Opposingeffectsoflowversushighconcentrationsofwatersoluble ...Cancer stem cells (CD44þ) and non-stem cancer cells (CD44 ) cells were sorted and collected in FBS containing media followed

Cell Biology International ISSN 1065-6995doi: 10.1002/cbin.10830

RESEARCH ARTICLE

Opposing effects of low versus high concentrations of water solublevitamins/dietary ingredients Vitamin C and niacin on colon cancerstem cells (CSCs)Utsav Sen, Sudheer Shenoy P and Bipasha Bose *

Stem Cells and Regenerative Medicine Centre, Yenepoya Research Centre, Yenepoya University, University Road, Mangalore 575018 Karnataka, India

Abstract

Colorectal cancer is one of the global causes of cancer deaths. Cancer stem cells (CSCs) inside the tumour niche responsible formetastasis and relapses, and hence need to be targeted for cancer therapeutics. Although dietary fibre and lifestyle changeshave been recommended asmeasures for colorectal cancer prevention, no such recommendations are available for using watersoluble vitamins as prophylaxis measure for colorectal cancers. High dose of Vitamin C has been proven to selectively killcolon cancer cells having BRAF and KRAS mutations by inducing oxidative stress. In this study, we show for the first time theopposing effects of the low and high dose of Vitamin C and vitamin B3 on colon CSCs isolated from HT-29 and HCT-15colorectal carcinoma cell lines. At small doses, both of these vitamins exerted a cell proliferative effect only on CSCs, whilethere was no change in the proliferation status of non-stem cancer cells and wild-type (WT) populations. On the other hand,the death effects induced by high doses of Vitamin C and B3 were of the order of 50–60% and�30% on CSCs fromHT-29 andHCT15, respectively. Interestingly, the control fibroblast cell line (NIH3T3) was highly refractory all the tested concentrationsof Vitamin C and B3, except for the highest dose – 10,000mg of Vitamin C that induced only 15% of cell death. Hence, theseresults indicate the future scope of use of therapeutic doses of Vitamin C and B3 especially in patients with advanced colorectalcancer.

Keywords: cancer stem cells; colorectal cancer; HT-29; non-stem cancer cells; vitamin B3; Vitamin C

Introduction

Amongst all the cancers, the third most common cancer ofall types is colorectal cancers which affect both men andwomen at an equal frequency (Jemal et al. 2009). Moreover,with the sedentary lifestyle and fad dietary habits for foodcontaining high fat and less fibre, colorectal cancer is on therise on a global scale especially in the developed countries(Marley and Nan, 2016). Hence, dietary fibres and lifestylechanges are recommended measures for preventing colo-rectal cancers (Hastert and White, 2016).

The patients with advanced colorectal cancers have only8% 5-year survival rates (Marley and Nan, 2016). Fiftypercent amongst all the patients undergoing aggressivetumour resection surgery, in combination with chemother-apy, suffer from a cancer recurrence (Yazilitas et al. 2016).Less than 1% population amongst the colorectal cancer niche

also known as the cancer stem cells (CSCs) or tumour-initiating cells (TICs) are the rogue population of cells thatare responsible for cancer relapse (Bagheri et al. 2017). Asthe CSCs/TICs are the ones which have survived during theinitial chemotherapy, such CSCs/TICs tend to acquirechemoresistance and hence, the patient fails to respond tothe same chemotherapeutic drugs after the cancer relapses.Moreover, the CSCs have the capacity to regenerate the bulkof cancer. Hence, 90% of the drug failures in metastaticcancer are due to chemoresistance (Longley and Johnston,2005; Eduati et al. 2017). Such chemoresistance is attributedto various features of CSCs such as presence of ABCtransporter proteins for effluxing the drug (Zhou et al. 2001;Lou and Dean, 2007; Hao et al., 2010; Gatti et al., 2011; Shafiand Jabeen, 2017); presence of prosurvival protein BCL2 andassociated proteins such as BCL-XL, BCL-W, BCL A1A,MCL-1 (Sinicrope et al. 1995; Abdullah and Chow, 2013;

�Corresponding author: e-mail: [email protected]; [email protected] Bose and Sudheer Shenoy P are co-senior authors.

1127Cell Biol Int 41 (2017) 1127–1145 © 2017 International Federation for Cell Biology

http://orcid.org/0000-0002-2312-9734

Carter et al. 2016); altered DNA damage/repair response(Abdullah and Chow, 2013; Yu et al., 2016a; Wang et al.,2016; Wang et al., 2017); and also structural changes in theendocrine receptors (Stender et al., 2017).

Human colorectal cancer cells (CRCs) carrying KRAS orBRAF mutations that are often refractory to targetedtherapies are reportedly killed by high doses of Vitamin Cin vitro (Yun et al., 2015). The mechanism of killing ofcancer cells through the high dose of Vitamin C treatmenthas been attributed to the increased uptake dehydroas-corbate, the oxidised form of Vitamin C which, in turn,results in oxidative stress. The ROS thus generated inhibitsthe glyceraldehyde 3-phosphate dehydrogenase (GAPDH)leading to an energy crisis and cell killing in KRAS andBRAF mutated colon cancer cells (Yun et al., 2015).However, the effect of varying doses of Vitamin C andother dietary ingredients in cancer stem cells has neverbeen elucidated. Here, we have studied the effects ofvarying doses of dietary ingredients such as Vitamin C andNiacin on colon cancer stem cells isolated from HT-29 andHCT-15 colon cancer cell line. HT-29 cell line hasmutations in BRAF, PIK3CA, SMAD4, APC and TP53while KRAS and PTEN genes are non-mutated in this cellline. On the other hand, HCT-15 cell line has mutations inKRAS, APC and PIK3CA (Ahmed et al., 2013). The detailsof the mutations of these two colorectal cancer cell linescan also be obtained in the ATCC document (https://www.atcc.org/~/media/PDFs/Culture%20Guides/Cell_Lines_by_Gene_Mutation.ashx).

In this study, we show for the first time that small dosesof Vitamin C and Niacin are promoting the proliferationof cancer stem cells isolated from HT-29 and HCT-15colon cancer cell lines. However, it is only the high dosesof these dietary ingredients that have a cell killing effecton the CSCs isolated from the HT-29 colon cancer cellline.

Materials and methods

Cell culture

The colon adenocarcinoma cell lines HT-29 and HCT-15and the control fibroblast cell line NIH3T3 were procuredfrom the National Centre for Cell Science, Pune, India. Thecells were cultured in media containing 10% fetal bovineserum of EU origin (Cat No-RM-9955, Hi-Media Laborato-ries, India), DMEMand 1%Penstrep and 1% L-Glutamine allfrom Thermo Fisher Scientific, USA in 5% CO2 andhumidified atmosphere. The cells were split by using0.25% trypsin-EDTA (Thermo Fisher Scientific, USA)upon reaching 70–80% confluency. The colon carcinomacell lines needed splitting at a ratio of 1:6 while the controlfibroblasts were split at a ratio of 1:3.

Enumeration of the percentage of cancer stem cells usinganalytical flow cytometry

Adherent cells were washed with DPBS/Ca2þ-Mg2þ free(Thermo Fisher Scientific) followed by dislodging the cellsusing 0.25% Trypsin-EDTA. Three to four million cells werecollected, the pellet washed using DPBS. Cells were fixedusing 4% paraformaldehyde for 30min to overnight beforeperforming analytical flow cytometry. 0.5� 106 cells werestained using Alexa Fluor

1

488 conjugated anti-human/mouse CD44 antibody (Biolegend, Catalog No- 103015) for45min. 10,000 events were acquired using Guava Easycyteflow cytometer, and the results were analysed using Prosortflow cytometry software. The cell morphologies werephotographed using the Zeiss Primovert microscope withthe camera attachment.

Isolation of cancer stem cells and non-stem cancer cellpopulations using flow cytometric cell sorting

All the procedure after cell harvesting was done at 4�C toensure live cells for cell sorting. From a growing dish at70–80% confluency, media was discarded, and the cellswere washed with DPBS without Ca2þ and Mg2þ. Thecells were harvested using 0.25% Trypsin-EDTA. 5.0million cells were suspended in 500mL of ice-cold FACSdissociation buffer for 15min at 4Deg C. FACS dissocia-tion buffer comprised of 3% FBS in PBS with 1% Penstrepand 7.5mM EDTA. Centrifugation at 1200 rpm at 4�C wascarried out for the removal of the FACS dissociationbuffer. Up to 5.0� 106 cells were stained at a concentra-tion of 0.5mg of antibody in a staining volume of 100ml inFACS staining/sorting buffer containing 2% FBS in PBS,5mM EDTA and 1% Penstrep for 1 h in the dark at 4�C.0.2� 106 cells served as the unstained control. Stained cellswere washed thrice, resuspended in 1mL of staining/sorting buffer and sorted using S3e Bio-Rad cell sorter.Cancer stem cells (CD44þ) and non-stem cancer cells(CD44�) cells were sorted and collected in FBS containingmedia followed by cell plating and expansion. CD44 hasbeen taken as a robust single marker to sort colon cancerstem cells as it has been reported to reprogram coloncancer cells into cancer stem cell phenotype (Su et al.,2011a) and is also a proven potent functional marker toisolate colon cancer stem cells (Du et al., 2008).

Population doubling time

WT, CD44þ and CD44� cells were plated into triplicatesonto 10 cm dish at a cell number of 0.5 million. Cellharvesting at different time points followed by cell countingusing a hemocytometer and tabulated. The doubling time ofeach of the cell populations was calculated using the online

CSC and vitamins-proliferation versus killing U. Sen et al.

