HeparanSulfateRegulatesHairFollicleandSebaceousGland ... · growth and hair shaft production. The absence of HS also impeded the formation of catagen follicles in EXT1StEpi /StEpi

Heparan Sulfate Regulates Hair Follicle and Sebaceous GlandMorphogenesis and Homeostasis*

Received for publication, April 11, 2014, and in revised form, July 18, 2014 Published, JBC Papers in Press, July 22, 2014, DOI 10.1074/jbc.M114.572511

Vivien Jane Coulson-Thomas‡1, Tarsis Ferreira Gesteira‡§, Jeffrey Esko¶, and Winston Kao‡

From the ‡Department of Ophthalmology, University of Cincinnati, Cincinnati, Ohio 45267-0838, §Division of DevelopmentalBiology, Cincinnati Children’s Hospital and Research, Cincinnati, Ohio 45229-3039, and ¶Department of Cellular and MolecularMedicine, Glycobiology Research and Training Center, University of California, San Diego, La Jolla, California 92093-0687

Background: Skin appendages play a vital role in keeping the skin healthy.Results: The loss of heparan sulfate (HS) induces morphogenesis of skin appendages in mature skin.Conclusion: HS plays an inhibitory role in skin appendage formation in mature skin.Significance: Manipulation of HS levels in skin could enable the formation of skin appendages after skin injuries and in agedskin.

Hair follicle (HF) morphogenesis and cycling are a resultof intricate autonomous epithelial-mesenchymal interactions.Once the first HF cycle is complete it repeatedly undergoescyclic transformations. Heparan sulfate (HS) proteoglycans arefound on the cell surface and in the extracellular matrix wherethey influence a variety of biological processes by interactingwith physiologically important proteins, such as growth factors.Inhibition of heparanase (an HS endoglycosidase) in in vitro cul-tured HFs has been shown to induce a catagen-like process.Therefore, this study aimed to elucidate the precise role of HS inHF morphogenesis and cycling. An inducible tetratransgenicmouse model was generated to excise exostosin glycosyltrans-ferase 1 (Ext1) in keratin 14-positive cells from P21. Interest-ingly, EXT1StEpi�/StEpi� mice presented solely anagen HFs.Moreover, waxing the fur to synchronize the HFs revealed accel-erated hair regrowth in the EXT1StEpi�/StEpi� mice and hinderedcycling into catagen. The ablation of HS in the interfollicularepidermal cells of mature skin led to the spontaneous forma-tion of new HFs and an increase in Sonic Hedgehog expres-sion resembling wild-type mice at P0, thereby indicating thatthe HS/Sonic Hedgehog signaling pathway regulates HF for-mation during embryogenesis and prevents HF formation inmature skin. Finally, the knock-out of HS also led to the mor-phogenesis and hyperplasia of sebaceous glands and sweatglands in mature mice, leading to exacerbated sebum produc-tion and accumulation on the skin surface. Therefore, ourfindings clearly show that an intricate control of HS levels isrequired for HF, sebaceous gland, and sweat gland morpho-genesis and HF cycling.

The hair follicle is a skin appendage that is a compositemicro-organ of both epithelial and dermal origin. The follicle iscomposed of various compartments, namely the outer root

sheath (ORS),2 inner root sheath (IRS), hair shaft, and extracel-lular matrix, which are all of epithelial origin, and the dermalpapilla and connective tissue sheath, which are both of mesen-chymal origin (1). The IRS and ORS are two concentric epithe-lial cell layers (keratinocytes) that surround the hair follicle.The bulb is located at the proximal end of the hair follicle andcontains melanocytes and proximal cells from the ORS. Thebulge is a convex extension of the distal ORS and containsthe hair follicle stem cells. The dermal papilla is composed ofclosely packed mesenchymal cells and is engulfed by the bulbduring anagen. The connective tissue sheath is composed offibroblasts, macrophages, and connective tissue tightlyattached to the outer side of the hair follicle.

The hair follicle develops as a result of intricate epithelial-mesenchymal interactions between epidermal keratinocytescommitted to hair-specific differentiation and a cluster of der-mal fibroblasts that form the follicular papilla (2). Hair folliclemorphogenesis occurs in three waves, each giving rise to thespecific types of fur, and all are completed by postnatal day 20(3, 4). Thereafter, the budding of new hair follicles ceases, andthe existent hair follicles undergo spontaneous hair cycling in awavelike manner. Hair follicle morphogenesis commences withthe appearance of epithelial placodes at embryonic day E14.5that develop into primary guard hair follicles comprising 1–5%of the adult mouse coat. At embryonic day E16.5, the secondwave of placode formation is initiated; these develop into awland auchene hairs that account for �20% of the adult mousecoat. The second wave of placodes develop with an even distri-bution between the established guard follicles. Finally, atembryonic day E18.5, the third and final wave of placode for-mation begins; these develop into the zigzag hairs comprisingthe bulk of the adult coat (3, 4). All mouse hair follicle typeshave the same basic arrangement.

* This work was supported, in whole or in part, by National Institutes of HealthGrant RO1 EY011845 from the NEI.

1 To whom correspondence should be addressed: John van Geest Centre forBrain Repair, Robinson Way, Cambridge CB2 0PY, UK. Tel.: 447841656102;E-mail: [email protected] or [email protected].

2 The abbreviations used are: ORS, outer root sheath; HF, hair follicle; Ext1,exostosin glycosyltransferase 1; K14, keratin 14; IRS, inner root sheath; HS,heparan sulfate; SHH, Sonic Hedgehog; BMP, bone morphogenetic pro-tein; EDA, ectodysplasin A; HSPG, heparan sulfate proteoglycan; MPS,mucopolysaccharidosis; GAG, glycosaminoglycan; K15, keratin 15; StEpi,stratified epithelium; ECM, extracellular matrix; mG, membrane-targetedenhanced GFP; mT, membrane-targeted red fluorescence.

THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 289, NO. 36, pp. 25211–25226, September 5, 2014© 2014 by The American Society for Biochemistry and Molecular Biology, Inc. Published in the U.S.A.

SEPTEMBER 5, 2014 • VOLUME 289 • NUMBER 36 JOURNAL OF BIOLOGICAL CHEMISTRY 25211

Once the first hair cycle is completed, hair follicles repeatedlyundergo cyclic transformations of rapid growth involving hairshaft production (anagen), apoptosis-driven regression (cata-gen), and quiescence of mature hair follicles (telogen) (4 – 6). Atthe onset of anagen, enhanced proliferation of stem cells (ofepithelial origin) located within the bulge region leads to thedownward growth of the hair follicles, leading to the formationof the hair shaft, which requires enzymatic degradation of thesurrounding extracellular matrix (7). Both hair follicle devel-opment and cycling rely on autonomous intricate epitheli-um-mesenchyme interactions controlled by similar signal-ing networks involving bone morphogenetic protein (BMP),transforming growth factor (TGF), epidermal growth factor(EGF), fibroblast growth factor (FGF), Sonic Hedgehog (SHH),tumor necrosis factor (TNF), ectodysplasin A (EDA), and Wntfamilies (8 –16). EDA, which is from the TNF superfamily, isinvolved in hair follicle morphogenesis and hair follicle cyclingand has been interposed downstream of inductive Wnt signal-ing (15). EDA is also required for scale formation in fish andsweat gland development in mammals (15). The growth factorsBMP, TGF, EGF, FGF, SHH, TNF, Wnt, and EDA have all beenshown previously to be modulated by heparan sulfate (HS)(17–20).

Heparan sulfate proteoglycans (HSPGs) are found on the cellsurface and in the extracellular matrix where they influence avariety of biological processes by interacting with physiologi-cally important proteins, such as growth factors, chemokinesand cytokines, extracellular matrix proteins, enzymes, andenzyme inhibitors (21). Their activity is due at large to the pat-tern of sulfated sugar residues along the HS chains covalentlybound to the core proteins of the proteoglycans (22–25). Theinteraction of growth factors with HS protects the growth fac-tors from degradation, creates a storage pool, acts as a co-re-ceptor facilitating the assembly of signaling complexes, regu-lates growth factor diffusion throughout the tissue, and enablesclearance of the growth factors by endocytosis (21, 26). Hepa-ranase is an endoglycosidase that cleaves HS, enabling therelease of growth factors bound to HS chains, removal of thephysical barrier imposed by HS, and interruption of HS-medi-ated cell-cell and cell-matrix contacts (27). Therefore, a delicatebalance is required between the expression of both HS andheparanase. Previous studies have shown that heparanaseexpression varies throughout hair follicle cycling withincreased expression in anagen hair follicles and decreasedexpression in catagen follicles (28). Moreover, the overexpres-sion of heparanase accelerates the rate of anagen, and theauthors speculate that the release of growth factors upon hepa-ranase cleavage would play a major role (29). Herein we showthat it is the absence of HS within the hair follicle that triggersthe enhanced rate of anagen and not the release of growth fac-tors upon heparanase digestion. Moreover, we show that HSplays a vital role in hair follicle morphogenesis. Further indica-tions of the important role HS plays in hair follicle homeostasisis that hair abnormalities are solely present in mucopolysaccha-ridosis types with accumulation of HS, such as MPS I, MPS II,and MPS III; however, no hair abnormalities are found in MPSIV or MPS VI in which HS catabolism is not affected (30).

