Neuritogenic Activity of Chondroitin/Dermatan Sulfate Hybrid Chains of Embryonic Pig Brain and Their Mimicry from Shark Liver

Accumulating evidence suggests the involvement of chondroitin sulfate (CS) and dermatan sulfate (DS) hybrid chains in the brain's development and critical roles for oversulfated disaccharides and IdoUA residues in the growth factor-binding and neuritogenic activities of these chains. In the pursuit of sources of CS/DS with unique structures, neuritogenic activity, and therapeutic potential, two novel CS/DS preparations were isolated from shark liver by anion exchange chromatography. The major (80%) low sulfated and minor (20%) highly sulfated fractions had an average molecular mass of 3.8–38.9 and 75.7 kDa, respectively. Digestion with various chondroitinases (CSases) revealed a large panel of disaccharides with either GlcUA or IdoUA scattered along the polysaccharide chains in both of the fractions. The higher Mr fraction, richer in IdoUA(2-O-sulfate)α1–3GalNAc(4-O-sulfate) and GlcUAβ/IdoUAα1–3GalNAc(4,6-O-disulfate) units, exerted greater neurite outgrowth-promoting (NOP) activity and better promoted the binding of various heparin-binding growth factors, including pleiotrophin (PTN), midkine, recombinant human heparin-binding epidermal growth factor-like growth factor, VEGF165, fibroblast growth factor-2, fibroblast growth factor-7, and hepatocyte growth factor (HGF). These activities were largely abolished by digestion with CSase ABC or B but only moderately affected by a mixture of CSases AC-I and AC-II. In addition, the NOP activity of the larger fraction was markedly reduced by desulfation with alkali, suggesting a role for the 2-O-sulfate of IdoUA(2-O-sulfate)α1–3GalNAc(4-O-sulfate). The NOP activity of the higher molecular weight fraction and that of the embryonic pig brain-derived CS/DS fraction were also sup pressed to a large extent by antibodies against HGF, PTN, and their individual receptors cMet and anaplastic lymphoma kinase, revealing the involvement of the HGF and PTN signaling pathways in the activity.

To search for sources of CS/DS chains with therapeutic potential, CS/DS hybrid chains were purified from shark liver and found to have a unique structure and strong NOP activity. Further study revealed the molecular mechanism of the NOP activity to involve the signaling pathway of not only PTN but also hepatocyte growth factor (HGF). . Embryonic pig brainderived CS/DS (E-CS/DS) and its high affinity fraction (E-CS/ DS-H) were prepared as described previously (32). Recombinant human (rh) pleiotrophin (PTN) expressed in E. coli and rh-vascular endothelial growth factor-165 (VEGF 165 ) expressed in insect cells were from RELIA Tech GmbH (Braunschweig, Germany). rh-Midkine (MK) expressed in Escherichia coli and rh-fibroblast growth factor (FGF)-1 (or acidic FGF) expressed in E. coli were from PeproTech EC Ltd. (London, UK). rh-FGF-2 (or basic FGF) expressed in E. coli was from Genzyme TECHNE (Minneapolis, MN). rh-Heparin-binding epidermal growth factor-like growth factor (HB-EGF) and rhhepatocyte growth factor/scatter factor (HGF/SF) expressed in Sf21 insect cells, rh-keratinocyte growth factor (KGF/FGF-7) expressed in E. coli, and anti-mouse HGF receptor IgG were obtained from R&D Systems (Minneapolis, MN). 1,9-Dimethylmethylene blue was from Aldrich. Polyclonal goat anti-rh-PTN IgG and polyclonal goat anti-rh-MK IgG were obtained from Genzyme/Techne (Cambridge, MA). Anti-FGF-2/basic FGF, clone bFM-1, was from Upstate Biotechnology, Inc. (New York, NY). Polyclonal goat IgG against mouse anaplastic lymphoma kinase (ALK T-18), was from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Polyclonal anti-rat HGF rabbit IgG was provided by Prof. Toshikazu Nakamura (Osaka University, Osaka, Japan). Purified serum IgG from mouse, goat, and rabbit were obtained from Sigma. Actinase E was from Kaken Pharmaceutical Co. (Tokyo, Japan). All other chemicals and reagents were of the highest quality available.