1128 Cell Biol Int 41 (2017) 1127–1145 © 2017 International Federation for Cell Biology

https://www.atcc.org/~~/media/PDFs/Culture%20Guides/Cell_Lines_by_Gene_Mutation.ashx



doubling time calculator. (http://www.doubling-time.com/compute.php) followed by graphical representation.

MTT assay

MTT assaywas performed as per the protocol freely availableon R andD systems’website. Briefly, 25,000 cells per well of a96-well plate were seeded followed by 4 h attachment time.The cells were treated with various doses of Vitamin C (lowconcentrations – 5, 10, 15, 20, 25mM; and high concen-trations 100, 200, 500, 1,000 and 10,000mM) for 2 h. Cellswere given a media change containing 0.5mg/mL of MTT[(3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazoliumbromide] and incubated for 4 h. The reaction was stopped

usingDMSO/incubated for 30min, and absorbance was readat 620 nm using amicroplate reader (Fluo Star Omega, BMGLabtech, Germany). The cell growth/inhibition was ex-pressed as the percentage of cell proliferation as compared tountreated controls run with each of the treatments andassigned a value of 100%.

Apoptosis versus live cell detection in live cells (live–deadcell staining)

Live–dead cell staining was performed for corroborating theMTT assay results. Only two concentrations of each vitamin(25mM – low dose and 1,000mM – high dose) were testedfollowed by the live-Hoechst versus dead-Propidium iodide

Figure 1 Analytical flow cytometry showing the gross percentage of cancer stem cells in HT-29 and HCT-15 colorectal cancer cell lines.Scatter plots of (a) HT-29 and (b) HCT-15 colorectal cell lines showing the cell size (forward scatter-FSC-Area) versus cell granularity (side scatter-SSC-Area). Unstained control cells from (b) HT-29 and (e) HCT-15 cell line in quadrant R2. CD44 stained cells/cancer stem cells from (c) HT-29 cell line and (f)HCT-15 cell line in quadrant R3.

U. Sen et al. CSC and vitamins-proliferation versus killing


http://www.doubling-time.com/compute.php

http://www.doubling-time.com/compute.php

Figure 2 Flow sorting of CD44+ cancer stem cells andCD44�non-stem cancer cells fromHT-29 andHCT-15 colorectal cancer cell lines. Scatterplots with forward and side scatter indicating cell size and granularity respectively from (a) HT-29 and (e) HCT-15 cell line. Dot plot with side scatter area(SSC-Area) versus side scatter height (SSC-Height) indicating singlets and doublets in (b) HT-29 and (f) HCT-15 cell line. The singlets have been gated inand included for sorting. Dot plot showing unstained cells (Unstained control) in (c) HT-29 cell line and (g) HCT-15 cell line. CD44 stained cells showingthe CD44 positive population in gate R4 and CD44 negative population in gate R3 in (d) HT-29 cell line and (h) HCT-15 cell line. Both the cell fractionsCD44 positive and CD44 negative were collected and expanded.



staining of all six cell types (HT29-WT, HT29-CD44þ andHT29-CD44�, HCT15-WT, HCT15-CD44þ, HCT15-CD44�) and NIH3T3 in the live cell cultures. Twentythousand cells per well were plated onto each well of 24-wellplate and allowed to grow for 18 h till 70% confluency. Eachcell type was treated with two concentrations (25 and1,000mM) of each Niacin and Vitamin C and controls (notreatment) in triplicates for 2 h at 37�C under 5% CO2. Thecells were then subjected to media change with stainingsolution containing DMEMþ 1% Penstrep, 5mg/mL ofHoechst 33342 dye (Sigma, Cat No-B2261) and 20mg/mLPropidium Iodide (Sigma, Cat No-P4170). After 10minincubation of the cells in staining solution, a fresh mediachange with complete media, visualisation and photomicro-graphs under dark using ZOE cell imager (Bio-Rad) wascarried out. The live cell nuclei stained blue (with Hoechst33342 dye), the apoptotic cells stained light blue withfragmented nuclei (with Hoechst 33342 dye) while the deadcells stained Red with PI. PI-stained cells were enumeratedfrom five different fields for each of the treatments andplotted graphically.

Tumourosphere 3D assay

2D cells of all types (HT29-WT, HT29-CD44þ and HT29-CD44�, HCT15-WT, HCT15-CD44þ, HCT15-CD44�) andNIH3T3 control fibroblasts were harvested by trypsiniza-tion. Cells were suspended in media in hanging drops/droplets. Hanging drops were allowed to form againstgravity on the lid of a 10 cm dish under humidifiedconditions. We have modified the protocol of hangingdrops by Foty (2011), regarding cell number and droplet sizeso as to obtain an average diameter of the 3D tumouro-spheres �200–250m for the cell types that successfullyformed spheroids by 48 h of seeding. The same cell numberfor hanging drops were maintained across all the six celltypes. The hanging drops were observed and photographedfor the formation of 3D spheroids/tumourosphere after4 days under ZOE cell imager (Bio-Rad). The spheroids withsharp boundaries were graded as tumourospheres, incontrast to, the cell aggregates that exhibited irregularboundaries.

Live immunostaining of 3D spheroids/tumourosphere forthe assessment of CD44þcancer stem cells

2D cells of all types (HT29-WT, HT29-CD44þ and HT29-CD44�, HCT15-WT, HCT15-CD44þ, HCT15-CD44�) andNIH3T3 control fibroblasts were harvested by trypsinisa-tion. Cells were suspended in media in hanging drops/droplets. Hanging drops were allowed to form againstgravity on the lid of a 10 cm dish under humidifiedconditions. We have modified the protocol of hanging

drops by Foty (2011), regarding cell number and droplet sizeso as to obtain the average diameter of the 3D tumouro-spheres �200–250m for the cell types that successfullyformed spheroids by 48 h of seeding. The same cell numberfor hanging drops were maintained across all the six celltypes. The hanging drops were observed and photographedfor the formation of 3D spheroids/tumourosphere after fourdays under ZOE cell imager (Bio-Rad). The spheroids withsharp boundaries were graded as tumourospheres, in

Figure 3 Gene expression analyses of cellular derivatives (WT, CSC-CD44+ and non-CSC-CD44�) from HT-29 and HCT-15 colorectalcarcinoma cell lines (a) CD44 expression (dCt values with respect toGAPDH). (b) EMT (Snail, Slug and Vimentin) and MET (EpCAM and E-Cadherin)markers in HT-29 cell lines (Fold change calculatedwith respectto WT) (c) EMT (Snail, Slug and Vimentin) and MET (EpCAM and E-Cadherin) markers in HCT-15 cell line (fold change calculated withrespect to WT).



contrast to, the cell aggregates that exhibited irregularborders.

Gene expression analysis

All the cell types (HT29-WT, HT29-CD44þ and HT29-CD44�, HCT15-WT, HCT15-CD44þ, HCT15-CD44�), 1million each were harvested by trypsinization. RNA wasisolated using the RNeasy mini kit (Qiagen, Germany,Catalog No-74104) as per manufacturer’s instructions. RNAwas quantified using Nanophotometer NP80 (Implen,

Germany). One microgram of RNA was used for transcrib-ing c-DNA using iSCRIPTTM c-DNA synthesis kit (Bio-RadCatalog Number-1708891), and qRT-PCR reactions wereperformed using SS-FastTM Eva Green Supermix (Bio-Rad,USA, Catalog No-172-5201) also as per manufacturer’sinstructions. Results were expressed as fold change to theWT cells for each of the cell lines (plotted as value 1 in the Yaxis). Fold change was calculated by the 2ddCt. Thehousekeeping gene GAPDH was found to express consis-tently in all the cell types (Ct value �20). The individualmRNA expressions of the tested genes of all the cell types (Ct

Figure 4 Live cell co-staining for CD44�AF488 (green); nuclear-Hoechst (blue) of 3Dpartial tumourospheresþcellular aggregates (WT cells);fully formed tumourospheres (CSCs), cellular aggregates (non-CSC) obtained from cellular derivates (WT, CSC-CD44+ and non-CSC-CD44�)of HT-29 andHCT-15 colorectal carcinoma cell lines. Phase contrast, Hoechst live nuclear stain (blue) and CD44�AF488 live cell staining and PhaseþCD44�AF488 merged image of partially formed tumourosphereþ cellular aggregates of (a) WT-HT-29 cells and (b) WT-HCT-15 cells. Phase contrast,Hoechst live nuclear stain (blue) and CD44�AF488 live cell staining and Phaseþ CD44�AF488 merged image of fully formed 3D tumourospheres of (b)CD44þ CSCs from HT-29 cell line and (e) CD44þ CSCs from HCT-15 cell line. Phase contrast, Hoechst live nuclear stain (blue) and CD44�AF488 live cellstaining and Phaseþ CD44�AF488 merged image of cellular aggregates of (c) non-CSC/CD44�HT-29 cells and (f) non-CSC-CD44�HCT-15 cells. Thecellular aggregates from CD44�/non-CSCs stained negative for CD44�AF488 (green). Scale bar – 100mM.



values) were first normalised with their respective GAPDHvalues for obtaining the dCt values. The absolute expressionof CD44 in all the six cell types (HT29-WT, HT29-CD44þ

and HT29-CD44�, HCT15-WT, HCT15-CD44þ, HCT15-CD44�) have been represented as dCt values, to, ensure avery high expression of this CSC marker in the desired celltypes. The primer sequences were designed by the authors

and were ordered from Sigma. The details of primers areprovided in Supplementary Table S1.