To study the role of HS in hair follicle homeostasis andcycling, a transgenic mouse model was generated to knock outexostosin glycosyltransferase 1 (Ext1) using a tissue-specificinducible system. Ext1 was knocked out of cells of epithelialorigin using a keratin 14 (K14)-rtTA driver system after thecompletion of the first hair cycle, generating EXT1StEpi�/StEpi�

mice (excision of Ext1 in stratified epithelium). Therefore, withthe Ext1 knock-out system, HS was specifically ablated from theORS, IRS, and hair shaft of these mice but remained in the hairfollicle components of dermal origin. Interestingly, the ablationof HS in the hair follicle presents a more severe phenotype whencompared with the heparanase overexpression mouse model byaccelerating anagen, enhancing active hair growth, and ena-bling faster hair regrowth upon waxing, thereby suggesting thatHS plays an important barrier function during rapid hair folliclegrowth and hair shaft production. The absence of HS alsoimpeded the formation of catagen follicles in EXT1StEpi�/StEpi�

mice, resulting in sequestration of all hair follicles in anagenupon cycling. Moreover, EXT1StEpi�/StEpi� mature miceinduced at all time points analyzed (P20, P23, P25, and P35)presented spontaneous hair follicle budding upon induction,revealing the important role HS plays in both hair follicle mor-phogenesis and homeostasis. Both manipulations of the HFcycling and the ability of inducing morphogenesis present greatpharmaceutical potential.

EXPERIMENTAL PROCEDURES

Mouse Strains and Genotyping—Transgenic mouse linesK14-rtTA (stock number 008099) (31), tetO-cre (TC) (stocknumber 006224) (32), and RosamTmG/mTmG (stock number007576) (33) were purchased from The Jackson Laboratory (BarHarbor, ME). The floxed mice utilized were Ext1flox/flox (34).Compound transgenic mice were generated by mating. All themice were bred at the Animal Facility of the University of Cin-cinnati Medical Center. Experimental procedures for handlingthe mice were approved by the Institutional Animal Care andUse Committee, University of Cincinnati/College of Medicine.The identification of each transgene allele was performed byPCR genotyping with tail DNA.

Induction by Administration of Doxycycline Chow—Admin-istration of doxycycline chow was utilized to induce K14-drivenpersistent and irreversible excision of Ext1 in tetratransgenicmice (K14-rtTA/TC/RosaLSL/Ext1). Transgenic mice at P20or older were fed with doxycycline chow (1 g of doxycycline/kgof chow; Custom Animal Diets LLC, Bangor, PA) ad libitum.Control animals were either double transgenic or triple trans-genic heterozygous littermates.

Skin Collection—Skin samples were obtained fromEXT1StEpi�/StEpi� mice and control littermates. Rectangularpieces of skin from the left and right sides of both the dorsal andabdominal regions were collected parallel to the vertebral line(�3.5 � 1 cm, length � width) and processed for further par-affin or cryoembedding or protein extraction.

Agarose Gel Electrophoresis—The skin samples were mincedin acetone and centrifuged. The precipitate was dried and sub-jected to proteolysis with subtilisin (Sigma-Aldrich), and pro-tein products were removed by trichloroacetic (TCA) acid (Sig-ma-Aldrich) precipitation. The GAGs were then precipitated

HS Regulates Hair Follicle Morphogenesis and Cycling

25212 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 289 • NUMBER 36 • SEPTEMBER 5, 2014

overnight with methanol followed by centrifugation. The skinlevels of GAG were first analyzed by 0.6% agarose gel electro-phoresis in 0.05 M propanediamine acetate buffer, pH 9 asdescribed previously (35). Following electrophoresis, the gelswere submerged in 0.2% Cetavlon (cetyltrimethylammoniumbromide, Sigma-Aldrich) for 1 h at room temperature; dried;stained with 0.1% toluidine blue prepared in a solution of 1%acetic acid, 50% ethanol, and 49% water; and destained with thesame solution without toluidine blue.

Dimethylmethylene Blue Assay—To quantify HS levels,chondroitin sulfate and dermatan sulfate were removed fromtotal skin GAGs by Chondroitinase ABC (Sigma) digestion fol-lowed by centrifugation using Microcon� centrifugal filters(10,000; Millipore). The HS content was then measured usingdimethylmethylene blue reagent. The HS content was calcu-lated using a standard curve prepared with porcine mucosa HS(Neoparin, Alameda, CA).

Histochemistry—Tissues were fixed for 12 h in 4% bufferedparaformaldehyde, washed five times with PBS, sequentiallydehydrated, immersed in paraffin overnight, and subsequentlymounted. The blocks were sectioned at 7 �m, and sections werecollected on poly-L-lysine-treated slides. Upon use, tissue sec-tions were deparaffinized, rehydrated in PBS, and stained withhematoxylin and eosin. Images were captured using a ZeissObserver Z1 inverted microscope or Zeiss LSM-710 confocalmicroscope, and images were analyzed using LSM ImageBrowser 3.2 software (Zeiss, Germany).

Immunohistochemistry—Tissues were fixed for 12 h in 4%buffered paraformaldehyde, treated for 15 min in 0.1% sodiumborohydride, and embedded in Tissue-Tek� embedding mediumfor cryosectioning. 10-�m sections were cut using a cryostat(Cryostar NX70, Thermo Scientific) and collected on Fisher-brand� Superfrost�Plus Gold microscope slides (Thermo Scien-tific). Upon use, sections were incubated for 30 min at 37 °C, andexcess tissue embedding medium was removed with PBS. Non-specific protein binding sites were blocked with 5% fetal bovineserum (FBS). Sections were then incubated with primary anti-body anti-K14 (Covance, PRB-155P), anti-keratin 15 (K15)(Thermo Scientific, MS-1068), anti-syndecan 1 (Abcam,ab34164), anti-syndecan 2 (Abcam, ab79978), anti-syndecan 3(Abcam, ab63932), anti-syndecan 4 (Abcam, ab24511), anti-EXT1 (HPA044394, Atlas), anti-EXT1 (TA323730, Origene),anti-HS (clone 10E4, US Biological) anti-Wnt1 (Abcam,ab15251), anti-Wnt2 (Abcam, ab27794), anti-BMP4 (Millipore,MAB1049), anti-EDA, anti-SHH (Sigma, AV44235), or anti-�-catenin (Cell Signaling Technology, 9582). Sections werewashed and incubated with appropriate secondary antibodiesproduced in donkey labeled with Alexa Fluor� 647 (Invitrogen)for 1 h at 18 °C. Subsequently, sections were washed, and nucleiwere stained with DAPI. Sections were mounted in Fluoro-mount-G� (Electron Microscopy Sciences). Images were cap-tured using a Zeiss Observer Z1 inverted microscope or ZeissLSM-710 confocal microscope, and images were analyzed usingLSM Image Browser 3.2 software (Zeiss).

Oil Red O Staining—For the detection of neutral lipids, cryo-sections were washed with PBS and then stained with 0.5% oilred O (Sigma) for 15 min, rinsed with PBS, and counterstainedwith hematoxylin. Sections were mounted in FluoromountG.

Images were captured using a Nikon Eclipse E800 microscopecoupled with Axiocam ICc5 and processed using Axiovision 4.8(Zeiss).

Induction of Hair Cycle—The left dorsal fur of mice wasremoved using unscented commercial hair-removing waxstrips, leading to the synchronized development of anagen hairfollicles. Skin tissue samples were collected 5, 7, and 25 daysafter hair removal. Five mice were used per group for each timepoint. The samples were fixed in 4% paraformaldehyde andprocessed for histology and immunohistochemistry.

Iodine Sweat Test—Iodine/alcohol (1 g of iodine/100 ml ofethanol) was smeared on the plantar surface of the rear pawwith the use of a paint brush and air-dried. Thereafter, a starchoil suspension (50 g of starch/100 ml of castor oil) was smearedover the iodine solution, and black dots formed, revealing theopening of sweat glands. Images were captured 5 s after thestarch oil suspension was applied to the foot pad.

Protein Extraction from Skin—Skin samples were homoge-nized in modified radioimmune precipitation assay lysis buffer(50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% Triton X-100, 1%sodium deoxycholate, 0.1% SDS, 1 mM EDTA) containing 1 mM

phenylmethylsulfonyl fluoride and 1� protease inhibitor mix-ture (Sigma, P8340). The samples were homogenized on ice andthen centrifuged at 13,000 � g for 10 min at 4 °C to obtain asoluble fraction. Protein content was assayed using the Brad-ford method (36). Western blot analysis was performed asdescribed previously (37, 38). Briefly, protein (25 �g) was sep-arated on a gradient SDS-polyacrylamide gels (4 –20%) by SDS-PAGE and transferred by electrical current to polyvinylidenedifluoride (PVDF) membranes. The membranes were devel-oped with anti-syndecan 1, anti-syndecan 3, anti-syndecan 4,anti-SHH, and anti-�-catenin followed by secondary antibod-ies coupled with Alexa Fluor 555. For syndecan, Western blot-ting analysis samples were previously treated with heparinase II(Hepase II) (IBEX) and heparinase III (Hepase III) (IBEX) for24 h at 37 °C. A loading control was performed with goat anti-�-actin followed by anti-goat secondary antibody coupled withAlexa Fluor 488.