Materials-Livers of
Extraction of SL-CS/DS-Livers of blue shark (P. glauca), were dehydrated and delipidated by extraction with acetone, air-dried, and used for extraction of GAGs essentially as described previously (28) with some modifications. Briefly, 65 g of the acetone powder, corresponding to 163 g of the wet tissue, was treated with actinase E, followed by 5% trichloroacetic acid to precipitate residual proteins and peptides and with ether to extract trichloroacetic acid. To extract GAGs exhaustively, the precipitate obtained with trichloroacetic acid was treated with 0.5 M NaOH at 4°C for 20 h and then neutralized with 1 M acetic acid before being precipitated with trichloroacetic acid, extracted with ether, and combined with the GAG extract obtained by actinase digestion. A crude GAG fraction was recovered from the combined extract by precipitation with 80% ethanol containing 5% sodium acetate at 4°C overnight. The yield was 9 g, containing 416 mg of GAG based on the carbazole reaction.
Purification of SL-CS/DS-The crude GAG fraction (1.5 g) was loaded on a DEAE-Sephadex column (15 ϫ 300 mm) preequilibrated with 0.3 M phosphate buffer, pH 6.0, containing 0.2 M NaCl. After the column was washed with the equilibration buffer, GAGs were eluted with the same buffer containing 2.0 M NaCl, dialyzed against water, and concentrated to dryness (the yield was 25 mg). This sample was subjected to a nitrous acid treatment (pH 1.5) to remove heparin/HS as described previously (34), and the resultant HS fragments were removed by gel filtration on a Superdex 75 column (10 ϫ 300 mm; Amersham Biosciences) eluted with 0.2 M NH 4 HCO 3 at a flow rate of 0.4 ml/min. The elution was monitored by absorption at 210 nm. The fraction eluted in the void volume was pooled and freezedried repeatedly by reconstituting in water to remove NH 4 HCO 3 . This CS/DS preparation was fractionated by anion exchange chromatography on an Accell QMA Plus cartridge (Waters, Milford, MA) and eluted stepwise with 0.3 M phosphate buffers (pH 6.0) containing 0.2, 1.0, 1.5, and 2.0 M NaCl. The fractions obtained by elution with 1.0 and 1.5 M NaCl, referred to as SL-CS/DS (1.0 M) and SL-CS/DS (1.5 M), respectively, were desalted using a PD-10 column (Amersham Biosciences), and each fraction was analyzed by the carbazole reaction for the amount of CS/DS (35). Finally, the SL-CS/DS preparations were passed through a Sep-PakC 18 cartridge (Waters) to remove peptides.
Determination of the Disaccharide Composition and Molecular Mass-An aliquot (1 g as GAG) of SL-CS/DS (1.0 M) or (1.5 M) was subjected to digestion with CSase ABC, a mixture of CSases AC-I and AC-II, or CSase B. Each digest was labeled with 2-aminobenzamide (2AB) and subjected to anion exchange HPLC on an amine-bound silica PA-03 column (YMC-Pack PA, Kyoto, Japan) as described previously (28). To determine the molecular mass, an aliquot (5 g as GlcUA) of SL-CS/DS (1.0 and 1.5 M fractions) was chromatographed by gel filtration on a Superdex TM 200 column (10 ϫ 300 mm; Amersham Biosciences), which had been calibrated using a series of size-defined commercial polysaccharides (36). The sample was eluted with 0.2 M ammonium bicarbonate at a flow rate of 0.3 ml/min for 90 min. Fractions were collected at 3-min intervals, freeze-dried, and dissolved in 100 l of water. An aliquot was utilized for estimating the total amount of GAGs using 1,9-di-methylmethylene blue according to the method of Chandrasekhar et al. (37), except that the absorbance was recorded at 525 nm.
