Biological Chemistry vol 279 no 53 pp55958ndash55968 2004
[27] Y-S Yoon N Lee and H Scadova ldquoMyocardial regenerationwith bone-marrow-derived stem cellsrdquo Biology of the Cell vol97 no 4 pp 253ndash263 2005
[28] Y S Yoon ldquoCardiac regeneration with novel bone marrow-derived multipotent stem cellsrdquo Discovery Medicine vol 5 pp204ndash208 2005
[29] Y-S Yoon A Wecker L Heyd et al ldquoClonally expanded novelmultipotent stem cells from human bone marrow regeneratemyocardium after myocardial infarctionrdquo Journal of ClinicalInvestigation vol 115 no 2 pp 326ndash338 2005
[30] GM BolandG Perkins D J Hall andR S Tuan ldquoWnt 3a pro-motes proliferation and suppresses osteogenic differentiationof adult human mesenchymal stem cellsrdquo Journal of CellularBiochemistry vol 93 no 6 pp 1210ndash1230 2004
[31] D A Cook S W Fellgett M E Pownall P J OrsquoShea and PG Genever ldquoWnt-dependent osteogenic commitment of bonemarrow stromal cells using a novel GSK3120573 inhibitorrdquo Stem CellResearch vol 12 no 2 pp 415ndash427 2014
[32] J De Boer R Siddappa C Gaspar A VanApeldoorn R Foddeand C Van Blitterswijk ldquoWnt signaling inhibits osteogenicdifferentiation of humanmesenchymal stem cellsrdquoBone vol 34no 5 pp 818ndash826 2004
[33] MAl-Toub A AlmusaM Almajed et al ldquoPleiotropic effects ofcancer cellsrsquo secreted factors on human stromal (mesenchymal)stem cellsrdquo Stem Cell Research and Therapy vol 4 article 1142013
[34] L Chen D Qanie A Jafari et al ldquoDelta-like 1fetal antigen-1(Dlk1FA1) is a novel regulator of chondrogenic cell differenti-ation via inhibition of the Akt kinase-dependent pathwayrdquoTheJournal of Biological Chemistry vol 286 no 37 pp 32140ndash321492011
[35] T Matsunobu K Torigoe M Ishikawa et al ldquoCritical roles ofthe TGF-120573 type I receptor ALK5 in perichondrial formationand function cartilage integrity and osteoblast differentiationduring growth plate developmentrdquo Developmental Biology vol332 no 2 pp 325ndash338 2009
[36] B C Nair D E Johnson R J Majeska J A Rodkey C DBennett and G A Rodan ldquoRat alkaline phosphatase II Struc-tural similarities between the osteosarcoma bone kidney andplacenta isoenzymesrdquo Archives of Biochemistry and Biophysicsvol 254 no 1 pp 28ndash34 1987
[37] P J Butterworth ldquoAlkaline phosphatase Biochemistry of mam-malian alkaline phosphatasesrdquo Cell Biochemistry and Functionvol 1 no 2 pp 66ndash70 1983
[38] A T Ali C B Penny J E Paiker et al ldquoAlkaline phosphataseis involved in the control of adipogenesis in the murinepreadipocyte cell line 3T3-L1rdquo Clinica Chimica Acta vol 354no 1-2 pp 101ndash109 2005
[39] G M Spaggiari and L Moretta ldquoCellular and molecularinteractions of mesenchymal stem cells in innate immunityrdquoImmunology and Cell Biology vol 91 no 1 pp 27ndash31 2012
adipocytes and chondrocytes (trilineage differentiation) andto form heterotopic bone and bone marrow organ whenimplanted in vivo subcutaneously in immune deficient mice[5] The identity of the remaining cells is not clarified butthey may represent lineage-committed cells [3] Therefore itis plausible that functional heterogeneity exists in culturedhBMSC reflecting the in vivo functional and developmentalheterogeneity of hBMSC [6]
In addition to their ability to differentiate into skeletaltissue cells (known as progenitor function) hBMSC possessimmunomodulatory characteristics (known as nonprogen-itor functions) [7] It is not clear whether these differentfunctions are mediated by a number of independent sub-populations within the hBMSC [2] Only a few studies havetried to identify the subpopulation within cultured hBMSCbased on surface markers for example STRO1 and alkalinephosphatase (ALP) but limited molecular phenotyping hasbeen conducted [8]
We have previously demonstrated the presence of mor-phological and functional heterogeneity of clones isolatedfrom telomerized hMSC (hMSC-TERT) cell line [3]The aimof the present study was therefore to further study in detailthe heterogeneity of cultured hBMSC as demonstrated bytwo clonal cell lines with opposite cellular and functionalphenotypeWe also employed theDNAmicroarrays to definetheir molecular signature and signaling pathways associatedwith their functional phenotype
2 Experimental Procedures21 Cell Culture As a model for hBMSC we employedimmortalized hBMSC-TERT cell line that is created fromnormal human BMSC by overexpression of human telom-erase reverse transcriptase gene (hTERT) [9] The hBMSC-TERT cells have been extensively characterized and theyexhibit similar cellular and molecular phenotype to primaryMSC [10] CL1 and CL2 cells are clonal cell populationsof hBMSC-TERT identified in long term culture (passagenumbers 15ndash25) of hBMSC-TERT and were chosen based ontheir distinct and different morphologies Cells were culturedin Dulbeccorsquos Modified Eagle Medium (DMEM) supple-mented with D-glucose 4500mgL 4mM L-glutamine and110mgL sodium pyruvate 10 Fetal Bovine Serum (FBS) 1xpenicillin-streptomycin (Pen-strep) and nonessential aminoacids (all purchased from Gibco-Invitrogen USA) For somecontrol experiments primary bone marrow derived MSC(phBMSC) were employed Sixty milliliters of bone marrowwas aspirated from the iliac crest bone of consenting healthydonors This procedure was approved by the King KhalidUniversity Hospital-King Saud University ethics committeephBMSCwere isolated from bonemarrowmononuclear cellsby plastic adherence as described previously [9]
22 Cell Proliferation Cell proliferation rate was determinedby counting cell number and calculating population doubling(PD) rate The cells were cultured in 25 cm2 tissue culturePetri dish at cell density 05 times 106 cells (28000 cellscm2) Atconfluence the cells were trypsinized and counted manuallyby hemocytometer At each passage population doubling wasdetermined by the following formula log119873 log 2 where 119873
is the number of cells at confluence divided by the initialcell number Cumulative PD level is the sum of populationdoublings and PD rate is PDtime in culture
23 Flow Cytometry Cells were trypsinized to a singlecell suspension were recovered by centrifugation at 200 gfor 5min washed twice in ice-cold PBS supplementedwith 2 BSA and resuspended at a concentration of105 cellsantibody After incubation with the preconjugatedantibodies or matched isotype controls for 30min on icein the dark cells were washed with PBS resuspended in500120583L of PBS and analyzed in the BD FACSCalibur flowcytometer (BD Biosciences) Living cells were gated in adot plot of forward versus side scatter signals acquired ona linear scale At least 10000 gated events were acquiredon a log fluorescence scale Positive staining was distinct asthe emission of a fluorescence signal that surpassed levelsachieved by gt99 of control cell population stained withcorresponding isotype antibodies The ratios of fluorescencesignals versus scatter signals were calculated and histogramswere generated using the software Cell Quest Pro SoftwareVersion 33 (BD Biosciences) The following antibodies wereused all from BD Biosciences FITC-PE-APC-Mouse IgG1kisotype control APC-Mouse Anti-Human CD44 (559942)FITC-Mouse Anti-HumanCD63 (557305) PE-Mouse Anti-Human CD73 (550257) PE-Mouse Anti-Human CD105(560839) PE-Mouse Anti-Human CD146 (550315) PE-Mouse Anti-HumanCD166 (560903) and Alexa Fluor 488Mouse Anti-Human Alkaline Phosphatase (561495)
24 Electron Microscopy After trypsinizing the hBMSC cellsfrom the flasks or 6-well plates were collected the sampleswere washed with PBS and the pellets were resuspendeddirectly in 25 glutaraldehyde in 01M phosphate buffer (pH72) and kept at 4∘C for 4 hr First the cells were washed with01Mphosphate buffer (pH72) and transferred to 1osmiumtetroxide (OsO4) solution in 01M phosphate buffer (pH 72)for two hr The cells were dehydrated in ascending grades ofethanolThe cells were then resuspended in acetone and werealiquoted into BEEM embedding capsules and infiltratedwith acetone resin mixture followed by embedding in apure resin mixture for two hr Semithin sections (05 120583mthickness) and ultrathin sections (70 nm thickness) wereprepared examined and photographed under a transmissionelectron microscope (TEM) (Jeol 1010 Jeol Tokyo Japan)
25 In Vitro Osteoblast Differentiation Cells were grownin standard DMEM growth medium in 6-well plates at20000 cellcm2 When 70ndash80 confluence was achieved testcells were cultured inDMEM supplementedwith osteoblasticinduction mixture (referred to as OS) containing 10 FBS1 Pen-strep 50120583gmL L-ascorbic acid (Wako ChemicalsNeuss Germany) 10mM 120573-glycerophosphate (Sigma) and10 nM calcitriol (112057225-dihydroxyvitamin D3 Sigma) and10 nM dexamethasone (Sigma) noninduced cells (referredto as Cont) were cultured in normal growth media for thesame duration as induced The media were replaced threetimes per week Cells cultured in standard culture mediumwere considered as control At day 14 of differentiation
Stem Cells International 3
mineralized nodules became apparent and were stained withAlizarin Red S and ALP
26 In Vitro Adipocyte Differentiation Cells were grown instandard DMEM growth medium in 6-well plates at 03times 106 cellsmL At 90ndash100 confluence cells were culturedin DMEM supplemented with adipogenic (Adip) inductionmixture containing 10 FBS 10 Horse Serum (Sigma)1 Pen-strep 100 nM dexamethasone 045mM isobutylmethylxanthine (Sigma) 3120583gmL insulin (Sigma) and 1 120583MRosiglitazone (Novo Nordisk Bagsvaerd Denmark) Themedia were replaced three times per week Cells cultured instandard culture medium were considered as control Fromday 3 of differentiation small lipid droplets became visibleand at day 7 they were stained with Oil Red-O and Nile red
27 In Vitro Chondrogenic Differentiation Both CL1 and CL2cells were trypsinized and counted around 1 times 106 cellstaken in each 15mL conical tube centrifuged at 400timesg for5 minutes For chondrocyte differentiation pellet culturesystem used chondrocyte induction was done in mediacontaining advanced DMEMF12 supplemented with 1 ITSPremix Tissue Culture Supplement 100 nM dexamethasoneGlutamax and 10 ngmL transforming growth factor-beta-3(TGF120573-3) Cells were maintained in chondrocyte differentia-tion media for 21 days and changed every two days
28 Cytochemical Staining281 Alkaline Phosphatase (ALP) Staining CL1 and CL2 cellswere stained before OS differentiation for the basal ALPexpression and after OB differentiation at day 7 of inductionCells cultured in 6-well plates were washed in PBS minusminus(-Ca -Mg) andfixed in acetonecitrate buffer 10mMat pH42for 5min at room temperature The NaphtholFast Red stain[02mgmL Naphthol AS-TR phosphate substrate (Sigma)][0417mgmL of Fast Red (Sigma)] was added for one hourat room temperature
Histological tissue blocks were sectioned at 4 micronsImmunohistochemical staining was performed on CL1 andCL2 chondrocyte 3D pellets using DAKO EnVision andPowerVision according to the manufacturerrsquos instructions(DAKO Glostrup Denmark) Briefly paraffin sections wereincubated for 1 hour at room temperature with primaryantibodies diluted in ChemMate (DAKO) (HumanAnti-Col-10 and Human Anti-Col-2 ABI) Sections were washed sub-sequently in Tris-buffered saline (TBS 005M pH 74) incu-bated for 30 minutes with secondary anti-mouse IgHRP-conjugated polymers (K4001 En Visionthorn DAKO) and visu-alizedwith 330-diaminobenzidine tetrahydrochloride (DABS3000 DAKO) or with 3-amino-9-ethylcarbazole (AECDAKO) according to manufacturerrsquos instruction Controlswere performed with nonimmune immunoglobulins of thesame isotype as the primary antibodies (negative controls)and processed under identical conditions Alcian blue stain-ing was used to detect chondrocytes Sections of paraffin-embedded implants were stained with Alcian blue (Sigma)solution pH 25 at this pH all the glycoproteins (neutral andacidic) will be stained blue
29 Alizarin Red S Staining for Mineralized Matrix Seven-day-old OS differentiated cells in 6-well plates were used forAlizarin Red S staining The cell layer was washed with PBSand then fixed with 70 ice-cold ethanol for 1 hr at minus20∘CAfter removing the ethanol the cell layer was rinsed withdistilled water and stained with 40 nM AR-S (Sigma) pH 42for 10 minutes at room temperature Excess dye was washedoff with water followed by a wash with PBS for few minutesto minimize nonspecific AR-S stain
For quantifying the Alizarin Red S staining the air-driedplates the Alizarin Red S dye was eluted in 800 120583L of aceticacid incubated in each well for 30 minutes at room temper-ature as described [11] and measured in spectrophotometer(BioTek Epoch) at 405 nm
210 Quantitative ALP Activity To quantify ALP activity inCL1 and CL2 hBMSC before and after OS differentiationwe used the BioVision ALP activity colorimetric assay kit(BioVision Inc CA USA) with some modifications Cellswere cultured in 96-well plates under normal conditionsthen on day of analysis wells were rinsed once with PBSand were fixed using 37 formaldehyde in 90 ethanol for30 seconds at room temperature Subsequently fixative wasremoved and 50120583L of pNPP solution was added to each welland incubated for 1 hour in the dark at room temperatureThe reaction was subsequently stopped by adding 20 120583L stopsolution and gently shaking the plate ODwas thenmeasuredat 405 nm
211 Oil Red-O Staining for Lipid Droplets CL1 and CL2 cellsdifferentiated to adipocytes with Adip induction media atday 7 were used Accumulated cytoplasmic lipid dropletswere visualized by staining with Oil Red-O After washingcells grown in 6-well plates with PBS the cells were fixedin 4 formaldehyde for 10min at room temperature andthen rinsed once with 3 isopropanol and stained for 1 hr atroom temperature with filtered Oil Red-O staining solution(prepared by dissolving 05 g Oil Red-O powder in 60isopropanol) To quantify staining of fat droplets Oil Red-O was used as a stain Oil Red-O was eluted by adding 100isopropanol to each well and color changes were measuredby spectrophotometer at 510 nm (BioTek SpectrophotometerEpoch)
212 Nile Red Fluorescence Determination and Quantificationof Adipogenesis A stock solution of Nile red (1mgmL) inDMSO was prepared and stored at minus20∘C protected fromlight Staining was performed on unfixed cells Culturedundifferentiated and day 7 adipocyte differentiated cells weregrown in Corning polystyrene flat bottom 96-well TC-treated black microplates (Corning NY USA) were washedonce with PBS The dye was then added directly to the cells(5 120583gmL in PBS) and the preparation was incubated for10min at room temperature and then washed twice with PBSFluorescent signal was measured using SpectraMaxM5 fluo-rescence spectrophotometer plate reader (Molecular DevicesCo Sunnyvale CA USA) using bottom well-scan modewhere nine readings were taken per well using Ex (485 nm)and Em (572 nm) spectra
4 Stem Cells International
213 Quantitative Real-Time PCR (qRT-PCR) Analysis TotalRNA was extracted using MagNA pure compact RNA iso-lation kit (Roche Applied Science Germany Cat number04802993001) in an automatedMagNA pure compact system(Roche Germany) as recommended by the manufacturerThe total RNAwas quantified by Nanodrop spectrophotome-ter (Nanodrop 2000 Thermo Scientific USA) Complemen-tary DNA (cDNA) was synthesized from 1 120583g of the RNAsamples usingHighCapacity cDNAReverse Transcription kit(Applied Biosystems USA) using Labnet Multigene thermo-cycler according to the manufacturerrsquos instructions Relativelevels of mRNA were determined from cDNA by real-timePCR (Applied Biosystems-Real-Time PCRDetection System)with Power SYBR Green PCR kit (Applied BiosystemsUK) according to the manufacturerrsquos instructions Followingnormalization to the reference gene GAPDH quantificationof gene expression was carried out using a comparative Ctmethod where ΔCt is the difference between the CT valuesof the target and the reference gene and fold induction wasperformed from the control (Cont) for the same time pointPrimers (Supplementary Table 1 in Supplementary Materialavailable online at httpdxdoiorg10115520169378081)were obtained from Applied Biosystems (USA) as TAQMANprimers or previously published primers were used (seeSupplementary Table 1)
214 DNAMicroarray Global Gene Expression Analysis Fourhundred ng of total RNA was used as input for gener-ating biotin-labeled cRNA (Ambion Austin TX UnitedStates) cRNA samples were then hybridized onto Illuminahuman-8 BeadChips version 3 Hybridization washing Cy3-streptavidin staining and scanning were performed on theIllumina BeadStation 500 platform (Illumina San DiegoCA USA) according to the manufacturerrsquos instructions andeverything was done in triplicate Expression data analysiswas carried out using the Partek genomic suite softwareRaw data were background-subtracted normalized usingthe ldquorank invariantrdquo algorithm and filtered for significantexpression on the basis of negative control beads Geneswere considered significantly expressed with detection 119901values le 001 Differential expression analysis was performedwith the Illumina custom method using freshly isolatedprimary hBMSC (used at passage 3) as a reference controlThe following parameters were set to identify statisticalsignificance differential 119901 values le 001 fold change ratiogt 15 Pathway analysis was performed using DAVID Bioin-formatics Resources 67 (httpdavidabccncifcrfgov) andGeneSpring GX software (Agilent Technologies) Pathwayanalysis for CL1 OS D14 versus CL2 OS D14 was conductedusing the Single Experiment Pathway analysis feature inGeneSpring 120 (Agilent Technologies)
215 Small Interfering (si)RNATransfection For transfectionhBMSC in logarithmic growth phase were transfected withSilencer Select Predesigned ALP siRNA (25 nM) (Assay IDs1298 and Cat number 4390824) (Ambion The RNA Com-pany USA) using Lipofectamine RNAiMAX Reagent (Invit-rogen CA USA) plus serum-free Opti-MEMI mediumunder the conditions described by the manufacturer At
day 3 of transfection the cells were induced for osteogenicdifferentiation for an additional 7 days ALP staining wasused as a control for the siRNA transfection efficiency andtimeline
216 Statistical Analysis All of the results were presentedas the mean and standard deviation (SD) of at least 3independent experiments with 3ndash5 technical repeats in eachexperiment Studentrsquos 119905-test (two-tailed) was used for testingdifferences between groups 119901 value lt005 was consideredstatistically significant
3 Results
31 Comparison between CL1 and CL2 Differences in Mor-phology Proliferation and Marker Expression Profile Weisolated two distinct clonal cell populations of hBMSC-TERThBMSC-CL1 and hBMSC-CL2 (for easiness will be termedhereafter CL1 and CL2) based on differences in cell morphol-ogy (Figure 1(a)) CL1 cells had cuboidalmorphologywhereasCL2 cells have spindle-shaped fibroblast-like morphologyCL1 cells had higher proliferation rate compared to CL2(Figure 1(b)) mean PD rates of CL1 and CL2 were 0714 and0429 PDday respectively (Figure 1(b)) Both CL1 and CL2expressed surface marker profiles characteristics of hBMSC(gt90) CD44+ CD63+ CD73+ CD105+ and CD166+(Figure 1(c)) However CL1 cells showed higher expressionof CD146 (927 versus 12) and ALP (98 versus 0)compared to CL2 (Figure 1(d)) TEM revealed the presenceof abundant pseudopodia in CL1 indicating high motility(Figure 1(e)(A)) as well as well-developed mitochondriaand rough endoplasmic reticulum (rER) suggesting highmetabolic activity CL2 cells contained abundant phagocyticvacuole (pv)microvilli (mi) and lysosomes (ly) (Figure 1(e))
We performed quantitative real-time PCR (RT-PCR)for genes expressed in mesodermal progenitor cells [12]CL1 expressed higher levels of BMP4 MIXL1 WNT3a andTWIST compared to CL2 (Figure 1(f) 119901 lt 001) In contrastCL2 expressed higher levels of Kinase Insert Domain Recep-tor (Type III Receptor Tyrosine Kinase) (KDR) expressed inendothelial cells and smoothmusclemyosin heavy chain gene(smMHC) expressed in smooth muscle cells (Figure 1(f))
32 CL1 Cells Exhibit Enhanced Osteoblast DifferentiationFollowing osteoblast (OB) differentiation induction ALPstaining andALP enzymatic activity were significantly higherin CL1 compared to CL2 cells (Figure 2(a) 119901 lt 001)Similarly Alizarin Red staining and quantitation of formedmineralized matrix were more pronounced in CL1 cells (Fig-ure 2(b) 119901 lt 001) In addition CL1 cells expressed higherlevels of osteoblastic genes ALP RUNX2 and osteopontin(OPN) (Figure 2(c) upper panel) compared to CL2 cells
Global gene expression microarray analysis of OB dif-ferentiated cells at day 14 showed around 1060 genes signif-icantly upregulated more than 2-fold (119901 lt 001) in CL1Among the upregulated genes 80 genes were annotated tobone development and osteoblast differentiation (Table 1)The highest upregulated genes included paired-like home-odomain 2 (PITX2) Insulin-like growth factor 1 (IGF1) and
Stem Cells International 5
CL1
CL2
(a)
Cum
ulat
ive P
D
Cumulative days
Growth curve
0 10 20 30 40
CL1 CL2
CL1 CL20
5
10
15
20
25
0
01
02
03
04
05
06
07
08
(PD
day
)
(b)
CL2
103
102
101
100
103
102
101
100
103
102
101
100
103
102
101
100
103
102
101
100
103
101
100
103
102
102
101
100
103
102
101
100
103
102
101
100
103
102
101
100
CL1
CD44 CD63 CD73 CD166 CD105
T3 CD44 T3 CD63 T3 CD73 T3 CD166 T3 CD105
0
20
40
60
80
Cou
nt
0
20
40
60
80
100
120
Cou
ntC
ount
Cou
nt
Cou
nt
Cou
nt
Cou
nt0
20
40
60
80
Cou
nt
0
20
40
60
80
100
Cou
nt
0
20
40
60
80
Cou
nt
80
60
40
20
0
100
50
0
120
100
80
60
40
20
0
120
100
80
60
40
20
0
100
80
60
40
20
0
(c)
ISO CD146 ALP
ISO CD146 ALP
FL2-
H
FL2-
H
FL2-
HPEPEPECL
2CL
1
FL1-H FL1-H FL1-H
FITCFITCFITC
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
(d)
Figure 1 Continued
6 Stem Cells International
(A) (B)
(e)
Fold
indu
ctio
n
CL1CL2
lowast
lowastlowast lowast
BMP4 MIXL1 WNT3a TWIST
lowastlowast
KDR smMHCminus01
01
03
05
07
09
11
13
15
mdash
050
100
150
200
250
300
CL1CL2
(f)
Figure 1 Cellular phenotype of two clonal populations of humanbonemarrow stromal stem cells CL1 andCL2 (a) Cellmorphology CL1 cellsshowed large cuboidal morphology while CL2 cells had spindle-shaped fibroblast-like morphology (b) Growth curves showing populationdoubling (PD) rate during long term culture (c) Flow cytometry analysis (FACS) shows expression of CD44 CD63 CD73 CD105 andCD166in CL1 and CL2 cells Matched isotype control was used for gating (d) Flow cytometry analysis presented as dot blot of CD146 and alkalinephosphatase (ALP) cell surface proteins (e) Transmission electron microscope (TEM) (A) CL1 (1200x) (B) CL2 (1500x) n nucleus nunucleolus rER rough endoplasmic reticulum ly lysosomes pv phagocytic vacuole and rer reticular stalk of rER (f) Gene expressionanalysis using RT-PCR for a group of mesodermal and stromal genes Gene expression was normalized to GAPDH and presented as foldchange Data is shown as mean plusmn SD of three independent experiments lowast119901 lt 005 119901 lt 0001
collagen type V alpha 3 (COL5A3) osteomodulin (OMD)and T-box 15 (TBX15) (Table 1) Furthermore several knownosteoblast-related genes were upregulated in CL1 cells suchas bone morphogenetic protein 6 (BMP6) fibroblast growthfactor receptor 3 (FGFR3) insulin-like growth factor bindingprotein 5 (IGFBP5) and vitamin D (125-dihydroxyvitaminD3) receptor (VDR) (Table 1) On the other hand 1200 geneswere upregulated in CL2 cells 255 genes were annotated toimmunity and immune response and defense This categoryincluded genes from complement system chemokine (C-Cmotif) ligands interferon family chemokine (C-X-C motif)ligands and receptor major histocompatibility complex classII molecules interleukins and tumor necrosis factor receptorsuperfamily (Table 2 and Supplementary Table 3)
33 CL1 Cells Exhibit Enhanced Adipocyte DifferentiationWe observed significant differences between CL1 and CL2in their response to adipocytic differentiation inductionCL1 differentiated readily to adipocytes compared to CL2(Figure 2(d) lower panel) evidenced by higher levels ofadipocyticmarkers gene expression LPL (lipoprotein lipase)and adiponectin as well as formation of mature lipid filledadipocytes visualized by Oil Red-O staining and quantitativeNile red staining (Figure 2(d))
34 CL1 Cells Differentiate to Chondrocytic Lineage Inpellet cultures CL1 cells formed 3D pellets containingproteoglycan-secreting chondrocytes which stained positivewith Alcian blue Limited chondrocyte differentiation wasvisible in cell pellets of CL2 cells The differentiated chon-drocytes in CL1 pellets expressed higher levels of collagen Xand collagen II which was overlapping the Alcian blue stain(Figure 2(e))
35 Molecular Signature of CL1 and CL2 Cells To define themolecular signature and molecular differences between CL1
and CL2 we compared the basal gene expression patternof CL1 and CL2 cells using DNA microarrays The PCAanalysis showed a clear separation between CL1 and CL2(Supplementary Figure 1) Comparison between CL1 withCL2 showed that 915 geneswere differentially expressed in thetwo cell lines (gt2-fold119901 lt 001) 462 genes were upregulatedand 452 were downregulated in CL1 versus CL2 The mostrelevant genes that were upregulated in CL1 are listed inTable 3(a) Among these 35 highly expressed genes in CL1 thefollowing 11 genes were present in skeletal and muscular sys-tem development and function FOLR3 CCL3L1 SERPINB2POSTN IGFBP5 CCL3 NOV ALP TNFRSF11B ACTG2and CDH11 (Table 3(a)) Functional annotation of the upreg-ulated genes in CL1 using the Ingenuity Pathway Analysis(IPA) revealed enrichment in the following categories ldquotissuedevelopmentrdquo ldquoskeletal and muscular system developmentand functionrdquo and ldquoorganismal developmentrdquo (Table 3(b))Furthermore the DAVID annotation tool was employed toassess the functional relationships of the upregulated genesin CL1 showing enrichment in ontologies ldquoskeletal andmuscular system development and functionrdquo that includedbone size osteoblast differentiation bone mineralizationand bone mineral density (Table 3(c)) CL1 exhibited upreg-ulation of WNT pathway ligands WNT5B (2-fold) andLRP5 (2-fold) (Table 1) Also ALP was among the highlyexpressed genes together with POSTN IGFBP5 SPP1 IL-6 and DKK1 (Tables 3(a) and 1) These genes are knownto play an important role in osteoblast differentiation andbone formation For CL2 inhibitors of WNT pathway wereupregulated and included SFRP1 (11-fold) DKK2 (32-fold)FGF2 (31-fold) and GBP2 (24-fold) Functional annotationof the upregulated genes in CL2 revealed enrichment in thefollowing categories ldquodevelopmental processrdquo ldquomulticellularorganismal processrdquo ldquobiological adhesionrdquo and ldquoimmunesystem processrdquo (Supplementary Table 2A) In-depth analysis
Stem Cells International 7
CL1
CL2
ALPCont D 14
Cont OB I
ALP activity
mdash100
200
300
400
500
600
Fold
chan
ge
CL1CL2
(a)
CL1
CL2
ALZCont D 14
Fold
chan
ge
25
20
15
10
5
0
Cont d14 d21
CL1CL2
ALZ quantification
(b)
CL1CL2
CL1CL2
LPL ADIPOQ
Cont ContADIP D14 ADIP D14
000
100
200
300
400
500
600
700
0
1000
2000
3000
4000
5000
CL1CL2
CL1CL2
CL1CL2
RUNX2 ALP OPN
Cont Cont ContOS D7 OS D7 OS D7
00
05
10
15
20
25
30
00
05
10
15
20
25
30
35
00
05
10
15
20
25
30
35
(c)
CL1CL2
CL1
CL2
Control Adip d7 Adip d14
d7 d140
005
01
015
02
025
03
035
04
045
Nile
red
fluor
esce
nt in
tens
ity
(d)
CL1
CL2
Alcian blue Col X Col II
(e)
Figure 2 In vitro osteoblastic and adipocytic differentiation of two clonal populations of human bone marrow stromal stem cells CL1 andCL2 cells Both cell lines were induced for osteoblast differentiation using standard protocol described in the Methods (a) ALP staining atday 14 in control noninduced (Cont) and osteoblast induced cells (D14) Right panel shows ALP activity (119899 = 3 independent experiments119901 lt 0001) (b) Mineralized matrix formation visualized by Alizarin Red S staining Right panel shows Alizarin Red quantification at day 14and 21 after osteoblast differentiation (119899 = 3 independent experiments 119901 lt 0001) (c) Quantitative RT-PCR of osteoblastic and adipocytegene markers in CL1 and CL2 during osteoblast (upper panel) and adipocyte (lower panel) differentiation ALP = alkaline phosphataseOPN = osteopontin LPL = lipoprotein lipase and ADIPOQ = adiponectin Data are presented as fold change in expression of each targetgene normalized to GAPDH (119899 = 3 independent experiments 119901 lt 005
119901 lt 0001) (d) CL1 and CL2 lines were induced for adipocytedifferentiation using standard protocol described in the Methods Adipocyte formation was visualized at day 7 (Adip d7) and day 14 (Adipd14) by Oil Red-O staining Lower panel presents quantification of Nile red staining (119899 = 3 independent experiments 119901 lt 0001) (e) CL1and CL2 lines were induced for chondrocyte differentiation using 21-day pellet culture method as described in the Methods The pellets werestained with Alcian blue collagen 10 (Col X) and collagen 2 (Col II) (original magnification 5x)
8 Stem Cells International
Table 1 Microarray data analysis showing genes related to bone development and osteoblast differentiation upregulated in CL1 versus CL2cells
Probe ID Genbank accession Gene name Gene symbol FCA 23 P167367 NM 153426 Paired-like homeodomain 2 PITX2 30884
of the biological processes revealed several immune-relatedpathways ldquoMAPKKK cascaderdquo ldquoimmunity and defenserdquoldquosignal transductionrdquo ldquoextracellular matrix protein-mediatedsignalingrdquo and ldquointerferon-mediated immunityrdquo among oth-ers that were upregulated (Supplementary Table 2B) Also40 genes related to immune system related factors wereidentified as significantly enriched in CL2 compared to CL1cells (Supplementary Table 2C)We chose the following genesfor validation of the microarray results NOV IGFBP5 ALPTAGLN and CDH11 as they were highly expressed in CL1RT-PCR analysis confirmed themicroarray results (Figure 3)
Furthermore we compared the molecular phenotype ofCL1 and CL2 cells with that of phBMSCWe found that morethan 80 of the genes expressed in CL1 and 90 in CL2 cellswere commonwith primary phBMSC (Supplementary Figure2) suggesting that CL1 and CL2 molecular phenotype existwithin the heterogeneous population of phBMSC cultures
36 ALP Knockdown Impairs Differentiation of CL1 CellsSince ALP has been suggested as a marker for hBMSCprogenitor cell lineage commitment [13] and was highlyupregulated in CL1 cells we tested its biological role in
CL1 cells ALP siRNA transfection decreased ALP proteinlevel ALP activity and mRNA gene expression compared tocontrol cells transfected with control siRNA (119901 lt 001) andthis inhibition was detectable up to day 7 days after osteoblastdifferentiation induction (Figures 4(A) and 4(B)) At day 14 ofdifferentiation mineralization ability of CL1 was significantlyimpaired (Figure 4(C)) In addition we found that thenumber of mature adipocyte formations was significantlyreduced to more than 75 (119901 lt 001) (Figure 4(D))
To identify relevant adipocyte differentiation associatedgenes that were targeted by ALP deficiency we compared thedownregulated genes of ALP deficient CL1 with the upreg-ulated genes identified during adipocytic differentiation ofCL1 We identified 62 genes that were common (Figure 5(a)Table 4) and among these genes were genes related tometabolism (primarily lipid and carbohydrate) and transportincluding CYB5B CHST1 TAP1 ATP8A1 LRP8 PLCD1and FABP5 (Table 4) We further performed quantitativereal-time PCR of ALP deficient CL1 cells during adipocytedifferentiation The following adipocyte-associated geneswere downregulated PPAR1205742 LPL and aP2 (Figure 5(b))
12 Stem Cells International
Table 3 Whole genome microarray analysis of two clonal popu-lations of human bone marrow stromal stem cells CL1 and CL2cells (a) Top 35 highly upregulated genes in CL1 cells versus CL2cells (b) Ingenuity Pathway Analysis (IPA) showing the differentphysiological system development and function genes found in eachcategory and the corresponding 119901 value (c) Upper part IPA analysisshowing the categories for skeletal and muscular system developmentand function upregulated in CL1 cells and lower part showing genesupregulated in CL2 cells NOV = nephroblastoma overexpressedIGFBP5 = insulin-like growth factor binding protein 5 ALP TAGLN= transgelin and CDH11 = OB-cadherin (osteoblast)
Name 119901 value moleculesPhysiological system development and function
Organismal development 705119864 minus 10ndash143119864 minus 03 181Embryonic development 100119864 minus 09ndash143119864 minus 03 154Organ development 100119864 minus 09ndash143119864 minus 03 145
(b) Continued
Name 119901 value moleculesSkeletal and muscularsystem development andfunction
100119864 minus 09ndash102119864 minus 03 123
Tissue development 100119864 minus 09ndash143119864 minus 03 236
(c)
Functions annotation 119901 value moleculesSkeletal and muscular system development and
function upregulated in CL1 cellsSize of bone 143119864 minus 06 24Differentiation ofosteoblasts 384119864 minus 06 25
Mineralization ofbone 493119864 minus 06 19
Bone mineral density 365119864 minus 05 19Skeletal and muscular system development and
function upregulated in CL2 cellsDevelopment ofmuscle 100119864 minus 09 44
Proliferation ofmuscle cells 229119864 minus 06 35
Remodeling of bone 363119864 minus 06 21Resorption of bone 394119864 minus 06 19
confirming impairment of adipocytic differentiation of ALPdeficient CL1 cells
4 Discussion
We extensively studied two cell populations within culturedhBMSC that were identified based on differences in mor-phology Cellular and molecular studies revealed differencesin growth differentiation capacity and molecular signatureOur data support the notion of the presence of cellular andfunctional heterogeneity among cultured hBMSC
Cellular heterogeneity of cultured hBMSC is recognizedin an increasing number of reports Several extrinsic andintrinsic factors may contribute to the observed hBMSCheterogeneity Extrinsic factors include donor-to-donor vari-ations in the number and quantity of initiating cells whichresult in differences in cell growth rate and differentia-tion capacity [14 15] Intrinsic factors have been examinedemploying single cell clonal analysis and revealed variationsin differentiation potential among individual colonies withinhBMSC cultures ranging from the presence of cells withtrilineage (osteoblast adipocytes and chondrocyte) potencyto cells with null potency [16] Also variations in the abilityof clonal cells to form heterotopic bone when implanted invivo have been reported [5] Our study corroborates thesefindings and provides more detailed cellular and molecularphenotyping of two examples of cell populations that existwithin the heterogeneous hBMSC cultures [17]
Determining the molecular signature of CL1 and CL2using whole genomemicroarray analysis showed enrichmentof lineage-commitment associated genes inCL1 For exampleinsulin-like growth factor 5 (IGFBP5) and interleukin 6 (IL6)
Stem Cells International 13
NOV0040
0030
0020
0010
mdash
IGFBP50015
0010
0005
mdash
ALP0100
0080
0060
0040
0020
mdash
TAGLN0800
0600
0400
0200
CDH11
0060
0040
0020
mdashmdash
LRP30015
0010
0005
mdash
CL1 CL2 CL1 CL2 CL1 CL2
CL1 CL2CL1 CL2CL1 CL2
Figure 3 Validation of whole genome microarray analysis of two clonal populations of human bone marrow stromal stem cells CL1 andCL2 cells Quantitative real-time PCR for highly expressed genes in CL1 cells NOV = nephroblastoma overexpressed IGFBP5 = insulin-likegrowth factor binding protein 5 ALP TAGLN = transgelin and CDH11 = OB-cadherin (osteoblast) Data are presented as fold change inexpression of each target gene normalized to GAPDH (119899 = three independent experiments 119901 lt 0001) (see also Table 3)
were 147- and 33-fold upregulated in CL1 cells respectivelyBoth factors are expressed in osteoprogenitor cells andimportant for osteoblast maturation [18] We also observedthat periostin (POSTN) gene was highly upregulated inCL1 cells (156-fold) POSTN is a 90 kDa secreted proteinoriginally identified in murine osteoblast-like cells and isupregulated by PTH [19] Several studies employing murineand human cells have revealed important role of POSTNin osteoblast differentiation and during development inintramembranous ossification [20ndash23] Another factor iden-tified in CL1 cells is nephroblastoma overexpression (NOV)which is a member of the Cyr 61 connective tissue growthfactor (CNN) family The CCN family of proteins promotesosteoblast differentiation through interaction with integrinsWNT BMP and NOTCH signaling pathways [24ndash26] Inaddition a large number of signaling molecules known tobe regulators of hBMSC lineage specific differentiation forexample insulin-like growth factors [27ndash29] WNT [30ndash32]and MAPK [33ndash35] were enriched in CL1 cells In contrastCL2 expressed high levels of immune-related genes whichmay explain the poor differentiation response to osteoblastor adipocyte lineage In a recent study the authors usedtelomerized hBMSC and showed clearly a clonal populationthat had very low in vitro and in vivo differentiation abilityhowever they had enhanced immune-related features includ-ing high IL7 expression These nullipotent cells expressedCD317 which was associated with remarkably high basal levelexpression of factors with a proinflammatory and antiviralfunction [17] We observed that this molecular phenotypewas associated with distinct ultrastructural characteristicsof the cells In particular CL2 had abundant phagocyticvacuole microvilli and lysosomes features reminiscent of
ultrastructure of immune-regulatory cells Our data thussupport the increasingly recognized feature that hBMSCexhibit immune modulatory functions and a part of theinnate immune response [17]
We observed that ALP protein expression and enzymaticactivity were significantly different between CL1 and CL2cell lines and were thus a potential marker that distinguishesdifferent cell populations with progenitor functions (CL1)from cells with nonprogenitor functions (CL2) ALP isexpressed in a wide variety of tissues including kidneysbone and liver [36 37] but tissue-nonspecific ALP (ALPL)is considered a commitment marker for osteoblastic lineage[13 38] However in a recent study the authors examined thedifferentiation potential of a number of hMSC clones in vitroand in vivo and reported that the hMSC clones with highlevels of ALP expression were committed to trilineage differ-entiation [13] Our data corroborate and extend these findingsby reporting the effects of siRNA-mediated inhibition ofALP that resulted in an impaired hBMSC differentiation notonly to osteoblasts but also to adipocytes Also our resultscorroborate earlier studies that demonstrated in human bonebiopsies the presence of ALP expression in bone marrowadipocytic cells [13 17] All these data suggest that ALP isa ldquostemnessrdquo marker of hBMSC and not just an indicator ofosteoblastic lineage commitment
While CL1 and CL2 were isolated from telomerizedhMSC cell line they are relevant to normal human physi-ology We observed that the molecular phenotypes of CL1and CL2 were contained within the molecular signatureof primary hBMSC suggesting that CL1 and CL2 repre-sent cell populations within the heterogeneous cultures ofhBMSCWe have also previously reported that the molecular
14 Stem Cells International
ALZ
stai
ning
Control siRNA ALP siRNA
ALP
Beta
-act
in
Con
trol
ALP
siRN
A d3
ALP
siRN
A d7
Nile red quantificationlowast
NI ADIP I ALP siRNA0
50
100
150
200
Con
trol s
iRN
AA
LP si
RNA
Day 3 Day 5 OS induction Day 7 OS induction
Fold
indu
ctio
n lo
g sc
ale
10000
1000
100
10
1
Day 0 OS day 7
Control siRNAALP siRNA
(A) (B)
(C)
(D)
Figure 4 Effect of alkaline phosphatase (ALP) gene silencing by small interfering RNA (siRNA) on a clonal population of human bonemarrow stromal stem cell CL1 (A) ALP staining at day 3 day 5 or day 7 days of osteoblast differentiation (OB induction) (B) Quantitativereal-time PCR for ALP gene following ALP siRNA transfection at day 0 OB and day 7 of OB Data are presented as fold change in expressionof each target gene normalized to GAPDH (119899 = three independent experiments 119901 lt 001) Western blotting analysis of day 3 and day 7after siRNA ALP transfection of CL1 cells ALPL specific antibody and B-actin was used (C) Mineralized matrix formation as visualizedby Alizarin Red S staining in siRNA transfected CL1 cells after 14 days of OB induction (D) Nile red quantification of mature lipid filledadipocyte in control noninduced (Cont) adipocyte induced (Adip I) and ALP siRNA transfected cells that are adipocyte induced (ALPsiRNA) Adipocyte induction was carried out for 7 days lowast119901 lt 005
phenotype and cellular responses of hMSC-TERT are similarto those of primary hMSC [10]While we have identified ALPas a marker that can be used for a prospective identificationof differentiation committed population of hBMSC we iden-tified additional distinctive molecular markers of the cellsFor example IGF-1 IGF-2 and IGF binding protein 5 wereenriched in CL1 compared to CL2 IGFs and their bindingproteins are very well-studied factors that play a role inhBMSC proliferation and osteoblast differentiation [18] On
the other hand annexin A3 as well as several immune-relatedgenes was highly enriched in CL2 compared to CL1 Futurestudies are needed to determine the functional significanceof these molecules in relation to the functional identity ofvarious cell populations within the hBMSC cultures andtheir usefulness as biomarkers to dissect the heterogeneouspopulation of cultured hBMSC
Our finding of the presence of functional diversity withinhBMSC cultures that contain progenitor and nonprogenitor
Stem Cells International 15
Table 4 Microarray data analysis showing genes found to be upregulated during adipogenic differentiation and downregulated after ALPKO
62 common elements inldquoAD uprdquo and ldquoALP downrdquo Gene name FC (ALP siRNA versus
control siRNA)APOBEC3G Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3G minus15938025IFI44L Interferon-induced protein 44-like minus11520283PAQR5 Progestin and adipoQ receptor family member V minus6868241PNMA2 Paraneoplastic antigen MA2 minus59695344DUSP23 Dual specificity phosphatase 23 minus54786854CLDN23 Claudin 23 minus51885047ANKDD1A Ankyrin repeat and death domain containing 1A minus51646647IL8 Interleukin 8 minus4887188LRRC23 Leucine rich repeat containing 23 minus47611775IL6 Interleukin 6 (interferon beta 2) minus4693139LIFR Leukemia inhibitory factor receptor alpha minus46540866PTGFR Prostaglandin F receptor (FP) minus4457529FAM134B Family with sequence similarity 134 member B minus4403495CYFIP2 Cytoplasmic FMR1 interacting protein 2 minus4260462METTL7A Methyltransferase-like 7A minus40480843APOBEC3F Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3F minus39502614CA5B Carbonic anhydrase VB mitochondrial minus393889ITGA10 Integrin alpha 10 minus39143775FMO3 Flavin containing monooxygenase 3 minus3852087IMPA2 Inositol monophosphatase 2 (human) minus38374884CDO1 Cysteine dioxygenase type I minus38181455CCDC68 Coiled-coil domain containing 68 minus37292893
HEXDC Hexosaminidase (glycosyl hydrolase family 20 catalytic domain) containing minus2094836SARM1 Sterile alpha and TIR motif containing 1 minus20797038BBS1 Bardet-Biedl syndrome 1 minus20146718SERPING1 Serpin peptidase inhibitor clade G (C1 inhibitor) member 1 minus20102212FAM162A Family with sequence similarity 162 member A minus2005807TCTN1 Tectonic family member 1 minus20033443
AD up ALP down
1006 62 1145
(a)
140
120
100
080
060
040
020
mdash
ALPL PPARg aP2 LPL
lowastlowastlowastlowastlowastlowast
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
lowast
(b)
Figure 5 Adipocyte related genes downregulated in ALP knockdown CL1 cells (a) Venn diagram of whole gene expression analysis of ADupregulated genes compared with ALP KD downregulated genes (b) Quantitative real-time PCR of four selected common genes from Venndiagram including ALPL PPARg2 aP2 and LPL Data are presented as fold change in expression of each target gene normalized to GAPDH(119899 = three independent experiments lowast119901 lt 005 lowastlowast119901 lt 001 and lowastlowastlowast119901 lt 0001)
Stem Cells International 17
cell populations has a clinical relevance It demonstrates thatthe progenitor function and the immune modulatory rolesof hBMSC [39] are mediated by specific and distinguishablepopulations of hBMSC Thus future clinical studies employ-ing hBMSC should attempt to administer the relevant sub-population of hBMSC dependent on the experimental aim asa novel approach to improving the clinical efficiency insteadof the current use of heterogeneous hBMSC populations
Competing Interests
The authors declare that they have no competing interests
Authorsrsquo Contributions
Mona Elsafadi and Amer Mahmood contributed equally tothis work
Acknowledgments
These studies are supported by a grant from the College ofMedicine Research Centre (CMRC) Deanship of ScientificResearch King Saud University Riyadh Saudi Arabia Theauthors would like to thank Mr Ali Al-Roalle for all histechnical help They thank Natalie Twine for advice on DNAmicroarray analysis
References
[1] W Zaher L Harkness A Jafari and M Kassem ldquoAn update ofhuman mesenchymal stem cell biology and their clinical usesrdquoArchives of Toxicology vol 88 no 5 pp 1069ndash1082 2014
[2] S Post B M Abdallah J F Bentzon andM Kassem ldquoDemon-stration of the presence of independent pre-osteoblastic andpre-adipocytic cell populations in bone marrow-derived mes-enchymal stem cellsrdquo Bone vol 43 no 1 pp 32ndash39 2008
[3] K H Larsen C M Frederiksen J S Burns B M Abdallahand M Kassem ldquoIdentifying a molecular phenotype for bonemarrow stromal cells with in vivo bone-forming capacityrdquoJournal of Bone and Mineral Research vol 25 no 4 pp 796ndash808 2010
[4] A Rosenzweig ldquoCardiac cell therapymdashmixed results frommixed cellsrdquoThe New England Journal of Medicine vol 355 no12 pp 1274ndash1277 2006
[5] S A Kuznetsov P H Krebsbach K Satomura et al ldquoSingle-colony derived strains of human marrow stromal fibroblastsform bone after transplantation in vivordquo Journal of Bone andMineral Research vol 12 no 9 pp 1335ndash1347 1997
[6] M Kassem and P Bianco ldquoSkeletal stem cells in space and timerdquoCell vol 160 no 1-2 pp 17ndash19 2015
[7] A I Caplan and D Correa ldquoPDGF in bone formation andregeneration new insights into a novel mechanism involvingMSCsrdquo Journal of Orthopaedic Research vol 29 no 12 pp 1795ndash1803 2011
[8] K Stewart S Walsh J Screen et al ldquoFurther characterizationof cells expressing STRO-1 in cultures of adult human bonemarrow stromal cellsrdquo Journal of Bone and Mineral Researchvol 14 no 8 pp 1345ndash1356 1999
[9] J L Simonsen C Rosada N Serakinci et al ldquoTelomeraseexpression extends the proliferative life-span and maintains
the osteogenic potential of human bone marrow stromal cellsrdquoNature Biotechnology vol 20 no 6 pp 592ndash596 2002
[10] M Al-Nbaheen R Vishnubalaji D Ali et al ldquoHuman stromal(mesenchymal) stem cells from bone marrow adipose tissueand skin exhibit differences in molecular phenotype and differ-entiation potentialrdquo Stem Cell Reviews and Reports vol 9 no 1pp 32ndash43 2013
[11] C A Gregory W G Gunn A Peister and D J Prockop ldquoAnAlizarin red-based assay of mineralization by adherent cells inculture comparison with cetylpyridinium chloride extractionrdquoAnalytical Biochemistry vol 329 no 1 pp 77ndash84 2004
[12] AMahmood L Harkness H D Schroslashder B M Abdallah andM Kassem ldquoEnhanced differentiation of human embryonicstem cells to mesenchymal progenitors by inhibition of TGF-120573activinnodal signaling using SB-431542rdquo Journal of Bone andMineral Research vol 25 no 6 pp 1216ndash1233 2010
[13] H-J Prins A K Braat D Gawlitta et al ldquoIn vitro inductionof alkaline phosphatase levels predicts in vivo bone formingcapacity of human bone marrow stromal cellsrdquo Stem CellResearch vol 12 no 2 pp 428ndash440 2014
[14] K Stenderup J Justesen C Clausen and M Kassem ldquoAgingis associated with decreased maximal life span and acceleratedsenescence of bone marrow stromal cellsrdquo Bone vol 33 no 6pp 919ndash926 2003
[15] D G Phinney G Kopen W Righter S Webster N Tremainand D J Prockop ldquoDonor variation in the growth propertiesand osteogenic potential of human marrow stromal cellsrdquoJournal of Cellular Biochemistry vol 75 no 3 pp 424ndash436 1999
[16] C M Digirolamo D Stokes D Colter D G Phinney R Classand D J Prockop ldquoPropagation and senescence of humanmarrow stromal cells in culture a simple colony-forming assayidentifies samples with the greatest potential to propagate anddifferentiaterdquo British Journal of Haematology vol 107 no 2 pp275ndash281 1999
[17] S James J Fox F Afsari et al ldquoMultiparameter analy-sis of human bone marrow stromal cells identifies distinctimmunomodulatory and differentiation-competent subtypesrdquoStem Cell Reports vol 4 no 6 pp 1004ndash1015 2015
[18] B Peruzzi A Cappariello A Del Fattore N Rucci F DeBenedetti and A Teti ldquoc-Src and IL-6 inhibit osteoblastdifferentiation and integrate IGFBP5 signallingrdquo Nature Com-munications vol 3 article 630 2012
[19] D Fortunati S Reppe A-K Fjeldheim M Nielsen V TGautvik and K M Gautvik ldquoPeriostin is a collagen associatedbonematrix protein regulated by parathyroid hormonerdquoMatrixBiology vol 29 no 7 pp 594ndash601 2010
[20] A Oshima H Tanabe T Yan G N Lowe C A Glackin andA Kudo ldquoA novel mechanism for the regulation of osteoblastdifferentiation transcription of periostin a member of thefasciclin I family is regulated by the bHLH transcription factorTwistrdquo Journal of Cellular Biochemistry vol 86 no 4 pp 792ndash804 2002
[21] T G Kashima T Nishiyama K Shimazu et al ldquoPeriostina novel marker of intramembranous ossification is expressedin fibrous dysplasia and in c-Fos-overexpressing bone lesionsrdquoHuman Pathology vol 40 no 2 pp 226ndash237 2009
[22] CGalliM PiergianniM Piemontese et al ldquoPeriostin improvescell adhesion to implantable biomaterials and osteoblasticdifferentiation on implant titanium surfaces in a topography-dependent fashionrdquo Journal of Biomedical Materials Research -Part A vol 102 no 11 pp 3855ndash3861 2014
18 Stem Cells International
[23] D L Coutu H W Jian A Monette G-E Rivard M DBlostein and J Galipeau ldquoPeriostin amember of a novel familyof vitamin K-dependent proteins is expressed by mesenchymalstromal cellsrdquo The Journal of Biological Chemistry vol 283 no26 pp 17991ndash18001 2008
[24] T Minamizato K Sakamoto T Liu et al ldquoCCN3NOV inhibitsBMP-2-induced osteoblast differentiation by interacting withBMP andNotch signaling pathwaysrdquo Biochemical and Biophysi-cal Research Communications vol 354 no 2 pp 567ndash573 2007
[25] W Si Q Kang H H Luu et al ldquoCCN1Cyr61 is regulated bythe canonicalWnt signal and plays an important role inWnt3A-induced osteoblast differentiation of mesenchymal stem cellsrdquoMolecular and Cellular Biology vol 26 no 8 pp 2955ndash29642006
[26] Q Luo Q Kang W Si et al ldquoConnective tissue growth factor(CTGF) is regulated by Wnt and bone morphogenetic proteinssignaling in osteoblast differentiation of mesenchymal stemcellsrdquo The Journal of Biological Chemistry vol 279 no 53 pp55958ndash55968 2004
[27] Y-S Yoon N Lee and H Scadova ldquoMyocardial regenerationwith bone-marrow-derived stem cellsrdquo Biology of the Cell vol97 no 4 pp 253ndash263 2005
[28] Y S Yoon ldquoCardiac regeneration with novel bone marrow-derived multipotent stem cellsrdquo Discovery Medicine vol 5 pp204ndash208 2005
[29] Y-S Yoon A Wecker L Heyd et al ldquoClonally expanded novelmultipotent stem cells from human bone marrow regeneratemyocardium after myocardial infarctionrdquo Journal of ClinicalInvestigation vol 115 no 2 pp 326ndash338 2005
[30] GM BolandG Perkins D J Hall andR S Tuan ldquoWnt 3a pro-motes proliferation and suppresses osteogenic differentiationof adult human mesenchymal stem cellsrdquo Journal of CellularBiochemistry vol 93 no 6 pp 1210ndash1230 2004
[31] D A Cook S W Fellgett M E Pownall P J OrsquoShea and PG Genever ldquoWnt-dependent osteogenic commitment of bonemarrow stromal cells using a novel GSK3120573 inhibitorrdquo Stem CellResearch vol 12 no 2 pp 415ndash427 2014
[32] J De Boer R Siddappa C Gaspar A VanApeldoorn R Foddeand C Van Blitterswijk ldquoWnt signaling inhibits osteogenicdifferentiation of humanmesenchymal stem cellsrdquoBone vol 34no 5 pp 818ndash826 2004
