Chondroitin 6-sulfate represses keratinocyte proliferation in mouse skin, which is associated with psoriasis

Chondroitin sulfates are implicated in epidermal biology, but functional significance of chondroitin sulfates remains unclear. Here, we report that chondroitin 6-sulfate is important for the maintenance of epidermal homeostasis. Mice deficient in chondroitin 6-O-sulfotransferase-1 (C6st-1), which is involved in biosynthesis of chondroitin 6-sulfate, exhibited keratinocyte hyperproliferation and impaired skin permeability barrier function. Chondroitin 6-sulfate directly interacted with the EGF receptor and negatively controlled ligand-induced EGF receptor signaling. Normal function of hyperproliferative C6st-1-knockout mouse-derived keratinocytes was rescued by treatment with exogenous chondroitin 6-sulfate. Epidermal hyperplasia, induced using imiquimod, was more severe in C6st-1-knockout mice than in C6st-1 wild-type mice. Taken together, these findings indicate that chondroitin 6-sulfate represses keratinocyte proliferation in normal skin, and that the expression level of C6st-1 may be associated with susceptibility to psoriasis. Kitazawa et al. show that chondroitin 6-sulfate represses keratinocyte proliferation in skin. They find that mice deficient in chondroitin-6-O-sulfotransferase-1 (C6st-1), which synthesizes chondroitin 6-sulfate, exhibit keratinocyte hyperproliferation and impaired skin permeability barrier. This study suggests that the expression level of C6st-1 may serve as a biomarker for the susceptibility to psoriasis.

M ammalian skin epithelium is a self-renewing tissue that constitutes the barrier between an organism and its environment. Epithelial keratinocyte proliferation is an essential element in wound healing, and abnormal epithelial proliferation is an intrinsic factor in the skin disorder psoriasis. The mechanisms that trigger epithelial proliferation are not completely understood. The inner basal layer adheres to an underlying basement membrane, which is rich in extracellular matrix components containing chondroitin sulfate (CS) proteoglycans. This layer contains proliferative keratinocytes that are typed by their expression of genes encoding keratins 5 and 14, and growth factor receptors such as the EGF receptor (EGFR) 1 .
Here we have shown that the expression level of chondroitin 6sulfate plays an important role in controlling the proliferation of basal keratinocytes via EGFR signaling.
Expression quantitative trait loci (eQTL) analysis shows that expression of the FAM20B gene is associated with psoriasis 2 . FAM20B is a kinase involved in xylose phosphorylation in the glycosaminoglycan (GAG)-protein linkage region 3 . FAM20B also regulates the number of GAG chains generated by a cell and plays an important role in the biosynthesis of GAG 3 . A previous study has demonstrated that fam20b-mutant zebrafishes have deficient production of cartilage CS proteoglycan and show chondrocyte hypertrophy in vivo 4 . In addition, Fam20B-knockout mouse embryos are severely stunted, showing multisystem organ hypoplasia, delayed development most evident in the skeletal system, eyes, lung, GI tract, and liver, and increased mortality at E13.5 5 . Koike et al. reported that FAM20B regulates the sulfation profile of CS chains and total amount of CS synthesized in cells 3 . Based on these findings, we hypothesized that a decrease in the expression of FAM20B would impact CS biosynthesis and would be associated with psoriasis.
CS chains consist of repeating disaccharide units [(−4GlcUAβ1-3GalNAcβ1-) n ], and are covalently linked to specific serine (Ser) residues in any of the core proteins via GAGprotein linkage region (GlcUAβ1-3 Galβ1-3Galβ1-4Xylβ1-O-Ser) 6,7 . The biosynthesis of the repeating disaccharide units is catalyzed by combined activity of six homologous glycosyltransferaseschondroitin (Chn) synthases (ChSy)−1, −2, and −3, Chn polymerizing factor (ChPF), and CS GalNAc transferases (CSGalNAcT)−1 and −2. The resulting backbones of CS chains are subsequently modified via sulfation and uronate epimerization 6 . Based on substrate preferences of Chn sulfotransferases identified to date, the biosynthetic scheme for CStype sulfation is classified into the initial "4-O-sulfation" and "6-O-sulfation" pathways. In the initial step, the non-sulfated O unit [GlcUA-GalNAc] serves as a common acceptor substrate for two types of sulfotransferases, chondroitin 4-O-sulfotransferases (C4ST-1 and C4ST-2) 8-10 and chondroitin 6-O-sulfotransferase-1 (C6ST-1), forming monosulfated A (GlcUA-GalNAc(4-O-sulfate)) and C (GlcUA-GalNAc(6-O-sulfate)) units, respectively. Subsequent sulfation of A and C units can also occur via activity of GalNAc 4-sulfate 6-O-sulfotransferase (GalNAc4S-6ST) or CSspecific uronyl 