Interaction Analysis-Inhibition of PTN binding to a PTN high affinity fraction derived from embryonic pig brain (E-CS/ DS-H) was examined using a BIAcore J system (BIAcore AB, Uppsala, Sweden). E-CS/DS-H was immobilized on a sensor chip as previously described (32). PTN (100 ng) was mixed with 0.5 g of each tested GAG preparation (SL-CS/DS fractions or commercial CS/DS preparations) and incubated for 15 min at room temperature prior to injection onto the surface of an E-CS/DS-H-immobilized sensor chip. Results are expressed as relative percentages of inhibition based on the binding of PTN to E-CS/DS-H in the absence of inhibitor as 100%. To examine the interaction with various growth factors, SL-CS/DS (1.5 M) was immobilized on a sensor chip as reported earlier (38). For kinetic analysis, various concentrations of growth factors were injected onto the surface of this sensor chip in the running buffer (HBS-EP, pH 7.4, BIAcore AB) with a medium flow rate (30 l/min) as per the manufacturer's protocol. Each growth factor was allowed to interact with SL-CS/DS (1.5 M) for 2 min each for association and for dissociation, after which the sensor chip was regenerated by injecting 1 M NaCl for 2 min before each injection. The kinetic parameters were evaluated with BIAevaluation 3.1 software (BIAcore AB) using a 1:1 binding model with mass transfer. To investigate the structural characteristics of the putative functional epitopes of SL-CS/DS (1.5 M) for the binding to various growth factors, an aliquot (10 g as GAG) was digested with 10 mIU each of CSase ABC, a mixture of CSases AC-I and AC-II, or CSase B, and a 2-g aliquot of each digest was used for the inhibition analysis as described above except that the SL-CS/DS (1.5 M)-immobilized sensor chip was used here.
Preparation of Partially Desulfated CS/DS Chains-The alkali treatment, which removes 2-O-sulfate from the IdoUA of heparin (39), was applied to the SL-CS/DS (1.5 M) fraction. SL-CS/DS (1.5 M) (20 g as GAG) was dissolved in 20 l of 0.1 M NaOH, frozen at Ϫ20°C for 2 h, neutralized with 0.5 M HCl to pH 7.0, desalted by gel filtration on a PD-10 column, and freeze-dried.
Assays for NOP Activity-Cultures of hippocampal neurons were established from E16 mice as previously described (25). Briefly, 2 g/well of a CS/DS preparation was individually coated onto coverslips precoated with poly-DL-ornithine (P-ORN) (Sigma) at 4°C overnight. The hippocampal neuronal cells were freshly isolated from E16 mouse embryos, suspended in Eagle's minimum essential medium containing N2 supplements, seeded on coverslips at a density of 10,000 cells/cm 2 , and allowed to grow in a humidified atmosphere for 24 h at 37°C with 5% CO 2 .
For the neutralization assay using antibodies, polyclonal anti-PTN antibody (10 g/ml), polyclonal anti-MK antibody (10 g/ml), monoclonal anti-bFGF antibody (10 g/ml), or polyclonal anti-HGF antibody (3 g/ml) was added to the medium 2 h after the seeding of the cells. After incubation overnight, the cells were fixed using 4% (w/v) paraformaldehyde for 30 min at room temperature, and the neurites were visualized by immunochemical staining using anti-microtubule-associ-ated protein-2 (Lieco Technologies Inc., St. Louis, MO) (40) and anti-neurofilament (Sigma) (41). The antibodies were then detected using a Vectastain ABC kit (Vector Laboratories Inc., Burlingame, CA) with 3,3Ј-diaminobenzidine as a chromogen. The immunostained cells on each coverslip were scanned and digitized with a ϫ20 objective lens on an optical microscope (BH-2; Olympus, Tokyo, Japan) equipped with a digital camera (HC-300Z/OL; Olympus). 100 clearly isolated cells with at least one neurite longer than the cell body were chosen at random to determine the length of the longest neurite using a morphological analysis software (Mac SCOPE; Mitani Corp., Tokyo, Japan). At least three independent experiments per parameter or condition were carried out.

