Inhibition of neural crest cell migration by aggregating chondroitin sulfate proteoglycans is mediated by their hyaluronan-binding region

DEVELOPMENTAL

137, 1-12 (1990)

BIOLOGY

Inhibition of Neural Crest Cell Migration by Aggregating Chondroitin Sulfate Proteoglycans Is Mediated by Their Hyaluronan-Binding Region ROBERTO PERRIS* Developmental

AND

STAFFAN

JoHANssoNt

Irvine, California 92717; and TDepartment of Medical and Biology Center, University of Cal(fornia, Physiological Chemistry, Biomedical Center Uppsala, Boz 575, S-751 23, Uppsala, Sweden Accepted September 7, 1989

We have recently shown that the large hyaluronan-aggregating chondroitin sulfate proteoglycan from cartilage (PG-LA) is unfavorable as a substrate for neural crest cell migration in vitro and that this macromolecule inhibits cell dispersion on fibronectin substrates when included in the medium (R. Perris and S. Johansson, 198’7, J. Cell Biol. 105, 2511-2521). In this study we present data on the specificity of the migration-repressing activity of PG-LA and data on the molecular mechanisms by which the proteoglycan might impair neural crest cell motility. Soluble PC-LA potently impaired cell migration on substrates of lamininllaminin-nidogen, vitronectin, and collagen types I, III, IV, and VI. When tested in solid-phase binding assays, PG-LA bound avidly to substrates of collagen types I-III and V. Conversely, minimal amounts of the proteoglycan bound to substrates of laminin-nidogen, vitronectin, collagen types IV and VI, and fibronectin or to a proteolytic fragment encompassing its cell-binding domain (105 kDa). Preincubation of these substrates with soluble PG-LA prior to plating of the cells had no effect on their locomotory behavior. These results indicate that PG-LA affects neural crest cell movement primarily through an interaction with the cell surface, rather than by association with the cell motility-promoting substrate molecules. The molecular interaction of soluble PG-LA with neural crest cells was further examined by analyzing the effects of isolated domains of the proteoglycan on cell migration on fibronectin. Addition of chondroitin sulfate chains, the core protein free of glycosaminoglycans, the isolated hyaluronan-binding region (HABr), or a proteolytic fragment corresponding to the keratan sulfate-enriched domain of the PG-LA to neural crest cells migrating on fibronectin or the 105kDa fibronectin fragment had no significant effect on their motility. After reduction and alkylation, PG-LA was considerably less efficient in perturbing cell movement on fibronectin substrates and virtually ineffective in altering migration on the 105kDa fragment. In the presence of hyaluronan fragments of 16-30 monosaccharides in length, or an antiserum against the HABr, the migration repressing activity of PG-LA was reduced in a dose-dependent fashion. Furthermore, the inhibitory action of PG-LA was significantly reduced by treatment of the cells with Streptomyces hyaluronidase. These findings indicate that hyaluronan-aggregating proteoglycans inhibit neural crest cell movement by interacting with the cell surface, at which the molecules binds via its HABr to cell membrane-anchored hyaluronan. We propose that by this mechanism, PG-LA-related proteoglycans may regulate neural crest cell migration by altering the function of cell surface receptors 0 1990 Academic Press, Inc. for motility-promoting molecules.

tion by contributing to the assembly of permissive interstitial/basement membrane matrices and directly involved by supporting neural crest cell movement (Lofberg et ah, 1980; Tucker and Erickson, 1984; Duband and Thiery, 1987; Perris et ah, 1989b,c; Perris and Bronner-Fraser, 198913). In contrast, there is evidence that chondroitin sulfate-rich proteoglycans may impose directionality of movement by locally restraining neural crest cell migration (Tucker and Erickson, 1984; Tucker, 1986; Newgreen et al, 1986; Tan et ah, 1987; Perris et cd, 1989b,c). In a recent investigation we examined the role of one class of chondroitin sulfate proteoglycans in neural crest cell migration by employing an in vitro assay. It was found that the large hyaluronan-aggregating chondroitin sulfate proteoglycan (PG-LA) from bovine cartilage was nonpermissive as a substrate for the initial migration of explanted axolotl neural crest cells. In addition, when added to the culture medium, the same