2-O-sulfotransferase (UST), resulting in the formation of disulfated disaccharide E (GlcUA-GalNAc(4,6-O-disulfate)) and D (GlcUA(2-O-sulfate)-GalNAc(6-O-sulfate)) units, respectively 6 . Specific sulfation patterns of CS have been shown to affect cell proliferation. Chondroitin 6-sulfate, which has a high content of C units, is incorporated into collagen gels and promotes the growth of keratinocytes 11 . Chondroitin 4-sulfate is necessary for chondrocyte proliferation mediated by Indian hedgehog signaling 12 . Previous reports have shown that an imbalance in the sulfation of CS causes a hyperproliferative phenotype in cells derived from patients with Costello syndrome 13 . Notably, abnormal proliferation is a critical step in the pathogenesis of Costello syndrome, which is a combination of distinctive multiple congenital anomalies associated with increased cellular proliferation and tumor development. Costello syndrome is associated with severely reduced expression of C4ST-1 and loss of chondroitin 4-sulfate via constitutively activated HRAS signaling 13 . Chondroitin 6-sulfate is accumulated in the hearts of patients with Costello syndrome 14 . Restored expression of C4ST-1 can rescue increased proliferation of fibroblasts acquired from patients with Costello syndrome 13 . These findings indicate that the sulfation pattern of CS is a crucial regulator of proliferation, and that an imbalance in 6-O-sulfation and 4-Osulfation of CS leads to pathological conditions. FAM20B, a psoriasis-risk gene, regulates the sulfation profile of CS chains and total amount of GAG synthesized in cells. Increased expression levels of FAM20B are concomitant with increased proportion of CS 6-O-sulfation to 4-O-sulfation 3 . Therefore, we hypothesized that decreased expression of FAM20B may cause pathological conditions via decreases in the proportion of CS 6-O-sulfation to 4-O-sulfation. In this study, we used C6st-1 knockout mice that lack the expression of CS 6-sulfate to investigate how an imbalance in 6-O-sulfation to 4-O-sulfation affects CS and skin pathology. Here we show that C6st-1 knockout mice exhibit keratinocyte hyperproliferation and impaired skin permeability barrier and that the expression level of chondroitin 6sulfate plays an important role in controlling the proliferation of basal keratinocytes via EGFR signaling. This study suggests that the expression level of C6st-1 may serve as a biomarker for the susceptibility to psoriasis.

Results
Percentage of 6-sulfated CS-disaccharides (C units) is decreased in psoriatic fibroblasts. Figure 1a, b show the structure and biosynthesis of CS. Because psoriatic human epidermal keratinocytes are not commercially available, we first analyzed the CS produced by psoriatic human skin fibroblasts and normal human skin fibroblasts. The percentage of 6-sulfated CS-disaccharides (C units) in total CS-disaccharide units was slightly, but significantly, decreased in psoriatic fibroblasts (Fig. 1c).
C6st-1 HE and KO mice show decreased expression of C and D units and epidermal hyperplasia. Next, we analyzed mice lacking in CS 6-sulfation to examine whether a decrease in CS 6-sulfation would be associated with psoriasis. C6st-1 is involved in formation of 6-O-sulfated CS disaccharide units (C unit and D unit) (Fig. 1b). All mice produced similar amounts of CS, while production of the C and D units was decreased in C6st-1 wild type (WT), C6st-1 hetero (HE), and C6st-1 knockout (KO) mice with respect to genotype (Fig. 1d, e, Supplementary Table 1). The expression levels of CS biosynthetic enzymes are shown in Fig. 1f. We examined the expression pattern of chondroitin 6-sulfate (CS-C) in the epidermis of WT mice (Fig. 1g). Immunohistochemical analysis using anti-CS-C antibody demonstrated that CS-C was expressed beneath the keratin 14 (K14)-positive stem cells of epidermal basal layer (Fig. 1g). In contrast, CS-C was not detected in the epidermal basal layer of C6st-1 HE and C6st-1 KO mice. Histological analyses revealed that in contrast with the thin epidermis of newborn and 8-week-old C6st-1 WT mice, C6st-1 HE, and C6st-1 KO mice showed a hyperthickened epidermis (Fig. 1h, i). In addition, the epidermal thickness in newborn was thicker than that in 8-week-old mice. In this regard, it is reported that neonatal mice epidermal thickness in adult skin is thinner than that in infant skin because the epidermis reaches maximum thickness in the late embryo and cell proliferation gradually decreases within a few weeks postpartum 15 . Although there were differences in proliferation activity of keratinocytes between newborn and adult epidermis, C6st-1 affected both newborn and Fig. 1 Downregulation of chondroitin 6-sulfate is associated with epidermal hyperplasia. a The structure of CS chains is