[33] MAl-Toub A AlmusaM Almajed et al ldquoPleiotropic effects ofcancer cellsrsquo secreted factors on human stromal (mesenchymal)stem cellsrdquo Stem Cell Research and Therapy vol 4 article 1142013
[34] L Chen D Qanie A Jafari et al ldquoDelta-like 1fetal antigen-1(Dlk1FA1) is a novel regulator of chondrogenic cell differenti-ation via inhibition of the Akt kinase-dependent pathwayrdquoTheJournal of Biological Chemistry vol 286 no 37 pp 32140ndash321492011
[35] T Matsunobu K Torigoe M Ishikawa et al ldquoCritical roles ofthe TGF-120573 type I receptor ALK5 in perichondrial formationand function cartilage integrity and osteoblast differentiationduring growth plate developmentrdquo Developmental Biology vol332 no 2 pp 325ndash338 2009
[36] B C Nair D E Johnson R J Majeska J A Rodkey C DBennett and G A Rodan ldquoRat alkaline phosphatase II Struc-tural similarities between the osteosarcoma bone kidney andplacenta isoenzymesrdquo Archives of Biochemistry and Biophysicsvol 254 no 1 pp 28ndash34 1987
[37] P J Butterworth ldquoAlkaline phosphatase Biochemistry of mam-malian alkaline phosphatasesrdquo Cell Biochemistry and Functionvol 1 no 2 pp 66ndash70 1983
[38] A T Ali C B Penny J E Paiker et al ldquoAlkaline phosphataseis involved in the control of adipogenesis in the murinepreadipocyte cell line 3T3-L1rdquo Clinica Chimica Acta vol 354no 1-2 pp 101ndash109 2005
[39] G M Spaggiari and L Moretta ldquoCellular and molecularinteractions of mesenchymal stem cells in innate immunityrdquoImmunology and Cell Biology vol 91 no 1 pp 27ndash31 2012
mineralized nodules became apparent and were stained withAlizarin Red S and ALP
26 In Vitro Adipocyte Differentiation Cells were grown instandard DMEM growth medium in 6-well plates at 03times 106 cellsmL At 90ndash100 confluence cells were culturedin DMEM supplemented with adipogenic (Adip) inductionmixture containing 10 FBS 10 Horse Serum (Sigma)1 Pen-strep 100 nM dexamethasone 045mM isobutylmethylxanthine (Sigma) 3120583gmL insulin (Sigma) and 1 120583MRosiglitazone (Novo Nordisk Bagsvaerd Denmark) Themedia were replaced three times per week Cells cultured instandard culture medium were considered as control Fromday 3 of differentiation small lipid droplets became visibleand at day 7 they were stained with Oil Red-O and Nile red
27 In Vitro Chondrogenic Differentiation Both CL1 and CL2cells were trypsinized and counted around 1 times 106 cellstaken in each 15mL conical tube centrifuged at 400timesg for5 minutes For chondrocyte differentiation pellet culturesystem used chondrocyte induction was done in mediacontaining advanced DMEMF12 supplemented with 1 ITSPremix Tissue Culture Supplement 100 nM dexamethasoneGlutamax and 10 ngmL transforming growth factor-beta-3(TGF120573-3) Cells were maintained in chondrocyte differentia-tion media for 21 days and changed every two days
28 Cytochemical Staining281 Alkaline Phosphatase (ALP) Staining CL1 and CL2 cellswere stained before OS differentiation for the basal ALPexpression and after OB differentiation at day 7 of inductionCells cultured in 6-well plates were washed in PBS minusminus(-Ca -Mg) andfixed in acetonecitrate buffer 10mMat pH42for 5min at room temperature The NaphtholFast Red stain[02mgmL Naphthol AS-TR phosphate substrate (Sigma)][0417mgmL of Fast Red (Sigma)] was added for one hourat room temperature
Histological tissue blocks were sectioned at 4 micronsImmunohistochemical staining was performed on CL1 andCL2 chondrocyte 3D pellets using DAKO EnVision andPowerVision according to the manufacturerrsquos instructions(DAKO Glostrup Denmark) Briefly paraffin sections wereincubated for 1 hour at room temperature with primaryantibodies diluted in ChemMate (DAKO) (HumanAnti-Col-10 and Human Anti-Col-2 ABI) Sections were washed sub-sequently in Tris-buffered saline (TBS 005M pH 74) incu-bated for 30 minutes with secondary anti-mouse IgHRP-conjugated polymers (K4001 En Visionthorn DAKO) and visu-alizedwith 330-diaminobenzidine tetrahydrochloride (DABS3000 DAKO) or with 3-amino-9-ethylcarbazole (AECDAKO) according to manufacturerrsquos instruction Controlswere performed with nonimmune immunoglobulins of thesame isotype as the primary antibodies (negative controls)and processed under identical conditions Alcian blue stain-ing was used to detect chondrocytes Sections of paraffin-embedded implants were stained with Alcian blue (Sigma)solution pH 25 at this pH all the glycoproteins (neutral andacidic) will be stained blue
29 Alizarin Red S Staining for Mineralized Matrix Seven-day-old OS differentiated cells in 6-well plates were used forAlizarin Red S staining The cell layer was washed with PBSand then fixed with 70 ice-cold ethanol for 1 hr at minus20∘CAfter removing the ethanol the cell layer was rinsed withdistilled water and stained with 40 nM AR-S (Sigma) pH 42for 10 minutes at room temperature Excess dye was washedoff with water followed by a wash with PBS for few minutesto minimize nonspecific AR-S stain
For quantifying the Alizarin Red S staining the air-driedplates the Alizarin Red S dye was eluted in 800 120583L of aceticacid incubated in each well for 30 minutes at room temper-ature as described [11] and measured in spectrophotometer(BioTek Epoch) at 405 nm
210 Quantitative ALP Activity To quantify ALP activity inCL1 and CL2 hBMSC before and after OS differentiationwe used the BioVision ALP activity colorimetric assay kit(BioVision Inc CA USA) with some modifications Cellswere cultured in 96-well plates under normal conditionsthen on day of analysis wells were rinsed once with PBSand were fixed using 37 formaldehyde in 90 ethanol for30 seconds at room temperature Subsequently fixative wasremoved and 50120583L of pNPP solution was added to each welland incubated for 1 hour in the dark at room temperatureThe reaction was subsequently stopped by adding 20 120583L stopsolution and gently shaking the plate ODwas thenmeasuredat 405 nm
211 Oil Red-O Staining for Lipid Droplets CL1 and CL2 cellsdifferentiated to adipocytes with Adip induction media atday 7 were used Accumulated cytoplasmic lipid dropletswere visualized by staining with Oil Red-O After washingcells grown in 6-well plates with PBS the cells were fixedin 4 formaldehyde for 10min at room temperature andthen rinsed once with 3 isopropanol and stained for 1 hr atroom temperature with filtered Oil Red-O staining solution(prepared by dissolving 05 g Oil Red-O powder in 60isopropanol) To quantify staining of fat droplets Oil Red-O was used as a stain Oil Red-O was eluted by adding 100isopropanol to each well and color changes were measuredby spectrophotometer at 510 nm (BioTek SpectrophotometerEpoch)
212 Nile Red Fluorescence Determination and Quantificationof Adipogenesis A stock solution of Nile red (1mgmL) inDMSO was prepared and stored at minus20∘C protected fromlight Staining was performed on unfixed cells Culturedundifferentiated and day 7 adipocyte differentiated cells weregrown in Corning polystyrene flat bottom 96-well TC-treated black microplates (Corning NY USA) were washedonce with PBS The dye was then added directly to the cells(5 120583gmL in PBS) and the preparation was incubated for10min at room temperature and then washed twice with PBSFluorescent signal was measured using SpectraMaxM5 fluo-rescence spectrophotometer plate reader (Molecular DevicesCo Sunnyvale CA USA) using bottom well-scan modewhere nine readings were taken per well using Ex (485 nm)and Em (572 nm) spectra
4 Stem Cells International
213 Quantitative Real-Time PCR (qRT-PCR) Analysis TotalRNA was extracted using MagNA pure compact RNA iso-lation kit (Roche Applied Science Germany Cat number04802993001) in an automatedMagNA pure compact system(Roche Germany) as recommended by the manufacturerThe total RNAwas quantified by Nanodrop spectrophotome-ter (Nanodrop 2000 Thermo Scientific USA) Complemen-tary DNA (cDNA) was synthesized from 1 120583g of the RNAsamples usingHighCapacity cDNAReverse Transcription kit(Applied Biosystems USA) using Labnet Multigene thermo-cycler according to the manufacturerrsquos instructions Relativelevels of mRNA were determined from cDNA by real-timePCR (Applied Biosystems-Real-Time PCRDetection System)with Power SYBR Green PCR kit (Applied BiosystemsUK) according to the manufacturerrsquos instructions Followingnormalization to the reference gene GAPDH quantificationof gene expression was carried out using a comparative Ctmethod where ΔCt is the difference between the CT valuesof the target and the reference gene and fold induction wasperformed from the control (Cont) for the same time pointPrimers (Supplementary Table 1 in Supplementary Materialavailable online at httpdxdoiorg10115520169378081)were obtained from Applied Biosystems (USA) as TAQMANprimers or previously published primers were used (seeSupplementary Table 1)
214 DNAMicroarray Global Gene Expression Analysis Fourhundred ng of total RNA was used as input for gener-ating biotin-labeled cRNA (Ambion Austin TX UnitedStates) cRNA samples were then hybridized onto Illuminahuman-8 BeadChips version 3 Hybridization washing Cy3-streptavidin staining and scanning were performed on theIllumina BeadStation 500 platform (Illumina San DiegoCA USA) according to the manufacturerrsquos instructions andeverything was done in triplicate Expression data analysiswas carried out using the Partek genomic suite softwareRaw data were background-subtracted normalized usingthe ldquorank invariantrdquo algorithm and filtered for significantexpression on the basis of negative control beads Geneswere considered significantly expressed with detection 119901values le 001 Differential expression analysis was performedwith the Illumina custom method using freshly isolatedprimary hBMSC (used at passage 3) as a reference controlThe following parameters were set to identify statisticalsignificance differential 119901 values le 001 fold change ratiogt 15 Pathway analysis was performed using DAVID Bioin-formatics Resources 67 (httpdavidabccncifcrfgov) andGeneSpring GX software (Agilent Technologies) Pathwayanalysis for CL1 OS D14 versus CL2 OS D14 was conductedusing the Single Experiment Pathway analysis feature inGeneSpring 120 (Agilent Technologies)
215 Small Interfering (si)RNATransfection For transfectionhBMSC in logarithmic growth phase were transfected withSilencer Select Predesigned ALP siRNA (25 nM) (Assay IDs1298 and Cat number 4390824) (Ambion The RNA Com-pany USA) using Lipofectamine RNAiMAX Reagent (Invit-rogen CA USA) plus serum-free Opti-MEMI mediumunder the conditions described by the manufacturer At
day 3 of transfection the cells were induced for osteogenicdifferentiation for an additional 7 days ALP staining wasused as a control for the siRNA transfection efficiency andtimeline
216 Statistical Analysis All of the results were presentedas the mean and standard deviation (SD) of at least 3independent experiments with 3ndash5 technical repeats in eachexperiment Studentrsquos 119905-test (two-tailed) was used for testingdifferences between groups 119901 value lt005 was consideredstatistically significant
3 Results
31 Comparison between CL1 and CL2 Differences in Mor-phology Proliferation and Marker Expression Profile Weisolated two distinct clonal cell populations of hBMSC-TERThBMSC-CL1 and hBMSC-CL2 (for easiness will be termedhereafter CL1 and CL2) based on differences in cell morphol-ogy (Figure 1(a)) CL1 cells had cuboidalmorphologywhereasCL2 cells have spindle-shaped fibroblast-like morphologyCL1 cells had higher proliferation rate compared to CL2(Figure 1(b)) mean PD rates of CL1 and CL2 were 0714 and0429 PDday respectively (Figure 1(b)) Both CL1 and CL2expressed surface marker profiles characteristics of hBMSC(gt90) CD44+ CD63+ CD73+ CD105+ and CD166+(Figure 1(c)) However CL1 cells showed higher expressionof CD146 (927 versus 12) and ALP (98 versus 0)compared to CL2 (Figure 1(d)) TEM revealed the presenceof abundant pseudopodia in CL1 indicating high motility(Figure 1(e)(A)) as well as well-developed mitochondriaand rough endoplasmic reticulum (rER) suggesting highmetabolic activity CL2 cells contained abundant phagocyticvacuole (pv)microvilli (mi) and lysosomes (ly) (Figure 1(e))
We performed quantitative real-time PCR (RT-PCR)for genes expressed in mesodermal progenitor cells [12]CL1 expressed higher levels of BMP4 MIXL1 WNT3a andTWIST compared to CL2 (Figure 1(f) 119901 lt 001) In contrastCL2 expressed higher levels of Kinase Insert Domain Recep-tor (Type III Receptor Tyrosine Kinase) (KDR) expressed inendothelial cells and smoothmusclemyosin heavy chain gene(smMHC) expressed in smooth muscle cells (Figure 1(f))
32 CL1 Cells Exhibit Enhanced Osteoblast DifferentiationFollowing osteoblast (OB) differentiation induction ALPstaining andALP enzymatic activity were significantly higherin CL1 compared to CL2 cells (Figure 2(a) 119901 lt 001)Similarly Alizarin Red staining and quantitation of formedmineralized matrix were more pronounced in CL1 cells (Fig-ure 2(b) 119901 lt 001) In addition CL1 cells expressed higherlevels of osteoblastic genes ALP RUNX2 and osteopontin(OPN) (Figure 2(c) upper panel) compared to CL2 cells
Global gene expression microarray analysis of OB dif-ferentiated cells at day 14 showed around 1060 genes signif-icantly upregulated more than 2-fold (119901 lt 001) in CL1Among the upregulated genes 80 genes were annotated tobone development and osteoblast differentiation (Table 1)The highest upregulated genes included paired-like home-odomain 2 (PITX2) Insulin-like growth factor 1 (IGF1) and
Stem Cells International 5
CL1
CL2
(a)
Cum
ulat
ive P
D
Cumulative days
Growth curve
0 10 20 30 40
CL1 CL2
CL1 CL20
5
10
15
20
25
0
01
02
03
04
05
06
07
08
(PD
day
)
(b)
CL2
103
102
101
100
103
102
101
100
103
102
101
100
103
102
101
100
103
102
101
100
103
101
100
103
102
102
101
100
103
102
101
100
103
102
101
100
103
102
101
100
CL1
CD44 CD63 CD73 CD166 CD105
T3 CD44 T3 CD63 T3 CD73 T3 CD166 T3 CD105
0
20
40
60
80
Cou
nt
0
20
40
60
80
100
120
Cou
ntC
ount
Cou
nt
Cou
nt
Cou
nt
Cou
nt0
20
40
60
80
Cou
nt
0
20
40
60
80
100
Cou
nt
0
20
40
60
80
Cou
nt
80
60
40
20
0
100
50
0
120
100
80
60
40
20
0
120
100
80
60
40
20
0
100
80
60
40
20
0
(c)
ISO CD146 ALP
ISO CD146 ALP
FL2-
H
FL2-
H
FL2-
HPEPEPECL
2CL
1
FL1-H FL1-H FL1-H
FITCFITCFITC
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
(d)
Figure 1 Continued
6 Stem Cells International
(A) (B)
(e)
Fold
indu
ctio
n
CL1CL2
lowast
lowastlowast lowast
BMP4 MIXL1 WNT3a TWIST
lowastlowast
KDR smMHCminus01
01
03
05
07
09
11
13
15
mdash
050
100
150
200
250
300
CL1CL2
(f)
Figure 1 Cellular phenotype of two clonal populations of humanbonemarrow stromal stem cells CL1 andCL2 (a) Cellmorphology CL1 cellsshowed large cuboidal morphology while CL2 cells had spindle-shaped fibroblast-like morphology (b) Growth curves showing populationdoubling (PD) rate during long term culture (c) Flow cytometry analysis (FACS) shows expression of CD44 CD63 CD73 CD105 andCD166in CL1 and CL2 cells Matched isotype control was used for gating (d) Flow cytometry analysis presented as dot blot of CD146 and alkalinephosphatase (ALP) cell surface proteins (e) Transmission electron microscope (TEM) (A) CL1 (1200x) (B) CL2 (1500x) n nucleus nunucleolus rER rough endoplasmic reticulum ly lysosomes pv phagocytic vacuole and rer reticular stalk of rER (f) Gene expressionanalysis using RT-PCR for a group of mesodermal and stromal genes Gene expression was normalized to GAPDH and presented as foldchange Data is shown as mean plusmn SD of three independent experiments lowast119901 lt 005 119901 lt 0001
collagen type V alpha 3 (COL5A3) osteomodulin (OMD)and T-box 15 (TBX15) (Table 1) Furthermore several knownosteoblast-related genes were upregulated in CL1 cells suchas bone morphogenetic protein 6 (BMP6) fibroblast growthfactor receptor 3 (FGFR3) insulin-like growth factor bindingprotein 5 (IGFBP5) and vitamin D (125-dihydroxyvitaminD3) receptor (VDR) (Table 1) On the other hand 1200 geneswere upregulated in CL2 cells 255 genes were annotated toimmunity and immune response and defense This categoryincluded genes from complement system chemokine (C-Cmotif) ligands interferon family chemokine (C-X-C motif)ligands and receptor major histocompatibility complex classII molecules interleukins and tumor necrosis factor receptorsuperfamily (Table 2 and Supplementary Table 3)
33 CL1 Cells Exhibit Enhanced Adipocyte DifferentiationWe observed significant differences between CL1 and CL2in their response to adipocytic differentiation inductionCL1 differentiated readily to adipocytes compared to CL2(Figure 2(d) lower panel) evidenced by higher levels ofadipocyticmarkers gene expression LPL (lipoprotein lipase)and adiponectin as well as formation of mature lipid filledadipocytes visualized by Oil Red-O staining and quantitativeNile red staining (Figure 2(d))
34 CL1 Cells Differentiate to Chondrocytic Lineage Inpellet cultures CL1 cells formed 3D pellets containingproteoglycan-secreting chondrocytes which stained positivewith Alcian blue Limited chondrocyte differentiation wasvisible in cell pellets of CL2 cells The differentiated chon-drocytes in CL1 pellets expressed higher levels of collagen Xand collagen II which was overlapping the Alcian blue stain(Figure 2(e))
35 Molecular Signature of CL1 and CL2 Cells To define themolecular signature and molecular differences between CL1
and CL2 we compared the basal gene expression patternof CL1 and CL2 cells using DNA microarrays The PCAanalysis showed a clear separation between CL1 and CL2(Supplementary Figure 1) Comparison between CL1 withCL2 showed that 915 geneswere differentially expressed in thetwo cell lines (gt2-fold119901 lt 001) 462 genes were upregulatedand 452 were downregulated in CL1 versus CL2 The mostrelevant genes that were upregulated in CL1 are listed inTable 3(a) Among these 35 highly expressed genes in CL1 thefollowing 11 genes were present in skeletal and muscular sys-tem development and function FOLR3 CCL3L1 SERPINB2POSTN IGFBP5 CCL3 NOV ALP TNFRSF11B ACTG2and CDH11 (Table 3(a)) Functional annotation of the upreg-ulated genes in CL1 using the Ingenuity Pathway Analysis(IPA) revealed enrichment in the following categories ldquotissuedevelopmentrdquo ldquoskeletal and muscular system developmentand functionrdquo and ldquoorganismal developmentrdquo (Table 3(b))Furthermore the DAVID annotation tool was employed toassess the functional relationships of the upregulated genesin CL1 showing enrichment in ontologies ldquoskeletal andmuscular system development and functionrdquo that includedbone size osteoblast differentiation bone mineralizationand bone mineral density (Table 3(c)) CL1 exhibited upreg-ulation of WNT pathway ligands WNT5B (2-fold) andLRP5 (2-fold) (Table 1) Also ALP was among the highlyexpressed genes together with POSTN IGFBP5 SPP1 IL-6 and DKK1 (Tables 3(a) and 1) These genes are knownto play an important role in osteoblast differentiation andbone formation For CL2 inhibitors of WNT pathway wereupregulated and included SFRP1 (11-fold) DKK2 (32-fold)FGF2 (31-fold) and GBP2 (24-fold) Functional annotationof the upregulated genes in CL2 revealed enrichment in thefollowing categories ldquodevelopmental processrdquo ldquomulticellularorganismal processrdquo ldquobiological adhesionrdquo and ldquoimmunesystem processrdquo (Supplementary Table 2A) In-depth analysis
Stem Cells International 7
CL1
CL2
ALPCont D 14
Cont OB I
ALP activity
mdash100
200
300
400
500
600
Fold
chan
ge
CL1CL2
(a)
CL1
CL2
ALZCont D 14
Fold
chan
ge
25
20
15
10
5
0
Cont d14 d21
CL1CL2
ALZ quantification
(b)
CL1CL2
CL1CL2
LPL ADIPOQ
Cont ContADIP D14 ADIP D14
000
100
200
300
400
500
600
700
0
1000
2000
3000
4000
5000
CL1CL2
CL1CL2
CL1CL2
RUNX2 ALP OPN
Cont Cont ContOS D7 OS D7 OS D7
00
05
10
15
20
25
30
00
05
10
15
20
25
30
35
00
05
10
15
20
25
30
35
(c)
CL1CL2
CL1
CL2
Control Adip d7 Adip d14
d7 d140
005
01
015
02
025
03
035
04
045
Nile
red
fluor
esce
nt in
tens
ity
(d)
CL1
CL2
Alcian blue Col X Col II
(e)
Figure 2 In vitro osteoblastic and adipocytic differentiation of two clonal populations of human bone marrow stromal stem cells CL1 andCL2 cells Both cell lines were induced for osteoblast differentiation using standard protocol described in the Methods (a) ALP staining atday 14 in control noninduced (Cont) and osteoblast induced cells (D14) Right panel shows ALP activity (119899 = 3 independent experiments119901 lt 0001) (b) Mineralized matrix formation visualized by Alizarin Red S staining Right panel shows Alizarin Red quantification at day 14and 21 after osteoblast differentiation (119899 = 3 independent experiments 119901 lt 0001) (c) Quantitative RT-PCR of osteoblastic and adipocytegene markers in CL1 and CL2 during osteoblast (upper panel) and adipocyte (lower panel) differentiation ALP = alkaline phosphataseOPN = osteopontin LPL = lipoprotein lipase and ADIPOQ = adiponectin Data are presented as fold change in expression of each targetgene normalized to GAPDH (119899 = 3 independent experiments 119901 lt 005
119901 lt 0001) (d) CL1 and CL2 lines were induced for adipocytedifferentiation using standard protocol described in the Methods Adipocyte formation was visualized at day 7 (Adip d7) and day 14 (Adipd14) by Oil Red-O staining Lower panel presents quantification of Nile red staining (119899 = 3 independent experiments 119901 lt 0001) (e) CL1and CL2 lines were induced for chondrocyte differentiation using 21-day pellet culture method as described in the Methods The pellets werestained with Alcian blue collagen 10 (Col X) and collagen 2 (Col II) (original magnification 5x)
8 Stem Cells International
Table 1 Microarray data analysis showing genes related to bone development and osteoblast differentiation upregulated in CL1 versus CL2cells
Probe ID Genbank accession Gene name Gene symbol FCA 23 P167367 NM 153426 Paired-like homeodomain 2 PITX2 30884
of the biological processes revealed several immune-relatedpathways ldquoMAPKKK cascaderdquo ldquoimmunity and defenserdquoldquosignal transductionrdquo ldquoextracellular matrix protein-mediatedsignalingrdquo and ldquointerferon-mediated immunityrdquo among oth-ers that were upregulated (Supplementary Table 2B) Also40 genes related to immune system related factors wereidentified as significantly enriched in CL2 compared to CL1cells (Supplementary Table 2C)We chose the following genesfor validation of the microarray results NOV IGFBP5 ALPTAGLN and CDH11 as they were highly expressed in CL1RT-PCR analysis confirmed themicroarray results (Figure 3)
Furthermore we compared the molecular phenotype ofCL1 and CL2 cells with that of phBMSCWe found that morethan 80 of the genes expressed in CL1 and 90 in CL2 cellswere commonwith primary phBMSC (Supplementary Figure2) suggesting that CL1 and CL2 molecular phenotype existwithin the heterogeneous population of phBMSC cultures
36 ALP Knockdown Impairs Differentiation of CL1 CellsSince ALP has been suggested as a marker for hBMSCprogenitor cell lineage commitment [13] and was highlyupregulated in CL1 cells we tested its biological role in
CL1 cells ALP siRNA transfection decreased ALP proteinlevel ALP activity and mRNA gene expression compared tocontrol cells transfected with control siRNA (119901 lt 001) andthis inhibition was detectable up to day 7 days after osteoblastdifferentiation induction (Figures 4(A) and 4(B)) At day 14 ofdifferentiation mineralization ability of CL1 was significantlyimpaired (Figure 4(C)) In addition we found that thenumber of mature adipocyte formations was significantlyreduced to more than 75 (119901 lt 001) (Figure 4(D))
To identify relevant adipocyte differentiation associatedgenes that were targeted by ALP deficiency we compared thedownregulated genes of ALP deficient CL1 with the upreg-ulated genes identified during adipocytic differentiation ofCL1 We identified 62 genes that were common (Figure 5(a)Table 4) and among these genes were genes related tometabolism (primarily lipid and carbohydrate) and transportincluding CYB5B CHST1 TAP1 ATP8A1 LRP8 PLCD1and FABP5 (Table 4) We further performed quantitativereal-time PCR of ALP deficient CL1 cells during adipocytedifferentiation The following adipocyte-associated geneswere downregulated PPAR1205742 LPL and aP2 (Figure 5(b))
12 Stem Cells International
Table 3 Whole genome microarray analysis of two clonal popu-lations of human bone marrow stromal stem cells CL1 and CL2cells (a) Top 35 highly upregulated genes in CL1 cells versus CL2cells (b) Ingenuity Pathway Analysis (IPA) showing the differentphysiological system development and function genes found in eachcategory and the corresponding 119901 value (c) Upper part IPA analysisshowing the categories for skeletal and muscular system developmentand function upregulated in CL1 cells and lower part showing genesupregulated in CL2 cells NOV = nephroblastoma overexpressedIGFBP5 = insulin-like growth factor binding protein 5 ALP TAGLN= transgelin and CDH11 = OB-cadherin (osteoblast)
Name 119901 value moleculesPhysiological system development and function
Organismal development 705119864 minus 10ndash143119864 minus 03 181Embryonic development 100119864 minus 09ndash143119864 minus 03 154Organ development 100119864 minus 09ndash143119864 minus 03 145
(b) Continued
Name 119901 value moleculesSkeletal and muscularsystem development andfunction
100119864 minus 09ndash102119864 minus 03 123
Tissue development 100119864 minus 09ndash143119864 minus 03 236
(c)
Functions annotation 119901 value moleculesSkeletal and muscular system development and
function upregulated in CL1 cellsSize of bone 143119864 minus 06 24Differentiation ofosteoblasts 384119864 minus 06 25
Mineralization ofbone 493119864 minus 06 19
Bone mineral density 365119864 minus 05 19Skeletal and muscular system development and
function upregulated in CL2 cellsDevelopment ofmuscle 100119864 minus 09 44
Proliferation ofmuscle cells 229119864 minus 06 35
Remodeling of bone 363119864 minus 06 21Resorption of bone 394119864 minus 06 19
confirming impairment of adipocytic differentiation of ALPdeficient CL1 cells
4 Discussion
We extensively studied two cell populations within culturedhBMSC that were identified based on differences in mor-phology Cellular and molecular studies revealed differencesin growth differentiation capacity and molecular signatureOur data support the notion of the presence of cellular andfunctional heterogeneity among cultured hBMSC
Cellular heterogeneity of cultured hBMSC is recognizedin an increasing number of reports Several extrinsic andintrinsic factors may contribute to the observed hBMSCheterogeneity Extrinsic factors include donor-to-donor vari-ations in the number and quantity of initiating cells whichresult in differences in cell growth rate and differentia-tion capacity [14 15] Intrinsic factors have been examinedemploying single cell clonal analysis and revealed variationsin differentiation potential among individual colonies withinhBMSC cultures ranging from the presence of cells withtrilineage (osteoblast adipocytes and chondrocyte) potencyto cells with null potency [16] Also variations in the abilityof clonal cells to form heterotopic bone when implanted invivo have been reported [5] Our study corroborates thesefindings and provides more detailed cellular and molecularphenotyping of two examples of cell populations that existwithin the heterogeneous hBMSC cultures [17]
Determining the molecular signature of CL1 and CL2using whole genomemicroarray analysis showed enrichmentof lineage-commitment associated genes inCL1 For exampleinsulin-like growth factor 5 (IGFBP5) and interleukin 6 (IL6)
Stem Cells International 13
NOV0040
0030
0020
0010
mdash
IGFBP50015
0010
0005
mdash
ALP0100
0080
0060
0040
0020
mdash
TAGLN0800
0600
0400
0200
CDH11
0060
0040
0020
mdashmdash
LRP30015
0010
0005
mdash
CL1 CL2 CL1 CL2 CL1 CL2
CL1 CL2CL1 CL2CL1 CL2
Figure 3 Validation of whole genome microarray analysis of two clonal populations of human bone marrow stromal stem cells CL1 andCL2 cells Quantitative real-time PCR for highly expressed genes in CL1 cells NOV = nephroblastoma overexpressed IGFBP5 = insulin-likegrowth factor binding protein 5 ALP TAGLN = transgelin and CDH11 = OB-cadherin (osteoblast) Data are presented as fold change inexpression of each target gene normalized to GAPDH (119899 = three independent experiments 119901 lt 0001) (see also Table 3)
were 147- and 33-fold upregulated in CL1 cells respectivelyBoth factors are expressed in osteoprogenitor cells andimportant for osteoblast maturation [18] We also observedthat periostin (POSTN) gene was highly upregulated inCL1 cells (156-fold) POSTN is a 90 kDa secreted proteinoriginally identified in murine osteoblast-like cells and isupregulated by PTH [19] Several studies employing murineand human cells have revealed important role of POSTNin osteoblast differentiation and during development inintramembranous ossification [20ndash23] Another factor iden-tified in CL1 cells is nephroblastoma overexpression (NOV)which is a member of the Cyr 61 connective tissue growthfactor (CNN) family The CCN family of proteins promotesosteoblast differentiation through interaction with integrinsWNT BMP and NOTCH signaling pathways [24ndash26] Inaddition a large number of signaling molecules known tobe regulators of hBMSC lineage specific differentiation forexample insulin-like growth factors [27ndash29] WNT [30ndash32]and MAPK [33ndash35] were enriched in CL1 cells In contrastCL2 expressed high levels of immune-related genes whichmay explain the poor differentiation response to osteoblastor adipocyte lineage In a recent study the authors usedtelomerized hBMSC and showed clearly a clonal populationthat had very low in vitro and in vivo differentiation abilityhowever they had enhanced immune-related features includ-ing high IL7 expression These nullipotent cells expressedCD317 which was associated with remarkably high basal levelexpression of factors with a proinflammatory and antiviralfunction [17] We observed that this molecular phenotypewas associated with distinct ultrastructural characteristicsof the cells In particular CL2 had abundant phagocyticvacuole microvilli and lysosomes features reminiscent of
ultrastructure of immune-regulatory cells Our data thussupport the increasingly recognized feature that hBMSCexhibit immune modulatory functions and a part of theinnate immune response [17]
We observed that ALP protein expression and enzymaticactivity were significantly different between CL1 and CL2cell lines and were thus a potential marker that distinguishesdifferent cell populations with progenitor functions (CL1)from cells with nonprogenitor functions (CL2) ALP isexpressed in a wide variety of tissues including kidneysbone and liver [36 37] but tissue-nonspecific ALP (ALPL)is considered a commitment marker for osteoblastic lineage[13 38] However in a recent study the authors examined thedifferentiation potential of a number of hMSC clones in vitroand in vivo and reported that the hMSC clones with highlevels of ALP expression were committed to trilineage differ-entiation [13] Our data corroborate and extend these findingsby reporting the effects of siRNA-mediated inhibition ofALP that resulted in an impaired hBMSC differentiation notonly to osteoblasts but also to adipocytes Also our resultscorroborate earlier studies that demonstrated in human bonebiopsies the presence of ALP expression in bone marrowadipocytic cells [13 17] All these data suggest that ALP isa ldquostemnessrdquo marker of hBMSC and not just an indicator ofosteoblastic lineage commitment
While CL1 and CL2 were isolated from telomerizedhMSC cell line they are relevant to normal human physi-ology We observed that the molecular phenotypes of CL1and CL2 were contained within the molecular signatureof primary hBMSC suggesting that CL1 and CL2 repre-sent cell populations within the heterogeneous cultures ofhBMSCWe have also previously reported that the molecular
14 Stem Cells International
ALZ
stai
ning
Control siRNA ALP siRNA
ALP
Beta
-act
in
Con
trol
ALP
siRN
A d3
ALP
siRN
A d7
Nile red quantificationlowast
NI ADIP I ALP siRNA0
50
100
150
200
Con
trol s
iRN
AA
LP si
RNA
Day 3 Day 5 OS induction Day 7 OS induction
Fold
indu
ctio
n lo
g sc
ale
10000
1000
100
10
1
Day 0 OS day 7
Control siRNAALP siRNA
(A) (B)
(C)
(D)
Figure 4 Effect of alkaline phosphatase (ALP) gene silencing by small interfering RNA (siRNA) on a clonal population of human bonemarrow stromal stem cell CL1 (A) ALP staining at day 3 day 5 or day 7 days of osteoblast differentiation (OB induction) (B) Quantitativereal-time PCR for ALP gene following ALP siRNA transfection at day 0 OB and day 7 of OB Data are presented as fold change in expressionof each target gene normalized to GAPDH (119899 = three independent experiments 119901 lt 001) Western blotting analysis of day 3 and day 7after siRNA ALP transfection of CL1 cells ALPL specific antibody and B-actin was used (C) Mineralized matrix formation as visualizedby Alizarin Red S staining in siRNA transfected CL1 cells after 14 days of OB induction (D) Nile red quantification of mature lipid filledadipocyte in control noninduced (Cont) adipocyte induced (Adip I) and ALP siRNA transfected cells that are adipocyte induced (ALPsiRNA) Adipocyte induction was carried out for 7 days lowast119901 lt 005
phenotype and cellular responses of hMSC-TERT are similarto those of primary hMSC [10]While we have identified ALPas a marker that can be used for a prospective identificationof differentiation committed population of hBMSC we iden-tified additional distinctive molecular markers of the cellsFor example IGF-1 IGF-2 and IGF binding protein 5 wereenriched in CL1 compared to CL2 IGFs and their bindingproteins are very well-studied factors that play a role inhBMSC proliferation and osteoblast differentiation [18] On
the other hand annexin A3 as well as several immune-relatedgenes was highly enriched in CL2 compared to CL1 Futurestudies are needed to determine the functional significanceof these molecules in relation to the functional identity ofvarious cell populations within the hBMSC cultures andtheir usefulness as biomarkers to dissect the heterogeneouspopulation of cultured hBMSC
Our finding of the presence of functional diversity withinhBMSC cultures that contain progenitor and nonprogenitor
Stem Cells International 15
Table 4 Microarray data analysis showing genes found to be upregulated during adipogenic differentiation and downregulated after ALPKO
62 common elements inldquoAD uprdquo and ldquoALP downrdquo Gene name FC (ALP siRNA versus
control siRNA)APOBEC3G Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3G minus15938025IFI44L Interferon-induced protein 44-like minus11520283PAQR5 Progestin and adipoQ receptor family member V minus6868241PNMA2 Paraneoplastic antigen MA2 minus59695344DUSP23 Dual specificity phosphatase 23 minus54786854CLDN23 Claudin 23 minus51885047ANKDD1A Ankyrin repeat and death domain containing 1A minus51646647IL8 Interleukin 8 minus4887188LRRC23 Leucine rich repeat containing 23 minus47611775IL6 Interleukin 6 (interferon beta 2) minus4693139LIFR Leukemia inhibitory factor receptor alpha minus46540866PTGFR Prostaglandin F receptor (FP) minus4457529FAM134B Family with sequence similarity 134 member B minus4403495CYFIP2 Cytoplasmic FMR1 interacting protein 2 minus4260462METTL7A Methyltransferase-like 7A minus40480843APOBEC3F Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3F minus39502614CA5B Carbonic anhydrase VB mitochondrial minus393889ITGA10 Integrin alpha 10 minus39143775FMO3 Flavin containing monooxygenase 3 minus3852087IMPA2 Inositol monophosphatase 2 (human) minus38374884CDO1 Cysteine dioxygenase type I minus38181455CCDC68 Coiled-coil domain containing 68 minus37292893
HEXDC Hexosaminidase (glycosyl hydrolase family 20 catalytic domain) containing minus2094836SARM1 Sterile alpha and TIR motif containing 1 minus20797038BBS1 Bardet-Biedl syndrome 1 minus20146718SERPING1 Serpin peptidase inhibitor clade G (C1 inhibitor) member 1 minus20102212FAM162A Family with sequence similarity 162 member A minus2005807TCTN1 Tectonic family member 1 minus20033443
AD up ALP down
1006 62 1145
(a)
140
120
100
080
060
040
020
mdash
ALPL PPARg aP2 LPL
lowastlowastlowastlowastlowastlowast
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
lowast
(b)
Figure 5 Adipocyte related genes downregulated in ALP knockdown CL1 cells (a) Venn diagram of whole gene expression analysis of ADupregulated genes compared with ALP KD downregulated genes (b) Quantitative real-time PCR of four selected common genes from Venndiagram including ALPL PPARg2 aP2 and LPL Data are presented as fold change in expression of each target gene normalized to GAPDH(119899 = three independent experiments lowast119901 lt 005 lowastlowast119901 lt 001 and lowastlowastlowast119901 lt 0001)
Stem Cells International 17
cell populations has a clinical relevance It demonstrates thatthe progenitor function and the immune modulatory rolesof hBMSC [39] are mediated by specific and distinguishablepopulations of hBMSC Thus future clinical studies employ-ing hBMSC should attempt to administer the relevant sub-population of hBMSC dependent on the experimental aim asa novel approach to improving the clinical efficiency insteadof the current use of heterogeneous hBMSC populations
Competing Interests
The authors declare that they have no competing interests
Authorsrsquo Contributions
Mona Elsafadi and Amer Mahmood contributed equally tothis work
Acknowledgments
These studies are supported by a grant from the College ofMedicine Research Centre (CMRC) Deanship of ScientificResearch King Saud University Riyadh Saudi Arabia Theauthors would like to thank Mr Ali Al-Roalle for all histechnical help They thank Natalie Twine for advice on DNAmicroarray analysis
References
[1] W Zaher L Harkness A Jafari and M Kassem ldquoAn update ofhuman mesenchymal stem cell biology and their clinical usesrdquoArchives of Toxicology vol 88 no 5 pp 1069ndash1082 2014
[2] S Post B M Abdallah J F Bentzon andM Kassem ldquoDemon-stration of the presence of independent pre-osteoblastic andpre-adipocytic cell populations in bone marrow-derived mes-enchymal stem cellsrdquo Bone vol 43 no 1 pp 32ndash39 2008
[3] K H Larsen C M Frederiksen J S Burns B M Abdallahand M Kassem ldquoIdentifying a molecular phenotype for bonemarrow stromal cells with in vivo bone-forming capacityrdquoJournal of Bone and Mineral Research vol 25 no 4 pp 796ndash808 2010
[4] A Rosenzweig ldquoCardiac cell therapymdashmixed results frommixed cellsrdquoThe New England Journal of Medicine vol 355 no12 pp 1274ndash1277 2006
[5] S A Kuznetsov P H Krebsbach K Satomura et al ldquoSingle-colony derived strains of human marrow stromal fibroblastsform bone after transplantation in vivordquo Journal of Bone andMineral Research vol 12 no 9 pp 1335ndash1347 1997
[6] M Kassem and P Bianco ldquoSkeletal stem cells in space and timerdquoCell vol 160 no 1-2 pp 17ndash19 2015
[7] A I Caplan and D Correa ldquoPDGF in bone formation andregeneration new insights into a novel mechanism involvingMSCsrdquo Journal of Orthopaedic Research vol 29 no 12 pp 1795ndash1803 2011
[8] K Stewart S Walsh J Screen et al ldquoFurther characterizationof cells expressing STRO-1 in cultures of adult human bonemarrow stromal cellsrdquo Journal of Bone and Mineral Researchvol 14 no 8 pp 1345ndash1356 1999
[9] J L Simonsen C Rosada N Serakinci et al ldquoTelomeraseexpression extends the proliferative life-span and maintains
the osteogenic potential of human bone marrow stromal cellsrdquoNature Biotechnology vol 20 no 6 pp 592ndash596 2002
[10] M Al-Nbaheen R Vishnubalaji D Ali et al ldquoHuman stromal(mesenchymal) stem cells from bone marrow adipose tissueand skin exhibit differences in molecular phenotype and differ-entiation potentialrdquo Stem Cell Reviews and Reports vol 9 no 1pp 32ndash43 2013
[11] C A Gregory W G Gunn A Peister and D J Prockop ldquoAnAlizarin red-based assay of mineralization by adherent cells inculture comparison with cetylpyridinium chloride extractionrdquoAnalytical Biochemistry vol 329 no 1 pp 77ndash84 2004
[12] AMahmood L Harkness H D Schroslashder B M Abdallah andM Kassem ldquoEnhanced differentiation of human embryonicstem cells to mesenchymal progenitors by inhibition of TGF-120573activinnodal signaling using SB-431542rdquo Journal of Bone andMineral Research vol 25 no 6 pp 1216ndash1233 2010
[13] H-J Prins A K Braat D Gawlitta et al ldquoIn vitro inductionof alkaline phosphatase levels predicts in vivo bone formingcapacity of human bone marrow stromal cellsrdquo Stem CellResearch vol 12 no 2 pp 428ndash440 2014
[14] K Stenderup J Justesen C Clausen and M Kassem ldquoAgingis associated with decreased maximal life span and acceleratedsenescence of bone marrow stromal cellsrdquo Bone vol 33 no 6pp 919ndash926 2003
[15] D G Phinney G Kopen W Righter S Webster N Tremainand D J Prockop ldquoDonor variation in the growth propertiesand osteogenic potential of human marrow stromal cellsrdquoJournal of Cellular Biochemistry vol 75 no 3 pp 424ndash436 1999
[16] C M Digirolamo D Stokes D Colter D G Phinney R Classand D J Prockop ldquoPropagation and senescence of humanmarrow stromal cells in culture a simple colony-forming assayidentifies samples with the greatest potential to propagate anddifferentiaterdquo British Journal of Haematology vol 107 no 2 pp275ndash281 1999
[17] S James J Fox F Afsari et al ldquoMultiparameter analy-sis of human bone marrow stromal cells identifies distinctimmunomodulatory and differentiation-competent subtypesrdquoStem Cell Reports vol 4 no 6 pp 1004ndash1015 2015
[18] B Peruzzi A Cappariello A Del Fattore N Rucci F DeBenedetti and A Teti ldquoc-Src and IL-6 inhibit osteoblastdifferentiation and integrate IGFBP5 signallingrdquo Nature Com-munications vol 3 article 630 2012
[19] D Fortunati S Reppe A-K Fjeldheim M Nielsen V TGautvik and K M Gautvik ldquoPeriostin is a collagen associatedbonematrix protein regulated by parathyroid hormonerdquoMatrixBiology vol 29 no 7 pp 594ndash601 2010
[20] A Oshima H Tanabe T Yan G N Lowe C A Glackin andA Kudo ldquoA novel mechanism for the regulation of osteoblastdifferentiation transcription of periostin a member of thefasciclin I family is regulated by the bHLH transcription factorTwistrdquo Journal of Cellular Biochemistry vol 86 no 4 pp 792ndash804 2002
[21] T G Kashima T Nishiyama K Shimazu et al ldquoPeriostina novel marker of intramembranous ossification is expressedin fibrous dysplasia and in c-Fos-overexpressing bone lesionsrdquoHuman Pathology vol 40 no 2 pp 226ndash237 2009
[22] CGalliM PiergianniM Piemontese et al ldquoPeriostin improvescell adhesion to implantable biomaterials and osteoblasticdifferentiation on implant titanium surfaces in a topography-dependent fashionrdquo Journal of Biomedical Materials Research -Part A vol 102 no 11 pp 3855ndash3861 2014
18 Stem Cells International
[23] D L Coutu H W Jian A Monette G-E Rivard M DBlostein and J Galipeau ldquoPeriostin amember of a novel familyof vitamin K-dependent proteins is expressed by mesenchymalstromal cellsrdquo The Journal of Biological Chemistry vol 283 no26 pp 17991ndash18001 2008
[24] T Minamizato K Sakamoto T Liu et al ldquoCCN3NOV inhibitsBMP-2-induced osteoblast differentiation by interacting withBMP andNotch signaling pathwaysrdquo Biochemical and Biophysi-cal Research Communications vol 354 no 2 pp 567ndash573 2007
[25] W Si Q Kang H H Luu et al ldquoCCN1Cyr61 is regulated bythe canonicalWnt signal and plays an important role inWnt3A-induced osteoblast differentiation of mesenchymal stem cellsrdquoMolecular and Cellular Biology vol 26 no 8 pp 2955ndash29642006
[26] Q Luo Q Kang W Si et al ldquoConnective tissue growth factor(CTGF) is regulated by Wnt and bone morphogenetic proteinssignaling in osteoblast differentiation of mesenchymal stemcellsrdquo The Journal of Biological Chemistry vol 279 no 53 pp55958ndash55968 2004
[27] Y-S Yoon N Lee and H Scadova ldquoMyocardial regenerationwith bone-marrow-derived stem cellsrdquo Biology of the Cell vol97 no 4 pp 253ndash263 2005
[28] Y S Yoon ldquoCardiac regeneration with novel bone marrow-derived multipotent stem cellsrdquo Discovery Medicine vol 5 pp204ndash208 2005
[29] Y-S Yoon A Wecker L Heyd et al ldquoClonally expanded novelmultipotent stem cells from human bone marrow regeneratemyocardium after myocardial infarctionrdquo Journal of ClinicalInvestigation vol 115 no 2 pp 326ndash338 2005
[30] GM BolandG Perkins D J Hall andR S Tuan ldquoWnt 3a pro-motes proliferation and suppresses osteogenic differentiationof adult human mesenchymal stem cellsrdquo Journal of CellularBiochemistry vol 93 no 6 pp 1210ndash1230 2004
[31] D A Cook S W Fellgett M E Pownall P J OrsquoShea and PG Genever ldquoWnt-dependent osteogenic commitment of bonemarrow stromal cells using a novel GSK3120573 inhibitorrdquo Stem CellResearch vol 12 no 2 pp 415ndash427 2014
[32] J De Boer R Siddappa C Gaspar A VanApeldoorn R Foddeand C Van Blitterswijk ldquoWnt signaling inhibits osteogenicdifferentiation of humanmesenchymal stem cellsrdquoBone vol 34no 5 pp 818ndash826 2004
[33] MAl-Toub A AlmusaM Almajed et al ldquoPleiotropic effects ofcancer cellsrsquo secreted factors on human stromal (mesenchymal)stem cellsrdquo Stem Cell Research and Therapy vol 4 article 1142013
[34] L Chen D Qanie A Jafari et al ldquoDelta-like 1fetal antigen-1(Dlk1FA1) is a novel regulator of chondrogenic cell differenti-ation via inhibition of the Akt kinase-dependent pathwayrdquoTheJournal of Biological Chemistry vol 286 no 37 pp 32140ndash321492011
[35] T Matsunobu K Torigoe M Ishikawa et al ldquoCritical roles ofthe TGF-120573 type I receptor ALK5 in perichondrial formationand function cartilage integrity and osteoblast differentiationduring growth plate developmentrdquo Developmental Biology vol332 no 2 pp 325ndash338 2009
[36] B C Nair D E Johnson R J Majeska J A Rodkey C DBennett and G A Rodan ldquoRat alkaline phosphatase II Struc-tural similarities between the osteosarcoma bone kidney andplacenta isoenzymesrdquo Archives of Biochemistry and Biophysicsvol 254 no 1 pp 28ndash34 1987
[37] P J Butterworth ldquoAlkaline phosphatase Biochemistry of mam-malian alkaline phosphatasesrdquo Cell Biochemistry and Functionvol 1 no 2 pp 66ndash70 1983
[38] A T Ali C B Penny J E Paiker et al ldquoAlkaline phosphataseis involved in the control of adipogenesis in the murinepreadipocyte cell line 3T3-L1rdquo Clinica Chimica Acta vol 354no 1-2 pp 101ndash109 2005
[39] G M Spaggiari and L Moretta ldquoCellular and molecularinteractions of mesenchymal stem cells in innate immunityrdquoImmunology and Cell Biology vol 91 no 1 pp 27ndash31 2012
213 Quantitative Real-Time PCR (qRT-PCR) Analysis TotalRNA was extracted using MagNA pure compact RNA iso-lation kit (Roche Applied Science Germany Cat number04802993001) in an automatedMagNA pure compact system(Roche Germany) as recommended by the manufacturerThe total RNAwas quantified by Nanodrop spectrophotome-ter (Nanodrop 2000 Thermo Scientific USA) Complemen-tary DNA (cDNA) was synthesized from 1 120583g of the RNAsamples usingHighCapacity cDNAReverse Transcription kit(Applied Biosystems USA) using Labnet Multigene thermo-cycler according to the manufacturerrsquos instructions Relativelevels of mRNA were determined from cDNA by real-timePCR (Applied Biosystems-Real-Time PCRDetection System)with Power SYBR Green PCR kit (Applied BiosystemsUK) according to the manufacturerrsquos instructions Followingnormalization to the reference gene GAPDH quantificationof gene expression was carried out using a comparative Ctmethod where ΔCt is the difference between the CT valuesof the target and the reference gene and fold induction wasperformed from the control (Cont) for the same time pointPrimers (Supplementary Table 1 in Supplementary Materialavailable online at httpdxdoiorg10115520169378081)were obtained from Applied Biosystems (USA) as TAQMANprimers or previously published primers were used (seeSupplementary Table 1)
214 DNAMicroarray Global Gene Expression Analysis Fourhundred ng of total RNA was used as input for gener-ating biotin-labeled cRNA (Ambion Austin TX UnitedStates) cRNA samples were then hybridized onto Illuminahuman-8 BeadChips version 3 Hybridization washing Cy3-streptavidin staining and scanning were performed on theIllumina BeadStation 500 platform (Illumina San DiegoCA USA) according to the manufacturerrsquos instructions andeverything was done in triplicate Expression data analysiswas carried out using the Partek genomic suite softwareRaw data were background-subtracted normalized usingthe ldquorank invariantrdquo algorithm and filtered for significantexpression on the basis of negative control beads Geneswere considered significantly expressed with detection 119901values le 001 Differential expression analysis was performedwith the Illumina custom method using freshly isolatedprimary hBMSC (used at passage 3) as a reference controlThe following parameters were set to identify statisticalsignificance differential 119901 values le 001 fold change ratiogt 15 Pathway analysis was performed using DAVID Bioin-formatics Resources 67 (httpdavidabccncifcrfgov) andGeneSpring GX software (Agilent Technologies) Pathwayanalysis for CL1 OS D14 versus CL2 OS D14 was conductedusing the Single Experiment Pathway analysis feature inGeneSpring 120 (Agilent Technologies)
215 Small Interfering (si)RNATransfection For transfectionhBMSC in logarithmic growth phase were transfected withSilencer Select Predesigned ALP siRNA (25 nM) (Assay IDs1298 and Cat number 4390824) (Ambion The RNA Com-pany USA) using Lipofectamine RNAiMAX Reagent (Invit-rogen CA USA) plus serum-free Opti-MEMI mediumunder the conditions described by the manufacturer At
day 3 of transfection the cells were induced for osteogenicdifferentiation for an additional 7 days ALP staining wasused as a control for the siRNA transfection efficiency andtimeline
216 Statistical Analysis All of the results were presentedas the mean and standard deviation (SD) of at least 3independent experiments with 3ndash5 technical repeats in eachexperiment Studentrsquos 119905-test (two-tailed) was used for testingdifferences between groups 119901 value lt005 was consideredstatistically significant
3 Results
31 Comparison between CL1 and CL2 Differences in Mor-phology Proliferation and Marker Expression Profile Weisolated two distinct clonal cell populations of hBMSC-TERThBMSC-CL1 and hBMSC-CL2 (for easiness will be termedhereafter CL1 and CL2) based on differences in cell morphol-ogy (Figure 1(a)) CL1 cells had cuboidalmorphologywhereasCL2 cells have spindle-shaped fibroblast-like morphologyCL1 cells had higher proliferation rate compared to CL2(Figure 1(b)) mean PD rates of CL1 and CL2 were 0714 and0429 PDday respectively (Figure 1(b)) Both CL1 and CL2expressed surface marker profiles characteristics of hBMSC(gt90) CD44+ CD63+ CD73+ CD105+ and CD166+(Figure 1(c)) However CL1 cells showed higher expressionof CD146 (927 versus 12) and ALP (98 versus 0)compared to CL2 (Figure 1(d)) TEM revealed the presenceof abundant pseudopodia in CL1 indicating high motility(Figure 1(e)(A)) as well as well-developed mitochondriaand rough endoplasmic reticulum (rER) suggesting highmetabolic activity CL2 cells contained abundant phagocyticvacuole (pv)microvilli (mi) and lysosomes (ly) (Figure 1(e))
We performed quantitative real-time PCR (RT-PCR)for genes expressed in mesodermal progenitor cells [12]CL1 expressed higher levels of BMP4 MIXL1 WNT3a andTWIST compared to CL2 (Figure 1(f) 119901 lt 001) In contrastCL2 expressed higher levels of Kinase Insert Domain Recep-tor (Type III Receptor Tyrosine Kinase) (KDR) expressed inendothelial cells and smoothmusclemyosin heavy chain gene(smMHC) expressed in smooth muscle cells (Figure 1(f))
32 CL1 Cells Exhibit Enhanced Osteoblast DifferentiationFollowing osteoblast (OB) differentiation induction ALPstaining andALP enzymatic activity were significantly higherin CL1 compared to CL2 cells (Figure 2(a) 119901 lt 001)Similarly Alizarin Red staining and quantitation of formedmineralized matrix were more pronounced in CL1 cells (Fig-ure 2(b) 119901 lt 001) In addition CL1 cells expressed higherlevels of osteoblastic genes ALP RUNX2 and osteopontin(OPN) (Figure 2(c) upper panel) compared to CL2 cells