Statistical Analysis—All values are presented as mean � S.D.The difference between two groups was compared by unpairedMann-Whitney test. p � 0.05 was considered to be statisticallysignificant. Statistical analysis was performed with theGraphPad Prism version 5 software package (GraphPad Soft-ware, San Diego, CA).

RESULTS

Ext1 Excision of the Stratified Epithelium and Hair Follicle—Tetratransgenic mice (K14-rtTA/tetO-cre-EXT1flox-mTmG)were generated to specifically excise Ext1 upon doxycyclineinduction in keratin 14-expressing cells; hence, HS was notexpressed in the stratified epithelium (StEpi), generatingEXT1StEpi�/StEpi� mice. Prior to doxycycline induction, theK14-rtTA/tetO-cre-EXT1flox-mTmG mice exhibited nomacroscopic or microscopic abnormalities of the skin.EXT1StEpi�/StEpi� mice at P60 began to lose weight, and by P120,the mice showed severe signs of dehydration and had to beculled.



K14 Expression in the Hair Follicle—K14 staining was per-formed to determine the population of hair follicle cells thatexpress K14 (Fig. 1). Anti-K14 staining of telogen hair folliclesrevealed that cells in the ORS, IRS, IRS cone, hair bulb, andsebaceous gland all express K14 (Fig. 1). A similar K14 expres-sion pattern was observed in developing hair follicles and ana-gen hair follicles in P0 and P20 mice. The hair follicles from theskin of EXT1StEpi�/StEpi� mice at P55 also showed K14 stainingin the ORS, IRS, IRS cone, hair bulb, and sebaceous gland (Fig.1). Therefore, all of the hair follicle components of epithelialorigin express K14, and thus, in our mouse model, these cellswould lack Ext1 and consequently be HS-deficient. EXT1 stain-ing was also performed on the skin of EXT1StEpi�/StEpi� and

littermate control mice at P55, revealing a loss in EXT1 expres-sion in the hair follicles of EXT1StEpi�/StEpi� mice (Fig. 1D).

To further evaluate which components of the hair follicle ofEXT1StEpi�/StEpi� mice would lack HS, the mice were bred withthe reporter mTmG gene, which expresses strong red fluores-cence in all tissues; however, in cells expressing Cre recombi-nase, the mT cassette is deleted, enabling the expression of thedownstream cassette, membrane-targeted enhanced GFP(mG). Therefore, in all cells that present K14-driven Crerecombinase activity, both the floxed Ext1 gene and mT cas-sette will be deleted, and cells will consequently express mem-brane-bound GFP. The reporter mG expression driven by K14expression presented an expression pattern similar to that ofanti-K14 staining wherein the IRS, IRS cone, hair bulb, andsebaceous gland all presented strong GFP expression (Fig. 2),further reinforcing that the EXT1StEpi�/StEpi� mouse modellacks HS expression in the hair follicle in all cells of epithelialorigin. Therefore, based on reporter expression, the hair folli-cles of EXT1StEpi�/StEpi� mice lack HS in the IRS, IRS cone, andhair bulb but retain normal HS expression in the hair folliclecomponents of mesenchymal origin (Fig. 2). HS staining wasfurther performed using anti-HS (clone 10E4) confirming theknock-out of HS in the hair follicle components of epithelialorigin (Fig. 2). Moreover, GAGs were extracted from the skin ofEXT1StEpi�/StEpi� and littermate control mice and subjected toChondroitinase ABC treatment followed by dimethylmethy-lene blue assay, which further confirmed the decrease in skinHS (Fig. 1C). Moreover, the extracted GAGs were also sub-jected to agarose gel electrophoresis, which revealed nochanges in the skin levels of dermatan sulfate betweenEXT1StEpi�/StEpi� and littermate control mice (Fig. 1B).

Role of HS in Hair Follicle Morphogenesis and Cycling—His-tology of the EXT1StEpi�/StEpi� mice induced from P20 to P55revealed a 4-fold increase in the number of hair follicles com-pared with littermate controls (Fig. 3, a, b, c, and e). Thus, hairfollicle morphogenesis occurs in EXT1StEpi�/StEpi� mice uponinduction at P20. The drastic increase in overall hair folliclenumber in animals induced at all time frames led to the irregu-lar spacing of hair follicles with some fused hair follicles thatmerged at the outer epithelial layers; however, each individualhair follicle possessed its own hair bulb, dermal papilla, hairshaft, and sebaceous gland. EXT1StEpi�/StEpi� mice beyond P55presented such a dense distribution of hair follicles that three orfour were fused together in some locations. Moreover, in vivo con-focal microscopy of the abdomen of the mice revealed an overall4-fold increase in the quantity of fur in the EXT1StEpi�/StEpi� micewhen compared with the littermate control (Fig. 3, f and g).These data indicate that the absence of HS in the epitheliumleads to the budding of hair follicles in the mature skin of mice.

Interestingly, the EXT1StEpi�/StEpi� mice at all time framessolely presented anagen hair follicles; however, the littermatecontrol mice presented telogen hair follicles at the same timeframe (Fig. 3). Previous studies have shown that the first mousefur cycle starts around P28 and is concluded by approximatelyP49, and the second fur cycle commences around P84 (1).Therefore, mice analyzed at P55 should present mainly telogenhair follicles as observed in the littermate controls. This sug-gests that the lack of HS in hair follicles impedes cycling into

FIGURE 1. A, K14 staining (red) can be observed in the bulb, ORS, and IRS ofhair follicles and in the sebaceous gland from littermate control mice at P0(panel a), P20 (panel b), and P55 (panel c) and EXT1StEpi�/StEpi� mice at P55(panels d and e). Nuclei were stained with DAPI (blue). Images were capturedusing a Zeiss LSM-710 confocal microscope. B, the dermatan sulfate level inthe skin of EXT1StEpi�/StEpi� and littermate control mice was determined byagarose gel electrophoresis of skin GAG extracts. C, the HS level in the skin ofEXT1StEpi�/StEpi� and littermate control mice was determined by dimethyl-methylene blue assay after Chondroitinase ABC treatment of skin GAGextracts. D, panels a– c, loss of EXT1 staining can be observed in the hair follicleof EXT1StEpi�/StEpi� mice (b and c) compared with littermate control (a). Scalebars, 20 �m. Error bars represent S.D. CS, chondroitin sulfate; DS, dermatansulfate.



the catagen phase and consequently the telogen phase inEXT1StEpi�/StEpi� mice.

Syndecan and Perlecan Expression during Hair Follicle Mor-phogenesis and Cycling—Previous studies have shown that hairfollicles express syndecan 1. To elucidate the expression profileof syndecan 1 in hair follicles, anti-syndecan 1 staining wasperformed on the skin of control mice at P0 and P20. Duringhair follicle morphogenesis (P0), high levels of syndecan 1expression were observed in the stroma proximal to the epithe-lium; hence, syndecan 1 could play a role in stroma-epitheliumcross-talk during hair follicle formation (Fig. 4A, panels a andb). Moreover, hair follicles at P0 and P20 also presented synde-can 1 expression in the IRS and ORS (Fig. 4A, panels a– d). ByP20, there was a drastic decrease in stromal syndecan 1 expres-sion, which was present solely adjacent to the epithelium and inthe ORS (Fig. 4A, panels c and d). In the mature hair follicles ofthe littermate control at P55, syndecan 1 expression was

restricted to the ORS and stroma adjacent to the epithelium(Fig. 4A, panels e and f). Interestingly, EXT1StEpi�/StEpi� miceinduced from P20 to P55 presented strong stromal syndecan 1expression similar to that observed in the P0 mice (Fig. 4A,panels g–j). Moreover, there was a 6-fold increase in syndecan 1expression in the ORS of EXT1StEpi�/StEpi� mice compared withthat of control mice at P0, P20, and P55 (quantification ofimmunofluorescence staining performed using Zen, Zeiss). Anincrease of 2-fold in syndecan 1 expression was further con-firmed by Western blotting analysis (Fig. 4, B and C).

The expression of syndecans 2, 3, and 4 was also investigatedin the hair follicles of EXT1StEpi�/StEpi� mice. The lack of HS inthe EXT1StEpi�/StEpi� mice led to an induction/increase in theexpression of syndecan 3 in the ORS and sebaceous glands ofEXT1StEpi�/StEpi� mice (Fig. 5A, panels g–j). Western blottinganalysis confirmed a 2-fold increase of syndecan 3 expression inEXT1StEpi�/StEpi� mice when compared with littermate con-

FIGURE 2. HS staining (white) was performed with anti-HS (clone 10E4) on skin cryosections from littermate control mice at P0 (a, b, and c), P20 (d, e,and f), and P55 (g, h, and i) and EXT1StEpi�/StEpi� mice at P55 (j, k, and l). Mice were bred with the reporter RosamTmG gene; thereby K14-positive cells thatexpressed Cre recombinase upon induction express mG representing EXT1-deficient cells in EXT1StEpi�/StEpi� mice. EXT1StEpi�/StEpi� mice present HS stainingsolely in the mT-positive cells, and therefore, Cre recombinase successfully excised Ext1 from the cells of epithelial origin. The interfollicular epithelium (Epi) isoriented upward in the images. a, d, g, and j, merged mT, mGFP, HS, and DAPI images. b, e, h, and k, HS and DAPI. c, f, i, and l, HS and mT. Nuclei were stainedwith DAPI (blue). Scale bars, 40 �m.



trols (Fig. 5, B and C). Syndecan 4 expression was not present incontrol mice at P0; however, weak staining was present in theORS and IRS at P20 and P55. EXT1StEpi�/StEpi� mice also pre-sented weak syndecan 4 expression in the ORS and IRS at P55(Fig. 6). No significant syndecan 2 expression was observed inany of the samples analyzed (results not shown).