mediated by various glycosyltransferases. The common glycosaminoglycan-protein linkage region, GlcUAβ1-3Galβ1-3Galβ1-4Xylβ1-, is built on specific serine (Ser) residue(s) of core proteins such as ACAN, VCAN, DCN, and BGN. After the linkage region is formed, the CS polymerase complex assembles the CS backbone (disaccharide repeating region). b Outline of sulfation pathways. The C6-position of the GalNAc residue in the O unit is sulfated by C6ST-1 to form the C unit. Subsequently, the C unit is converted to a D unit by uronyl 2-O-sulfotransferase UST. c Expression of CS chains in normal human skin fibroblasts (n = 3) and psoriatic fibroblasts (n = 3) was analyzed using high pressure liquid chromatography (HPLC) to measure the composition of CS-disaccharides. Sulfated CS chains, isolated from the epidermis of C6st-1 WT (n = 3), C6st-1 HE (n = 4), and C6st-1 KO mice (n = 3), were analyzed using HPLC to measure the total amount (d) and composition of CS-disaccharides (e). f Gene expression levels of CS biosynthetic enzymes in the epidermis of C6st-1 WT (n = 4-6), C6st-1 HE (n = 4-7), and C6st-1 KO mice (n = 3) were analyzed using real-time PCR. g The expression pattern of 6-O-sulfated CS (CS-C) in the tail epidermis of C6st-1 WT, C6st-1 HE, and C6st-1 KO adult mice was examined by immunohistochemical analysis using anti-keratin 14 (K14) and anti-CS-C antibody. h Hematoxylin-eosin staining of paraffin-embedded skin samples from newborn and 8-week-old C6st-1 WT, C6st-1 HE, and C6st-1 KO mouse skin. Yellow arrows indicate epidermis. i Epidermal thickness of newborn and 8-week-old C6st-1 WT, C6st-1 HE, and C6st-1 KO mice. Newborn mice, C6st-1 WT (n = 4), C6st-1 HE (n = 4), and C6st-1 KO (n = 3) ; 8-week-old mice, C6st-1 WT (n = 4), C6st-1 HE (n = 3), and C6st-1 KO (n = 4) . Statistical significance was determined using one-way ANOVA with Tukey's HSD test.
Proliferating K14-positive basal keratinocytes is increased in newborn and 13-week-old HE and KO mice. We next examined whether epidermal proliferation and differentiation are affected by C6st-1 ablation. C6st-1-null epidermis appeared morphologically normal; the expression patterns of basal keratin 14 (K14), which is a marker protein for non-differentiated and proliferative keratinocytes, were indistinguishable from those in C6st-1 WT skin (Fig. 2). Proliferating K14-positive basal keratinocytes were examined by immunolabeling for Ki67, which is expressed in the nucleus of cycling cells. The number of proliferating K14-positive basal keratinocytes was increased in newborn and 13-week-old C6st-1 HE and C6st-1 KO mice compared with that in C6st-1 WT mice (Fig. 2). Consistent with this result, the K14-positive cells in the basal layer were mitotically more active in newborn C6st-1 HE and C6st-1 KO epidermis than in C6st-1 WT epidermis (Fig. 2b). These results suggest that proliferation is enhanced by loss of C6st-1 in the basal layer.
WT mice (Fig. 3d). These results suggest that loss of C6st-1 affects epidermal differentiation and epidermal proliferation. We next examined the gene expression levels of differentiation markers (K14, K10, Flg, Ivl, and Lor) in primary keratinocytes (Fig. 3g). The expression level of K14 was elevated in C6st-1 KO keratinocytes compared with C6st-1 WT keratinocytes, whereas Ivl was down-regulated in C6st-1 KO keratinocytes. However, altered differentiation marker expression levels could not be rescued by addition of CS-C to C6st-1 KO keratinocytes. There were no significant differences in the expression levels of K10, Flg, and Lor between C6st-1 WT and C6st-1 KO keratinocytes (Fig. 3g). Thus, gene expression levels of differentiation markers may be low during in vitro culture even in C6st-1 WT keratinocytes. These results suggest that differentiation status of C6st-1 KO keratinocytes is slightly altered by loss of CS-C. However, it is thought that CS-C does not directly regulate keratinocyte differentiation.
Epidermal permeability barrier function is impaired by loss of C6st-1. Loss of filaggrin impacts the skin barrier function 18 . Decreased levels of proteins required for formation of tight junctions and the cornified envelope can also disrupt the skin barrier function. Thus, we performed a dye penetration assay using toluidine blue dye on newborn C6st-1 WT, C6st-1 HE, and C6st-1 KO pups (Fig. 3e). C6st-1 HE and C6st-1 KO pups showed increased dye penetration compared with that observed in C6st-1 WT pups (Fig. 3e). Transepidermal water loss (TEWL) measurements indicated that C6st-1 HE and C6st-1 KO mice showed a significant increase in TEWL values compared with those of C6st-1 WT mice (Fig. 3f). These results suggest that both the "outside-in" and "inside-out" functions of the epidermal permeability barrier were impaired by loss of C6st-1.