Global gene expression microarray analysis of OB dif-ferentiated cells at day 14 showed around 1060 genes signif-icantly upregulated more than 2-fold (119901 lt 001) in CL1Among the upregulated genes 80 genes were annotated tobone development and osteoblast differentiation (Table 1)The highest upregulated genes included paired-like home-odomain 2 (PITX2) Insulin-like growth factor 1 (IGF1) and
Stem Cells International 5
CL1
CL2
(a)
Cum
ulat
ive P
D
Cumulative days
Growth curve
0 10 20 30 40
CL1 CL2
CL1 CL20
5
10
15
20
25
0
01
02
03
04
05
06
07
08
(PD
day
)
(b)
CL2
103
102
101
100
103
102
101
100
103
102
101
100
103
102
101
100
103
102
101
100
103
101
100
103
102
102
101
100
103
102
101
100
103
102
101
100
103
102
101
100
CL1
CD44 CD63 CD73 CD166 CD105
T3 CD44 T3 CD63 T3 CD73 T3 CD166 T3 CD105
0
20
40
60
80
Cou
nt
0
20
40
60
80
100
120
Cou
ntC
ount
Cou
nt
Cou
nt
Cou
nt
Cou
nt0
20
40
60
80
Cou
nt
0
20
40
60
80
100
Cou
nt
0
20
40
60
80
Cou
nt
80
60
40
20
0
100
50
0
120
100
80
60
40
20
0
120
100
80
60
40
20
0
100
80
60
40
20
0
(c)
ISO CD146 ALP
ISO CD146 ALP
FL2-
H
FL2-
H
FL2-
HPEPEPECL
2CL
1
FL1-H FL1-H FL1-H
FITCFITCFITC
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
(d)
Figure 1 Continued
6 Stem Cells International
(A) (B)
(e)
Fold
indu
ctio
n
CL1CL2
lowast
lowastlowast lowast
BMP4 MIXL1 WNT3a TWIST
lowastlowast
KDR smMHCminus01
01
03
05
07
09
11
13
15
mdash
050
100
150
200
250
300
CL1CL2
(f)
Figure 1 Cellular phenotype of two clonal populations of humanbonemarrow stromal stem cells CL1 andCL2 (a) Cellmorphology CL1 cellsshowed large cuboidal morphology while CL2 cells had spindle-shaped fibroblast-like morphology (b) Growth curves showing populationdoubling (PD) rate during long term culture (c) Flow cytometry analysis (FACS) shows expression of CD44 CD63 CD73 CD105 andCD166in CL1 and CL2 cells Matched isotype control was used for gating (d) Flow cytometry analysis presented as dot blot of CD146 and alkalinephosphatase (ALP) cell surface proteins (e) Transmission electron microscope (TEM) (A) CL1 (1200x) (B) CL2 (1500x) n nucleus nunucleolus rER rough endoplasmic reticulum ly lysosomes pv phagocytic vacuole and rer reticular stalk of rER (f) Gene expressionanalysis using RT-PCR for a group of mesodermal and stromal genes Gene expression was normalized to GAPDH and presented as foldchange Data is shown as mean plusmn SD of three independent experiments lowast119901 lt 005 119901 lt 0001
collagen type V alpha 3 (COL5A3) osteomodulin (OMD)and T-box 15 (TBX15) (Table 1) Furthermore several knownosteoblast-related genes were upregulated in CL1 cells suchas bone morphogenetic protein 6 (BMP6) fibroblast growthfactor receptor 3 (FGFR3) insulin-like growth factor bindingprotein 5 (IGFBP5) and vitamin D (125-dihydroxyvitaminD3) receptor (VDR) (Table 1) On the other hand 1200 geneswere upregulated in CL2 cells 255 genes were annotated toimmunity and immune response and defense This categoryincluded genes from complement system chemokine (C-Cmotif) ligands interferon family chemokine (C-X-C motif)ligands and receptor major histocompatibility complex classII molecules interleukins and tumor necrosis factor receptorsuperfamily (Table 2 and Supplementary Table 3)
33 CL1 Cells Exhibit Enhanced Adipocyte DifferentiationWe observed significant differences between CL1 and CL2in their response to adipocytic differentiation inductionCL1 differentiated readily to adipocytes compared to CL2(Figure 2(d) lower panel) evidenced by higher levels ofadipocyticmarkers gene expression LPL (lipoprotein lipase)and adiponectin as well as formation of mature lipid filledadipocytes visualized by Oil Red-O staining and quantitativeNile red staining (Figure 2(d))
34 CL1 Cells Differentiate to Chondrocytic Lineage Inpellet cultures CL1 cells formed 3D pellets containingproteoglycan-secreting chondrocytes which stained positivewith Alcian blue Limited chondrocyte differentiation wasvisible in cell pellets of CL2 cells The differentiated chon-drocytes in CL1 pellets expressed higher levels of collagen Xand collagen II which was overlapping the Alcian blue stain(Figure 2(e))
35 Molecular Signature of CL1 and CL2 Cells To define themolecular signature and molecular differences between CL1
and CL2 we compared the basal gene expression patternof CL1 and CL2 cells using DNA microarrays The PCAanalysis showed a clear separation between CL1 and CL2(Supplementary Figure 1) Comparison between CL1 withCL2 showed that 915 geneswere differentially expressed in thetwo cell lines (gt2-fold119901 lt 001) 462 genes were upregulatedand 452 were downregulated in CL1 versus CL2 The mostrelevant genes that were upregulated in CL1 are listed inTable 3(a) Among these 35 highly expressed genes in CL1 thefollowing 11 genes were present in skeletal and muscular sys-tem development and function FOLR3 CCL3L1 SERPINB2POSTN IGFBP5 CCL3 NOV ALP TNFRSF11B ACTG2and CDH11 (Table 3(a)) Functional annotation of the upreg-ulated genes in CL1 using the Ingenuity Pathway Analysis(IPA) revealed enrichment in the following categories ldquotissuedevelopmentrdquo ldquoskeletal and muscular system developmentand functionrdquo and ldquoorganismal developmentrdquo (Table 3(b))Furthermore the DAVID annotation tool was employed toassess the functional relationships of the upregulated genesin CL1 showing enrichment in ontologies ldquoskeletal andmuscular system development and functionrdquo that includedbone size osteoblast differentiation bone mineralizationand bone mineral density (Table 3(c)) CL1 exhibited upreg-ulation of WNT pathway ligands WNT5B (2-fold) andLRP5 (2-fold) (Table 1) Also ALP was among the highlyexpressed genes together with POSTN IGFBP5 SPP1 IL-6 and DKK1 (Tables 3(a) and 1) These genes are knownto play an important role in osteoblast differentiation andbone formation For CL2 inhibitors of WNT pathway wereupregulated and included SFRP1 (11-fold) DKK2 (32-fold)FGF2 (31-fold) and GBP2 (24-fold) Functional annotationof the upregulated genes in CL2 revealed enrichment in thefollowing categories ldquodevelopmental processrdquo ldquomulticellularorganismal processrdquo ldquobiological adhesionrdquo and ldquoimmunesystem processrdquo (Supplementary Table 2A) In-depth analysis
Stem Cells International 7
CL1
CL2
ALPCont D 14
Cont OB I
ALP activity
mdash100
200
300
400
500
600
Fold
chan
ge
CL1CL2
(a)
CL1
CL2
ALZCont D 14
Fold
chan
ge
25
20
15
10
5
0
Cont d14 d21
CL1CL2
ALZ quantification
(b)
CL1CL2
CL1CL2
LPL ADIPOQ
Cont ContADIP D14 ADIP D14
000
100
200
300
400
500
600
700
0
1000
2000
3000
4000
5000
CL1CL2
CL1CL2
CL1CL2
RUNX2 ALP OPN
Cont Cont ContOS D7 OS D7 OS D7
00
05
10
15
20
25
30
00
05
10
15
20
25
30
35
00
05
10
15
20
25
30
35
(c)
CL1CL2
CL1
CL2
Control Adip d7 Adip d14
d7 d140
005
01
015
02
025
03
035
04
045
Nile
red
fluor
esce
nt in
tens
ity
(d)
CL1
CL2
Alcian blue Col X Col II
(e)
Figure 2 In vitro osteoblastic and adipocytic differentiation of two clonal populations of human bone marrow stromal stem cells CL1 andCL2 cells Both cell lines were induced for osteoblast differentiation using standard protocol described in the Methods (a) ALP staining atday 14 in control noninduced (Cont) and osteoblast induced cells (D14) Right panel shows ALP activity (119899 = 3 independent experiments119901 lt 0001) (b) Mineralized matrix formation visualized by Alizarin Red S staining Right panel shows Alizarin Red quantification at day 14and 21 after osteoblast differentiation (119899 = 3 independent experiments 119901 lt 0001) (c) Quantitative RT-PCR of osteoblastic and adipocytegene markers in CL1 and CL2 during osteoblast (upper panel) and adipocyte (lower panel) differentiation ALP = alkaline phosphataseOPN = osteopontin LPL = lipoprotein lipase and ADIPOQ = adiponectin Data are presented as fold change in expression of each targetgene normalized to GAPDH (119899 = 3 independent experiments 119901 lt 005
119901 lt 0001) (d) CL1 and CL2 lines were induced for adipocytedifferentiation using standard protocol described in the Methods Adipocyte formation was visualized at day 7 (Adip d7) and day 14 (Adipd14) by Oil Red-O staining Lower panel presents quantification of Nile red staining (119899 = 3 independent experiments 119901 lt 0001) (e) CL1and CL2 lines were induced for chondrocyte differentiation using 21-day pellet culture method as described in the Methods The pellets werestained with Alcian blue collagen 10 (Col X) and collagen 2 (Col II) (original magnification 5x)
8 Stem Cells International
Table 1 Microarray data analysis showing genes related to bone development and osteoblast differentiation upregulated in CL1 versus CL2cells
Probe ID Genbank accession Gene name Gene symbol FCA 23 P167367 NM 153426 Paired-like homeodomain 2 PITX2 30884
of the biological processes revealed several immune-relatedpathways ldquoMAPKKK cascaderdquo ldquoimmunity and defenserdquoldquosignal transductionrdquo ldquoextracellular matrix protein-mediatedsignalingrdquo and ldquointerferon-mediated immunityrdquo among oth-ers that were upregulated (Supplementary Table 2B) Also40 genes related to immune system related factors wereidentified as significantly enriched in CL2 compared to CL1cells (Supplementary Table 2C)We chose the following genesfor validation of the microarray results NOV IGFBP5 ALPTAGLN and CDH11 as they were highly expressed in CL1RT-PCR analysis confirmed themicroarray results (Figure 3)
Furthermore we compared the molecular phenotype ofCL1 and CL2 cells with that of phBMSCWe found that morethan 80 of the genes expressed in CL1 and 90 in CL2 cellswere commonwith primary phBMSC (Supplementary Figure2) suggesting that CL1 and CL2 molecular phenotype existwithin the heterogeneous population of phBMSC cultures
36 ALP Knockdown Impairs Differentiation of CL1 CellsSince ALP has been suggested as a marker for hBMSCprogenitor cell lineage commitment [13] and was highlyupregulated in CL1 cells we tested its biological role in
CL1 cells ALP siRNA transfection decreased ALP proteinlevel ALP activity and mRNA gene expression compared tocontrol cells transfected with control siRNA (119901 lt 001) andthis inhibition was detectable up to day 7 days after osteoblastdifferentiation induction (Figures 4(A) and 4(B)) At day 14 ofdifferentiation mineralization ability of CL1 was significantlyimpaired (Figure 4(C)) In addition we found that thenumber of mature adipocyte formations was significantlyreduced to more than 75 (119901 lt 001) (Figure 4(D))
To identify relevant adipocyte differentiation associatedgenes that were targeted by ALP deficiency we compared thedownregulated genes of ALP deficient CL1 with the upreg-ulated genes identified during adipocytic differentiation ofCL1 We identified 62 genes that were common (Figure 5(a)Table 4) and among these genes were genes related tometabolism (primarily lipid and carbohydrate) and transportincluding CYB5B CHST1 TAP1 ATP8A1 LRP8 PLCD1and FABP5 (Table 4) We further performed quantitativereal-time PCR of ALP deficient CL1 cells during adipocytedifferentiation The following adipocyte-associated geneswere downregulated PPAR1205742 LPL and aP2 (Figure 5(b))
12 Stem Cells International
Table 3 Whole genome microarray analysis of two clonal popu-lations of human bone marrow stromal stem cells CL1 and CL2cells (a) Top 35 highly upregulated genes in CL1 cells versus CL2cells (b) Ingenuity Pathway Analysis (IPA) showing the differentphysiological system development and function genes found in eachcategory and the corresponding 119901 value (c) Upper part IPA analysisshowing the categories for skeletal and muscular system developmentand function upregulated in CL1 cells and lower part showing genesupregulated in CL2 cells NOV = nephroblastoma overexpressedIGFBP5 = insulin-like growth factor binding protein 5 ALP TAGLN= transgelin and CDH11 = OB-cadherin (osteoblast)
Name 119901 value moleculesPhysiological system development and function
Organismal development 705119864 minus 10ndash143119864 minus 03 181Embryonic development 100119864 minus 09ndash143119864 minus 03 154Organ development 100119864 minus 09ndash143119864 minus 03 145
(b) Continued
Name 119901 value moleculesSkeletal and muscularsystem development andfunction
100119864 minus 09ndash102119864 minus 03 123
Tissue development 100119864 minus 09ndash143119864 minus 03 236
(c)
Functions annotation 119901 value moleculesSkeletal and muscular system development and
function upregulated in CL1 cellsSize of bone 143119864 minus 06 24Differentiation ofosteoblasts 384119864 minus 06 25
Mineralization ofbone 493119864 minus 06 19
Bone mineral density 365119864 minus 05 19Skeletal and muscular system development and
function upregulated in CL2 cellsDevelopment ofmuscle 100119864 minus 09 44
Proliferation ofmuscle cells 229119864 minus 06 35
Remodeling of bone 363119864 minus 06 21Resorption of bone 394119864 minus 06 19
confirming impairment of adipocytic differentiation of ALPdeficient CL1 cells
4 Discussion
We extensively studied two cell populations within culturedhBMSC that were identified based on differences in mor-phology Cellular and molecular studies revealed differencesin growth differentiation capacity and molecular signatureOur data support the notion of the presence of cellular andfunctional heterogeneity among cultured hBMSC
Cellular heterogeneity of cultured hBMSC is recognizedin an increasing number of reports Several extrinsic andintrinsic factors may contribute to the observed hBMSCheterogeneity Extrinsic factors include donor-to-donor vari-ations in the number and quantity of initiating cells whichresult in differences in cell growth rate and differentia-tion capacity [14 15] Intrinsic factors have been examinedemploying single cell clonal analysis and revealed variationsin differentiation potential among individual colonies withinhBMSC cultures ranging from the presence of cells withtrilineage (osteoblast adipocytes and chondrocyte) potencyto cells with null potency [16] Also variations in the abilityof clonal cells to form heterotopic bone when implanted invivo have been reported [5] Our study corroborates thesefindings and provides more detailed cellular and molecularphenotyping of two examples of cell populations that existwithin the heterogeneous hBMSC cultures [17]
Determining the molecular signature of CL1 and CL2using whole genomemicroarray analysis showed enrichmentof lineage-commitment associated genes inCL1 For exampleinsulin-like growth factor 5 (IGFBP5) and interleukin 6 (IL6)
Stem Cells International 13
NOV0040
0030
0020
0010
mdash
IGFBP50015
0010
0005
mdash
ALP0100
0080
0060
0040
0020
mdash
TAGLN0800
0600
0400
0200
CDH11
0060
0040
0020
mdashmdash
LRP30015
0010
0005
mdash
CL1 CL2 CL1 CL2 CL1 CL2
CL1 CL2CL1 CL2CL1 CL2
Figure 3 Validation of whole genome microarray analysis of two clonal populations of human bone marrow stromal stem cells CL1 andCL2 cells Quantitative real-time PCR for highly expressed genes in CL1 cells NOV = nephroblastoma overexpressed IGFBP5 = insulin-likegrowth factor binding protein 5 ALP TAGLN = transgelin and CDH11 = OB-cadherin (osteoblast) Data are presented as fold change inexpression of each target gene normalized to GAPDH (119899 = three independent experiments 119901 lt 0001) (see also Table 3)
were 147- and 33-fold upregulated in CL1 cells respectivelyBoth factors are expressed in osteoprogenitor cells andimportant for osteoblast maturation [18] We also observedthat periostin (POSTN) gene was highly upregulated inCL1 cells (156-fold) POSTN is a 90 kDa secreted proteinoriginally identified in murine osteoblast-like cells and isupregulated by PTH [19] Several studies employing murineand human cells have revealed important role of POSTNin osteoblast differentiation and during development inintramembranous ossification [20ndash23] Another factor iden-tified in CL1 cells is nephroblastoma overexpression (NOV)which is a member of the Cyr 61 connective tissue growthfactor (CNN) family The CCN family of proteins promotesosteoblast differentiation through interaction with integrinsWNT BMP and NOTCH signaling pathways [24ndash26] Inaddition a large number of signaling molecules known tobe regulators of hBMSC lineage specific differentiation forexample insulin-like growth factors [27ndash29] WNT [30ndash32]and MAPK [33ndash35] were enriched in CL1 cells In contrastCL2 expressed high levels of immune-related genes whichmay explain the poor differentiation response to osteoblastor adipocyte lineage In a recent study the authors usedtelomerized hBMSC and showed clearly a clonal populationthat had very low in vitro and in vivo differentiation abilityhowever they had enhanced immune-related features includ-ing high IL7 expression These nullipotent cells expressedCD317 which was associated with remarkably high basal levelexpression of factors with a proinflammatory and antiviralfunction [17] We observed that this molecular phenotypewas associated with distinct ultrastructural characteristicsof the cells In particular CL2 had abundant phagocyticvacuole microvilli and lysosomes features reminiscent of
ultrastructure of immune-regulatory cells Our data thussupport the increasingly recognized feature that hBMSCexhibit immune modulatory functions and a part of theinnate immune response [17]
We observed that ALP protein expression and enzymaticactivity were significantly different between CL1 and CL2cell lines and were thus a potential marker that distinguishesdifferent cell populations with progenitor functions (CL1)from cells with nonprogenitor functions (CL2) ALP isexpressed in a wide variety of tissues including kidneysbone and liver [36 37] but tissue-nonspecific ALP (ALPL)is considered a commitment marker for osteoblastic lineage[13 38] However in a recent study the authors examined thedifferentiation potential of a number of hMSC clones in vitroand in vivo and reported that the hMSC clones with highlevels of ALP expression were committed to trilineage differ-entiation [13] Our data corroborate and extend these findingsby reporting the effects of siRNA-mediated inhibition ofALP that resulted in an impaired hBMSC differentiation notonly to osteoblasts but also to adipocytes Also our resultscorroborate earlier studies that demonstrated in human bonebiopsies the presence of ALP expression in bone marrowadipocytic cells [13 17] All these data suggest that ALP isa ldquostemnessrdquo marker of hBMSC and not just an indicator ofosteoblastic lineage commitment
While CL1 and CL2 were isolated from telomerizedhMSC cell line they are relevant to normal human physi-ology We observed that the molecular phenotypes of CL1and CL2 were contained within the molecular signatureof primary hBMSC suggesting that CL1 and CL2 repre-sent cell populations within the heterogeneous cultures ofhBMSCWe have also previously reported that the molecular
14 Stem Cells International
ALZ
stai
ning
Control siRNA ALP siRNA
ALP
Beta
-act
in
Con
trol
ALP
siRN
A d3
ALP
siRN
A d7
Nile red quantificationlowast
NI ADIP I ALP siRNA0
50
100
150
200
Con
trol s
iRN
AA
LP si
RNA
Day 3 Day 5 OS induction Day 7 OS induction
Fold
indu
ctio
n lo
g sc
ale
10000
1000
100
10
1
Day 0 OS day 7
Control siRNAALP siRNA
(A) (B)
(C)
(D)
Figure 4 Effect of alkaline phosphatase (ALP) gene silencing by small interfering RNA (siRNA) on a clonal population of human bonemarrow stromal stem cell CL1 (A) ALP staining at day 3 day 5 or day 7 days of osteoblast differentiation (OB induction) (B) Quantitativereal-time PCR for ALP gene following ALP siRNA transfection at day 0 OB and day 7 of OB Data are presented as fold change in expressionof each target gene normalized to GAPDH (119899 = three independent experiments 119901 lt 001) Western blotting analysis of day 3 and day 7after siRNA ALP transfection of CL1 cells ALPL specific antibody and B-actin was used (C) Mineralized matrix formation as visualizedby Alizarin Red S staining in siRNA transfected CL1 cells after 14 days of OB induction (D) Nile red quantification of mature lipid filledadipocyte in control noninduced (Cont) adipocyte induced (Adip I) and ALP siRNA transfected cells that are adipocyte induced (ALPsiRNA) Adipocyte induction was carried out for 7 days lowast119901 lt 005
phenotype and cellular responses of hMSC-TERT are similarto those of primary hMSC [10]While we have identified ALPas a marker that can be used for a prospective identificationof differentiation committed population of hBMSC we iden-tified additional distinctive molecular markers of the cellsFor example IGF-1 IGF-2 and IGF binding protein 5 wereenriched in CL1 compared to CL2 IGFs and their bindingproteins are very well-studied factors that play a role inhBMSC proliferation and osteoblast differentiation [18] On
the other hand annexin A3 as well as several immune-relatedgenes was highly enriched in CL2 compared to CL1 Futurestudies are needed to determine the functional significanceof these molecules in relation to the functional identity ofvarious cell populations within the hBMSC cultures andtheir usefulness as biomarkers to dissect the heterogeneouspopulation of cultured hBMSC
Our finding of the presence of functional diversity withinhBMSC cultures that contain progenitor and nonprogenitor
Stem Cells International 15
Table 4 Microarray data analysis showing genes found to be upregulated during adipogenic differentiation and downregulated after ALPKO
62 common elements inldquoAD uprdquo and ldquoALP downrdquo Gene name FC (ALP siRNA versus
control siRNA)APOBEC3G Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3G minus15938025IFI44L Interferon-induced protein 44-like minus11520283PAQR5 Progestin and adipoQ receptor family member V minus6868241PNMA2 Paraneoplastic antigen MA2 minus59695344DUSP23 Dual specificity phosphatase 23 minus54786854CLDN23 Claudin 23 minus51885047ANKDD1A Ankyrin repeat and death domain containing 1A minus51646647IL8 Interleukin 8 minus4887188LRRC23 Leucine rich repeat containing 23 minus47611775IL6 Interleukin 6 (interferon beta 2) minus4693139LIFR Leukemia inhibitory factor receptor alpha minus46540866PTGFR Prostaglandin F receptor (FP) minus4457529FAM134B Family with sequence similarity 134 member B minus4403495CYFIP2 Cytoplasmic FMR1 interacting protein 2 minus4260462METTL7A Methyltransferase-like 7A minus40480843APOBEC3F Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3F minus39502614CA5B Carbonic anhydrase VB mitochondrial minus393889ITGA10 Integrin alpha 10 minus39143775FMO3 Flavin containing monooxygenase 3 minus3852087IMPA2 Inositol monophosphatase 2 (human) minus38374884CDO1 Cysteine dioxygenase type I minus38181455CCDC68 Coiled-coil domain containing 68 minus37292893
HEXDC Hexosaminidase (glycosyl hydrolase family 20 catalytic domain) containing minus2094836SARM1 Sterile alpha and TIR motif containing 1 minus20797038BBS1 Bardet-Biedl syndrome 1 minus20146718SERPING1 Serpin peptidase inhibitor clade G (C1 inhibitor) member 1 minus20102212FAM162A Family with sequence similarity 162 member A minus2005807TCTN1 Tectonic family member 1 minus20033443
AD up ALP down
1006 62 1145
(a)
140
120
100
080
060
040
020
mdash
ALPL PPARg aP2 LPL
lowastlowastlowastlowastlowastlowast
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
lowast
(b)
Figure 5 Adipocyte related genes downregulated in ALP knockdown CL1 cells (a) Venn diagram of whole gene expression analysis of ADupregulated genes compared with ALP KD downregulated genes (b) Quantitative real-time PCR of four selected common genes from Venndiagram including ALPL PPARg2 aP2 and LPL Data are presented as fold change in expression of each target gene normalized to GAPDH(119899 = three independent experiments lowast119901 lt 005 lowastlowast119901 lt 001 and lowastlowastlowast119901 lt 0001)
Stem Cells International 17
cell populations has a clinical relevance It demonstrates thatthe progenitor function and the immune modulatory rolesof hBMSC [39] are mediated by specific and distinguishablepopulations of hBMSC Thus future clinical studies employ-ing hBMSC should attempt to administer the relevant sub-population of hBMSC dependent on the experimental aim asa novel approach to improving the clinical efficiency insteadof the current use of heterogeneous hBMSC populations
Competing Interests
The authors declare that they have no competing interests
Authorsrsquo Contributions
Mona Elsafadi and Amer Mahmood contributed equally tothis work
Acknowledgments
These studies are supported by a grant from the College ofMedicine Research Centre (CMRC) Deanship of ScientificResearch King Saud University Riyadh Saudi Arabia Theauthors would like to thank Mr Ali Al-Roalle for all histechnical help They thank Natalie Twine for advice on DNAmicroarray analysis
References
[1] W Zaher L Harkness A Jafari and M Kassem ldquoAn update ofhuman mesenchymal stem cell biology and their clinical usesrdquoArchives of Toxicology vol 88 no 5 pp 1069ndash1082 2014
[2] S Post B M Abdallah J F Bentzon andM Kassem ldquoDemon-stration of the presence of independent pre-osteoblastic andpre-adipocytic cell populations in bone marrow-derived mes-enchymal stem cellsrdquo Bone vol 43 no 1 pp 32ndash39 2008
[3] K H Larsen C M Frederiksen J S Burns B M Abdallahand M Kassem ldquoIdentifying a molecular phenotype for bonemarrow stromal cells with in vivo bone-forming capacityrdquoJournal of Bone and Mineral Research vol 25 no 4 pp 796ndash808 2010
[4] A Rosenzweig ldquoCardiac cell therapymdashmixed results frommixed cellsrdquoThe New England Journal of Medicine vol 355 no12 pp 1274ndash1277 2006
[5] S A Kuznetsov P H Krebsbach K Satomura et al ldquoSingle-colony derived strains of human marrow stromal fibroblastsform bone after transplantation in vivordquo Journal of Bone andMineral Research vol 12 no 9 pp 1335ndash1347 1997
[6] M Kassem and P Bianco ldquoSkeletal stem cells in space and timerdquoCell vol 160 no 1-2 pp 17ndash19 2015
[7] A I Caplan and D Correa ldquoPDGF in bone formation andregeneration new insights into a novel mechanism involvingMSCsrdquo Journal of Orthopaedic Research vol 29 no 12 pp 1795ndash1803 2011
[8] K Stewart S Walsh J Screen et al ldquoFurther characterizationof cells expressing STRO-1 in cultures of adult human bonemarrow stromal cellsrdquo Journal of Bone and Mineral Researchvol 14 no 8 pp 1345ndash1356 1999
[9] J L Simonsen C Rosada N Serakinci et al ldquoTelomeraseexpression extends the proliferative life-span and maintains
the osteogenic potential of human bone marrow stromal cellsrdquoNature Biotechnology vol 20 no 6 pp 592ndash596 2002
[10] M Al-Nbaheen R Vishnubalaji D Ali et al ldquoHuman stromal(mesenchymal) stem cells from bone marrow adipose tissueand skin exhibit differences in molecular phenotype and differ-entiation potentialrdquo Stem Cell Reviews and Reports vol 9 no 1pp 32ndash43 2013
[11] C A Gregory W G Gunn A Peister and D J Prockop ldquoAnAlizarin red-based assay of mineralization by adherent cells inculture comparison with cetylpyridinium chloride extractionrdquoAnalytical Biochemistry vol 329 no 1 pp 77ndash84 2004
[12] AMahmood L Harkness H D Schroslashder B M Abdallah andM Kassem ldquoEnhanced differentiation of human embryonicstem cells to mesenchymal progenitors by inhibition of TGF-120573activinnodal signaling using SB-431542rdquo Journal of Bone andMineral Research vol 25 no 6 pp 1216ndash1233 2010
[13] H-J Prins A K Braat D Gawlitta et al ldquoIn vitro inductionof alkaline phosphatase levels predicts in vivo bone formingcapacity of human bone marrow stromal cellsrdquo Stem CellResearch vol 12 no 2 pp 428ndash440 2014
[14] K Stenderup J Justesen C Clausen and M Kassem ldquoAgingis associated with decreased maximal life span and acceleratedsenescence of bone marrow stromal cellsrdquo Bone vol 33 no 6pp 919ndash926 2003
[15] D G Phinney G Kopen W Righter S Webster N Tremainand D J Prockop ldquoDonor variation in the growth propertiesand osteogenic potential of human marrow stromal cellsrdquoJournal of Cellular Biochemistry vol 75 no 3 pp 424ndash436 1999
[16] C M Digirolamo D Stokes D Colter D G Phinney R Classand D J Prockop ldquoPropagation and senescence of humanmarrow stromal cells in culture a simple colony-forming assayidentifies samples with the greatest potential to propagate anddifferentiaterdquo British Journal of Haematology vol 107 no 2 pp275ndash281 1999
[17] S James J Fox F Afsari et al ldquoMultiparameter analy-sis of human bone marrow stromal cells identifies distinctimmunomodulatory and differentiation-competent subtypesrdquoStem Cell Reports vol 4 no 6 pp 1004ndash1015 2015
[18] B Peruzzi A Cappariello A Del Fattore N Rucci F DeBenedetti and A Teti ldquoc-Src and IL-6 inhibit osteoblastdifferentiation and integrate IGFBP5 signallingrdquo Nature Com-munications vol 3 article 630 2012
[19] D Fortunati S Reppe A-K Fjeldheim M Nielsen V TGautvik and K M Gautvik ldquoPeriostin is a collagen associatedbonematrix protein regulated by parathyroid hormonerdquoMatrixBiology vol 29 no 7 pp 594ndash601 2010
[20] A Oshima H Tanabe T Yan G N Lowe C A Glackin andA Kudo ldquoA novel mechanism for the regulation of osteoblastdifferentiation transcription of periostin a member of thefasciclin I family is regulated by the bHLH transcription factorTwistrdquo Journal of Cellular Biochemistry vol 86 no 4 pp 792ndash804 2002
[21] T G Kashima T Nishiyama K Shimazu et al ldquoPeriostina novel marker of intramembranous ossification is expressedin fibrous dysplasia and in c-Fos-overexpressing bone lesionsrdquoHuman Pathology vol 40 no 2 pp 226ndash237 2009
[22] CGalliM PiergianniM Piemontese et al ldquoPeriostin improvescell adhesion to implantable biomaterials and osteoblasticdifferentiation on implant titanium surfaces in a topography-dependent fashionrdquo Journal of Biomedical Materials Research -Part A vol 102 no 11 pp 3855ndash3861 2014
18 Stem Cells International
[23] D L Coutu H W Jian A Monette G-E Rivard M DBlostein and J Galipeau ldquoPeriostin amember of a novel familyof vitamin K-dependent proteins is expressed by mesenchymalstromal cellsrdquo The Journal of Biological Chemistry vol 283 no26 pp 17991ndash18001 2008
[24] T Minamizato K Sakamoto T Liu et al ldquoCCN3NOV inhibitsBMP-2-induced osteoblast differentiation by interacting withBMP andNotch signaling pathwaysrdquo Biochemical and Biophysi-cal Research Communications vol 354 no 2 pp 567ndash573 2007
[25] W Si Q Kang H H Luu et al ldquoCCN1Cyr61 is regulated bythe canonicalWnt signal and plays an important role inWnt3A-induced osteoblast differentiation of mesenchymal stem cellsrdquoMolecular and Cellular Biology vol 26 no 8 pp 2955ndash29642006
[26] Q Luo Q Kang W Si et al ldquoConnective tissue growth factor(CTGF) is regulated by Wnt and bone morphogenetic proteinssignaling in osteoblast differentiation of mesenchymal stemcellsrdquo The Journal of Biological Chemistry vol 279 no 53 pp55958ndash55968 2004
[27] Y-S Yoon N Lee and H Scadova ldquoMyocardial regenerationwith bone-marrow-derived stem cellsrdquo Biology of the Cell vol97 no 4 pp 253ndash263 2005
[28] Y S Yoon ldquoCardiac regeneration with novel bone marrow-derived multipotent stem cellsrdquo Discovery Medicine vol 5 pp204ndash208 2005
[29] Y-S Yoon A Wecker L Heyd et al ldquoClonally expanded novelmultipotent stem cells from human bone marrow regeneratemyocardium after myocardial infarctionrdquo Journal of ClinicalInvestigation vol 115 no 2 pp 326ndash338 2005
[30] GM BolandG Perkins D J Hall andR S Tuan ldquoWnt 3a pro-motes proliferation and suppresses osteogenic differentiationof adult human mesenchymal stem cellsrdquo Journal of CellularBiochemistry vol 93 no 6 pp 1210ndash1230 2004
[31] D A Cook S W Fellgett M E Pownall P J OrsquoShea and PG Genever ldquoWnt-dependent osteogenic commitment of bonemarrow stromal cells using a novel GSK3120573 inhibitorrdquo Stem CellResearch vol 12 no 2 pp 415ndash427 2014
[32] J De Boer R Siddappa C Gaspar A VanApeldoorn R Foddeand C Van Blitterswijk ldquoWnt signaling inhibits osteogenicdifferentiation of humanmesenchymal stem cellsrdquoBone vol 34no 5 pp 818ndash826 2004
[33] MAl-Toub A AlmusaM Almajed et al ldquoPleiotropic effects ofcancer cellsrsquo secreted factors on human stromal (mesenchymal)stem cellsrdquo Stem Cell Research and Therapy vol 4 article 1142013
[34] L Chen D Qanie A Jafari et al ldquoDelta-like 1fetal antigen-1(Dlk1FA1) is a novel regulator of chondrogenic cell differenti-ation via inhibition of the Akt kinase-dependent pathwayrdquoTheJournal of Biological Chemistry vol 286 no 37 pp 32140ndash321492011
[35] T Matsunobu K Torigoe M Ishikawa et al ldquoCritical roles ofthe TGF-120573 type I receptor ALK5 in perichondrial formationand function cartilage integrity and osteoblast differentiationduring growth plate developmentrdquo Developmental Biology vol332 no 2 pp 325ndash338 2009
[36] B C Nair D E Johnson R J Majeska J A Rodkey C DBennett and G A Rodan ldquoRat alkaline phosphatase II Struc-tural similarities between the osteosarcoma bone kidney andplacenta isoenzymesrdquo Archives of Biochemistry and Biophysicsvol 254 no 1 pp 28ndash34 1987
[37] P J Butterworth ldquoAlkaline phosphatase Biochemistry of mam-malian alkaline phosphatasesrdquo Cell Biochemistry and Functionvol 1 no 2 pp 66ndash70 1983
[38] A T Ali C B Penny J E Paiker et al ldquoAlkaline phosphataseis involved in the control of adipogenesis in the murinepreadipocyte cell line 3T3-L1rdquo Clinica Chimica Acta vol 354no 1-2 pp 101ndash109 2005
[39] G M Spaggiari and L Moretta ldquoCellular and molecularinteractions of mesenchymal stem cells in innate immunityrdquoImmunology and Cell Biology vol 91 no 1 pp 27ndash31 2012
Figure 1 Cellular phenotype of two clonal populations of humanbonemarrow stromal stem cells CL1 andCL2 (a) Cellmorphology CL1 cellsshowed large cuboidal morphology while CL2 cells had spindle-shaped fibroblast-like morphology (b) Growth curves showing populationdoubling (PD) rate during long term culture (c) Flow cytometry analysis (FACS) shows expression of CD44 CD63 CD73 CD105 andCD166in CL1 and CL2 cells Matched isotype control was used for gating (d) Flow cytometry analysis presented as dot blot of CD146 and alkalinephosphatase (ALP) cell surface proteins (e) Transmission electron microscope (TEM) (A) CL1 (1200x) (B) CL2 (1500x) n nucleus nunucleolus rER rough endoplasmic reticulum ly lysosomes pv phagocytic vacuole and rer reticular stalk of rER (f) Gene expressionanalysis using RT-PCR for a group of mesodermal and stromal genes Gene expression was normalized to GAPDH and presented as foldchange Data is shown as mean plusmn SD of three independent experiments lowast119901 lt 005 119901 lt 0001
collagen type V alpha 3 (COL5A3) osteomodulin (OMD)and T-box 15 (TBX15) (Table 1) Furthermore several knownosteoblast-related genes were upregulated in CL1 cells suchas bone morphogenetic protein 6 (BMP6) fibroblast growthfactor receptor 3 (FGFR3) insulin-like growth factor bindingprotein 5 (IGFBP5) and vitamin D (125-dihydroxyvitaminD3) receptor (VDR) (Table 1) On the other hand 1200 geneswere upregulated in CL2 cells 255 genes were annotated toimmunity and immune response and defense This categoryincluded genes from complement system chemokine (C-Cmotif) ligands interferon family chemokine (C-X-C motif)ligands and receptor major histocompatibility complex classII molecules interleukins and tumor necrosis factor receptorsuperfamily (Table 2 and Supplementary Table 3)
33 CL1 Cells Exhibit Enhanced Adipocyte DifferentiationWe observed significant differences between CL1 and CL2in their response to adipocytic differentiation inductionCL1 differentiated readily to adipocytes compared to CL2(Figure 2(d) lower panel) evidenced by higher levels ofadipocyticmarkers gene expression LPL (lipoprotein lipase)and adiponectin as well as formation of mature lipid filledadipocytes visualized by Oil Red-O staining and quantitativeNile red staining (Figure 2(d))
34 CL1 Cells Differentiate to Chondrocytic Lineage Inpellet cultures CL1 cells formed 3D pellets containingproteoglycan-secreting chondrocytes which stained positivewith Alcian blue Limited chondrocyte differentiation wasvisible in cell pellets of CL2 cells The differentiated chon-drocytes in CL1 pellets expressed higher levels of collagen Xand collagen II which was overlapping the Alcian blue stain(Figure 2(e))
35 Molecular Signature of CL1 and CL2 Cells To define themolecular signature and molecular differences between CL1
and CL2 we compared the basal gene expression patternof CL1 and CL2 cells using DNA microarrays The PCAanalysis showed a clear separation between CL1 and CL2(Supplementary Figure 1) Comparison between CL1 withCL2 showed that 915 geneswere differentially expressed in thetwo cell lines (gt2-fold119901 lt 001) 462 genes were upregulatedand 452 were downregulated in CL1 versus CL2 The mostrelevant genes that were upregulated in CL1 are listed inTable 3(a) Among these 35 highly expressed genes in CL1 thefollowing 11 genes were present in skeletal and muscular sys-tem development and function FOLR3 CCL3L1 SERPINB2POSTN IGFBP5 CCL3 NOV ALP TNFRSF11B ACTG2and CDH11 (Table 3(a)) Functional annotation of the upreg-ulated genes in CL1 using the Ingenuity Pathway Analysis(IPA) revealed enrichment in the following categories ldquotissuedevelopmentrdquo ldquoskeletal and muscular system developmentand functionrdquo and ldquoorganismal developmentrdquo (Table 3(b))Furthermore the DAVID annotation tool was employed toassess the functional relationships of the upregulated genesin CL1 showing enrichment in ontologies ldquoskeletal andmuscular system development and functionrdquo that includedbone size osteoblast differentiation bone mineralizationand bone mineral density (Table 3(c)) CL1 exhibited upreg-ulation of WNT pathway ligands WNT5B (2-fold) andLRP5 (2-fold) (Table 1) Also ALP was among the highlyexpressed genes together with POSTN IGFBP5 SPP1 IL-6 and DKK1 (Tables 3(a) and 1) These genes are knownto play an important role in osteoblast differentiation andbone formation For CL2 inhibitors of WNT pathway wereupregulated and included SFRP1 (11-fold) DKK2 (32-fold)FGF2 (31-fold) and GBP2 (24-fold) Functional annotationof the upregulated genes in CL2 revealed enrichment in thefollowing categories ldquodevelopmental processrdquo ldquomulticellularorganismal processrdquo ldquobiological adhesionrdquo and ldquoimmunesystem processrdquo (Supplementary Table 2A) In-depth analysis
Stem Cells International 7
CL1
CL2
ALPCont D 14
Cont OB I
ALP activity
mdash100
200
300
400
500
600
Fold
chan
ge
CL1CL2
(a)
CL1
CL2
ALZCont D 14
Fold
chan
ge
25
20
15
10
5
0
Cont d14 d21
CL1CL2
ALZ quantification
(b)
CL1CL2
CL1CL2
LPL ADIPOQ
Cont ContADIP D14 ADIP D14
000
100
200
300
400
500
600
700
0
1000
2000
3000
4000
5000
CL1CL2
CL1CL2
CL1CL2
RUNX2 ALP OPN
Cont Cont ContOS D7 OS D7 OS D7
00
05
10
15
20
25
30
00
05
10
15
20
25
30
35
00
05
10
15
20
25
30
35
(c)
CL1CL2
CL1
CL2
Control Adip d7 Adip d14
d7 d140
005
01
015
02
025
03
035
04
045
Nile
red
fluor
esce
nt in
tens
ity
(d)
CL1
CL2
Alcian blue Col X Col II
(e)
Figure 2 In vitro osteoblastic and adipocytic differentiation of two clonal populations of human bone marrow stromal stem cells CL1 andCL2 cells Both cell lines were induced for osteoblast differentiation using standard protocol described in the Methods (a) ALP staining atday 14 in control noninduced (Cont) and osteoblast induced cells (D14) Right panel shows ALP activity (119899 = 3 independent experiments119901 lt 0001) (b) Mineralized matrix formation visualized by Alizarin Red S staining Right panel shows Alizarin Red quantification at day 14and 21 after osteoblast differentiation (119899 = 3 independent experiments 119901 lt 0001) (c) Quantitative RT-PCR of osteoblastic and adipocytegene markers in CL1 and CL2 during osteoblast (upper panel) and adipocyte (lower panel) differentiation ALP = alkaline phosphataseOPN = osteopontin LPL = lipoprotein lipase and ADIPOQ = adiponectin Data are presented as fold change in expression of each targetgene normalized to GAPDH (119899 = 3 independent experiments 119901 lt 005
119901 lt 0001) (d) CL1 and CL2 lines were induced for adipocytedifferentiation using standard protocol described in the Methods Adipocyte formation was visualized at day 7 (Adip d7) and day 14 (Adipd14) by Oil Red-O staining Lower panel presents quantification of Nile red staining (119899 = 3 independent experiments 119901 lt 0001) (e) CL1and CL2 lines were induced for chondrocyte differentiation using 21-day pellet culture method as described in the Methods The pellets werestained with Alcian blue collagen 10 (Col X) and collagen 2 (Col II) (original magnification 5x)
8 Stem Cells International
Table 1 Microarray data analysis showing genes related to bone development and osteoblast differentiation upregulated in CL1 versus CL2cells
Probe ID Genbank accession Gene name Gene symbol FCA 23 P167367 NM 153426 Paired-like homeodomain 2 PITX2 30884
of the biological processes revealed several immune-relatedpathways ldquoMAPKKK cascaderdquo ldquoimmunity and defenserdquoldquosignal transductionrdquo ldquoextracellular matrix protein-mediatedsignalingrdquo and ldquointerferon-mediated immunityrdquo among oth-ers that were upregulated (Supplementary Table 2B) Also40 genes related to immune system related factors wereidentified as significantly enriched in CL2 compared to CL1cells (Supplementary Table 2C)We chose the following genesfor validation of the microarray results NOV IGFBP5 ALPTAGLN and CDH11 as they were highly expressed in CL1RT-PCR analysis confirmed themicroarray results (Figure 3)
Furthermore we compared the molecular phenotype ofCL1 and CL2 cells with that of phBMSCWe found that morethan 80 of the genes expressed in CL1 and 90 in CL2 cellswere commonwith primary phBMSC (Supplementary Figure2) suggesting that CL1 and CL2 molecular phenotype existwithin the heterogeneous population of phBMSC cultures
36 ALP Knockdown Impairs Differentiation of CL1 CellsSince ALP has been suggested as a marker for hBMSCprogenitor cell lineage commitment [13] and was highlyupregulated in CL1 cells we tested its biological role in
CL1 cells ALP siRNA transfection decreased ALP proteinlevel ALP activity and mRNA gene expression compared tocontrol cells transfected with control siRNA (119901 lt 001) andthis inhibition was detectable up to day 7 days after osteoblastdifferentiation induction (Figures 4(A) and 4(B)) At day 14 ofdifferentiation mineralization ability of CL1 was significantlyimpaired (Figure 4(C)) In addition we found that thenumber of mature adipocyte formations was significantlyreduced to more than 75 (119901 lt 001) (Figure 4(D))
To identify relevant adipocyte differentiation associatedgenes that were targeted by ALP deficiency we compared thedownregulated genes of ALP deficient CL1 with the upreg-ulated genes identified during adipocytic differentiation ofCL1 We identified 62 genes that were common (Figure 5(a)Table 4) and among these genes were genes related tometabolism (primarily lipid and carbohydrate) and transportincluding CYB5B CHST1 TAP1 ATP8A1 LRP8 PLCD1and FABP5 (Table 4) We further performed quantitativereal-time PCR of ALP deficient CL1 cells during adipocytedifferentiation The following adipocyte-associated geneswere downregulated PPAR1205742 LPL and aP2 (Figure 5(b))
12 Stem Cells International
Table 3 Whole genome microarray analysis of two clonal popu-lations of human bone marrow stromal stem cells CL1 and CL2cells (a) Top 35 highly upregulated genes in CL1 cells versus CL2cells (b) Ingenuity Pathway Analysis (IPA) showing the differentphysiological system development and function genes found in eachcategory and the corresponding 119901 value (c) Upper part IPA analysisshowing the categories for skeletal and muscular system developmentand function upregulated in CL1 cells and lower part showing genesupregulated in CL2 cells NOV = nephroblastoma overexpressedIGFBP5 = insulin-like growth factor binding protein 5 ALP TAGLN= transgelin and CDH11 = OB-cadherin (osteoblast)
Name 119901 value moleculesPhysiological system development and function
Organismal development 705119864 minus 10ndash143119864 minus 03 181Embryonic development 100119864 minus 09ndash143119864 minus 03 154Organ development 100119864 minus 09ndash143119864 minus 03 145
(b) Continued
Name 119901 value moleculesSkeletal and muscularsystem development andfunction
100119864 minus 09ndash102119864 minus 03 123
Tissue development 100119864 minus 09ndash143119864 minus 03 236
(c)
Functions annotation 119901 value moleculesSkeletal and muscular system development and
function upregulated in CL1 cellsSize of bone 143119864 minus 06 24Differentiation ofosteoblasts 384119864 minus 06 25
Mineralization ofbone 493119864 minus 06 19
Bone mineral density 365119864 minus 05 19Skeletal and muscular system development and
function upregulated in CL2 cellsDevelopment ofmuscle 100119864 minus 09 44
Proliferation ofmuscle cells 229119864 minus 06 35
Remodeling of bone 363119864 minus 06 21Resorption of bone 394119864 minus 06 19
confirming impairment of adipocytic differentiation of ALPdeficient CL1 cells
4 Discussion
We extensively studied two cell populations within culturedhBMSC that were identified based on differences in mor-phology Cellular and molecular studies revealed differencesin growth differentiation capacity and molecular signatureOur data support the notion of the presence of cellular andfunctional heterogeneity among cultured hBMSC
Cellular heterogeneity of cultured hBMSC is recognizedin an increasing number of reports Several extrinsic andintrinsic factors may contribute to the observed hBMSCheterogeneity Extrinsic factors include donor-to-donor vari-ations in the number and quantity of initiating cells whichresult in differences in cell growth rate and differentia-tion capacity [14 15] Intrinsic factors have been examinedemploying single cell clonal analysis and revealed variationsin differentiation potential among individual colonies withinhBMSC cultures ranging from the presence of cells withtrilineage (osteoblast adipocytes and chondrocyte) potencyto cells with null potency [16] Also variations in the abilityof clonal cells to form heterotopic bone when implanted invivo have been reported [5] Our study corroborates thesefindings and provides more detailed cellular and molecularphenotyping of two examples of cell populations that existwithin the heterogeneous hBMSC cultures [17]
Determining the molecular signature of CL1 and CL2using whole genomemicroarray analysis showed enrichmentof lineage-commitment associated genes inCL1 For exampleinsulin-like growth factor 5 (IGFBP5) and interleukin 6 (IL6)
Stem Cells International 13
NOV0040
0030
0020
0010
mdash
IGFBP50015
0010
0005
mdash
ALP0100
0080
0060
0040
0020
mdash
TAGLN0800
0600
0400
0200
CDH11
0060
0040
0020
mdashmdash
LRP30015
0010
0005
mdash
CL1 CL2 CL1 CL2 CL1 CL2
CL1 CL2CL1 CL2CL1 CL2
Figure 3 Validation of whole genome microarray analysis of two clonal populations of human bone marrow stromal stem cells CL1 andCL2 cells Quantitative real-time PCR for highly expressed genes in CL1 cells NOV = nephroblastoma overexpressed IGFBP5 = insulin-likegrowth factor binding protein 5 ALP TAGLN = transgelin and CDH11 = OB-cadherin (osteoblast) Data are presented as fold change inexpression of each target gene normalized to GAPDH (119899 = three independent experiments 119901 lt 0001) (see also Table 3)
were 147- and 33-fold upregulated in CL1 cells respectivelyBoth factors are expressed in osteoprogenitor cells andimportant for osteoblast maturation [18] We also observedthat periostin (POSTN) gene was highly upregulated inCL1 cells (156-fold) POSTN is a 90 kDa secreted proteinoriginally identified in murine osteoblast-like cells and isupregulated by PTH [19] Several studies employing murineand human cells have revealed important role of POSTNin osteoblast differentiation and during development inintramembranous ossification [20ndash23] Another factor iden-tified in CL1 cells is nephroblastoma overexpression (NOV)which is a member of the Cyr 61 connective tissue growthfactor (CNN) family The CCN family of proteins promotesosteoblast differentiation through interaction with integrinsWNT BMP and NOTCH signaling pathways [24ndash26] Inaddition a large number of signaling molecules known tobe regulators of hBMSC lineage specific differentiation forexample insulin-like growth factors [27ndash29] WNT [30ndash32]and MAPK [33ndash35] were enriched in CL1 cells In contrastCL2 expressed high levels of immune-related genes whichmay explain the poor differentiation response to osteoblastor adipocyte lineage In a recent study the authors usedtelomerized hBMSC and showed clearly a clonal populationthat had very low in vitro and in vivo differentiation abilityhowever they had enhanced immune-related features includ-ing high IL7 expression These nullipotent cells expressedCD317 which was associated with remarkably high basal levelexpression of factors with a proinflammatory and antiviralfunction [17] We observed that this molecular phenotypewas associated with distinct ultrastructural characteristicsof the cells In particular CL2 had abundant phagocyticvacuole microvilli and lysosomes features reminiscent of
ultrastructure of immune-regulatory cells Our data thussupport the increasingly recognized feature that hBMSCexhibit immune modulatory functions and a part of theinnate immune response [17]
We observed that ALP protein expression and enzymaticactivity were significantly different between CL1 and CL2cell lines and were thus a potential marker that distinguishesdifferent cell populations with progenitor functions (CL1)from cells with nonprogenitor functions (CL2) ALP isexpressed in a wide variety of tissues including kidneysbone and liver [36 37] but tissue-nonspecific ALP (ALPL)is considered a commitment marker for osteoblastic lineage[13 38] However in a recent study the authors examined thedifferentiation potential of a number of hMSC clones in vitroand in vivo and reported that the hMSC clones with highlevels of ALP expression were committed to trilineage differ-entiation [13] Our data corroborate and extend these findingsby reporting the effects of siRNA-mediated inhibition ofALP that resulted in an impaired hBMSC differentiation notonly to osteoblasts but also to adipocytes Also our resultscorroborate earlier studies that demonstrated in human bonebiopsies the presence of ALP expression in bone marrowadipocytic cells [13 17] All these data suggest that ALP isa ldquostemnessrdquo marker of hBMSC and not just an indicator ofosteoblastic lineage commitment
While CL1 and CL2 were isolated from telomerizedhMSC cell line they are relevant to normal human physi-ology We observed that the molecular phenotypes of CL1and CL2 were contained within the molecular signatureof primary hBMSC suggesting that CL1 and CL2 repre-sent cell populations within the heterogeneous cultures ofhBMSCWe have also previously reported that the molecular
14 Stem Cells International
ALZ
stai
ning
Control siRNA ALP siRNA
ALP
Beta
-act
in
Con
trol
ALP
siRN
A d3
ALP
siRN
A d7
Nile red quantificationlowast
NI ADIP I ALP siRNA0
50
100
150
200
Con
trol s
iRN
AA
LP si
RNA
Day 3 Day 5 OS induction Day 7 OS induction
Fold
indu
ctio
n lo
g sc
ale
10000
1000
100
10
1
Day 0 OS day 7
Control siRNAALP siRNA
(A) (B)
(C)
(D)
Figure 4 Effect of alkaline phosphatase (ALP) gene silencing by small interfering RNA (siRNA) on a clonal population of human bonemarrow stromal stem cell CL1 (A) ALP staining at day 3 day 5 or day 7 days of osteoblast differentiation (OB induction) (B) Quantitativereal-time PCR for ALP gene following ALP siRNA transfection at day 0 OB and day 7 of OB Data are presented as fold change in expressionof each target gene normalized to GAPDH (119899 = three independent experiments 119901 lt 001) Western blotting analysis of day 3 and day 7after siRNA ALP transfection of CL1 cells ALPL specific antibody and B-actin was used (C) Mineralized matrix formation as visualizedby Alizarin Red S staining in siRNA transfected CL1 cells after 14 days of OB induction (D) Nile red quantification of mature lipid filledadipocyte in control noninduced (Cont) adipocyte induced (Adip I) and ALP siRNA transfected cells that are adipocyte induced (ALPsiRNA) Adipocyte induction was carried out for 7 days lowast119901 lt 005
phenotype and cellular responses of hMSC-TERT are similarto those of primary hMSC [10]While we have identified ALPas a marker that can be used for a prospective identificationof differentiation committed population of hBMSC we iden-tified additional distinctive molecular markers of the cellsFor example IGF-1 IGF-2 and IGF binding protein 5 wereenriched in CL1 compared to CL2 IGFs and their bindingproteins are very well-studied factors that play a role inhBMSC proliferation and osteoblast differentiation [18] On