Strong perlecan staining was present in the stroma of controlmice at P0; by P20, the expression decreased and was solelypresent in the stroma cells adjacent to the hair follicles; and byP55, there was an overall decrease in perlecan staining.EXT1StEpi�/StEpi� mice presented strong perlecan staining sim-ilar to that observed in littermate control mice at P0 (Figure 7).The perlecan staining colocalized with the mT-expressing cellssurrounding the hair follicles.

Hair Follicle Differentiation—K15 is at large considered to bea marker for hair follicle cells (39). To further investigate therole of HS in hair follicle differentiation, and consequentlycycling, anti-K15 staining was performed. Hair follicles at P0presented faint K15 staining along the hair follicle shaft; how-ever, by P20, the hair follicles displayed dense K15 staining inthe ORS, IRS, and sebaceous gland (Fig. 8, a– d). The telogenhair follicles in the littermate control mice also presentedintense K15 staining in the ORS, IRS, bulge, and sebaceousgland (Fig. 8, e and f). In the EXT1StEpi�/StEpi� mice inducedfrom P20 to P55, there was an overall loss of K15 expression inthe hair follicles with subtle staining solely in the ORS and some

FIGURE 3. Skin sections from littermate control (a and c) and EXT1StEpi�/StEpi�

mice (b and e) at P55 were double stained with hematoxylin and eosin andvisualized under a light microscope. Telogen hair follicles can be observed inthe skin sections from littermate control mice, and anagen hair follicles can beobserved in the skin sections from EXT1StEpi�/StEpi� mice. EXT1StEpi�/StEpi� miceat P55 present excessive accumulation of sebum on the skin and fur (d) due toan increase in the number and size of sebaceous glands. f and g, in vivo con-focal microscopy of the mouse abdomen was performed to assess the furdensity of littermate control (f) and EXT1StEpi�/StEpi� (g) mice at P55. An overallincrease in hair density number was observed in EXT1StEpi�/StEpi� mice com-pared with littermate controls.

FIGURE 4. A, syndecan 1 staining (white) was performed in skin cryosectionsfrom littermate control mice at P0 (panels a and b), P20 (panels c and d), andP55 (panels e and f) and EXT1StEpi�/StEpi� mice at P55 (panels g–j). The interfol-licular epithelium (Epi) is oriented upward in the images. Control mice at P0and P20 present syndecan 1 expression in the stroma adjacent to the epithe-lium, ORS, IRS, and bulb; however, mature skin with telogen hair follicles pres-ents syndecan 1 distributed throughout the interfollicular epithelium andORS. EXT1StEpi�/StEpi� mice present syndecan 1 distribution similar to that of P0and P20 control mice; however, a 20-fold increase is observed in the ORS.Panels a, c, e, g, and i, merged mT (red), mGFP (green), syndecan 1, and DAPIimages. Panels b, d, f, h, and j, syndecan 1 and DAPI. Nuclei were stained withDAPI (blue). Scale bars, 40 �m. B, protein was extracted from the skin of EXT1f/f

and EXT1StEpi�/StEpi� mice, submitted to separation by SDS-PAGE, and trans-ferred by current to a PVDF membrane. The membrane was developed withanti-syndecan 1 followed by anti-rabbit secondary antibody conjugated withAlexa Fluor 555. C, the syndecan 1-positive band from the Western blot wasquantified, and overall content was divided by that quantified for �-actin,which was developed using anti-�-actin followed by anti-goat secondaryantibody conjugated with Alexa Fluor 488. Results are presented as a graph.Error bars represent S.D. RelUF, relative units of fluorescence.



newly developing sebaceous glands (Fig. 8, g–i). Therefore, thelack of HS in the EXT1StEpi�/StEpi� mice leads to altered differ-entiation of the hair follicle cells. This supports the notion thatthe lack of HS leads to altered differentiation of anagen hairfollicles that could hinder the cycling into catagen.

Wnt1, EDA, and SHH Signaling during Hair Follicle Morpho-genesis and Cycling—In an attempt to elucidate the role of HS inhair follicle morphogenesis and cycling, immunostaining wasperformed for Wnt1, Wnt2, BMP4, �-catenin, EDA, and SHH.No changes in Wnt1, Wnt2, BMP4, and EDA expression andlocalization were detected between the EXT1StEpi�/StEpi� miceand littermate controls (results not shown). EDA expressionwas detected in the ORS, and Wnt1 expression was detected inthe IRS and bulb of the hair follicles in adult EXT1StEpi�/StEpi�

mice (Fig. 9). Wnt1 expression was localized to the bulb, in thematrix, in the IRS cone in the bulb region, and along the prox-imal IRS of the anagen hair follicles in the EXT1StEpi�/StEpi�

mice (Fig. 9). Wnt2 staining was very weak and localized pri-

FIGURE 5. A, syndecan 3 staining (white) was performed in skin cryosectionsfrom littermate control mice at P0 (panels a and b), P20 (panels c and d), andP55 (panels e and f) and EXT1StEpi�/StEpi� mice at P55 (panels g–j). The interfol-licular epithelium (Epi) is oriented upward in the images. Syndecan 3 expres-sion is observed primarily in the sebaceous glands, and EXT1StEpi�/StEpi� micepresent an increase in expression. Panels a, c, e, g, and i, merged mT (red),mGFP (green), syndecan 3, and DAPI (blue) images. Panels b, d, f, h, and j,syndecan 3 and DAPI. Scale bars, 40 �m. B, protein was extracted from the skinof EXT1f/f and EXT1StEpi�/StEpi� mice, submitted to separation by SDS-PAGE,and transferred by current to a PVDF membrane. The membrane was devel-oped with anti-syndecan 3 followed by anti-rabbit secondary antibody con-jugated with Alexa Fluor 555. C, the syndecan 3-positive band from the West-ern blot was quantified, and overall content was divided by that quantified for�-actin, which was developed using anti-�-actin followed by anti-goat sec-ondary antibody conjugated with Alexa Fluor 488. Results are presented as agraph. Error bars represent S.D. RelUF, relative units of fluorescence.

FIGURE 6. A, syndecan 4 staining (white) was performed in skin cryosectionsfrom littermate control mice at P0 (panels a and b), P20 (panels c and d), andP55 (panels e and f) and EXT1StEpi�/StEpi� mice at P55 (panels g– h). The interfol-licular epithelium (Epi) is oriented upward in the images. Panels a, c, e, and g,merged mT (red), mGFP (green), syndecan 4, and DAPI (blue) images. Panels b,d, f, and h, syndecan 4 and DAPI (blue). Scale bars, 40 �m. B, protein wasextracted from the skin of EXT1f/f and EXT1StEpi�/StEpi� mice, submitted to sep-aration by SDS-PAGE, and transferred by current to a PVDF membrane. Themembrane was developed with anti-syndecan 4 followed by anti-rabbit sec-ondary antibody conjugated with Alexa Fluor 555.



marily to the ORS of the EXT1StEpi�/StEpi� mice anagen hairfollicles (results not shown). EDA expression was predomi-nantly located along the proximal IRS of the EXT1StEpi�/StEpi�

mice anagen hair follicles.A drastic increase in SHH expression was observed in the

EXT1StEpi�/StEpi� mice (Fig. 10). Littermate control mice at P0presented SHH expression in the interfollicular epidermal cells,whereas SHH was no longer detected in the interfollicular epi-thelium of mice at P20 or P55, corroborating previous findingsthat SHH plays a role in hair follicle formation (Fig. 10A, panelsa–f). In contrast, there was a 20-fold increase in SHH expres-sion in the interfollicular epidermal cells of EXT1StEpi�/StEpi�

mice when compared with littermate control mice at P0 (deter-mined by quantification of immunofluorescence) (Fig. 10A,panels a, b, and g–j). Moreover, the dermal stroma adjacent tothe epithelium and the sebaceous glands also presented strongSHH staining in EXT1StEpi�/StEpi� mice (Fig. 10A, panels g–j).Therefore the loss of HS leads to an increase in SHH expression(Fig. 10). These results were further confirmed through West-ern blotting analysis, which revealed an �7-fold increase inSHH expression in the EXT1StEpi�/StEpi� mice when comparedwith littermate control mice at P55.

The expression and localization of �-catenin was also inves-tigated in the EXT1StEpi�/StEpi� mice (Fig. 11). Control mice atP0 presented strong �-catenin staining in the IRS of the hairfollicles and in the interfollicular epidermal cells. The expres-sion of �-catenin decreases once hair follicle morphogenesisceases, and the hair follicles of control mice at P20 and P55presented low levels of �-catenin staining. Conversely,EXT1StEpi�/StEpi� mice presented strong �-catenin in the der-

FIGURE 7. Perlecan staining (white) was performed in skin cryosectionsfrom littermate control mice at P0 (a and b), P20 (c and d), and P55 (e andf) and EXT1StEpi�/StEpi� mice at P55 (g and h). The interfollicular epithelium(Epi) is oriented upward in the images. Perlecan expression is observed pri-marily in connective tissue sheath, sebaceous glands, and stroma of bothEXT1f/f mice at P0 and EXT1StEpi�/StEpi� mice at P55. a, c, e, and g, merged mT(red), mGFP (green), perlecan, and DAPI (blue) images. b, d, f, and h, perlecanand DAPI (blue). Nuclei were stained with DAPI (blue). Scale bars, 40 �m.