Ablation of C6st-1 affects proliferative and inflammatory signaling in the epidermis. We hypothesized that the loss of C6st-1 can cause an abnormality in the proliferative signaling of the epidermis. EGFR signaling is particularly important in regulating proliferation in the epidermis 19 . Phosphorylation of EGFR was significantly elevated in C6st-1 HE and C6st-1 KO epidermis compared with that in C6st-1 WT epidermis ( Fig. 4a, b, and Supplementary Fig. 1). Stimulation of EGFR typically leads to downstream activation of ERK1/2 and STAT3 1, 20 . ERK1/2 and STAT3 were activated in C6st-1 HE and C6st-1 KO epidermis at a higher level than in the C6st-1 WT epidermis ( Fig. 4a, b, and Supplementary Fig. 1). Additionally, the major ligands for EGFR in C6st-1 HE and C6st-1 KO mice are expressed at levels similar to those in C6st-1 WT mice (Fig. 4c). Furthermore, the gene expression levels of several pro-inflammatory cytokines, associated with STAT3 activation, were examined (Fig. 4d). Expression levels of Il1b, Il6, Il7, and Il23 were unaffected by C6st-1 expression level in the dorsal skin of newborn mice (Fig. 4d). In adult mouse skin, expression levels of Il6, Il7, and Il23 were elevated by loss of C6st-1 (Fig. 4d). These results suggest that ablation of C6st-1 does not directly activate intrinsic proinflammatory signaling but may impair epidermal barrier function, producing a state more primed for inflammation.
Keratinocyte proliferation is cell-autonomously stimulated in an EGFR-dependent manner in the absence of C6st-1. We next examined whether EGFR signaling was cell-autonomously activated in C6st-1 HE and C6st-1 KO keratinocytes. As shown in Supplementary Table 2, 7 mol% of C units were contained in CS chains produced by primary keratinocytes. The expression levels of phosphorylated EGFR and phosphorylated STAT3 were significantly elevated in primary keratinocytes derived from C6st-1 HE and C6st-1 KO mice compared with the levels in keratinocytes derived from C6st-1 WT mice (Fig. 5b, c, and Supplementary Fig. 1). Furthermore, C6st-1 HE and C6st-1 KO keratinocytes showed a highly proliferative phenotype compared with that of C6st-1 WT keratinocytes (Fig. 5a). PD153035, a specific inhibitor of EGFR, effectively blocked the phosphorylation of EGFR and STAT3 (Fig. 5b, c, and Supplementary Fig. 1), and proliferation of C6st-1 KO keratinocytes was lowered to the same level as that of C6st-1 WT keratinocytes (Fig. 5a). Notably, the expression levels of phosphorylated EGFR and phosphorylated STAT3 were returned to WT levels by exogenously added CS-C but not by exogenously added chondroitin 4-sulfates (CS-A) (Fig. 5b, c, and Supplementary Fig. 1). Consistent with these results, the hyperproliferative phenotype of C6st-1 KO keratinocytes was recovered after treatment with CS-C (Fig. 5a). These results indicate that proliferation of keratinocytes is cellautonomously stimulated in an EGFR-dependent manner in the absence of C6st-1.
CS-C regulates EGFR signaling. We investigated the mechanism underlying the regulation of EGFR signaling by CS-C. A direct interaction between EGFR and CS-C was visualized and quantified using proximity ligation assay (Fig. 6a, c). Signals indicating direct interactions between EGFR with chondroitin 6-sulfate (EGFR-CS) were detected in C6st-1 WT keratinocytes (Fig. 6a). These signals were diminished after C6st-1 WT keratinocytes were pre-treated with chondroitinase ABC (Chase ABC) (Fig. 6a, c). EGFR-CS signals were decreased in C6st-1 HE and C6st-1 KO keratinocytes (Fig. 6a, c); however, EGFR homodimers in C6st-1 HE and C6st-1 KO keratinocytes were expressed at levels similar to those in C6st-1 WT keratinocytes (Fig. 6b, d). These results suggest that chondroitin 6-sulfates directly bind to EGFR and block EGFR signaling.
Furthermore, we investigated direct interaction of CS-C with EGFR (binding of these proteins to each other) using surface plasmon resonance (SPR; Fig. 6e). While CS-C could bind to EGFR with high affinity, CS-A could not. In addition, we investigated whether CS-C could interact with EGF ligands (Fig. 6f). Neither CS-C nor CS-A bound to EGF.
CS-C regulates human keratinocytes via EGFR signaling. We confirmed that CS-C regulates the proliferation of human keratinocytes via EGFR signaling 21 (Fig. 7). CS chains produced in two human keratinocyte clones, HaCaT and PSVK1, contained C units (Supplementary Table 2). Depletion of CS chains by digestion with Chase ABC accelerated the proliferation of two human keratinocyte clones, HaCaT and PSVK1 (Fig. 7d, e) with concomitant activation of EGFR (Fig. 7b, c, and Supplementary  Fig. 1). In addition, Chase ABC-mediated enhancement of proliferation could be rescued by treatment with the EGFR inhibitor PD153035 (Fig. 7d, e). Furthermore, interactions between CS-C and EGFR were detected in both HaCaT and PSVK1 cells (Fig. 7f,  g).