Preparation of the CS/DS Fractions from Shark Liver-GAGs
were extracted from shark liver by protease digestion and alkali treatment, recovered by ethanol precipitation, and purified by anion exchange chromatography on a DEAE-Sephadex column. The GAG preparation thus obtained was treated with nitrous acid followed by gel filtration to remove heparin/HS (ϳ50% of all GAG). No significant hyaluronic acid-derived oligosaccharides were found by anion exchange HPLC in the digests of the CS/DS preparations treated with CSase ABC or hyaluronidase SH (data not shown). This preparation was further fractionated by anion exchange chromatography using an Accell TM Plus QMA cartridge, which was eluted stepwise with buffers containing 0.2, 1.0, 1.5, and 2.0 M NaCl. Only trace amounts of CS/DS were detected in the fractions eluted with buffers containing 0.2 and 2.0 M NaCl. 80 and 20% of the total CS/DS were detected in the fractions eluted with buffers containing 1.0 and 1.5 M NaCl and designated as SL-CS/DS (1.0 M) and SL-CS/DS (1.5 M), respectively. These preparations were passed through a C 18 cartridge to remove peptides.
Determination of the Molecular Mass of Shark Liver CS/DS-The molecular mass of SL-CS/DS preparations was determined by gel filtration (Fig. 1). Using the calibration curve generated with standard polysaccharides, the average molecular mass of SL-CS/DS (1.5 M) was estimated to be 75.7 kDa, whereas SL-CS/DS (1.0 M) gave a broader peak with a molecular mass ranging from 3.8 to 38.9 kDa. The distinct sizes of these two preparations may suggest different structures and functions. Interestingly, the mass of SL-CS/DS (1.  (Table 1). Thus, an unique composition was revealed for SL-CS/DS (1.0 M) with three kinds of disulfated disaccharide units (⌬B, ⌬D, and ⌬E) and for SL-CS/DS (1.5 M) with small proportions of ⌬D and ⌬T in addition to significant proportions of ⌬B and ⌬E.
To discriminate GlcUA-or IdoUA-containing disaccharides in CS/DS chains, an analysis was also carried out using a mixture of CSases AC-I and AC-II, which specifically digest Gal-NAc-GlcUA linkages in the CS structure (42), and CSase B, which specifically attacks GalNAc-IdoUA linkages in the DS structure (43). Both digestions yielded most of the unsaturated disaccharide units except for ⌬O, ⌬C, and ⌬T. Nonsulfated units appear to exist as O units rather than iO units. However, the fact that no ⌬O was observed in the CSase B digests of the two preparations may be partially due to the resistant nature of the iO unit to CSase B (44) and may not necessarily indicate the absence of this unit. Interestingly, the rare B unit was demonstrated for both preparations, as in the case of SS-CS/DS (28), and may be a unique feature of shark CS/DS. Another interesting feature is the obvious presence of iC units in the 1.0 M fraction but not in the 1.5 M fraction, although no C units were found in either. It is interesting that the ⌬T unit for SL-CS/DS (1.5 M) was almost completely recovered in the CSase B digest and not in the digest obtained with a mixture of CSases AC-I and AC-II, suggesting that ⌬T was derived exclusively from IdoUA(2S)␣1-3GalNAc(4S,6S) (iT unit), not from GlcUA(2S)␤1-3GalNAc(4S,6S). Thus, the composition of both preparations, although highly heterogenous, is distinct.
These results also clearly revealed a significantly higher proportion of GlcUA than IdoUA in SL-CS/DS (1.0 M) with a molar ratio of 1.56:1 (GlcUA/IdoUA), whereas SL-CS/DS (1.5 M) showed a higher proportion of IdoUA with a molar ratio of 3.35:1 (IdoUA/GlcUA). In this context, these two SL-CS/DS    Fig. 4. These sensorgrams were analyzed collectively by using "the 1:1 Langmuir binding model with mass transfer" of the BIAevaluation 3.1 software to calculate the kinetic parameters. The kinetic parameters are summarized in Table 2 for all the growth factors except FGF-1, which exhibited only a weak binding response.