FIGURE 8. K15 staining (white) was performed in skin cryosectionsfrom EXT1f/f mice at P0 (a and b), P20 (c and d), and P55 (e and f) andEXT1StEpi�/StEpi� mice at P55 (g–i). The interfollicular epithelium (Epi) is ori-ented upward in the images. Control mice at P0 present dispersed K15expression in the epithelium; however, P0 and P55 control mice presentK15 expression in the dermal epithelium, ORS, IRS, and bulb. EXT1StEpi�/StEpi�

mice present a loss of K15 expression in the ORS, IRS, and bulb but retain K15expression in the dermal epithelium and sebaceous glands and in the ORSsolely in some areas. a, c, e, g, and i, merged mT (red), mGFP (green), K15, andDAPI images. b, d, f, and h, K15 and DAPI. Nuclei were stained with DAPI (blue).Scale bars, 40 �m.



mal epithelium, stroma, bulb, dermal papilla, and ORS at P55.The increase in �-catenin expression in EXT1StEpi�/StEpi� micewas further confirmed by Western blotting, which revealed a�4.5-fold increase.

Role of HS in Sebaceous Gland Morphogenesis—EXT1StEpi�/StEpi�

mice induced from P20 to P55 presented a 4-fold increase insebaceous gland number compared with littermate controls,which is in accordance with the increase in hair follicle number.Moreover, EXT1StEpi�/StEpi� mice induced at all time framespresented sebaceous gland hyperplasia. Interestingly,EXT1StEpi�/StEpi� mice induced from P20 presented excessivesebum production to the extent that by P55 macroscopicallythe fur looked wet (Fig. 3d). Oil red O staining of the skin fur-ther revealed the hyperplastic sebaceous glands with alteredmorphology presenting irregular shapes and thickening of thesebaceous gland canal (Fig. 12A).

Role of HS in Sweat Gland Morphogenesis—Mice presentsweat glands solely on the plantar surface of their paws; there-fore, to determine whether HS plays a role in sweat gland mor-phogenesis, the iodine/starch test was performed on the footpads of the hind paws. There was an overall increase in thenumber of sweat glands in EXT1StEpi�/StEpi� mice when com-pared with littermate control mice (Fig. 12B). Moreover, therewas a gradual increase in overall sweat gland number over timein the EXT1StEpi�/StEpi� mice (Fig. 12B). These results werefurther confirmed by histochemistry of the hind paws thatrevealed a 2.5-fold increase in sweat gland number inEXT1StEpi�/StEpi� mice compared with littermate control miceat P55 (Fig. 12C).

Induction of Hair Cycle—To determine whether anagen hairfollicles present in the EXT1StEpi�/StEpi� mice are able to cycle

into the telogen phase, fur was removed by waxing to synchro-nize hair follicles and induce entry of the hair follicles into theanagen phase in mutant and littermate mice. Thereafter, theanimals were left for 25 days to complete a full hair cycle andenter the telogen phase. Interestingly, EXT1StEpi�/StEpi� mice atall time points analyzed presented black skin upon waxing,which represents the anagen phase, whereas EXT1f/f waxed atP35 and P55 presented pink skin, which is representative of thetelogen phase, and at P40 presented gray skin, which is repre-sentative of the catagen phase (Fig. 13), thereby indicating thatall EXT1StEpi�/StEpi� mice were arrested in the anagen phase.Within 5 days of waxing, EXT1StEpi�/StEpi� mice waxed at alltime points presented fully grown fur in the waxed region,thereby no longer presenting a bald patch, whereas littermatecontrol mice only presented signs of fur growing back 15 daysafter waxing (Fig. 14). The skin of the waxed animals was har-vested 25 days after waxing upon completion of a full hair cycle.The waxed skin of littermate control mice presented primarilytelogen hair follicles; however, EXT1StEpi�/StEpi� mice pre-sented solely anagen hair follicles, further demonstrating thatthe absence of HS in hair follicle epithelial cells impedes theprogression of hair follicles into the catagen and, consequently,telogen phases (Fig. 14). Moreover, a gradual increase in synde-can 1 expression was detected upon waxing, further demon-strating that syndecan 1 plays an important role in hair folliclecycling (Fig. 14C). To evaluate whether the ablation of HS instratified epithelial cells accelerates fur regrowth during theanagen phase, animals were induced at P23 or P30 and waxedeither 7 days or immediately after induction, respectively, andthe skin was analyzed 5 days after waxing. Interestingly, whenthe mice were induced 7 days before waxing, the fur of P30

FIGURE 9. EDA (a, b, and c) and Wnt 1 (d, e, and f) staining (white) was performed in skin cryosections from EXT1StEpi�/StEpi� mice at P55. EXT1StEpi�/StEpi�

mice present EDA staining in the ORS and Wnt1 staining in the proximal IRS and bulb. a, merged mT (red), mGFP (green), EDA, and DAPI images. b and e, mergedmT, mGFP, and DAPI images. c, EDA and DAPI images. d, merged mT (red), mGFP (green), Wnt1, and DAPI images. f, Wnt1 and DAPI images. Nuclei were stainedwith DAPI (blue).



EXT1StEpi�/StEpi� mice grew back within 5 days (results notshown), and histology revealed a high density of anagen hairfollicles throughout the waxed region, whereas control mice

FIGURE 10. A, SHH staining (white) was performed in skin cryosections fromlittermate control mice at P0 (panels a and b), P20 (panels c and d), and P55(panels e and f) and EXT1StEpi�/StEpi� mice at P55 (panels g–j). Panels i and j areenlarged images of the boxed area in panel g. The interfollicular epithelium(Epi) is oriented upward in the images. Control mice at P0 present SHH expres-sion in the dermal epithelium that is absent in the P20 and P55 control mice.EXT1StEpi�/StEpi� mice present strong SHH expression in the dermal epithe-lium, stroma adjacent to the epithelium, ORS, IRS, and sebaceous glands.Panels a, c, e, g, and i, merged mT, mGFP, SHH, and DAPI images. Panels b, d, f,h, and j, SHH and DAPI. Panels i and j, enlarged images from the boxed area inpanel g. Nuclei were stained with DAPI (blue). Scale bars, 40 �m. B, protein wasextracted from the skin of EXT1f/f and EXT1StEpi�/StEpi� mice, submitted to sep-aration by SDS-PAGE, and transferred by current to a PVDF membrane. Themembrane was developed with anti-SHH followed by anti-rabbit secondaryantibody conjugated with Alexa Fluor 555. C, the SHH positive band from theWestern blot was quantified, and overall content was divided by that quanti-fied for �-actin, which was developed using anti-�-actin followed by anti-goat secondary antibody conjugated with Alexa Fluor 488. Results are pre-sented as a graph. Error bars represent S.D. RelUF, relative units offluorescence.

FIGURE 11. A, �-catenin staining (white) was performed in skin cryosectionsfrom littermate control mice at P0 (panels a and b), P20 (panels c and d), andP55 (panels e and f) and EXT1StEpi�/StEpi� mice at P55 (panels g and h). Theinterfollicular epithelium (Epi) is oriented upward in the images. Control miceat P0 present strong �-catenin expression in the interfollicular epithelium,IRS, and bulb that is absent in the P20 and P55 control mice. EXT1StEpi�/StEpi�

mice present strong �-catenin expression in the dermal epithelium, stromaadjacent to the epithelium, ORS, IRS, and sebaceous glands. Panels a, c, e, andg, merged mT (red), mGFP (green), �-catenin, and DAPI images. Panels b, d, f,and h, �-catenin and DAPI. Nuclei were stained with DAPI (blue). Scale bars, 40�m. B, protein was extracted from the skin of EXT1f/f and EXT1StEpi�/StEpi� mice,submitted to separation by SDS-PAGE, and transferred by current to a PVDFmembrane. The membrane was developed with anti-�-catenin followed byanti-rabbit secondary antibody conjugated with Alexa Fluor 555. C, the�-catenin-positive band from the Western blot was quantified, and overallcontent was divided by that quantified for �-actin, which was developedusing anti-�-actin followed by anti-goat secondary antibody conjugated withAlexa Fluor 488. Results are presented as a graph. Error bars represent S.D.RelUF, relative units of fluorescence.



presented no visual signs of fur regrowth, and histology revealedregressing hair follicles as well as hair follicles in the initial anagenphase (results not shown). When the mice were waxed immedi-ately after induction, the fur of the EXT1StEpi�/StEpi� mice grewback within 12 days, whereas control mice presented no visualsigns of fur regrowth until 15 days after waxing (results not shown).