Imiquimod (IMQ)-treated C6st-1 HE and C6st-1 KO mice are more vulnerable to treatment with IMQ than are C6st-1 WT mice. Psoriasis did not develop as a result of enhanced proliferation of keratinocytes induced by lowered expression of CS-C. We used the imiquimod (IMQ)-induced psoriasis model in WT, C6st-1 HE, and C6st-1 KO mice to investigate the role of CS-C in the epidermis under conditions of inflammation. For this, we swabbed IMQcontaining cream daily for 4 days onto the shaved backs of 6-weekold WT, C6st-1 HE, and C6st-1 KO mice; skin was obtained for analysis on day 3. As shown in Fig. 8, C6st-1 HE and C6st-1 KO mice showed increased levels of acute skin inflammation (manifested by erythema, scaling, and skin induration) and thicker epidermis at day 3 of IMQ treatment compared with those of WT mice (Fig. 8a, b, c). Psoriasis is characterized by dysregulated proliferation and differentiation of keratinocytes. K14 and Ki67 doublepositive keratinocytes were increased in basal layer of WT, C6st-1 HE, and C6st-1 KO mice at 1 day after the first application of IMQ (Fig. 8d, e). Under IMQ-induced inflammation, keratinocyte proliferation showed increased activation in C6st-1 HE and C6st-1 KO mice compared with that in C6st-1 WT mice (Fig. 8e). After WT, C6st-1 HE, and C6st-1 KO mice were treated with IMQ for 3 days, the K14-positive basal layer was markedly thicker in C6st-1 HE and C6st-1 KO mice than in WT mice (Fig. 8f, g) Fig. 4 The expression of phospho-EGFR, phospho-ERK1/2, and phospho-STAT3 is augmented in the epidermis of C6st-1 HE and C6st-1 KO mice compared with that in C6st-1 WT mice. a Proteins extracted from the epidermis of newborn C6st-1 WT, C6st-1 HE, and C6st-1 KO mice were analyzed by immunoblotting using anti-phospho-EGFR (p-EGFR), anti-EGFR, anti-phospho-ERK1/2 (p-ERK1/2), anti-ERK1/2, and anti-phospho-STAT3 (p-STAT3), anti-STAT3, and anti-tubulin antibodies. b Densitometric analysis of the levels of phospho-EGFR, phospho-ERK1/2, and phospho-STAT3 compared with levels of total EGFR, ERK1/2, and STAT3 is shown. Bars represent the means ± S.D. from three independent biological replicates. Statistical significance was determined using one-way ANOVA with Tukey's HSD test. c The expression levels of EGFR ligands (Egf, Areg, Hbegf, and Tgfa) in C6st-1 WT, C6st-1 HE, and C6st-1 KO keratinocytes were analyzed using real-time PCR (n = 4). Values are mean ± S.D. d Expression levels of pro-inflammatory cytokines (Il1b, Il6, Il17, and Il23) in dorsal skin of C6st-1 WT, C6st-1 HE, and C6st-1 KO newborn and adult (13-week-old) mice were analyzed using real-time PCR (n = 3-7). Values represent the mean ± S.D. Statistical significance was determined using one-way ANOVA with Tukey's HSD test.
KO epidermis at 8 h after application of IMQ compared with those in untreated epidermis; EGFR was not activated in C6st-1 WT epidermis ( Fig. 8h and Supplementary Fig. 1). Furthermore, the number of K14 and Ki67 double-positive keratinocytes in C6st-1 HE and C6st-1 KO mice was higher than that in C6st-1 WT mice at 8, 24, and 96 h after application of IMQ (Fig. 8i). IMQ-treated skin in C6st-1 HE and C6st-1 KO mice showed increased epidermal thickening as early as 24 h after application of IMQ, and became further thicker between 24 and 72 h post application of IMQ (Fig. 8j). In contrast, C6st-1 WT epidermis increased in thickness at 72 h after application of IMQ, and became much thicker between 72 and 120 h post application (Fig. 8j). These results indicate that skin hypertrophy in the IMQ-treated C6st-1 HE and C6st-1 KO mice was accelerated compared with C6st-1 WT mice.