the other hand annexin A3 as well as several immune-relatedgenes was highly enriched in CL2 compared to CL1 Futurestudies are needed to determine the functional significanceof these molecules in relation to the functional identity ofvarious cell populations within the hBMSC cultures andtheir usefulness as biomarkers to dissect the heterogeneouspopulation of cultured hBMSC
Our finding of the presence of functional diversity withinhBMSC cultures that contain progenitor and nonprogenitor
Stem Cells International 15
Table 4 Microarray data analysis showing genes found to be upregulated during adipogenic differentiation and downregulated after ALPKO
62 common elements inldquoAD uprdquo and ldquoALP downrdquo Gene name FC (ALP siRNA versus
control siRNA)APOBEC3G Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3G minus15938025IFI44L Interferon-induced protein 44-like minus11520283PAQR5 Progestin and adipoQ receptor family member V minus6868241PNMA2 Paraneoplastic antigen MA2 minus59695344DUSP23 Dual specificity phosphatase 23 minus54786854CLDN23 Claudin 23 minus51885047ANKDD1A Ankyrin repeat and death domain containing 1A minus51646647IL8 Interleukin 8 minus4887188LRRC23 Leucine rich repeat containing 23 minus47611775IL6 Interleukin 6 (interferon beta 2) minus4693139LIFR Leukemia inhibitory factor receptor alpha minus46540866PTGFR Prostaglandin F receptor (FP) minus4457529FAM134B Family with sequence similarity 134 member B minus4403495CYFIP2 Cytoplasmic FMR1 interacting protein 2 minus4260462METTL7A Methyltransferase-like 7A minus40480843APOBEC3F Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3F minus39502614CA5B Carbonic anhydrase VB mitochondrial minus393889ITGA10 Integrin alpha 10 minus39143775FMO3 Flavin containing monooxygenase 3 minus3852087IMPA2 Inositol monophosphatase 2 (human) minus38374884CDO1 Cysteine dioxygenase type I minus38181455CCDC68 Coiled-coil domain containing 68 minus37292893
HEXDC Hexosaminidase (glycosyl hydrolase family 20 catalytic domain) containing minus2094836SARM1 Sterile alpha and TIR motif containing 1 minus20797038BBS1 Bardet-Biedl syndrome 1 minus20146718SERPING1 Serpin peptidase inhibitor clade G (C1 inhibitor) member 1 minus20102212FAM162A Family with sequence similarity 162 member A minus2005807TCTN1 Tectonic family member 1 minus20033443
AD up ALP down
1006 62 1145
(a)
140
120
100
080
060
040
020
mdash
ALPL PPARg aP2 LPL
lowastlowastlowastlowastlowastlowast
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
lowast
(b)
Figure 5 Adipocyte related genes downregulated in ALP knockdown CL1 cells (a) Venn diagram of whole gene expression analysis of ADupregulated genes compared with ALP KD downregulated genes (b) Quantitative real-time PCR of four selected common genes from Venndiagram including ALPL PPARg2 aP2 and LPL Data are presented as fold change in expression of each target gene normalized to GAPDH(119899 = three independent experiments lowast119901 lt 005 lowastlowast119901 lt 001 and lowastlowastlowast119901 lt 0001)
Stem Cells International 17
cell populations has a clinical relevance It demonstrates thatthe progenitor function and the immune modulatory rolesof hBMSC [39] are mediated by specific and distinguishablepopulations of hBMSC Thus future clinical studies employ-ing hBMSC should attempt to administer the relevant sub-population of hBMSC dependent on the experimental aim asa novel approach to improving the clinical efficiency insteadof the current use of heterogeneous hBMSC populations
Competing Interests
The authors declare that they have no competing interests
Authorsrsquo Contributions
Mona Elsafadi and Amer Mahmood contributed equally tothis work
Acknowledgments
These studies are supported by a grant from the College ofMedicine Research Centre (CMRC) Deanship of ScientificResearch King Saud University Riyadh Saudi Arabia Theauthors would like to thank Mr Ali Al-Roalle for all histechnical help They thank Natalie Twine for advice on DNAmicroarray analysis
References
[1] W Zaher L Harkness A Jafari and M Kassem ldquoAn update ofhuman mesenchymal stem cell biology and their clinical usesrdquoArchives of Toxicology vol 88 no 5 pp 1069ndash1082 2014
[2] S Post B M Abdallah J F Bentzon andM Kassem ldquoDemon-stration of the presence of independent pre-osteoblastic andpre-adipocytic cell populations in bone marrow-derived mes-enchymal stem cellsrdquo Bone vol 43 no 1 pp 32ndash39 2008
[3] K H Larsen C M Frederiksen J S Burns B M Abdallahand M Kassem ldquoIdentifying a molecular phenotype for bonemarrow stromal cells with in vivo bone-forming capacityrdquoJournal of Bone and Mineral Research vol 25 no 4 pp 796ndash808 2010
[4] A Rosenzweig ldquoCardiac cell therapymdashmixed results frommixed cellsrdquoThe New England Journal of Medicine vol 355 no12 pp 1274ndash1277 2006
[5] S A Kuznetsov P H Krebsbach K Satomura et al ldquoSingle-colony derived strains of human marrow stromal fibroblastsform bone after transplantation in vivordquo Journal of Bone andMineral Research vol 12 no 9 pp 1335ndash1347 1997
[6] M Kassem and P Bianco ldquoSkeletal stem cells in space and timerdquoCell vol 160 no 1-2 pp 17ndash19 2015
[7] A I Caplan and D Correa ldquoPDGF in bone formation andregeneration new insights into a novel mechanism involvingMSCsrdquo Journal of Orthopaedic Research vol 29 no 12 pp 1795ndash1803 2011
[8] K Stewart S Walsh J Screen et al ldquoFurther characterizationof cells expressing STRO-1 in cultures of adult human bonemarrow stromal cellsrdquo Journal of Bone and Mineral Researchvol 14 no 8 pp 1345ndash1356 1999
[9] J L Simonsen C Rosada N Serakinci et al ldquoTelomeraseexpression extends the proliferative life-span and maintains
the osteogenic potential of human bone marrow stromal cellsrdquoNature Biotechnology vol 20 no 6 pp 592ndash596 2002
[10] M Al-Nbaheen R Vishnubalaji D Ali et al ldquoHuman stromal(mesenchymal) stem cells from bone marrow adipose tissueand skin exhibit differences in molecular phenotype and differ-entiation potentialrdquo Stem Cell Reviews and Reports vol 9 no 1pp 32ndash43 2013
[11] C A Gregory W G Gunn A Peister and D J Prockop ldquoAnAlizarin red-based assay of mineralization by adherent cells inculture comparison with cetylpyridinium chloride extractionrdquoAnalytical Biochemistry vol 329 no 1 pp 77ndash84 2004
[12] AMahmood L Harkness H D Schroslashder B M Abdallah andM Kassem ldquoEnhanced differentiation of human embryonicstem cells to mesenchymal progenitors by inhibition of TGF-120573activinnodal signaling using SB-431542rdquo Journal of Bone andMineral Research vol 25 no 6 pp 1216ndash1233 2010
[13] H-J Prins A K Braat D Gawlitta et al ldquoIn vitro inductionof alkaline phosphatase levels predicts in vivo bone formingcapacity of human bone marrow stromal cellsrdquo Stem CellResearch vol 12 no 2 pp 428ndash440 2014
[14] K Stenderup J Justesen C Clausen and M Kassem ldquoAgingis associated with decreased maximal life span and acceleratedsenescence of bone marrow stromal cellsrdquo Bone vol 33 no 6pp 919ndash926 2003
[15] D G Phinney G Kopen W Righter S Webster N Tremainand D J Prockop ldquoDonor variation in the growth propertiesand osteogenic potential of human marrow stromal cellsrdquoJournal of Cellular Biochemistry vol 75 no 3 pp 424ndash436 1999
[16] C M Digirolamo D Stokes D Colter D G Phinney R Classand D J Prockop ldquoPropagation and senescence of humanmarrow stromal cells in culture a simple colony-forming assayidentifies samples with the greatest potential to propagate anddifferentiaterdquo British Journal of Haematology vol 107 no 2 pp275ndash281 1999
[17] S James J Fox F Afsari et al ldquoMultiparameter analy-sis of human bone marrow stromal cells identifies distinctimmunomodulatory and differentiation-competent subtypesrdquoStem Cell Reports vol 4 no 6 pp 1004ndash1015 2015
[18] B Peruzzi A Cappariello A Del Fattore N Rucci F DeBenedetti and A Teti ldquoc-Src and IL-6 inhibit osteoblastdifferentiation and integrate IGFBP5 signallingrdquo Nature Com-munications vol 3 article 630 2012
[19] D Fortunati S Reppe A-K Fjeldheim M Nielsen V TGautvik and K M Gautvik ldquoPeriostin is a collagen associatedbonematrix protein regulated by parathyroid hormonerdquoMatrixBiology vol 29 no 7 pp 594ndash601 2010
[20] A Oshima H Tanabe T Yan G N Lowe C A Glackin andA Kudo ldquoA novel mechanism for the regulation of osteoblastdifferentiation transcription of periostin a member of thefasciclin I family is regulated by the bHLH transcription factorTwistrdquo Journal of Cellular Biochemistry vol 86 no 4 pp 792ndash804 2002
[21] T G Kashima T Nishiyama K Shimazu et al ldquoPeriostina novel marker of intramembranous ossification is expressedin fibrous dysplasia and in c-Fos-overexpressing bone lesionsrdquoHuman Pathology vol 40 no 2 pp 226ndash237 2009
[22] CGalliM PiergianniM Piemontese et al ldquoPeriostin improvescell adhesion to implantable biomaterials and osteoblasticdifferentiation on implant titanium surfaces in a topography-dependent fashionrdquo Journal of Biomedical Materials Research -Part A vol 102 no 11 pp 3855ndash3861 2014
18 Stem Cells International
[23] D L Coutu H W Jian A Monette G-E Rivard M DBlostein and J Galipeau ldquoPeriostin amember of a novel familyof vitamin K-dependent proteins is expressed by mesenchymalstromal cellsrdquo The Journal of Biological Chemistry vol 283 no26 pp 17991ndash18001 2008
[24] T Minamizato K Sakamoto T Liu et al ldquoCCN3NOV inhibitsBMP-2-induced osteoblast differentiation by interacting withBMP andNotch signaling pathwaysrdquo Biochemical and Biophysi-cal Research Communications vol 354 no 2 pp 567ndash573 2007
[25] W Si Q Kang H H Luu et al ldquoCCN1Cyr61 is regulated bythe canonicalWnt signal and plays an important role inWnt3A-induced osteoblast differentiation of mesenchymal stem cellsrdquoMolecular and Cellular Biology vol 26 no 8 pp 2955ndash29642006
[26] Q Luo Q Kang W Si et al ldquoConnective tissue growth factor(CTGF) is regulated by Wnt and bone morphogenetic proteinssignaling in osteoblast differentiation of mesenchymal stemcellsrdquo The Journal of Biological Chemistry vol 279 no 53 pp55958ndash55968 2004
[27] Y-S Yoon N Lee and H Scadova ldquoMyocardial regenerationwith bone-marrow-derived stem cellsrdquo Biology of the Cell vol97 no 4 pp 253ndash263 2005
[28] Y S Yoon ldquoCardiac regeneration with novel bone marrow-derived multipotent stem cellsrdquo Discovery Medicine vol 5 pp204ndash208 2005
[29] Y-S Yoon A Wecker L Heyd et al ldquoClonally expanded novelmultipotent stem cells from human bone marrow regeneratemyocardium after myocardial infarctionrdquo Journal of ClinicalInvestigation vol 115 no 2 pp 326ndash338 2005
[30] GM BolandG Perkins D J Hall andR S Tuan ldquoWnt 3a pro-motes proliferation and suppresses osteogenic differentiationof adult human mesenchymal stem cellsrdquo Journal of CellularBiochemistry vol 93 no 6 pp 1210ndash1230 2004
[31] D A Cook S W Fellgett M E Pownall P J OrsquoShea and PG Genever ldquoWnt-dependent osteogenic commitment of bonemarrow stromal cells using a novel GSK3120573 inhibitorrdquo Stem CellResearch vol 12 no 2 pp 415ndash427 2014
[32] J De Boer R Siddappa C Gaspar A VanApeldoorn R Foddeand C Van Blitterswijk ldquoWnt signaling inhibits osteogenicdifferentiation of humanmesenchymal stem cellsrdquoBone vol 34no 5 pp 818ndash826 2004
[33] MAl-Toub A AlmusaM Almajed et al ldquoPleiotropic effects ofcancer cellsrsquo secreted factors on human stromal (mesenchymal)stem cellsrdquo Stem Cell Research and Therapy vol 4 article 1142013
[34] L Chen D Qanie A Jafari et al ldquoDelta-like 1fetal antigen-1(Dlk1FA1) is a novel regulator of chondrogenic cell differenti-ation via inhibition of the Akt kinase-dependent pathwayrdquoTheJournal of Biological Chemistry vol 286 no 37 pp 32140ndash321492011
[35] T Matsunobu K Torigoe M Ishikawa et al ldquoCritical roles ofthe TGF-120573 type I receptor ALK5 in perichondrial formationand function cartilage integrity and osteoblast differentiationduring growth plate developmentrdquo Developmental Biology vol332 no 2 pp 325ndash338 2009
[36] B C Nair D E Johnson R J Majeska J A Rodkey C DBennett and G A Rodan ldquoRat alkaline phosphatase II Struc-tural similarities between the osteosarcoma bone kidney andplacenta isoenzymesrdquo Archives of Biochemistry and Biophysicsvol 254 no 1 pp 28ndash34 1987
[37] P J Butterworth ldquoAlkaline phosphatase Biochemistry of mam-malian alkaline phosphatasesrdquo Cell Biochemistry and Functionvol 1 no 2 pp 66ndash70 1983
[38] A T Ali C B Penny J E Paiker et al ldquoAlkaline phosphataseis involved in the control of adipogenesis in the murinepreadipocyte cell line 3T3-L1rdquo Clinica Chimica Acta vol 354no 1-2 pp 101ndash109 2005
[39] G M Spaggiari and L Moretta ldquoCellular and molecularinteractions of mesenchymal stem cells in innate immunityrdquoImmunology and Cell Biology vol 91 no 1 pp 27ndash31 2012
Figure 1 Cellular phenotype of two clonal populations of humanbonemarrow stromal stem cells CL1 andCL2 (a) Cellmorphology CL1 cellsshowed large cuboidal morphology while CL2 cells had spindle-shaped fibroblast-like morphology (b) Growth curves showing populationdoubling (PD) rate during long term culture (c) Flow cytometry analysis (FACS) shows expression of CD44 CD63 CD73 CD105 andCD166in CL1 and CL2 cells Matched isotype control was used for gating (d) Flow cytometry analysis presented as dot blot of CD146 and alkalinephosphatase (ALP) cell surface proteins (e) Transmission electron microscope (TEM) (A) CL1 (1200x) (B) CL2 (1500x) n nucleus nunucleolus rER rough endoplasmic reticulum ly lysosomes pv phagocytic vacuole and rer reticular stalk of rER (f) Gene expressionanalysis using RT-PCR for a group of mesodermal and stromal genes Gene expression was normalized to GAPDH and presented as foldchange Data is shown as mean plusmn SD of three independent experiments lowast119901 lt 005 119901 lt 0001
collagen type V alpha 3 (COL5A3) osteomodulin (OMD)and T-box 15 (TBX15) (Table 1) Furthermore several knownosteoblast-related genes were upregulated in CL1 cells suchas bone morphogenetic protein 6 (BMP6) fibroblast growthfactor receptor 3 (FGFR3) insulin-like growth factor bindingprotein 5 (IGFBP5) and vitamin D (125-dihydroxyvitaminD3) receptor (VDR) (Table 1) On the other hand 1200 geneswere upregulated in CL2 cells 255 genes were annotated toimmunity and immune response and defense This categoryincluded genes from complement system chemokine (C-Cmotif) ligands interferon family chemokine (C-X-C motif)ligands and receptor major histocompatibility complex classII molecules interleukins and tumor necrosis factor receptorsuperfamily (Table 2 and Supplementary Table 3)
33 CL1 Cells Exhibit Enhanced Adipocyte DifferentiationWe observed significant differences between CL1 and CL2in their response to adipocytic differentiation inductionCL1 differentiated readily to adipocytes compared to CL2(Figure 2(d) lower panel) evidenced by higher levels ofadipocyticmarkers gene expression LPL (lipoprotein lipase)and adiponectin as well as formation of mature lipid filledadipocytes visualized by Oil Red-O staining and quantitativeNile red staining (Figure 2(d))
34 CL1 Cells Differentiate to Chondrocytic Lineage Inpellet cultures CL1 cells formed 3D pellets containingproteoglycan-secreting chondrocytes which stained positivewith Alcian blue Limited chondrocyte differentiation wasvisible in cell pellets of CL2 cells The differentiated chon-drocytes in CL1 pellets expressed higher levels of collagen Xand collagen II which was overlapping the Alcian blue stain(Figure 2(e))
35 Molecular Signature of CL1 and CL2 Cells To define themolecular signature and molecular differences between CL1
and CL2 we compared the basal gene expression patternof CL1 and CL2 cells using DNA microarrays The PCAanalysis showed a clear separation between CL1 and CL2(Supplementary Figure 1) Comparison between CL1 withCL2 showed that 915 geneswere differentially expressed in thetwo cell lines (gt2-fold119901 lt 001) 462 genes were upregulatedand 452 were downregulated in CL1 versus CL2 The mostrelevant genes that were upregulated in CL1 are listed inTable 3(a) Among these 35 highly expressed genes in CL1 thefollowing 11 genes were present in skeletal and muscular sys-tem development and function FOLR3 CCL3L1 SERPINB2POSTN IGFBP5 CCL3 NOV ALP TNFRSF11B ACTG2and CDH11 (Table 3(a)) Functional annotation of the upreg-ulated genes in CL1 using the Ingenuity Pathway Analysis(IPA) revealed enrichment in the following categories ldquotissuedevelopmentrdquo ldquoskeletal and muscular system developmentand functionrdquo and ldquoorganismal developmentrdquo (Table 3(b))Furthermore the DAVID annotation tool was employed toassess the functional relationships of the upregulated genesin CL1 showing enrichment in ontologies ldquoskeletal andmuscular system development and functionrdquo that includedbone size osteoblast differentiation bone mineralizationand bone mineral density (Table 3(c)) CL1 exhibited upreg-ulation of WNT pathway ligands WNT5B (2-fold) andLRP5 (2-fold) (Table 1) Also ALP was among the highlyexpressed genes together with POSTN IGFBP5 SPP1 IL-6 and DKK1 (Tables 3(a) and 1) These genes are knownto play an important role in osteoblast differentiation andbone formation For CL2 inhibitors of WNT pathway wereupregulated and included SFRP1 (11-fold) DKK2 (32-fold)FGF2 (31-fold) and GBP2 (24-fold) Functional annotationof the upregulated genes in CL2 revealed enrichment in thefollowing categories ldquodevelopmental processrdquo ldquomulticellularorganismal processrdquo ldquobiological adhesionrdquo and ldquoimmunesystem processrdquo (Supplementary Table 2A) In-depth analysis
Stem Cells International 7
CL1
CL2
ALPCont D 14
Cont OB I
ALP activity
mdash100
200
300
400
500
600
Fold
chan
ge
CL1CL2
(a)
CL1
CL2
ALZCont D 14
Fold
chan
ge
25
20
15
10
5
0
Cont d14 d21
CL1CL2
ALZ quantification
(b)
CL1CL2
CL1CL2
LPL ADIPOQ
Cont ContADIP D14 ADIP D14
000
100
200
300
400
500
600
700
0
1000
2000
3000
4000
5000
CL1CL2
CL1CL2
CL1CL2
RUNX2 ALP OPN
Cont Cont ContOS D7 OS D7 OS D7
00
05
10
15
20
25
30
00
05
10
15
20
25
30
35
00
05
10
15
20
25
30
35
(c)
CL1CL2
CL1
CL2
Control Adip d7 Adip d14
d7 d140
005
01
015
02
025
03
035
04
045
Nile
red
fluor
esce
nt in
tens
ity
(d)
CL1
CL2
Alcian blue Col X Col II
(e)
Figure 2 In vitro osteoblastic and adipocytic differentiation of two clonal populations of human bone marrow stromal stem cells CL1 andCL2 cells Both cell lines were induced for osteoblast differentiation using standard protocol described in the Methods (a) ALP staining atday 14 in control noninduced (Cont) and osteoblast induced cells (D14) Right panel shows ALP activity (119899 = 3 independent experiments119901 lt 0001) (b) Mineralized matrix formation visualized by Alizarin Red S staining Right panel shows Alizarin Red quantification at day 14and 21 after osteoblast differentiation (119899 = 3 independent experiments 119901 lt 0001) (c) Quantitative RT-PCR of osteoblastic and adipocytegene markers in CL1 and CL2 during osteoblast (upper panel) and adipocyte (lower panel) differentiation ALP = alkaline phosphataseOPN = osteopontin LPL = lipoprotein lipase and ADIPOQ = adiponectin Data are presented as fold change in expression of each targetgene normalized to GAPDH (119899 = 3 independent experiments 119901 lt 005
119901 lt 0001) (d) CL1 and CL2 lines were induced for adipocytedifferentiation using standard protocol described in the Methods Adipocyte formation was visualized at day 7 (Adip d7) and day 14 (Adipd14) by Oil Red-O staining Lower panel presents quantification of Nile red staining (119899 = 3 independent experiments 119901 lt 0001) (e) CL1and CL2 lines were induced for chondrocyte differentiation using 21-day pellet culture method as described in the Methods The pellets werestained with Alcian blue collagen 10 (Col X) and collagen 2 (Col II) (original magnification 5x)
8 Stem Cells International
Table 1 Microarray data analysis showing genes related to bone development and osteoblast differentiation upregulated in CL1 versus CL2cells
Probe ID Genbank accession Gene name Gene symbol FCA 23 P167367 NM 153426 Paired-like homeodomain 2 PITX2 30884
of the biological processes revealed several immune-relatedpathways ldquoMAPKKK cascaderdquo ldquoimmunity and defenserdquoldquosignal transductionrdquo ldquoextracellular matrix protein-mediatedsignalingrdquo and ldquointerferon-mediated immunityrdquo among oth-ers that were upregulated (Supplementary Table 2B) Also40 genes related to immune system related factors wereidentified as significantly enriched in CL2 compared to CL1cells (Supplementary Table 2C)We chose the following genesfor validation of the microarray results NOV IGFBP5 ALPTAGLN and CDH11 as they were highly expressed in CL1RT-PCR analysis confirmed themicroarray results (Figure 3)
Furthermore we compared the molecular phenotype ofCL1 and CL2 cells with that of phBMSCWe found that morethan 80 of the genes expressed in CL1 and 90 in CL2 cellswere commonwith primary phBMSC (Supplementary Figure2) suggesting that CL1 and CL2 molecular phenotype existwithin the heterogeneous population of phBMSC cultures
36 ALP Knockdown Impairs Differentiation of CL1 CellsSince ALP has been suggested as a marker for hBMSCprogenitor cell lineage commitment [13] and was highlyupregulated in CL1 cells we tested its biological role in
CL1 cells ALP siRNA transfection decreased ALP proteinlevel ALP activity and mRNA gene expression compared tocontrol cells transfected with control siRNA (119901 lt 001) andthis inhibition was detectable up to day 7 days after osteoblastdifferentiation induction (Figures 4(A) and 4(B)) At day 14 ofdifferentiation mineralization ability of CL1 was significantlyimpaired (Figure 4(C)) In addition we found that thenumber of mature adipocyte formations was significantlyreduced to more than 75 (119901 lt 001) (Figure 4(D))
To identify relevant adipocyte differentiation associatedgenes that were targeted by ALP deficiency we compared thedownregulated genes of ALP deficient CL1 with the upreg-ulated genes identified during adipocytic differentiation ofCL1 We identified 62 genes that were common (Figure 5(a)Table 4) and among these genes were genes related tometabolism (primarily lipid and carbohydrate) and transportincluding CYB5B CHST1 TAP1 ATP8A1 LRP8 PLCD1and FABP5 (Table 4) We further performed quantitativereal-time PCR of ALP deficient CL1 cells during adipocytedifferentiation The following adipocyte-associated geneswere downregulated PPAR1205742 LPL and aP2 (Figure 5(b))
12 Stem Cells International
Table 3 Whole genome microarray analysis of two clonal popu-lations of human bone marrow stromal stem cells CL1 and CL2cells (a) Top 35 highly upregulated genes in CL1 cells versus CL2cells (b) Ingenuity Pathway Analysis (IPA) showing the differentphysiological system development and function genes found in eachcategory and the corresponding 119901 value (c) Upper part IPA analysisshowing the categories for skeletal and muscular system developmentand function upregulated in CL1 cells and lower part showing genesupregulated in CL2 cells NOV = nephroblastoma overexpressedIGFBP5 = insulin-like growth factor binding protein 5 ALP TAGLN= transgelin and CDH11 = OB-cadherin (osteoblast)
Name 119901 value moleculesPhysiological system development and function
Organismal development 705119864 minus 10ndash143119864 minus 03 181Embryonic development 100119864 minus 09ndash143119864 minus 03 154Organ development 100119864 minus 09ndash143119864 minus 03 145
(b) Continued
Name 119901 value moleculesSkeletal and muscularsystem development andfunction
100119864 minus 09ndash102119864 minus 03 123
Tissue development 100119864 minus 09ndash143119864 minus 03 236
(c)
Functions annotation 119901 value moleculesSkeletal and muscular system development and
function upregulated in CL1 cellsSize of bone 143119864 minus 06 24Differentiation ofosteoblasts 384119864 minus 06 25
Mineralization ofbone 493119864 minus 06 19
Bone mineral density 365119864 minus 05 19Skeletal and muscular system development and
function upregulated in CL2 cellsDevelopment ofmuscle 100119864 minus 09 44
Proliferation ofmuscle cells 229119864 minus 06 35
Remodeling of bone 363119864 minus 06 21Resorption of bone 394119864 minus 06 19
confirming impairment of adipocytic differentiation of ALPdeficient CL1 cells
4 Discussion
We extensively studied two cell populations within culturedhBMSC that were identified based on differences in mor-phology Cellular and molecular studies revealed differencesin growth differentiation capacity and molecular signatureOur data support the notion of the presence of cellular andfunctional heterogeneity among cultured hBMSC
Cellular heterogeneity of cultured hBMSC is recognizedin an increasing number of reports Several extrinsic andintrinsic factors may contribute to the observed hBMSCheterogeneity Extrinsic factors include donor-to-donor vari-ations in the number and quantity of initiating cells whichresult in differences in cell growth rate and differentia-tion capacity [14 15] Intrinsic factors have been examinedemploying single cell clonal analysis and revealed variationsin differentiation potential among individual colonies withinhBMSC cultures ranging from the presence of cells withtrilineage (osteoblast adipocytes and chondrocyte) potencyto cells with null potency [16] Also variations in the abilityof clonal cells to form heterotopic bone when implanted invivo have been reported [5] Our study corroborates thesefindings and provides more detailed cellular and molecularphenotyping of two examples of cell populations that existwithin the heterogeneous hBMSC cultures [17]
Determining the molecular signature of CL1 and CL2using whole genomemicroarray analysis showed enrichmentof lineage-commitment associated genes inCL1 For exampleinsulin-like growth factor 5 (IGFBP5) and interleukin 6 (IL6)
Stem Cells International 13
NOV0040
0030
0020
0010
mdash
IGFBP50015
0010
0005
mdash
ALP0100
0080
0060
0040
0020
mdash
TAGLN0800
0600
0400
0200
CDH11
0060
0040
0020
mdashmdash
LRP30015
0010
0005
mdash
CL1 CL2 CL1 CL2 CL1 CL2
CL1 CL2CL1 CL2CL1 CL2
Figure 3 Validation of whole genome microarray analysis of two clonal populations of human bone marrow stromal stem cells CL1 andCL2 cells Quantitative real-time PCR for highly expressed genes in CL1 cells NOV = nephroblastoma overexpressed IGFBP5 = insulin-likegrowth factor binding protein 5 ALP TAGLN = transgelin and CDH11 = OB-cadherin (osteoblast) Data are presented as fold change inexpression of each target gene normalized to GAPDH (119899 = three independent experiments 119901 lt 0001) (see also Table 3)
were 147- and 33-fold upregulated in CL1 cells respectivelyBoth factors are expressed in osteoprogenitor cells andimportant for osteoblast maturation [18] We also observedthat periostin (POSTN) gene was highly upregulated inCL1 cells (156-fold) POSTN is a 90 kDa secreted proteinoriginally identified in murine osteoblast-like cells and isupregulated by PTH [19] Several studies employing murineand human cells have revealed important role of POSTNin osteoblast differentiation and during development inintramembranous ossification [20ndash23] Another factor iden-tified in CL1 cells is nephroblastoma overexpression (NOV)which is a member of the Cyr 61 connective tissue growthfactor (CNN) family The CCN family of proteins promotesosteoblast differentiation through interaction with integrinsWNT BMP and NOTCH signaling pathways [24ndash26] Inaddition a large number of signaling molecules known tobe regulators of hBMSC lineage specific differentiation forexample insulin-like growth factors [27ndash29] WNT [30ndash32]and MAPK [33ndash35] were enriched in CL1 cells In contrastCL2 expressed high levels of immune-related genes whichmay explain the poor differentiation response to osteoblastor adipocyte lineage In a recent study the authors usedtelomerized hBMSC and showed clearly a clonal populationthat had very low in vitro and in vivo differentiation abilityhowever they had enhanced immune-related features includ-ing high IL7 expression These nullipotent cells expressedCD317 which was associated with remarkably high basal levelexpression of factors with a proinflammatory and antiviralfunction [17] We observed that this molecular phenotypewas associated with distinct ultrastructural characteristicsof the cells In particular CL2 had abundant phagocyticvacuole microvilli and lysosomes features reminiscent of
ultrastructure of immune-regulatory cells Our data thussupport the increasingly recognized feature that hBMSCexhibit immune modulatory functions and a part of theinnate immune response [17]
We observed that ALP protein expression and enzymaticactivity were significantly different between CL1 and CL2cell lines and were thus a potential marker that distinguishesdifferent cell populations with progenitor functions (CL1)from cells with nonprogenitor functions (CL2) ALP isexpressed in a wide variety of tissues including kidneysbone and liver [36 37] but tissue-nonspecific ALP (ALPL)is considered a commitment marker for osteoblastic lineage[13 38] However in a recent study the authors examined thedifferentiation potential of a number of hMSC clones in vitroand in vivo and reported that the hMSC clones with highlevels of ALP expression were committed to trilineage differ-entiation [13] Our data corroborate and extend these findingsby reporting the effects of siRNA-mediated inhibition ofALP that resulted in an impaired hBMSC differentiation notonly to osteoblasts but also to adipocytes Also our resultscorroborate earlier studies that demonstrated in human bonebiopsies the presence of ALP expression in bone marrowadipocytic cells [13 17] All these data suggest that ALP isa ldquostemnessrdquo marker of hBMSC and not just an indicator ofosteoblastic lineage commitment
While CL1 and CL2 were isolated from telomerizedhMSC cell line they are relevant to normal human physi-ology We observed that the molecular phenotypes of CL1and CL2 were contained within the molecular signatureof primary hBMSC suggesting that CL1 and CL2 repre-sent cell populations within the heterogeneous cultures ofhBMSCWe have also previously reported that the molecular
14 Stem Cells International
ALZ
stai
ning
Control siRNA ALP siRNA
ALP
Beta
-act
in
Con
trol
ALP
siRN
A d3
ALP
siRN
A d7
Nile red quantificationlowast
NI ADIP I ALP siRNA0
50
100
150
200
Con
trol s
iRN
AA
LP si
RNA
Day 3 Day 5 OS induction Day 7 OS induction
Fold
indu
ctio
n lo
g sc
ale
10000
1000
100
10
1
Day 0 OS day 7
Control siRNAALP siRNA
(A) (B)
(C)
(D)
Figure 4 Effect of alkaline phosphatase (ALP) gene silencing by small interfering RNA (siRNA) on a clonal population of human bonemarrow stromal stem cell CL1 (A) ALP staining at day 3 day 5 or day 7 days of osteoblast differentiation (OB induction) (B) Quantitativereal-time PCR for ALP gene following ALP siRNA transfection at day 0 OB and day 7 of OB Data are presented as fold change in expressionof each target gene normalized to GAPDH (119899 = three independent experiments 119901 lt 001) Western blotting analysis of day 3 and day 7after siRNA ALP transfection of CL1 cells ALPL specific antibody and B-actin was used (C) Mineralized matrix formation as visualizedby Alizarin Red S staining in siRNA transfected CL1 cells after 14 days of OB induction (D) Nile red quantification of mature lipid filledadipocyte in control noninduced (Cont) adipocyte induced (Adip I) and ALP siRNA transfected cells that are adipocyte induced (ALPsiRNA) Adipocyte induction was carried out for 7 days lowast119901 lt 005
phenotype and cellular responses of hMSC-TERT are similarto those of primary hMSC [10]While we have identified ALPas a marker that can be used for a prospective identificationof differentiation committed population of hBMSC we iden-tified additional distinctive molecular markers of the cellsFor example IGF-1 IGF-2 and IGF binding protein 5 wereenriched in CL1 compared to CL2 IGFs and their bindingproteins are very well-studied factors that play a role inhBMSC proliferation and osteoblast differentiation [18] On
the other hand annexin A3 as well as several immune-relatedgenes was highly enriched in CL2 compared to CL1 Futurestudies are needed to determine the functional significanceof these molecules in relation to the functional identity ofvarious cell populations within the hBMSC cultures andtheir usefulness as biomarkers to dissect the heterogeneouspopulation of cultured hBMSC
Our finding of the presence of functional diversity withinhBMSC cultures that contain progenitor and nonprogenitor
Stem Cells International 15
Table 4 Microarray data analysis showing genes found to be upregulated during adipogenic differentiation and downregulated after ALPKO
62 common elements inldquoAD uprdquo and ldquoALP downrdquo Gene name FC (ALP siRNA versus
control siRNA)APOBEC3G Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3G minus15938025IFI44L Interferon-induced protein 44-like minus11520283PAQR5 Progestin and adipoQ receptor family member V minus6868241PNMA2 Paraneoplastic antigen MA2 minus59695344DUSP23 Dual specificity phosphatase 23 minus54786854CLDN23 Claudin 23 minus51885047ANKDD1A Ankyrin repeat and death domain containing 1A minus51646647IL8 Interleukin 8 minus4887188LRRC23 Leucine rich repeat containing 23 minus47611775IL6 Interleukin 6 (interferon beta 2) minus4693139LIFR Leukemia inhibitory factor receptor alpha minus46540866PTGFR Prostaglandin F receptor (FP) minus4457529FAM134B Family with sequence similarity 134 member B minus4403495CYFIP2 Cytoplasmic FMR1 interacting protein 2 minus4260462METTL7A Methyltransferase-like 7A minus40480843APOBEC3F Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3F minus39502614CA5B Carbonic anhydrase VB mitochondrial minus393889ITGA10 Integrin alpha 10 minus39143775FMO3 Flavin containing monooxygenase 3 minus3852087IMPA2 Inositol monophosphatase 2 (human) minus38374884CDO1 Cysteine dioxygenase type I minus38181455CCDC68 Coiled-coil domain containing 68 minus37292893
HEXDC Hexosaminidase (glycosyl hydrolase family 20 catalytic domain) containing minus2094836SARM1 Sterile alpha and TIR motif containing 1 minus20797038BBS1 Bardet-Biedl syndrome 1 minus20146718SERPING1 Serpin peptidase inhibitor clade G (C1 inhibitor) member 1 minus20102212FAM162A Family with sequence similarity 162 member A minus2005807TCTN1 Tectonic family member 1 minus20033443
AD up ALP down
1006 62 1145
(a)
140
120
100
080
060
040
020
mdash
ALPL PPARg aP2 LPL
lowastlowastlowastlowastlowastlowast
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
lowast
(b)
Figure 5 Adipocyte related genes downregulated in ALP knockdown CL1 cells (a) Venn diagram of whole gene expression analysis of ADupregulated genes compared with ALP KD downregulated genes (b) Quantitative real-time PCR of four selected common genes from Venndiagram including ALPL PPARg2 aP2 and LPL Data are presented as fold change in expression of each target gene normalized to GAPDH(119899 = three independent experiments lowast119901 lt 005 lowastlowast119901 lt 001 and lowastlowastlowast119901 lt 0001)
Stem Cells International 17
cell populations has a clinical relevance It demonstrates thatthe progenitor function and the immune modulatory rolesof hBMSC [39] are mediated by specific and distinguishablepopulations of hBMSC Thus future clinical studies employ-ing hBMSC should attempt to administer the relevant sub-population of hBMSC dependent on the experimental aim asa novel approach to improving the clinical efficiency insteadof the current use of heterogeneous hBMSC populations
Competing Interests
The authors declare that they have no competing interests
Authorsrsquo Contributions
Mona Elsafadi and Amer Mahmood contributed equally tothis work
Acknowledgments
These studies are supported by a grant from the College ofMedicine Research Centre (CMRC) Deanship of ScientificResearch King Saud University Riyadh Saudi Arabia Theauthors would like to thank Mr Ali Al-Roalle for all histechnical help They thank Natalie Twine for advice on DNAmicroarray analysis
References
[1] W Zaher L Harkness A Jafari and M Kassem ldquoAn update ofhuman mesenchymal stem cell biology and their clinical usesrdquoArchives of Toxicology vol 88 no 5 pp 1069ndash1082 2014
[2] S Post B M Abdallah J F Bentzon andM Kassem ldquoDemon-stration of the presence of independent pre-osteoblastic andpre-adipocytic cell populations in bone marrow-derived mes-enchymal stem cellsrdquo Bone vol 43 no 1 pp 32ndash39 2008
[3] K H Larsen C M Frederiksen J S Burns B M Abdallahand M Kassem ldquoIdentifying a molecular phenotype for bonemarrow stromal cells with in vivo bone-forming capacityrdquoJournal of Bone and Mineral Research vol 25 no 4 pp 796ndash808 2010
[4] A Rosenzweig ldquoCardiac cell therapymdashmixed results frommixed cellsrdquoThe New England Journal of Medicine vol 355 no12 pp 1274ndash1277 2006
[5] S A Kuznetsov P H Krebsbach K Satomura et al ldquoSingle-colony derived strains of human marrow stromal fibroblastsform bone after transplantation in vivordquo Journal of Bone andMineral Research vol 12 no 9 pp 1335ndash1347 1997
[6] M Kassem and P Bianco ldquoSkeletal stem cells in space and timerdquoCell vol 160 no 1-2 pp 17ndash19 2015
[7] A I Caplan and D Correa ldquoPDGF in bone formation andregeneration new insights into a novel mechanism involvingMSCsrdquo Journal of Orthopaedic Research vol 29 no 12 pp 1795ndash1803 2011
[8] K Stewart S Walsh J Screen et al ldquoFurther characterizationof cells expressing STRO-1 in cultures of adult human bonemarrow stromal cellsrdquo Journal of Bone and Mineral Researchvol 14 no 8 pp 1345ndash1356 1999
[9] J L Simonsen C Rosada N Serakinci et al ldquoTelomeraseexpression extends the proliferative life-span and maintains
the osteogenic potential of human bone marrow stromal cellsrdquoNature Biotechnology vol 20 no 6 pp 592ndash596 2002
[10] M Al-Nbaheen R Vishnubalaji D Ali et al ldquoHuman stromal(mesenchymal) stem cells from bone marrow adipose tissueand skin exhibit differences in molecular phenotype and differ-entiation potentialrdquo Stem Cell Reviews and Reports vol 9 no 1pp 32ndash43 2013
[11] C A Gregory W G Gunn A Peister and D J Prockop ldquoAnAlizarin red-based assay of mineralization by adherent cells inculture comparison with cetylpyridinium chloride extractionrdquoAnalytical Biochemistry vol 329 no 1 pp 77ndash84 2004
[12] AMahmood L Harkness H D Schroslashder B M Abdallah andM Kassem ldquoEnhanced differentiation of human embryonicstem cells to mesenchymal progenitors by inhibition of TGF-120573activinnodal signaling using SB-431542rdquo Journal of Bone andMineral Research vol 25 no 6 pp 1216ndash1233 2010
[13] H-J Prins A K Braat D Gawlitta et al ldquoIn vitro inductionof alkaline phosphatase levels predicts in vivo bone formingcapacity of human bone marrow stromal cellsrdquo Stem CellResearch vol 12 no 2 pp 428ndash440 2014
[14] K Stenderup J Justesen C Clausen and M Kassem ldquoAgingis associated with decreased maximal life span and acceleratedsenescence of bone marrow stromal cellsrdquo Bone vol 33 no 6pp 919ndash926 2003
[15] D G Phinney G Kopen W Righter S Webster N Tremainand D J Prockop ldquoDonor variation in the growth propertiesand osteogenic potential of human marrow stromal cellsrdquoJournal of Cellular Biochemistry vol 75 no 3 pp 424ndash436 1999
[16] C M Digirolamo D Stokes D Colter D G Phinney R Classand D J Prockop ldquoPropagation and senescence of humanmarrow stromal cells in culture a simple colony-forming assayidentifies samples with the greatest potential to propagate anddifferentiaterdquo British Journal of Haematology vol 107 no 2 pp275ndash281 1999
[17] S James J Fox F Afsari et al ldquoMultiparameter analy-sis of human bone marrow stromal cells identifies distinctimmunomodulatory and differentiation-competent subtypesrdquoStem Cell Reports vol 4 no 6 pp 1004ndash1015 2015
[18] B Peruzzi A Cappariello A Del Fattore N Rucci F DeBenedetti and A Teti ldquoc-Src and IL-6 inhibit osteoblastdifferentiation and integrate IGFBP5 signallingrdquo Nature Com-munications vol 3 article 630 2012
[19] D Fortunati S Reppe A-K Fjeldheim M Nielsen V TGautvik and K M Gautvik ldquoPeriostin is a collagen associatedbonematrix protein regulated by parathyroid hormonerdquoMatrixBiology vol 29 no 7 pp 594ndash601 2010
[20] A Oshima H Tanabe T Yan G N Lowe C A Glackin andA Kudo ldquoA novel mechanism for the regulation of osteoblastdifferentiation transcription of periostin a member of thefasciclin I family is regulated by the bHLH transcription factorTwistrdquo Journal of Cellular Biochemistry vol 86 no 4 pp 792ndash804 2002
[21] T G Kashima T Nishiyama K Shimazu et al ldquoPeriostina novel marker of intramembranous ossification is expressedin fibrous dysplasia and in c-Fos-overexpressing bone lesionsrdquoHuman Pathology vol 40 no 2 pp 226ndash237 2009
[22] CGalliM PiergianniM Piemontese et al ldquoPeriostin improvescell adhesion to implantable biomaterials and osteoblasticdifferentiation on implant titanium surfaces in a topography-dependent fashionrdquo Journal of Biomedical Materials Research -Part A vol 102 no 11 pp 3855ndash3861 2014
18 Stem Cells International
[23] D L Coutu H W Jian A Monette G-E Rivard M DBlostein and J Galipeau ldquoPeriostin amember of a novel familyof vitamin K-dependent proteins is expressed by mesenchymalstromal cellsrdquo The Journal of Biological Chemistry vol 283 no26 pp 17991ndash18001 2008
[24] T Minamizato K Sakamoto T Liu et al ldquoCCN3NOV inhibitsBMP-2-induced osteoblast differentiation by interacting withBMP andNotch signaling pathwaysrdquo Biochemical and Biophysi-cal Research Communications vol 354 no 2 pp 567ndash573 2007
[25] W Si Q Kang H H Luu et al ldquoCCN1Cyr61 is regulated bythe canonicalWnt signal and plays an important role inWnt3A-induced osteoblast differentiation of mesenchymal stem cellsrdquoMolecular and Cellular Biology vol 26 no 8 pp 2955ndash29642006
[26] Q Luo Q Kang W Si et al ldquoConnective tissue growth factor(CTGF) is regulated by Wnt and bone morphogenetic proteinssignaling in osteoblast differentiation of mesenchymal stemcellsrdquo The Journal of Biological Chemistry vol 279 no 53 pp55958ndash55968 2004
[27] Y-S Yoon N Lee and H Scadova ldquoMyocardial regenerationwith bone-marrow-derived stem cellsrdquo Biology of the Cell vol97 no 4 pp 253ndash263 2005
[28] Y S Yoon ldquoCardiac regeneration with novel bone marrow-derived multipotent stem cellsrdquo Discovery Medicine vol 5 pp204ndash208 2005
[29] Y-S Yoon A Wecker L Heyd et al ldquoClonally expanded novelmultipotent stem cells from human bone marrow regeneratemyocardium after myocardial infarctionrdquo Journal of ClinicalInvestigation vol 115 no 2 pp 326ndash338 2005
[30] GM BolandG Perkins D J Hall andR S Tuan ldquoWnt 3a pro-motes proliferation and suppresses osteogenic differentiationof adult human mesenchymal stem cellsrdquo Journal of CellularBiochemistry vol 93 no 6 pp 1210ndash1230 2004
[31] D A Cook S W Fellgett M E Pownall P J OrsquoShea and PG Genever ldquoWnt-dependent osteogenic commitment of bonemarrow stromal cells using a novel GSK3120573 inhibitorrdquo Stem CellResearch vol 12 no 2 pp 415ndash427 2014
[32] J De Boer R Siddappa C Gaspar A VanApeldoorn R Foddeand C Van Blitterswijk ldquoWnt signaling inhibits osteogenicdifferentiation of humanmesenchymal stem cellsrdquoBone vol 34no 5 pp 818ndash826 2004
[33] MAl-Toub A AlmusaM Almajed et al ldquoPleiotropic effects ofcancer cellsrsquo secreted factors on human stromal (mesenchymal)stem cellsrdquo Stem Cell Research and Therapy vol 4 article 1142013
[34] L Chen D Qanie A Jafari et al ldquoDelta-like 1fetal antigen-1(Dlk1FA1) is a novel regulator of chondrogenic cell differenti-ation via inhibition of the Akt kinase-dependent pathwayrdquoTheJournal of Biological Chemistry vol 286 no 37 pp 32140ndash321492011
[35] T Matsunobu K Torigoe M Ishikawa et al ldquoCritical roles ofthe TGF-120573 type I receptor ALK5 in perichondrial formationand function cartilage integrity and osteoblast differentiationduring growth plate developmentrdquo Developmental Biology vol332 no 2 pp 325ndash338 2009
[36] B C Nair D E Johnson R J Majeska J A Rodkey C DBennett and G A Rodan ldquoRat alkaline phosphatase II Struc-tural similarities between the osteosarcoma bone kidney andplacenta isoenzymesrdquo Archives of Biochemistry and Biophysicsvol 254 no 1 pp 28ndash34 1987
[37] P J Butterworth ldquoAlkaline phosphatase Biochemistry of mam-malian alkaline phosphatasesrdquo Cell Biochemistry and Functionvol 1 no 2 pp 66ndash70 1983
[38] A T Ali C B Penny J E Paiker et al ldquoAlkaline phosphataseis involved in the control of adipogenesis in the murinepreadipocyte cell line 3T3-L1rdquo Clinica Chimica Acta vol 354no 1-2 pp 101ndash109 2005
[39] G M Spaggiari and L Moretta ldquoCellular and molecularinteractions of mesenchymal stem cells in innate immunityrdquoImmunology and Cell Biology vol 91 no 1 pp 27ndash31 2012
Figure 2 In vitro osteoblastic and adipocytic differentiation of two clonal populations of human bone marrow stromal stem cells CL1 andCL2 cells Both cell lines were induced for osteoblast differentiation using standard protocol described in the Methods (a) ALP staining atday 14 in control noninduced (Cont) and osteoblast induced cells (D14) Right panel shows ALP activity (119899 = 3 independent experiments119901 lt 0001) (b) Mineralized matrix formation visualized by Alizarin Red S staining Right panel shows Alizarin Red quantification at day 14and 21 after osteoblast differentiation (119899 = 3 independent experiments 119901 lt 0001) (c) Quantitative RT-PCR of osteoblastic and adipocytegene markers in CL1 and CL2 during osteoblast (upper panel) and adipocyte (lower panel) differentiation ALP = alkaline phosphataseOPN = osteopontin LPL = lipoprotein lipase and ADIPOQ = adiponectin Data are presented as fold change in expression of each targetgene normalized to GAPDH (119899 = 3 independent experiments 119901 lt 005
119901 lt 0001) (d) CL1 and CL2 lines were induced for adipocytedifferentiation using standard protocol described in the Methods Adipocyte formation was visualized at day 7 (Adip d7) and day 14 (Adipd14) by Oil Red-O staining Lower panel presents quantification of Nile red staining (119899 = 3 independent experiments 119901 lt 0001) (e) CL1and CL2 lines were induced for chondrocyte differentiation using 21-day pellet culture method as described in the Methods The pellets werestained with Alcian blue collagen 10 (Col X) and collagen 2 (Col II) (original magnification 5x)
8 Stem Cells International
Table 1 Microarray data analysis showing genes related to bone development and osteoblast differentiation upregulated in CL1 versus CL2cells
Probe ID Genbank accession Gene name Gene symbol FCA 23 P167367 NM 153426 Paired-like homeodomain 2 PITX2 30884
of the biological processes revealed several immune-relatedpathways ldquoMAPKKK cascaderdquo ldquoimmunity and defenserdquoldquosignal transductionrdquo ldquoextracellular matrix protein-mediatedsignalingrdquo and ldquointerferon-mediated immunityrdquo among oth-ers that were upregulated (Supplementary Table 2B) Also40 genes related to immune system related factors wereidentified as significantly enriched in CL2 compared to CL1cells (Supplementary Table 2C)We chose the following genesfor validation of the microarray results NOV IGFBP5 ALPTAGLN and CDH11 as they were highly expressed in CL1RT-PCR analysis confirmed themicroarray results (Figure 3)
Furthermore we compared the molecular phenotype ofCL1 and CL2 cells with that of phBMSCWe found that morethan 80 of the genes expressed in CL1 and 90 in CL2 cellswere commonwith primary phBMSC (Supplementary Figure2) suggesting that CL1 and CL2 molecular phenotype existwithin the heterogeneous population of phBMSC cultures
36 ALP Knockdown Impairs Differentiation of CL1 CellsSince ALP has been suggested as a marker for hBMSCprogenitor cell lineage commitment [13] and was highlyupregulated in CL1 cells we tested its biological role in
CL1 cells ALP siRNA transfection decreased ALP proteinlevel ALP activity and mRNA gene expression compared tocontrol cells transfected with control siRNA (119901 lt 001) andthis inhibition was detectable up to day 7 days after osteoblastdifferentiation induction (Figures 4(A) and 4(B)) At day 14 ofdifferentiation mineralization ability of CL1 was significantlyimpaired (Figure 4(C)) In addition we found that thenumber of mature adipocyte formations was significantlyreduced to more than 75 (119901 lt 001) (Figure 4(D))
To identify relevant adipocyte differentiation associatedgenes that were targeted by ALP deficiency we compared thedownregulated genes of ALP deficient CL1 with the upreg-ulated genes identified during adipocytic differentiation ofCL1 We identified 62 genes that were common (Figure 5(a)Table 4) and among these genes were genes related tometabolism (primarily lipid and carbohydrate) and transportincluding CYB5B CHST1 TAP1 ATP8A1 LRP8 PLCD1and FABP5 (Table 4) We further performed quantitativereal-time PCR of ALP deficient CL1 cells during adipocytedifferentiation The following adipocyte-associated geneswere downregulated PPAR1205742 LPL and aP2 (Figure 5(b))
12 Stem Cells International
Table 3 Whole genome microarray analysis of two clonal popu-lations of human bone marrow stromal stem cells CL1 and CL2cells (a) Top 35 highly upregulated genes in CL1 cells versus CL2cells (b) Ingenuity Pathway Analysis (IPA) showing the differentphysiological system development and function genes found in eachcategory and the corresponding 119901 value (c) Upper part IPA analysisshowing the categories for skeletal and muscular system developmentand function upregulated in CL1 cells and lower part showing genesupregulated in CL2 cells NOV = nephroblastoma overexpressedIGFBP5 = insulin-like growth factor binding protein 5 ALP TAGLN= transgelin and CDH11 = OB-cadherin (osteoblast)
Name 119901 value moleculesPhysiological system development and function
Organismal development 705119864 minus 10ndash143119864 minus 03 181Embryonic development 100119864 minus 09ndash143119864 minus 03 154Organ development 100119864 minus 09ndash143119864 minus 03 145
(b) Continued
Name 119901 value moleculesSkeletal and muscularsystem development andfunction
100119864 minus 09ndash102119864 minus 03 123
Tissue development 100119864 minus 09ndash143119864 minus 03 236
(c)
Functions annotation 119901 value moleculesSkeletal and muscular system development and
function upregulated in CL1 cellsSize of bone 143119864 minus 06 24Differentiation ofosteoblasts 384119864 minus 06 25
Mineralization ofbone 493119864 minus 06 19
Bone mineral density 365119864 minus 05 19Skeletal and muscular system development and
function upregulated in CL2 cellsDevelopment ofmuscle 100119864 minus 09 44
Proliferation ofmuscle cells 229119864 minus 06 35
Remodeling of bone 363119864 minus 06 21Resorption of bone 394119864 minus 06 19
confirming impairment of adipocytic differentiation of ALPdeficient CL1 cells
4 Discussion
We extensively studied two cell populations within culturedhBMSC that were identified based on differences in mor-phology Cellular and molecular studies revealed differencesin growth differentiation capacity and molecular signatureOur data support the notion of the presence of cellular andfunctional heterogeneity among cultured hBMSC
Cellular heterogeneity of cultured hBMSC is recognizedin an increasing number of reports Several extrinsic andintrinsic factors may contribute to the observed hBMSCheterogeneity Extrinsic factors include donor-to-donor vari-ations in the number and quantity of initiating cells whichresult in differences in cell growth rate and differentia-tion capacity [14 15] Intrinsic factors have been examinedemploying single cell clonal analysis and revealed variationsin differentiation potential among individual colonies withinhBMSC cultures ranging from the presence of cells withtrilineage (osteoblast adipocytes and chondrocyte) potencyto cells with null potency [16] Also variations in the abilityof clonal cells to form heterotopic bone when implanted invivo have been reported [5] Our study corroborates thesefindings and provides more detailed cellular and molecularphenotyping of two examples of cell populations that existwithin the heterogeneous hBMSC cultures [17]