INTRODUCTION

A wealth of previous investigations have affirmed the role of the extracellular matrix for the support and guidance of neural crest cell migration (Newgreen and Erickson, 1986; Liifberg et ah, 1989a; Perris and Bronner-Fraser, 1989a). More recently, it has also been demonstrated that the initiation of neural crest cell movement in viva is correlated with a developmental process occurring within the surrounding extracellular matrix (Liifberg et ah, 1985; 1989b). Cell adhesion glycoproteins such as fibronectin, vitronectin, and laminin support neural crest cell motility in vitro and are also likely to participate in the process of neural crest cell migration in vivo (Boucaut et ah, 1984; Bronner-Fraser, 1986; Newgreen and Erickson, 1986; Perris and Johansson, 1987; Dufour et al, 1988; Perris et al, 1989a). Presumably, various types of collagens are indirectly involved in neural crest cell migra1

0012-1606/90 $3.00 Copyright All rights

0 1990 by Academic Press, Inc. of reproduction in any form reserved.

2

DEVELOPMENTAL BIOLOGY

VOLUME 137.1990

proteoglycan prevented neural crest cell migration on washes of PBS at room temperature. An aliquot of 100 substrates of intact fibronectin, and even more potently ~1 of ovomucoid-agarose bead suspension was mixed on substrates composed of a 105-kDa fragment com- with 1 ml of Streptwmyces hyaluronidase dissolved in prising the cell-binding domain of fibronectin (Perris PBS at 1050 U/ml. The mixture was then incubated at and Johansson, 1987). Preincubation of this latter sub- 4°C overnight under constant shaking. The ovomucoidstrate with soluble PG-LA prior to cell plating did not agarose beads were finally separated from the enzyme affect migration, indicating that the proteoglycan could solution by mild centrifugation. For control experiinhibit cell movement by a mechanism independent of ments, the enzyme was heat-inactivated for 30 min at binding to the substrate. In the present study we have 60°C. further analyzed the mechanism by which this class of Puri,fcation of Various Domains of the PG-LA proteoglycans may inhibit neural crest cell migration. Cartilage proteoglycan complexes were isolated from MATERIAL AND METHODS bovine nasal cartilage, digested with TPCK-treated trypsin and fractionated by CsCl-density gradient cenExtracellular Matrix Molecules trifugation into a protein rich fraction and a glycosamiHuman plasma fibronectin and a lO&kDa fragment noglycan rich fraction as previously described comprising the cell-binding domain of the molecule (HeinegHrd and Sommarin, 1987). The complex of hyalwere purified according to previously described proceuronan/hyaluronan-binding region (HABr)/link produres (Woods et al., 1986; Perris and Johansson, 1987). A tein was then isolated from the protein-rich, top frac11.5-kDa fragment encompassing the RGDS recognition site of fibronectin (Pierschbacher et al, 1982) was tion by chromatography on Sepharose 2B (Pharmacia, purchased from Telios Pharmaceuticals and Research Sweden) columns under associative conditions, and Products, Inc. (San Diego, CA). Hyaluronan fragments subsequently the HABr was separated from hyalof 16-30 monosaccharides in length were isolated after uronan and link protein by further chromatography on partial digestion of hyaluronan with testicular hyal- Sepharose 6B (Pharmacia, Sweden) columns under disuronidase (Sigma) as previously described (Hascall and sociative conditions according to previous procedures Heineglrd, 1979). Other matrix components were ob- (Heinegird and Sommarin, 1987). The bottom fraction tained as follows: purified laminin from Rupert Timpl, of the density gradient was used for the isolation of the keratan sulfate-rich fragment after fragmentation of Max-Planck Institut fur Biochemie, Martiensried, Munich, West Germany; a laminin-nidogen complex the chondroitin sulfate side chains by chondroitinase (1:l ratio) from Dr. Mats Paulsson, Biozentrum, Univer- ABC digestion. The digest was further chromatosity of Basel, Basel, Switzerland; human and bovine vi- graphed on Sepharose 6B columns (Heinegard and tronectin from Dr. Erkki Ruoslahti, La Jolla Cancer Sommarin, 1987). The core protein free of glycosaminoglycan side Research Foundation, La Jolla, California, and Telios chains was obtained by Smith’s degradation of a bovine Pharmaceuticals and Research Products, Inc.; human, bovine, and murine collagen types I-III from Dr. Kris- nasal proteoglycan monomer (Al-Dl) fraction (Baker et al., 1972). All samples were converted to their sodium tofer Rubin, Department of Medical and Physiological Chemistry, Biomedical Center, Uppsala, Sweden, Telios form before use. Pharmaceuticals, Inc., and Chemicon Biochemicals, Inc.; human collagen types IV and VI from Dr. Jorgen Wieslander, Biocarb AB, Lund, Sweden, Telios Pharmaceuticals, Inc., and Collaborative Research Lab.; chondroitin sulfate polysaccharides from Dr. Ulf Lindahl, Department of Veterinary Medical Chemistry, Swedish University of Agricultural Science, Uppsala, Sweden; and a high molecular mass, hyaluronan-binding chondroitin sulfate proteoglycan (PG-LA), purified from bovine nasal cartilage under dissociative conditions (Dl) from Dr. Bruce Caterson, Department of Biochemistry, West Virginia University, West Virginia. Ovomucoid-Puri@ation