Discussion
Proliferation of epidermal keratinocytes is tightly controlled under normal physiological conditions. In many common skin diseases, such as in psoriasis, inflammation disrupts the controlled proliferation of keratinocytes and causes pathological epidermal hyperplasia [22][23][24] . Numerous studies have shown that abnormal proliferation of keratinocytes is triggered by inflammation under pathological conditions. In this study, we examined whether CS, expressed in the extracellular matrix and on cell surface, control proliferation of keratinocytes under normal conditions. The proliferative EGFR-expressing keratinocytes reside in the basal layer, which adheres to an underlying basement membrane that is rich in 6-sulfated CS (Fig. 1g). The expression of  Values are mean ± S.D. with n = 3-4 per group. Statistical significance was determined using one-way ANOVA with Tukey's HSD test. Keratinocytes were treated with 5 μM PD153035, 100 μg/ml CS-A, and 100 μg/ml CS-C as indicated. Cells were lysed and subjected to immunoblotting using b anti-phospho-EGFR and anti-tubulin antibodies, and c anti-phospho-STAT3 and anti-STAT3 antibodies. Tubulin and STAT3 were used as loading controls. The graphs below the blots show relative expression levels of phospho-EGFR and phosphor STAT3 in C6st-1 WT, C6st-1 HE, and C6st-1 KO keratinocytes subjected to the indicated treatments. Values are mean ± S.D. with n = 3 per group. Statistical significance was determined using one-way ANOVA with Tukey's HSD test. chondroitin 6-sulfate (CS-C) in the basement membrane has previously been shown by immunohistochemical analysis using a murine monoclonal antibody against CS-C 25 . Therefore, we analyzed the epidermis of mice lacking C6st-1, which encodes the sulfotransferase involved in biosynthesis of CS-C. Loss of CS-C caused keratinocyte hyperproliferation mediated by hyperactivation of EGFR (Figs. 2 and 4). Keratinocyte proliferation is regulated in a cell-autonomous manner under normal conditions because CS-C produced by basal keratinocytes negatively controls the activation of EGFR (Fig. 5). Normal levels of keratinocyte hyperproliferation were reestablished by exogenously added CS-C (Fig. 5). In addition, hyperproliferation caused by loss of CS-C may indirectly disturb keratinocyte differentiation and thus epidermal barrier function (Fig. 3). Such a sequence of events initiated by loss of CS-C may make keratinocytes more vulnerable to environmental stress (Fig. 9). Thus, we consider that C6st-1KO keratinocytes exist in a primed activation state following inflammatory stimuli. A primed activation state may not necessarily involve increased expression of pro-inflammatory cytokines. This primed state can be conceptualized as "readiness"  state. Thus, C6st-1KO keratinocytes may rapidly respond or have a low-threshold response to inflammatory stimuli. The mitotically active basal layer of the epidermis contains stem cells (basal cells). The basal layer produces and secretes extracellular matrix constituents, including CS-C, which contributes to stem cell niche maintenance. As shown, CS-C localizes primarily to the basement membrane (Fig. 1g). Contact of epidermal stem cells with basement membrane is of fundamental importance in regulating their proliferation and differentiation. Thus, the present study identifies CS-C as a regulatory stem cell niche component within skin.
One manifestation of aging is the decline in epidermal barrier function. In human skin, CS-C is specifically expressed in the basal lamina, and decreases in an age-dependent manner 26 . Results of the present study suggest that a decrease in CS-C impairs barrier function via dysregulating keratinocyte proliferation. Thus, agerelated changes in CS-C may be causally associated with functional changes intrinsic to the aging process of human skin.
CS-C, hyaluronic acid, and collagens serve as substrates supporting the normal functioning of human keratinocytes and fibroblasts. These biomaterials can promote wound healing 27  keratinocytes, our study shows that CS-C suppressed EGFR signaling, thereby regulating keratinocyte cell fate to maintain epidermal homeostasis. Sulfation can occur at various positions within the chondroitin sulfate backbone. Because different sulfation patterns are involved in different biological processes, sulfation patterns are referred to as the sulfation code 28 . We previously reported that C6ST-1 transgenic mice, which have upregulated CS-C, show impaired accumulation of aggrecan in perineuronal nets and lattice-like extracellular matrix structures composed of CS-PGs 29, 30 . Abnormal formation of perineuronal nets decreases the accumulation of orthodenticle homeobox 2 (Otx2) and causes persistent cortical plasticity 29,30 . Additionally, CS-C chains bound to aggrecan contribute to perineuronal nets and glial abnormalities in schizophrenia and bipolar disorders 31 . In this study, CS-C abrogated the EGFR-ERK1/2 and EGFR-STAT3 signaling pathway, suppressed keratinocyte proliferation, and directed epidermal differentiation. As shown in Fig. 5, chondroitin 4-sulfate (CS-A) did not inhibit activation of EGFR and proliferation of keratinocytes. Thus, epidermal homeostasis may require information from the chondroitin 6-sulfation code. Although our in situ proximity ligation assay and SPR assay Fig. 8 CS-C regulates keratinocyte proliferation and skin thickness in imiquimod-induced inflammation model of psoriasis. a Macroscopic phenotypical representation of psoriatic lesions in C6st-1 WT, C6st-1 HE, and C6st-1 