Statistical analysis

All the experiments were performed as three biologicalreplicates using three independently sorted batches of cells

Figure 5 Morphologies and population doubling times for various populations (cancer stem cells-CD44+; non-stem cancer cells-CD44� andmixed population-Wild type,WT) obtained from theHT-29 andHCT-15 colorectal cancer cell lines. Phase contrast images of (a)WT-HT-29 cells;(e) WT-HCT-15 cells; CD44þ/cancer stem cells from (b) HT-29 cell line and (f) HCT-15 cell line; CD44�/ non-stem cancer cells from (c) HT-29 cell line and(g) HCT-15 cell line. The inset (marked as white arrow head in the gross image) for each of the cell types shows the details of the respective cellularmorphologies. Population doubling time of all three cell populations from (d) HT-29 cell line and (h) HCT-15 cell line with duration of cultures in X-axisand cell number in millions in Y-axis. Scale bar – 100mm.



(CD44þ and CD44�). The results were expressed asMean� SEM.

Results

Cancer stem cells exhibited a high percentage�60% in theHT-29 cell line versus HCT-115–35%

The colorectal carcinoma cell lines upon subjection toanalytical flow cytometry for the assessment of thepercentage of cancer stem cells (CSCs) prior sorting, asCD44þ cells exhibited a high �62% in HT-29 (Figure 1c)versus �35% in HCT-15 cell lines (Figure 1f), respectively.The populations gated as R3 (CD44 negative), and R4 (CD44positive) were sorted and expanded from HT-29 (Figures2c and 2d for) and HCT-15 (Figures 2g and 2h) colorectalcarcinoma cell lines, respectively, for the entire study. Only52.3% cells were sorted as HT-29 CD44þ cells that weresinglets (Figures 2b amd 2d) and then expanded for all futureexperiments. 0.45% cells gated as R3 were sorted as HT-29CD44� singlets (Figures 2b and 2d) and expanded for allfuture experiments. Thirty-one percent of cells gated as R4were sorted as CD44þ singlets (Figures 2f and 2h), while 25%of cells gated as R3 were sorted as CD44� singlets (Figures2f and 2h) from theHCT-15 cell line. The sorted cells CD44þ

and CD44�were only used within two passages of culturing,so that the cells maintain their distinct characteristics. Therespective parent cell lines containing mixed populations ofCD44þ and CD44� and other uncharacterised cell typeswere termed as the wild type (WT).

The sorted CSC populations from both the cell linesmaintained a high CSC expression in 2D cultures andexhibited high EMT markers

The sorted cells, CD44þ and CD44�, were maintained in 2Dcultures for a maximum of two passages. The sorted CSC-CD44þ cells from both the cell types, HT-29 and HCT-15maintained a very high expression of cancer stem cell marker

CD44 as evident from the gene expression analysis for CD44gene (Figure 3a). The CD44� non-stem cancer cells sortedfrom both the cell lines HT-29 andHCT-15 exhibited little tonil expression of CD44 as expected (Figure 3a). The CD44þ

cancer stem cells showed high expression of the epithelial tomesenchymal transition (EMT) genes such as Snail, Slug andVimentin as analysed by qRT-PCR gene expression analysis(Figures 3b and 3c). However, mesenchymal to epithelialmarkers (MET) such as EpCAM and E-Cadherin wereequally expressed in WT, CD44þ CSCs and non-CSCs fromboth HT-29 and HCT-15 cell lines probably owing to theirorigin from epithelial cells of the colon (Figures 3b and 3c).In the case of WT cells, both the cell lines exhibited similarexpression levels of all the EMT (Snail, Slug and Vimentin)and MET (EpCAM and E-Cadherin) markers (Figures3b and 3c).

The sorted CSC populations from both the cell linesformed 3D spheroids/tumourospheres which could bedistinguished from the non-spheroid aggregates of non-stem cells and control fibroblasts

All the six cell types (HT-29 WT, HT-29 CD44þ, HT-29CD44�, HCT-15WT,HCT-15CD44þ andHCT-15 CD44�)were subjected to 3D tumoursphere assay as described in themethods section. Live 3D spheroids and cellular aggregateswere stained with CD44 and Hoechst. Only CD44þ cellsobtained from both the cell lines formed tumourospherewith distinct boundaries which also stained strongly positivefor CD44 (Figures 4b and 4e). However, the CD44�

populations from both the cell types did form cellularaggregates and stained negative for CD44 as expected(Figures 4c and 4f). NIH3T3 control fibroblasts also formedcellular aggregates (data not shown). WT HT-29, as well asHCT-15 cells, also formed cellular aggregates and partialspheroids with sparse staining with the CD44 antibody(Figures 4a and 4d). All the 3D tumourosphere and cellularaggregates stained negative for propidium iodide therebyconfirming the absence of dead cells (data not shown).

Table 1 Population doubling time for WT, CD44+ and CD44� cell populations from HT-29 and HCT15 cell lines. Table showing the cellnumber inmillions of the cell populationsWT, CSCs (CD44þ) andnon-stem cancer cells (CD44�) fromHT-29 andHCT-15 colorectal carcinomacell lines at different time points for estimation of the respective doubling times of the aforesaid cell populations.

Hours in culture and cell number

Cell type 0 h 24h 48h 72h 96 h 120 h Doubling time

HT-29/ WT (cell number in millions) 0.5 1 2.1 3.96 8.3 16.11 23.6 hHCT-15/WT (cell number in millions) 1.0 2.597 5.19 7.79 10.38 12.98 18.48 hHT-29/CD44þ (cell number in millions) 0.5 1.3 3 9 18 Not counted 18 hHCT-15/CD44þ (cell number in millions) 1.0 2.92 5.85 8.77 11.70 14.62 16.41 hHT-29/CD44� (cell number in millions) 0.5 1 2.01 4.12 8.1 15.75 24.11 hHCT-15/CD44� (cell number in millions) 1.0 1.775 3.55 5.32 7.10 8.87 27.04 h



Cancer stem cells, non-stem cancer cells and WTpopulations exhibited a variable doubling time andmorphological behavior

The sorted populations as Cancer stem cells CSCs (CD44þ),non-stem cancer cells (CD44�) from both the colorectalcarcinoma cell lines HT-29 and HCT-15 showed distinctmorphologies in (Figure 5). The WT and CSC populationsexhibited a mixed mesenchymal and epithelial morphology

(Figures 5a, 5b, 5e and 5f) in contrast to, the epithelialmorphology of non-CSC/CD44� cells (Figures 5c and 5g).The CSC populations from the HT-29 cell line exhibited adoubling time of 18 h, in contrast, to the CD44� and WTpopulations that took around 24 h thereby indicating aninfluence of slow growing CD44� non-stem cancer cells onthe overall growth of WT population (Figure 5d, Table 1).Similarly, the CSCCD44þ populations from theHCT-15 cellline grew fastest with a population doubling time of �16 h,

Figure 6 Graphical representation for the percentage of cell proliferation of all three populations (WT, CD44+ and CD44�) and also NIH3T3control fibroblast cell line with respect to low (5–25mM) and high (100–10,000mM) concentration ranges of Vitamin C and Niacin.Percentage cell proliferationwith respect to varying concentrations ofVitamin C for all three cellular derivatives (WT, CD44þ, CD44�) of (a) HT-29 cell line(b) HCT-15 cell line. Percentage cell proliferation with respect to varying concentrations of Niacin for all three cellular sorted derivatives (WT, CD44þ,CD44�) of (c) HT-29 cell line (d) HCT-15 cell line. (e) Percentage cell proliferations of NIH3T3 control fibroblast, in response to, varying concentrations ofVitaminC andNiacin. VitaminC/Niacin concentrations inmM inX-axis are plotted versus percentage of cell proliferation plotted in the Y-axis. An arbitraryvalue of 100% proliferation has been assigned to the respective untreated control cells. The values for this plot have been taken as a mean� SE of threeindependent experiments.



while the non-CSCs (CD44�) grew slowest with a popula-tion doubling time of 27 h and WT had intermediate celldoubling time of �18 h (Figure 5h, Table 1).

Proliferative effects of cancer stem cells, in response to,low concentrations of Vitamin C and vitamin B3/niacin

CSC/CD44þ populations isolated from HT-29 and HCT-15cell lines, respectively, when subjected to low concentrations(5,10, 15, 20 and 25mM) of Vitamin C (Figures 6a and 6b,Table 2) and that of Niacin (Figures 6c and 6d, Table 3),exhibited an increase in cell proliferation. In the case of the

CSCs isolated from the HT-29 cell line, the growth in cellproliferation, in response to the aforementioned small dosesof both the vitamins was more pronounced, as compared tothe CSCs isolated from HCT-15 cell line (Figures 6a–6d).

In the case of Vitamin C exposure, the CSC/CD44þ

populations from HT-29 cell line exhibited about 170%increase in cell proliferation as compared to the untreatedcontrols (Fgure 6a, Table 2). Also, at all the five lowconcentrations of Vitamin C from 5 to 25mM, the CSCsexhibited similar proliferation rates �160–170% in HT-29CSCs and�107–116% in HCT-15 CSCs cell proliferation ascompared to their respective untreated controls (Figures

Table 2 Percentage of cell proliferation upon exposure to low (5–25mM) and high concentration ranges (100–10,000mM) of vitamin C/ascorbic acid in various cell populations obtained from HT-29 and HCT-15 colorectal carcinoma cell lines respectively. Table showing therespective percentages of cell proliferation of the cell populations WT, CSCs (CD44þ) and non-stem cancer cells (CD44�) with respect tovarious concentrations (5–10,000mM) of Vitamin C/Vitamin C from HT-29 and HCT-15 cell lines. The untreated control cells for each of thecell type WT, CSCs (CD44þ) and non-stem cancer cells (CD44�) have been assigned an arbitrary value of 100% cell proliferation.

