Using Sugar Remodeling to Study Chondroitin Sulfate Function

Abstract

Chondroitin sulfate (CS) chains constitute a class of glycosaminoglycans (GAGs). CS chains are distributed on the surfaces of virtually all cells and throughout most extracellular matrices; they are covalently attached to serine residues of core proteoglycan proteins. CS proteoglycans have been implicated as regulators of a variety of biological events, including cell–cell and cell–matrix adhesion, cell proliferation, morphogenesis, and neurite outgrowth. The functional diversity of CS proteoglycans is mainly attributed to the structural variability of the GAG chains, specifically the CS chains. Despite their relatively simple polysaccharide backbones, CS chains acquire remarkable structural variability via several types of enzymatic modifications, including sulfation. Moreover, the sulfation status of CS chains, chain length, number of CS chains per core protein, or combinations thereof can be finely tuned via CS biosynthetic machinery to specify the structure and function of CS proteoglycans. The term “sugar remodeling” refers to the experimental or therapeutic structural alteration of CS chains via perturbation of specific CS biosynthetic enzymes in cells or living organisms; sugar remodeling is a promising approach to the study of CS chain function. This review focuses on our recent findings regarding CS function which have resulted from studies involving sugar remodeling.

INTRODUCTION

Chondroitin sulfate (CS) and heparan sulfate (HS) are two distinct classes of glycosaminoglycans (GAGs); both are distributed on the surfaces of virtually all cells and throughout most extracellular matrices (ECMs). Individual CS and HS chains are covalently linked to specific serine residues of core proteins; these proteoglycans (PGs) occur as either CSPGs or HSPGs.^1–4) Many of the physiological roles of CSPGs and of HSPGs are attributed to the CS or HS side chains; the core proteins may primarily play the role of scaffolds that present the functional CS or HS chains to potential binding partners (ligands or receptors).^1–4) Gene-targeting technologies for vertebrates and invertebrates have facilitated the elucidation of the physiological functions of HS in animal development and morphogenesis and in regulation of signaling molecules. In contrast, until recently, much less was known about the roles of CS, mainly because of the unexpected redundancy of CS-synthesizing enzymes (described below). This redundancy made the functional analysis of CS more difficult than that of HS.^1–4)

1. BIOSYNTHESIS OF CS CHAINS

CS is a linear polysaccharide comprising repeating disaccharide units (4GlcAβ1-3GalNAcβ1), where GlcA and GalNAc represent glucuronic acid and N-acetylgalactosamine, respectively (Fig. 1). The assembly of CS chains occurs in the endoplasmic reticulum and Golgi compartments and is initiated by the synthesis of the so-called GAG-protein linkage region, GlcAβ1-3Galβ1-3Galβ1-4Xylβ1-O-Ser, which is covalently linked to specific serine residues embedded in any of many core proteins⁴⁾ (Fig. 1). The tetrasaccharide structure of the linkage region is assembled through the sequential, stepwise addition of individual monosaccharide units; a single Xyl (xylose), a Gal (galactose), another Gal, and a single GlcA residue are added successively by the corresponding specific glycosyltransferases: xylosyltransferase (XylT), β1,4-galactosyltransferase I (GalT-I), β1,3-galactosyltransferase II (GalT-II), and β1,3-glucuronyltransferase I (GlcAT-I), respectively⁴⁾ (Fig. 1). Transfer of one GalNAc residue to the nonreducing terminal GlcA residue in the tetrasaccharide linkage region triggers the synthesis of the chondroitin (Chn) backbone; GalNAc transferase I (GalNAcT-I) catalyzes this reaction, which constitutes the rate-limiting step in CS synthesis and the branching point between CSPG and HSPG synthesis.⁴⁾ The repetitive disaccharide region characteristic of CS is synthesized by the alternate additions of GlcA and GalNAc residues; GlcA transferase II (GlcAT-II) and GalNAc transferase II (GalNAcT-II), respectively, catalyze these additions.⁴⁾ Notably, the tetrasaccharide linkage region is shared with HS, another sulfated GAG; moreover, HS also comprises repetitive disaccharide units [(-4GlcAβ1-4GlcNAcα1-)_n], including N-acetylglucosamine (GlcNAc).³⁾ The addition of GlcNAc, instead of GalNAc, to the linkage region triggers the synthesis of HS; therefore, the first N-acetylhexosamine transfer is critical for the selective assembly of CS and HS chains.^1–4)

Fig. 1. Schematic Representation of Deduced and Proposed CS Biosynthetic Machinery

Several glycosyltransferases participate in the synthesis of the common tetrasaccharide linkage region and repeating disaccharide region characteristic of CS chains.⁴⁾ The Chn backbone is further modified by specific sulfotransferases.⁴⁾ The inset shows that a disaccharide unit found in CS chains comprises a GlcA and a GalNAc residue. These sugar residues can be esterified by sulfate at various positions that are indicated by an “S” enclosed by a circle. On the basis of the substrate specificities of the sulfotransferases, the biosynthetic pathways for CS-type disaccharide units can be classified into the initial “4-O-sulfation” or “6-O-sulfation” pathways. The disaccharide units constituting CS chains are classified as O, A, C, D, or E units based on the individual sulfation pattern. 2S, 4S, and 6S represent the 2-O-, 4-O-, and 6-O-sulfate groups, respectively. XYLK, xylosylkinase (Fam20B); XYLP, xylosylphosphatase⁶⁵⁾; XylT, xylosyltransferase; GalT-I, β1,4-galactosyltransferase-I; GalT-II, β1,3-galactosyltransferase-II, GlcAT-I, β1,3-glucuronyltransferase-I; GalNAcT-I, GalNAc transferase-I; GlcAT-II, GlcA transferase-II; GalNAcT-II, GalNAc transferase-II; ChSy, chondroitin synthase; ChPF, chondroitin polymerizing factor; ChGn, chondroitin GalNAc transferase; C4ST, chondroitin 4-O-sulfotransferase; C6ST, chondroitin 6-O-sulfotransferase; UST, uronosyl 2-O-sulfotransferase; GalNAc4S-6ST, GalNAc 4-sulfate 6-O-sulfotransferase.