KO mice, treated daily with imiquimod. Images acquired at 1, 2, 3, and 4 days after initial treatment. b Back skin sections from C6st-1 WT, C6st-1 HE, and C6st-1 KO, mice treated or untreated with imiquimod on day 3 after initial treatment, stained using H&E. Yellow arrows indicate epidermis. c Graph shows epidermal thickness. Values are mean ± S.D. from four specimens. Statistical significance was determined using one-way ANOVA with Tukey's HSD test. d Dorsal skin sections from C6st-1 WT, C6st-1 HE, and C6st-1 KO mice, treated or untreated with imiquimod for 1 day, were labeled with anti-K14 and anti-Ki67 antibodies. Nuclei were counter-stained with Hoechst33342. Arrows indicate Ki67-positive cells within K14-positive basal layer. e Ki67-positive cells in the K14-positive layer were quantified. Bars represent means ± S.D. from three specimens. Statistical significance was determined using one-way ANOVA with Tukey's HSD test. f Dorsal skin sections from C6st-1 WT, C6st-1 HE, and C6st-1 KO mice treated or untreated with imiquimod on 3 day after initial treatment, were stained with anti-K14 antibody. Nuclei were counter-stained with Hoechst33342. Yellow arrows show K14-positive basal layer. g Thickness of the K14-positive basal layer was measured. Bars represent means ± S.D. from four specimens. Statistical significance was determined using one-way ANOVA with Tukey's HSD test. h Epidermis from C6st-1 WT, C6st-1 HE, and C6st-1 KO mice, treated or untreated with imiquimod for 8 h, analyzed via immunoblotting using anti-phospho-EGFR and anti-tubulin. Graph under blots shows relative expression of phospho-EGFR in C6st-1 WT, C6st-1 HE, and C6st-1 KO epidermis. Values represent the mean ± S.D. with n = 3 or 4 per group. Statistical significance was determined using one-way ANOVA with Tukey's HSD test. i Ki67-positive cells in epidermis at 8, 24, and 96 h after application of imiquimod, quantified as percentage of K14-positive cells. j Epidermal thickness was measured at 8, 24, and 96 h after imiquimod application. Values are mean ± S.D. with n = 3 per group. Statistical significance was determined using one-way ANOVA with Tukey's HSD test.

Normal
A decrease in chondroitin 6-sulfates  indicated that CS-C can interact with EGFR but not EGF (Fig. 6), we cannot exclude the possibility that other EGF ligands such as amphiregulin bind to CS-C. Further studies are needed to identify the proteins involved in the sulfation code of chondroitin sulfates in epidermal homeostasis.
This study will lead to improved understanding of the mechanisms involved in chondroitin 6-sulfate-mediated regulation of keratinocyte proliferation. Further studies will generate new therapeutic strategies, such as topical creams containing chondroitin 6-sulfate, for alleviating psoriasis.
Animals. C6ST-1 knockout mice 33 were housed under specific pathogen-free conditions in an environmentally controlled, bio-clean room at the Institute of Laboratory Animals, Kobe Pharmaceutical University. The mice were maintained on standard mouse/rat chow and on 12-h light/12-h dark cycle. All experiments were conducted according to the institutional ethics guidelines for animal experiments and safety guidelines for gene manipulation experiments of Kobe Pharmaceutical University. All animal procedures were approved by the Kobe Pharmaceutical University Committee on Animal Research and Ethics.
Preparation of mouse keratinocytes. Post-natal day 0-5 neonates from C6ST-1 heterogeneous (HE) mice were humanely euthanized using a CO 2 chamber, after which whole skin was peeled off. The peeled skin was washed with PBS, placed into a tube filled with ice-cold dispase digestion buffer [0.5 mg/ml dispase II (#383-02281, Wako)] in Keratinocyte Media 2 without growth factors (#C-20111, Pro-moCell), and incubated overnight at 4°C. The epidermis was separated from the dermis using forceps, placed into a Petri dish containing keratinocyte Growth Media 2 supplemented with supplementMix (#C-20111, PromoCell), and the cell suspension was filtered. Cells were seeded at the density 0.5 × 10 6 cells/well and cultured at 37°C and 5% CO 2 .
Real-time PCR. Total RNA was extracted from the peeled frozen mouse skin as follows. Frozen mouse skin (20-50 mg) was crushed using a Precellys 24 tissue homogenizer with small beads (CK14) at 5000 rpm for 15 s, using 3 cycles with intervals of 20 s. Crushed tissues were homogenized in 1 ml TRIzole reagent (Invitrogen), and total RNA was isolated according to the manufacturer's protocol. Aliquots (1 μg) of total RNA were digested with 2 IU of RQ1 RNase-free DNase (Promega) for 30 min at 37°C and then incubated for 10 min at 65°C with stop solution (Promega). For reverse transcription, total RNA (0.75 μg) was treated with M-MLV reverse transcriptase (Invitrogen) using random primers (nonadeoxyribonucleotide mixture; pd(N) 9 ) (Takara bio Inc., Shiga, Japan). Quantitative real-time PCR was conducted using FastStart Essential DNA Green Master and a LightCycler 96 (Roche Applied Science) according to the manufacturer's protocols. Housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as an internal control for quantification. The primers used for real-time PCR are listed in Supplementary Table 3.
Measurement of transepidermal water loss (TEWL). TEWL from the dorsal skin of 8-week-old mice was determined with a VAPO METER SWL5001 (Delfin Technologies Ltd.). Mean values of five measurements from each animal were determined. Data are expressed in g/hm 2 as means ± SD from WT and HE mice (n = 4), and KO mice (n = 3).