Cell types and percentage of cell proliferation

Vitamin C conc. mM HT-29 WT HCT-15 WT HT-29-CD44þ HCT-15 CD44þ HT-29-CD44� HCT-15 CD44�

0 100 100 100 100 100 1005 114 108 172 107 86 12210 104.7 114 160.8 107.26 93.11 121.0115 95.81 121.03 172 110.15 97.1 10620 90.57 94.19 169.73 115.01 99.2 93.7925 88.48 90.01 169.73 116.32 86.59 98.13100 21.51 89.01 68.28 93.79 14.23 96.15200 21.81 88.06 69.34 73.58 14.13 96.52500 23.56 85.42 67.23 72.01 12.68 95.161,000 24.34 83.12 65.81 69.56 11.59 88.0610,000 27.06 226 61.18 304.10 14.13 233.03

Table 3 Percentage of cell proliferation upon exposure to low (5–25mM) and high concentration ranges (100–10,000mM) of Niacin invarious cell populations obtained from HT-29 and HCT-15 colorectal cancer cell lines, respectively. Table showing the respectivepercentages of cell proliferation of the cell populations WT, CSCs (CD44þ)and non-stem cancer cells (CD44�) with respect to variousconcentrations (5–10,000mM) of Vitamin B3/Niacin from HT-29 and HCT-15 cell lines. The untreated control cells for each of the cell typeWT, CSCs (CD44þ) and non-stem cancer cells (CD44-) have been assigned an arbitrary value of 100% cell proliferation.

Cell types and percentage of cell proliferation

Niacin conc. mM HT-29 WT HCT-15 WT HT-29-CD44þ HCT-15 CD44þ HT-29-CD44- HCT-15 CD44�

0 100 100 100 100 100 1005 98 110 163 72 96 11810 103.06 110 150 76 92.52 111.9015 98.46 112 137.5 84.2 85.4 10820 98.97 133 131.25 105.1 78.29 114.5025 94.89 135 128.75 113.2 78.29 114.56100 22.95 97 65 77.65 15.3 103.40200 22.95 95 55 75 12.45 113.40500 22.48 89 52.5 72.1 13.52 109.401,000 23.97 87.12 52.5 70.12 11.74 111.0610,000 19.89 87.06 56.25 69.16 11.74 112.97



6a and 6b, Table 2). Similarly, low (25mM) concentrations ofNiacin also induced an increased proliferation of the CSC/CD44þ cell populations to �130% in HT-29 CSCs and�113% in HCT-15 CSCs (Figures 6c and 6d, Table 3).Interestingly, in the Niacin low dose group (5–25mM),5–15mM doses were inhibitory to the growth and 20 and25mM induced proliferation of CSC/CD44þ populationsfrom the HCT-15 cell line. In the case of, the HT-29 CSC/CD44þ cell populations exhibited continued over-prolifera-tion in all the low tested low doses (Figures 6c and 6d,Table 3).

In the case ofWT populations of both the cell lines, HT-29and HCT-15 Vitamin C treatment of 5mM failed to exhibitany significant difference in the percentage of cellproliferation as compared to the untreated controls (Figures6a and 6b, Table 2). However, the 25mM concentration ofVitamin Cwas able to exert a cell killing effect of the order ofonly�10 and �12%, respectively, in the WT populations ofHCT-15 and HT-29 cell lines (Figures 6a and 6b, Table 2).Interestingly, the non-stem cancer cells (CD44�) from HT-29 and HCT-15 exhibited different behaviour regardingtheir sensitivities/cell killings with both Niacin and VitaminC all tested doses from 5 to 10,000mM. Doses as small as25mMof Vitamin C induced a dose-dependent reduction incell proliferation of CD44�HT-29 cells of the order of�86%(14% cell killing) (Figures 6a, Table 2) and �78% (22% cellkilling) in the case of Niacin (Figure 6c, Table 3). However,the CD44� cells fromHCT-15 cell line were refractory to thelow doses (5–15mM) of Vitamin C and all the small doses ofNiacin (5–25mM) thereby indicating a probable distinctregulation of such cells.

Cell killing effects of cancer stem cells to highconcentrations of Vitamin C and Vitamin B3/Niacin

In contrast to the small doses of Vitamin C and Niacin, thehigh doses of these dietary ingredients from 100 to 1,000mMwere capable of inducing cell killing effects in CSCs. The cellkillings caused by the high dose of Vitamin C (1,000mM)were of the order of �35 and �29% (percentage of cellkilling is indicated as 100 minus percentage of cellproliferation) in CD44þ/CSC from HT-29 and HCT-15,respectively (Figures 6a and 6c, Table 2). However, in thecase of 1,000mM of Niacin treatment, we observed a cellkilling in CD44þ/CSC from HT-29 and HCT-15 cell lines ofthe order of 48 and 30%, respectively (Figures 6c and 6d andTables 2 and 3). Precisely, there was a dose-dependentreduction in the proliferation of CSC population from HT-29 cell line from �68 to 61% upon treatment with 100–10,000mM Vitamin C (Figure 6a, Table 2) as compared tothe untreated controls. However, we observed a Vitamin Cdose-dependent reduction in HCT-15/CD44þ CSC prolif-eration from �93 to 69%. One significant observation was

the highest proliferation of CD44þ/CSC from HCT-15, inresponse to, the highest tested concentration of Vitamin C(10,000mM). Similarly, the CSC populations also exhibited aNiacin dose (100–10,000mM) dependent reduction in cellproliferation from�65 to 56% in HT-29/CD44þ and�77 to69% in HCT-15/CD44þ as compared to the untreatedcontrols (Figure 5, Table 3). However, the CSCs from boththe cell types still failed to achieve LD50 at any of the studieddoses of both Vitamin C and Niacin.

The WT and non-CSC populations only from the HT-29cell line exhibited greater sensitivities/higher cell killingeffects of both Vitamin C and Niacin. Precisely, doses from100 to 10,000mM (Figures 6a–6d and Tables 2 and 3)induced greater cell killings in HT-29 cell line derivatives(WT and CD44�) as compared to the WT and non-CSCpopulations from HCT-15 cell line. The HT-29 WT cellpopulations exhibited a cell proliferation of the orderof �21 to 27% in case of Vitamin C treatment(100–10,000mM) (Figure 6a, Table 2). HT-29 WTpopulation also exhibited �22 to 19% reduction in cellproliferation in case of Niacin treatment (100–10,000mM)(Figure 6c, Table 3) as compared to the untreated controls.The highest sensitivity towards Vitamin C and Niacin wasindeed observed in the case of the HT-29 non-stem cancercell populations (CD44�). The dose-dependent reducedcell proliferation of HT-29 non-CSCs, in response to highdoses (100–10,000mM) Vitamin C was �14–10%(Figure 6a, Table 2) and to that of Niacin was�15–11%, respectively (Figure 6c, Table 3) as comparedto untreated controls. This dose-dependent reduced cellproliferation of non-CSCs from HT-29 indicates theachievability of a value of more than LD50 in WT andnon-CSC cell populations.

TheWT and non-CSC populations fromHCT-15 cell linefailed to exhibit any dose-dependent behaviour in cellproliferation/inhibition through the doses 100–10,000mMof both Vitamin C and Niacin. Vitamin C doses from 100 to1,000mM, however, induced a dose-dependent reduction incell proliferation from �89 to �83% in HCT15-WT and�96 to �88% in HCT-14 non-CSC (CD44�) cell popula-tions. We, however, observed a high degree of cellproliferation �119 and �233% in HCT-15 WT and non-CSC (CD44�) populations, respectively, only at the highestconcentration (10,000mM) of Vitamin C thereby indicatingan increased cell proliferation after initial dose-dependentdecreased proliferation with Vitamin C levels from 100 to1,000mM.

In response to Niacin concentrations from 100 to10,000mM, HCT-15 WT exhibited a dose-dependentdecrease in cell proliferation from �97 to �87%(Figure 6d, Table 3). However, the non-CSC populationshowed an increase in cell proliferation from �103 to�113% in response to Niacin doses from 100 to 10,000mM



(Figure 6d, Table 3) thereby indicating a differentialresponse of HCT-15 non-CSC cell populations.

NIH3T3 control cells responded in a proliferative mannerto all the low to high dose (5–10,000mM) of Niacin whileproliferative only to low dose of Vitamin C (5–25mM)

The NIH3T3 control fibroblasts, when subjected to all theindividual doses (5–10,000mM) of Vitamin C and Niacin,behaved proliferative only to small doses (5–25mM) ofVitamin C and all the tested doses of Niacin (Figure 6e). Theproliferation exhibited by the NIH3T3, in response to,Niacin was rather dose dependent (Figure 6e) with highestcell proliferation (�250%) observed with the highest dose(10,000mM) of Niacin. However, the highest dose(10,000mM) of Vitamin C was able to induce only aminimal of 15% cell killing of NIH3T3 cells therebyindicating a non-attainment of LD50 value, even in responseto, the highest dose of Vitamin C.