The biosynthesis of the Chn backbone (i.e., the repeating disaccharide region [(-4GlcAβ1-3GalNAcβ1-)_n]) is catalyzed by six homologous glycosyltransferases, and each has been cloned.^5–15) Based on in vitro substrate specificities, the members of this family are designated Chn synthases (ChSy)-1, -2, and -3, Chn polymerizing factor (ChPF), and Chn GalNAc transferases (ChGn)-1 and -2 (Fig. 1). Each of the three proteins ChSy-1, ChSy-2, and ChSy-3 is a dual-function enzyme with GlcAT-II and GaNAcT-II activity (Fig. 1), but none of these three can polymerize Chn chains alone.^5,7,10,11) However, coexpression of any two of the four proteins (ChSy-1, ChSy-2, ChSy-3, and ChPF) markedly increases CS-synthesizing GlcAT-II and GaNAcT-II activities.⁷⁾ Such coexpression can lead to Chn polymerization on the tetrasaccharide linkage region of α-thrombomodulin,¹⁶⁾ GlcAβ1-3Galβ1-3Galβ1-4Xyl, where the resultant chain lengths vary depending on the particular protein combination.^7,10,11) Thus, Chn polymerization can be achieved, at least in part, by several enzyme complexes, each constituting a Chn polymerase and comprising some pairwise combination of ChSy-1, ChSy-2, ChSy-3, and ChPF (Fig. 1).

In contrast, each of the two other enzymes, ChGn-1 and ChGn-2, possesses both GalNAcT-I and GalNAcT-II activities (Fig. 1) and therefore are believed to catalyze chain initiation and elongation of the Chn backbone.^12–15) The above-mentioned activities of Chn polymerases might indicate that ChGn-1 and ChGn-2 are apparently dispensable during the biosynthesis of a Chn backbone; however, recent evidence has demonstrated that each is essential for working with Chn sulfotransferases to regulate the number or chain length of CS^17,18) (Fig. 1).

The Chn backbone is further modified by sulfation.⁴⁾ Several sulfotransferases responsible for the sulfation of CS chains have been identified.⁴⁾ Each catalyzes the transfer of a sulfate group from 3′-phosphoadenosine 5′-phosphosulfate (PAPS), the universal donor in sulfation reactions, to the respective sulfation site on GalNAc or GlcA residues in the CS chain. A nonsulfated O unit of GalNAc residues serves as the common acceptor substrate for two types of sulfotransferases: chondroitin 4-O-sulfotransferase (C4ST) catalyzes 4-O-sulfation; and chondroitin 6-O-sulfotransferase (C6ST) catalyzes 6-O-sulfation, thereby generating monosulfated A and C units, respectively (Fig. 1). Subsequent sulfation of the A and C units can also occur via GalNAc 4-sulfate 6-O-sulfotransferase (GalNAc4S-6ST) or uronosyl 2-O-sulfotransferase, forming disulfated disaccharide E and D units, respectively (Fig. 1). Therefore, the sulfation of the Chn backbone can be classified into the initial “4-O-sulfation” or “6-O-sulfation” pathways. Each biosynthetic enzyme responsible for CS biosynthesis exhibits a cell type-restricted and spatiotemporally regulated pattern of expression. Consequently, substantial heterogeneity exists even in CS chains isolated from a single species; this heterogeneity may lead to the functional diversity of CSPGs. Therefore, sugar remodeling, defined here as the structural alteration of CS chains via perturbation of specific enzymes in cells or living organisms, is a promising approach to the study of CS chain function (Fig. 2).

Fig. 2. Sugar Remodeling and CS Chains

Depletion or structural alteration of CS chains by perturbation of specific enzymes in cells and living organisms (sugar remodeling) is a promising approach to the study of CS chain function.

2. DISRUPTION OF GLYCOSYLTRANSFERASES TO REMODEL SUGARS IN CELLS OR LIVING ORGANISMS

2.1 Chn/CS Has an Essential Role in Embryonic Cell Division

Caenorhabditis elegans is a genetically tractable model organism that synthesizes HS and Chn, which is identical to CS except that Chn lacks sulfated residues.^19,20) Therefore, we reasoned that C. elegans would have ChSy homologues. Interrogation of C. elegans sequence data with a human ChSy-1 sequence identified an orthologous gene, ChSy (sqv-5, one of eight squashed vulva mutants defective in the formation of the vulva in C. elegans),²¹⁾ in the C. elegans genome. SQV-5, the C. elegans ChSy protein, has GalNAc-I and GaNAcT-II activities and can catalyze Chn chain initiation and elongation.²²⁾ C. elegans seems to have only one ChSy-like gene (sqv-5), and it is orthologous to human ChSy-1; therefore, we reasoned that RNAi-mediated knockdown of sqv-5 would disrupt Chn production. Chn production is significantly reduced in sqv-5 RNA interference (RNAi)-treated worms, and analysis involving four-dimensional microscopy revealed that fertilized eggs laid within 11 to 12 h of RNAi treatment do not complete cell division, particularly cytokinesis.²²⁾ Nuclear division proceeds, but cytokinesis is incomplete, and sister cells remain connected to each other, forming multinucleated cells.^21,22) Speculating that the fundamental mechanism underlying the biosynthesis of Chn in C. elegans might be similar to that in humans, we hypothesized that an orthologue of ChPF might exist in C. elegans. As expected, interrogation of the C. elegans genome sequence with human ChPF sequences identified an orthologue, ChPF (mig-22).²³⁾ As with RNAi-mediated sqv-5 knockdown, functional knockdown of mig-22 causes defects indicative of ChSy (sqv-5) disruption, including a marked decline in Chn levels and cytokinetic regression in early embryogenesis.²³⁾ Phenotypes resulting from sqv-5 or mig-22 knockdown indicate that Chn is absolutely required for early embryonic cytokinesis, but HS is not.

Apparently, the C. elegans genome encodes only two enzymes, SQV-5 and MIG-22 (the C. elegans ChPF protein), that are together necessary and sufficient for Chn biosynthesis. In contrast, multiple glycosyltransferases involved in CS biosynthesis in mammals are highly redundant, and functional analysis of CS biosynthesis is more difficult in mammals than in C. elegans, in part because of this redundancy. Each of three single-gene knockouts ChSy-1 (Chsy1), ChPF (Chpf), or ChGn-1 (Csgalnact1) result in viable, fertile mice, although each knockout genotype results in reduced CS production and/or a sulfation imbalance in CS chains.^24–27) Thus, to clarify the functions of CS in early mammalian embryogenesis, we focused on GlcAT-I (Fig. 1). Because GlcAT-I is a single enzyme that transfers GlcA to the trisaccharide-serine, Galβ1-3Galβ1-4Xylβ1-O-Ser, finalizing the formation of the common linkage region,^28,29) GlcAT-I knockout would result in mutant mice completely lacking CS and HS. Gene knockout of GlcAT-I (B3gat3) results in embryonic lethality before the 8-cell stage due to cytokinetic failure. A complementary analysis in which the bacterial CS-degrading enzyme chondroitinase ABC (ChABC) was used to remove CS selectively from wild-type, 2-cell embryos indicated that the failure of cytokinesis is causally linked to the loss of CS.³⁰⁾ Taken together, these findings show that Chn/CS chains are indispensable for embryonic cell division in a manner that is conserved from worms to mammals.