Neuritogenic Activity of CS/DS Hybrid Chains
The tested growth factors varied in their ability to bind SL-CS/DS (1.5 M). HGF, FGF-2, MK, and PTN displayed quick binding and a slow dissociation, giving K d values in the low nanomolar range and signifying their strong affinity for SL-CS/DS (1.5 M) ( Table 2). In contrast, HB-EGF, VEGF 165 , and FGF-7 showed weaker affinity for SL-CS/DS (1.5) as reflected in the K d values listed in Table 2. These differences in affinity support the specificity of the interactions between SL-CS/DS (1.5 M) chains and various heparin-binding growth factors, suggesting a biological significance of these interactions.
NOP Activity of the SL-CS/DS Preparation-That SL-CS/DS (1.5 M) chains specifically interacted with some of the growth factors involved in the brain's development suggests that they may possess NOP activity. To evaluate the NOP activity of SL-CS/DS preparations, embryonic day 16 mouse hippocampal  neuronal cells were utilized. The cells were cultured on a substrate coated with SL-CS/DS (1.5 M), SL-CS/DS (1.0 M), CS-E (a positive control), or P-ORN alone (a negative control). The length of the longest neurite of each of 100 randomly selected cells cultured on each substrate was measured. The neuronal cells cultured on the P-ORN substrate had some short neurites; however, their length was not significant (Fig. 5, A and bottom). In contrast, the neuronal cells cultured on coverslips precoated with SL-CS/DS (1.5 M) exhibited striking NOP activity (Fig. 5, D  and bottom), showing neurites axonic in nature with stronger activity than a positive control CS-E (Fig. 5, B and bottom). It is interesting that although SL-CS/DS (1.5 M) and CS-E have comparable sulfate/unit ratios, 1.43 and 1.53, respectively, the former displayed stronger activity, suggesting that the types and sequential arrangement of oversulfated disaccharides are important. In contrast, SL-CS/DS (1.0 M) (Fig. 5, C and bottom) showed weak yet significant NOP activity stronger than the negative control of P-ORN (Fig. 5, A and bottom). These results suggest that the two preparations from SL-CS/DS exert stimulatory effects on hippocampal neurons to different extents.

TABLE 2 Kinetic parameters for the interaction of growth factors with immobilized SL-CS/DS (1.5 M)
The k a , k d , and K d values were determined using a 1:1 Languimuir binding model with mass transfer as described under "Experimental Procedures." The value for each growth factor is expressed as the mean Ϯ S.E. of five different concentrations.  FEBRUARY 2, 2007 • VOLUME 282 • NUMBER 5 tional structure for the growth factor-binding and NOP activities. Therefore, to investigate the structural basis of the growth factor-binding and NOP activities of SL-CS/DS (1.5 M), the preparation was digested with various CSases differing in specificity, and a comparison of the resultant fragments was made by gel filtration. CSase ABC catalyzes the eliminative cleavage of most if not all the galactosaminidic linkages in CS/DS chains to produce disaccharides, resulting in a nearly complete digestion of SL-CS/DS (1.5 M) (Fig. 6A). In contrast, a mixture of CSases AC-I and AC-II, which specifically cleaves N-acetylgalactosaminidic linkages to give GlcUA, released 13.2% of all disaccharides compared with CSase ABC (Table 1; note the difference in the scale of Fig. 6A), and the majority of the resistant fragments ranged greatly in size from tetrasaccharides to polysaccharides (Fig. 6B). In contrast, SL-CS/DS (1.5 M) was highly sensitive to CSase B, which cleaves the linkage between N-acetylgalactosamine and IdoUA, and the digest mainly gave disaccharides and tetrasaccharides with a small hexasaccharide peak, suggesting a higher proportion of IdoUA than GlcUA (Fig. 6C) Fig. 7, top, digestion of SL-CS/DS (1.5 M) with CSase ABC or B almost completely abolished its inhibitory activity against most of the growth factors binding to SL-CS/DS. In contrast, only 10 -30% inhibition was observed with a CSase AC-I/AC-II digest for all of the