CD44þ CSC from HT-29 cell line, as compared to, theCD44þ CSC from HCT-15 cell line, responded well to thecell killing effects of high dose (1,000mM) Vitamin C andNiacin

We next attempted to have a live assessment of cellproliferation versus cell killing effects of low and highconcentrations of Vitamin C and Niacin, all six cell types(HT-29-WT, HT-29-CD44þ, HT-29 CD44�, HCT-15-WT,HCT-15-CD44� and HCT-15-CD44�). One single low dose(25mM) and one single high dose (1000mM) of each ofVitamin C and Niacin were tested. The maximum dose10,000mMwas avoided because it had earlier induced�300and �230% cell proliferation in CD44þ CSC and non-CSC(CD44�) cells, respectively, from HCT-15 cells as estimatedby MTT assay (Figures 6b and 6d, Table 2). The live cellsstained simultaneously with Hoechst and PI upon quantifi-cation exhibited the high cell killing (percentage of PIpositive cells), upon treatment with 1,000mM of Vitamin CandNiacin in CD44þ-CSC fromHT-29 cells, as compared tothe CD44þ-CSC from HCT-15 cells (Figure 7a). Thepercentage of cell killings by 1,000mM of Vitamin C andNiacin in HT-29-CD44þ-CSC were of the order of �35 and�50%, respectively (Figure 7a) which corroborates with theMTT-assay results (Figures 6a, 6c and Tables 2 and 3).However, the percentage of cell killings by 1,000mM ofVitamin C and Niacin in HCT-15 CD44þ-CSC were only ofthe order of �30% (Figure 7a). Figures 7b–7d, respectively,represent high cell killing (�80–90% PI positive cells),moderate cell killing (30–50% PI positive cells) and low cellkilling (5–10% PI positive cells).

However, with regard to, the low concentration (25mM)of Vitamin C and Niacin, almost negligible PI positive cells

were detected in CD44þ-CSC from HT-29 and HCT-15 celllines, respectively (Figures 7a and 7d). No cell killings, withresponse to, 25mM of Vitamin C and Niacin in the CD44þ-CSC from HT-29 and HCT-15 cell lines also corroboratewell with the MTT data (Figures 6a–6d, Tables 2 and 3) inwhich we had observed an increase in cell proliferation at aconcentration of 25mM. Amongst all the six cell types, HT-29-WT and HT-29-CD44-(non-CSC) exhibited highest cellkilling of the order of�80–90%, in response to, 1,000mM ofboth Vitamin C and Niacin thereby again corroboratingwith the MTT data (Figures 6a, 6c and Table 3).

Discussion

In all forms of cancer, the first appearance either detectedearly or at the second stage appears to be less complicatedand is well managed with the help of conventional therapiessuch as surgery, chemotherapy and radiotherapy. However,minimal interventions are available, in response to, possiblerelapse of the cancers. It is mostly a relapsed cancer that isresponsible for the major deaths of the cancer patients.Colorectal cancers have very high rates of relapses (Guerreraet al., 2016). Moreover, it is now known that cancer stemcells-CSCs/tumour initiating cells-TICs are the ones thatform the bulk of a relapsed tumour (Yu et al., 2016b).Moreover, the cells that home to the metastatic sites are alsolikely to be predominated with cancer stem cells. Therefore,it is important to target the cancer stem cells mainly after thefirst phase of conventional therapy for any cancer patient.

Colorectal cancer stem cells are often isolated asCD44high/CD133high (Zhou et al., 2016). Additionalmarkers such as CD24, CD166 (ALCAM), Oct4, Nanog,Sox2, ALDH1, CD166, CD26, EpCAM are reportedlyexpressed in colon CSC (Abetov et al., 2015, Wahab et al.,2017). The classical Nature paper had the first time reportedthe isolation of human colorectal CSC as CD133 positive in2007 (Ricci-Vitiani et al., 2007). However, CD44 has beenproved to be a robust functionalmarker for colon CSC by theloss of colony forming abilities of CD44þ colon CSC thatwere knocked down for CD44 in vitro and in vivo in APC/Minþ mice (the Mouse model for colorectal cancers).However, CD133þ colon CSC, despite knockdown ofCD133 could still form tumour colonies (Du et al., 2008).Later on, Su et al. (2011a) could directly reprogram coloncancer cells into colon cancer stem cells using a singlemarker CD44 (Su et al., 2011b). Hence, we decided to isolatethe colon cancer stem cells by the unique, robust markerCD44. Also, the CD44þ has been reported to preserve theparental stem cell characteristics in culture, as compared toCD44� cells (Du et al., 2008). However, as we furthercharacterised the flow-sorted CSCs (CD44þ) and non-CSCs(CD44�) cells up to two passages of culture, we found highexpression of CD44 and stemness markers/EMT markers in



CSCs versus low expression of CD44 in non-CSCs usingqRT-PCR (Figure 3). Most important, the CSCs from bothHT-29 and HCT-15 cell lines successfully formed 3Dtumourospheres (Figure 4) thereby validating the stemnessof CD44þ cells isolated in this study. Hence, we confirm theuse of bonafide colon CSC (CD44þ) for all our experimentsfor studying the effects of Vitamin C and Niacin.

The period between cancer remission and cancer relapseis the time which is largely unmonitored. Hence, it isimportant to identify individual molecules or dietaryingredients that can target the small populations of cancerstem cells left behind after first round of conventional cancer

therapy, surgery, chemotherapy and radiotherapy. Also,various vitamins are used as nutritional and epigeneticapproaches for cancer control and prevention (Verma,2012). Often, the early on and terminal cancer patients areadvised to take various forms of dietary supplementation(Cuhls et al., 2016). Vitamin C and Niacin are such commondietary ingredients that have been recommended forconsumption by the cancer patients (Starr, 2015; Cuhlset al., 2016). The specific logic of choosing the vitaminsniacin and ascorbic acid in this study is because of making abeginning in studying the effects of the vitamins that play asignificant role in cellular energetics/metabolism (niacin)

Figure 7 Graphical representation of enumeration for the percentage of cell death (% PI positive cells) from five microscopic fields of livecells (WT, CD44+ and CD44�) co-stained with Hoechst and PI from HT-29 and HCT-15 cell lines and representative images of microscopicimages showing high, medium and low cell death in response to low (25mM) and high (1,000mM) concentrations of Vitamin C and Niacin,respectively. (a) Graphical representation of percentage PI positive cells in WT, CD44þ and CD44� from HT-29 and HCT-15 cell lines treated with low(25mM) and high (1,000mM) concentrations of Vitamin C and Niacin, respectively. Representative phase contrast, PI dead cells and Hoechst live cellsshowing (b) high (80–90%) PI positive cells, (c) moderate number (30–50%) of PI positive cells and (d) low (5–10%) of PI positive cells. Scalebar – 100mM.



and oxidative stress reduction (vitamin C) on healthy cells.Moreover, cancer cells have an altered cellular energetics(Uetaki et al., 2015; Yang and Sauve, 2016). Hence, wewanted to study if there is an alteration of cell behaviourupon treatment with niacin and ascorbic acid on colorectalcancer stem cells. In our next study, we intend to furtherwork on cellular energetics upon treatment of colorectalcancer stem cells with niacin and ascorbic acid respectively.Secondly, we had simultaneously found the literature on theeffect of high doses of vitamin C on the killing of colorectalcancer cells by inducing oxidative stress. Hence, we wereinterested in studying the status of colorectal cancer stemcells using high versus low doses of vitamin C.

Reports regarding the usage and research with Vitamin Chave been rather exhaustive, as compared to that of Niacin.Vitamin C has been reported to reduce the malignantpotential of malignant melanoma cell lines (Fischer andMiles, 2017); reduce cancer in patients when administered inhigh doses via intravenous route (Padayatty et al., 2006;Sawant et al., 2012; Sharma et al., 2012). However,intravenous administration of palmitoyl ascorbate lip-osomes proved to be more potent, as compared to, freeVitamin C injection in Balb/c mice model of mammarycarcinoma (Sawant et al., 2012). Also, Vitamin C reportedlytargets the cancer cells thereby killing them throughmultiplemechanisms (Du et al., 2012). A recently decipheredmechanism regarding Vitamin C induced selective killingof prostate cancer cells is via enhanced expression of NFkbtranscription factor RelB, which in turn decreased theexpression of Sirtuin and intracellular antioxidant MnSOD(Wei et al., 2017). However, in normal prostate epithelialcells, Vitamin C-mediated enhanced expression of RelBrather improved antioxidant defences of the cells therebyleading to better health of the cells and enhancedproliferation (Wei et al., 2017). In our context, it would behence, interesting to decipher the status of RelB inVitaminCand Niacin treated CSC, non-stem cells and WT cells whichform the future scope of this study.

The observations viz. low versus high concentrations ofvitamin C and niacin on proliferations and cell death,respectively, in cancer stem cells is being solely elucidatedunder in vitro culture conditions in this study. However,further corroboration of these findings under in-vivoconditions in tumour xenografts model of CD44þ coloncancer stem cells from HT-29 and HCT-15 colon carcinomacell lines will be able to provide a clearer picture regardingthe effect of such water-soluble vitamins. Secondly, withregard to, the different effects of Vitamin C on the cell types,attributes such as sodium-dependent vitamin C transporterproteins (SVCT) might be responsible. Various levels ofabsorption of vitamin C has been reported in vitro inMCDK-MRD1 kidney cells (Luo et al., 2008), in response to,different concentrations of sodium in the media, pH, as well

as, temperature conditions. However, in our case, since themedia compositions and culture conditions of all the celltypes was same, the sodium, pH or temperature dependentvariation in SVCT activity, and hence, vitamin C absorptionis a remote possibility. Nevertheless, it forms the scope of thefuture study to assess the levels of SVCT in various cell linescultured in the given media conditions with or without thepresence of the tested water-soluble vitamins, vitamin C andniacin. Moreover, the presence of trace amounts oftransition metals mostly iron in the synthetic media suchas DMEM might also be responsible for inducing variationin the uptake of the tested water soluble vitamins (Yao andAsayama, 2017). However, that is a given situation under in-vitro conditions and corroboration of the experiments inxenografts mouse models for cancer stem cells in the bestalternative.