2.2 CS Has an Essential Role in Pluripotency and Differentiation of Mouse Embryonic Stem Cells

CS is apparently required for proper cell division in early mammalian embryos based on our analyses of GlcAT-I-deficient and ChABC-treated embryos; nevertheless, approximately 7% of the homozygous GlcAT-I-deficient embryos resulting from crosses between GlcAT-I^+/− heterozygotes survive to the implantation stage.³⁰⁾ Thus, we reasoned that these implantation-stage, GlcAT-I-deficient embryos could be used to produce embryonic stem cells that lack GAGs. GlcAT-I-deficient mouse embryonic stem cells have been established, and they completely lack CS and HS.³¹⁾ GlcAT-I-deficient mouse embryonic stem cells fail to exit the self-renewal program and they cannot initiate differentiation even in the absence of leukemia inhibitory factor.³¹⁾ Additionally, GlcAT-I-deficient mouse embryonic stem cells are unable to form embryoid bodies or differentiate into multiple lineages. Moreover, treatment of wild-type mouse embryonic stem cells with ChABC has marked effects on embryoid body formation; in contrast, heparitinase-treated mouse embryonic stem cells seem to develop normally into embryoid bodies.³¹⁾ Taken together, these findings indicate that CS is required to maintain the pluripotency of embryonic stem cells, and that CS differs from HS because it promotes the initial commitment of embryonic stem cells to differentiation. Notably, polysaccharides rich in A disaccharide units (CS-A) and those rich in E disaccharide units (CS-E), but not those rich in C disaccharide units (CS-C), bind to E-cadherin and enhance embryonic stem cell differentiation.³¹⁾ Furthermore, levels of active RhoA and extracellular signal-regulated kinase (ERK) 1/2 phosphorylation are lower in ChABC-treated embryonic stem cells than in controls.³¹⁾ In GlcAT-I-deficient mouse embryonic stem cells, the levels of active RhoA and ERK1/2 phosphorylation are reduced, whereas the levels of active RhoA and ERK1/2 phosphorylation are rescued when GlcAT-I-deficient mouse embryonic stem cells are cultured in the presence of CS-A or CS-E, but not CS-C.³¹⁾ Therefore, CS-A and CS-E are selective ligands for a potential CS receptor, E-cadherin, and these ligands lead to embryonic stem cell differentiation by activating the Rho signaling pathway (Fig. 3). Collectively, these results indicate that CS controls the functional integrity of embryonic stem cells via binding to E-cadherin.

Fig. 3. Mechanisms of Action of CS Chains

1) CS chains act as cell surface receptors, especially in the case of microbial infections. 2) CS chains interact with humoral factors, such as cytokines and morphogens, to present them to the respective receptors or to sequester them away from receptors. These coreceptor/modulator functions may also be important in the generation of morphogenic concentration gradients, such as Wnt-3a gradients. 3) CS chains bind to specific CS receptors on cell surfaces and therefore function as extracellular signaling molecules that can induce intracellular signaling pathways.

2.3 CS Abundance in Myogenic Differentiation/Regeneration

As described above, insufficient CS production at very early stages of embryonic cell division results in abnormal cell death due to frequent reversion of cytokinesis accompanied by the formation of abnormal multinucleated cells.^21–23,30) This finding suggests the paradoxical notion that a temporal decline in CS levels may be required for normal cell fusion processes that occur during the formation of multinucleated syncytia in somatic cells, including the formation of skeletal myogenesis.^32,33) The levels of extracellular and/or pericellular CS chains in differentiating C2C12 myoblast culture are dramatically reduced at the stage when multinucleated myotube formation begins in earnest, and this temporal reduction in CS chain abundance seems to be cell autonomously regulated by hyaluronidase-1, one of the CS catabolic enzymes.³⁴⁾ Forced downregulation of CS, but not of hyaluronan, levels enhances myogenic differentiation in vitro and during myofiber regeneration in vivo in a mouse cardiotoxin-induced injury model.³⁴⁾ Notably, in dystrophin-deficient mdx mice, a model of Duchenne muscular dystrophy, intramuscular injection of ChABC can ameliorate the dystrophic pathology that is characterized by widespread degeneration and recuperative regeneration of myofibers.³⁴⁾ These findings indicate that CS abundance is a crucial determinant of skeletal muscle differentiation and regeneration and for improvement of muscular dystrophy; moreover, these findings may help elucidate the underlying mechanisms of Chn/CS-dependent embryonic cell division.

2.4 Liver Regeneration and Aortic Calcification Are Controlled by Exostosin-Like 2 (EXTL2)

EXTL2 is one of three EXT-like gene products that share significant sequence homology with EXT1 and EXT2, which are HS-synthesizing enzymes.³⁾ Unlike the other family members, EXTL2 can function as an α1,4-N-acetylhexosaminyltransferase in vitro and catalyze the transfer of α-GalNAc and α-GlcNAc residues to the tetrasaccharide linkage region.^35,36) Although no such GalNAcα-capped structure has yet been found in any natural GAG chains, the laborious isolation and analysis of a biosynthetic intermediate of immature GAG chains have recently made it possible to isolate a characteristic oligosaccharide, GlcNAcα1-4GlcAβ1-3Galβ1-3Galβ1-4Xyl(2-O-phosphate), from wild-type mouse liver. Notably, this oligosaccharide could not be isolated from EXTL2-knockout mice.³⁷⁾ Additionally, EXTL2 can transfer a GlcNAc residue to the tetrasaccharide linkage region when this region is phosphorylated by a xylose kinase 1 (FAM20B),³⁸⁾ and the resulting phosphorylated pentasaccharide is not available as an acceptor for CS or HS polymerases; therefore, EXTL2 can terminate chain elongation.³⁷⁾ In addition, the production of GAGs is significantly higher in EXTL2-knockout mice than in wild-type mice.³⁷⁾ Therefore, EXTL2 functions as a suppressor of GAG biosynthesis that is enhanced by the Xyl kinase FAM20B, and this EXTL2-dependent regulation might be a quality control system for CSPG and HSPG synthesis. Although EXTL2-knockout mice are viable and apparently healthy during development and after birth, they are more sensitive than wild-type mice to experimental induction of two separate pathological conditions: 1) liver failure induced by carbon tetrachloride (CCl₄) treatment; and 2) chronic kidney disease (CKD) induced by 5/6th nephrectomy in combination with a high-phosphate diet. Under conditions of CCl₄-induced liver failure, hepatocyte proliferation following CCl₄ treatment was lower in EXTL2-knockout mice than in wild-type mice; consequently, liver regeneration was impaired in EXTL2-knockout mice.³⁹⁾ This reduction in hepatocyte proliferation resulted partially because EXTL2-knockout mice experience less hepatocyte growth factor-mediated signaling than do wild-type mice.³⁹⁾ Under conditions of induced CKD, matrix mineralization in vascular smooth muscle cells (VSMCs) in aortic rings of EXTL2-knockout mice is enhanced relative to that in wild-type mice.⁴⁰⁾ Altered biosynthesis of GAGs in EXTL2-knockout mice affects bone-morphogenetic protein signaling and consequently enhances the differentiation of VSMCs into osteoblasts.⁴⁰⁾ Thus, the EXTL2-dependent mechanism that regulates GAG biosynthesis is important for the maintenance of tissue homeostasis under pathological conditions, and lack of EXTL2 causes GAG overproduction and structural changes in GAGs associated with pathological processes.⁴¹⁾