ligands used. Subsequently, NOP assays were carried out to confirm the different contributions of the CS and DS moieties in SL-CS/DS (1.5 M). Enzyme digests were individually coated on P-ORN-precoated coverslips, and then hippocampal neuronal cells were cultured (Fig.  7, bottom). As expected, the NOP activity of SL-CS/DS (1.5 M) was completely eliminated by digestion with CSase ABC or B, and the activity was dramatically reduced to the basal level, comparable with that for P-ORN alone, whereas digestion with a mixture of CSases AC-I and AC-II resulted in only a partial loss of the activity (Fig. 7, bottom) (1.5 M) was desulfated by the alkali treatment (see "Experimental Procedures"), and the resultant preparation was used for coating P-ORN-precoated coverslips for the NOP assay (Fig. 8). The alkali treatment, which removes 2-O-sulfate from the IdoUA of heparin (39), was used to remove 2-O-sulfate groups from the SL-CS/DS (1.5 M) fraction as described under "Experimental Procedures." The disaccharide analysis of the treated sample showed 88.3% desulfation at 2-O-sulfate of ⌬B and a corresponding increase in the ⌬A unit, with a concomitant decrease (45.8%) in ⌬E resulting in the corresponding increase in ⌬C (Table 3). Considering that monosulfated disaccharide units are resistant to the alkaline treatment, the changes in proportion of ⌬ 4,5 HexUA␣1-3GalNAc(6S) and ⌬ 4,5 HexUA␣1-3GalNAc (4S) seem to be primarily due to the desulfation of disulfated units ( Table 3). The NOP of SL-CS/DS (1.5 M) toward hippocampal neurons was significantly suppressed by desulfation with the alkali treatment (Fig. 8) as compared with that of the untreated SL-CS/DS (1.5 M). These results suggested that oversulfated disaccharides, especially iB as well as iE/E units, are key elements for the NOP activity of SL-CS/DS (1.5 M), with  (45)(46)(47)(48)(49). Hence, these endogenous heparin-binding growth factors may be involved in the mechanism of the expression of the NOP activity of SL-CS/DS (1.5 M). To examine this possibility, antibodies against HGF, PTN, MK, or bFGF were individually added to the system for the neutralization assay of the NOP activity of SL-CS/DS (1.5 M) toward hippocampal neurons from an embryonic mouse brain. The addition of anti-HGF or anti-PTN antibody markedly suppressed the NOP activity of SL-CS/DS (1.5 M) to the basal level of P-ORN (Fig. 9A). In strong contrast, the anti-MK or bFGF antibody showed no significant inhibition. These results suggest

TABLE 3 Comparison of disaccharide composition and degree of sulfation of SL-CS/DS (1.5 M) before and after alkali treatment
SL-CS/DS (1.5 M) (20 g as GAGs) was treated with 0.1 M NaOH, and one-tenth of the resultant product and 0.5 g of native preparation were individually digested with CSase ABC. Each digest was labeled with 2AB and analyzed by anion exchange HPLC as described under "Experimental Procedures." ND, not determined.  To prove the involvement of the signaling pathways of HGF and PTN in the NOP activity of the CS/DS chains, effects of the antibodies against the receptors for HGF (c-Met) and for PTN (ALK) were investigated. The anti-c-Met and anti-ALK anti-bodies inhibited the NOP activity of SL-CS/DS (1.5 M) by 60 and 56%, respectively, revealing the involvement of the HGF and PTN signaling pathways in the NOP activity of the sugar chains (Fig. 9A).

Unsaturated disaccharide a A b B c B ؊ A (B ؊ A)/A
To investigate the in vivo mechanism of the NOP activity of brain CS/DS, SL-CS/DS (1.5 M) was replaced by E-CS/DS, which was prepared from embryonic pig brain (26), in the above mentioned inhibition assay using the antibodies against the growth factors and receptors. The NOP activity of E-CS/DS was also strongly suppressed by the antibody against HGF, PTN, c-Met, and ALK by 58,45,47, and 38%, respectively, but barely by the antibody against MK or bFGF (Fig. 9B). These results suggest the involvement of the signaling pathways of HGF and PTN in the expression of the NOP activity of the embryonic brain CS/DS in vivo.