Thirdly, the different response of CSCs towards lowversus high concentrations of the tested water solublevitamins also can be attributed to the differential expressionof the cell’s internal antioxidant defence mechanisms. It ispossible that elevated levels of these vitamins are inhibitoryto the cellular antioxidant defence machinery and isresponsible for the reduced expression of catalase andsuperoxide dismutase. As a result of reduced expression ofsuperoxide dismutase and catalase, the ROS generatedmightbe responsible for the killing of the cells at a highconcentration. In other words, homeostatic levels of catalaseand superoxide dismutase in a cell prevent the cell fromundergoing ROS-mediated apoptosis. Hence, it would be ofgreat importance to understand whether the different effectsof cell proliferation versus cell killing, in response to, lowversus high concentrations of the tested water solublevitamins is due to the effect of ROS versus ROS quenching.Accordingly, assessment of ROS and hydrogen peroxide inCSCs treated with varying concentrations of Vitamin C andNiacin makes the future scope of this study.

Fourthly, regarding the mechanisms of action of Niacin,with regard to, cell proliferation versus cell killing, in highand low doses respectively, we speculate a shift in theattenuation of ROS. Niacin at concentrations 100–300 nmreportedly quenches ROS, Nitric oxide (NO) and Nitricoxide synthase 2 (NOS2) in LPS treated human maturemacrophages (Montserrat-de la Paz et al., 2017). In ourstudies, the concentrations of Niacin used were much higherthan the ones reported to quench ROS. Precisely, our lowerconcentrations of Niacin ranged from 5 to 100mM that is�50 to 500 times more than the levels reported byMontserrat-de la Paz et al. (2017), that caused proliferationof the CSCs. Also, long-term (8 weeks) administration ofvery high doses of Niacin [�30mg/kg body weight] throughdrinking water to the mice diet resulted in ROS induction,hyperglycemia, reduced glucose-induced insulin release inpancreatic islets (Chen et al., 2015). Hence, we can conclude



that lower concentrations of Niacin up to 100mM canquench the ROS, while the higher levels (200–10,000mM) ofNiacin might be inducing ROS and resulting in cell killingsof CSCs. However, the levels of ROS generated via varyingdose of Niacin need further validation. We also propose amechanism of action of high versus low concentrations ofNiacin, in response to, cell proliferation versus cell killing(Figure 8).

Interestingly, cancer cells exhibit higher levels of oxidativestress, as compared to, normal cells, and hence thismechanism can be used as a strategy to induce selectivekilling of cancer cells (Lee et al., 2017). Hence, increasing theamounts of pro-oxidants in the tumour milieu can inducesensitization of cancer cells. In the context of the currentwork, it would be interesting to further elucidate the levels ofoxidative stress in CSC, non-CSC and wild-type cells andalso NIH3T3 control cell line under normal untreatedconditions versus Vitamin C and Niacin treatments. Hence,pre-exposure of cancer cells to oxidative agents furtherincreases the oxidative stress of cancer cells, as compared tonormal healthy cells. For example, exposure ofMCF-7 breastcancer cells with triethylenetetramine (TETA) enhanced theautooxidation of Vitamin C thereby inducing elevatedoxidative stress, and H2O2 production, finally leading to celldeath (Wang et al., 2017). With particular reference toascorbic acid, it is known as an antioxidant vitamin andhence, is non-toxic to normal healthy cells. Moreover, as theoxidative stress in cancer cells is already elevated tomoderately high levels, the cancer cells remain proliferative.However, a further increase in the oxidative stress in cancer

cells to very high concentrations would lead to cellularapoptosis. Hence, in our case, small doses (5–25mM) ofVitamin C and Niacin are inductive to cell proliferation ofCSCs, but high doses (100–1,000mM) of Vitamin C couldinduce cell death in CSC (Figures 6a–6d, 7a–7d, Tables 2 and3). Moreover, the Vitamin C and Niacin were tolerated wellby the normal fibroblastic cell line NIH3T3, and the highestconcentration of Vitamin C could induce only 15% of celldeath in this cell line (Figure 6e). Surprisingly, the highestdose (10,000mM) of Vitamin C-induced cell proliferation ofHCT-15- WT, CD44þ CSC and CD44� non-CSC to a greatextent to the order of �226, �300 and �230%, respectively(Table 2). This very high increase in cell proliferation in theHCT-15 cell line and its flow-sorted derivatives might be dueto further change in the cell energetics. Additional pathwaysinvolved in such shift in cellular energetics need furtherinvestigation.

Moreover, regarding the suggested dose of such dietaryingredients, not much information is available. Hence, it islikely that patients might consume from low to high doses ofsuch vitamins either in the form of natural products ordietary supplements. Also, cancer patients are oftenintravenously administered with high doses of Vitamin C.Very high dose (Gramme dose 0.5–5mM) of Vitamin Cupon intravenous injection reportedly causes increasedprocoagulant activity of freshly isolated human red bloodcells. The mechanism of such procoagulant activity involvesexternalisation of phosphatidylserine (PS) to the external cellmembrane and formation of PS-containing microvesicles(Kim et al., 2015). Also, intravenous injection of Vitamin Cof the order of 0.5–1 g/kg body weight induced thrombosisin rat models (Kim et al., 2015). However, in our case, thehighest dose used in 10,000mM (10mM) which is, hence notsafe to be administered intravenously to cancer patients.Also, as per an earlier report (Uetaki et al., 2015) LD50 ofHT-29 WT cells were achieved at Vitamin C dose >10mM(10,000mM). Accordingly, in this study, the high testeddoses from 100 to < 500mM (0.1–0.5mM) can still beconsidered safe for intravenous injection without posing anyrisk of thrombosis to the cancer patient. However, higherdoses of Vitamin C injected locally inside the tumour in-vivotumour xenograft models can be further tested for safety andprothrombin activity.

From our study, it is evident that small doses of Vitamin Cand Niacin are enhancing the cell proliferation of CD44þ

colon CSCs. Hence, it can be inferred that if a colorectalcancer patient, after the first phase of treatment involvingsurgery, chemotherapy and radiotherapy happen to con-sume small doses of Vitamin C and Niacin is likely to havehis/her cancer stem cell proliferate faster leading to a morerapid relapse of disease. This phenomenon of cancer relapsecan be due to the induction of moderate levels of reactiveoxygen species by the tested vitamins which are conducive to

Figure 8 Proposed scheme for the mechanism of action Niacintowards of cell proliferation versus cell killings, in response to, lowand high concentrations respectively. Low levels of Niacin (5–25mM)leading to the inhibition of PPARg, thereby resulting in the reduction inROS production and hence cell proliferation is represented on the left sideof the diagram. High concentrations of Niacin (100–10,000mM) leadingto the activation of PPARg, accumulation of ROS leading to cell death isthe proposed mechanism for the action of Niacin on colon CSCs.



the growth of CSC. Furthermore, very high levels of ROS areindeed inductive of apoptosis of cancer cells (Lee et al.,2017).

One interesting observation in this study involveddifferential killing rates of cancer stem cells, non-stemcancer cells and the mixed populations (WT) cells, inresponse to a high dose of Vitamin C and Niacin. The non-stem cancer cells from HT-29 cell line (CD44�) exhibitedvery high cell killings both with the maximum dose ofVitamin C and Niacin. The order of cell killings were �86and �89%, respectively, when treated with the highest doseof Vitamin C (14% cell proliferation) and Niacin (11% cellproliferation) compared to a baseline of 100% for untreatedcontrols) (Figures 6a, 6c, Tables 2 and 3). However, the non-stem cancer cells fromHCT-15 cell line (CD44�) did exhibitcell killing only up to 1,000mM of Vitamin C to the order of�12% (88% cell proliferation) (Figure 6b, Table 2). Thehighest dose (10,000mM) of Vitamin C, on the contrary, wasconducive to very high proliferation (�233%) in HCT-15/CD44� (Table 2). Moreover, the standard fibroblast cell lineNIH3T3 at a 10,000mM dose of Vitamin C also exhibited acell killing of the order of �20% (Figure 6d). Also, thehighest dose of Vitamin C (10,000mM) induced a cellproliferation to the extent of 226% in HCT-15 WT cells(Table 2). Moreover, the maximum dose of Niacin did notproduce extremely high levels of cell proliferation in any ofthe sorted derivatives from HCT-15 cell line (Table 3).

In the HT-29 CSCs moderate cell killing only �39% cellkilling (61.18% cell proliferation) with highest dose ofVitamin C (Figure 6a, Table 2) and �43% cell killing upontreatment with the maximum dose of Niacin (�56% cellproliferation) was observed (Figure 6c, Table 3). Also, HCT-15-CSCs exhibited comparatively lower cell killings �30%each with 1,000mM of Vitamin C and Niacin. Hence, themaximum cell killing in CSC by any dose of Vitamin C orNiacin did not attain an LD50. So, attainment of LD50 valuesin these CSC provides a scope for optimisation for thecombinatorial use of both the vitamins.