2.5 Human Peripheral Neuropathies Caused by Mutation in ChGn-1 or ChSy-1

Heterozygous missense mutations in the human ChGn-1 gene (CSGALNACT1) have been reported in patients with either of two peripheral neuropathies: 1) Bell’s palsy; and 2) a type of hereditary motor and sensory neuropathy (HMSN). The causative mutations are H234R in exon 5 and M509R in exon 10, respectively.⁴²⁾ Recombinant enzymes carrying the respective missense mutations exhibit no GalNAcT-II activity.⁴²⁾ This finding indicates that defects in biosynthetic glycosyltransferases such as ChGn-1 may be associated with the pathogenesis of peripheral neuropathies because the genetic defects may compromise recovery from minor trauma. Moreover, a recent survey of ChSy-1 sequences among 310 patients with neurological disorders identified a novel missense mutation (F362S) in exon 3 of ChSy-1.⁴³⁾ The patient had motor neuropathy without any apparent family history of the condition and was diagnosed with HMSN of unknown type. When expressed in cultured cells, a ChSy-1 enzyme with this mutation exhibited decreased CS-synthesizing GlcAT-II and GaNAcT-II activities relative to controls.⁴³⁾ Moreover, the mutant ChSy-1 protein cannot cooperate with wild-type ChGn-1 to regulate the number of CS chains properly. Therefore, these results indicate that the elongation of CS chains may be tightly regulated by cooperation between ChSy-1 and ChGn-1 in peripheral neurons, and that peripheral neuropathies may result from the synthesis of abnormally truncated CS chains.⁴³⁾

3. REMODELING SUGARS IN CELLS OR LIVING ORGANISMS BY PERTURBATING SULFOTRANSFERASES

3.1 Cell Surface Receptors for Pathogens

Several pathogens including parasites, bacteria, and viruses use CS chains on the surfaces of host cells to attach to and infect the respective host cells (Fig. 3). For example, CS chains rich in E units serve as cell surface receptors for herpes simplex virus (HSV) during infection.^44,45) CS-E exhibits potent antiviral activity.⁴⁵⁾ Susceptibility to HSV-1 infection is reduced in sog9-mutant cells that exhibit reduced chondroitin 4-O-sulfotransferase-1 (C4ST-1) expression and production of E units.⁴⁴⁾ Notably, the introduction of C4ST-1 into sog9-mutant cells leads to a significant increase in E disaccharide expression and renders these cells more susceptible to HSV-1 infection,⁴⁴⁾ suggesting that C4ST-1 expression is critical for the synthesis of functional, HSV-sensitive CS chains that contain E units.

3.2 Coreceptors and/or Signal Modulators

Analyses of mouse L cells and sog9-mutants derived from these cells demonstrate that the CS-E-like structure synthesized by C4ST-1 binds with high affinity to Wnt-3a and modulates canonical, β-catenin-dependent Wnt signaling.^46,47) Compared with the parental L cells, sog9-mutant cells are less responsive to Wnt-3a; however, their responsiveness is restored with introduction of C4ST-1. In contrast, the introduction of EXT1, an HS-synthesizing enzyme that is also defective in sog9-mutant cells, does not confer Wnt-3a responsiveness on cells, suggesting that the CS chains formed by C4ST-1 contribute substantially to Wnt-3a signaling.⁴⁶⁾ Molecular interaction analyses showed that CS-E binds strongly to Wnt-3a, and exogenously added CS-E potently inhibited the Wnt-3a-induced accumulation of β-catenin.⁴⁶⁾ Interestingly, constitutive Wnt signaling downregulates C4ST-1 expression in a cell-autonomous and a non-cell-autonomous fashion; this downregulation may lead to structural alterations of extra/pericellular CS chains that reduce the affinity between CS chains and Wnt-3a.⁴⁷⁾ These findings are consistent with a plausible mechanism that could mediate the observed diffusion of Wnt molecules from Wnt-producing cells and highlight the roles of CS chains as coreceptors and/or signal modulators (Fig. 3).

3.3 Human Chondrodysplasia Is Caused by Mutation in Chondroitin 6-O-Sulfotransferase-1

A loss-of-function mutation in CHST3, which encodes the human Chn 6-O-sulfotransferase-1 (C6ST-1) enzyme, is tightly associated with spondyloepiphyseal dysplasia (SED, Omani type), which is a severe chondrodysplasia with progressive spinal involvement.⁴⁸⁾ The identified missense mutation changes an arginine into a glutamine (R304Q) in the putative PAPS binding site and completely abolishes C6ST-1 activity.⁴⁸⁾ As expected, CS chains isolated from fibroblasts derived from patients with SED have significantly fewer C and D units (6-O-sulfated disaccharide units, see Fig. 1) proportionally than those from control fibroblasts.⁴⁸⁾ The proportion of 6-O-sulfated disaccharide units is remarkably decreased, but not zero, which may be attributable to C6ST-2.⁴⁹⁾ Disaccharide composition in urine is consistent with that in fibroblasts; specifically, the patient samples have some C and D units, but significantly fewer than control samples.⁴⁸⁾ Thus, 6-O-sulfated CS formed by C6ST-1 plays critical roles in skeletal development and maintenance in humans.