of Streptomyces hyaluronidase

Streptomyces hyaluronidase (Sigma) was purified from potential protease contamination by chromatography over agarose-coupled ovomucoid (Sigma). Briefly, ovomucoid-agarose beads were rinsed in three

Antibodies

A monospecific antiserum against the HABr was raised in rabbits by four intramuscular injections (2week interval) of the purified bovine HABr (0.1 mg/injection) emulsified in Freund’s adjuvant. The specificity of the antiserum for the HABr of the PG-LA was assessedby RIA competition assay (Table 1) and ELISA (data not shown). The monoclonal antibody 7-D-4 specific for native chondroitin chains of cartilage-type proteoglycans (Sorrel1 et al., 1988, 1989) was obtained from Dr. Michael Sorrell, Department of Biochemistry, University of West Virginia, Morgantown, West Virginia. Reductive Alkylation

of the PG-LA

The isolated PG-LA was incubated in 50 mM dithiothreitol (Sigma) and 6 Mguanidium chloride (Sigma) in

PERRIS AND JOHANSSON

Inhibition

of Neural

TABLE 1 ABILITY OF UNLABELLEDLIGANDS TO COMPETEWITH '=I-LABELED PG-LA FORBINDING TOANTI-HABra

Antigen PG-LA PG-LA, reduced + alkylated PG-LA, chondroitinase ABC-digested HABr

Maximal inhibition obtained (% )

Amount of protein (ng) giving 50% inhibition of precipitation

97

41

10

24000

98 93

44 21

a The radioimmunoassay was carried according to previously puhlished procedures (Caterson et a& 1983).

0.1 MTris buffer, pH 8.0, for 6 hr at 22°C. Subsequently, a 25% molar excess of iodoacetamide (Sigma) was added and the solution was further incubated for 2 hr at 22°C. Finally, the reduced and alkylated PG-LA was extensively dialyzed against phosphate-buffered saline. By this treatment, PG-LA looses its ability to form stable complexes with hyaluronan as determined by chromatography on Sepharose CL-2B columns (Johansson et al., 1985). Solid-Phase Binding

Assays

To determine the ability of soluble PG-LA to bind to the various matrix substrates we carried out enzymelinked binding assays with biotinylated PG-LA (Perris and Johansson, 1987; Perris et aL, 1989b). Briefly, microwell plates were coated with fibronectin, the 105kDa fragment, vitronectin, laminin-nidogen, and collagen types I-VI at their maximal coating concentration in 0.05 M bicarbonate buffer, pH 9.6 (Perris and Johansson, 1987; Perris et al., 1989b; Perris and BronnerFraser, 1989b). Uncoated surface areas of the plastic were then blocked with 2% BSA/l% ovalbumin in the same buffer. PG-LA was biotinylated by incubation with biotin-/3-aminocaproic acid-N-hydroxysuccimide ester and applied to the precoated wells at 0.05-500 pg/ml. Binding of PG-LA to these matrix substrates was assessed calorimetrically by using streptavidinhorseradish peroxidase and an indamine dye catalyzed by the peroxidase (Perris et al., 1989a). Cell Culture and Assessment of Migration Culture substrates containing various matrix proteins were prepared as previously described (Perris and Johansson, 1987; Perris et al., 1989a). Collagens were adsorbed directly onto plastic or in the case of types I-III gelled at 250-600 pg/ml in culture medium at 37°C