To evaluate the long term effects of excessive hair follicleformation and hair follicles being sequestered in the anagen

phase, animals were induced at P20 and waxed at P81. Indeed,fur had regrown on the EXT1StEpi�/StEpi� mice 5 days after waxing;however, the fur grew back in patches, and several areas remainedfurless. 25 days after waxing (P105), the EXT1StEpi�/StEpi� micestill presented irregular fur and bald patches in the waxedregion. Animals at P120 presented general hair loss and areas ofmelanin incontinence within the balding regions (Fig. 15).EXT1StEpi�/StEpi� mice beyond P55 presented a dense distribu-tion of hair follicles throughout the skin with significantlyenlarged sebaceous glands, which seemed to constrict the hairshaft and could possibly have led to the hair loss in mice beyondP55 (Fig. 15). However, the lack of hair cycling due to hair fol-licles being sequestered in the anagen phase could lead to hairloss from an exhausted anagen hair follicle, and an overall sat-urated number of hair follicles could impede the developmentof new hair follicles.

DISCUSSION

Epithelial-mesenchymal interactions trigger the develop-ment of hair follicles during embryogenesis. The extracellularmatrix composition and the coordination between cells and theextracellular matrix play a major role in epithelial-mesenchy-mal interactions (40). HS, a major component of the extracel-

FIGURE 12. A, oil red O staining of EXT1f/f mice at P0 and P55 and EXT1StEpi�/StEpi� mice at P55. Sections were counterstained with hematoxylin. EXT1StEpi�/StEpi�

mice present an increase in oil red O staining, revealing hyperplastic sebaceous glands with an irregular shape. B, iodine/starch test reveals sweat glands oflittermate control mice at P55 and EXT1StEpi�/StEpi� mice at P35, P55, and P120. EXT1StEpi�/StEpi� mice present an increase in sweat gland duct openings. C,sections of the hind paws from EXT1f/f and EXT1StEpi�/StEpi� mice at P55 were double stained with hematoxylin and eosin and visualized under a light micro-scope. A 2.5-fold increase in sweat gland (asterisks) number was observed in EXT1StEpi�/StEpi� mice compared with littermate controls. Scale bars, 100 �m.

FIGURE 13. EXT1f/f and EXT1StEpi�/StEpi� mice were waxed at P55, exposingthe skin and revealing that EXT1f/f mice present pink skin andEXT1StEpi�/StEpi� present black/gray skin.



lular matrix (ECM), is present on the cell surface covalentlybound to syndecan 1, 2, 3, or 4 or in the ECM, for example,bound to perlecan. HS present on the cell surface plays a fun-damental role in cell-cell and cell-matrix interactions (21). HShas been shown to modulate various cytokines involved in celldifferentiation and/or proliferation, such as FGFs, VEGF, SHH,BMP, and Wnts (41, 42). During hair follicle morphogenesis,there is the formation of an epithelial placode, after which ini-tially fibroblasts from the underlying mesenchyme differentiateinto rudimentary dermal papillae (43). For this process, intri-cate cross-talk between the epithelial cells and underlying mes-enchyme must take place. The precise events that regulate hairfollicle morphogenesis remain elusive; however, EDA/EDAreceptor (member of the FGF family), SHH, Wnt1, Wnt2, and�-catenin have been shown to play a role in promoting placodeformation, whereas BMP2 and BMP4 have been shown to playan inhibiting role in placode formation (44 – 48). Interestingly,the ablation of HS in the interfollicular epidermal cells of miceat P20, P23, P25, and P35 led to the formation of new hairfollicles. EDA plays an important role in the early stages of hairfollicle morphogenesis during initial mesenchymal-ectodermalinteractions. After induction, the EXT1StEpi�/StEpi� mice pre-sented persistent formation of new hair follicles, leading to anexcessive distribution of hair follicles throughout the skin. Thedeveloping hair follicles in the EXT1StEpi�/StEpi� mice presentedEDA expression at P55, confirming the formation of hair folli-cles in mature skin with epithelial cells lacking HS. Moreover,the population of stem cells that maintain the epidermis arealso K14-positive and therefore in our mouse model alsoundergo the ablation of the Ext1 gene and therefore lack HS (49,50). Previous studies have shown that during homeostasis theepidermal stem cells simply maintain tissue integrity; however,upon wounding or specific genetic modifications, the stem cellscan give rise to any differentiated epidermal lineage (49, 50).Therefore, our studies demonstrate that upon HS ablation epi-dermal stem cells differentiate into hair follicle, sebaceousgland cells, and sweat glands; thereby HS is an important regu-

FIGURE 14. A, accelerated fur regrowth in EXT1StEpi�/StEpi� mice. EXT1StEpi�/StEpi� andlittermate control mice were induced at P23, waxed at P30, and imaged atP35. EXT1StEpi�/StEpi� mice present fur regrowth 5 days after waxing; however,littermate control mice present no signs of hair regrowth at the same timeframe. B, EXT1StEpi�/StEpi� mice present hair follicles sequestered in anagen.EXT1StEpi�/StEpi� and littermate control mice were waxed to synchronize hairfollicles and left to complete a full hair cycle. EXT1StEpi�/StEpi� mice present nocatagen or telogen hair follicles, whereas littermate control mice present pri-marily telogen hair follicles. Tissue sections were double stained with hema-toxylin and eosin and visualized under a Nikon light microscope. C, proteinwas extracted from naïve skin and skin 25 days after waxing from EXT1f/f andEXT1StEpi�/StEpi� mice, submitted to separation by SDS-PAGE, and transferredby current to a PVDF membrane. The membrane was developed with anti-syndecan 1 followed by anti-rabbit secondary antibody conjugated withAlexa Fluor 555. D, the syndecan 1-positive band from the Western blot wasquantified, and overall content was divided by that quantified for �-actin,

which was developed using anti-�-actin followed by anti-goat secondaryantibody conjugated with Alexa Fluor 488. Results are presented as a graph.Error bars represent S.D. RelUF, relative units of fluorescence.

FIGURE 15. EXT1StEpi�/StEpi� mice present hair loss beyond P55 (a and b),melanin incontinence beyond P120 (c), and fused hair follicles and seba-ceous glands (d and e). e represents the boxed area in d at higher magnifica-tion. Tissue sections were double stained with hematoxylin and eosin andvisualized under a Nikon light microscope.



lator of the epidermal stem cells, and HSPGs are promisingstem cell markers for the epidermis and epidermal appendages.

Previous studies have shown that hair follicles express highlevels of syndecan 1 (51, 52). We hereby show that during hairfollicle morphogenesis high levels of syndecan 1 expressionwere observed in the stroma proximal to the epithelium, andhence, syndecan 1 could play a role in stroma-epithelium cross-talk during hair follicle formation. Syndecan 1 is the major syn-decan expressed by epithelial cells, and it plays a vital role dur-ing wound healing; however, no hair phenotype has beendescribed in syndecan 1-overexpressing or knock-out mice todate (38). Syndecan 1, syndecan 3, and syndecan 4 knock-outstudies have revealed that these mice present subtle phenotypeswhen compared with knock-out mice lacking enzymes involvedin HS biosynthesis that could be attributed to compensatorymechanisms between syndecans (53). Accordingly, we herebyshow that the ablation of HS in solely the interfollicular epider-mis led to a phenotype not present in the syndecan knock-out

mice. In contrast, the perlecan knock-out mice (Hspg2�/�)present a severe phenotype, and the mice die from embryonicday 10.5 to just after birth. It remains to be elucidated whetherknocking out perlecan using an inducible system would repro-duce the phenotype observed in the EXT1StEpi�/StEpi� mice.

HS chains can bind a plethora of growth factors, acting as areservoir of soluble factors in the ECM. To investigate whichsignaling pathways were affected by the absence of HS, we ana-lyzed the expression profile of cytokines that govern hair folliclemorphogenesis. HS has been shown previously to bind andmodulate the activity of Wnt and BMP family cytokines; how-ever, no alteration in their expression profile was observed.SHH is a morphogen that mediates many developmental pro-cesses, including hair follicle morphogenesis and cycling, andrequires HS for normal distribution and signaling activity (54 –56). During embryogenesis, SHH is expressed at the distal tip ofthe developing follicle by the proliferating cells. Interestingly,EXT1StEpi�/StEpi� mice presented strong SHH staining in the

FIGURE 16. Schematic of the hair follicle morphogenesis and cycling in EXT1StEpi�/StEpi� mice. Top left panel, events that take place during placodeformation during normal hair follicle morphogenesis. Bottom left panel, signaling pathways and protein expression in telogen hair follicles at P55 in littermatecontrol mice. Right panel, signaling pathways and protein expression present in hair follicles of EXT1StEpi�/StEpi� mice at P55 revealing similarities with earlymorphogenesis. Syn, syndecan.



interfollicular epidermal cells at P55 that in control mice wassolely present during hair follicle morphogenesis. Thus,SHH expressed by the interfollicular epidermal cells mayplay an intricate role in the epithelial-mesenchymal cross-talk regulating the formation of hair follicles, which is medi-ated through HSPGs. Thereby, we can hypothesize that theHS/SHH balance in the interfollicular epidermal cells ofmature skin hinders the formation of new hair follicles, andsolely the ablation of HS from these cells in the adult micetriggered formation of new hair follicles. Thus, pharmaceu-tically targeting HS levels in skin epithelial cells could enableengineering of de novo skin appendages in mature skin afterskin burns and scarring.