3.4 CS Sulfation Pattern-Dependent Neuronal Plasticity

Neuronal circuitry can be reorganized based on new experiences. Such experience-dependent neuronal plasticity is most evident during the “critical period,” a limited time in early postnatal development. Thereafter, plasticity declines; this decline is accompanied by the emergence of perineuronal nets (PNNs), which are specialized ECMs that encapsulate cell soma and proximal dendrites of parvalbumin (PV)-positive interneurons.⁵⁰⁾ CSPGs are major components of PNNs.⁵⁰⁾ The digestion of CS chains with ChABC disrupts PNNs and permits reactivation of ocular dominance plasticity in adult animals even after closure of the critical period.⁵¹⁾ These findings indicate that CS moieties of CSPGs are involved in PNN formation and control of critical-period plasticity. However, the importance of CS chain sulfation patterns in this plasticity has not been examined in previous studies because ChABC, which degrades all CS chains regardless of sulfation status, was used exclusively.

Sulfation profiles of CS chains change dramatically during brain development.⁵²⁾ Notably, the proportion of 6-O-sulfation gradually decreases, whereas that of 4-O-sulfation progressively increases. Together, these changes result in incremental change in the 4S/6S ratio during development. Presumably, C4ST and C6ST compete for acceptor Chn disaccharide structures (Fig. 1), and consistent with this notion, transgenic mice that overexpress human C6ST-1 maintain abnormally low 4S/6S ratios. These abnormal ratios result because the transgenic mice have an increased proportion of 6-O-sulfation and a decreased proportion of 4-O-sulfation throughout development relative to wild-type mice.⁵³⁾ Intriguingly, C6ST-1 transgenic mice retain ocular dominance plasticity even in adulthood and have reduced PNN formation and decreased accumulation of Otx2 homeoprotein around PV-positive interneurons.⁵³⁾ Otx2 regulates ocular dominance plasticity via its effect on the maturation of PV-positive interneurons.⁵⁴⁾ Therefore, the 4S:6S ratio of CSPGs may regulate the Otx2-mediated maturation of PV-positive interneurons, which determines the critical period of cortical plasticity. Because the number of PNNs is reduced in individuals with some psychological disorders, such as schizophrenia,⁵⁵⁾ it will be of particular interest to examine the involvement of the 4S : 6S ratio of CSPGs in models of these diseases.

4. EXOGENOUS ADDITION OF CS CHAINS AS EXTRACELLULAR SIGNALING MOLECULES CAN DRIVE SUGAR REMODELING

4.1 Neuronal Extension and Regeneration

CSPGs are the principal components of ECMs in the mammalian central nervous system (CNS) and because they can inhibit axon growth after CNS injury, they have attracted considerable attention.⁵⁰⁾ ChABC treatments can remove CS chains from CSPGs around sites of spinal cord injury and improve axonal regeneration and functional recovery. These effects indicate that the CS moieties are a critical component that determines the inhibitory nature of CSPGs.^56,57) However, CS does not always impede neurite outgrowth. Several CS preparations such as CS-E serve as stimulatory substrates for neurite outgrowth of cultured primary neurons in a cell type-dependent manner.^4,58,59) Such apparently contradictory functions are thought to arise from the structural diversity of CS chains. Thus, we hypothesized that neuronal cells may have distinct CS-recognition mechanisms and consequently multiple CS receptors (Fig. 3). Contactin (CNTN)-1 was identified as the neuronal cell surface receptor for CS-E which is responsible for CS-E-mediated neurite outgrowth. CNTN-1 is a glycosylphosphatidylinositol-anchored cell adhesion molecule of the immunoglobulin superfamily.⁶⁰⁾ CS-E, but not CS-A, CS-C, or HS, binds CNTN-1 and induces intracellular signaling downstream of CNTN-1, thereby causing neurite outgrowth.⁶⁰⁾ This original study provided important evidence that specific CS chains behave as extracellular signaling molecules that can induce intracellular signaling cascades.

4.2 Osteogenesis

Mouse MC3T3-E1 cells constitute a well-characterized osteoblastic cell line, and MC3T3-E1 cultures can mimic intramembranous ossification processes in vitro. In particular, MC3T3-E1 cells express N-cadherin and cadherin-11, and such cadherin-mediated cell–cell contact is critical for the onset of osteogenic differentiation in vitro and in vivo.^61,62) Biochemical analysis of CS chains in differentiating MC3T3-E1 cultures revealed an increase in CS chains with a relatively high proportion of E units.^63,64) CS-E, but not CS-A, binds to N-cadherin and cadherin-11 in the presence of divalent cations and enhances osteogenic differentiation.⁶³⁾ Importantly, even in a low-density culture of MC3T3-E1 cells, where cadherin-mediated cell–cell contact does not occur, exogenous application of CS-E polysacchride and its defined hexasaccharides can activate the intracellular signaling required for osteogenesis; importantly, these reactions are completely inhibited by antibodies that bind N-cadherin and cadherin-11.⁶³⁾ Moreover, the introduction of GalNAc4S-6ST via transfection results in the overproduction of E units in MC3T3-E1 cells and renders the cells more adhesive to N-cadherin/cadherin-11-coated plates.⁶³⁾ These findings indicate that CS-E is a selective ligand for the potential CS receptors, N-cadherin and cadherin-11, leading to osteogenesis in MC3T3-E1 cells, and thus can be useful as an osteogenesis-promoting agent in patients with osteoporosis (Fig. 3). CS chains bind to specific CS receptors on cell surfaces and therefore function as extracellular signaling molecules that can induce intracellular signaling pathways. This new mode of action of CS chains enables us to establish innovative methods for pharmacological and/or therapeutic manipulations of CS chains.

Acknowledgment

This review of our recent findings commemorates my acceptance of the Pharmaceutical Society of Japan Award for Divisional Scientific Promotions. I would like to thank all collaborators, especially Drs. Satomi Nadanaka, Tadahisa Mikami, Tomomi Izumikawa, Shinji Miyata, and Toshiyasu Koike at the Department of Biochemistry, Kobe Pharmaceutical University, Japan, for major contributions to the results described in this review. This work was supported in part by Grants-in-Aid for Scientific Research (B), for Challenging Exploratory Research, and for Scientific Research on Innovative Areas and by the Supported Program for the Strategic Research Foundation at Private Universities from the Ministry of Education, Culture, Sports, Science and Technology of Japan.