Determining the molecular signature of CL1 and CL2using whole genomemicroarray analysis showed enrichmentof lineage-commitment associated genes inCL1 For exampleinsulin-like growth factor 5 (IGFBP5) and interleukin 6 (IL6)
Stem Cells International 13
NOV0040
0030
0020
0010
mdash
IGFBP50015
0010
0005
mdash
ALP0100
0080
0060
0040
0020
mdash
TAGLN0800
0600
0400
0200
CDH11
0060
0040
0020
mdashmdash
LRP30015
0010
0005
mdash
CL1 CL2 CL1 CL2 CL1 CL2
CL1 CL2CL1 CL2CL1 CL2
Figure 3 Validation of whole genome microarray analysis of two clonal populations of human bone marrow stromal stem cells CL1 andCL2 cells Quantitative real-time PCR for highly expressed genes in CL1 cells NOV = nephroblastoma overexpressed IGFBP5 = insulin-likegrowth factor binding protein 5 ALP TAGLN = transgelin and CDH11 = OB-cadherin (osteoblast) Data are presented as fold change inexpression of each target gene normalized to GAPDH (119899 = three independent experiments 119901 lt 0001) (see also Table 3)
were 147- and 33-fold upregulated in CL1 cells respectivelyBoth factors are expressed in osteoprogenitor cells andimportant for osteoblast maturation [18] We also observedthat periostin (POSTN) gene was highly upregulated inCL1 cells (156-fold) POSTN is a 90 kDa secreted proteinoriginally identified in murine osteoblast-like cells and isupregulated by PTH [19] Several studies employing murineand human cells have revealed important role of POSTNin osteoblast differentiation and during development inintramembranous ossification [20ndash23] Another factor iden-tified in CL1 cells is nephroblastoma overexpression (NOV)which is a member of the Cyr 61 connective tissue growthfactor (CNN) family The CCN family of proteins promotesosteoblast differentiation through interaction with integrinsWNT BMP and NOTCH signaling pathways [24ndash26] Inaddition a large number of signaling molecules known tobe regulators of hBMSC lineage specific differentiation forexample insulin-like growth factors [27ndash29] WNT [30ndash32]and MAPK [33ndash35] were enriched in CL1 cells In contrastCL2 expressed high levels of immune-related genes whichmay explain the poor differentiation response to osteoblastor adipocyte lineage In a recent study the authors usedtelomerized hBMSC and showed clearly a clonal populationthat had very low in vitro and in vivo differentiation abilityhowever they had enhanced immune-related features includ-ing high IL7 expression These nullipotent cells expressedCD317 which was associated with remarkably high basal levelexpression of factors with a proinflammatory and antiviralfunction [17] We observed that this molecular phenotypewas associated with distinct ultrastructural characteristicsof the cells In particular CL2 had abundant phagocyticvacuole microvilli and lysosomes features reminiscent of
ultrastructure of immune-regulatory cells Our data thussupport the increasingly recognized feature that hBMSCexhibit immune modulatory functions and a part of theinnate immune response [17]
We observed that ALP protein expression and enzymaticactivity were significantly different between CL1 and CL2cell lines and were thus a potential marker that distinguishesdifferent cell populations with progenitor functions (CL1)from cells with nonprogenitor functions (CL2) ALP isexpressed in a wide variety of tissues including kidneysbone and liver [36 37] but tissue-nonspecific ALP (ALPL)is considered a commitment marker for osteoblastic lineage[13 38] However in a recent study the authors examined thedifferentiation potential of a number of hMSC clones in vitroand in vivo and reported that the hMSC clones with highlevels of ALP expression were committed to trilineage differ-entiation [13] Our data corroborate and extend these findingsby reporting the effects of siRNA-mediated inhibition ofALP that resulted in an impaired hBMSC differentiation notonly to osteoblasts but also to adipocytes Also our resultscorroborate earlier studies that demonstrated in human bonebiopsies the presence of ALP expression in bone marrowadipocytic cells [13 17] All these data suggest that ALP isa ldquostemnessrdquo marker of hBMSC and not just an indicator ofosteoblastic lineage commitment
While CL1 and CL2 were isolated from telomerizedhMSC cell line they are relevant to normal human physi-ology We observed that the molecular phenotypes of CL1and CL2 were contained within the molecular signatureof primary hBMSC suggesting that CL1 and CL2 repre-sent cell populations within the heterogeneous cultures ofhBMSCWe have also previously reported that the molecular
14 Stem Cells International
ALZ
stai
ning
Control siRNA ALP siRNA
ALP
Beta
-act
in
Con
trol
ALP
siRN
A d3
ALP
siRN
A d7
Nile red quantificationlowast
NI ADIP I ALP siRNA0
50
100
150
200
Con
trol s
iRN
AA
LP si
RNA
Day 3 Day 5 OS induction Day 7 OS induction
Fold
indu
ctio
n lo
g sc
ale
10000
1000
100
10
1
Day 0 OS day 7
Control siRNAALP siRNA
(A) (B)
(C)
(D)
Figure 4 Effect of alkaline phosphatase (ALP) gene silencing by small interfering RNA (siRNA) on a clonal population of human bonemarrow stromal stem cell CL1 (A) ALP staining at day 3 day 5 or day 7 days of osteoblast differentiation (OB induction) (B) Quantitativereal-time PCR for ALP gene following ALP siRNA transfection at day 0 OB and day 7 of OB Data are presented as fold change in expressionof each target gene normalized to GAPDH (119899 = three independent experiments 119901 lt 001) Western blotting analysis of day 3 and day 7after siRNA ALP transfection of CL1 cells ALPL specific antibody and B-actin was used (C) Mineralized matrix formation as visualizedby Alizarin Red S staining in siRNA transfected CL1 cells after 14 days of OB induction (D) Nile red quantification of mature lipid filledadipocyte in control noninduced (Cont) adipocyte induced (Adip I) and ALP siRNA transfected cells that are adipocyte induced (ALPsiRNA) Adipocyte induction was carried out for 7 days lowast119901 lt 005
phenotype and cellular responses of hMSC-TERT are similarto those of primary hMSC [10]While we have identified ALPas a marker that can be used for a prospective identificationof differentiation committed population of hBMSC we iden-tified additional distinctive molecular markers of the cellsFor example IGF-1 IGF-2 and IGF binding protein 5 wereenriched in CL1 compared to CL2 IGFs and their bindingproteins are very well-studied factors that play a role inhBMSC proliferation and osteoblast differentiation [18] On
the other hand annexin A3 as well as several immune-relatedgenes was highly enriched in CL2 compared to CL1 Futurestudies are needed to determine the functional significanceof these molecules in relation to the functional identity ofvarious cell populations within the hBMSC cultures andtheir usefulness as biomarkers to dissect the heterogeneouspopulation of cultured hBMSC
Our finding of the presence of functional diversity withinhBMSC cultures that contain progenitor and nonprogenitor
Stem Cells International 15
Table 4 Microarray data analysis showing genes found to be upregulated during adipogenic differentiation and downregulated after ALPKO
62 common elements inldquoAD uprdquo and ldquoALP downrdquo Gene name FC (ALP siRNA versus
control siRNA)APOBEC3G Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3G minus15938025IFI44L Interferon-induced protein 44-like minus11520283PAQR5 Progestin and adipoQ receptor family member V minus6868241PNMA2 Paraneoplastic antigen MA2 minus59695344DUSP23 Dual specificity phosphatase 23 minus54786854CLDN23 Claudin 23 minus51885047ANKDD1A Ankyrin repeat and death domain containing 1A minus51646647IL8 Interleukin 8 minus4887188LRRC23 Leucine rich repeat containing 23 minus47611775IL6 Interleukin 6 (interferon beta 2) minus4693139LIFR Leukemia inhibitory factor receptor alpha minus46540866PTGFR Prostaglandin F receptor (FP) minus4457529FAM134B Family with sequence similarity 134 member B minus4403495CYFIP2 Cytoplasmic FMR1 interacting protein 2 minus4260462METTL7A Methyltransferase-like 7A minus40480843APOBEC3F Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3F minus39502614CA5B Carbonic anhydrase VB mitochondrial minus393889ITGA10 Integrin alpha 10 minus39143775FMO3 Flavin containing monooxygenase 3 minus3852087IMPA2 Inositol monophosphatase 2 (human) minus38374884CDO1 Cysteine dioxygenase type I minus38181455CCDC68 Coiled-coil domain containing 68 minus37292893
HEXDC Hexosaminidase (glycosyl hydrolase family 20 catalytic domain) containing minus2094836SARM1 Sterile alpha and TIR motif containing 1 minus20797038BBS1 Bardet-Biedl syndrome 1 minus20146718SERPING1 Serpin peptidase inhibitor clade G (C1 inhibitor) member 1 minus20102212FAM162A Family with sequence similarity 162 member A minus2005807TCTN1 Tectonic family member 1 minus20033443
AD up ALP down
1006 62 1145
(a)
140
120
100
080
060
040
020
mdash
ALPL PPARg aP2 LPL
lowastlowastlowastlowastlowastlowast
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
lowast
(b)
Figure 5 Adipocyte related genes downregulated in ALP knockdown CL1 cells (a) Venn diagram of whole gene expression analysis of ADupregulated genes compared with ALP KD downregulated genes (b) Quantitative real-time PCR of four selected common genes from Venndiagram including ALPL PPARg2 aP2 and LPL Data are presented as fold change in expression of each target gene normalized to GAPDH(119899 = three independent experiments lowast119901 lt 005 lowastlowast119901 lt 001 and lowastlowastlowast119901 lt 0001)
Stem Cells International 17
cell populations has a clinical relevance It demonstrates thatthe progenitor function and the immune modulatory rolesof hBMSC [39] are mediated by specific and distinguishablepopulations of hBMSC Thus future clinical studies employ-ing hBMSC should attempt to administer the relevant sub-population of hBMSC dependent on the experimental aim asa novel approach to improving the clinical efficiency insteadof the current use of heterogeneous hBMSC populations
Competing Interests
The authors declare that they have no competing interests
Authorsrsquo Contributions
Mona Elsafadi and Amer Mahmood contributed equally tothis work
Acknowledgments
These studies are supported by a grant from the College ofMedicine Research Centre (CMRC) Deanship of ScientificResearch King Saud University Riyadh Saudi Arabia Theauthors would like to thank Mr Ali Al-Roalle for all histechnical help They thank Natalie Twine for advice on DNAmicroarray analysis
References
[1] W Zaher L Harkness A Jafari and M Kassem ldquoAn update ofhuman mesenchymal stem cell biology and their clinical usesrdquoArchives of Toxicology vol 88 no 5 pp 1069ndash1082 2014
[2] S Post B M Abdallah J F Bentzon andM Kassem ldquoDemon-stration of the presence of independent pre-osteoblastic andpre-adipocytic cell populations in bone marrow-derived mes-enchymal stem cellsrdquo Bone vol 43 no 1 pp 32ndash39 2008
[3] K H Larsen C M Frederiksen J S Burns B M Abdallahand M Kassem ldquoIdentifying a molecular phenotype for bonemarrow stromal cells with in vivo bone-forming capacityrdquoJournal of Bone and Mineral Research vol 25 no 4 pp 796ndash808 2010
[4] A Rosenzweig ldquoCardiac cell therapymdashmixed results frommixed cellsrdquoThe New England Journal of Medicine vol 355 no12 pp 1274ndash1277 2006
[5] S A Kuznetsov P H Krebsbach K Satomura et al ldquoSingle-colony derived strains of human marrow stromal fibroblastsform bone after transplantation in vivordquo Journal of Bone andMineral Research vol 12 no 9 pp 1335ndash1347 1997
[6] M Kassem and P Bianco ldquoSkeletal stem cells in space and timerdquoCell vol 160 no 1-2 pp 17ndash19 2015
[7] A I Caplan and D Correa ldquoPDGF in bone formation andregeneration new insights into a novel mechanism involvingMSCsrdquo Journal of Orthopaedic Research vol 29 no 12 pp 1795ndash1803 2011
[8] K Stewart S Walsh J Screen et al ldquoFurther characterizationof cells expressing STRO-1 in cultures of adult human bonemarrow stromal cellsrdquo Journal of Bone and Mineral Researchvol 14 no 8 pp 1345ndash1356 1999
[9] J L Simonsen C Rosada N Serakinci et al ldquoTelomeraseexpression extends the proliferative life-span and maintains
the osteogenic potential of human bone marrow stromal cellsrdquoNature Biotechnology vol 20 no 6 pp 592ndash596 2002
[10] M Al-Nbaheen R Vishnubalaji D Ali et al ldquoHuman stromal(mesenchymal) stem cells from bone marrow adipose tissueand skin exhibit differences in molecular phenotype and differ-entiation potentialrdquo Stem Cell Reviews and Reports vol 9 no 1pp 32ndash43 2013
[11] C A Gregory W G Gunn A Peister and D J Prockop ldquoAnAlizarin red-based assay of mineralization by adherent cells inculture comparison with cetylpyridinium chloride extractionrdquoAnalytical Biochemistry vol 329 no 1 pp 77ndash84 2004
[12] AMahmood L Harkness H D Schroslashder B M Abdallah andM Kassem ldquoEnhanced differentiation of human embryonicstem cells to mesenchymal progenitors by inhibition of TGF-120573activinnodal signaling using SB-431542rdquo Journal of Bone andMineral Research vol 25 no 6 pp 1216ndash1233 2010
[13] H-J Prins A K Braat D Gawlitta et al ldquoIn vitro inductionof alkaline phosphatase levels predicts in vivo bone formingcapacity of human bone marrow stromal cellsrdquo Stem CellResearch vol 12 no 2 pp 428ndash440 2014
[14] K Stenderup J Justesen C Clausen and M Kassem ldquoAgingis associated with decreased maximal life span and acceleratedsenescence of bone marrow stromal cellsrdquo Bone vol 33 no 6pp 919ndash926 2003
[15] D G Phinney G Kopen W Righter S Webster N Tremainand D J Prockop ldquoDonor variation in the growth propertiesand osteogenic potential of human marrow stromal cellsrdquoJournal of Cellular Biochemistry vol 75 no 3 pp 424ndash436 1999
[16] C M Digirolamo D Stokes D Colter D G Phinney R Classand D J Prockop ldquoPropagation and senescence of humanmarrow stromal cells in culture a simple colony-forming assayidentifies samples with the greatest potential to propagate anddifferentiaterdquo British Journal of Haematology vol 107 no 2 pp275ndash281 1999
[17] S James J Fox F Afsari et al ldquoMultiparameter analy-sis of human bone marrow stromal cells identifies distinctimmunomodulatory and differentiation-competent subtypesrdquoStem Cell Reports vol 4 no 6 pp 1004ndash1015 2015
[18] B Peruzzi A Cappariello A Del Fattore N Rucci F DeBenedetti and A Teti ldquoc-Src and IL-6 inhibit osteoblastdifferentiation and integrate IGFBP5 signallingrdquo Nature Com-munications vol 3 article 630 2012
[19] D Fortunati S Reppe A-K Fjeldheim M Nielsen V TGautvik and K M Gautvik ldquoPeriostin is a collagen associatedbonematrix protein regulated by parathyroid hormonerdquoMatrixBiology vol 29 no 7 pp 594ndash601 2010
[20] A Oshima H Tanabe T Yan G N Lowe C A Glackin andA Kudo ldquoA novel mechanism for the regulation of osteoblastdifferentiation transcription of periostin a member of thefasciclin I family is regulated by the bHLH transcription factorTwistrdquo Journal of Cellular Biochemistry vol 86 no 4 pp 792ndash804 2002
[21] T G Kashima T Nishiyama K Shimazu et al ldquoPeriostina novel marker of intramembranous ossification is expressedin fibrous dysplasia and in c-Fos-overexpressing bone lesionsrdquoHuman Pathology vol 40 no 2 pp 226ndash237 2009
[22] CGalliM PiergianniM Piemontese et al ldquoPeriostin improvescell adhesion to implantable biomaterials and osteoblasticdifferentiation on implant titanium surfaces in a topography-dependent fashionrdquo Journal of Biomedical Materials Research -Part A vol 102 no 11 pp 3855ndash3861 2014
18 Stem Cells International
[23] D L Coutu H W Jian A Monette G-E Rivard M DBlostein and J Galipeau ldquoPeriostin amember of a novel familyof vitamin K-dependent proteins is expressed by mesenchymalstromal cellsrdquo The Journal of Biological Chemistry vol 283 no26 pp 17991ndash18001 2008
[24] T Minamizato K Sakamoto T Liu et al ldquoCCN3NOV inhibitsBMP-2-induced osteoblast differentiation by interacting withBMP andNotch signaling pathwaysrdquo Biochemical and Biophysi-cal Research Communications vol 354 no 2 pp 567ndash573 2007
[25] W Si Q Kang H H Luu et al ldquoCCN1Cyr61 is regulated bythe canonicalWnt signal and plays an important role inWnt3A-induced osteoblast differentiation of mesenchymal stem cellsrdquoMolecular and Cellular Biology vol 26 no 8 pp 2955ndash29642006
[26] Q Luo Q Kang W Si et al ldquoConnective tissue growth factor(CTGF) is regulated by Wnt and bone morphogenetic proteinssignaling in osteoblast differentiation of mesenchymal stemcellsrdquo The Journal of Biological Chemistry vol 279 no 53 pp55958ndash55968 2004
[27] Y-S Yoon N Lee and H Scadova ldquoMyocardial regenerationwith bone-marrow-derived stem cellsrdquo Biology of the Cell vol97 no 4 pp 253ndash263 2005
[28] Y S Yoon ldquoCardiac regeneration with novel bone marrow-derived multipotent stem cellsrdquo Discovery Medicine vol 5 pp204ndash208 2005
[29] Y-S Yoon A Wecker L Heyd et al ldquoClonally expanded novelmultipotent stem cells from human bone marrow regeneratemyocardium after myocardial infarctionrdquo Journal of ClinicalInvestigation vol 115 no 2 pp 326ndash338 2005
[30] GM BolandG Perkins D J Hall andR S Tuan ldquoWnt 3a pro-motes proliferation and suppresses osteogenic differentiationof adult human mesenchymal stem cellsrdquo Journal of CellularBiochemistry vol 93 no 6 pp 1210ndash1230 2004
[31] D A Cook S W Fellgett M E Pownall P J OrsquoShea and PG Genever ldquoWnt-dependent osteogenic commitment of bonemarrow stromal cells using a novel GSK3120573 inhibitorrdquo Stem CellResearch vol 12 no 2 pp 415ndash427 2014
[32] J De Boer R Siddappa C Gaspar A VanApeldoorn R Foddeand C Van Blitterswijk ldquoWnt signaling inhibits osteogenicdifferentiation of humanmesenchymal stem cellsrdquoBone vol 34no 5 pp 818ndash826 2004
[33] MAl-Toub A AlmusaM Almajed et al ldquoPleiotropic effects ofcancer cellsrsquo secreted factors on human stromal (mesenchymal)stem cellsrdquo Stem Cell Research and Therapy vol 4 article 1142013
[34] L Chen D Qanie A Jafari et al ldquoDelta-like 1fetal antigen-1(Dlk1FA1) is a novel regulator of chondrogenic cell differenti-ation via inhibition of the Akt kinase-dependent pathwayrdquoTheJournal of Biological Chemistry vol 286 no 37 pp 32140ndash321492011
[35] T Matsunobu K Torigoe M Ishikawa et al ldquoCritical roles ofthe TGF-120573 type I receptor ALK5 in perichondrial formationand function cartilage integrity and osteoblast differentiationduring growth plate developmentrdquo Developmental Biology vol332 no 2 pp 325ndash338 2009
[36] B C Nair D E Johnson R J Majeska J A Rodkey C DBennett and G A Rodan ldquoRat alkaline phosphatase II Struc-tural similarities between the osteosarcoma bone kidney andplacenta isoenzymesrdquo Archives of Biochemistry and Biophysicsvol 254 no 1 pp 28ndash34 1987
[37] P J Butterworth ldquoAlkaline phosphatase Biochemistry of mam-malian alkaline phosphatasesrdquo Cell Biochemistry and Functionvol 1 no 2 pp 66ndash70 1983
[38] A T Ali C B Penny J E Paiker et al ldquoAlkaline phosphataseis involved in the control of adipogenesis in the murinepreadipocyte cell line 3T3-L1rdquo Clinica Chimica Acta vol 354no 1-2 pp 101ndash109 2005
[39] G M Spaggiari and L Moretta ldquoCellular and molecularinteractions of mesenchymal stem cells in innate immunityrdquoImmunology and Cell Biology vol 91 no 1 pp 27ndash31 2012
of the biological processes revealed several immune-relatedpathways ldquoMAPKKK cascaderdquo ldquoimmunity and defenserdquoldquosignal transductionrdquo ldquoextracellular matrix protein-mediatedsignalingrdquo and ldquointerferon-mediated immunityrdquo among oth-ers that were upregulated (Supplementary Table 2B) Also40 genes related to immune system related factors wereidentified as significantly enriched in CL2 compared to CL1cells (Supplementary Table 2C)We chose the following genesfor validation of the microarray results NOV IGFBP5 ALPTAGLN and CDH11 as they were highly expressed in CL1RT-PCR analysis confirmed themicroarray results (Figure 3)
Furthermore we compared the molecular phenotype ofCL1 and CL2 cells with that of phBMSCWe found that morethan 80 of the genes expressed in CL1 and 90 in CL2 cellswere commonwith primary phBMSC (Supplementary Figure2) suggesting that CL1 and CL2 molecular phenotype existwithin the heterogeneous population of phBMSC cultures
36 ALP Knockdown Impairs Differentiation of CL1 CellsSince ALP has been suggested as a marker for hBMSCprogenitor cell lineage commitment [13] and was highlyupregulated in CL1 cells we tested its biological role in
CL1 cells ALP siRNA transfection decreased ALP proteinlevel ALP activity and mRNA gene expression compared tocontrol cells transfected with control siRNA (119901 lt 001) andthis inhibition was detectable up to day 7 days after osteoblastdifferentiation induction (Figures 4(A) and 4(B)) At day 14 ofdifferentiation mineralization ability of CL1 was significantlyimpaired (Figure 4(C)) In addition we found that thenumber of mature adipocyte formations was significantlyreduced to more than 75 (119901 lt 001) (Figure 4(D))
To identify relevant adipocyte differentiation associatedgenes that were targeted by ALP deficiency we compared thedownregulated genes of ALP deficient CL1 with the upreg-ulated genes identified during adipocytic differentiation ofCL1 We identified 62 genes that were common (Figure 5(a)Table 4) and among these genes were genes related tometabolism (primarily lipid and carbohydrate) and transportincluding CYB5B CHST1 TAP1 ATP8A1 LRP8 PLCD1and FABP5 (Table 4) We further performed quantitativereal-time PCR of ALP deficient CL1 cells during adipocytedifferentiation The following adipocyte-associated geneswere downregulated PPAR1205742 LPL and aP2 (Figure 5(b))
12 Stem Cells International
Table 3 Whole genome microarray analysis of two clonal popu-lations of human bone marrow stromal stem cells CL1 and CL2cells (a) Top 35 highly upregulated genes in CL1 cells versus CL2cells (b) Ingenuity Pathway Analysis (IPA) showing the differentphysiological system development and function genes found in eachcategory and the corresponding 119901 value (c) Upper part IPA analysisshowing the categories for skeletal and muscular system developmentand function upregulated in CL1 cells and lower part showing genesupregulated in CL2 cells NOV = nephroblastoma overexpressedIGFBP5 = insulin-like growth factor binding protein 5 ALP TAGLN= transgelin and CDH11 = OB-cadherin (osteoblast)
Name 119901 value moleculesPhysiological system development and function
Organismal development 705119864 minus 10ndash143119864 minus 03 181Embryonic development 100119864 minus 09ndash143119864 minus 03 154Organ development 100119864 minus 09ndash143119864 minus 03 145
(b) Continued
Name 119901 value moleculesSkeletal and muscularsystem development andfunction
100119864 minus 09ndash102119864 minus 03 123
Tissue development 100119864 minus 09ndash143119864 minus 03 236
(c)
Functions annotation 119901 value moleculesSkeletal and muscular system development and
function upregulated in CL1 cellsSize of bone 143119864 minus 06 24Differentiation ofosteoblasts 384119864 minus 06 25
Mineralization ofbone 493119864 minus 06 19
Bone mineral density 365119864 minus 05 19Skeletal and muscular system development and
function upregulated in CL2 cellsDevelopment ofmuscle 100119864 minus 09 44
Proliferation ofmuscle cells 229119864 minus 06 35
Remodeling of bone 363119864 minus 06 21Resorption of bone 394119864 minus 06 19
confirming impairment of adipocytic differentiation of ALPdeficient CL1 cells
4 Discussion
We extensively studied two cell populations within culturedhBMSC that were identified based on differences in mor-phology Cellular and molecular studies revealed differencesin growth differentiation capacity and molecular signatureOur data support the notion of the presence of cellular andfunctional heterogeneity among cultured hBMSC
Cellular heterogeneity of cultured hBMSC is recognizedin an increasing number of reports Several extrinsic andintrinsic factors may contribute to the observed hBMSCheterogeneity Extrinsic factors include donor-to-donor vari-ations in the number and quantity of initiating cells whichresult in differences in cell growth rate and differentia-tion capacity [14 15] Intrinsic factors have been examinedemploying single cell clonal analysis and revealed variationsin differentiation potential among individual colonies withinhBMSC cultures ranging from the presence of cells withtrilineage (osteoblast adipocytes and chondrocyte) potencyto cells with null potency [16] Also variations in the abilityof clonal cells to form heterotopic bone when implanted invivo have been reported [5] Our study corroborates thesefindings and provides more detailed cellular and molecularphenotyping of two examples of cell populations that existwithin the heterogeneous hBMSC cultures [17]
Determining the molecular signature of CL1 and CL2using whole genomemicroarray analysis showed enrichmentof lineage-commitment associated genes inCL1 For exampleinsulin-like growth factor 5 (IGFBP5) and interleukin 6 (IL6)
Stem Cells International 13
NOV0040
0030
0020
0010
mdash
IGFBP50015
0010
0005
mdash
ALP0100
0080
0060
0040
0020
mdash
TAGLN0800
0600
0400
0200
CDH11
0060
0040
0020
mdashmdash
LRP30015
0010
0005
mdash
CL1 CL2 CL1 CL2 CL1 CL2
CL1 CL2CL1 CL2CL1 CL2
Figure 3 Validation of whole genome microarray analysis of two clonal populations of human bone marrow stromal stem cells CL1 andCL2 cells Quantitative real-time PCR for highly expressed genes in CL1 cells NOV = nephroblastoma overexpressed IGFBP5 = insulin-likegrowth factor binding protein 5 ALP TAGLN = transgelin and CDH11 = OB-cadherin (osteoblast) Data are presented as fold change inexpression of each target gene normalized to GAPDH (119899 = three independent experiments 119901 lt 0001) (see also Table 3)
were 147- and 33-fold upregulated in CL1 cells respectivelyBoth factors are expressed in osteoprogenitor cells andimportant for osteoblast maturation [18] We also observedthat periostin (POSTN) gene was highly upregulated inCL1 cells (156-fold) POSTN is a 90 kDa secreted proteinoriginally identified in murine osteoblast-like cells and isupregulated by PTH [19] Several studies employing murineand human cells have revealed important role of POSTNin osteoblast differentiation and during development inintramembranous ossification [20ndash23] Another factor iden-tified in CL1 cells is nephroblastoma overexpression (NOV)which is a member of the Cyr 61 connective tissue growthfactor (CNN) family The CCN family of proteins promotesosteoblast differentiation through interaction with integrinsWNT BMP and NOTCH signaling pathways [24ndash26] Inaddition a large number of signaling molecules known tobe regulators of hBMSC lineage specific differentiation forexample insulin-like growth factors [27ndash29] WNT [30ndash32]and MAPK [33ndash35] were enriched in CL1 cells In contrastCL2 expressed high levels of immune-related genes whichmay explain the poor differentiation response to osteoblastor adipocyte lineage In a recent study the authors usedtelomerized hBMSC and showed clearly a clonal populationthat had very low in vitro and in vivo differentiation abilityhowever they had enhanced immune-related features includ-ing high IL7 expression These nullipotent cells expressedCD317 which was associated with remarkably high basal levelexpression of factors with a proinflammatory and antiviralfunction [17] We observed that this molecular phenotypewas associated with distinct ultrastructural characteristicsof the cells In particular CL2 had abundant phagocyticvacuole microvilli and lysosomes features reminiscent of
ultrastructure of immune-regulatory cells Our data thussupport the increasingly recognized feature that hBMSCexhibit immune modulatory functions and a part of theinnate immune response [17]
We observed that ALP protein expression and enzymaticactivity were significantly different between CL1 and CL2cell lines and were thus a potential marker that distinguishesdifferent cell populations with progenitor functions (CL1)from cells with nonprogenitor functions (CL2) ALP isexpressed in a wide variety of tissues including kidneysbone and liver [36 37] but tissue-nonspecific ALP (ALPL)is considered a commitment marker for osteoblastic lineage[13 38] However in a recent study the authors examined thedifferentiation potential of a number of hMSC clones in vitroand in vivo and reported that the hMSC clones with highlevels of ALP expression were committed to trilineage differ-entiation [13] Our data corroborate and extend these findingsby reporting the effects of siRNA-mediated inhibition ofALP that resulted in an impaired hBMSC differentiation notonly to osteoblasts but also to adipocytes Also our resultscorroborate earlier studies that demonstrated in human bonebiopsies the presence of ALP expression in bone marrowadipocytic cells [13 17] All these data suggest that ALP isa ldquostemnessrdquo marker of hBMSC and not just an indicator ofosteoblastic lineage commitment
While CL1 and CL2 were isolated from telomerizedhMSC cell line they are relevant to normal human physi-ology We observed that the molecular phenotypes of CL1and CL2 were contained within the molecular signatureof primary hBMSC suggesting that CL1 and CL2 repre-sent cell populations within the heterogeneous cultures ofhBMSCWe have also previously reported that the molecular
14 Stem Cells International
ALZ
stai
ning
Control siRNA ALP siRNA
ALP
Beta
-act
in
Con
trol
ALP
siRN
A d3
ALP
siRN
A d7
Nile red quantificationlowast
NI ADIP I ALP siRNA0
50
100
150
200
Con
trol s
iRN
AA
LP si
RNA
Day 3 Day 5 OS induction Day 7 OS induction
Fold
indu
ctio
n lo
g sc
ale
10000
1000
100
10
1
Day 0 OS day 7
Control siRNAALP siRNA
(A) (B)
(C)
(D)
Figure 4 Effect of alkaline phosphatase (ALP) gene silencing by small interfering RNA (siRNA) on a clonal population of human bonemarrow stromal stem cell CL1 (A) ALP staining at day 3 day 5 or day 7 days of osteoblast differentiation (OB induction) (B) Quantitativereal-time PCR for ALP gene following ALP siRNA transfection at day 0 OB and day 7 of OB Data are presented as fold change in expressionof each target gene normalized to GAPDH (119899 = three independent experiments 119901 lt 001) Western blotting analysis of day 3 and day 7after siRNA ALP transfection of CL1 cells ALPL specific antibody and B-actin was used (C) Mineralized matrix formation as visualizedby Alizarin Red S staining in siRNA transfected CL1 cells after 14 days of OB induction (D) Nile red quantification of mature lipid filledadipocyte in control noninduced (Cont) adipocyte induced (Adip I) and ALP siRNA transfected cells that are adipocyte induced (ALPsiRNA) Adipocyte induction was carried out for 7 days lowast119901 lt 005
phenotype and cellular responses of hMSC-TERT are similarto those of primary hMSC [10]While we have identified ALPas a marker that can be used for a prospective identificationof differentiation committed population of hBMSC we iden-tified additional distinctive molecular markers of the cellsFor example IGF-1 IGF-2 and IGF binding protein 5 wereenriched in CL1 compared to CL2 IGFs and their bindingproteins are very well-studied factors that play a role inhBMSC proliferation and osteoblast differentiation [18] On
the other hand annexin A3 as well as several immune-relatedgenes was highly enriched in CL2 compared to CL1 Futurestudies are needed to determine the functional significanceof these molecules in relation to the functional identity ofvarious cell populations within the hBMSC cultures andtheir usefulness as biomarkers to dissect the heterogeneouspopulation of cultured hBMSC
Our finding of the presence of functional diversity withinhBMSC cultures that contain progenitor and nonprogenitor
Stem Cells International 15
Table 4 Microarray data analysis showing genes found to be upregulated during adipogenic differentiation and downregulated after ALPKO
62 common elements inldquoAD uprdquo and ldquoALP downrdquo Gene name FC (ALP siRNA versus
control siRNA)APOBEC3G Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3G minus15938025IFI44L Interferon-induced protein 44-like minus11520283PAQR5 Progestin and adipoQ receptor family member V minus6868241PNMA2 Paraneoplastic antigen MA2 minus59695344DUSP23 Dual specificity phosphatase 23 minus54786854CLDN23 Claudin 23 minus51885047ANKDD1A Ankyrin repeat and death domain containing 1A minus51646647IL8 Interleukin 8 minus4887188LRRC23 Leucine rich repeat containing 23 minus47611775IL6 Interleukin 6 (interferon beta 2) minus4693139LIFR Leukemia inhibitory factor receptor alpha minus46540866PTGFR Prostaglandin F receptor (FP) minus4457529FAM134B Family with sequence similarity 134 member B minus4403495CYFIP2 Cytoplasmic FMR1 interacting protein 2 minus4260462METTL7A Methyltransferase-like 7A minus40480843APOBEC3F Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3F minus39502614CA5B Carbonic anhydrase VB mitochondrial minus393889ITGA10 Integrin alpha 10 minus39143775FMO3 Flavin containing monooxygenase 3 minus3852087IMPA2 Inositol monophosphatase 2 (human) minus38374884CDO1 Cysteine dioxygenase type I minus38181455CCDC68 Coiled-coil domain containing 68 minus37292893
HEXDC Hexosaminidase (glycosyl hydrolase family 20 catalytic domain) containing minus2094836SARM1 Sterile alpha and TIR motif containing 1 minus20797038BBS1 Bardet-Biedl syndrome 1 minus20146718SERPING1 Serpin peptidase inhibitor clade G (C1 inhibitor) member 1 minus20102212FAM162A Family with sequence similarity 162 member A minus2005807TCTN1 Tectonic family member 1 minus20033443
AD up ALP down
1006 62 1145
(a)
140
120
100
080
060
040
020
mdash
ALPL PPARg aP2 LPL
lowastlowastlowastlowastlowastlowast
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
lowast
(b)
Figure 5 Adipocyte related genes downregulated in ALP knockdown CL1 cells (a) Venn diagram of whole gene expression analysis of ADupregulated genes compared with ALP KD downregulated genes (b) Quantitative real-time PCR of four selected common genes from Venndiagram including ALPL PPARg2 aP2 and LPL Data are presented as fold change in expression of each target gene normalized to GAPDH(119899 = three independent experiments lowast119901 lt 005 lowastlowast119901 lt 001 and lowastlowastlowast119901 lt 0001)
Stem Cells International 17
cell populations has a clinical relevance It demonstrates thatthe progenitor function and the immune modulatory rolesof hBMSC [39] are mediated by specific and distinguishablepopulations of hBMSC Thus future clinical studies employ-ing hBMSC should attempt to administer the relevant sub-population of hBMSC dependent on the experimental aim asa novel approach to improving the clinical efficiency insteadof the current use of heterogeneous hBMSC populations
Competing Interests
The authors declare that they have no competing interests
Authorsrsquo Contributions
Mona Elsafadi and Amer Mahmood contributed equally tothis work
Acknowledgments
These studies are supported by a grant from the College ofMedicine Research Centre (CMRC) Deanship of ScientificResearch King Saud University Riyadh Saudi Arabia Theauthors would like to thank Mr Ali Al-Roalle for all histechnical help They thank Natalie Twine for advice on DNAmicroarray analysis
References
[1] W Zaher L Harkness A Jafari and M Kassem ldquoAn update ofhuman mesenchymal stem cell biology and their clinical usesrdquoArchives of Toxicology vol 88 no 5 pp 1069ndash1082 2014
[2] S Post B M Abdallah J F Bentzon andM Kassem ldquoDemon-stration of the presence of independent pre-osteoblastic andpre-adipocytic cell populations in bone marrow-derived mes-enchymal stem cellsrdquo Bone vol 43 no 1 pp 32ndash39 2008
[3] K H Larsen C M Frederiksen J S Burns B M Abdallahand M Kassem ldquoIdentifying a molecular phenotype for bonemarrow stromal cells with in vivo bone-forming capacityrdquoJournal of Bone and Mineral Research vol 25 no 4 pp 796ndash808 2010
[4] A Rosenzweig ldquoCardiac cell therapymdashmixed results frommixed cellsrdquoThe New England Journal of Medicine vol 355 no12 pp 1274ndash1277 2006
[5] S A Kuznetsov P H Krebsbach K Satomura et al ldquoSingle-colony derived strains of human marrow stromal fibroblastsform bone after transplantation in vivordquo Journal of Bone andMineral Research vol 12 no 9 pp 1335ndash1347 1997
[6] M Kassem and P Bianco ldquoSkeletal stem cells in space and timerdquoCell vol 160 no 1-2 pp 17ndash19 2015
[7] A I Caplan and D Correa ldquoPDGF in bone formation andregeneration new insights into a novel mechanism involvingMSCsrdquo Journal of Orthopaedic Research vol 29 no 12 pp 1795ndash1803 2011
[8] K Stewart S Walsh J Screen et al ldquoFurther characterizationof cells expressing STRO-1 in cultures of adult human bonemarrow stromal cellsrdquo Journal of Bone and Mineral Researchvol 14 no 8 pp 1345ndash1356 1999
[9] J L Simonsen C Rosada N Serakinci et al ldquoTelomeraseexpression extends the proliferative life-span and maintains
the osteogenic potential of human bone marrow stromal cellsrdquoNature Biotechnology vol 20 no 6 pp 592ndash596 2002
[10] M Al-Nbaheen R Vishnubalaji D Ali et al ldquoHuman stromal(mesenchymal) stem cells from bone marrow adipose tissueand skin exhibit differences in molecular phenotype and differ-entiation potentialrdquo Stem Cell Reviews and Reports vol 9 no 1pp 32ndash43 2013
[11] C A Gregory W G Gunn A Peister and D J Prockop ldquoAnAlizarin red-based assay of mineralization by adherent cells inculture comparison with cetylpyridinium chloride extractionrdquoAnalytical Biochemistry vol 329 no 1 pp 77ndash84 2004
[12] AMahmood L Harkness H D Schroslashder B M Abdallah andM Kassem ldquoEnhanced differentiation of human embryonicstem cells to mesenchymal progenitors by inhibition of TGF-120573activinnodal signaling using SB-431542rdquo Journal of Bone andMineral Research vol 25 no 6 pp 1216ndash1233 2010
[13] H-J Prins A K Braat D Gawlitta et al ldquoIn vitro inductionof alkaline phosphatase levels predicts in vivo bone formingcapacity of human bone marrow stromal cellsrdquo Stem CellResearch vol 12 no 2 pp 428ndash440 2014
[14] K Stenderup J Justesen C Clausen and M Kassem ldquoAgingis associated with decreased maximal life span and acceleratedsenescence of bone marrow stromal cellsrdquo Bone vol 33 no 6pp 919ndash926 2003
[15] D G Phinney G Kopen W Righter S Webster N Tremainand D J Prockop ldquoDonor variation in the growth propertiesand osteogenic potential of human marrow stromal cellsrdquoJournal of Cellular Biochemistry vol 75 no 3 pp 424ndash436 1999
[16] C M Digirolamo D Stokes D Colter D G Phinney R Classand D J Prockop ldquoPropagation and senescence of humanmarrow stromal cells in culture a simple colony-forming assayidentifies samples with the greatest potential to propagate anddifferentiaterdquo British Journal of Haematology vol 107 no 2 pp275ndash281 1999
[17] S James J Fox F Afsari et al ldquoMultiparameter analy-sis of human bone marrow stromal cells identifies distinctimmunomodulatory and differentiation-competent subtypesrdquoStem Cell Reports vol 4 no 6 pp 1004ndash1015 2015
[18] B Peruzzi A Cappariello A Del Fattore N Rucci F DeBenedetti and A Teti ldquoc-Src and IL-6 inhibit osteoblastdifferentiation and integrate IGFBP5 signallingrdquo Nature Com-munications vol 3 article 630 2012
[19] D Fortunati S Reppe A-K Fjeldheim M Nielsen V TGautvik and K M Gautvik ldquoPeriostin is a collagen associatedbonematrix protein regulated by parathyroid hormonerdquoMatrixBiology vol 29 no 7 pp 594ndash601 2010
[20] A Oshima H Tanabe T Yan G N Lowe C A Glackin andA Kudo ldquoA novel mechanism for the regulation of osteoblastdifferentiation transcription of periostin a member of thefasciclin I family is regulated by the bHLH transcription factorTwistrdquo Journal of Cellular Biochemistry vol 86 no 4 pp 792ndash804 2002
[21] T G Kashima T Nishiyama K Shimazu et al ldquoPeriostina novel marker of intramembranous ossification is expressedin fibrous dysplasia and in c-Fos-overexpressing bone lesionsrdquoHuman Pathology vol 40 no 2 pp 226ndash237 2009
[22] CGalliM PiergianniM Piemontese et al ldquoPeriostin improvescell adhesion to implantable biomaterials and osteoblasticdifferentiation on implant titanium surfaces in a topography-dependent fashionrdquo Journal of Biomedical Materials Research -Part A vol 102 no 11 pp 3855ndash3861 2014
18 Stem Cells International
[23] D L Coutu H W Jian A Monette G-E Rivard M DBlostein and J Galipeau ldquoPeriostin amember of a novel familyof vitamin K-dependent proteins is expressed by mesenchymalstromal cellsrdquo The Journal of Biological Chemistry vol 283 no26 pp 17991ndash18001 2008
[24] T Minamizato K Sakamoto T Liu et al ldquoCCN3NOV inhibitsBMP-2-induced osteoblast differentiation by interacting withBMP andNotch signaling pathwaysrdquo Biochemical and Biophysi-cal Research Communications vol 354 no 2 pp 567ndash573 2007
[25] W Si Q Kang H H Luu et al ldquoCCN1Cyr61 is regulated bythe canonicalWnt signal and plays an important role inWnt3A-induced osteoblast differentiation of mesenchymal stem cellsrdquoMolecular and Cellular Biology vol 26 no 8 pp 2955ndash29642006
[26] Q Luo Q Kang W Si et al ldquoConnective tissue growth factor(CTGF) is regulated by Wnt and bone morphogenetic proteinssignaling in osteoblast differentiation of mesenchymal stemcellsrdquo The Journal of Biological Chemistry vol 279 no 53 pp55958ndash55968 2004
[27] Y-S Yoon N Lee and H Scadova ldquoMyocardial regenerationwith bone-marrow-derived stem cellsrdquo Biology of the Cell vol97 no 4 pp 253ndash263 2005
[28] Y S Yoon ldquoCardiac regeneration with novel bone marrow-derived multipotent stem cellsrdquo Discovery Medicine vol 5 pp204ndash208 2005
[29] Y-S Yoon A Wecker L Heyd et al ldquoClonally expanded novelmultipotent stem cells from human bone marrow regeneratemyocardium after myocardial infarctionrdquo Journal of ClinicalInvestigation vol 115 no 2 pp 326ndash338 2005
[30] GM BolandG Perkins D J Hall andR S Tuan ldquoWnt 3a pro-motes proliferation and suppresses osteogenic differentiationof adult human mesenchymal stem cellsrdquo Journal of CellularBiochemistry vol 93 no 6 pp 1210ndash1230 2004
[31] D A Cook S W Fellgett M E Pownall P J OrsquoShea and PG Genever ldquoWnt-dependent osteogenic commitment of bonemarrow stromal cells using a novel GSK3120573 inhibitorrdquo Stem CellResearch vol 12 no 2 pp 415ndash427 2014
[32] J De Boer R Siddappa C Gaspar A VanApeldoorn R Foddeand C Van Blitterswijk ldquoWnt signaling inhibits osteogenicdifferentiation of humanmesenchymal stem cellsrdquoBone vol 34no 5 pp 818ndash826 2004
[33] MAl-Toub A AlmusaM Almajed et al ldquoPleiotropic effects ofcancer cellsrsquo secreted factors on human stromal (mesenchymal)stem cellsrdquo Stem Cell Research and Therapy vol 4 article 1142013
[34] L Chen D Qanie A Jafari et al ldquoDelta-like 1fetal antigen-1(Dlk1FA1) is a novel regulator of chondrogenic cell differenti-ation via inhibition of the Akt kinase-dependent pathwayrdquoTheJournal of Biological Chemistry vol 286 no 37 pp 32140ndash321492011
[35] T Matsunobu K Torigoe M Ishikawa et al ldquoCritical roles ofthe TGF-120573 type I receptor ALK5 in perichondrial formationand function cartilage integrity and osteoblast differentiationduring growth plate developmentrdquo Developmental Biology vol332 no 2 pp 325ndash338 2009
[36] B C Nair D E Johnson R J Majeska J A Rodkey C DBennett and G A Rodan ldquoRat alkaline phosphatase II Struc-tural similarities between the osteosarcoma bone kidney andplacenta isoenzymesrdquo Archives of Biochemistry and Biophysicsvol 254 no 1 pp 28ndash34 1987
[37] P J Butterworth ldquoAlkaline phosphatase Biochemistry of mam-malian alkaline phosphatasesrdquo Cell Biochemistry and Functionvol 1 no 2 pp 66ndash70 1983
[38] A T Ali C B Penny J E Paiker et al ldquoAlkaline phosphataseis involved in the control of adipogenesis in the murinepreadipocyte cell line 3T3-L1rdquo Clinica Chimica Acta vol 354no 1-2 pp 101ndash109 2005
[39] G M Spaggiari and L Moretta ldquoCellular and molecularinteractions of mesenchymal stem cells in innate immunityrdquoImmunology and Cell Biology vol 91 no 1 pp 27ndash31 2012
of the biological processes revealed several immune-relatedpathways ldquoMAPKKK cascaderdquo ldquoimmunity and defenserdquoldquosignal transductionrdquo ldquoextracellular matrix protein-mediatedsignalingrdquo and ldquointerferon-mediated immunityrdquo among oth-ers that were upregulated (Supplementary Table 2B) Also40 genes related to immune system related factors wereidentified as significantly enriched in CL2 compared to CL1cells (Supplementary Table 2C)We chose the following genesfor validation of the microarray results NOV IGFBP5 ALPTAGLN and CDH11 as they were highly expressed in CL1RT-PCR analysis confirmed themicroarray results (Figure 3)
Furthermore we compared the molecular phenotype ofCL1 and CL2 cells with that of phBMSCWe found that morethan 80 of the genes expressed in CL1 and 90 in CL2 cellswere commonwith primary phBMSC (Supplementary Figure2) suggesting that CL1 and CL2 molecular phenotype existwithin the heterogeneous population of phBMSC cultures
36 ALP Knockdown Impairs Differentiation of CL1 CellsSince ALP has been suggested as a marker for hBMSCprogenitor cell lineage commitment [13] and was highlyupregulated in CL1 cells we tested its biological role in
CL1 cells ALP siRNA transfection decreased ALP proteinlevel ALP activity and mRNA gene expression compared tocontrol cells transfected with control siRNA (119901 lt 001) andthis inhibition was detectable up to day 7 days after osteoblastdifferentiation induction (Figures 4(A) and 4(B)) At day 14 ofdifferentiation mineralization ability of CL1 was significantlyimpaired (Figure 4(C)) In addition we found that thenumber of mature adipocyte formations was significantlyreduced to more than 75 (119901 lt 001) (Figure 4(D))
To identify relevant adipocyte differentiation associatedgenes that were targeted by ALP deficiency we compared thedownregulated genes of ALP deficient CL1 with the upreg-ulated genes identified during adipocytic differentiation ofCL1 We identified 62 genes that were common (Figure 5(a)Table 4) and among these genes were genes related tometabolism (primarily lipid and carbohydrate) and transportincluding CYB5B CHST1 TAP1 ATP8A1 LRP8 PLCD1and FABP5 (Table 4) We further performed quantitativereal-time PCR of ALP deficient CL1 cells during adipocytedifferentiation The following adipocyte-associated geneswere downregulated PPAR1205742 LPL and aP2 (Figure 5(b))
12 Stem Cells International
Table 3 Whole genome microarray analysis of two clonal popu-lations of human bone marrow stromal stem cells CL1 and CL2cells (a) Top 35 highly upregulated genes in CL1 cells versus CL2cells (b) Ingenuity Pathway Analysis (IPA) showing the differentphysiological system development and function genes found in eachcategory and the corresponding 119901 value (c) Upper part IPA analysisshowing the categories for skeletal and muscular system developmentand function upregulated in CL1 cells and lower part showing genesupregulated in CL2 cells NOV = nephroblastoma overexpressedIGFBP5 = insulin-like growth factor binding protein 5 ALP TAGLN= transgelin and CDH11 = OB-cadherin (osteoblast)
Name 119901 value moleculesPhysiological system development and function
Organismal development 705119864 minus 10ndash143119864 minus 03 181Embryonic development 100119864 minus 09ndash143119864 minus 03 154Organ development 100119864 minus 09ndash143119864 minus 03 145
(b) Continued
Name 119901 value moleculesSkeletal and muscularsystem development andfunction
100119864 minus 09ndash102119864 minus 03 123
Tissue development 100119864 minus 09ndash143119864 minus 03 236
(c)
Functions annotation 119901 value moleculesSkeletal and muscular system development and
function upregulated in CL1 cellsSize of bone 143119864 minus 06 24Differentiation ofosteoblasts 384119864 minus 06 25
Mineralization ofbone 493119864 minus 06 19
Bone mineral density 365119864 minus 05 19Skeletal and muscular system development and
function upregulated in CL2 cellsDevelopment ofmuscle 100119864 minus 09 44
Proliferation ofmuscle cells 229119864 minus 06 35
Remodeling of bone 363119864 minus 06 21Resorption of bone 394119864 minus 06 19
confirming impairment of adipocytic differentiation of ALPdeficient CL1 cells
4 Discussion
We extensively studied two cell populations within culturedhBMSC that were identified based on differences in mor-phology Cellular and molecular studies revealed differencesin growth differentiation capacity and molecular signatureOur data support the notion of the presence of cellular andfunctional heterogeneity among cultured hBMSC
Cellular heterogeneity of cultured hBMSC is recognizedin an increasing number of reports Several extrinsic andintrinsic factors may contribute to the observed hBMSCheterogeneity Extrinsic factors include donor-to-donor vari-ations in the number and quantity of initiating cells whichresult in differences in cell growth rate and differentia-tion capacity [14 15] Intrinsic factors have been examinedemploying single cell clonal analysis and revealed variationsin differentiation potential among individual colonies withinhBMSC cultures ranging from the presence of cells withtrilineage (osteoblast adipocytes and chondrocyte) potencyto cells with null potency [16] Also variations in the abilityof clonal cells to form heterotopic bone when implanted invivo have been reported [5] Our study corroborates thesefindings and provides more detailed cellular and molecularphenotyping of two examples of cell populations that existwithin the heterogeneous hBMSC cultures [17]
Determining the molecular signature of CL1 and CL2using whole genomemicroarray analysis showed enrichmentof lineage-commitment associated genes inCL1 For exampleinsulin-like growth factor 5 (IGFBP5) and interleukin 6 (IL6)
Stem Cells International 13
NOV0040
0030
0020
0010
mdash
IGFBP50015
0010
0005
mdash
ALP0100
0080
0060
0040
0020
mdash
TAGLN0800
0600
0400
0200
CDH11
0060
0040
0020
mdashmdash
LRP30015
0010
0005
mdash
CL1 CL2 CL1 CL2 CL1 CL2
CL1 CL2CL1 CL2CL1 CL2
Figure 3 Validation of whole genome microarray analysis of two clonal populations of human bone marrow stromal stem cells CL1 andCL2 cells Quantitative real-time PCR for highly expressed genes in CL1 cells NOV = nephroblastoma overexpressed IGFBP5 = insulin-likegrowth factor binding protein 5 ALP TAGLN = transgelin and CDH11 = OB-cadherin (osteoblast) Data are presented as fold change inexpression of each target gene normalized to GAPDH (119899 = three independent experiments 119901 lt 0001) (see also Table 3)
were 147- and 33-fold upregulated in CL1 cells respectivelyBoth factors are expressed in osteoprogenitor cells andimportant for osteoblast maturation [18] We also observedthat periostin (POSTN) gene was highly upregulated inCL1 cells (156-fold) POSTN is a 90 kDa secreted proteinoriginally identified in murine osteoblast-like cells and isupregulated by PTH [19] Several studies employing murineand human cells have revealed important role of POSTNin osteoblast differentiation and during development inintramembranous ossification [20ndash23] Another factor iden-tified in CL1 cells is nephroblastoma overexpression (NOV)which is a member of the Cyr 61 connective tissue growthfactor (CNN) family The CCN family of proteins promotesosteoblast differentiation through interaction with integrinsWNT BMP and NOTCH signaling pathways [24ndash26] Inaddition a large number of signaling molecules known tobe regulators of hBMSC lineage specific differentiation forexample insulin-like growth factors [27ndash29] WNT [30ndash32]and MAPK [33ndash35] were enriched in CL1 cells In contrastCL2 expressed high levels of immune-related genes whichmay explain the poor differentiation response to osteoblastor adipocyte lineage In a recent study the authors usedtelomerized hBMSC and showed clearly a clonal populationthat had very low in vitro and in vivo differentiation abilityhowever they had enhanced immune-related features includ-ing high IL7 expression These nullipotent cells expressedCD317 which was associated with remarkably high basal levelexpression of factors with a proinflammatory and antiviralfunction [17] We observed that this molecular phenotypewas associated with distinct ultrastructural characteristicsof the cells In particular CL2 had abundant phagocyticvacuole microvilli and lysosomes features reminiscent of