Hair follicles maintain hair health by undergoing repeatedcycles of growth (anagen), regression (catagen), and finally qui-escence of mature hair follicles (telogen). During anagen, thematrix cells at the base of hair follicles derived from stem cells inthe bulge region undergo rapid proliferation. Thereafter, thedownward growth of anagen hair follicles requires enzymaticdegradation of the surrounding ECM. Therefore, previousstudies with heparanase (HS endoglycosidase) overexpressionin mice revealed an increased hair growth rate during the ana-gen phase, suggesting that heparanase plays an important rolein degrading the ECM during this downward growth (27, 28).Interestingly, EXT1StEpi�/StEpi� mice showed a drastic increasein hair growth rate; however, these mice have intact HS in theECM surrounding the hair follicle. However, the bulge regionand the stem cells in the bulge region, which undergo rapidproliferation and are responsible for the downward growth,lack HS, which could potentially eliminate a physical barrierimposed by HS within the hair follicle, thereby facilitating themigration of stem cells undergoing rapid proliferation from thebulge region to the base of the hair follicle.

The absence of HS in hair follicles also led to altered differ-entiation of the cells during anagen, which could hinder regres-sion into catagen. SHH and �-catenin have been shown to alsoplay an important role in regulating hair follicle cycling (10).The absence of HS in the hair follicles led to an increase in SHHand �-catenin expression, which could directly affect the differ-entiation of these cells, hindering them in the anagen phase byaffecting the signaling pathways of SHH and �-catenin, thusaltering the signaling cues for catagen.

Heparanase has been detected previously in the ORS ofmurine hair follicles. Heparanase overexpression has beenreported to improve mouse hair regrowth (28). Heparanase hasbeen shown to be distributed in hair follicles, primarily locatedin the IRS of human hair follicles strictly in the anagen phase(28). Moreover, the inhibition of heparanase in in vitro culturedhair follicles induces a catagen-like process (28). Our findingsare in accordance with the absence of HS impeding progressioninto catagen. Moreover, Malgouries et al. (29) showed previ-ously that there is an increase in heparanase expression inhuman anagen hair follicles; however, at the onset of catagen,hair follicle heparanase expression ceases, further supportingour findings. Therefore, our results clearly reveal the impor-tance of fluctuations in the HS levels of hair follicle epithelialcells, which dictate hair follicle cycling and thus play a vital rolein hair follicle homeostasis.

The sebaceous gland plays a vital role in hair and skin homeo-stasis by producing sebum, which is deposited on the hairwithin the follicle and is brought to the surface through the hairshaft. Sebum protects the hair and skin by maintaining lubrica-tion and preventing dryness, infections, and irritation. Lipidcomponents of sebum have also been speculated to play a rolein waterproofing the skin and hair. The effects of hair folliclecycling on sebaceous gland homeostasis and whether the seba-ceous gland itself plays a role during hair follicle cycling remainelusive. EXT1StEpi�/StEpi� mice presented a drastic increase inthe overall number of hair follicles, which consequently led toan increase in the number of sebaceous glands. The ablation ofHS in sebaceous glands led to hyperplasia, and the glands pre-sented an overall altered morphology (enlarged with an irregu-lar shape). From P55, the animals presented a significantincrease in sebum secretion and excessive sebum accumulationon the skin and fur. As the animals aged, the size of sebaceousglands increased to the extent that they constrained the hairshaft, possibly blocking the surfacing of hair from the hair fol-licles, and this could have led to the hair loss observed beyondP55.

Taken together, our findings clearly show that an intricatecontrol of HS levels is required for hair follicle and sebaceousgland cycling. A gradual decrease in HS is required for anagen,and thereafter an increase in HS expression is required for cata-gen. Moreover, HS plays a vital role in the early stages of hairfollicle formation during mesenchymal-ectodermal interac-tions. Our findings are summarized in Fig. 16. Studies on hairfollicle morphogenesis may be transposed to other appendages,such as tooth formation and mammary gland formation, whichrequire similar mesenchymal-epithelial interactions (57). Thefundamental signaling cues for hair morphogenesis are evolu-tionarily conserved across species and similar for other types ofskin appendages, such as feathers and scales (58).

Acknowledgment—We thank Shao-Hsuan Chang for technicalassistance.

REFERENCES1. Müller-Röver, S., Handjiski, B., van der Veen, C., Eichmüller, S., Foitzik, K.,

McKay, I. A., Stenn, K. S., and Paus, R. (2001) A comprehensive guide forthe accurate classification of murine hair follicles in distinct hair cyclestages. J. Invest. Dermatol. 117, 3–15

2. Sennett, R., and Rendl, M. (2012) Mesenchymal-epithelial interactionsduring hair follicle morphogenesis and cycling. Semin. Cell Dev. Biol. 23,917–927

3. Schlake, T. (2007) Determination of hair structure and shape. Semin. CellDev. Biol. 18, 267–273

4. Dry, F. W. (1926) The coat of the mouse (Mus musculus). J. Genet. 16,287–344

5. Paus, R., and Cotsarelis, G. (1999) The biology of hair follicles. N. Engl.J. Med. 341, 491– 497

6. Stenn, K. S., and Paus, R. (2001) Controls of hair follicle cycling. Physiol.Rev. 81, 449 – 494

7. Paus, R., Krejci-Papa, N., Li, L., Czarnetzki, B. M., and Hoffman, R. M.(1994) Correlation of proteolytic activities of organ cultured intact mouseskin with defined hair cycle stages. J. Dermatol. Sci. 7, 202–209

8. Ye, J., Yang, T., Guo, H., Tang, Y., Deng, F., Li, Y., Xing, Y., Yang, L., andYang, K. (2013) Wnt10b promotes differentiation of mouse hair folliclemelanocytes. Int. J. Med. Sci. 10, 691– 698



9. Woo, W. M., Zhen, H. H., and Oro, A. E. (2012) Shh maintains dermalpapilla identity and hair morphogenesis via a Noggin-Shh regulatory loop.Genes Dev. 26, 1235–1246

10. Cui, C. Y., Kunisada, M., Childress, V., Michel, M., and Schlessinger, D.(2011) Shh is required for Tabby hair follicle development. Cell Cycle 10,3379 –3386

11. Oshimori, N., and Fuchs, E. (2012) Paracrine TGF-� signaling counterbal-ances BMP-mediated repression in hair follicle stem cell activation. CellStem Cell 10, 63–75

12. Lee, J., and Tumbar, T. (2012) Hairy tale of signaling in hair follicle devel-opment and cycling. Semin. Cell Dev. Biol. 23, 906 –916

13. Nanba, D., Toki, F., Barrandon, Y., and Higashiyama, S. (2013) Recentadvances in the epidermal growth factor receptor/ligand system biologyon skin homeostasis and keratinocyte stem cell regulation. J. Dermatol.Sci. 72, 81– 86

14. Sugawara, K., Schneider, M. R., Dahlhoff, M., Kloepper, J. E., and Paus, R.(2010) Cutaneous consequences of inhibiting EGF receptor signaling invivo: normal hair follicle development, but retarded hair cycle inductionand inhibition of adipocyte growth in Egfr(Wa5) mice. J. Dermatol. Sci. 57,155–161

15. Srivastava, A. K., Durmowicz, M. C., Hartung, A. J., Hudson, J., Ouzts,L. V., Donovan, D. M., Cui, C. Y., and Schlessinger, D. (2001) Ectodyspla-sin-A1 is sufficient to rescue both hair growth and sweat glands in Tabbymice. Hum. Mol. Genet. 10, 2973–2981

16. Lin, X., and Perrimon, N. (1999) Dally cooperates with Drosophila Frizzled2 to transduce Wingless signalling. Nature 400, 281–284

17. Swee, L. K., Ingold-Salamin, K., Tardivel, A., Willen, L., Gaide, O., Favre,M., Demotz, S., Mikkola, M., and Schneider, P. (2009) Biological activity ofectodysplasin A is conditioned by its collagen and heparan sulfate pro-teoglycan-binding domains. J. Biol. Chem. 284, 27567–27576

18. Kuo, W. J., Digman, M. A., and Lander, A. D. (2010) Heparan sulfate actsas a bone morphogenetic protein coreceptor by facilitating ligand-in-duced receptor hetero-oligomerization. Mol. Biol. Cell 21, 4028 – 4041

19. Park, Y., Rangel, C., Reynolds, M. M., Caldwell, M. C., Johns, M., Nayak,M., Welsh, C. J., McDermott, S., and Datta, S. (2003) Drosophila perlecanmodulates FGF and Hedgehog signals to activate neural stem cell division.Dev. Biol. 253, 247–257

20. Yan, D., Wu, Y., Yang, Y., Belenkaya, T. Y., Tang, X., and Lin, X. (2010) Thecell-surface proteins Dally-like and Ihog differentially regulate Hedgehogsignaling strength and range during development. Development 137,2033–2044

21. Bishop, J. R., Schuksz, M., and Esko, J. D. (2007) Heparan sulphate pro-teoglycans fine-tune mammalian physiology. Nature 446, 1030 –1037

22. Meyers, J. R., Planamento, J., Ebrom, P., Krulewitz, N., Wade, E., andPownall, M. E. (2013) Sulf1 modulates BMP signaling and is required forsomite morphogenesis and development of the horizontal myoseptum.Dev. Biol. 378, 107–121

23. Bush, K. T., Crawford, B. E., Garner, O. B., Nigam, K. B., Esko, J. D., andNigam, S. K. (2012) N-Sulfation of heparan sulfate regulates early branch-ing events in the developing mammary gland. J. Biol. Chem. 287,42064 – 42070