Outside-in barrier function assay. For Toluidine Blue staining, newborn mice were killed, washed in methanol for 5 min, washed in PBS, incubated for 15 min in 0.1% Toluidine Blue 34 .
Histology. Tissues were fixed in 4% paraformaldehyde and embedded in paraffin. Tissues were sectioned at 5 μm, and stained with hematoxylin and eosin. Images were obtained with a Keyence BZ-8000 fluorescent microscope. For immunostaining, mouse dorsal and tail skin were cryopreserved in OCT compound, sectioned at 10 μm using a cryostat microtome (Leica CM3050S), hybridized to an anti-CS-C antibody (CS-56), and mounted on slides.
Epidermal thickness measurement. Images were acquired using the all-in-one BZ-X700 microscope (Keyence), and vertical distance between the skin surface and dermal epidermal junction (Fig. 1e(a), yellow lines) was measured using ImageJ (NIH). Measurements were performed on nine randomly selected tissue sections per animal. Three randomly chosen fields per tissue sections were used for the analysis and the mean of epidermal thickness was calculated for each section. Epidermal thickness of each animal was then calculated based on the mean values.
Cells were cultured on 13-mm cover glass (Matsunami Glass Ind., Ltd., Osaka, Japan) in a 24-well plate at the density of 1.0 × 10 5 cells/well. Cells were fixed with prechilled methanol for 20 min on ice and then washed with PBS containing 0.1% Tween 20. After blocking with PBS containing 1% BSA and 0.3% Triton X-100 for 1 h at 25°C, cells were incubated overnight at 4°C with the following antibodies: Ki67 (dilution 1:250), and anti-K14 monoclonal antibody (dilution 1:250). Cells were then washed with PBST and incubated with Alexa488-conjugated anti-rabbit IgG antibody (dilution 1:250; Molecular Probes, Eugene, OR) for detection of K14, or with Alexa594-conjugated anti-mouse IgG (dilution 1:250; Molecular Probes, Eugene, OR) for detection of Ki67. Hoechst33342 was used as nuclear counterstain. Images were acquired with a Zeiss LSM 700 confocal laser scanning system.
For the cell proliferation assay, HaCaT and PSVK1 cells were seeded into a 24well plate (for immunofluorescence analysis) and 6-cm dishes (for immunoblotting). The cell proliferation assay proceeded as shown in Fig. 7a. Twenty-four hours post-seeding, cell starvation commenced via replacement of culture medium with serum-free growth medium (for HaCaT cells) or keratinocyte basal medium 2 containing 30 μM Ca 2+ but without supplement mix (for PSVK1 cells). After another 24 h incubation interval, the cell cycle was permitted to progress via addition of serum in the presence of 2.5 munits/mL of chondrotinase ABC and 5 μM PD153035. Cells were subjected to immunofluorescence analysis and immunoblotting at the indicated time (Fig. 7a).
Disaccharide analysis of CS. Isolation and purification of CS from mouse skin, epidermal splits, primary mouse keratinocytes, and human keratinocyte cell lines were performed 8,9,35 . Cells were homogenized in acetone, extracted with acetone three times, and air-dried thoroughly. The dried materials were digested with heatactivated actinase E (10% by weight of dried materials) in 0.1 M borate-sodium, pH 8.0, containing 10 mM CaCl 2 at 55°C for 48 h. The samples were adjusted to 5% v/ v in trichloroacetic acid and centrifuged. The resulting supernatants were extracted with diethyl ether three times to remove the trichloroacetic acid. The aqueous phase was evaporated to dryness, dissolved in 50 mM pyridine acetate, pH 5.0, and subjected to gel filtration on a PD-10 column (GE Healthcare) using 50 mM pyridine acetate, pH 5.0, as an eluent. The flow-through fractions were collected and evaporated to dryness. Purified GAGs were digested with 5 munits of chondroitinase ABC at 37°C for 4 h. Reactions were terminated by boiling for 1 min. The digests were derivatized with a fluorophore 2-aminobenzamide and then analyzed by high performance liquid chromatography.
Imiquimod Treatment. Mice aged 6-10 weeks of age were anesthetized using isoflurane and denuded by shaving. Mice received a daily topical dose of 62.5 mg of 5% imiquimod cream (Beselna Cream; Mochida Pharmaceutical, Tokyo, Japan) on shaved lower backs for 1 day or 3 consecutive days. Consequently, the mice were humanely euthanized and tissues were harvested. Untreated upper back skin was used as a control. Skin adjacent to that collected for histology was used for isolation of total RNA and extraction of proteins.
Statistics and reproducibility. Data are expressed as mean ± standard deviation of the mean. Statistical significance was determined using Tukey-Kramer multiple comparison method and Student's t test as noted in the figure legends. Statistical analyses were performed by KaleidaGraph version 4.5.1. For all analyses, n = number of experiments, and differences are considered statistically significant at p < 0.05. All experiments were reproduced with similar results a minimum of three times, and the exact number of repetitions is provided in the figure legends.