REFERENCES

• 1) Sugahara K, Kitagawa H. Recent advances in the study of the biosynthesis and functions of sulfated glycosaminoglycans. Curr. Opin. Struct. Biol., 10, 518–527 (2000).
• 2) Sugahara K, Mikami T, Uyama T, Mizuguchi S, Nomura K, Kitagawa H. Recent advances in the structural biology of chondroitin sulfate and dermatan sulfate. Curr. Opin. Struct. Biol., 13, 612–620 (2003).
• 3) Nadanaka S, Kitagawa H. Heparan sulphate biosynthesis and disease. J. Biochem., 144, 7–14 (2008).
• 4) Mikami T, Kitagawa H. Biosynthesis and function of chondroitin sulfate. Biochim. Biophys. Acta, 1830, 4719–4733 (2013).
• 5) Kitagawa H, Uyama T, Sugahara K. Molecular cloning and expression of a human chondroitin synthase. J. Biol. Chem., 276, 38721–38726 (2001).
• 6) Gotoh M, Yada T, Sato T, Akashima T, Iwasaki H, Mochizuki H, Inaba N, Togayachi A, Kudo T, Watanabe H, Kimata K, Narimatsu H. Molecular cloning and characterization of a novel chondroitin sulfate glucuronyltransferase that transfers glucuronic acid to N-acetylgalactosamine. J. Biol. Chem., 277, 38179–38188 (2002).
• 7) Kitagawa H, Izumikawa T, Uyama T, Sugahara K. Molecular cloning of a chondroitin polymerizing factor that cooperates with chondroitin synthase for chondroitin polymerization. J. Biol. Chem., 278, 23666–23671 (2003).
• 8) Yada T, Gotoh M, Sato T, Shionyu M, Go M, Kaseyama H, Iwasaki H, Kikuchi N, Kwon YD, Togayachi A, Kudo T, Watanabe H, Narimatsu H, Kimata K. Chondroitin sulfate synthase-2. Molecular cloning and characterization of a novel human glycosyltransferase homologous to chondroitin sulfate glucuronyltransferase, which has dual enzymatic activities. J. Biol. Chem., 278, 30235–30247 (2003).
• 9) Yada T, Sato T, Kaseyama H, Gotoh M, Iwasaki H, Kikuchi N, Kwon YD, Togayachi A, Kudo T, Watanabe H, Narimatsu H, Kimata K. Chondroitin sulfate synthase-3. Molecular cloning and characterization. J. Biol. Chem., 278, 39711–39725 (2003).
• 10) Izumikawa T, Uyama T, Okuura Y, Sugahara K, Kitagawa H. Involvement of chondroitin sulfate synthase-3 (chondroitin synthase-2) in chondroitin polymerization through its interaction with chondroitin synthase-1 or chondroitin-polymerizing factor. Biochem. J., 403, 545–552 (2007).
• 11) Izumikawa T, Koike T, Shiozawa S, Sugahara K, Tamura J, Kitagawa H. Identification of chondroitin sulfate glucuronyltransferase as chondroitin synthase-3 involved in chondroitin polymerization: chondroitin polymerization is achieved by multiple enzyme complexes consisting of chondroitin synthase family members. J. Biol. Chem., 283, 11396–11406 (2008).
• 12) Uyama T, Kitagawa H, Tamura J, Sugahara K. Molecular cloning and expression of human chondroitin N-acetylgalactosaminyltransferase: the key enzyme for chain initiation and elongation of chondroitin/dermatan sulfate on the protein linkage region tetrasaccharide shared by heparin/heparan sulfate. J. Biol. Chem., 277, 8841–8846 (2002).
• 13) Gotoh M, Sato T, Akashima T, Iwasaki H, Kameyama A, Mochizuki H, Yada T, Inaba N, Zhang Y, Kikuchi N, Kwon YD, Togayachi A, Kudo T, Nishihara S, Watanabe H, Kimata K, Narimatsu H. Enzymatic synthesis of chondroitin with a novel chondroitin sulfate N-acetylgalactosaminyltransferase that transfers N-acetylgalactosamine to glucuronic acid in initiation and elongation of chondroitin sulfate synthesis. J. Biol. Chem., 277, 38189–38196 (2002).
• 14) Sato T, Gotoh M, Kiyohara K, Akashima T, Iwasaki H, Kameyama A, Mochizuki H, Yada T, Inaba N, Togayachi A, Kudo T, Asada M, Watanabe H, Imamura T, Kimata K, Narimatsu H. Differential roles of two N-acetylgalactosaminyltransferases, CSGalNAcT-1, and a novel enzyme, CSGalNAcT-2. Initiation and elongation in synthesis of chondroitin sulfate. J. Biol. Chem., 278, 3063–3071 (2003).
• 15) Uyama T, Kitagawa H, Tanaka J, Tamura J, Ogawa T, Sugahara K. Molecular cloning and expression of a second chondroitin N-acetylgalactosaminyltransferase involved in the initiation and elongation of chondroitin/dermatan sulfate. J. Biol. Chem., 278, 3072–3078 (2003).
• 16) Nadanaka S, Kitagawa H, Sugahara K. Demonstration of the immature glycosaminoglycan tetrasaccharide sequence GlcAβ1-3Galβ1-3Galβ1-4Xyl on recombinant soluble human α-thrombomodulin. A possible mechanism generating ‘part-time’ proteoglycan. J. Biol. Chem., 273, 33728–33734 (1998).
• 17) Izumikawa T, Okuura Y, Koike T, Sakoda N, Kitagawa H. Chondroitin 4-O-sulfotransferase-1 regulates the chain length of chondroitin sulfate in co-operation with chondroitin N-acetylgalactosaminyltransferase-2. Biochem. J., 434, 321–331 (2011).
• 18) Izumikawa T, Koike T, Kitagawa H. Chondroitin 4-O-sulfotransferase-2 regulates the number of chondroitin sulfate chains initiated by chondroitin N-acetylgalactosaminyltransferase-1. Biochem. J., 441, 697–705 (2012).
• 19) Yamada S, Van Die I, Van den Eijnden DH, Yokota A, Kitagawa H, Sugahara K. Demonstration of glycosaminoglycans in Caenorhabditis elegans. FEBS Lett., 459, 327–331 (1999).
• 20) Toyoda H, Kinoshita-Toyoda A, Selleck SB. Structural analysis of glycosaminoglycans in Drosophila and Caenorhabditis elegans and demonstration that tout-velu, a Drosophila gene related to EXT tumor suppressors, affects heparan sulfate in vivo. J. Biol. Chem., 275, 2269–2275 (2000).