Wild-type populations could achieve high cell killingsonly inHT-29 cell line with themaximum dose of Vitamin C�73% (Cell proliferation �27%) (Figure 6a, Table 2) and�82% with the highest dose of Niacin (Cell proliferation�19.89%) (Figure 6c, Table 3). On the other hand, the HCT-15-WT cells attained maximum cell death of the order of�17% (Cell proliferation – 83.12%) with 1,000mMVitaminC and �13% with 1,000 and 10,000mM Niacin (Cellproliferation 87.12 and 87.06%, respectively) (Figures 6b, 6dand Tables 3 and 4). The cell killings of the order of>60% inHT-29 WT populations are possibly an indicator of apredominance of non-stem cancer cells/cell populations thatare sensitive to high doses of Vitamin C and Niacin.However, the WT and non-CSC fromHCT-15 cell line wereless sensitive as compared to their HT-29 counterparts that

again form an interesting observation with regards to, HCT-15 cell line that harbours KRAS mutation. Hence, molecularinteractions of mutated KRAS protein of HCT-15 cell lineand its cellular derivatives (WT, CSC and non-CSC), inresponse to various doses of Vitamin C and Niacin forms anew future scope of this study.

This phenomenon of differential cell killings is probablyan indicative of different cell survival mechanisms andcellular energetics of non-stem cancer cells and cancer stemcells. Cancer cells mainly rely on aerobic glycolysis, incontrast to oxidative phosphorylation in the case of somatic/healthy cells (Ooi and Gomperts, 2015; Xu et al., 2015).Moreover, transcriptome analyses of metastatic competentcolon cancer cell lines such as CTC-MCC-41 and HT-29harbouring high percentages of cancer stem cells haveindicated differential expression of energy metabolismgenes. Examples of energymetabolism genes are peroxisomeproliferator-activated receptor g coactivator 1B(PPARGC1B), peroxisome proliferator-activated receptorg coactivator 1A (PPARGC1A), fatty acid binding protein 1(FABP1), DNA repair [BRCA1 interacting protein C-terminal helicase 1 (BRIP1), aldehyde dehydrogenase threefamily member A1 (ALDH3A1), Fanconi anemia comple-mentation group B and M (FANCB, FANCM), stemnessglutaminase 2 (GLS2), cystathionine-beta-synthase (CBS)and cystathionine gamma-lyase (CTH) (Alix-Panabi�ereset al., 2016). Hence, targeting cancer stem metabolism hasbeen discussed as an effective strategy for cancer therapeu-tics in the review by Deshmukh et al. (2016). Accordingly,the results of the current work, regarding enhancedproliferation and cell killings of CSCs by respectively lowand high doses of Vitamin C and Niacin also has also a scopefor evaluation regarding various marker genes for energymetabolism. Differences in cellular proliferation/cell deathbehaviour of CSC, non-stem cancer cells and normal cellscan also be explained by probable differences in the intrinsicmetabolic reprogramming (Shen et al., 2015) and thepresence of certain programmed cell death ligands (Wuet al., 2017) that need further validation.

Mechanisms of targeting colon cancer cells having BRAFand KRAS mutations using high concentrations of VitaminC involve the inhibition of GAPDH thereby leading toenergy crisis has already been worked out by Yun et al.(2015). Moreover, the colorectal cancer cell line HT-29harbours only BRAF mutation and has a wild-type KRASgene. On the other hand, comparatively less sensitive cell lineHCT-15 has a mutated KRAS and wild-type BRAF. Themechanisms for the differences in sensitivities towards dosesof Vitamin C and Niacin in KRAS and BRAF mutated celllines are rather unknown. However, it is likely to furtherhigh cell killing can be achieved by treatment of the highdose of Vitamin C in the colorectal cancer cells harbouringmutations in both KRAS and BRAF. Also, there might be



additional mechanisms involved in cell survival of HCT-15colorectal cancer stem cells upon treatment with high dose ofVitamin C and Niacin.

Amongst both the vitamins, our results indicate a bettercell killing effect of Niacin on HT-19 CSC populations withthe highest individual dose of these two vitamins. Indeed,reports regarding cancer inhibitory effects of Niacin arescanty as compared to the reports on Vitamin C. Theremight be additional mechanisms for greater sensitivity ofCSCs to Niacin that needs to be elucidated. Nevertheless, wehave hypothesised pathway (Figure 8) involving themechanism of action of Niacin on colorectal CSC thatneeds further validation. Moreover, this observation opensthe scope for studying the GAPDH pathways of the energycrisis and various genes involved in cellular energetics andcell death upon treatment with concentrations of Niacin aswell in the HT-29 and HCT-15 colorectal cancer stem cells.

Finally, small doses of the vitamins, Vitamin C andNiacinare conducive towards cell proliferation of colon cancer stemcells as indicated in this study, in contrast to, not enoughliterature available regarding the dose-dependent behaviourof cancer/cancer stem cells towards Vitamin C and Niacin.However, dose responsive cellular response namely loss ofcardioprotection, an unrelated context, has been reportedabout the high dose of vitamin A (Csepanyi et al., 2015).

Conclusion

High doses of Vitamin C and Niacin-induced cell killings, incontrast to small doses of VitaminC andNiacin that inducedcell proliferation of colon cancer stem cells isolated fromHT-29 and HCT-15 colon cancer cell lines. The detailedmechanism for this different phenomenon needs to beworked out concerning cellular energetics upon low versusthe high dose of exposure to Vitamin C and Niacin.

Acknowledgements and funding

The authors would like to acknowledge the YenepoyaResearch Centre, YenepoyaUniversity for the infrastructure,core facility and funding to carry out this research. Authorsacknowledge the help of Mr Surya Prakash, Junior ResearchFellow for his contribution to imaging the live–dead stainedcells and 3D tumourspheres and Ms Deeksha K for MTTassay optimisation.

Authors’ contributions

BB and SS conceptualized and designed the project,performed cell sorting, 3D tumourosphere assay, analysedthe data, compiled the manuscript figures using AdobeIllustrator, wrote and approved the manuscript. US helpedin cell sorting, maintained the cells, carried out MTT

assays with HT-29 and HCT-15 colorectal carcinoma celllines and its flow-sorted sorted derivatives and NIH3T3cell line, performed live–dead cell staining experimentsand imaging.

Conflict of interest

The authors declare no conflict of interest.

References

Abdullah LN, Chow EKH (2013)Mechanisms of chemoresistancein cancer stem cells. Clin Transl Med 2(1): 1.

Abetov D, Mustapova Z, Saliev T, Bulanin D (2015) Biomarkersand signaling pathways of colorectal cancer stem cells. TumorBiol 36(3): 133.

Ahmed D, Eide PW, Eilertsen IA, Danielsen SA, Eknæs M,HektoenM, Lothe RA (2013) Epigenetic and genetic features of24 colon cancer cell lines. Oncogenesis 2(9): e71.

Alix-Panabi�eres C, Cayrefourcq L, Mazard T, Maudelonde T,Assenat E, Assou S (2016) Molecular portrait of metastasis-competent circulating tumor cells in colon cancer reveals thecrucial role of genes regulating energy metabolism and DNArepair. Clin Chem 2016: 700–13.

Bagheri V, Razavi MS, Momtazi AA, Sahebkar A, Abbasza-degan MR, Gholamin M (2017) Isolation, identification,and characterization of cancer stem cells: a review. J CellPhysiol 232(8): 2008–18.

Carter BZ,Mak PY,MuH, ZhouH,MakDH, SchoberW,Cortes J(2016) Combined targeting of BCL-2 and BCR-ABL tyrosinekinase eradicates chronic myeloid leukemia stem cells. SciTransl Med 8(355): 355ra117.

Chen L, So WY, Li SY, Cheng Q, Boucher BJ, Leung PS (2015)Niacin-induced hyperglycemia is partially mediated via niacinreceptor GPR109a in pancreatic islets. Mol Cell Endocrinol404: 56–66.

Csepanyi E, Czompa A, Haines D, Lekli I, Bakondi E, Balla G, BakI (2015) Cardiovascular effects of low versus high-dose beta-carotene in a rat model. Pharmacol Res 100: 148–56.

Cuhls H, Marinova M, Kaasa S, Stieber C, Conrad R, Radbruch L,M€ucke M (2016) A systematic review on the role of vitamins,minerals, proteins, and other supplements for the treatment ofcachexia in cancer: a European Palliative Care Research Centrecachexia project. J Cachexia Sarcopenia Muscle 8(1): 25–39.

Deshmukh A, Deshpande K, Arfuso F, Newsholme P, Dharmar-ajan A (2016) Cancer stem cell metabolism: a potential targetfor cancer therapy. Mol Cancer 15(1): 69.

Du J, Cullen JJ, Buettner GR (2012) Vitamin C: chemistry, biologyand the treatment of cancer. Biochim Biophys Acta 1826(2):443–57.

Du L, Wang H, He L, Zhang J, Ni B, Wang X, Chen Q (2008)CD44 is of functional importance for colorectal cancer stemcells. Clin Cancer Res 14(21): 6751–60.

Eduati F, Dold�an-Martelli V, Klinger B, Cokelaer T, Sieber A,Kogera F, Saez-Rodriguez J (2017) Drug resistance



mechanisms in colorectal cancer dissected with cell type-specific dynamic logic models. Cancer Res 77(12): 3364–75.

Fischer AP, Miles SL (2017) Vitamin C, but not dehydroVitaminC increases intracellular Vitamin C content to decreaseHypoxia Inducible Factor-1 alpha activity and reduce malig-nant potential in humanmelanoma. Biomed Pharmacother 86:502–13.

Foty R (2011) A simple hanging drop cell culture protocol forgeneration of 3D spheroids. J Vis Exp 51: e2720.