ultrastructure of immune-regulatory cells Our data thussupport the increasingly recognized feature that hBMSCexhibit immune modulatory functions and a part of theinnate immune response [17]
We observed that ALP protein expression and enzymaticactivity were significantly different between CL1 and CL2cell lines and were thus a potential marker that distinguishesdifferent cell populations with progenitor functions (CL1)from cells with nonprogenitor functions (CL2) ALP isexpressed in a wide variety of tissues including kidneysbone and liver [36 37] but tissue-nonspecific ALP (ALPL)is considered a commitment marker for osteoblastic lineage[13 38] However in a recent study the authors examined thedifferentiation potential of a number of hMSC clones in vitroand in vivo and reported that the hMSC clones with highlevels of ALP expression were committed to trilineage differ-entiation [13] Our data corroborate and extend these findingsby reporting the effects of siRNA-mediated inhibition ofALP that resulted in an impaired hBMSC differentiation notonly to osteoblasts but also to adipocytes Also our resultscorroborate earlier studies that demonstrated in human bonebiopsies the presence of ALP expression in bone marrowadipocytic cells [13 17] All these data suggest that ALP isa ldquostemnessrdquo marker of hBMSC and not just an indicator ofosteoblastic lineage commitment
While CL1 and CL2 were isolated from telomerizedhMSC cell line they are relevant to normal human physi-ology We observed that the molecular phenotypes of CL1and CL2 were contained within the molecular signatureof primary hBMSC suggesting that CL1 and CL2 repre-sent cell populations within the heterogeneous cultures ofhBMSCWe have also previously reported that the molecular
14 Stem Cells International
ALZ
stai
ning
Control siRNA ALP siRNA
ALP
Beta
-act
in
Con
trol
ALP
siRN
A d3
ALP
siRN
A d7
Nile red quantificationlowast
NI ADIP I ALP siRNA0
50
100
150
200
Con
trol s
iRN
AA
LP si
RNA
Day 3 Day 5 OS induction Day 7 OS induction
Fold
indu
ctio
n lo
g sc
ale
10000
1000
100
10
1
Day 0 OS day 7
Control siRNAALP siRNA
(A) (B)
(C)
(D)
Figure 4 Effect of alkaline phosphatase (ALP) gene silencing by small interfering RNA (siRNA) on a clonal population of human bonemarrow stromal stem cell CL1 (A) ALP staining at day 3 day 5 or day 7 days of osteoblast differentiation (OB induction) (B) Quantitativereal-time PCR for ALP gene following ALP siRNA transfection at day 0 OB and day 7 of OB Data are presented as fold change in expressionof each target gene normalized to GAPDH (119899 = three independent experiments 119901 lt 001) Western blotting analysis of day 3 and day 7after siRNA ALP transfection of CL1 cells ALPL specific antibody and B-actin was used (C) Mineralized matrix formation as visualizedby Alizarin Red S staining in siRNA transfected CL1 cells after 14 days of OB induction (D) Nile red quantification of mature lipid filledadipocyte in control noninduced (Cont) adipocyte induced (Adip I) and ALP siRNA transfected cells that are adipocyte induced (ALPsiRNA) Adipocyte induction was carried out for 7 days lowast119901 lt 005
phenotype and cellular responses of hMSC-TERT are similarto those of primary hMSC [10]While we have identified ALPas a marker that can be used for a prospective identificationof differentiation committed population of hBMSC we iden-tified additional distinctive molecular markers of the cellsFor example IGF-1 IGF-2 and IGF binding protein 5 wereenriched in CL1 compared to CL2 IGFs and their bindingproteins are very well-studied factors that play a role inhBMSC proliferation and osteoblast differentiation [18] On
the other hand annexin A3 as well as several immune-relatedgenes was highly enriched in CL2 compared to CL1 Futurestudies are needed to determine the functional significanceof these molecules in relation to the functional identity ofvarious cell populations within the hBMSC cultures andtheir usefulness as biomarkers to dissect the heterogeneouspopulation of cultured hBMSC
Our finding of the presence of functional diversity withinhBMSC cultures that contain progenitor and nonprogenitor
Stem Cells International 15
Table 4 Microarray data analysis showing genes found to be upregulated during adipogenic differentiation and downregulated after ALPKO
62 common elements inldquoAD uprdquo and ldquoALP downrdquo Gene name FC (ALP siRNA versus
control siRNA)APOBEC3G Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3G minus15938025IFI44L Interferon-induced protein 44-like minus11520283PAQR5 Progestin and adipoQ receptor family member V minus6868241PNMA2 Paraneoplastic antigen MA2 minus59695344DUSP23 Dual specificity phosphatase 23 minus54786854CLDN23 Claudin 23 minus51885047ANKDD1A Ankyrin repeat and death domain containing 1A minus51646647IL8 Interleukin 8 minus4887188LRRC23 Leucine rich repeat containing 23 minus47611775IL6 Interleukin 6 (interferon beta 2) minus4693139LIFR Leukemia inhibitory factor receptor alpha minus46540866PTGFR Prostaglandin F receptor (FP) minus4457529FAM134B Family with sequence similarity 134 member B minus4403495CYFIP2 Cytoplasmic FMR1 interacting protein 2 minus4260462METTL7A Methyltransferase-like 7A minus40480843APOBEC3F Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3F minus39502614CA5B Carbonic anhydrase VB mitochondrial minus393889ITGA10 Integrin alpha 10 minus39143775FMO3 Flavin containing monooxygenase 3 minus3852087IMPA2 Inositol monophosphatase 2 (human) minus38374884CDO1 Cysteine dioxygenase type I minus38181455CCDC68 Coiled-coil domain containing 68 minus37292893
HEXDC Hexosaminidase (glycosyl hydrolase family 20 catalytic domain) containing minus2094836SARM1 Sterile alpha and TIR motif containing 1 minus20797038BBS1 Bardet-Biedl syndrome 1 minus20146718SERPING1 Serpin peptidase inhibitor clade G (C1 inhibitor) member 1 minus20102212FAM162A Family with sequence similarity 162 member A minus2005807TCTN1 Tectonic family member 1 minus20033443
AD up ALP down
1006 62 1145
(a)
140
120
100
080
060
040
020
mdash
ALPL PPARg aP2 LPL
lowastlowastlowastlowastlowastlowast
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
lowast
(b)
Figure 5 Adipocyte related genes downregulated in ALP knockdown CL1 cells (a) Venn diagram of whole gene expression analysis of ADupregulated genes compared with ALP KD downregulated genes (b) Quantitative real-time PCR of four selected common genes from Venndiagram including ALPL PPARg2 aP2 and LPL Data are presented as fold change in expression of each target gene normalized to GAPDH(119899 = three independent experiments lowast119901 lt 005 lowastlowast119901 lt 001 and lowastlowastlowast119901 lt 0001)
Stem Cells International 17
cell populations has a clinical relevance It demonstrates thatthe progenitor function and the immune modulatory rolesof hBMSC [39] are mediated by specific and distinguishablepopulations of hBMSC Thus future clinical studies employ-ing hBMSC should attempt to administer the relevant sub-population of hBMSC dependent on the experimental aim asa novel approach to improving the clinical efficiency insteadof the current use of heterogeneous hBMSC populations
Competing Interests
The authors declare that they have no competing interests
Authorsrsquo Contributions
Mona Elsafadi and Amer Mahmood contributed equally tothis work
Acknowledgments
These studies are supported by a grant from the College ofMedicine Research Centre (CMRC) Deanship of ScientificResearch King Saud University Riyadh Saudi Arabia Theauthors would like to thank Mr Ali Al-Roalle for all histechnical help They thank Natalie Twine for advice on DNAmicroarray analysis
References
[1] W Zaher L Harkness A Jafari and M Kassem ldquoAn update ofhuman mesenchymal stem cell biology and their clinical usesrdquoArchives of Toxicology vol 88 no 5 pp 1069ndash1082 2014
[2] S Post B M Abdallah J F Bentzon andM Kassem ldquoDemon-stration of the presence of independent pre-osteoblastic andpre-adipocytic cell populations in bone marrow-derived mes-enchymal stem cellsrdquo Bone vol 43 no 1 pp 32ndash39 2008
[3] K H Larsen C M Frederiksen J S Burns B M Abdallahand M Kassem ldquoIdentifying a molecular phenotype for bonemarrow stromal cells with in vivo bone-forming capacityrdquoJournal of Bone and Mineral Research vol 25 no 4 pp 796ndash808 2010
[4] A Rosenzweig ldquoCardiac cell therapymdashmixed results frommixed cellsrdquoThe New England Journal of Medicine vol 355 no12 pp 1274ndash1277 2006
[5] S A Kuznetsov P H Krebsbach K Satomura et al ldquoSingle-colony derived strains of human marrow stromal fibroblastsform bone after transplantation in vivordquo Journal of Bone andMineral Research vol 12 no 9 pp 1335ndash1347 1997
[6] M Kassem and P Bianco ldquoSkeletal stem cells in space and timerdquoCell vol 160 no 1-2 pp 17ndash19 2015
[7] A I Caplan and D Correa ldquoPDGF in bone formation andregeneration new insights into a novel mechanism involvingMSCsrdquo Journal of Orthopaedic Research vol 29 no 12 pp 1795ndash1803 2011
[8] K Stewart S Walsh J Screen et al ldquoFurther characterizationof cells expressing STRO-1 in cultures of adult human bonemarrow stromal cellsrdquo Journal of Bone and Mineral Researchvol 14 no 8 pp 1345ndash1356 1999
[9] J L Simonsen C Rosada N Serakinci et al ldquoTelomeraseexpression extends the proliferative life-span and maintains
the osteogenic potential of human bone marrow stromal cellsrdquoNature Biotechnology vol 20 no 6 pp 592ndash596 2002
[10] M Al-Nbaheen R Vishnubalaji D Ali et al ldquoHuman stromal(mesenchymal) stem cells from bone marrow adipose tissueand skin exhibit differences in molecular phenotype and differ-entiation potentialrdquo Stem Cell Reviews and Reports vol 9 no 1pp 32ndash43 2013
[11] C A Gregory W G Gunn A Peister and D J Prockop ldquoAnAlizarin red-based assay of mineralization by adherent cells inculture comparison with cetylpyridinium chloride extractionrdquoAnalytical Biochemistry vol 329 no 1 pp 77ndash84 2004
[12] AMahmood L Harkness H D Schroslashder B M Abdallah andM Kassem ldquoEnhanced differentiation of human embryonicstem cells to mesenchymal progenitors by inhibition of TGF-120573activinnodal signaling using SB-431542rdquo Journal of Bone andMineral Research vol 25 no 6 pp 1216ndash1233 2010
[13] H-J Prins A K Braat D Gawlitta et al ldquoIn vitro inductionof alkaline phosphatase levels predicts in vivo bone formingcapacity of human bone marrow stromal cellsrdquo Stem CellResearch vol 12 no 2 pp 428ndash440 2014
[14] K Stenderup J Justesen C Clausen and M Kassem ldquoAgingis associated with decreased maximal life span and acceleratedsenescence of bone marrow stromal cellsrdquo Bone vol 33 no 6pp 919ndash926 2003
[15] D G Phinney G Kopen W Righter S Webster N Tremainand D J Prockop ldquoDonor variation in the growth propertiesand osteogenic potential of human marrow stromal cellsrdquoJournal of Cellular Biochemistry vol 75 no 3 pp 424ndash436 1999
[16] C M Digirolamo D Stokes D Colter D G Phinney R Classand D J Prockop ldquoPropagation and senescence of humanmarrow stromal cells in culture a simple colony-forming assayidentifies samples with the greatest potential to propagate anddifferentiaterdquo British Journal of Haematology vol 107 no 2 pp275ndash281 1999
[17] S James J Fox F Afsari et al ldquoMultiparameter analy-sis of human bone marrow stromal cells identifies distinctimmunomodulatory and differentiation-competent subtypesrdquoStem Cell Reports vol 4 no 6 pp 1004ndash1015 2015
[18] B Peruzzi A Cappariello A Del Fattore N Rucci F DeBenedetti and A Teti ldquoc-Src and IL-6 inhibit osteoblastdifferentiation and integrate IGFBP5 signallingrdquo Nature Com-munications vol 3 article 630 2012
[19] D Fortunati S Reppe A-K Fjeldheim M Nielsen V TGautvik and K M Gautvik ldquoPeriostin is a collagen associatedbonematrix protein regulated by parathyroid hormonerdquoMatrixBiology vol 29 no 7 pp 594ndash601 2010
[20] A Oshima H Tanabe T Yan G N Lowe C A Glackin andA Kudo ldquoA novel mechanism for the regulation of osteoblastdifferentiation transcription of periostin a member of thefasciclin I family is regulated by the bHLH transcription factorTwistrdquo Journal of Cellular Biochemistry vol 86 no 4 pp 792ndash804 2002
[21] T G Kashima T Nishiyama K Shimazu et al ldquoPeriostina novel marker of intramembranous ossification is expressedin fibrous dysplasia and in c-Fos-overexpressing bone lesionsrdquoHuman Pathology vol 40 no 2 pp 226ndash237 2009
[22] CGalliM PiergianniM Piemontese et al ldquoPeriostin improvescell adhesion to implantable biomaterials and osteoblasticdifferentiation on implant titanium surfaces in a topography-dependent fashionrdquo Journal of Biomedical Materials Research -Part A vol 102 no 11 pp 3855ndash3861 2014
18 Stem Cells International
[23] D L Coutu H W Jian A Monette G-E Rivard M DBlostein and J Galipeau ldquoPeriostin amember of a novel familyof vitamin K-dependent proteins is expressed by mesenchymalstromal cellsrdquo The Journal of Biological Chemistry vol 283 no26 pp 17991ndash18001 2008
[24] T Minamizato K Sakamoto T Liu et al ldquoCCN3NOV inhibitsBMP-2-induced osteoblast differentiation by interacting withBMP andNotch signaling pathwaysrdquo Biochemical and Biophysi-cal Research Communications vol 354 no 2 pp 567ndash573 2007
[25] W Si Q Kang H H Luu et al ldquoCCN1Cyr61 is regulated bythe canonicalWnt signal and plays an important role inWnt3A-induced osteoblast differentiation of mesenchymal stem cellsrdquoMolecular and Cellular Biology vol 26 no 8 pp 2955ndash29642006
[26] Q Luo Q Kang W Si et al ldquoConnective tissue growth factor(CTGF) is regulated by Wnt and bone morphogenetic proteinssignaling in osteoblast differentiation of mesenchymal stemcellsrdquo The Journal of Biological Chemistry vol 279 no 53 pp55958ndash55968 2004
[27] Y-S Yoon N Lee and H Scadova ldquoMyocardial regenerationwith bone-marrow-derived stem cellsrdquo Biology of the Cell vol97 no 4 pp 253ndash263 2005
[28] Y S Yoon ldquoCardiac regeneration with novel bone marrow-derived multipotent stem cellsrdquo Discovery Medicine vol 5 pp204ndash208 2005
[29] Y-S Yoon A Wecker L Heyd et al ldquoClonally expanded novelmultipotent stem cells from human bone marrow regeneratemyocardium after myocardial infarctionrdquo Journal of ClinicalInvestigation vol 115 no 2 pp 326ndash338 2005
[30] GM BolandG Perkins D J Hall andR S Tuan ldquoWnt 3a pro-motes proliferation and suppresses osteogenic differentiationof adult human mesenchymal stem cellsrdquo Journal of CellularBiochemistry vol 93 no 6 pp 1210ndash1230 2004
[31] D A Cook S W Fellgett M E Pownall P J OrsquoShea and PG Genever ldquoWnt-dependent osteogenic commitment of bonemarrow stromal cells using a novel GSK3120573 inhibitorrdquo Stem CellResearch vol 12 no 2 pp 415ndash427 2014
[32] J De Boer R Siddappa C Gaspar A VanApeldoorn R Foddeand C Van Blitterswijk ldquoWnt signaling inhibits osteogenicdifferentiation of humanmesenchymal stem cellsrdquoBone vol 34no 5 pp 818ndash826 2004
[33] MAl-Toub A AlmusaM Almajed et al ldquoPleiotropic effects ofcancer cellsrsquo secreted factors on human stromal (mesenchymal)stem cellsrdquo Stem Cell Research and Therapy vol 4 article 1142013
[34] L Chen D Qanie A Jafari et al ldquoDelta-like 1fetal antigen-1(Dlk1FA1) is a novel regulator of chondrogenic cell differenti-ation via inhibition of the Akt kinase-dependent pathwayrdquoTheJournal of Biological Chemistry vol 286 no 37 pp 32140ndash321492011
[35] T Matsunobu K Torigoe M Ishikawa et al ldquoCritical roles ofthe TGF-120573 type I receptor ALK5 in perichondrial formationand function cartilage integrity and osteoblast differentiationduring growth plate developmentrdquo Developmental Biology vol332 no 2 pp 325ndash338 2009
[36] B C Nair D E Johnson R J Majeska J A Rodkey C DBennett and G A Rodan ldquoRat alkaline phosphatase II Struc-tural similarities between the osteosarcoma bone kidney andplacenta isoenzymesrdquo Archives of Biochemistry and Biophysicsvol 254 no 1 pp 28ndash34 1987
[37] P J Butterworth ldquoAlkaline phosphatase Biochemistry of mam-malian alkaline phosphatasesrdquo Cell Biochemistry and Functionvol 1 no 2 pp 66ndash70 1983
[38] A T Ali C B Penny J E Paiker et al ldquoAlkaline phosphataseis involved in the control of adipogenesis in the murinepreadipocyte cell line 3T3-L1rdquo Clinica Chimica Acta vol 354no 1-2 pp 101ndash109 2005
[39] G M Spaggiari and L Moretta ldquoCellular and molecularinteractions of mesenchymal stem cells in innate immunityrdquoImmunology and Cell Biology vol 91 no 1 pp 27ndash31 2012
of the biological processes revealed several immune-relatedpathways ldquoMAPKKK cascaderdquo ldquoimmunity and defenserdquoldquosignal transductionrdquo ldquoextracellular matrix protein-mediatedsignalingrdquo and ldquointerferon-mediated immunityrdquo among oth-ers that were upregulated (Supplementary Table 2B) Also40 genes related to immune system related factors wereidentified as significantly enriched in CL2 compared to CL1cells (Supplementary Table 2C)We chose the following genesfor validation of the microarray results NOV IGFBP5 ALPTAGLN and CDH11 as they were highly expressed in CL1RT-PCR analysis confirmed themicroarray results (Figure 3)
Furthermore we compared the molecular phenotype ofCL1 and CL2 cells with that of phBMSCWe found that morethan 80 of the genes expressed in CL1 and 90 in CL2 cellswere commonwith primary phBMSC (Supplementary Figure2) suggesting that CL1 and CL2 molecular phenotype existwithin the heterogeneous population of phBMSC cultures
36 ALP Knockdown Impairs Differentiation of CL1 CellsSince ALP has been suggested as a marker for hBMSCprogenitor cell lineage commitment [13] and was highlyupregulated in CL1 cells we tested its biological role in
CL1 cells ALP siRNA transfection decreased ALP proteinlevel ALP activity and mRNA gene expression compared tocontrol cells transfected with control siRNA (119901 lt 001) andthis inhibition was detectable up to day 7 days after osteoblastdifferentiation induction (Figures 4(A) and 4(B)) At day 14 ofdifferentiation mineralization ability of CL1 was significantlyimpaired (Figure 4(C)) In addition we found that thenumber of mature adipocyte formations was significantlyreduced to more than 75 (119901 lt 001) (Figure 4(D))
To identify relevant adipocyte differentiation associatedgenes that were targeted by ALP deficiency we compared thedownregulated genes of ALP deficient CL1 with the upreg-ulated genes identified during adipocytic differentiation ofCL1 We identified 62 genes that were common (Figure 5(a)Table 4) and among these genes were genes related tometabolism (primarily lipid and carbohydrate) and transportincluding CYB5B CHST1 TAP1 ATP8A1 LRP8 PLCD1and FABP5 (Table 4) We further performed quantitativereal-time PCR of ALP deficient CL1 cells during adipocytedifferentiation The following adipocyte-associated geneswere downregulated PPAR1205742 LPL and aP2 (Figure 5(b))
12 Stem Cells International
Table 3 Whole genome microarray analysis of two clonal popu-lations of human bone marrow stromal stem cells CL1 and CL2cells (a) Top 35 highly upregulated genes in CL1 cells versus CL2cells (b) Ingenuity Pathway Analysis (IPA) showing the differentphysiological system development and function genes found in eachcategory and the corresponding 119901 value (c) Upper part IPA analysisshowing the categories for skeletal and muscular system developmentand function upregulated in CL1 cells and lower part showing genesupregulated in CL2 cells NOV = nephroblastoma overexpressedIGFBP5 = insulin-like growth factor binding protein 5 ALP TAGLN= transgelin and CDH11 = OB-cadherin (osteoblast)
Name 119901 value moleculesPhysiological system development and function
Organismal development 705119864 minus 10ndash143119864 minus 03 181Embryonic development 100119864 minus 09ndash143119864 minus 03 154Organ development 100119864 minus 09ndash143119864 minus 03 145
(b) Continued
Name 119901 value moleculesSkeletal and muscularsystem development andfunction
100119864 minus 09ndash102119864 minus 03 123
Tissue development 100119864 minus 09ndash143119864 minus 03 236
(c)
Functions annotation 119901 value moleculesSkeletal and muscular system development and
function upregulated in CL1 cellsSize of bone 143119864 minus 06 24Differentiation ofosteoblasts 384119864 minus 06 25
Mineralization ofbone 493119864 minus 06 19
Bone mineral density 365119864 minus 05 19Skeletal and muscular system development and
function upregulated in CL2 cellsDevelopment ofmuscle 100119864 minus 09 44
Proliferation ofmuscle cells 229119864 minus 06 35
Remodeling of bone 363119864 minus 06 21Resorption of bone 394119864 minus 06 19
confirming impairment of adipocytic differentiation of ALPdeficient CL1 cells
4 Discussion
We extensively studied two cell populations within culturedhBMSC that were identified based on differences in mor-phology Cellular and molecular studies revealed differencesin growth differentiation capacity and molecular signatureOur data support the notion of the presence of cellular andfunctional heterogeneity among cultured hBMSC
Cellular heterogeneity of cultured hBMSC is recognizedin an increasing number of reports Several extrinsic andintrinsic factors may contribute to the observed hBMSCheterogeneity Extrinsic factors include donor-to-donor vari-ations in the number and quantity of initiating cells whichresult in differences in cell growth rate and differentia-tion capacity [14 15] Intrinsic factors have been examinedemploying single cell clonal analysis and revealed variationsin differentiation potential among individual colonies withinhBMSC cultures ranging from the presence of cells withtrilineage (osteoblast adipocytes and chondrocyte) potencyto cells with null potency [16] Also variations in the abilityof clonal cells to form heterotopic bone when implanted invivo have been reported [5] Our study corroborates thesefindings and provides more detailed cellular and molecularphenotyping of two examples of cell populations that existwithin the heterogeneous hBMSC cultures [17]
Determining the molecular signature of CL1 and CL2using whole genomemicroarray analysis showed enrichmentof lineage-commitment associated genes inCL1 For exampleinsulin-like growth factor 5 (IGFBP5) and interleukin 6 (IL6)
Stem Cells International 13
NOV0040
0030
0020
0010
mdash
IGFBP50015
0010
0005
mdash
ALP0100
0080
0060
0040
0020
mdash
TAGLN0800
0600
0400
0200
CDH11
0060
0040
0020
mdashmdash
LRP30015
0010
0005
mdash
CL1 CL2 CL1 CL2 CL1 CL2
CL1 CL2CL1 CL2CL1 CL2
Figure 3 Validation of whole genome microarray analysis of two clonal populations of human bone marrow stromal stem cells CL1 andCL2 cells Quantitative real-time PCR for highly expressed genes in CL1 cells NOV = nephroblastoma overexpressed IGFBP5 = insulin-likegrowth factor binding protein 5 ALP TAGLN = transgelin and CDH11 = OB-cadherin (osteoblast) Data are presented as fold change inexpression of each target gene normalized to GAPDH (119899 = three independent experiments 119901 lt 0001) (see also Table 3)
were 147- and 33-fold upregulated in CL1 cells respectivelyBoth factors are expressed in osteoprogenitor cells andimportant for osteoblast maturation [18] We also observedthat periostin (POSTN) gene was highly upregulated inCL1 cells (156-fold) POSTN is a 90 kDa secreted proteinoriginally identified in murine osteoblast-like cells and isupregulated by PTH [19] Several studies employing murineand human cells have revealed important role of POSTNin osteoblast differentiation and during development inintramembranous ossification [20ndash23] Another factor iden-tified in CL1 cells is nephroblastoma overexpression (NOV)which is a member of the Cyr 61 connective tissue growthfactor (CNN) family The CCN family of proteins promotesosteoblast differentiation through interaction with integrinsWNT BMP and NOTCH signaling pathways [24ndash26] Inaddition a large number of signaling molecules known tobe regulators of hBMSC lineage specific differentiation forexample insulin-like growth factors [27ndash29] WNT [30ndash32]and MAPK [33ndash35] were enriched in CL1 cells In contrastCL2 expressed high levels of immune-related genes whichmay explain the poor differentiation response to osteoblastor adipocyte lineage In a recent study the authors usedtelomerized hBMSC and showed clearly a clonal populationthat had very low in vitro and in vivo differentiation abilityhowever they had enhanced immune-related features includ-ing high IL7 expression These nullipotent cells expressedCD317 which was associated with remarkably high basal levelexpression of factors with a proinflammatory and antiviralfunction [17] We observed that this molecular phenotypewas associated with distinct ultrastructural characteristicsof the cells In particular CL2 had abundant phagocyticvacuole microvilli and lysosomes features reminiscent of
ultrastructure of immune-regulatory cells Our data thussupport the increasingly recognized feature that hBMSCexhibit immune modulatory functions and a part of theinnate immune response [17]
We observed that ALP protein expression and enzymaticactivity were significantly different between CL1 and CL2cell lines and were thus a potential marker that distinguishesdifferent cell populations with progenitor functions (CL1)from cells with nonprogenitor functions (CL2) ALP isexpressed in a wide variety of tissues including kidneysbone and liver [36 37] but tissue-nonspecific ALP (ALPL)is considered a commitment marker for osteoblastic lineage[13 38] However in a recent study the authors examined thedifferentiation potential of a number of hMSC clones in vitroand in vivo and reported that the hMSC clones with highlevels of ALP expression were committed to trilineage differ-entiation [13] Our data corroborate and extend these findingsby reporting the effects of siRNA-mediated inhibition ofALP that resulted in an impaired hBMSC differentiation notonly to osteoblasts but also to adipocytes Also our resultscorroborate earlier studies that demonstrated in human bonebiopsies the presence of ALP expression in bone marrowadipocytic cells [13 17] All these data suggest that ALP isa ldquostemnessrdquo marker of hBMSC and not just an indicator ofosteoblastic lineage commitment
While CL1 and CL2 were isolated from telomerizedhMSC cell line they are relevant to normal human physi-ology We observed that the molecular phenotypes of CL1and CL2 were contained within the molecular signatureof primary hBMSC suggesting that CL1 and CL2 repre-sent cell populations within the heterogeneous cultures ofhBMSCWe have also previously reported that the molecular
14 Stem Cells International
ALZ
stai
ning
Control siRNA ALP siRNA
ALP
Beta
-act
in
Con
trol
ALP
siRN
A d3
ALP
siRN
A d7
Nile red quantificationlowast
NI ADIP I ALP siRNA0
50
100
150
200
Con
trol s
iRN
AA
LP si
RNA
Day 3 Day 5 OS induction Day 7 OS induction
Fold
indu
ctio
n lo
g sc
ale
10000
1000
100
10
1
Day 0 OS day 7
Control siRNAALP siRNA
(A) (B)
(C)
(D)
Figure 4 Effect of alkaline phosphatase (ALP) gene silencing by small interfering RNA (siRNA) on a clonal population of human bonemarrow stromal stem cell CL1 (A) ALP staining at day 3 day 5 or day 7 days of osteoblast differentiation (OB induction) (B) Quantitativereal-time PCR for ALP gene following ALP siRNA transfection at day 0 OB and day 7 of OB Data are presented as fold change in expressionof each target gene normalized to GAPDH (119899 = three independent experiments 119901 lt 001) Western blotting analysis of day 3 and day 7after siRNA ALP transfection of CL1 cells ALPL specific antibody and B-actin was used (C) Mineralized matrix formation as visualizedby Alizarin Red S staining in siRNA transfected CL1 cells after 14 days of OB induction (D) Nile red quantification of mature lipid filledadipocyte in control noninduced (Cont) adipocyte induced (Adip I) and ALP siRNA transfected cells that are adipocyte induced (ALPsiRNA) Adipocyte induction was carried out for 7 days lowast119901 lt 005
phenotype and cellular responses of hMSC-TERT are similarto those of primary hMSC [10]While we have identified ALPas a marker that can be used for a prospective identificationof differentiation committed population of hBMSC we iden-tified additional distinctive molecular markers of the cellsFor example IGF-1 IGF-2 and IGF binding protein 5 wereenriched in CL1 compared to CL2 IGFs and their bindingproteins are very well-studied factors that play a role inhBMSC proliferation and osteoblast differentiation [18] On
the other hand annexin A3 as well as several immune-relatedgenes was highly enriched in CL2 compared to CL1 Futurestudies are needed to determine the functional significanceof these molecules in relation to the functional identity ofvarious cell populations within the hBMSC cultures andtheir usefulness as biomarkers to dissect the heterogeneouspopulation of cultured hBMSC
Our finding of the presence of functional diversity withinhBMSC cultures that contain progenitor and nonprogenitor
Stem Cells International 15
Table 4 Microarray data analysis showing genes found to be upregulated during adipogenic differentiation and downregulated after ALPKO
62 common elements inldquoAD uprdquo and ldquoALP downrdquo Gene name FC (ALP siRNA versus
control siRNA)APOBEC3G Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3G minus15938025IFI44L Interferon-induced protein 44-like minus11520283PAQR5 Progestin and adipoQ receptor family member V minus6868241PNMA2 Paraneoplastic antigen MA2 minus59695344DUSP23 Dual specificity phosphatase 23 minus54786854CLDN23 Claudin 23 minus51885047ANKDD1A Ankyrin repeat and death domain containing 1A minus51646647IL8 Interleukin 8 minus4887188LRRC23 Leucine rich repeat containing 23 minus47611775IL6 Interleukin 6 (interferon beta 2) minus4693139LIFR Leukemia inhibitory factor receptor alpha minus46540866PTGFR Prostaglandin F receptor (FP) minus4457529FAM134B Family with sequence similarity 134 member B minus4403495CYFIP2 Cytoplasmic FMR1 interacting protein 2 minus4260462METTL7A Methyltransferase-like 7A minus40480843APOBEC3F Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3F minus39502614CA5B Carbonic anhydrase VB mitochondrial minus393889ITGA10 Integrin alpha 10 minus39143775FMO3 Flavin containing monooxygenase 3 minus3852087IMPA2 Inositol monophosphatase 2 (human) minus38374884CDO1 Cysteine dioxygenase type I minus38181455CCDC68 Coiled-coil domain containing 68 minus37292893
HEXDC Hexosaminidase (glycosyl hydrolase family 20 catalytic domain) containing minus2094836SARM1 Sterile alpha and TIR motif containing 1 minus20797038BBS1 Bardet-Biedl syndrome 1 minus20146718SERPING1 Serpin peptidase inhibitor clade G (C1 inhibitor) member 1 minus20102212FAM162A Family with sequence similarity 162 member A minus2005807TCTN1 Tectonic family member 1 minus20033443
AD up ALP down
1006 62 1145
(a)
140
120
100
080
060
040
020
mdash
ALPL PPARg aP2 LPL
lowastlowastlowastlowastlowastlowast
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
lowast
(b)
Figure 5 Adipocyte related genes downregulated in ALP knockdown CL1 cells (a) Venn diagram of whole gene expression analysis of ADupregulated genes compared with ALP KD downregulated genes (b) Quantitative real-time PCR of four selected common genes from Venndiagram including ALPL PPARg2 aP2 and LPL Data are presented as fold change in expression of each target gene normalized to GAPDH(119899 = three independent experiments lowast119901 lt 005 lowastlowast119901 lt 001 and lowastlowastlowast119901 lt 0001)
Stem Cells International 17
cell populations has a clinical relevance It demonstrates thatthe progenitor function and the immune modulatory rolesof hBMSC [39] are mediated by specific and distinguishablepopulations of hBMSC Thus future clinical studies employ-ing hBMSC should attempt to administer the relevant sub-population of hBMSC dependent on the experimental aim asa novel approach to improving the clinical efficiency insteadof the current use of heterogeneous hBMSC populations
Competing Interests
The authors declare that they have no competing interests
Authorsrsquo Contributions
Mona Elsafadi and Amer Mahmood contributed equally tothis work
Acknowledgments
These studies are supported by a grant from the College ofMedicine Research Centre (CMRC) Deanship of ScientificResearch King Saud University Riyadh Saudi Arabia Theauthors would like to thank Mr Ali Al-Roalle for all histechnical help They thank Natalie Twine for advice on DNAmicroarray analysis
References
[1] W Zaher L Harkness A Jafari and M Kassem ldquoAn update ofhuman mesenchymal stem cell biology and their clinical usesrdquoArchives of Toxicology vol 88 no 5 pp 1069ndash1082 2014
[2] S Post B M Abdallah J F Bentzon andM Kassem ldquoDemon-stration of the presence of independent pre-osteoblastic andpre-adipocytic cell populations in bone marrow-derived mes-enchymal stem cellsrdquo Bone vol 43 no 1 pp 32ndash39 2008
[3] K H Larsen C M Frederiksen J S Burns B M Abdallahand M Kassem ldquoIdentifying a molecular phenotype for bonemarrow stromal cells with in vivo bone-forming capacityrdquoJournal of Bone and Mineral Research vol 25 no 4 pp 796ndash808 2010
[4] A Rosenzweig ldquoCardiac cell therapymdashmixed results frommixed cellsrdquoThe New England Journal of Medicine vol 355 no12 pp 1274ndash1277 2006
[5] S A Kuznetsov P H Krebsbach K Satomura et al ldquoSingle-colony derived strains of human marrow stromal fibroblastsform bone after transplantation in vivordquo Journal of Bone andMineral Research vol 12 no 9 pp 1335ndash1347 1997
[6] M Kassem and P Bianco ldquoSkeletal stem cells in space and timerdquoCell vol 160 no 1-2 pp 17ndash19 2015
[7] A I Caplan and D Correa ldquoPDGF in bone formation andregeneration new insights into a novel mechanism involvingMSCsrdquo Journal of Orthopaedic Research vol 29 no 12 pp 1795ndash1803 2011
[8] K Stewart S Walsh J Screen et al ldquoFurther characterizationof cells expressing STRO-1 in cultures of adult human bonemarrow stromal cellsrdquo Journal of Bone and Mineral Researchvol 14 no 8 pp 1345ndash1356 1999
[9] J L Simonsen C Rosada N Serakinci et al ldquoTelomeraseexpression extends the proliferative life-span and maintains
the osteogenic potential of human bone marrow stromal cellsrdquoNature Biotechnology vol 20 no 6 pp 592ndash596 2002
[10] M Al-Nbaheen R Vishnubalaji D Ali et al ldquoHuman stromal(mesenchymal) stem cells from bone marrow adipose tissueand skin exhibit differences in molecular phenotype and differ-entiation potentialrdquo Stem Cell Reviews and Reports vol 9 no 1pp 32ndash43 2013
[11] C A Gregory W G Gunn A Peister and D J Prockop ldquoAnAlizarin red-based assay of mineralization by adherent cells inculture comparison with cetylpyridinium chloride extractionrdquoAnalytical Biochemistry vol 329 no 1 pp 77ndash84 2004
[12] AMahmood L Harkness H D Schroslashder B M Abdallah andM Kassem ldquoEnhanced differentiation of human embryonicstem cells to mesenchymal progenitors by inhibition of TGF-120573activinnodal signaling using SB-431542rdquo Journal of Bone andMineral Research vol 25 no 6 pp 1216ndash1233 2010
[13] H-J Prins A K Braat D Gawlitta et al ldquoIn vitro inductionof alkaline phosphatase levels predicts in vivo bone formingcapacity of human bone marrow stromal cellsrdquo Stem CellResearch vol 12 no 2 pp 428ndash440 2014
[14] K Stenderup J Justesen C Clausen and M Kassem ldquoAgingis associated with decreased maximal life span and acceleratedsenescence of bone marrow stromal cellsrdquo Bone vol 33 no 6pp 919ndash926 2003
[15] D G Phinney G Kopen W Righter S Webster N Tremainand D J Prockop ldquoDonor variation in the growth propertiesand osteogenic potential of human marrow stromal cellsrdquoJournal of Cellular Biochemistry vol 75 no 3 pp 424ndash436 1999
[16] C M Digirolamo D Stokes D Colter D G Phinney R Classand D J Prockop ldquoPropagation and senescence of humanmarrow stromal cells in culture a simple colony-forming assayidentifies samples with the greatest potential to propagate anddifferentiaterdquo British Journal of Haematology vol 107 no 2 pp275ndash281 1999
[17] S James J Fox F Afsari et al ldquoMultiparameter analy-sis of human bone marrow stromal cells identifies distinctimmunomodulatory and differentiation-competent subtypesrdquoStem Cell Reports vol 4 no 6 pp 1004ndash1015 2015
[18] B Peruzzi A Cappariello A Del Fattore N Rucci F DeBenedetti and A Teti ldquoc-Src and IL-6 inhibit osteoblastdifferentiation and integrate IGFBP5 signallingrdquo Nature Com-munications vol 3 article 630 2012
[19] D Fortunati S Reppe A-K Fjeldheim M Nielsen V TGautvik and K M Gautvik ldquoPeriostin is a collagen associatedbonematrix protein regulated by parathyroid hormonerdquoMatrixBiology vol 29 no 7 pp 594ndash601 2010
[20] A Oshima H Tanabe T Yan G N Lowe C A Glackin andA Kudo ldquoA novel mechanism for the regulation of osteoblastdifferentiation transcription of periostin a member of thefasciclin I family is regulated by the bHLH transcription factorTwistrdquo Journal of Cellular Biochemistry vol 86 no 4 pp 792ndash804 2002
[21] T G Kashima T Nishiyama K Shimazu et al ldquoPeriostina novel marker of intramembranous ossification is expressedin fibrous dysplasia and in c-Fos-overexpressing bone lesionsrdquoHuman Pathology vol 40 no 2 pp 226ndash237 2009
[22] CGalliM PiergianniM Piemontese et al ldquoPeriostin improvescell adhesion to implantable biomaterials and osteoblasticdifferentiation on implant titanium surfaces in a topography-dependent fashionrdquo Journal of Biomedical Materials Research -Part A vol 102 no 11 pp 3855ndash3861 2014
18 Stem Cells International
[23] D L Coutu H W Jian A Monette G-E Rivard M DBlostein and J Galipeau ldquoPeriostin amember of a novel familyof vitamin K-dependent proteins is expressed by mesenchymalstromal cellsrdquo The Journal of Biological Chemistry vol 283 no26 pp 17991ndash18001 2008
[24] T Minamizato K Sakamoto T Liu et al ldquoCCN3NOV inhibitsBMP-2-induced osteoblast differentiation by interacting withBMP andNotch signaling pathwaysrdquo Biochemical and Biophysi-cal Research Communications vol 354 no 2 pp 567ndash573 2007
[25] W Si Q Kang H H Luu et al ldquoCCN1Cyr61 is regulated bythe canonicalWnt signal and plays an important role inWnt3A-induced osteoblast differentiation of mesenchymal stem cellsrdquoMolecular and Cellular Biology vol 26 no 8 pp 2955ndash29642006
[26] Q Luo Q Kang W Si et al ldquoConnective tissue growth factor(CTGF) is regulated by Wnt and bone morphogenetic proteinssignaling in osteoblast differentiation of mesenchymal stemcellsrdquo The Journal of Biological Chemistry vol 279 no 53 pp55958ndash55968 2004
[27] Y-S Yoon N Lee and H Scadova ldquoMyocardial regenerationwith bone-marrow-derived stem cellsrdquo Biology of the Cell vol97 no 4 pp 253ndash263 2005
[28] Y S Yoon ldquoCardiac regeneration with novel bone marrow-derived multipotent stem cellsrdquo Discovery Medicine vol 5 pp204ndash208 2005
[29] Y-S Yoon A Wecker L Heyd et al ldquoClonally expanded novelmultipotent stem cells from human bone marrow regeneratemyocardium after myocardial infarctionrdquo Journal of ClinicalInvestigation vol 115 no 2 pp 326ndash338 2005
[30] GM BolandG Perkins D J Hall andR S Tuan ldquoWnt 3a pro-motes proliferation and suppresses osteogenic differentiationof adult human mesenchymal stem cellsrdquo Journal of CellularBiochemistry vol 93 no 6 pp 1210ndash1230 2004
[31] D A Cook S W Fellgett M E Pownall P J OrsquoShea and PG Genever ldquoWnt-dependent osteogenic commitment of bonemarrow stromal cells using a novel GSK3120573 inhibitorrdquo Stem CellResearch vol 12 no 2 pp 415ndash427 2014
[32] J De Boer R Siddappa C Gaspar A VanApeldoorn R Foddeand C Van Blitterswijk ldquoWnt signaling inhibits osteogenicdifferentiation of humanmesenchymal stem cellsrdquoBone vol 34no 5 pp 818ndash826 2004
[33] MAl-Toub A AlmusaM Almajed et al ldquoPleiotropic effects ofcancer cellsrsquo secreted factors on human stromal (mesenchymal)stem cellsrdquo Stem Cell Research and Therapy vol 4 article 1142013
[34] L Chen D Qanie A Jafari et al ldquoDelta-like 1fetal antigen-1(Dlk1FA1) is a novel regulator of chondrogenic cell differenti-ation via inhibition of the Akt kinase-dependent pathwayrdquoTheJournal of Biological Chemistry vol 286 no 37 pp 32140ndash321492011
[35] T Matsunobu K Torigoe M Ishikawa et al ldquoCritical roles ofthe TGF-120573 type I receptor ALK5 in perichondrial formationand function cartilage integrity and osteoblast differentiationduring growth plate developmentrdquo Developmental Biology vol332 no 2 pp 325ndash338 2009
[36] B C Nair D E Johnson R J Majeska J A Rodkey C DBennett and G A Rodan ldquoRat alkaline phosphatase II Struc-tural similarities between the osteosarcoma bone kidney andplacenta isoenzymesrdquo Archives of Biochemistry and Biophysicsvol 254 no 1 pp 28ndash34 1987
[37] P J Butterworth ldquoAlkaline phosphatase Biochemistry of mam-malian alkaline phosphatasesrdquo Cell Biochemistry and Functionvol 1 no 2 pp 66ndash70 1983
[38] A T Ali C B Penny J E Paiker et al ldquoAlkaline phosphataseis involved in the control of adipogenesis in the murinepreadipocyte cell line 3T3-L1rdquo Clinica Chimica Acta vol 354no 1-2 pp 101ndash109 2005
[39] G M Spaggiari and L Moretta ldquoCellular and molecularinteractions of mesenchymal stem cells in innate immunityrdquoImmunology and Cell Biology vol 91 no 1 pp 27ndash31 2012
of the biological processes revealed several immune-relatedpathways ldquoMAPKKK cascaderdquo ldquoimmunity and defenserdquoldquosignal transductionrdquo ldquoextracellular matrix protein-mediatedsignalingrdquo and ldquointerferon-mediated immunityrdquo among oth-ers that were upregulated (Supplementary Table 2B) Also40 genes related to immune system related factors wereidentified as significantly enriched in CL2 compared to CL1cells (Supplementary Table 2C)We chose the following genesfor validation of the microarray results NOV IGFBP5 ALPTAGLN and CDH11 as they were highly expressed in CL1RT-PCR analysis confirmed themicroarray results (Figure 3)
Furthermore we compared the molecular phenotype ofCL1 and CL2 cells with that of phBMSCWe found that morethan 80 of the genes expressed in CL1 and 90 in CL2 cellswere commonwith primary phBMSC (Supplementary Figure2) suggesting that CL1 and CL2 molecular phenotype existwithin the heterogeneous population of phBMSC cultures
36 ALP Knockdown Impairs Differentiation of CL1 CellsSince ALP has been suggested as a marker for hBMSCprogenitor cell lineage commitment [13] and was highlyupregulated in CL1 cells we tested its biological role in
CL1 cells ALP siRNA transfection decreased ALP proteinlevel ALP activity and mRNA gene expression compared tocontrol cells transfected with control siRNA (119901 lt 001) andthis inhibition was detectable up to day 7 days after osteoblastdifferentiation induction (Figures 4(A) and 4(B)) At day 14 ofdifferentiation mineralization ability of CL1 was significantlyimpaired (Figure 4(C)) In addition we found that thenumber of mature adipocyte formations was significantlyreduced to more than 75 (119901 lt 001) (Figure 4(D))
To identify relevant adipocyte differentiation associatedgenes that were targeted by ALP deficiency we compared thedownregulated genes of ALP deficient CL1 with the upreg-ulated genes identified during adipocytic differentiation ofCL1 We identified 62 genes that were common (Figure 5(a)Table 4) and among these genes were genes related tometabolism (primarily lipid and carbohydrate) and transportincluding CYB5B CHST1 TAP1 ATP8A1 LRP8 PLCD1and FABP5 (Table 4) We further performed quantitativereal-time PCR of ALP deficient CL1 cells during adipocytedifferentiation The following adipocyte-associated geneswere downregulated PPAR1205742 LPL and aP2 (Figure 5(b))
12 Stem Cells International
Table 3 Whole genome microarray analysis of two clonal popu-lations of human bone marrow stromal stem cells CL1 and CL2cells (a) Top 35 highly upregulated genes in CL1 cells versus CL2cells (b) Ingenuity Pathway Analysis (IPA) showing the differentphysiological system development and function genes found in eachcategory and the corresponding 119901 value (c) Upper part IPA analysisshowing the categories for skeletal and muscular system developmentand function upregulated in CL1 cells and lower part showing genesupregulated in CL2 cells NOV = nephroblastoma overexpressedIGFBP5 = insulin-like growth factor binding protein 5 ALP TAGLN= transgelin and CDH11 = OB-cadherin (osteoblast)
Name 119901 value moleculesPhysiological system development and function
Organismal development 705119864 minus 10ndash143119864 minus 03 181Embryonic development 100119864 minus 09ndash143119864 minus 03 154Organ development 100119864 minus 09ndash143119864 minus 03 145
(b) Continued
Name 119901 value moleculesSkeletal and muscularsystem development andfunction
100119864 minus 09ndash102119864 minus 03 123
Tissue development 100119864 minus 09ndash143119864 minus 03 236
(c)
Functions annotation 119901 value moleculesSkeletal and muscular system development and
function upregulated in CL1 cellsSize of bone 143119864 minus 06 24Differentiation ofosteoblasts 384119864 minus 06 25
Mineralization ofbone 493119864 minus 06 19
Bone mineral density 365119864 minus 05 19Skeletal and muscular system development and
function upregulated in CL2 cellsDevelopment ofmuscle 100119864 minus 09 44
Proliferation ofmuscle cells 229119864 minus 06 35
Remodeling of bone 363119864 minus 06 21Resorption of bone 394119864 minus 06 19
confirming impairment of adipocytic differentiation of ALPdeficient CL1 cells
4 Discussion
We extensively studied two cell populations within culturedhBMSC that were identified based on differences in mor-phology Cellular and molecular studies revealed differencesin growth differentiation capacity and molecular signatureOur data support the notion of the presence of cellular andfunctional heterogeneity among cultured hBMSC
Cellular heterogeneity of cultured hBMSC is recognizedin an increasing number of reports Several extrinsic andintrinsic factors may contribute to the observed hBMSCheterogeneity Extrinsic factors include donor-to-donor vari-ations in the number and quantity of initiating cells whichresult in differences in cell growth rate and differentia-tion capacity [14 15] Intrinsic factors have been examinedemploying single cell clonal analysis and revealed variationsin differentiation potential among individual colonies withinhBMSC cultures ranging from the presence of cells withtrilineage (osteoblast adipocytes and chondrocyte) potencyto cells with null potency [16] Also variations in the abilityof clonal cells to form heterotopic bone when implanted invivo have been reported [5] Our study corroborates thesefindings and provides more detailed cellular and molecularphenotyping of two examples of cell populations that existwithin the heterogeneous hBMSC cultures [17]
Determining the molecular signature of CL1 and CL2using whole genomemicroarray analysis showed enrichmentof lineage-commitment associated genes inCL1 For exampleinsulin-like growth factor 5 (IGFBP5) and interleukin 6 (IL6)
Stem Cells International 13
NOV0040
0030
0020
0010
mdash
IGFBP50015
0010
0005
mdash
ALP0100
0080
0060
0040
0020
mdash
TAGLN0800
0600
0400
0200
CDH11
0060
0040
0020
mdashmdash
LRP30015
0010
0005
mdash
CL1 CL2 CL1 CL2 CL1 CL2
CL1 CL2CL1 CL2CL1 CL2
Figure 3 Validation of whole genome microarray analysis of two clonal populations of human bone marrow stromal stem cells CL1 andCL2 cells Quantitative real-time PCR for highly expressed genes in CL1 cells NOV = nephroblastoma overexpressed IGFBP5 = insulin-likegrowth factor binding protein 5 ALP TAGLN = transgelin and CDH11 = OB-cadherin (osteoblast) Data are presented as fold change inexpression of each target gene normalized to GAPDH (119899 = three independent experiments 119901 lt 0001) (see also Table 3)
were 147- and 33-fold upregulated in CL1 cells respectivelyBoth factors are expressed in osteoprogenitor cells andimportant for osteoblast maturation [18] We also observedthat periostin (POSTN) gene was highly upregulated inCL1 cells (156-fold) POSTN is a 90 kDa secreted proteinoriginally identified in murine osteoblast-like cells and isupregulated by PTH [19] Several studies employing murineand human cells have revealed important role of POSTNin osteoblast differentiation and during development inintramembranous ossification [20ndash23] Another factor iden-tified in CL1 cells is nephroblastoma overexpression (NOV)which is a member of the Cyr 61 connective tissue growthfactor (CNN) family The CCN family of proteins promotesosteoblast differentiation through interaction with integrinsWNT BMP and NOTCH signaling pathways [24ndash26] Inaddition a large number of signaling molecules known tobe regulators of hBMSC lineage specific differentiation forexample insulin-like growth factors [27ndash29] WNT [30ndash32]and MAPK [33ndash35] were enriched in CL1 cells In contrastCL2 expressed high levels of immune-related genes whichmay explain the poor differentiation response to osteoblastor adipocyte lineage In a recent study the authors usedtelomerized hBMSC and showed clearly a clonal populationthat had very low in vitro and in vivo differentiation abilityhowever they had enhanced immune-related features includ-ing high IL7 expression These nullipotent cells expressedCD317 which was associated with remarkably high basal levelexpression of factors with a proinflammatory and antiviralfunction [17] We observed that this molecular phenotypewas associated with distinct ultrastructural characteristicsof the cells In particular CL2 had abundant phagocyticvacuole microvilli and lysosomes features reminiscent of
ultrastructure of immune-regulatory cells Our data thussupport the increasingly recognized feature that hBMSCexhibit immune modulatory functions and a part of theinnate immune response [17]
We observed that ALP protein expression and enzymaticactivity were significantly different between CL1 and CL2cell lines and were thus a potential marker that distinguishesdifferent cell populations with progenitor functions (CL1)from cells with nonprogenitor functions (CL2) ALP isexpressed in a wide variety of tissues including kidneysbone and liver [36 37] but tissue-nonspecific ALP (ALPL)is considered a commitment marker for osteoblastic lineage[13 38] However in a recent study the authors examined thedifferentiation potential of a number of hMSC clones in vitroand in vivo and reported that the hMSC clones with highlevels of ALP expression were committed to trilineage differ-entiation [13] Our data corroborate and extend these findingsby reporting the effects of siRNA-mediated inhibition ofALP that resulted in an impaired hBMSC differentiation notonly to osteoblasts but also to adipocytes Also our resultscorroborate earlier studies that demonstrated in human bonebiopsies the presence of ALP expression in bone marrowadipocytic cells [13 17] All these data suggest that ALP isa ldquostemnessrdquo marker of hBMSC and not just an indicator ofosteoblastic lineage commitment
While CL1 and CL2 were isolated from telomerizedhMSC cell line they are relevant to normal human physi-ology We observed that the molecular phenotypes of CL1and CL2 were contained within the molecular signatureof primary hBMSC suggesting that CL1 and CL2 repre-sent cell populations within the heterogeneous cultures ofhBMSCWe have also previously reported that the molecular
14 Stem Cells International
ALZ
stai
ning
Control siRNA ALP siRNA
ALP
Beta
-act
in
Con
trol
ALP
siRN
A d3
ALP
siRN
A d7
Nile red quantificationlowast
NI ADIP I ALP siRNA0
50
100
150
200
Con
trol s
iRN
AA
LP si
RNA
Day 3 Day 5 OS induction Day 7 OS induction
Fold
indu
ctio
n lo
g sc
ale
10000
1000
100
10
1
Day 0 OS day 7
Control siRNAALP siRNA
(A) (B)
(C)
(D)
Figure 4 Effect of alkaline phosphatase (ALP) gene silencing by small interfering RNA (siRNA) on a clonal population of human bonemarrow stromal stem cell CL1 (A) ALP staining at day 3 day 5 or day 7 days of osteoblast differentiation (OB induction) (B) Quantitativereal-time PCR for ALP gene following ALP siRNA transfection at day 0 OB and day 7 of OB Data are presented as fold change in expressionof each target gene normalized to GAPDH (119899 = three independent experiments 119901 lt 001) Western blotting analysis of day 3 and day 7after siRNA ALP transfection of CL1 cells ALPL specific antibody and B-actin was used (C) Mineralized matrix formation as visualizedby Alizarin Red S staining in siRNA transfected CL1 cells after 14 days of OB induction (D) Nile red quantification of mature lipid filledadipocyte in control noninduced (Cont) adipocyte induced (Adip I) and ALP siRNA transfected cells that are adipocyte induced (ALPsiRNA) Adipocyte induction was carried out for 7 days lowast119901 lt 005