• 21) Hwang HY, Olson SK, Esko JD, Horvitz HR. Caenorhabditis elegans early embryogenesis and vulval morphogenesis require chondroitin biosynthesis. Nature, 423, 439–443 (2003).
• 22) Mizuguchi S, Uyama T, Kitagawa H, Nomura KH, Dejima K, Gengyo-Ando K, Mitani S, Sugahara K, Nomura K. Chondroitin proteoglycans are involved in cell division of Caenorhabditis elegans. Nature, 423, 443–448 (2003).
• 23) Izumikawa T, Kitagawa H, Mizuguchi S, Nomura KH, Nomura K, Tamura J, Gengyo-Ando K, Mitani S, Sugahara K. Nematode chondroitin polymerizing factor showing cell-/organ-specific expression is indispensable for chondroitin synthesis and embryonic cell division. J. Biol. Chem., 279, 53755–53761 (2004).
• 24) Wilson DG, Phamluong K, Lin WY, Barck K, Carano RA, Diehl L, Peterson AS, Martin F, Solloway MJ. Chondroitin sulfate synthase 1 (Chsy1) is required for bone development and digit patterning. Dev. Biol., 363, 413–425 (2012).
• 25) Ogawa H, Hatano S, Sugiura N, Nagai N, Sato T, Shimizu K, Kimata K, Narimatsu H, Watanabe H. Chondroitin sulfate synthase-2 is necessary for chain extension of chondroitin sulfate but not critical for skeletal development. PLoS ONE, 7, e43806 (2012).
• 26) Watanabe Y, Takeuchi K, Higa Onaga S, Sato M, Tsujita M, Abe M, Natsume R, Li M, Furuichi T, Saeki M, Izumikawa T, Hasegawa A, Yokoyama M, Ikegawa S, Sakimura K, Amizuka N, Kitagawa H, Igarashi M. Chondroitin sulfate N-acetylgalactosaminyltransferase-1 is required for normal cartilage development. Biochem. J., 432, 47–55 (2010).
• 27) Sato T, Kudo T, Ikehara Y, Ogawa H, Hirano T, Kiyohara K, Hagiwara K, Togayachi A, Ema M, Takahashi S, Kimata K, Watanabe H, Narimatsu H. Chondroitin sulfate N-acetylgalactosaminyltransferase 1 is necessary for normal endochondral ossification and aggrecan metabolism. J. Biol. Chem., 286, 5803–5812 (2011).
• 28) Kitagawa H, Tone Y, Tamura J, Neumann KW, Ogawa T, Oka S, Kawasaki T, Sugahara K. Molecular cloning and expression of glucuronyltransferase I involved in the biosynthesis of the glycosaminoglycan–protein linkage region of proteoglycans. J. Biol. Chem., 273, 6615–6618 (1998).
• 29) Wei G, Bai X, Sarkar AK, Esko JD. Formation of HNK-1 determinants and the glycosaminoglycan tetrasaccharide linkage region by UDP-GlcUA:Galactose beta1, 3-glucuronosyltransferases. J. Biol. Chem., 274, 7857–7864 (1999).
• 30) Izumikawa T, Kanagawa N, Watamoto Y, Okada M, Saeki M, Sakano M, Sugahara K, Sugihara K, Asano M, Kitagawa H. Impairment of embryonic cell division and glycosaminoglycan biosynthesis in glucuronyltransferase-I-deficient mice. J. Biol. Chem., 285, 12190–12196 (2010).
• 31) Izumikawa T, Sato B, Kitagawa H. Chondroitin sulfate is indispensable for pluripotency and differentiation of mouse embryonic stem cells. Sci. Rep., 4, 3701 (2014).
• 32) Rochlin K, Yu S, Roy S, Baylies MK. Myoblast fusion: when it takes more to make one. Dev. Biol., 341, 66–83 (2010).
• 33) Pavlath GK. Spatial and functional restriction of regulatory molecules during mammalian myoblast fusion. Exp. Cell Res., 316, 3067–3072 (2010).
• 34) Mikami T, Koyama S, Yabuta Y, Kitagawa H. Chondroitin sulfate is a crucial determinant for skeletal muscle development/regeneration and improvement of muscular dystrophies. J. Biol. Chem., 287, 38531–38542 (2012).
• 35) Kitagawa H, Shimakawa H, Sugahara K. The tumor suppressor EXT-like gene EXTL2 encodes an α1,4-N-acetylhexosaminyltransferase that transfers N-acetylgalactosamine and N-acetylglucosamine to the common glycosaminoglycan–protein linkage region. The key enzyme for the chain initiation of heparan sulfate. J. Biol. Chem., 274, 13933–13937 (1999).
• 36) Kitagawa H, Kano Y, Shimakawa H, Goto F, Ogawa T, Okabe H, Sugahara K. Identification and characterization of a novel UDP-GalNAc : GlcAbeta-R alpha1,4-N-acetylgalactosaminyltransferase from a human sarcoma cell line. Glycobiology, 9, 697–703 (1999).
• 37) Nadanaka S, Zhou S, Kagiyama S, Shoji N, Sugahara K, Sugihara K, Asano M, Kitagawa H. EXTL2, a member of EXT family of tumor suppressors, controls glycosaminoglycan biosynthesis in a xylose kinase-dependent manner. J. Biol. Chem., 288, 9321–9333 (2013).
• 38) Koike T, Izumikawa T, Tamura J, Kitagawa H. FAM20B is a kinase that phosphorylates xylose in the glycosaminoglycan–protein linkage region. Biochem. J., 421, 157–162 (2009).
• 39) Nadanaka S, Kagiyama S, Kitagawa H. Roles of EXTL2, a member of the EXT family of tumour suppressors, in liver injury and regeneration processes. Biochem. J., 454, 133–145 (2013).
• 40) Purnomo E, Emoto N, Nugrahaningsih DAA, Nakayama K, Yagi K, Heiden S, Nadanaka S, Kitagawa H, Hirata K. Glycosaminoglycan overproduction in the aorta increases aortic calcification in murine chronic kidney disease. J. Am. Heart Assoc., 2, e000405 (2013).
• 41) Nadanaka S, Kitagawa H. EXTL2 controls liver regeneration and aortic calcification through xylose kinase-dependent regulation of glycosaminoglycan biosynthesis. Matrix Biol., 35, 18–24 (2014).