24. Shah, M. M., Sakurai, H., Gallegos, T. F., Sweeney, D. E., Bush, K. T., Esko,J. D., and Nigam, S. K. (2011) Growth factor-dependent branching of theureteric bud is modulated by selective 6-O sulfation of heparan sulfate.Dev. Biol. 356, 19 –27

25. Qu, X., Carbe, C., Tao, C., Powers, A., Lawrence, R., van Kuppevelt, T. H.,Cardoso, W. V., Grobe, K., Esko, J. D., and Zhang, X. (2011) Lacrimal glanddevelopment and Fgf10-Fgfr2b signaling are controlled by 2-O- and 6-O-sulfated heparan Sulfate. J. Biol. Chem. 286, 14435–14444

26. Delehedde, M., Lyon, M., Sergeant, N., Rahmoune, H., and Fernig, D. G.(2001) Proteoglycans: pericellular and cell surface multireceptors that in-tegrate external stimuli in the mammary gland. J. Mammary Gland Biol.Neoplasia 6, 253–273

27. Vlodavsky, I., Friedmann, Y., Elkin, M., Aingorn, H., Atzmon, R., Ishai-Michaeli, R., Bitan, M., Pappo, O., Peretz, T., Michal, I., Spector, L., andPecker, I. (1999) Mammalian heparanase: gene cloning, expression andfunction in tumor progression and metastasis. Nat. Med. 5, 793– 802

28. Zcharia, E., Zilka, R., Yaar, A., Yacoby-Zeevi, O., Zetser, A., Metzger, S.,

Sarid, R., Naggi, A., Casu, B., Ilan, N., Vlodavsky, I., and Abramovitch, R.(2005) Heparanase accelerates wound angiogenesis and wound healing inmouse and rat models. FASEB J. 19, 211–221

29. Malgouries, S., Donovan, M., Thibaut, S., and Bernard, B. A. (2008) Hepa-ranase 1: a key participant of inner root sheath differentiation programand hair follicle homeostasis. Exp. Dermatol. 17, 1017–1023

30. Malinowska, M., Jakóbkiewicz-Banecka, J., Kloska, A., Tylki-Szymanska,A., Czartoryska, B., Piotrowska, E., Wegrzyn, A., and Wegrzyn, G. (2008)Abnormalities in the hair morphology of patients with some but not alltypes of mucopolysaccharidoses. Eur. J. Pediatr. 167, 203–209

31. Nguyen, H., Rendl, M., and Fuchs, E. (2006) Tcf3 governs stem cell fea-tures and represses cell fate determination in skin. Cell 127, 171–183

32. Perl, A. K., Wert, S. E., Nagy, A., Lobe, C. G., and Whitsett, J. A. (2002)Early restriction of peripheral and proximal cell lineages during formationof the lung. Proc. Natl. Acad. Sci. U.S.A. 99, 10482–10487

33. Muzumdar, M. D., Tasic, B., Miyamichi, K., Li, L., and Luo, L. (2007) Aglobal double-fluorescent Cre reporter mouse. Genesis 45, 593– 605

34. Inatani, M., Irie, F., Plump, A. S., Tessier-Lavigne, M., and Yamaguchi, Y.(2003) Mammalian brain morphogenesis and midline axon guidance re-quire heparan sulfate. Science 302, 1044 –1046

35. Dietrich, C. P., McDuffie, N. M., and Sampaio, L. O. (1977) Identificationof acidic mucopolysaccharides by agarose gel electrophoresis. J. Chro-matogr. 130, 299 –304

36. Bradford, M. M. (1976) A rapid and sensitive method for the quantitationof microgram quantities of protein utilizing the principle of protein-dyebinding. Anal. Biochem. 72, 248 –254

37. Podlasek, C. A., Zelner, D. J., Jiang, H. B., Tang, Y., Houston, J., McKenna,K. E., and McVary, K. T. (2003) Sonic hedgehog cascade is required forpenile postnatal morphogenesis, differentiation, and adult homeostasis.Biol. Reprod. 68, 423– 438

38. Elenius, V., Götte, M., Reizes, O., Elenius, K., and Bernfield, M. (2004)Inhibition by the soluble syndecan-1 ectodomains delays wound repair inmice overexpressing syndecan-1. J. Biol. Chem. 279, 41928 – 41935

39. Whitbread, L. A., and Powell, B. C. (1998) Expression of the intermediatefilament keratin gene, K15, in the basal cell layers of epithelia and the hairfollicle. Exp. Cell Res. 244, 448 – 459

40. Watt, F. M., and Huck, W. T. (2013) Role of the extracellular matrix inregulating stem cell fate. Nat Rev. Mol. Cell Biol. 14, 467– 473

41. Esko, J. D., and Selleck, S. B. (2002) Order out of chaos: assembly of ligandbinding sites in heparan sulfate. Annu. Rev. Biochem. 71, 435– 471

42. Lin, X. (2004) Functions of heparan sulfate proteoglycans in cell signalingduring development. Development 131, 6009 – 6021

43. Biggs, L. C., and Mikkola, M. L. (2014) Early inductive events in ectoder-mal appendage morphogenesis. Semin. Cell Dev. Biol. 25–26, 11–21

44. Mou, C., Jackson, B., Schneider, P., Overbeek, P. A., and Headon, D. J.(2006) Generation of the primary hair follicle pattern. Proc. Natl. Acad.Sci. U.S.A. 103, 9075–9080

45. Fliniaux, I., Mikkola, M. L., Lefebvre, S., and Thesleff, I. (2008) Identifica-tion of dkk4 as a target of Eda-A1/Edar pathway reveals an unexpectedrole of ectodysplasin as inhibitor of Wnt signalling in ectodermal placodes.Dev. Biol. 320, 60 –71

46. Pispa, J., Pummila, M., Barker, P. A., Thesleff, I., and Mikkola, M. L. (2008)Edar and Troy signalling pathways act redundantly to regulate initiation ofhair follicle development. Hum. Mol. Genet. 17, 3380 –3391

47. Pummila, M., Fliniaux, I., Jaatinen, R., James, M. J., Laurikkala, J., Schnei-der, P., Thesleff, I., and Mikkola, M. L. (2007) Ectodysplasin has a dual rolein ectodermal organogenesis: inhibition of Bmp activity and induction ofShh expression. Development 134, 117–125

48. Botchkarev, V. A., Botchkareva, N. V., Roth, W., Nakamura, M., Chen,L. H., Herzog, W., Lindner, G., McMahon, J. A., Peters, C., Lauster, R.,McMahon, A. P., and Paus, R. (1999) Noggin is a mesenchymally derivedstimulator of hair-follicle induction. Nat. Cell Biol. 1, 158 –164

49. Arwert, E. N., Hoste, E., and Watt, F. M. (2012) Epithelial stem cells,wound healing and cancer. Nat. Rev. Cancer 12, 170 –180

50. Jensen, K. B., Driskell, R. R., and Watt, F. M. (2010) Assaying proliferationand differentiation capacity of stem cells using disaggregated adult mouseepidermis. Nat. Protoc. 5, 898 –911



51. Richardson, G. D., Fantauzzo, K. A., Bazzi, H., Määttä, A., and Jahoda,C. A. (2009) Dynamic expression of Syndecan-1 during hair follicle mor-phogenesis. Gene Expr. Patterns 9, 454 – 460

52. Bayer-Garner, I. B., Sanderson, R. D., and Smoller, B. R. (2002) Syndecan-1is strongly expressed in the anagen hair follicle outer root sheath and in thedermal papilla but expression diminishes with involution of the hair folli-cle. Am. J. Dermatopathol. 24, 484 – 489

53. Alexopoulou, A. N., Multhaupt, H. A., and Couchman, J. R. (2007) Synde-cans in wound healing, inflammation and vascular biology. Int. J. Biochem.Cell Biol. 39, 505–528

54. Ohlig, S., Pickhinke, U., Sirko, S., Bandari, S., Hoffmann, D., Dreier, R.,Farshi, P., Götz, M., and Grobe, K. (2012) An emerging role of Sonichedgehog shedding as a modulator of heparan sulfate interactions. J. Biol.

Chem. 287, 43708 – 4371955. Chang, S. C., Mulloy, B., Magee, A. I., and Couchman, J. R. (2011) Two

distinct sites in sonic Hedgehog combine for heparan sulfate interactionsand cell signaling functions. J. Biol. Chem. 286, 44391– 44402

56. Zhang, F., McLellan, J. S., Ayala, A. M., Leahy, D. J., and Linhardt, R. J.(2007) Kinetic and structural studies on interactions between heparin orheparan sulfate and proteins of the hedgehog signaling pathway. Biochem-istry 46, 3933–3941

57. Mikkola, M. L. (2007) Genetic basis of skin appendage development. Se-min. Cell Dev. Biol. 18, 225–236

58. Wu, P., Hou, L., Plikus, M., Hughes, M., Scehnet, J., Suksaweang, S., Widel-itz, R., Jiang, T. X., and Chuong, C. M. (2004) Evo-Devo of amniote integ-uments and appendages. Int. J. Dev. Biol. 48, 249 –270



HeparanSulfateRegulatesHairFollicleandSebaceousGland ... · growth and hair shaft production. The absence of HS also impeded the formation of catagen follicles in EXT1StEpi /StEpi

Documents