phenotype and cellular responses of hMSC-TERT are similarto those of primary hMSC [10]While we have identified ALPas a marker that can be used for a prospective identificationof differentiation committed population of hBMSC we iden-tified additional distinctive molecular markers of the cellsFor example IGF-1 IGF-2 and IGF binding protein 5 wereenriched in CL1 compared to CL2 IGFs and their bindingproteins are very well-studied factors that play a role inhBMSC proliferation and osteoblast differentiation [18] On
the other hand annexin A3 as well as several immune-relatedgenes was highly enriched in CL2 compared to CL1 Futurestudies are needed to determine the functional significanceof these molecules in relation to the functional identity ofvarious cell populations within the hBMSC cultures andtheir usefulness as biomarkers to dissect the heterogeneouspopulation of cultured hBMSC
Our finding of the presence of functional diversity withinhBMSC cultures that contain progenitor and nonprogenitor
Stem Cells International 15
Table 4 Microarray data analysis showing genes found to be upregulated during adipogenic differentiation and downregulated after ALPKO
62 common elements inldquoAD uprdquo and ldquoALP downrdquo Gene name FC (ALP siRNA versus
control siRNA)APOBEC3G Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3G minus15938025IFI44L Interferon-induced protein 44-like minus11520283PAQR5 Progestin and adipoQ receptor family member V minus6868241PNMA2 Paraneoplastic antigen MA2 minus59695344DUSP23 Dual specificity phosphatase 23 minus54786854CLDN23 Claudin 23 minus51885047ANKDD1A Ankyrin repeat and death domain containing 1A minus51646647IL8 Interleukin 8 minus4887188LRRC23 Leucine rich repeat containing 23 minus47611775IL6 Interleukin 6 (interferon beta 2) minus4693139LIFR Leukemia inhibitory factor receptor alpha minus46540866PTGFR Prostaglandin F receptor (FP) minus4457529FAM134B Family with sequence similarity 134 member B minus4403495CYFIP2 Cytoplasmic FMR1 interacting protein 2 minus4260462METTL7A Methyltransferase-like 7A minus40480843APOBEC3F Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3F minus39502614CA5B Carbonic anhydrase VB mitochondrial minus393889ITGA10 Integrin alpha 10 minus39143775FMO3 Flavin containing monooxygenase 3 minus3852087IMPA2 Inositol monophosphatase 2 (human) minus38374884CDO1 Cysteine dioxygenase type I minus38181455CCDC68 Coiled-coil domain containing 68 minus37292893
HEXDC Hexosaminidase (glycosyl hydrolase family 20 catalytic domain) containing minus2094836SARM1 Sterile alpha and TIR motif containing 1 minus20797038BBS1 Bardet-Biedl syndrome 1 minus20146718SERPING1 Serpin peptidase inhibitor clade G (C1 inhibitor) member 1 minus20102212FAM162A Family with sequence similarity 162 member A minus2005807TCTN1 Tectonic family member 1 minus20033443
AD up ALP down
1006 62 1145
(a)
140
120
100
080
060
040
020
mdash
ALPL PPARg aP2 LPL
lowastlowastlowastlowastlowastlowast
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
lowast
(b)
Figure 5 Adipocyte related genes downregulated in ALP knockdown CL1 cells (a) Venn diagram of whole gene expression analysis of ADupregulated genes compared with ALP KD downregulated genes (b) Quantitative real-time PCR of four selected common genes from Venndiagram including ALPL PPARg2 aP2 and LPL Data are presented as fold change in expression of each target gene normalized to GAPDH(119899 = three independent experiments lowast119901 lt 005 lowastlowast119901 lt 001 and lowastlowastlowast119901 lt 0001)
Stem Cells International 17
cell populations has a clinical relevance It demonstrates thatthe progenitor function and the immune modulatory rolesof hBMSC [39] are mediated by specific and distinguishablepopulations of hBMSC Thus future clinical studies employ-ing hBMSC should attempt to administer the relevant sub-population of hBMSC dependent on the experimental aim asa novel approach to improving the clinical efficiency insteadof the current use of heterogeneous hBMSC populations
Competing Interests
The authors declare that they have no competing interests
Authorsrsquo Contributions
Mona Elsafadi and Amer Mahmood contributed equally tothis work
Acknowledgments
These studies are supported by a grant from the College ofMedicine Research Centre (CMRC) Deanship of ScientificResearch King Saud University Riyadh Saudi Arabia Theauthors would like to thank Mr Ali Al-Roalle for all histechnical help They thank Natalie Twine for advice on DNAmicroarray analysis
References
[1] W Zaher L Harkness A Jafari and M Kassem ldquoAn update ofhuman mesenchymal stem cell biology and their clinical usesrdquoArchives of Toxicology vol 88 no 5 pp 1069ndash1082 2014
[2] S Post B M Abdallah J F Bentzon andM Kassem ldquoDemon-stration of the presence of independent pre-osteoblastic andpre-adipocytic cell populations in bone marrow-derived mes-enchymal stem cellsrdquo Bone vol 43 no 1 pp 32ndash39 2008
[3] K H Larsen C M Frederiksen J S Burns B M Abdallahand M Kassem ldquoIdentifying a molecular phenotype for bonemarrow stromal cells with in vivo bone-forming capacityrdquoJournal of Bone and Mineral Research vol 25 no 4 pp 796ndash808 2010
[4] A Rosenzweig ldquoCardiac cell therapymdashmixed results frommixed cellsrdquoThe New England Journal of Medicine vol 355 no12 pp 1274ndash1277 2006
[5] S A Kuznetsov P H Krebsbach K Satomura et al ldquoSingle-colony derived strains of human marrow stromal fibroblastsform bone after transplantation in vivordquo Journal of Bone andMineral Research vol 12 no 9 pp 1335ndash1347 1997
[6] M Kassem and P Bianco ldquoSkeletal stem cells in space and timerdquoCell vol 160 no 1-2 pp 17ndash19 2015
[7] A I Caplan and D Correa ldquoPDGF in bone formation andregeneration new insights into a novel mechanism involvingMSCsrdquo Journal of Orthopaedic Research vol 29 no 12 pp 1795ndash1803 2011
[8] K Stewart S Walsh J Screen et al ldquoFurther characterizationof cells expressing STRO-1 in cultures of adult human bonemarrow stromal cellsrdquo Journal of Bone and Mineral Researchvol 14 no 8 pp 1345ndash1356 1999
[9] J L Simonsen C Rosada N Serakinci et al ldquoTelomeraseexpression extends the proliferative life-span and maintains
the osteogenic potential of human bone marrow stromal cellsrdquoNature Biotechnology vol 20 no 6 pp 592ndash596 2002
[10] M Al-Nbaheen R Vishnubalaji D Ali et al ldquoHuman stromal(mesenchymal) stem cells from bone marrow adipose tissueand skin exhibit differences in molecular phenotype and differ-entiation potentialrdquo Stem Cell Reviews and Reports vol 9 no 1pp 32ndash43 2013
[11] C A Gregory W G Gunn A Peister and D J Prockop ldquoAnAlizarin red-based assay of mineralization by adherent cells inculture comparison with cetylpyridinium chloride extractionrdquoAnalytical Biochemistry vol 329 no 1 pp 77ndash84 2004
[12] AMahmood L Harkness H D Schroslashder B M Abdallah andM Kassem ldquoEnhanced differentiation of human embryonicstem cells to mesenchymal progenitors by inhibition of TGF-120573activinnodal signaling using SB-431542rdquo Journal of Bone andMineral Research vol 25 no 6 pp 1216ndash1233 2010
[13] H-J Prins A K Braat D Gawlitta et al ldquoIn vitro inductionof alkaline phosphatase levels predicts in vivo bone formingcapacity of human bone marrow stromal cellsrdquo Stem CellResearch vol 12 no 2 pp 428ndash440 2014
[14] K Stenderup J Justesen C Clausen and M Kassem ldquoAgingis associated with decreased maximal life span and acceleratedsenescence of bone marrow stromal cellsrdquo Bone vol 33 no 6pp 919ndash926 2003
[15] D G Phinney G Kopen W Righter S Webster N Tremainand D J Prockop ldquoDonor variation in the growth propertiesand osteogenic potential of human marrow stromal cellsrdquoJournal of Cellular Biochemistry vol 75 no 3 pp 424ndash436 1999
[16] C M Digirolamo D Stokes D Colter D G Phinney R Classand D J Prockop ldquoPropagation and senescence of humanmarrow stromal cells in culture a simple colony-forming assayidentifies samples with the greatest potential to propagate anddifferentiaterdquo British Journal of Haematology vol 107 no 2 pp275ndash281 1999
[17] S James J Fox F Afsari et al ldquoMultiparameter analy-sis of human bone marrow stromal cells identifies distinctimmunomodulatory and differentiation-competent subtypesrdquoStem Cell Reports vol 4 no 6 pp 1004ndash1015 2015
[18] B Peruzzi A Cappariello A Del Fattore N Rucci F DeBenedetti and A Teti ldquoc-Src and IL-6 inhibit osteoblastdifferentiation and integrate IGFBP5 signallingrdquo Nature Com-munications vol 3 article 630 2012
[19] D Fortunati S Reppe A-K Fjeldheim M Nielsen V TGautvik and K M Gautvik ldquoPeriostin is a collagen associatedbonematrix protein regulated by parathyroid hormonerdquoMatrixBiology vol 29 no 7 pp 594ndash601 2010
[20] A Oshima H Tanabe T Yan G N Lowe C A Glackin andA Kudo ldquoA novel mechanism for the regulation of osteoblastdifferentiation transcription of periostin a member of thefasciclin I family is regulated by the bHLH transcription factorTwistrdquo Journal of Cellular Biochemistry vol 86 no 4 pp 792ndash804 2002
[21] T G Kashima T Nishiyama K Shimazu et al ldquoPeriostina novel marker of intramembranous ossification is expressedin fibrous dysplasia and in c-Fos-overexpressing bone lesionsrdquoHuman Pathology vol 40 no 2 pp 226ndash237 2009
[22] CGalliM PiergianniM Piemontese et al ldquoPeriostin improvescell adhesion to implantable biomaterials and osteoblasticdifferentiation on implant titanium surfaces in a topography-dependent fashionrdquo Journal of Biomedical Materials Research -Part A vol 102 no 11 pp 3855ndash3861 2014
18 Stem Cells International
[23] D L Coutu H W Jian A Monette G-E Rivard M DBlostein and J Galipeau ldquoPeriostin amember of a novel familyof vitamin K-dependent proteins is expressed by mesenchymalstromal cellsrdquo The Journal of Biological Chemistry vol 283 no26 pp 17991ndash18001 2008
[24] T Minamizato K Sakamoto T Liu et al ldquoCCN3NOV inhibitsBMP-2-induced osteoblast differentiation by interacting withBMP andNotch signaling pathwaysrdquo Biochemical and Biophysi-cal Research Communications vol 354 no 2 pp 567ndash573 2007
[25] W Si Q Kang H H Luu et al ldquoCCN1Cyr61 is regulated bythe canonicalWnt signal and plays an important role inWnt3A-induced osteoblast differentiation of mesenchymal stem cellsrdquoMolecular and Cellular Biology vol 26 no 8 pp 2955ndash29642006
[26] Q Luo Q Kang W Si et al ldquoConnective tissue growth factor(CTGF) is regulated by Wnt and bone morphogenetic proteinssignaling in osteoblast differentiation of mesenchymal stemcellsrdquo The Journal of Biological Chemistry vol 279 no 53 pp55958ndash55968 2004
[27] Y-S Yoon N Lee and H Scadova ldquoMyocardial regenerationwith bone-marrow-derived stem cellsrdquo Biology of the Cell vol97 no 4 pp 253ndash263 2005
[28] Y S Yoon ldquoCardiac regeneration with novel bone marrow-derived multipotent stem cellsrdquo Discovery Medicine vol 5 pp204ndash208 2005
[29] Y-S Yoon A Wecker L Heyd et al ldquoClonally expanded novelmultipotent stem cells from human bone marrow regeneratemyocardium after myocardial infarctionrdquo Journal of ClinicalInvestigation vol 115 no 2 pp 326ndash338 2005
[30] GM BolandG Perkins D J Hall andR S Tuan ldquoWnt 3a pro-motes proliferation and suppresses osteogenic differentiationof adult human mesenchymal stem cellsrdquo Journal of CellularBiochemistry vol 93 no 6 pp 1210ndash1230 2004
[31] D A Cook S W Fellgett M E Pownall P J OrsquoShea and PG Genever ldquoWnt-dependent osteogenic commitment of bonemarrow stromal cells using a novel GSK3120573 inhibitorrdquo Stem CellResearch vol 12 no 2 pp 415ndash427 2014
[32] J De Boer R Siddappa C Gaspar A VanApeldoorn R Foddeand C Van Blitterswijk ldquoWnt signaling inhibits osteogenicdifferentiation of humanmesenchymal stem cellsrdquoBone vol 34no 5 pp 818ndash826 2004
[33] MAl-Toub A AlmusaM Almajed et al ldquoPleiotropic effects ofcancer cellsrsquo secreted factors on human stromal (mesenchymal)stem cellsrdquo Stem Cell Research and Therapy vol 4 article 1142013
[34] L Chen D Qanie A Jafari et al ldquoDelta-like 1fetal antigen-1(Dlk1FA1) is a novel regulator of chondrogenic cell differenti-ation via inhibition of the Akt kinase-dependent pathwayrdquoTheJournal of Biological Chemistry vol 286 no 37 pp 32140ndash321492011
[35] T Matsunobu K Torigoe M Ishikawa et al ldquoCritical roles ofthe TGF-120573 type I receptor ALK5 in perichondrial formationand function cartilage integrity and osteoblast differentiationduring growth plate developmentrdquo Developmental Biology vol332 no 2 pp 325ndash338 2009
[36] B C Nair D E Johnson R J Majeska J A Rodkey C DBennett and G A Rodan ldquoRat alkaline phosphatase II Struc-tural similarities between the osteosarcoma bone kidney andplacenta isoenzymesrdquo Archives of Biochemistry and Biophysicsvol 254 no 1 pp 28ndash34 1987
[37] P J Butterworth ldquoAlkaline phosphatase Biochemistry of mam-malian alkaline phosphatasesrdquo Cell Biochemistry and Functionvol 1 no 2 pp 66ndash70 1983
[38] A T Ali C B Penny J E Paiker et al ldquoAlkaline phosphataseis involved in the control of adipogenesis in the murinepreadipocyte cell line 3T3-L1rdquo Clinica Chimica Acta vol 354no 1-2 pp 101ndash109 2005
[39] G M Spaggiari and L Moretta ldquoCellular and molecularinteractions of mesenchymal stem cells in innate immunityrdquoImmunology and Cell Biology vol 91 no 1 pp 27ndash31 2012
Table 3 Whole genome microarray analysis of two clonal popu-lations of human bone marrow stromal stem cells CL1 and CL2cells (a) Top 35 highly upregulated genes in CL1 cells versus CL2cells (b) Ingenuity Pathway Analysis (IPA) showing the differentphysiological system development and function genes found in eachcategory and the corresponding 119901 value (c) Upper part IPA analysisshowing the categories for skeletal and muscular system developmentand function upregulated in CL1 cells and lower part showing genesupregulated in CL2 cells NOV = nephroblastoma overexpressedIGFBP5 = insulin-like growth factor binding protein 5 ALP TAGLN= transgelin and CDH11 = OB-cadherin (osteoblast)
Name 119901 value moleculesPhysiological system development and function
Organismal development 705119864 minus 10ndash143119864 minus 03 181Embryonic development 100119864 minus 09ndash143119864 minus 03 154Organ development 100119864 minus 09ndash143119864 minus 03 145
(b) Continued
Name 119901 value moleculesSkeletal and muscularsystem development andfunction
100119864 minus 09ndash102119864 minus 03 123
Tissue development 100119864 minus 09ndash143119864 minus 03 236
(c)
Functions annotation 119901 value moleculesSkeletal and muscular system development and
function upregulated in CL1 cellsSize of bone 143119864 minus 06 24Differentiation ofosteoblasts 384119864 minus 06 25
Mineralization ofbone 493119864 minus 06 19
Bone mineral density 365119864 minus 05 19Skeletal and muscular system development and
function upregulated in CL2 cellsDevelopment ofmuscle 100119864 minus 09 44
Proliferation ofmuscle cells 229119864 minus 06 35
Remodeling of bone 363119864 minus 06 21Resorption of bone 394119864 minus 06 19
confirming impairment of adipocytic differentiation of ALPdeficient CL1 cells
4 Discussion
We extensively studied two cell populations within culturedhBMSC that were identified based on differences in mor-phology Cellular and molecular studies revealed differencesin growth differentiation capacity and molecular signatureOur data support the notion of the presence of cellular andfunctional heterogeneity among cultured hBMSC
Cellular heterogeneity of cultured hBMSC is recognizedin an increasing number of reports Several extrinsic andintrinsic factors may contribute to the observed hBMSCheterogeneity Extrinsic factors include donor-to-donor vari-ations in the number and quantity of initiating cells whichresult in differences in cell growth rate and differentia-tion capacity [14 15] Intrinsic factors have been examinedemploying single cell clonal analysis and revealed variationsin differentiation potential among individual colonies withinhBMSC cultures ranging from the presence of cells withtrilineage (osteoblast adipocytes and chondrocyte) potencyto cells with null potency [16] Also variations in the abilityof clonal cells to form heterotopic bone when implanted invivo have been reported [5] Our study corroborates thesefindings and provides more detailed cellular and molecularphenotyping of two examples of cell populations that existwithin the heterogeneous hBMSC cultures [17]
Determining the molecular signature of CL1 and CL2using whole genomemicroarray analysis showed enrichmentof lineage-commitment associated genes inCL1 For exampleinsulin-like growth factor 5 (IGFBP5) and interleukin 6 (IL6)
Stem Cells International 13
NOV0040
0030
0020
0010
mdash
IGFBP50015
0010
0005
mdash
ALP0100
0080
0060
0040
0020
mdash
TAGLN0800
0600
0400
0200
CDH11
0060
0040
0020
mdashmdash
LRP30015
0010
0005
mdash
CL1 CL2 CL1 CL2 CL1 CL2
CL1 CL2CL1 CL2CL1 CL2
Figure 3 Validation of whole genome microarray analysis of two clonal populations of human bone marrow stromal stem cells CL1 andCL2 cells Quantitative real-time PCR for highly expressed genes in CL1 cells NOV = nephroblastoma overexpressed IGFBP5 = insulin-likegrowth factor binding protein 5 ALP TAGLN = transgelin and CDH11 = OB-cadherin (osteoblast) Data are presented as fold change inexpression of each target gene normalized to GAPDH (119899 = three independent experiments 119901 lt 0001) (see also Table 3)
were 147- and 33-fold upregulated in CL1 cells respectivelyBoth factors are expressed in osteoprogenitor cells andimportant for osteoblast maturation [18] We also observedthat periostin (POSTN) gene was highly upregulated inCL1 cells (156-fold) POSTN is a 90 kDa secreted proteinoriginally identified in murine osteoblast-like cells and isupregulated by PTH [19] Several studies employing murineand human cells have revealed important role of POSTNin osteoblast differentiation and during development inintramembranous ossification [20ndash23] Another factor iden-tified in CL1 cells is nephroblastoma overexpression (NOV)which is a member of the Cyr 61 connective tissue growthfactor (CNN) family The CCN family of proteins promotesosteoblast differentiation through interaction with integrinsWNT BMP and NOTCH signaling pathways [24ndash26] Inaddition a large number of signaling molecules known tobe regulators of hBMSC lineage specific differentiation forexample insulin-like growth factors [27ndash29] WNT [30ndash32]and MAPK [33ndash35] were enriched in CL1 cells In contrastCL2 expressed high levels of immune-related genes whichmay explain the poor differentiation response to osteoblastor adipocyte lineage In a recent study the authors usedtelomerized hBMSC and showed clearly a clonal populationthat had very low in vitro and in vivo differentiation abilityhowever they had enhanced immune-related features includ-ing high IL7 expression These nullipotent cells expressedCD317 which was associated with remarkably high basal levelexpression of factors with a proinflammatory and antiviralfunction [17] We observed that this molecular phenotypewas associated with distinct ultrastructural characteristicsof the cells In particular CL2 had abundant phagocyticvacuole microvilli and lysosomes features reminiscent of
ultrastructure of immune-regulatory cells Our data thussupport the increasingly recognized feature that hBMSCexhibit immune modulatory functions and a part of theinnate immune response [17]
We observed that ALP protein expression and enzymaticactivity were significantly different between CL1 and CL2cell lines and were thus a potential marker that distinguishesdifferent cell populations with progenitor functions (CL1)from cells with nonprogenitor functions (CL2) ALP isexpressed in a wide variety of tissues including kidneysbone and liver [36 37] but tissue-nonspecific ALP (ALPL)is considered a commitment marker for osteoblastic lineage[13 38] However in a recent study the authors examined thedifferentiation potential of a number of hMSC clones in vitroand in vivo and reported that the hMSC clones with highlevels of ALP expression were committed to trilineage differ-entiation [13] Our data corroborate and extend these findingsby reporting the effects of siRNA-mediated inhibition ofALP that resulted in an impaired hBMSC differentiation notonly to osteoblasts but also to adipocytes Also our resultscorroborate earlier studies that demonstrated in human bonebiopsies the presence of ALP expression in bone marrowadipocytic cells [13 17] All these data suggest that ALP isa ldquostemnessrdquo marker of hBMSC and not just an indicator ofosteoblastic lineage commitment
While CL1 and CL2 were isolated from telomerizedhMSC cell line they are relevant to normal human physi-ology We observed that the molecular phenotypes of CL1and CL2 were contained within the molecular signatureof primary hBMSC suggesting that CL1 and CL2 repre-sent cell populations within the heterogeneous cultures ofhBMSCWe have also previously reported that the molecular
14 Stem Cells International
ALZ
stai
ning
Control siRNA ALP siRNA
ALP
Beta
-act
in
Con
trol
ALP
siRN
A d3
ALP
siRN
A d7
Nile red quantificationlowast
NI ADIP I ALP siRNA0
50
100
150
200
Con
trol s
iRN
AA
LP si
RNA
Day 3 Day 5 OS induction Day 7 OS induction
Fold
indu
ctio
n lo
g sc
ale
10000
1000
100
10
1
Day 0 OS day 7
Control siRNAALP siRNA
(A) (B)
(C)
(D)
Figure 4 Effect of alkaline phosphatase (ALP) gene silencing by small interfering RNA (siRNA) on a clonal population of human bonemarrow stromal stem cell CL1 (A) ALP staining at day 3 day 5 or day 7 days of osteoblast differentiation (OB induction) (B) Quantitativereal-time PCR for ALP gene following ALP siRNA transfection at day 0 OB and day 7 of OB Data are presented as fold change in expressionof each target gene normalized to GAPDH (119899 = three independent experiments 119901 lt 001) Western blotting analysis of day 3 and day 7after siRNA ALP transfection of CL1 cells ALPL specific antibody and B-actin was used (C) Mineralized matrix formation as visualizedby Alizarin Red S staining in siRNA transfected CL1 cells after 14 days of OB induction (D) Nile red quantification of mature lipid filledadipocyte in control noninduced (Cont) adipocyte induced (Adip I) and ALP siRNA transfected cells that are adipocyte induced (ALPsiRNA) Adipocyte induction was carried out for 7 days lowast119901 lt 005
phenotype and cellular responses of hMSC-TERT are similarto those of primary hMSC [10]While we have identified ALPas a marker that can be used for a prospective identificationof differentiation committed population of hBMSC we iden-tified additional distinctive molecular markers of the cellsFor example IGF-1 IGF-2 and IGF binding protein 5 wereenriched in CL1 compared to CL2 IGFs and their bindingproteins are very well-studied factors that play a role inhBMSC proliferation and osteoblast differentiation [18] On
the other hand annexin A3 as well as several immune-relatedgenes was highly enriched in CL2 compared to CL1 Futurestudies are needed to determine the functional significanceof these molecules in relation to the functional identity ofvarious cell populations within the hBMSC cultures andtheir usefulness as biomarkers to dissect the heterogeneouspopulation of cultured hBMSC
Our finding of the presence of functional diversity withinhBMSC cultures that contain progenitor and nonprogenitor
Stem Cells International 15
Table 4 Microarray data analysis showing genes found to be upregulated during adipogenic differentiation and downregulated after ALPKO
62 common elements inldquoAD uprdquo and ldquoALP downrdquo Gene name FC (ALP siRNA versus
control siRNA)APOBEC3G Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3G minus15938025IFI44L Interferon-induced protein 44-like minus11520283PAQR5 Progestin and adipoQ receptor family member V minus6868241PNMA2 Paraneoplastic antigen MA2 minus59695344DUSP23 Dual specificity phosphatase 23 minus54786854CLDN23 Claudin 23 minus51885047ANKDD1A Ankyrin repeat and death domain containing 1A minus51646647IL8 Interleukin 8 minus4887188LRRC23 Leucine rich repeat containing 23 minus47611775IL6 Interleukin 6 (interferon beta 2) minus4693139LIFR Leukemia inhibitory factor receptor alpha minus46540866PTGFR Prostaglandin F receptor (FP) minus4457529FAM134B Family with sequence similarity 134 member B minus4403495CYFIP2 Cytoplasmic FMR1 interacting protein 2 minus4260462METTL7A Methyltransferase-like 7A minus40480843APOBEC3F Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3F minus39502614CA5B Carbonic anhydrase VB mitochondrial minus393889ITGA10 Integrin alpha 10 minus39143775FMO3 Flavin containing monooxygenase 3 minus3852087IMPA2 Inositol monophosphatase 2 (human) minus38374884CDO1 Cysteine dioxygenase type I minus38181455CCDC68 Coiled-coil domain containing 68 minus37292893
HEXDC Hexosaminidase (glycosyl hydrolase family 20 catalytic domain) containing minus2094836SARM1 Sterile alpha and TIR motif containing 1 minus20797038BBS1 Bardet-Biedl syndrome 1 minus20146718SERPING1 Serpin peptidase inhibitor clade G (C1 inhibitor) member 1 minus20102212FAM162A Family with sequence similarity 162 member A minus2005807TCTN1 Tectonic family member 1 minus20033443
AD up ALP down
1006 62 1145
(a)
140
120
100
080
060
040
020
mdash
ALPL PPARg aP2 LPL
lowastlowastlowastlowastlowastlowast
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
lowast
(b)
Figure 5 Adipocyte related genes downregulated in ALP knockdown CL1 cells (a) Venn diagram of whole gene expression analysis of ADupregulated genes compared with ALP KD downregulated genes (b) Quantitative real-time PCR of four selected common genes from Venndiagram including ALPL PPARg2 aP2 and LPL Data are presented as fold change in expression of each target gene normalized to GAPDH(119899 = three independent experiments lowast119901 lt 005 lowastlowast119901 lt 001 and lowastlowastlowast119901 lt 0001)
Stem Cells International 17
cell populations has a clinical relevance It demonstrates thatthe progenitor function and the immune modulatory rolesof hBMSC [39] are mediated by specific and distinguishablepopulations of hBMSC Thus future clinical studies employ-ing hBMSC should attempt to administer the relevant sub-population of hBMSC dependent on the experimental aim asa novel approach to improving the clinical efficiency insteadof the current use of heterogeneous hBMSC populations
Competing Interests
The authors declare that they have no competing interests
Authorsrsquo Contributions
Mona Elsafadi and Amer Mahmood contributed equally tothis work
Acknowledgments
These studies are supported by a grant from the College ofMedicine Research Centre (CMRC) Deanship of ScientificResearch King Saud University Riyadh Saudi Arabia Theauthors would like to thank Mr Ali Al-Roalle for all histechnical help They thank Natalie Twine for advice on DNAmicroarray analysis
References
[1] W Zaher L Harkness A Jafari and M Kassem ldquoAn update ofhuman mesenchymal stem cell biology and their clinical usesrdquoArchives of Toxicology vol 88 no 5 pp 1069ndash1082 2014
[2] S Post B M Abdallah J F Bentzon andM Kassem ldquoDemon-stration of the presence of independent pre-osteoblastic andpre-adipocytic cell populations in bone marrow-derived mes-enchymal stem cellsrdquo Bone vol 43 no 1 pp 32ndash39 2008
[3] K H Larsen C M Frederiksen J S Burns B M Abdallahand M Kassem ldquoIdentifying a molecular phenotype for bonemarrow stromal cells with in vivo bone-forming capacityrdquoJournal of Bone and Mineral Research vol 25 no 4 pp 796ndash808 2010
[4] A Rosenzweig ldquoCardiac cell therapymdashmixed results frommixed cellsrdquoThe New England Journal of Medicine vol 355 no12 pp 1274ndash1277 2006
[5] S A Kuznetsov P H Krebsbach K Satomura et al ldquoSingle-colony derived strains of human marrow stromal fibroblastsform bone after transplantation in vivordquo Journal of Bone andMineral Research vol 12 no 9 pp 1335ndash1347 1997
[6] M Kassem and P Bianco ldquoSkeletal stem cells in space and timerdquoCell vol 160 no 1-2 pp 17ndash19 2015
[7] A I Caplan and D Correa ldquoPDGF in bone formation andregeneration new insights into a novel mechanism involvingMSCsrdquo Journal of Orthopaedic Research vol 29 no 12 pp 1795ndash1803 2011
[8] K Stewart S Walsh J Screen et al ldquoFurther characterizationof cells expressing STRO-1 in cultures of adult human bonemarrow stromal cellsrdquo Journal of Bone and Mineral Researchvol 14 no 8 pp 1345ndash1356 1999
[9] J L Simonsen C Rosada N Serakinci et al ldquoTelomeraseexpression extends the proliferative life-span and maintains
the osteogenic potential of human bone marrow stromal cellsrdquoNature Biotechnology vol 20 no 6 pp 592ndash596 2002
[10] M Al-Nbaheen R Vishnubalaji D Ali et al ldquoHuman stromal(mesenchymal) stem cells from bone marrow adipose tissueand skin exhibit differences in molecular phenotype and differ-entiation potentialrdquo Stem Cell Reviews and Reports vol 9 no 1pp 32ndash43 2013
[11] C A Gregory W G Gunn A Peister and D J Prockop ldquoAnAlizarin red-based assay of mineralization by adherent cells inculture comparison with cetylpyridinium chloride extractionrdquoAnalytical Biochemistry vol 329 no 1 pp 77ndash84 2004
[12] AMahmood L Harkness H D Schroslashder B M Abdallah andM Kassem ldquoEnhanced differentiation of human embryonicstem cells to mesenchymal progenitors by inhibition of TGF-120573activinnodal signaling using SB-431542rdquo Journal of Bone andMineral Research vol 25 no 6 pp 1216ndash1233 2010
[13] H-J Prins A K Braat D Gawlitta et al ldquoIn vitro inductionof alkaline phosphatase levels predicts in vivo bone formingcapacity of human bone marrow stromal cellsrdquo Stem CellResearch vol 12 no 2 pp 428ndash440 2014
[14] K Stenderup J Justesen C Clausen and M Kassem ldquoAgingis associated with decreased maximal life span and acceleratedsenescence of bone marrow stromal cellsrdquo Bone vol 33 no 6pp 919ndash926 2003
[15] D G Phinney G Kopen W Righter S Webster N Tremainand D J Prockop ldquoDonor variation in the growth propertiesand osteogenic potential of human marrow stromal cellsrdquoJournal of Cellular Biochemistry vol 75 no 3 pp 424ndash436 1999
[16] C M Digirolamo D Stokes D Colter D G Phinney R Classand D J Prockop ldquoPropagation and senescence of humanmarrow stromal cells in culture a simple colony-forming assayidentifies samples with the greatest potential to propagate anddifferentiaterdquo British Journal of Haematology vol 107 no 2 pp275ndash281 1999
[17] S James J Fox F Afsari et al ldquoMultiparameter analy-sis of human bone marrow stromal cells identifies distinctimmunomodulatory and differentiation-competent subtypesrdquoStem Cell Reports vol 4 no 6 pp 1004ndash1015 2015
[18] B Peruzzi A Cappariello A Del Fattore N Rucci F DeBenedetti and A Teti ldquoc-Src and IL-6 inhibit osteoblastdifferentiation and integrate IGFBP5 signallingrdquo Nature Com-munications vol 3 article 630 2012
[19] D Fortunati S Reppe A-K Fjeldheim M Nielsen V TGautvik and K M Gautvik ldquoPeriostin is a collagen associatedbonematrix protein regulated by parathyroid hormonerdquoMatrixBiology vol 29 no 7 pp 594ndash601 2010
[20] A Oshima H Tanabe T Yan G N Lowe C A Glackin andA Kudo ldquoA novel mechanism for the regulation of osteoblastdifferentiation transcription of periostin a member of thefasciclin I family is regulated by the bHLH transcription factorTwistrdquo Journal of Cellular Biochemistry vol 86 no 4 pp 792ndash804 2002
[21] T G Kashima T Nishiyama K Shimazu et al ldquoPeriostina novel marker of intramembranous ossification is expressedin fibrous dysplasia and in c-Fos-overexpressing bone lesionsrdquoHuman Pathology vol 40 no 2 pp 226ndash237 2009
[22] CGalliM PiergianniM Piemontese et al ldquoPeriostin improvescell adhesion to implantable biomaterials and osteoblasticdifferentiation on implant titanium surfaces in a topography-dependent fashionrdquo Journal of Biomedical Materials Research -Part A vol 102 no 11 pp 3855ndash3861 2014
18 Stem Cells International
[23] D L Coutu H W Jian A Monette G-E Rivard M DBlostein and J Galipeau ldquoPeriostin amember of a novel familyof vitamin K-dependent proteins is expressed by mesenchymalstromal cellsrdquo The Journal of Biological Chemistry vol 283 no26 pp 17991ndash18001 2008
[24] T Minamizato K Sakamoto T Liu et al ldquoCCN3NOV inhibitsBMP-2-induced osteoblast differentiation by interacting withBMP andNotch signaling pathwaysrdquo Biochemical and Biophysi-cal Research Communications vol 354 no 2 pp 567ndash573 2007
[25] W Si Q Kang H H Luu et al ldquoCCN1Cyr61 is regulated bythe canonicalWnt signal and plays an important role inWnt3A-induced osteoblast differentiation of mesenchymal stem cellsrdquoMolecular and Cellular Biology vol 26 no 8 pp 2955ndash29642006
[26] Q Luo Q Kang W Si et al ldquoConnective tissue growth factor(CTGF) is regulated by Wnt and bone morphogenetic proteinssignaling in osteoblast differentiation of mesenchymal stemcellsrdquo The Journal of Biological Chemistry vol 279 no 53 pp55958ndash55968 2004
[27] Y-S Yoon N Lee and H Scadova ldquoMyocardial regenerationwith bone-marrow-derived stem cellsrdquo Biology of the Cell vol97 no 4 pp 253ndash263 2005
[28] Y S Yoon ldquoCardiac regeneration with novel bone marrow-derived multipotent stem cellsrdquo Discovery Medicine vol 5 pp204ndash208 2005
[29] Y-S Yoon A Wecker L Heyd et al ldquoClonally expanded novelmultipotent stem cells from human bone marrow regeneratemyocardium after myocardial infarctionrdquo Journal of ClinicalInvestigation vol 115 no 2 pp 326ndash338 2005
[30] GM BolandG Perkins D J Hall andR S Tuan ldquoWnt 3a pro-motes proliferation and suppresses osteogenic differentiationof adult human mesenchymal stem cellsrdquo Journal of CellularBiochemistry vol 93 no 6 pp 1210ndash1230 2004
[31] D A Cook S W Fellgett M E Pownall P J OrsquoShea and PG Genever ldquoWnt-dependent osteogenic commitment of bonemarrow stromal cells using a novel GSK3120573 inhibitorrdquo Stem CellResearch vol 12 no 2 pp 415ndash427 2014
[32] J De Boer R Siddappa C Gaspar A VanApeldoorn R Foddeand C Van Blitterswijk ldquoWnt signaling inhibits osteogenicdifferentiation of humanmesenchymal stem cellsrdquoBone vol 34no 5 pp 818ndash826 2004
[33] MAl-Toub A AlmusaM Almajed et al ldquoPleiotropic effects ofcancer cellsrsquo secreted factors on human stromal (mesenchymal)stem cellsrdquo Stem Cell Research and Therapy vol 4 article 1142013
[34] L Chen D Qanie A Jafari et al ldquoDelta-like 1fetal antigen-1(Dlk1FA1) is a novel regulator of chondrogenic cell differenti-ation via inhibition of the Akt kinase-dependent pathwayrdquoTheJournal of Biological Chemistry vol 286 no 37 pp 32140ndash321492011
[35] T Matsunobu K Torigoe M Ishikawa et al ldquoCritical roles ofthe TGF-120573 type I receptor ALK5 in perichondrial formationand function cartilage integrity and osteoblast differentiationduring growth plate developmentrdquo Developmental Biology vol332 no 2 pp 325ndash338 2009
[36] B C Nair D E Johnson R J Majeska J A Rodkey C DBennett and G A Rodan ldquoRat alkaline phosphatase II Struc-tural similarities between the osteosarcoma bone kidney andplacenta isoenzymesrdquo Archives of Biochemistry and Biophysicsvol 254 no 1 pp 28ndash34 1987
[37] P J Butterworth ldquoAlkaline phosphatase Biochemistry of mam-malian alkaline phosphatasesrdquo Cell Biochemistry and Functionvol 1 no 2 pp 66ndash70 1983
[38] A T Ali C B Penny J E Paiker et al ldquoAlkaline phosphataseis involved in the control of adipogenesis in the murinepreadipocyte cell line 3T3-L1rdquo Clinica Chimica Acta vol 354no 1-2 pp 101ndash109 2005
[39] G M Spaggiari and L Moretta ldquoCellular and molecularinteractions of mesenchymal stem cells in innate immunityrdquoImmunology and Cell Biology vol 91 no 1 pp 27ndash31 2012
Figure 3 Validation of whole genome microarray analysis of two clonal populations of human bone marrow stromal stem cells CL1 andCL2 cells Quantitative real-time PCR for highly expressed genes in CL1 cells NOV = nephroblastoma overexpressed IGFBP5 = insulin-likegrowth factor binding protein 5 ALP TAGLN = transgelin and CDH11 = OB-cadherin (osteoblast) Data are presented as fold change inexpression of each target gene normalized to GAPDH (119899 = three independent experiments 119901 lt 0001) (see also Table 3)
were 147- and 33-fold upregulated in CL1 cells respectivelyBoth factors are expressed in osteoprogenitor cells andimportant for osteoblast maturation [18] We also observedthat periostin (POSTN) gene was highly upregulated inCL1 cells (156-fold) POSTN is a 90 kDa secreted proteinoriginally identified in murine osteoblast-like cells and isupregulated by PTH [19] Several studies employing murineand human cells have revealed important role of POSTNin osteoblast differentiation and during development inintramembranous ossification [20ndash23] Another factor iden-tified in CL1 cells is nephroblastoma overexpression (NOV)which is a member of the Cyr 61 connective tissue growthfactor (CNN) family The CCN family of proteins promotesosteoblast differentiation through interaction with integrinsWNT BMP and NOTCH signaling pathways [24ndash26] Inaddition a large number of signaling molecules known tobe regulators of hBMSC lineage specific differentiation forexample insulin-like growth factors [27ndash29] WNT [30ndash32]and MAPK [33ndash35] were enriched in CL1 cells In contrastCL2 expressed high levels of immune-related genes whichmay explain the poor differentiation response to osteoblastor adipocyte lineage In a recent study the authors usedtelomerized hBMSC and showed clearly a clonal populationthat had very low in vitro and in vivo differentiation abilityhowever they had enhanced immune-related features includ-ing high IL7 expression These nullipotent cells expressedCD317 which was associated with remarkably high basal levelexpression of factors with a proinflammatory and antiviralfunction [17] We observed that this molecular phenotypewas associated with distinct ultrastructural characteristicsof the cells In particular CL2 had abundant phagocyticvacuole microvilli and lysosomes features reminiscent of
ultrastructure of immune-regulatory cells Our data thussupport the increasingly recognized feature that hBMSCexhibit immune modulatory functions and a part of theinnate immune response [17]
We observed that ALP protein expression and enzymaticactivity were significantly different between CL1 and CL2cell lines and were thus a potential marker that distinguishesdifferent cell populations with progenitor functions (CL1)from cells with nonprogenitor functions (CL2) ALP isexpressed in a wide variety of tissues including kidneysbone and liver [36 37] but tissue-nonspecific ALP (ALPL)is considered a commitment marker for osteoblastic lineage[13 38] However in a recent study the authors examined thedifferentiation potential of a number of hMSC clones in vitroand in vivo and reported that the hMSC clones with highlevels of ALP expression were committed to trilineage differ-entiation [13] Our data corroborate and extend these findingsby reporting the effects of siRNA-mediated inhibition ofALP that resulted in an impaired hBMSC differentiation notonly to osteoblasts but also to adipocytes Also our resultscorroborate earlier studies that demonstrated in human bonebiopsies the presence of ALP expression in bone marrowadipocytic cells [13 17] All these data suggest that ALP isa ldquostemnessrdquo marker of hBMSC and not just an indicator ofosteoblastic lineage commitment
While CL1 and CL2 were isolated from telomerizedhMSC cell line they are relevant to normal human physi-ology We observed that the molecular phenotypes of CL1and CL2 were contained within the molecular signatureof primary hBMSC suggesting that CL1 and CL2 repre-sent cell populations within the heterogeneous cultures ofhBMSCWe have also previously reported that the molecular
14 Stem Cells International
ALZ
stai
ning
Control siRNA ALP siRNA
ALP
Beta
-act
in
Con
trol
ALP
siRN
A d3
ALP
siRN
A d7
Nile red quantificationlowast
NI ADIP I ALP siRNA0
50
100
150
200
Con
trol s
iRN
AA
LP si
RNA
Day 3 Day 5 OS induction Day 7 OS induction
Fold
indu
ctio
n lo
g sc
ale
10000
1000
100
10
1
Day 0 OS day 7
Control siRNAALP siRNA
(A) (B)
(C)
(D)
Figure 4 Effect of alkaline phosphatase (ALP) gene silencing by small interfering RNA (siRNA) on a clonal population of human bonemarrow stromal stem cell CL1 (A) ALP staining at day 3 day 5 or day 7 days of osteoblast differentiation (OB induction) (B) Quantitativereal-time PCR for ALP gene following ALP siRNA transfection at day 0 OB and day 7 of OB Data are presented as fold change in expressionof each target gene normalized to GAPDH (119899 = three independent experiments 119901 lt 001) Western blotting analysis of day 3 and day 7after siRNA ALP transfection of CL1 cells ALPL specific antibody and B-actin was used (C) Mineralized matrix formation as visualizedby Alizarin Red S staining in siRNA transfected CL1 cells after 14 days of OB induction (D) Nile red quantification of mature lipid filledadipocyte in control noninduced (Cont) adipocyte induced (Adip I) and ALP siRNA transfected cells that are adipocyte induced (ALPsiRNA) Adipocyte induction was carried out for 7 days lowast119901 lt 005
phenotype and cellular responses of hMSC-TERT are similarto those of primary hMSC [10]While we have identified ALPas a marker that can be used for a prospective identificationof differentiation committed population of hBMSC we iden-tified additional distinctive molecular markers of the cellsFor example IGF-1 IGF-2 and IGF binding protein 5 wereenriched in CL1 compared to CL2 IGFs and their bindingproteins are very well-studied factors that play a role inhBMSC proliferation and osteoblast differentiation [18] On
the other hand annexin A3 as well as several immune-relatedgenes was highly enriched in CL2 compared to CL1 Futurestudies are needed to determine the functional significanceof these molecules in relation to the functional identity ofvarious cell populations within the hBMSC cultures andtheir usefulness as biomarkers to dissect the heterogeneouspopulation of cultured hBMSC
Our finding of the presence of functional diversity withinhBMSC cultures that contain progenitor and nonprogenitor
Stem Cells International 15
Table 4 Microarray data analysis showing genes found to be upregulated during adipogenic differentiation and downregulated after ALPKO
62 common elements inldquoAD uprdquo and ldquoALP downrdquo Gene name FC (ALP siRNA versus
control siRNA)APOBEC3G Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3G minus15938025IFI44L Interferon-induced protein 44-like minus11520283PAQR5 Progestin and adipoQ receptor family member V minus6868241PNMA2 Paraneoplastic antigen MA2 minus59695344DUSP23 Dual specificity phosphatase 23 minus54786854CLDN23 Claudin 23 minus51885047ANKDD1A Ankyrin repeat and death domain containing 1A minus51646647IL8 Interleukin 8 minus4887188LRRC23 Leucine rich repeat containing 23 minus47611775IL6 Interleukin 6 (interferon beta 2) minus4693139LIFR Leukemia inhibitory factor receptor alpha minus46540866PTGFR Prostaglandin F receptor (FP) minus4457529FAM134B Family with sequence similarity 134 member B minus4403495CYFIP2 Cytoplasmic FMR1 interacting protein 2 minus4260462METTL7A Methyltransferase-like 7A minus40480843APOBEC3F Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3F minus39502614CA5B Carbonic anhydrase VB mitochondrial minus393889ITGA10 Integrin alpha 10 minus39143775FMO3 Flavin containing monooxygenase 3 minus3852087IMPA2 Inositol monophosphatase 2 (human) minus38374884CDO1 Cysteine dioxygenase type I minus38181455CCDC68 Coiled-coil domain containing 68 minus37292893
HEXDC Hexosaminidase (glycosyl hydrolase family 20 catalytic domain) containing minus2094836SARM1 Sterile alpha and TIR motif containing 1 minus20797038BBS1 Bardet-Biedl syndrome 1 minus20146718SERPING1 Serpin peptidase inhibitor clade G (C1 inhibitor) member 1 minus20102212FAM162A Family with sequence similarity 162 member A minus2005807TCTN1 Tectonic family member 1 minus20033443
AD up ALP down
1006 62 1145
(a)
140
120
100
080
060
040
020
mdash
ALPL PPARg aP2 LPL
lowastlowastlowastlowastlowastlowast
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
lowast
(b)
Figure 5 Adipocyte related genes downregulated in ALP knockdown CL1 cells (a) Venn diagram of whole gene expression analysis of ADupregulated genes compared with ALP KD downregulated genes (b) Quantitative real-time PCR of four selected common genes from Venndiagram including ALPL PPARg2 aP2 and LPL Data are presented as fold change in expression of each target gene normalized to GAPDH(119899 = three independent experiments lowast119901 lt 005 lowastlowast119901 lt 001 and lowastlowastlowast119901 lt 0001)
Stem Cells International 17
cell populations has a clinical relevance It demonstrates thatthe progenitor function and the immune modulatory rolesof hBMSC [39] are mediated by specific and distinguishablepopulations of hBMSC Thus future clinical studies employ-ing hBMSC should attempt to administer the relevant sub-population of hBMSC dependent on the experimental aim asa novel approach to improving the clinical efficiency insteadof the current use of heterogeneous hBMSC populations
Competing Interests
The authors declare that they have no competing interests
Authorsrsquo Contributions
Mona Elsafadi and Amer Mahmood contributed equally tothis work
Acknowledgments
These studies are supported by a grant from the College ofMedicine Research Centre (CMRC) Deanship of ScientificResearch King Saud University Riyadh Saudi Arabia Theauthors would like to thank Mr Ali Al-Roalle for all histechnical help They thank Natalie Twine for advice on DNAmicroarray analysis
References
[1] W Zaher L Harkness A Jafari and M Kassem ldquoAn update ofhuman mesenchymal stem cell biology and their clinical usesrdquoArchives of Toxicology vol 88 no 5 pp 1069ndash1082 2014
[2] S Post B M Abdallah J F Bentzon andM Kassem ldquoDemon-stration of the presence of independent pre-osteoblastic andpre-adipocytic cell populations in bone marrow-derived mes-enchymal stem cellsrdquo Bone vol 43 no 1 pp 32ndash39 2008
[3] K H Larsen C M Frederiksen J S Burns B M Abdallahand M Kassem ldquoIdentifying a molecular phenotype for bonemarrow stromal cells with in vivo bone-forming capacityrdquoJournal of Bone and Mineral Research vol 25 no 4 pp 796ndash808 2010
[4] A Rosenzweig ldquoCardiac cell therapymdashmixed results frommixed cellsrdquoThe New England Journal of Medicine vol 355 no12 pp 1274ndash1277 2006
[5] S A Kuznetsov P H Krebsbach K Satomura et al ldquoSingle-colony derived strains of human marrow stromal fibroblastsform bone after transplantation in vivordquo Journal of Bone andMineral Research vol 12 no 9 pp 1335ndash1347 1997
[6] M Kassem and P Bianco ldquoSkeletal stem cells in space and timerdquoCell vol 160 no 1-2 pp 17ndash19 2015
[7] A I Caplan and D Correa ldquoPDGF in bone formation andregeneration new insights into a novel mechanism involvingMSCsrdquo Journal of Orthopaedic Research vol 29 no 12 pp 1795ndash1803 2011
[8] K Stewart S Walsh J Screen et al ldquoFurther characterizationof cells expressing STRO-1 in cultures of adult human bonemarrow stromal cellsrdquo Journal of Bone and Mineral Researchvol 14 no 8 pp 1345ndash1356 1999
[9] J L Simonsen C Rosada N Serakinci et al ldquoTelomeraseexpression extends the proliferative life-span and maintains
the osteogenic potential of human bone marrow stromal cellsrdquoNature Biotechnology vol 20 no 6 pp 592ndash596 2002
[10] M Al-Nbaheen R Vishnubalaji D Ali et al ldquoHuman stromal(mesenchymal) stem cells from bone marrow adipose tissueand skin exhibit differences in molecular phenotype and differ-entiation potentialrdquo Stem Cell Reviews and Reports vol 9 no 1pp 32ndash43 2013
[11] C A Gregory W G Gunn A Peister and D J Prockop ldquoAnAlizarin red-based assay of mineralization by adherent cells inculture comparison with cetylpyridinium chloride extractionrdquoAnalytical Biochemistry vol 329 no 1 pp 77ndash84 2004
[12] AMahmood L Harkness H D Schroslashder B M Abdallah andM Kassem ldquoEnhanced differentiation of human embryonicstem cells to mesenchymal progenitors by inhibition of TGF-120573activinnodal signaling using SB-431542rdquo Journal of Bone andMineral Research vol 25 no 6 pp 1216ndash1233 2010
[13] H-J Prins A K Braat D Gawlitta et al ldquoIn vitro inductionof alkaline phosphatase levels predicts in vivo bone formingcapacity of human bone marrow stromal cellsrdquo Stem CellResearch vol 12 no 2 pp 428ndash440 2014
[14] K Stenderup J Justesen C Clausen and M Kassem ldquoAgingis associated with decreased maximal life span and acceleratedsenescence of bone marrow stromal cellsrdquo Bone vol 33 no 6pp 919ndash926 2003
[15] D G Phinney G Kopen W Righter S Webster N Tremainand D J Prockop ldquoDonor variation in the growth propertiesand osteogenic potential of human marrow stromal cellsrdquoJournal of Cellular Biochemistry vol 75 no 3 pp 424ndash436 1999
[16] C M Digirolamo D Stokes D Colter D G Phinney R Classand D J Prockop ldquoPropagation and senescence of humanmarrow stromal cells in culture a simple colony-forming assayidentifies samples with the greatest potential to propagate anddifferentiaterdquo British Journal of Haematology vol 107 no 2 pp275ndash281 1999
[17] S James J Fox F Afsari et al ldquoMultiparameter analy-sis of human bone marrow stromal cells identifies distinctimmunomodulatory and differentiation-competent subtypesrdquoStem Cell Reports vol 4 no 6 pp 1004ndash1015 2015
[18] B Peruzzi A Cappariello A Del Fattore N Rucci F DeBenedetti and A Teti ldquoc-Src and IL-6 inhibit osteoblastdifferentiation and integrate IGFBP5 signallingrdquo Nature Com-munications vol 3 article 630 2012
[19] D Fortunati S Reppe A-K Fjeldheim M Nielsen V TGautvik and K M Gautvik ldquoPeriostin is a collagen associatedbonematrix protein regulated by parathyroid hormonerdquoMatrixBiology vol 29 no 7 pp 594ndash601 2010
[20] A Oshima H Tanabe T Yan G N Lowe C A Glackin andA Kudo ldquoA novel mechanism for the regulation of osteoblastdifferentiation transcription of periostin a member of thefasciclin I family is regulated by the bHLH transcription factorTwistrdquo Journal of Cellular Biochemistry vol 86 no 4 pp 792ndash804 2002
[21] T G Kashima T Nishiyama K Shimazu et al ldquoPeriostina novel marker of intramembranous ossification is expressedin fibrous dysplasia and in c-Fos-overexpressing bone lesionsrdquoHuman Pathology vol 40 no 2 pp 226ndash237 2009
[22] CGalliM PiergianniM Piemontese et al ldquoPeriostin improvescell adhesion to implantable biomaterials and osteoblasticdifferentiation on implant titanium surfaces in a topography-dependent fashionrdquo Journal of Biomedical Materials Research -Part A vol 102 no 11 pp 3855ndash3861 2014
18 Stem Cells International
[23] D L Coutu H W Jian A Monette G-E Rivard M DBlostein and J Galipeau ldquoPeriostin amember of a novel familyof vitamin K-dependent proteins is expressed by mesenchymalstromal cellsrdquo The Journal of Biological Chemistry vol 283 no26 pp 17991ndash18001 2008
[24] T Minamizato K Sakamoto T Liu et al ldquoCCN3NOV inhibitsBMP-2-induced osteoblast differentiation by interacting withBMP andNotch signaling pathwaysrdquo Biochemical and Biophysi-cal Research Communications vol 354 no 2 pp 567ndash573 2007
[25] W Si Q Kang H H Luu et al ldquoCCN1Cyr61 is regulated bythe canonicalWnt signal and plays an important role inWnt3A-induced osteoblast differentiation of mesenchymal stem cellsrdquoMolecular and Cellular Biology vol 26 no 8 pp 2955ndash29642006
[26] Q Luo Q Kang W Si et al ldquoConnective tissue growth factor(CTGF) is regulated by Wnt and bone morphogenetic proteinssignaling in osteoblast differentiation of mesenchymal stemcellsrdquo The Journal of Biological Chemistry vol 279 no 53 pp55958ndash55968 2004
[27] Y-S Yoon N Lee and H Scadova ldquoMyocardial regenerationwith bone-marrow-derived stem cellsrdquo Biology of the Cell vol97 no 4 pp 253ndash263 2005