• 42) Saigoh K, Izumikawa T, Koike T, Shimizu J, Kitagawa H, Kusunoki S. Chondroitin beta-1,4-N-acetylgalactosaminyltransferase-1 missense mutations are associated with neuropathies. J. Hum. Genet., 56, 143–146 (2011).
• 43) Izumikawa T, Saigoh K, Shimizu J, Tsuji S, Kusunoki S, Kitagawa H. Chondroitin synthase-1 (ChSy-1) missense mutation in a patient with neuropathy impairs the elongation of chondroitin sulfate chains initiated by chondroitin N-acetylgalactosaminyltransferase-1. Biochim. Biophys. Acta, 1830, 4806–4812 (2013).
• 44) Uyama T, Ishida M, Izumikawa T, Trybala E, Tufaro F, Bergström T, Sugahara K, Kitagawa H. Chondroitin 4-O-sulfotransferase-1 regulates E disaccharide expression of chondroitin sulfate required for herpes simplex virus infectivity. J. Biol. Chem., 281, 38668–38674 (2006).
• 45) Bergefall K, Trybala E, Johansson M, Uyama T, Naito S, Yamada S, Kitagawa H, Sugahara K, Bergström T. Chondroitin sulfate characterized by the E-disaccharide unit is a potent inhibitor of herpes simplex virus infectivity and provides the virus binding sites on gro2C cells. J. Biol. Chem., 280, 32193–32199 (2005).
• 46) Nadanaka S, Ishida M, Ikegami M, Kitagawa H. Chondroitin 4-O-sulfotransferase-1 modulates Wnt-3a signaling through control of E disaccharide expression of chondroitin sulfate. J. Biol. Chem., 283, 27333–27343 (2008).
• 47) Nadanaka S, Kinouchi H, Taniguchi-Morita K, Tamura J, Kitagawa H. Down-regulation of chondroitin 4-O-sulfotransferase-1 by Wnt signaling triggers diffusion of Wnt-3a. J. Biol. Chem., 286, 4199–4208 (2011).
• 48) Thiele H, Sakano M, Kitagawa H, Sugahara K, Rajab A, Höhne W, Ritter H, Leschik G, Nürnberg P, Mundlos S. Loss of chondroitin 6-O-sulfotransferase-1 function results in severe human chondrodysplasia with progressive spinal involvement. Proc. Natl. Acad. Sci. U.S.A., 101, 10155–10160 (2004).
• 49) Kitagawa H, Fujita M, Ito N, Sugahara K. Molecular cloning and expression of a novel chondroitin 6-O-sulfotransferase. J. Biol. Chem., 275, 21075–21080 (2000).
• 50) Galtrey CM, Fawcett JW. The role of chondroitin sulfate proteoglycans in regeneration and plasticity in the central nervous system. Brain Res. Rev., 54, 1–18 (2007).
• 51) Pizzorusso T, Medini P, Berardi N, Chierzi S, Fawcett JW, Maffei L. Reactivation of ocular dominance plasticity in the adult visual cortex. Science, 298, 1248–1251 (2002).
• 52) Kitagawa H, Tsutsumi K, Tone Y, Sugahara K. Developmental regulation of the sulfation profile of chondroitin sulfate chains in the chicken embryo brain. J. Biol. Chem., 272, 31377–31381 (1997).
• 53) Miyata S, Komatsu Y, Yoshimura Y, Taya C, Kitagawa H. Persistent cortical plasticity by upregulation of chondroitin 6-sulfation. Nat. Neurosci., 15, 414–422, S1–S2 (2012).
• 54) Sugiyama S, Di Nardo AA, Aizawa S, Matsuo I, Volovitch M, Prochiantz A, Hensch TK. Experience-dependent transfer of Otx2 homeoprotein into the visual cortex activates postnatal plasticity. Cell, 134, 508–520 (2008).
• 55) Lewis DA, Hashimoto T, Volk DW. Cortical inhibitory neurons and schizophrenia. Nat. Rev. Neurosci., 6, 312–324 (2005).
• 56) Bradbury EJ, Moon LD, Popat RJ, King VR, Bennett GS, Patel PN, Fawcett JW, McMahon SB. Chondroitinase ABC promotes functional recovery after spinal cord injury. Nature, 416, 636–640 (2002).
• 57) Silver J, Miller JH. Regeneration beyond the glial scar. Nat. Rev. Neurosci., 5, 146–156 (2004).
• 58) Gama CI, Tully SE, Sotogaku N, Clark PM, Rawat M, Vaidehi N, Goddard WA 3rd, Nishi A, Hsieh-Wilson LC. Sulfation patterns of glycosaminoglycans encode molecular recognition and activity. Nat. Chem. Biol., 2, 467–473 (2006).
• 59) Purushothaman A, Fukuda J, Mizumoto S, ten Dam GB, van Kuppevelt TH, Kitagawa H, Mikami T, Sugahara K. Functions of chondroitin sulfate/dermatan sulfate chains in brain development: critical roles of E and iE disaccharide units recognized by a single chain antibody GD3G7. J. Biol. Chem., 282, 19442–19452 (2007).
• 60) Mikami T, Yasunaga D, Kitagawa H. Contactin-1 is a functional receptor for neuroregulatory chondroitin sulfate-E. J. Biol. Chem., 284, 4494–4499 (2009).
• 61) Cheng SL, Shin CS, Towler DA, Civitelli R. A dominant negative cadherin inhibits osteoblast differentiation. J. Bone Miner. Res., 15, 2362–2370 (2000).
• 62) Kii I, Amizuka N, Shimomura J, Saga Y, Kudo A. Cell–cell interaction mediated by cadherin-11 directly regulates the differentiation of mesenchymal cells into the cells of the osteo-lineage and the chondro-lineage. J. Bone Miner. Res., 19, 1840–1849 (2004).
• 63) Koike T, Izumikawa T, Tamura J, Kitagawa H. Chondroitin sulfate-E fine-tunes osteoblast differentiation via ERK1/2, Smad3 and Smad1/5/8 signaling by binding to N-cadherin and cadherin-11. Biochem. Biophys. Res. Commun., 420, 523–529 (2012).
• 64) Miyazaki T, Miyauchi S, Tawada A, Anada T, Matsuzaka S, Suzuki O. Oversulfated chondroitin sulfate-E binds to BMP-4 and enhances osteoblast differentiation. J. Cell. Physiol., 217, 769–777 (2008).
• 65) Koike T, Izumikawa T, Sato B, Kitagawa H. Identification of phosphatase that dephosphorylates xylose in the glycosaminoglycan–protein linkage region of proteoglycans. J. Biol. Chem., 289, 6695–6708 (2014).