DISCUSSION
The growth factor-binding activities of CS/DS chains exhibit a positive correlation with their NOP activity. Some oversulfated CS and DS, which can interact with various brain-derived growth factors and neurotrophic factors, show significant NOP activities (24,25,27,28). In this study, it was demonstrated for the first time that both the signaling pathways of PTN-ALK and HGF-cMet are involved in the NOP activities of CS/DS hybrid chains isolated from E16 embryonic mouse brain and of SL-CS/DS (1.5 M). Pleiotrophin uses protein-tyrosine phosphatase and ALK as its receptors. Although the former, which is a CS proteoglycan, has been shown to be involved in neuritogenesis through the CS chains (31), ALK has also been demonstrated here to be involved in the NOP activity of E-CS/DS. Notably, E-CS/DS and SL-CS/DS chains immobilized on the P-ORNcoated substrate served as a scaffold to recruit endogenous PTN and HGF in the culture system and to stimulate neuronal cells, promoting neuritogenesis.
PTN, MK, HGF, and bFGF, examined in this study, are broadly expressed in the brain, including hippocampal neuronal progenitor cells, and implicated in the development of the brain (45)(46)(47)(48)(49). The NOP activity of SL-CS/DS (1.5 M) and E-CS/DS was significantly suppressed by antibodies against PTN or HGF and also by antibodies against their respective receptors (ALK and cMet) but not by anti-MK or anti-bFGF antibody. These results suggested that endogenous PTN and HGF were recruited by E-CS/DS and SL-CS/DS and mediated their NOP activities. Although ALK has been identified as a common receptor for PTN (50) and MK (51), it appears to be involved in the NOP activity of CS/DS through PTN signaling. MK and bFGF may be important for the survival or adhesion of neuronal cells. Anti-HGF antibody showed much stronger inhibitory activity than anti-PTN antibody, suggesting that HGF plays a more crucial role in the CS/DS-generated signaling to promote neuritogenesis.
HGF is a pleiotropic factor. It binds to and activates the tyrosine kinase receptor cMet and is also an axonal chemoattractant and neurotrophic factor for motor neurons (52,53). It is required for the axonal growth of dorsal root ganglia sensory neurons (54) and elicits multiple functions in sympathetic neurons (55)(56)(57). HGF signaling potentiates the response of different neurons to specific signals (52). GAGs are essential co-re- antibodies were also tested separately. In parallel, the same amount of a IgG fraction purified from the corresponding host animal (mouse, goat, or rabbit) was used as a control, which showed no significant inhibition in the mean length of the longest neurite measured for SL-CS/DS (1.5 M) or E-CS/DS without the antibodies (data not shown). **, p Ͻ 0.01; ***, p Ͻ 0.001 (For the significance of differences, see the legend to Fig. 5).
ceptors for the activation of cMet (58), since HGF can bind HS (59) or DS (60) chains of proteoglycans and also interacts with cMet to form an active ternary complex (61). In this study, CS/DS hybrid chains were demonstrated to recruit a minute amount of endogenous HGF to stimulate the outgrowth of neurites in hippocampal neurons, which may suggest that CS/DS presented HGF to cMet or both cMet and glycan co-receptors (HS or CS/DS) on the neuronal surface. The structure of the HGF-binding sites on HS and CS/DS chains remains to be investigated.
PTN is expressed at the intracellular matrix of axonal tracts in the developing brain (30,62,63) and involved in the development of axons in vivo and the outgrowth of neurites in vitro (29,64). PTN induces neurite outgrowth through specific interaction with the HS side chains of syndecan-3 (65), although our studies have shown that PTN also interacts with both endogenous and exogenous CS/DS chains (21,26,27,38). The oversulfated disaccharide units in CS/DS chains are critical to such interaction. However, it remains unclear which oversulfated disaccharide plays a major role in the binding of PTN. An analysis of the PTN-binding fractions of E-CS/DS (32) or SS-CS/ DS 4 showed a significant increase in the proportion of ⌬B and ⌬E with increased affinity for PTN, whereas the proportion of ⌬D decreased. Similarly, SL-CS/DS (1.5 M), which showed high proportions of ⌬B and ⌬E units, strongly inhibited the binding of PTN to E-CS/DS-H, whereas CS-D displayed weak inhibition. Hence, the B/iB and E/iE units in CS/DS chains, which are the parental structures for ⌬B and ⌬E, are more important to the binding of PTN than D/iD units.