[28] Y S Yoon ldquoCardiac regeneration with novel bone marrow-derived multipotent stem cellsrdquo Discovery Medicine vol 5 pp204ndash208 2005
[29] Y-S Yoon A Wecker L Heyd et al ldquoClonally expanded novelmultipotent stem cells from human bone marrow regeneratemyocardium after myocardial infarctionrdquo Journal of ClinicalInvestigation vol 115 no 2 pp 326ndash338 2005
[30] GM BolandG Perkins D J Hall andR S Tuan ldquoWnt 3a pro-motes proliferation and suppresses osteogenic differentiationof adult human mesenchymal stem cellsrdquo Journal of CellularBiochemistry vol 93 no 6 pp 1210ndash1230 2004
[31] D A Cook S W Fellgett M E Pownall P J OrsquoShea and PG Genever ldquoWnt-dependent osteogenic commitment of bonemarrow stromal cells using a novel GSK3120573 inhibitorrdquo Stem CellResearch vol 12 no 2 pp 415ndash427 2014
[32] J De Boer R Siddappa C Gaspar A VanApeldoorn R Foddeand C Van Blitterswijk ldquoWnt signaling inhibits osteogenicdifferentiation of humanmesenchymal stem cellsrdquoBone vol 34no 5 pp 818ndash826 2004
[33] MAl-Toub A AlmusaM Almajed et al ldquoPleiotropic effects ofcancer cellsrsquo secreted factors on human stromal (mesenchymal)stem cellsrdquo Stem Cell Research and Therapy vol 4 article 1142013
[34] L Chen D Qanie A Jafari et al ldquoDelta-like 1fetal antigen-1(Dlk1FA1) is a novel regulator of chondrogenic cell differenti-ation via inhibition of the Akt kinase-dependent pathwayrdquoTheJournal of Biological Chemistry vol 286 no 37 pp 32140ndash321492011
[35] T Matsunobu K Torigoe M Ishikawa et al ldquoCritical roles ofthe TGF-120573 type I receptor ALK5 in perichondrial formationand function cartilage integrity and osteoblast differentiationduring growth plate developmentrdquo Developmental Biology vol332 no 2 pp 325ndash338 2009
[36] B C Nair D E Johnson R J Majeska J A Rodkey C DBennett and G A Rodan ldquoRat alkaline phosphatase II Struc-tural similarities between the osteosarcoma bone kidney andplacenta isoenzymesrdquo Archives of Biochemistry and Biophysicsvol 254 no 1 pp 28ndash34 1987
[37] P J Butterworth ldquoAlkaline phosphatase Biochemistry of mam-malian alkaline phosphatasesrdquo Cell Biochemistry and Functionvol 1 no 2 pp 66ndash70 1983
[38] A T Ali C B Penny J E Paiker et al ldquoAlkaline phosphataseis involved in the control of adipogenesis in the murinepreadipocyte cell line 3T3-L1rdquo Clinica Chimica Acta vol 354no 1-2 pp 101ndash109 2005
[39] G M Spaggiari and L Moretta ldquoCellular and molecularinteractions of mesenchymal stem cells in innate immunityrdquoImmunology and Cell Biology vol 91 no 1 pp 27ndash31 2012
Figure 4 Effect of alkaline phosphatase (ALP) gene silencing by small interfering RNA (siRNA) on a clonal population of human bonemarrow stromal stem cell CL1 (A) ALP staining at day 3 day 5 or day 7 days of osteoblast differentiation (OB induction) (B) Quantitativereal-time PCR for ALP gene following ALP siRNA transfection at day 0 OB and day 7 of OB Data are presented as fold change in expressionof each target gene normalized to GAPDH (119899 = three independent experiments 119901 lt 001) Western blotting analysis of day 3 and day 7after siRNA ALP transfection of CL1 cells ALPL specific antibody and B-actin was used (C) Mineralized matrix formation as visualizedby Alizarin Red S staining in siRNA transfected CL1 cells after 14 days of OB induction (D) Nile red quantification of mature lipid filledadipocyte in control noninduced (Cont) adipocyte induced (Adip I) and ALP siRNA transfected cells that are adipocyte induced (ALPsiRNA) Adipocyte induction was carried out for 7 days lowast119901 lt 005
phenotype and cellular responses of hMSC-TERT are similarto those of primary hMSC [10]While we have identified ALPas a marker that can be used for a prospective identificationof differentiation committed population of hBMSC we iden-tified additional distinctive molecular markers of the cellsFor example IGF-1 IGF-2 and IGF binding protein 5 wereenriched in CL1 compared to CL2 IGFs and their bindingproteins are very well-studied factors that play a role inhBMSC proliferation and osteoblast differentiation [18] On
the other hand annexin A3 as well as several immune-relatedgenes was highly enriched in CL2 compared to CL1 Futurestudies are needed to determine the functional significanceof these molecules in relation to the functional identity ofvarious cell populations within the hBMSC cultures andtheir usefulness as biomarkers to dissect the heterogeneouspopulation of cultured hBMSC
Our finding of the presence of functional diversity withinhBMSC cultures that contain progenitor and nonprogenitor
Stem Cells International 15
Table 4 Microarray data analysis showing genes found to be upregulated during adipogenic differentiation and downregulated after ALPKO
62 common elements inldquoAD uprdquo and ldquoALP downrdquo Gene name FC (ALP siRNA versus
control siRNA)APOBEC3G Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3G minus15938025IFI44L Interferon-induced protein 44-like minus11520283PAQR5 Progestin and adipoQ receptor family member V minus6868241PNMA2 Paraneoplastic antigen MA2 minus59695344DUSP23 Dual specificity phosphatase 23 minus54786854CLDN23 Claudin 23 minus51885047ANKDD1A Ankyrin repeat and death domain containing 1A minus51646647IL8 Interleukin 8 minus4887188LRRC23 Leucine rich repeat containing 23 minus47611775IL6 Interleukin 6 (interferon beta 2) minus4693139LIFR Leukemia inhibitory factor receptor alpha minus46540866PTGFR Prostaglandin F receptor (FP) minus4457529FAM134B Family with sequence similarity 134 member B minus4403495CYFIP2 Cytoplasmic FMR1 interacting protein 2 minus4260462METTL7A Methyltransferase-like 7A minus40480843APOBEC3F Apolipoprotein B mRNA editing enzyme catalytic polypeptide-like 3F minus39502614CA5B Carbonic anhydrase VB mitochondrial minus393889ITGA10 Integrin alpha 10 minus39143775FMO3 Flavin containing monooxygenase 3 minus3852087IMPA2 Inositol monophosphatase 2 (human) minus38374884CDO1 Cysteine dioxygenase type I minus38181455CCDC68 Coiled-coil domain containing 68 minus37292893
HEXDC Hexosaminidase (glycosyl hydrolase family 20 catalytic domain) containing minus2094836SARM1 Sterile alpha and TIR motif containing 1 minus20797038BBS1 Bardet-Biedl syndrome 1 minus20146718SERPING1 Serpin peptidase inhibitor clade G (C1 inhibitor) member 1 minus20102212FAM162A Family with sequence similarity 162 member A minus2005807TCTN1 Tectonic family member 1 minus20033443
AD up ALP down
1006 62 1145
(a)
140
120
100
080
060
040
020
mdash
ALPL PPARg aP2 LPL
lowastlowastlowastlowastlowastlowast
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
lowast
(b)
Figure 5 Adipocyte related genes downregulated in ALP knockdown CL1 cells (a) Venn diagram of whole gene expression analysis of ADupregulated genes compared with ALP KD downregulated genes (b) Quantitative real-time PCR of four selected common genes from Venndiagram including ALPL PPARg2 aP2 and LPL Data are presented as fold change in expression of each target gene normalized to GAPDH(119899 = three independent experiments lowast119901 lt 005 lowastlowast119901 lt 001 and lowastlowastlowast119901 lt 0001)
Stem Cells International 17
cell populations has a clinical relevance It demonstrates thatthe progenitor function and the immune modulatory rolesof hBMSC [39] are mediated by specific and distinguishablepopulations of hBMSC Thus future clinical studies employ-ing hBMSC should attempt to administer the relevant sub-population of hBMSC dependent on the experimental aim asa novel approach to improving the clinical efficiency insteadof the current use of heterogeneous hBMSC populations
Competing Interests
The authors declare that they have no competing interests
Authorsrsquo Contributions
Mona Elsafadi and Amer Mahmood contributed equally tothis work
Acknowledgments
These studies are supported by a grant from the College ofMedicine Research Centre (CMRC) Deanship of ScientificResearch King Saud University Riyadh Saudi Arabia Theauthors would like to thank Mr Ali Al-Roalle for all histechnical help They thank Natalie Twine for advice on DNAmicroarray analysis
References
[1] W Zaher L Harkness A Jafari and M Kassem ldquoAn update ofhuman mesenchymal stem cell biology and their clinical usesrdquoArchives of Toxicology vol 88 no 5 pp 1069ndash1082 2014
[2] S Post B M Abdallah J F Bentzon andM Kassem ldquoDemon-stration of the presence of independent pre-osteoblastic andpre-adipocytic cell populations in bone marrow-derived mes-enchymal stem cellsrdquo Bone vol 43 no 1 pp 32ndash39 2008
[3] K H Larsen C M Frederiksen J S Burns B M Abdallahand M Kassem ldquoIdentifying a molecular phenotype for bonemarrow stromal cells with in vivo bone-forming capacityrdquoJournal of Bone and Mineral Research vol 25 no 4 pp 796ndash808 2010
[4] A Rosenzweig ldquoCardiac cell therapymdashmixed results frommixed cellsrdquoThe New England Journal of Medicine vol 355 no12 pp 1274ndash1277 2006
[5] S A Kuznetsov P H Krebsbach K Satomura et al ldquoSingle-colony derived strains of human marrow stromal fibroblastsform bone after transplantation in vivordquo Journal of Bone andMineral Research vol 12 no 9 pp 1335ndash1347 1997
[6] M Kassem and P Bianco ldquoSkeletal stem cells in space and timerdquoCell vol 160 no 1-2 pp 17ndash19 2015
[7] A I Caplan and D Correa ldquoPDGF in bone formation andregeneration new insights into a novel mechanism involvingMSCsrdquo Journal of Orthopaedic Research vol 29 no 12 pp 1795ndash1803 2011
[8] K Stewart S Walsh J Screen et al ldquoFurther characterizationof cells expressing STRO-1 in cultures of adult human bonemarrow stromal cellsrdquo Journal of Bone and Mineral Researchvol 14 no 8 pp 1345ndash1356 1999
[9] J L Simonsen C Rosada N Serakinci et al ldquoTelomeraseexpression extends the proliferative life-span and maintains
the osteogenic potential of human bone marrow stromal cellsrdquoNature Biotechnology vol 20 no 6 pp 592ndash596 2002
[10] M Al-Nbaheen R Vishnubalaji D Ali et al ldquoHuman stromal(mesenchymal) stem cells from bone marrow adipose tissueand skin exhibit differences in molecular phenotype and differ-entiation potentialrdquo Stem Cell Reviews and Reports vol 9 no 1pp 32ndash43 2013
[11] C A Gregory W G Gunn A Peister and D J Prockop ldquoAnAlizarin red-based assay of mineralization by adherent cells inculture comparison with cetylpyridinium chloride extractionrdquoAnalytical Biochemistry vol 329 no 1 pp 77ndash84 2004
[12] AMahmood L Harkness H D Schroslashder B M Abdallah andM Kassem ldquoEnhanced differentiation of human embryonicstem cells to mesenchymal progenitors by inhibition of TGF-120573activinnodal signaling using SB-431542rdquo Journal of Bone andMineral Research vol 25 no 6 pp 1216ndash1233 2010
[13] H-J Prins A K Braat D Gawlitta et al ldquoIn vitro inductionof alkaline phosphatase levels predicts in vivo bone formingcapacity of human bone marrow stromal cellsrdquo Stem CellResearch vol 12 no 2 pp 428ndash440 2014
[14] K Stenderup J Justesen C Clausen and M Kassem ldquoAgingis associated with decreased maximal life span and acceleratedsenescence of bone marrow stromal cellsrdquo Bone vol 33 no 6pp 919ndash926 2003
[15] D G Phinney G Kopen W Righter S Webster N Tremainand D J Prockop ldquoDonor variation in the growth propertiesand osteogenic potential of human marrow stromal cellsrdquoJournal of Cellular Biochemistry vol 75 no 3 pp 424ndash436 1999
[16] C M Digirolamo D Stokes D Colter D G Phinney R Classand D J Prockop ldquoPropagation and senescence of humanmarrow stromal cells in culture a simple colony-forming assayidentifies samples with the greatest potential to propagate anddifferentiaterdquo British Journal of Haematology vol 107 no 2 pp275ndash281 1999
[17] S James J Fox F Afsari et al ldquoMultiparameter analy-sis of human bone marrow stromal cells identifies distinctimmunomodulatory and differentiation-competent subtypesrdquoStem Cell Reports vol 4 no 6 pp 1004ndash1015 2015
[18] B Peruzzi A Cappariello A Del Fattore N Rucci F DeBenedetti and A Teti ldquoc-Src and IL-6 inhibit osteoblastdifferentiation and integrate IGFBP5 signallingrdquo Nature Com-munications vol 3 article 630 2012
[19] D Fortunati S Reppe A-K Fjeldheim M Nielsen V TGautvik and K M Gautvik ldquoPeriostin is a collagen associatedbonematrix protein regulated by parathyroid hormonerdquoMatrixBiology vol 29 no 7 pp 594ndash601 2010
[20] A Oshima H Tanabe T Yan G N Lowe C A Glackin andA Kudo ldquoA novel mechanism for the regulation of osteoblastdifferentiation transcription of periostin a member of thefasciclin I family is regulated by the bHLH transcription factorTwistrdquo Journal of Cellular Biochemistry vol 86 no 4 pp 792ndash804 2002
[21] T G Kashima T Nishiyama K Shimazu et al ldquoPeriostina novel marker of intramembranous ossification is expressedin fibrous dysplasia and in c-Fos-overexpressing bone lesionsrdquoHuman Pathology vol 40 no 2 pp 226ndash237 2009
[22] CGalliM PiergianniM Piemontese et al ldquoPeriostin improvescell adhesion to implantable biomaterials and osteoblasticdifferentiation on implant titanium surfaces in a topography-dependent fashionrdquo Journal of Biomedical Materials Research -Part A vol 102 no 11 pp 3855ndash3861 2014
18 Stem Cells International
[23] D L Coutu H W Jian A Monette G-E Rivard M DBlostein and J Galipeau ldquoPeriostin amember of a novel familyof vitamin K-dependent proteins is expressed by mesenchymalstromal cellsrdquo The Journal of Biological Chemistry vol 283 no26 pp 17991ndash18001 2008
[24] T Minamizato K Sakamoto T Liu et al ldquoCCN3NOV inhibitsBMP-2-induced osteoblast differentiation by interacting withBMP andNotch signaling pathwaysrdquo Biochemical and Biophysi-cal Research Communications vol 354 no 2 pp 567ndash573 2007
[25] W Si Q Kang H H Luu et al ldquoCCN1Cyr61 is regulated bythe canonicalWnt signal and plays an important role inWnt3A-induced osteoblast differentiation of mesenchymal stem cellsrdquoMolecular and Cellular Biology vol 26 no 8 pp 2955ndash29642006
[26] Q Luo Q Kang W Si et al ldquoConnective tissue growth factor(CTGF) is regulated by Wnt and bone morphogenetic proteinssignaling in osteoblast differentiation of mesenchymal stemcellsrdquo The Journal of Biological Chemistry vol 279 no 53 pp55958ndash55968 2004
[27] Y-S Yoon N Lee and H Scadova ldquoMyocardial regenerationwith bone-marrow-derived stem cellsrdquo Biology of the Cell vol97 no 4 pp 253ndash263 2005
[28] Y S Yoon ldquoCardiac regeneration with novel bone marrow-derived multipotent stem cellsrdquo Discovery Medicine vol 5 pp204ndash208 2005
[29] Y-S Yoon A Wecker L Heyd et al ldquoClonally expanded novelmultipotent stem cells from human bone marrow regeneratemyocardium after myocardial infarctionrdquo Journal of ClinicalInvestigation vol 115 no 2 pp 326ndash338 2005
[30] GM BolandG Perkins D J Hall andR S Tuan ldquoWnt 3a pro-motes proliferation and suppresses osteogenic differentiationof adult human mesenchymal stem cellsrdquo Journal of CellularBiochemistry vol 93 no 6 pp 1210ndash1230 2004
[31] D A Cook S W Fellgett M E Pownall P J OrsquoShea and PG Genever ldquoWnt-dependent osteogenic commitment of bonemarrow stromal cells using a novel GSK3120573 inhibitorrdquo Stem CellResearch vol 12 no 2 pp 415ndash427 2014
[32] J De Boer R Siddappa C Gaspar A VanApeldoorn R Foddeand C Van Blitterswijk ldquoWnt signaling inhibits osteogenicdifferentiation of humanmesenchymal stem cellsrdquoBone vol 34no 5 pp 818ndash826 2004
[33] MAl-Toub A AlmusaM Almajed et al ldquoPleiotropic effects ofcancer cellsrsquo secreted factors on human stromal (mesenchymal)stem cellsrdquo Stem Cell Research and Therapy vol 4 article 1142013
[34] L Chen D Qanie A Jafari et al ldquoDelta-like 1fetal antigen-1(Dlk1FA1) is a novel regulator of chondrogenic cell differenti-ation via inhibition of the Akt kinase-dependent pathwayrdquoTheJournal of Biological Chemistry vol 286 no 37 pp 32140ndash321492011
[35] T Matsunobu K Torigoe M Ishikawa et al ldquoCritical roles ofthe TGF-120573 type I receptor ALK5 in perichondrial formationand function cartilage integrity and osteoblast differentiationduring growth plate developmentrdquo Developmental Biology vol332 no 2 pp 325ndash338 2009
[36] B C Nair D E Johnson R J Majeska J A Rodkey C DBennett and G A Rodan ldquoRat alkaline phosphatase II Struc-tural similarities between the osteosarcoma bone kidney andplacenta isoenzymesrdquo Archives of Biochemistry and Biophysicsvol 254 no 1 pp 28ndash34 1987
[37] P J Butterworth ldquoAlkaline phosphatase Biochemistry of mam-malian alkaline phosphatasesrdquo Cell Biochemistry and Functionvol 1 no 2 pp 66ndash70 1983
[38] A T Ali C B Penny J E Paiker et al ldquoAlkaline phosphataseis involved in the control of adipogenesis in the murinepreadipocyte cell line 3T3-L1rdquo Clinica Chimica Acta vol 354no 1-2 pp 101ndash109 2005
[39] G M Spaggiari and L Moretta ldquoCellular and molecularinteractions of mesenchymal stem cells in innate immunityrdquoImmunology and Cell Biology vol 91 no 1 pp 27ndash31 2012
HEXDC Hexosaminidase (glycosyl hydrolase family 20 catalytic domain) containing minus2094836SARM1 Sterile alpha and TIR motif containing 1 minus20797038BBS1 Bardet-Biedl syndrome 1 minus20146718SERPING1 Serpin peptidase inhibitor clade G (C1 inhibitor) member 1 minus20102212FAM162A Family with sequence similarity 162 member A minus2005807TCTN1 Tectonic family member 1 minus20033443
AD up ALP down
1006 62 1145
(a)
140
120
100
080
060
040
020
mdash
ALPL PPARg aP2 LPL
lowastlowastlowastlowastlowastlowast
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
lowast
(b)
Figure 5 Adipocyte related genes downregulated in ALP knockdown CL1 cells (a) Venn diagram of whole gene expression analysis of ADupregulated genes compared with ALP KD downregulated genes (b) Quantitative real-time PCR of four selected common genes from Venndiagram including ALPL PPARg2 aP2 and LPL Data are presented as fold change in expression of each target gene normalized to GAPDH(119899 = three independent experiments lowast119901 lt 005 lowastlowast119901 lt 001 and lowastlowastlowast119901 lt 0001)
Stem Cells International 17
cell populations has a clinical relevance It demonstrates thatthe progenitor function and the immune modulatory rolesof hBMSC [39] are mediated by specific and distinguishablepopulations of hBMSC Thus future clinical studies employ-ing hBMSC should attempt to administer the relevant sub-population of hBMSC dependent on the experimental aim asa novel approach to improving the clinical efficiency insteadof the current use of heterogeneous hBMSC populations
Competing Interests
The authors declare that they have no competing interests
Authorsrsquo Contributions
Mona Elsafadi and Amer Mahmood contributed equally tothis work
Acknowledgments
These studies are supported by a grant from the College ofMedicine Research Centre (CMRC) Deanship of ScientificResearch King Saud University Riyadh Saudi Arabia Theauthors would like to thank Mr Ali Al-Roalle for all histechnical help They thank Natalie Twine for advice on DNAmicroarray analysis
References
[1] W Zaher L Harkness A Jafari and M Kassem ldquoAn update ofhuman mesenchymal stem cell biology and their clinical usesrdquoArchives of Toxicology vol 88 no 5 pp 1069ndash1082 2014
[2] S Post B M Abdallah J F Bentzon andM Kassem ldquoDemon-stration of the presence of independent pre-osteoblastic andpre-adipocytic cell populations in bone marrow-derived mes-enchymal stem cellsrdquo Bone vol 43 no 1 pp 32ndash39 2008
[3] K H Larsen C M Frederiksen J S Burns B M Abdallahand M Kassem ldquoIdentifying a molecular phenotype for bonemarrow stromal cells with in vivo bone-forming capacityrdquoJournal of Bone and Mineral Research vol 25 no 4 pp 796ndash808 2010
[4] A Rosenzweig ldquoCardiac cell therapymdashmixed results frommixed cellsrdquoThe New England Journal of Medicine vol 355 no12 pp 1274ndash1277 2006
[5] S A Kuznetsov P H Krebsbach K Satomura et al ldquoSingle-colony derived strains of human marrow stromal fibroblastsform bone after transplantation in vivordquo Journal of Bone andMineral Research vol 12 no 9 pp 1335ndash1347 1997
[6] M Kassem and P Bianco ldquoSkeletal stem cells in space and timerdquoCell vol 160 no 1-2 pp 17ndash19 2015
[7] A I Caplan and D Correa ldquoPDGF in bone formation andregeneration new insights into a novel mechanism involvingMSCsrdquo Journal of Orthopaedic Research vol 29 no 12 pp 1795ndash1803 2011
[8] K Stewart S Walsh J Screen et al ldquoFurther characterizationof cells expressing STRO-1 in cultures of adult human bonemarrow stromal cellsrdquo Journal of Bone and Mineral Researchvol 14 no 8 pp 1345ndash1356 1999
[9] J L Simonsen C Rosada N Serakinci et al ldquoTelomeraseexpression extends the proliferative life-span and maintains
the osteogenic potential of human bone marrow stromal cellsrdquoNature Biotechnology vol 20 no 6 pp 592ndash596 2002
[10] M Al-Nbaheen R Vishnubalaji D Ali et al ldquoHuman stromal(mesenchymal) stem cells from bone marrow adipose tissueand skin exhibit differences in molecular phenotype and differ-entiation potentialrdquo Stem Cell Reviews and Reports vol 9 no 1pp 32ndash43 2013
[11] C A Gregory W G Gunn A Peister and D J Prockop ldquoAnAlizarin red-based assay of mineralization by adherent cells inculture comparison with cetylpyridinium chloride extractionrdquoAnalytical Biochemistry vol 329 no 1 pp 77ndash84 2004
[12] AMahmood L Harkness H D Schroslashder B M Abdallah andM Kassem ldquoEnhanced differentiation of human embryonicstem cells to mesenchymal progenitors by inhibition of TGF-120573activinnodal signaling using SB-431542rdquo Journal of Bone andMineral Research vol 25 no 6 pp 1216ndash1233 2010
[13] H-J Prins A K Braat D Gawlitta et al ldquoIn vitro inductionof alkaline phosphatase levels predicts in vivo bone formingcapacity of human bone marrow stromal cellsrdquo Stem CellResearch vol 12 no 2 pp 428ndash440 2014
[14] K Stenderup J Justesen C Clausen and M Kassem ldquoAgingis associated with decreased maximal life span and acceleratedsenescence of bone marrow stromal cellsrdquo Bone vol 33 no 6pp 919ndash926 2003
[15] D G Phinney G Kopen W Righter S Webster N Tremainand D J Prockop ldquoDonor variation in the growth propertiesand osteogenic potential of human marrow stromal cellsrdquoJournal of Cellular Biochemistry vol 75 no 3 pp 424ndash436 1999
[16] C M Digirolamo D Stokes D Colter D G Phinney R Classand D J Prockop ldquoPropagation and senescence of humanmarrow stromal cells in culture a simple colony-forming assayidentifies samples with the greatest potential to propagate anddifferentiaterdquo British Journal of Haematology vol 107 no 2 pp275ndash281 1999
[17] S James J Fox F Afsari et al ldquoMultiparameter analy-sis of human bone marrow stromal cells identifies distinctimmunomodulatory and differentiation-competent subtypesrdquoStem Cell Reports vol 4 no 6 pp 1004ndash1015 2015
[18] B Peruzzi A Cappariello A Del Fattore N Rucci F DeBenedetti and A Teti ldquoc-Src and IL-6 inhibit osteoblastdifferentiation and integrate IGFBP5 signallingrdquo Nature Com-munications vol 3 article 630 2012
[19] D Fortunati S Reppe A-K Fjeldheim M Nielsen V TGautvik and K M Gautvik ldquoPeriostin is a collagen associatedbonematrix protein regulated by parathyroid hormonerdquoMatrixBiology vol 29 no 7 pp 594ndash601 2010
[20] A Oshima H Tanabe T Yan G N Lowe C A Glackin andA Kudo ldquoA novel mechanism for the regulation of osteoblastdifferentiation transcription of periostin a member of thefasciclin I family is regulated by the bHLH transcription factorTwistrdquo Journal of Cellular Biochemistry vol 86 no 4 pp 792ndash804 2002
[21] T G Kashima T Nishiyama K Shimazu et al ldquoPeriostina novel marker of intramembranous ossification is expressedin fibrous dysplasia and in c-Fos-overexpressing bone lesionsrdquoHuman Pathology vol 40 no 2 pp 226ndash237 2009
[22] CGalliM PiergianniM Piemontese et al ldquoPeriostin improvescell adhesion to implantable biomaterials and osteoblasticdifferentiation on implant titanium surfaces in a topography-dependent fashionrdquo Journal of Biomedical Materials Research -Part A vol 102 no 11 pp 3855ndash3861 2014
18 Stem Cells International
[23] D L Coutu H W Jian A Monette G-E Rivard M DBlostein and J Galipeau ldquoPeriostin amember of a novel familyof vitamin K-dependent proteins is expressed by mesenchymalstromal cellsrdquo The Journal of Biological Chemistry vol 283 no26 pp 17991ndash18001 2008
[24] T Minamizato K Sakamoto T Liu et al ldquoCCN3NOV inhibitsBMP-2-induced osteoblast differentiation by interacting withBMP andNotch signaling pathwaysrdquo Biochemical and Biophysi-cal Research Communications vol 354 no 2 pp 567ndash573 2007
[25] W Si Q Kang H H Luu et al ldquoCCN1Cyr61 is regulated bythe canonicalWnt signal and plays an important role inWnt3A-induced osteoblast differentiation of mesenchymal stem cellsrdquoMolecular and Cellular Biology vol 26 no 8 pp 2955ndash29642006
[26] Q Luo Q Kang W Si et al ldquoConnective tissue growth factor(CTGF) is regulated by Wnt and bone morphogenetic proteinssignaling in osteoblast differentiation of mesenchymal stemcellsrdquo The Journal of Biological Chemistry vol 279 no 53 pp55958ndash55968 2004
[27] Y-S Yoon N Lee and H Scadova ldquoMyocardial regenerationwith bone-marrow-derived stem cellsrdquo Biology of the Cell vol97 no 4 pp 253ndash263 2005
[28] Y S Yoon ldquoCardiac regeneration with novel bone marrow-derived multipotent stem cellsrdquo Discovery Medicine vol 5 pp204ndash208 2005
[29] Y-S Yoon A Wecker L Heyd et al ldquoClonally expanded novelmultipotent stem cells from human bone marrow regeneratemyocardium after myocardial infarctionrdquo Journal of ClinicalInvestigation vol 115 no 2 pp 326ndash338 2005
[30] GM BolandG Perkins D J Hall andR S Tuan ldquoWnt 3a pro-motes proliferation and suppresses osteogenic differentiationof adult human mesenchymal stem cellsrdquo Journal of CellularBiochemistry vol 93 no 6 pp 1210ndash1230 2004
[31] D A Cook S W Fellgett M E Pownall P J OrsquoShea and PG Genever ldquoWnt-dependent osteogenic commitment of bonemarrow stromal cells using a novel GSK3120573 inhibitorrdquo Stem CellResearch vol 12 no 2 pp 415ndash427 2014
[32] J De Boer R Siddappa C Gaspar A VanApeldoorn R Foddeand C Van Blitterswijk ldquoWnt signaling inhibits osteogenicdifferentiation of humanmesenchymal stem cellsrdquoBone vol 34no 5 pp 818ndash826 2004
[33] MAl-Toub A AlmusaM Almajed et al ldquoPleiotropic effects ofcancer cellsrsquo secreted factors on human stromal (mesenchymal)stem cellsrdquo Stem Cell Research and Therapy vol 4 article 1142013
[34] L Chen D Qanie A Jafari et al ldquoDelta-like 1fetal antigen-1(Dlk1FA1) is a novel regulator of chondrogenic cell differenti-ation via inhibition of the Akt kinase-dependent pathwayrdquoTheJournal of Biological Chemistry vol 286 no 37 pp 32140ndash321492011
[35] T Matsunobu K Torigoe M Ishikawa et al ldquoCritical roles ofthe TGF-120573 type I receptor ALK5 in perichondrial formationand function cartilage integrity and osteoblast differentiationduring growth plate developmentrdquo Developmental Biology vol332 no 2 pp 325ndash338 2009
[36] B C Nair D E Johnson R J Majeska J A Rodkey C DBennett and G A Rodan ldquoRat alkaline phosphatase II Struc-tural similarities between the osteosarcoma bone kidney andplacenta isoenzymesrdquo Archives of Biochemistry and Biophysicsvol 254 no 1 pp 28ndash34 1987
[37] P J Butterworth ldquoAlkaline phosphatase Biochemistry of mam-malian alkaline phosphatasesrdquo Cell Biochemistry and Functionvol 1 no 2 pp 66ndash70 1983
[38] A T Ali C B Penny J E Paiker et al ldquoAlkaline phosphataseis involved in the control of adipogenesis in the murinepreadipocyte cell line 3T3-L1rdquo Clinica Chimica Acta vol 354no 1-2 pp 101ndash109 2005
[39] G M Spaggiari and L Moretta ldquoCellular and molecularinteractions of mesenchymal stem cells in innate immunityrdquoImmunology and Cell Biology vol 91 no 1 pp 27ndash31 2012
HEXDC Hexosaminidase (glycosyl hydrolase family 20 catalytic domain) containing minus2094836SARM1 Sterile alpha and TIR motif containing 1 minus20797038BBS1 Bardet-Biedl syndrome 1 minus20146718SERPING1 Serpin peptidase inhibitor clade G (C1 inhibitor) member 1 minus20102212FAM162A Family with sequence similarity 162 member A minus2005807TCTN1 Tectonic family member 1 minus20033443
AD up ALP down
1006 62 1145
(a)
140
120
100
080
060
040
020
mdash
ALPL PPARg aP2 LPL
lowastlowastlowastlowastlowastlowast
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
SCR
siRN
AA
LPL
siRN
A
lowast
(b)
Figure 5 Adipocyte related genes downregulated in ALP knockdown CL1 cells (a) Venn diagram of whole gene expression analysis of ADupregulated genes compared with ALP KD downregulated genes (b) Quantitative real-time PCR of four selected common genes from Venndiagram including ALPL PPARg2 aP2 and LPL Data are presented as fold change in expression of each target gene normalized to GAPDH(119899 = three independent experiments lowast119901 lt 005 lowastlowast119901 lt 001 and lowastlowastlowast119901 lt 0001)
Stem Cells International 17
cell populations has a clinical relevance It demonstrates thatthe progenitor function and the immune modulatory rolesof hBMSC [39] are mediated by specific and distinguishablepopulations of hBMSC Thus future clinical studies employ-ing hBMSC should attempt to administer the relevant sub-population of hBMSC dependent on the experimental aim asa novel approach to improving the clinical efficiency insteadof the current use of heterogeneous hBMSC populations
Competing Interests
The authors declare that they have no competing interests
Authorsrsquo Contributions
Mona Elsafadi and Amer Mahmood contributed equally tothis work
Acknowledgments
These studies are supported by a grant from the College ofMedicine Research Centre (CMRC) Deanship of ScientificResearch King Saud University Riyadh Saudi Arabia Theauthors would like to thank Mr Ali Al-Roalle for all histechnical help They thank Natalie Twine for advice on DNAmicroarray analysis
References
[1] W Zaher L Harkness A Jafari and M Kassem ldquoAn update ofhuman mesenchymal stem cell biology and their clinical usesrdquoArchives of Toxicology vol 88 no 5 pp 1069ndash1082 2014
[2] S Post B M Abdallah J F Bentzon andM Kassem ldquoDemon-stration of the presence of independent pre-osteoblastic andpre-adipocytic cell populations in bone marrow-derived mes-enchymal stem cellsrdquo Bone vol 43 no 1 pp 32ndash39 2008
[3] K H Larsen C M Frederiksen J S Burns B M Abdallahand M Kassem ldquoIdentifying a molecular phenotype for bonemarrow stromal cells with in vivo bone-forming capacityrdquoJournal of Bone and Mineral Research vol 25 no 4 pp 796ndash808 2010
[4] A Rosenzweig ldquoCardiac cell therapymdashmixed results frommixed cellsrdquoThe New England Journal of Medicine vol 355 no12 pp 1274ndash1277 2006
[5] S A Kuznetsov P H Krebsbach K Satomura et al ldquoSingle-colony derived strains of human marrow stromal fibroblastsform bone after transplantation in vivordquo Journal of Bone andMineral Research vol 12 no 9 pp 1335ndash1347 1997
[6] M Kassem and P Bianco ldquoSkeletal stem cells in space and timerdquoCell vol 160 no 1-2 pp 17ndash19 2015
[7] A I Caplan and D Correa ldquoPDGF in bone formation andregeneration new insights into a novel mechanism involvingMSCsrdquo Journal of Orthopaedic Research vol 29 no 12 pp 1795ndash1803 2011
[8] K Stewart S Walsh J Screen et al ldquoFurther characterizationof cells expressing STRO-1 in cultures of adult human bonemarrow stromal cellsrdquo Journal of Bone and Mineral Researchvol 14 no 8 pp 1345ndash1356 1999
[9] J L Simonsen C Rosada N Serakinci et al ldquoTelomeraseexpression extends the proliferative life-span and maintains
the osteogenic potential of human bone marrow stromal cellsrdquoNature Biotechnology vol 20 no 6 pp 592ndash596 2002
[10] M Al-Nbaheen R Vishnubalaji D Ali et al ldquoHuman stromal(mesenchymal) stem cells from bone marrow adipose tissueand skin exhibit differences in molecular phenotype and differ-entiation potentialrdquo Stem Cell Reviews and Reports vol 9 no 1pp 32ndash43 2013
[11] C A Gregory W G Gunn A Peister and D J Prockop ldquoAnAlizarin red-based assay of mineralization by adherent cells inculture comparison with cetylpyridinium chloride extractionrdquoAnalytical Biochemistry vol 329 no 1 pp 77ndash84 2004
[12] AMahmood L Harkness H D Schroslashder B M Abdallah andM Kassem ldquoEnhanced differentiation of human embryonicstem cells to mesenchymal progenitors by inhibition of TGF-120573activinnodal signaling using SB-431542rdquo Journal of Bone andMineral Research vol 25 no 6 pp 1216ndash1233 2010
[13] H-J Prins A K Braat D Gawlitta et al ldquoIn vitro inductionof alkaline phosphatase levels predicts in vivo bone formingcapacity of human bone marrow stromal cellsrdquo Stem CellResearch vol 12 no 2 pp 428ndash440 2014
[14] K Stenderup J Justesen C Clausen and M Kassem ldquoAgingis associated with decreased maximal life span and acceleratedsenescence of bone marrow stromal cellsrdquo Bone vol 33 no 6pp 919ndash926 2003
[15] D G Phinney G Kopen W Righter S Webster N Tremainand D J Prockop ldquoDonor variation in the growth propertiesand osteogenic potential of human marrow stromal cellsrdquoJournal of Cellular Biochemistry vol 75 no 3 pp 424ndash436 1999
[16] C M Digirolamo D Stokes D Colter D G Phinney R Classand D J Prockop ldquoPropagation and senescence of humanmarrow stromal cells in culture a simple colony-forming assayidentifies samples with the greatest potential to propagate anddifferentiaterdquo British Journal of Haematology vol 107 no 2 pp275ndash281 1999
[17] S James J Fox F Afsari et al ldquoMultiparameter analy-sis of human bone marrow stromal cells identifies distinctimmunomodulatory and differentiation-competent subtypesrdquoStem Cell Reports vol 4 no 6 pp 1004ndash1015 2015
[18] B Peruzzi A Cappariello A Del Fattore N Rucci F DeBenedetti and A Teti ldquoc-Src and IL-6 inhibit osteoblastdifferentiation and integrate IGFBP5 signallingrdquo Nature Com-munications vol 3 article 630 2012
[19] D Fortunati S Reppe A-K Fjeldheim M Nielsen V TGautvik and K M Gautvik ldquoPeriostin is a collagen associatedbonematrix protein regulated by parathyroid hormonerdquoMatrixBiology vol 29 no 7 pp 594ndash601 2010
[20] A Oshima H Tanabe T Yan G N Lowe C A Glackin andA Kudo ldquoA novel mechanism for the regulation of osteoblastdifferentiation transcription of periostin a member of thefasciclin I family is regulated by the bHLH transcription factorTwistrdquo Journal of Cellular Biochemistry vol 86 no 4 pp 792ndash804 2002
[21] T G Kashima T Nishiyama K Shimazu et al ldquoPeriostina novel marker of intramembranous ossification is expressedin fibrous dysplasia and in c-Fos-overexpressing bone lesionsrdquoHuman Pathology vol 40 no 2 pp 226ndash237 2009
[22] CGalliM PiergianniM Piemontese et al ldquoPeriostin improvescell adhesion to implantable biomaterials and osteoblasticdifferentiation on implant titanium surfaces in a topography-dependent fashionrdquo Journal of Biomedical Materials Research -Part A vol 102 no 11 pp 3855ndash3861 2014
18 Stem Cells International
[23] D L Coutu H W Jian A Monette G-E Rivard M DBlostein and J Galipeau ldquoPeriostin amember of a novel familyof vitamin K-dependent proteins is expressed by mesenchymalstromal cellsrdquo The Journal of Biological Chemistry vol 283 no26 pp 17991ndash18001 2008
[24] T Minamizato K Sakamoto T Liu et al ldquoCCN3NOV inhibitsBMP-2-induced osteoblast differentiation by interacting withBMP andNotch signaling pathwaysrdquo Biochemical and Biophysi-cal Research Communications vol 354 no 2 pp 567ndash573 2007
[25] W Si Q Kang H H Luu et al ldquoCCN1Cyr61 is regulated bythe canonicalWnt signal and plays an important role inWnt3A-induced osteoblast differentiation of mesenchymal stem cellsrdquoMolecular and Cellular Biology vol 26 no 8 pp 2955ndash29642006
[26] Q Luo Q Kang W Si et al ldquoConnective tissue growth factor(CTGF) is regulated by Wnt and bone morphogenetic proteinssignaling in osteoblast differentiation of mesenchymal stemcellsrdquo The Journal of Biological Chemistry vol 279 no 53 pp55958ndash55968 2004
[27] Y-S Yoon N Lee and H Scadova ldquoMyocardial regenerationwith bone-marrow-derived stem cellsrdquo Biology of the Cell vol97 no 4 pp 253ndash263 2005
[28] Y S Yoon ldquoCardiac regeneration with novel bone marrow-derived multipotent stem cellsrdquo Discovery Medicine vol 5 pp204ndash208 2005
[29] Y-S Yoon A Wecker L Heyd et al ldquoClonally expanded novelmultipotent stem cells from human bone marrow regeneratemyocardium after myocardial infarctionrdquo Journal of ClinicalInvestigation vol 115 no 2 pp 326ndash338 2005
[30] GM BolandG Perkins D J Hall andR S Tuan ldquoWnt 3a pro-motes proliferation and suppresses osteogenic differentiationof adult human mesenchymal stem cellsrdquo Journal of CellularBiochemistry vol 93 no 6 pp 1210ndash1230 2004
[31] D A Cook S W Fellgett M E Pownall P J OrsquoShea and PG Genever ldquoWnt-dependent osteogenic commitment of bonemarrow stromal cells using a novel GSK3120573 inhibitorrdquo Stem CellResearch vol 12 no 2 pp 415ndash427 2014
[32] J De Boer R Siddappa C Gaspar A VanApeldoorn R Foddeand C Van Blitterswijk ldquoWnt signaling inhibits osteogenicdifferentiation of humanmesenchymal stem cellsrdquoBone vol 34no 5 pp 818ndash826 2004
[33] MAl-Toub A AlmusaM Almajed et al ldquoPleiotropic effects ofcancer cellsrsquo secreted factors on human stromal (mesenchymal)stem cellsrdquo Stem Cell Research and Therapy vol 4 article 1142013
[34] L Chen D Qanie A Jafari et al ldquoDelta-like 1fetal antigen-1(Dlk1FA1) is a novel regulator of chondrogenic cell differenti-ation via inhibition of the Akt kinase-dependent pathwayrdquoTheJournal of Biological Chemistry vol 286 no 37 pp 32140ndash321492011
[35] T Matsunobu K Torigoe M Ishikawa et al ldquoCritical roles ofthe TGF-120573 type I receptor ALK5 in perichondrial formationand function cartilage integrity and osteoblast differentiationduring growth plate developmentrdquo Developmental Biology vol332 no 2 pp 325ndash338 2009
[36] B C Nair D E Johnson R J Majeska J A Rodkey C DBennett and G A Rodan ldquoRat alkaline phosphatase II Struc-tural similarities between the osteosarcoma bone kidney andplacenta isoenzymesrdquo Archives of Biochemistry and Biophysicsvol 254 no 1 pp 28ndash34 1987
[37] P J Butterworth ldquoAlkaline phosphatase Biochemistry of mam-malian alkaline phosphatasesrdquo Cell Biochemistry and Functionvol 1 no 2 pp 66ndash70 1983
[38] A T Ali C B Penny J E Paiker et al ldquoAlkaline phosphataseis involved in the control of adipogenesis in the murinepreadipocyte cell line 3T3-L1rdquo Clinica Chimica Acta vol 354no 1-2 pp 101ndash109 2005
[39] G M Spaggiari and L Moretta ldquoCellular and molecularinteractions of mesenchymal stem cells in innate immunityrdquoImmunology and Cell Biology vol 91 no 1 pp 27ndash31 2012
cell populations has a clinical relevance It demonstrates thatthe progenitor function and the immune modulatory rolesof hBMSC [39] are mediated by specific and distinguishablepopulations of hBMSC Thus future clinical studies employ-ing hBMSC should attempt to administer the relevant sub-population of hBMSC dependent on the experimental aim asa novel approach to improving the clinical efficiency insteadof the current use of heterogeneous hBMSC populations
Competing Interests
The authors declare that they have no competing interests
Authorsrsquo Contributions
Mona Elsafadi and Amer Mahmood contributed equally tothis work
Acknowledgments
These studies are supported by a grant from the College ofMedicine Research Centre (CMRC) Deanship of ScientificResearch King Saud University Riyadh Saudi Arabia Theauthors would like to thank Mr Ali Al-Roalle for all histechnical help They thank Natalie Twine for advice on DNAmicroarray analysis
References
[1] W Zaher L Harkness A Jafari and M Kassem ldquoAn update ofhuman mesenchymal stem cell biology and their clinical usesrdquoArchives of Toxicology vol 88 no 5 pp 1069ndash1082 2014
[2] S Post B M Abdallah J F Bentzon andM Kassem ldquoDemon-stration of the presence of independent pre-osteoblastic andpre-adipocytic cell populations in bone marrow-derived mes-enchymal stem cellsrdquo Bone vol 43 no 1 pp 32ndash39 2008
[3] K H Larsen C M Frederiksen J S Burns B M Abdallahand M Kassem ldquoIdentifying a molecular phenotype for bonemarrow stromal cells with in vivo bone-forming capacityrdquoJournal of Bone and Mineral Research vol 25 no 4 pp 796ndash808 2010
[4] A Rosenzweig ldquoCardiac cell therapymdashmixed results frommixed cellsrdquoThe New England Journal of Medicine vol 355 no12 pp 1274ndash1277 2006
[5] S A Kuznetsov P H Krebsbach K Satomura et al ldquoSingle-colony derived strains of human marrow stromal fibroblastsform bone after transplantation in vivordquo Journal of Bone andMineral Research vol 12 no 9 pp 1335ndash1347 1997
[6] M Kassem and P Bianco ldquoSkeletal stem cells in space and timerdquoCell vol 160 no 1-2 pp 17ndash19 2015
[7] A I Caplan and D Correa ldquoPDGF in bone formation andregeneration new insights into a novel mechanism involvingMSCsrdquo Journal of Orthopaedic Research vol 29 no 12 pp 1795ndash1803 2011
[8] K Stewart S Walsh J Screen et al ldquoFurther characterizationof cells expressing STRO-1 in cultures of adult human bonemarrow stromal cellsrdquo Journal of Bone and Mineral Researchvol 14 no 8 pp 1345ndash1356 1999
[9] J L Simonsen C Rosada N Serakinci et al ldquoTelomeraseexpression extends the proliferative life-span and maintains
the osteogenic potential of human bone marrow stromal cellsrdquoNature Biotechnology vol 20 no 6 pp 592ndash596 2002
[10] M Al-Nbaheen R Vishnubalaji D Ali et al ldquoHuman stromal(mesenchymal) stem cells from bone marrow adipose tissueand skin exhibit differences in molecular phenotype and differ-entiation potentialrdquo Stem Cell Reviews and Reports vol 9 no 1pp 32ndash43 2013
[11] C A Gregory W G Gunn A Peister and D J Prockop ldquoAnAlizarin red-based assay of mineralization by adherent cells inculture comparison with cetylpyridinium chloride extractionrdquoAnalytical Biochemistry vol 329 no 1 pp 77ndash84 2004
[12] AMahmood L Harkness H D Schroslashder B M Abdallah andM Kassem ldquoEnhanced differentiation of human embryonicstem cells to mesenchymal progenitors by inhibition of TGF-120573activinnodal signaling using SB-431542rdquo Journal of Bone andMineral Research vol 25 no 6 pp 1216ndash1233 2010
[13] H-J Prins A K Braat D Gawlitta et al ldquoIn vitro inductionof alkaline phosphatase levels predicts in vivo bone formingcapacity of human bone marrow stromal cellsrdquo Stem CellResearch vol 12 no 2 pp 428ndash440 2014
[14] K Stenderup J Justesen C Clausen and M Kassem ldquoAgingis associated with decreased maximal life span and acceleratedsenescence of bone marrow stromal cellsrdquo Bone vol 33 no 6pp 919ndash926 2003
[15] D G Phinney G Kopen W Righter S Webster N Tremainand D J Prockop ldquoDonor variation in the growth propertiesand osteogenic potential of human marrow stromal cellsrdquoJournal of Cellular Biochemistry vol 75 no 3 pp 424ndash436 1999
[16] C M Digirolamo D Stokes D Colter D G Phinney R Classand D J Prockop ldquoPropagation and senescence of humanmarrow stromal cells in culture a simple colony-forming assayidentifies samples with the greatest potential to propagate anddifferentiaterdquo British Journal of Haematology vol 107 no 2 pp275ndash281 1999
[17] S James J Fox F Afsari et al ldquoMultiparameter analy-sis of human bone marrow stromal cells identifies distinctimmunomodulatory and differentiation-competent subtypesrdquoStem Cell Reports vol 4 no 6 pp 1004ndash1015 2015
[18] B Peruzzi A Cappariello A Del Fattore N Rucci F DeBenedetti and A Teti ldquoc-Src and IL-6 inhibit osteoblastdifferentiation and integrate IGFBP5 signallingrdquo Nature Com-munications vol 3 article 630 2012
[19] D Fortunati S Reppe A-K Fjeldheim M Nielsen V TGautvik and K M Gautvik ldquoPeriostin is a collagen associatedbonematrix protein regulated by parathyroid hormonerdquoMatrixBiology vol 29 no 7 pp 594ndash601 2010
[20] A Oshima H Tanabe T Yan G N Lowe C A Glackin andA Kudo ldquoA novel mechanism for the regulation of osteoblastdifferentiation transcription of periostin a member of thefasciclin I family is regulated by the bHLH transcription factorTwistrdquo Journal of Cellular Biochemistry vol 86 no 4 pp 792ndash804 2002
[21] T G Kashima T Nishiyama K Shimazu et al ldquoPeriostina novel marker of intramembranous ossification is expressedin fibrous dysplasia and in c-Fos-overexpressing bone lesionsrdquoHuman Pathology vol 40 no 2 pp 226ndash237 2009
[22] CGalliM PiergianniM Piemontese et al ldquoPeriostin improvescell adhesion to implantable biomaterials and osteoblasticdifferentiation on implant titanium surfaces in a topography-dependent fashionrdquo Journal of Biomedical Materials Research -Part A vol 102 no 11 pp 3855ndash3861 2014
18 Stem Cells International
[23] D L Coutu H W Jian A Monette G-E Rivard M DBlostein and J Galipeau ldquoPeriostin amember of a novel familyof vitamin K-dependent proteins is expressed by mesenchymalstromal cellsrdquo The Journal of Biological Chemistry vol 283 no26 pp 17991ndash18001 2008
[24] T Minamizato K Sakamoto T Liu et al ldquoCCN3NOV inhibitsBMP-2-induced osteoblast differentiation by interacting withBMP andNotch signaling pathwaysrdquo Biochemical and Biophysi-cal Research Communications vol 354 no 2 pp 567ndash573 2007
[25] W Si Q Kang H H Luu et al ldquoCCN1Cyr61 is regulated bythe canonicalWnt signal and plays an important role inWnt3A-induced osteoblast differentiation of mesenchymal stem cellsrdquoMolecular and Cellular Biology vol 26 no 8 pp 2955ndash29642006
[26] Q Luo Q Kang W Si et al ldquoConnective tissue growth factor(CTGF) is regulated by Wnt and bone morphogenetic proteinssignaling in osteoblast differentiation of mesenchymal stemcellsrdquo The Journal of Biological Chemistry vol 279 no 53 pp55958ndash55968 2004
[27] Y-S Yoon N Lee and H Scadova ldquoMyocardial regenerationwith bone-marrow-derived stem cellsrdquo Biology of the Cell vol97 no 4 pp 253ndash263 2005
[28] Y S Yoon ldquoCardiac regeneration with novel bone marrow-derived multipotent stem cellsrdquo Discovery Medicine vol 5 pp204ndash208 2005
[29] Y-S Yoon A Wecker L Heyd et al ldquoClonally expanded novelmultipotent stem cells from human bone marrow regeneratemyocardium after myocardial infarctionrdquo Journal of ClinicalInvestigation vol 115 no 2 pp 326ndash338 2005
[30] GM BolandG Perkins D J Hall andR S Tuan ldquoWnt 3a pro-motes proliferation and suppresses osteogenic differentiationof adult human mesenchymal stem cellsrdquo Journal of CellularBiochemistry vol 93 no 6 pp 1210ndash1230 2004
[31] D A Cook S W Fellgett M E Pownall P J OrsquoShea and PG Genever ldquoWnt-dependent osteogenic commitment of bonemarrow stromal cells using a novel GSK3120573 inhibitorrdquo Stem CellResearch vol 12 no 2 pp 415ndash427 2014
[32] J De Boer R Siddappa C Gaspar A VanApeldoorn R Foddeand C Van Blitterswijk ldquoWnt signaling inhibits osteogenicdifferentiation of humanmesenchymal stem cellsrdquoBone vol 34no 5 pp 818ndash826 2004
[33] MAl-Toub A AlmusaM Almajed et al ldquoPleiotropic effects ofcancer cellsrsquo secreted factors on human stromal (mesenchymal)stem cellsrdquo Stem Cell Research and Therapy vol 4 article 1142013
[34] L Chen D Qanie A Jafari et al ldquoDelta-like 1fetal antigen-1(Dlk1FA1) is a novel regulator of chondrogenic cell differenti-ation via inhibition of the Akt kinase-dependent pathwayrdquoTheJournal of Biological Chemistry vol 286 no 37 pp 32140ndash321492011
[35] T Matsunobu K Torigoe M Ishikawa et al ldquoCritical roles ofthe TGF-120573 type I receptor ALK5 in perichondrial formationand function cartilage integrity and osteoblast differentiationduring growth plate developmentrdquo Developmental Biology vol332 no 2 pp 325ndash338 2009
[36] B C Nair D E Johnson R J Majeska J A Rodkey C DBennett and G A Rodan ldquoRat alkaline phosphatase II Struc-tural similarities between the osteosarcoma bone kidney andplacenta isoenzymesrdquo Archives of Biochemistry and Biophysicsvol 254 no 1 pp 28ndash34 1987
[37] P J Butterworth ldquoAlkaline phosphatase Biochemistry of mam-malian alkaline phosphatasesrdquo Cell Biochemistry and Functionvol 1 no 2 pp 66ndash70 1983
[38] A T Ali C B Penny J E Paiker et al ldquoAlkaline phosphataseis involved in the control of adipogenesis in the murinepreadipocyte cell line 3T3-L1rdquo Clinica Chimica Acta vol 354no 1-2 pp 101ndash109 2005
[39] G M Spaggiari and L Moretta ldquoCellular and molecularinteractions of mesenchymal stem cells in innate immunityrdquoImmunology and Cell Biology vol 91 no 1 pp 27ndash31 2012
[23] D L Coutu H W Jian A Monette G-E Rivard M DBlostein and J Galipeau ldquoPeriostin amember of a novel familyof vitamin K-dependent proteins is expressed by mesenchymalstromal cellsrdquo The Journal of Biological Chemistry vol 283 no26 pp 17991ndash18001 2008
[24] T Minamizato K Sakamoto T Liu et al ldquoCCN3NOV inhibitsBMP-2-induced osteoblast differentiation by interacting withBMP andNotch signaling pathwaysrdquo Biochemical and Biophysi-cal Research Communications vol 354 no 2 pp 567ndash573 2007
[25] W Si Q Kang H H Luu et al ldquoCCN1Cyr61 is regulated bythe canonicalWnt signal and plays an important role inWnt3A-induced osteoblast differentiation of mesenchymal stem cellsrdquoMolecular and Cellular Biology vol 26 no 8 pp 2955ndash29642006
[26] Q Luo Q Kang W Si et al ldquoConnective tissue growth factor(CTGF) is regulated by Wnt and bone morphogenetic proteinssignaling in osteoblast differentiation of mesenchymal stemcellsrdquo The Journal of Biological Chemistry vol 279 no 53 pp55958ndash55968 2004
[27] Y-S Yoon N Lee and H Scadova ldquoMyocardial regenerationwith bone-marrow-derived stem cellsrdquo Biology of the Cell vol97 no 4 pp 253ndash263 2005
[28] Y S Yoon ldquoCardiac regeneration with novel bone marrow-derived multipotent stem cellsrdquo Discovery Medicine vol 5 pp204ndash208 2005
[29] Y-S Yoon A Wecker L Heyd et al ldquoClonally expanded novelmultipotent stem cells from human bone marrow regeneratemyocardium after myocardial infarctionrdquo Journal of ClinicalInvestigation vol 115 no 2 pp 326ndash338 2005
[30] GM BolandG Perkins D J Hall andR S Tuan ldquoWnt 3a pro-motes proliferation and suppresses osteogenic differentiationof adult human mesenchymal stem cellsrdquo Journal of CellularBiochemistry vol 93 no 6 pp 1210ndash1230 2004
[31] D A Cook S W Fellgett M E Pownall P J OrsquoShea and PG Genever ldquoWnt-dependent osteogenic commitment of bonemarrow stromal cells using a novel GSK3120573 inhibitorrdquo Stem CellResearch vol 12 no 2 pp 415ndash427 2014
[32] J De Boer R Siddappa C Gaspar A VanApeldoorn R Foddeand C Van Blitterswijk ldquoWnt signaling inhibits osteogenicdifferentiation of humanmesenchymal stem cellsrdquoBone vol 34no 5 pp 818ndash826 2004
[33] MAl-Toub A AlmusaM Almajed et al ldquoPleiotropic effects ofcancer cellsrsquo secreted factors on human stromal (mesenchymal)stem cellsrdquo Stem Cell Research and Therapy vol 4 article 1142013
[34] L Chen D Qanie A Jafari et al ldquoDelta-like 1fetal antigen-1(Dlk1FA1) is a novel regulator of chondrogenic cell differenti-ation via inhibition of the Akt kinase-dependent pathwayrdquoTheJournal of Biological Chemistry vol 286 no 37 pp 32140ndash321492011
[35] T Matsunobu K Torigoe M Ishikawa et al ldquoCritical roles ofthe TGF-120573 type I receptor ALK5 in perichondrial formationand function cartilage integrity and osteoblast differentiationduring growth plate developmentrdquo Developmental Biology vol332 no 2 pp 325ndash338 2009
[36] B C Nair D E Johnson R J Majeska J A Rodkey C DBennett and G A Rodan ldquoRat alkaline phosphatase II Struc-tural similarities between the osteosarcoma bone kidney andplacenta isoenzymesrdquo Archives of Biochemistry and Biophysicsvol 254 no 1 pp 28ndash34 1987
[37] P J Butterworth ldquoAlkaline phosphatase Biochemistry of mam-malian alkaline phosphatasesrdquo Cell Biochemistry and Functionvol 1 no 2 pp 66ndash70 1983
[38] A T Ali C B Penny J E Paiker et al ldquoAlkaline phosphataseis involved in the control of adipogenesis in the murinepreadipocyte cell line 3T3-L1rdquo Clinica Chimica Acta vol 354no 1-2 pp 101ndash109 2005
[39] G M Spaggiari and L Moretta ldquoCellular and molecularinteractions of mesenchymal stem cells in innate immunityrdquoImmunology and Cell Biology vol 91 no 1 pp 27ndash31 2012