Gatti L, Cossa G, Beretta L, Zaffaroni G, Perego P (2011) Novelinsights into targeting ATP-binding cassette transporters forantitumor therapy. Curr Med Chem 18(27): 4237–49.

Guerrera F, Mossetti C, Ceccarelli M, Bruna MC, Bora G, OlivettiS, Oliaro A (2016) Surgery of colorectal cancer lungmetastases:analysis of survival, recurrence and re-surgery. J Thorac Dis8(7): 1764.

Hao J, Chen H, Madigan MC, Cozzi PJ, Beretov J, Xiao W,Delprado WJ, Russell PJ, Li Y (2010) Co-expression of CD147(EMMPRIN), CD44v3-10, MDR1 andmonocarboxylate trans-porters is associated with prostate cancer drug resistance andprogression. Br J Cancer 103: 1008–18.

Hastert TA, White E (2016) Association between meeting theWCRF/AICR cancer prevention recommendations and colo-rectal cancer incidence: results from the VITAL cohort. CancerCauses Control 27(11): 1347–59.

Jemal A, Siegel R, Ward E, Hao Y, Xu J, Thun MJ (2009) Cancerstatistics, 2009. CA Cancer J Clin 59(4): 225–49.

Kim K, Bae ON, Koh SH, Kang S, Lim KM, Noh JY, Chung JH(2015) High-dose Vitamin C injection to cancer patients maypromote thrombosis through procoagulant activation oferythrocytes. Toxicol Sci 147(2): 350–9.

Lee HY, Parkinson EI, Granchi C, Paterni I, Panigrahy D, Seth P,Hergenrother PJ (2017) Reactive oxygen species synergize topotently and selectively induce cancer cell death. ACS ChemBiol 12(5): 1416–24.

Longley DB, Johnston PG (2005) Molecular mechanisms of drugresistance. J Pathol 205(2): 275–92.

Lou H, DeanM (2007) Targeted therapy for cancer stem cells: thepatched pathway and ABC transporters. Oncogene 26(9):1357–60.

Luo S, Wang Z, Kansara V, Pal D, Mitra AK (2008) Activity of asodium-dependent vitamin C transporter (SVCT) in MDCK-MDR1 cells and mechanism of ascorbate uptake. Int J Pharm358(1): 168–76.

Marley AR, NanH (2016) Epidemiology of colorectal cancer. Int JMol Epidemiol Genet 7(3): 105.

Montserrat-de la Paz S, Naranjo MC, Lopez S, Abia R, MurianaFJG, Bermudez B (2017) Niacin and its metabolites as masterregulators of macrophage activation. J Nutr Biochem 39: 40–7.

Ooi AT, Gomperts BN (2015) Molecular pathways: targetingcellular energy metabolism in cancer via inhibition of SLC2A1and LDHA. Clin Cancer Res 21(11): 2440–4.

Padayatty SJ, Riordan HD, Hewitt SM, Katz A, Hoffer LJ, LevineM (2006) Intravenously administered Vitamin C as cancertherapy: three cases. CMAJ 174(7): 937–42.

Ricci-Vitiani L, Lombardi DG, Pilozzi E, Biffoni M, Todaro M,Peschle C, De Maria R (2007) Identification and expansion ofhuman colon-cancer-initiating cells. Nature 445(7123):111–5.

Sawant RR, Vaze OS, Wang T, D’Souza GG, Rockwell K, Gada K,Torchilin VP (2012) Palmitoyl ascorbate liposomes and freeascorbic acid: comparison of anticancer therapeutic effectsupon parenteral administration. Pharm Res 29(2): 375–83.

Shafi T, Jabeen I (2017) Grid-independent descriptors (GRIND)analysis and SAR guided molecular docking studies to probeselectivity profiles of inhibitors of multidrug resistance trans-porters ABCB1 and ABCG2. Curr Cancer Drug Targets 17(2):177–90.

Sharma M, Khan H, Thall PF, Orlowski RZ, Bassett RL, Shah N,Hosing CM (2012) A randomized phase 2 trial of a preparativeregimen of bortezomib, high-dose melphalan, arsenic trioxide,and ascorbic acid. Cancer 118(9): 2507–15.

Shen YA,WangCY,Hsieh YT, ChenYJ,Wei YH (2015)Metabolicreprogramming orchestrates cancer stem cell properties innasopharyngeal carcinoma. Cell Cycle 14(1): 86–98.

Sinicrope FA, Cleary KR, Stephens LC, Lee JJ, Levin B (1995) Bcl-2 and p53 oncoprotein expression during colorectal tumori-genesis. Cancer Res 55(2): 237–41.

Starr P (2015) Oral niacin prevents common skin cancers in high-risk patients, reduces costs. Am Health Drug Benefits 8(SpecIssue): 13.

Stender JD, Nwachukwu JC, Kastrati I, Kim Y, Strid T, Yakir M,Carlson KE (2017) Structural and molecular mechanisms ofcytokine-mediated endocrine resistance in human breastcancer cells. Mol Cell 65(6): 1122–35.

Su YJ, Lai HM, Chang YW, Chen GY, Lee JL (2011a) Directreprogramming of stem cell properties in colon cancer cells byCD44. EMBO J 30(15): 3186–99.

Su YJ, Lai HM, Chang YW, Chen GY, Lee JL (2011b) Directreprogramming of stem cell properties in colon cancer cells byCD44. EMBO J 30(15): 3186–99.

Uetaki M, Tabata S, Nakasuka F, Soga T, Tomita M (2015)Metabolomic alterations in human cancer cells by VitaminC-induced oxidative stress. Sci Rep 5: 13896.

Verma M (2012) Cancer control and prevention by nutrition andepigenetic approaches. Antioxid Redox Signal 17(2): 355–64.

Wahab SR, Islam F, Gopalan V, Lam AKY (2017) Theidentifications and clinical implications of cancer stem cellsin colorectal cancer. Clin Colorectal Cancer 16(2): 93–102.

Wang L, Luo X, Li C, Huang Y, Xu P, Lloyd-Davies LH, C Tong(2017) Triethylenetetramine synergizes with pharmacologicascorbic acid in hydrogen peroxide mediated selective toxicityto breast cancer cell. Oxid Med Cell Longev 2017: 3481710.https://doi.org/10.1155/2017/3481710

Wang W, Zhang G, Yang J, Gu H, Ding L, Yu H, Li M (2016)Digital gene expression profiling analysis of DNA repairpathways in colon cancer stem population of HT29 cells. ActaBiochim Biophys Sin 49(1): 90–100.

Wei X, Xu Y, Xu FF, Chaiswing L, Schnell D, Noel T, Clair WHS(2017) RelB expression determines the differential effects of



https://doi.org/10.1155/2017/3481710

ascorbic acid in normal and cancer cells. Cancer Res 77(6):1345–56.

Wu Y, Chen M, Wu P, Chen C, Xu ZP, Gu W (2017) IncreasedPD-L1 expression in breast and colon cancer stem cells. ClinExp Pharmacol Physiol 44(5): 602–4.

Xu X, Li J, Sun X, Guo Y, Chu D, Wei L, Zhang J (2015) Tumorsuppressor NDRG2 inhibits glycolysis and glutaminolysis incolorectal cancer cells by repressing c-Myc expression.Oncotarget 6(28): 26161.

Yao T, Asayama Y (2017) Animal-cell culture media: history,characteristics, and current issues. Reprod Med Biol. 16:99–117. https://doi.org/10.1016/j.jnutbio.2016.09.008

Yang Y, Sauve AA (2016) NADþ metabolism: bioenergetics,signaling and manipulation for therapy. Biochim Biophys Acta1864(12): 1787–1800.

Yazilitas D, €Ozdemir N, Hocazade C, Demirci NS, Zengin N(2016) The clinical and pathological features affecting the timeof relapse in patients with early stage colorectal cancer.J Cancer Res Ther 12(4): 1257.

Yu L, Fan Z, Fang S, Yang J, Gao T, Sim~oes BM, Clarke RB (2016a)Cisplatin selects for stem-like cells in osteosarcoma byactivating notch signaling. Oncotarget 7(22): 33055–68.

Yu WK, Wang Z, Fong CC, Liu D, Yip TC, Au SK, Yang M(2016b) Chemoresistant lung cancer stem cells display high

DNA repair capability to remove cisplatin-DNA damage. Br JPharmacol 174(4): 302–13.

Yun J,Mullarky E, Lu C, Bosch KN, Kavalier A, Rivera K,Muley A(2015) Vitamin C selectively kills KRAS and BRAF mutantcolorectal cancer cells by targeting GAPDH. Science 350(6266):1391–6.

Zhou JY, ChenM,Ma L,Wang X, Chen YG, Liu SL (2016) Role ofCD44high/CD133high HCT-116 cells in the tumorigenesis ofcolon cancer. Oncotarget 7(7): 7657.

Zhou S, Schuetz JD, Bunting KD, Colapietro AM, Sampath J,Morris JJ, Sorrentino BP (2001) The ABC transporter Bcrp1/ABCG2 is expressed in a wide variety of stem cells and is amolecular determinant of the side-population phenotype. NatMed 7(9): 1028–34.

Received 16 January 2017; accepted 26 July 2017.Final version published online 24 August 2017.

Supporting Information

Additional Supporting Information should be found in theonline version of this article at the publisher’s web-site.

Table S1. List and sequence of primers.



https://doi.org/10.1016/j.jnutbio.2016.09.008

Opposingeffectsoflowversushighconcentrationsofwatersoluble ...Cancer stem cells (CD44þ) and non-stem cancer cells (CD44 ) cells were sorted and collected in FBS containing media followed

Documents