The drastic abolishment of the growth factor-binding and NOP activities (Fig. 7) of SL-CS/DS (1.5 M) by digestion with CSase B rather than a mixture of CSases AC-I and AC-II revealed the key role of IdoUA-containing disaccharides. The critical role of IdoUA has also been demonstrated for the growth factor-binding and NOP activities of relatively low sulfated CS/DS preparations, including E-CS/DS (sulfate/disaccharide unit ratio ϭ 0.83) (26) and SS-CS/DS (sulfate/disaccharide unit ratio ϭ 1.17) (28). These and other findings together support the notion that disulfated units and IdoUA are critical factors for these activities. Although the data strongly suggest their involvement, the final proof awaits a demonstration of IdoUA in the functional domains for the growth factor-binding and neurite outgrowth-promoting activities. Compared with GlcUA, IdoUA can form various conformations, resulting in an inherent plasticity for interaction with various protein partners (66,67). However, the low sulfated porcine skin DS, or CS-B, which is rich in iA units (IdoUA␣1-3GalNAc(4S)), is a weak inhibitor for the binding of PTN to E-CS/DS-H, which in turn indicates the requirement for IdoUA-containing oversulfated disaccharides, such as IdoUA(2S)␣1-3GalNAc(4S) (iB) and IdoUA␣1-3GalNAc(4S,6S) (iE). In this context, adult sea urchin DS, which contains proportions of iB and iE comparable with those of SL-CS/DS (1.5 M), exhibited weaker NOP activity than SL-CS/DS (1.5 M) and CS-E, which contains 56% E units but no iB units (25). Interestingly, adult sea urchin DS contains 59% iC but only less than 1% iA, whereas SL-CS/DS (1.5 M) contains 42.6 -49.4 mol % of iA. In addition, ⌬A was observed with increasing proportions of ⌬B and ⌬E in the high PTN affinity fraction of E-CS/DS and SS-CS/DS 4 and with a decrease in ⌬C. The greater NOP activity of SL-CS/DS (1.5 M) than CS-E (Fig. 5) suggests that iB, iE, and iA units are preferred for the PTN binding activity and also for the NOP activity of CS/DS chains.
Previously, it was demonstrated that anti-PTN antibody strongly suppressed the endogenous PTN-mediated NOP activity of E-CS/DS chains with low affinity for PTN (E-CS/ DS-L) but did not significantly inhibit that of E-CS/DS with high affinity for PTN (E-CS/DS-H) (32). Hence, it seems that PTN mediates the NOP activity of E-CS/DS-L but not E-CS/ DS-H. A series of octasaccharides containing at least one D unit have been isolated from E-CS/DS-L (32). Although a signaling molecule responsible for the NOP activity of E-CS/DS-H remains to be identified, the possibility exists that PTN and other growth factors share the binding sites in E-CS/DS-H chains. The binding of other growth factors to the putative overlapping binding sites on E-CS/DS-H chains may be involved in the neuritogenesis. The specific interaction of SL-CS/DS (1.5 M) with PTN, MK, HGF, and bFGF implies the existence of such overlapping binding sites for these growth factors. HGF is a strong candidate for a factor involved in the NOP activity of E-CS/DS-H.
SL-CS/DS (1.5 M) is a potential candidate for a non-mammalderived therapeutic agent. Further investigation of the functional domains of SL-CS/DS (1.5 M) involved in PTN and HGF signaling should provide a basis for developing specific oligosaccharide drugs with fewer side effects than the parental polysaccharides.