Crest Cell Migration

by Proteog&ans

3

for 1 hr and subsequently air-dried overnight at room temperature. In some cases, matrix substrates were incubated with 500 pg/ml PG-LA at 4°C overnight. Excess nonbound PG-LA was then removed and the substrates were extensively rinsed with PBS. Neural crest cell cultures were prepared from stage 25 Mexican axolotl embryos (Ambystoma mexicanum) as described elsewhere (Perris and Johansson, 1987). Briefly, neural tube-neural crest complexes of a defined size were excised microsurgically from the midtrunk region of the embryo. Isolated explants were allowed to recover in BSA-containing medium for 30-45 min at room temperature, rinsed in serum-free medium, and subsequently deposited onto the various matrix substrates. The neural crest-neural tube explants were grown in serum-free PL-85 medium (Perris and Johansson, 1987) at room temperature. After 18 hr of culture, emigration of the neural crest cells from the neural tube was estimated by adopting two quantitative parameters, migratory distance and total number of cells dispersed (Perris and Johansson, 1987). Statistic significance was determined by using the parametrical Student t test. The criterion for significant difference between two experimental situations was defined as P < 0.005. Exogenously administered molecules were added to the medium at different concentrations at the start of the culture. In a set of experiments a concentration of PG-LA causing >50% inhibition of movement (500 pg/ml) was preincubated in PBS for 2 hr at 37°C with hyaluronan oligosaccharides, various concentrations of the anti-HABr antiserum, the anti-HABr antiserum preblocked with 25 pg/ml isolated HABr, or the monoclonal antibody 7-D-4 (l:lOO), before addition to the cultures. Streptomyces hyaluronidase pretreatment of neural crest cells was carried out for 1 hr at room temperature by incubating freshly isolated neural tube-neural crest explants in culture medium containing 210 U/ml Streptomyces hyaluronidase, 1 mg/ml of a mixture of chondroitin sulfates A and C, and 150 pg/ml keratan sulfate. The explants were then rinsed once in culture medium and plated on fibronectin substrates in the presence or absence of soluble PG-LA. Explants were additionally supplemented with 210 U/ml enzyme after 1, 6, and 12 hr of culture. RESULTS

Inhibition of Neural Crest Cell Migration on Laminin/Laminin-Nidogen, Vitronectin, and Collagen Types I- VI by Soluble PG-LA We have previously found that PG-LA causes a dosedependent inhibition of initial neural crest cell migration on fibronectin substrates and that it can arrest the

4

DEVELOPMENTAL BIOLOGY

progressing cell movement when included in the culture (Perris and Johansson, 1987). To study the specificity of the inhibition, the effect of PG-LA on cell movement on laminin/laminin-nidogen, vitronectin, and collagen types I-VI was investigated. Of the different collagens tested the cells dispersed most efficiently on types I, IV, and VI (Fig. 1). The migration on these substrates was somewhat lower than that on laminin-nidogen and corresponded to 35-40% of that observed on fibronectin/ vitronectin (Perris and Johansson, 1987). Neural crest cells migrated very poorly on collagen type V, the extent of migration being indistinguishable from the background movement on BSA/ovalbumin (Fig. 1). Soluble PG-LA inhibited initial neural crest cell migration on laminin (data not shown), laminin-nidogen, vitronectin, and collagen types I, IV, and VI (Fig. 2, Table II), in a fashion comparable to that previously reported for fibronectin (Perris and Johansson, 1987). In addition, PG-LA also inhibited migration on the 11.5-kDa cell-binding fragment of fibronectin (Table 3) and on collagen type III. The average numbers of cells emigrated from the neural tube on collagen type III in the absence versus the presence of PG-LA (0.5 mg/ml) were 100.7 + 12.2 cells (19 explants) and 55.6 t- 15.8 (9 explants) cells, respectively. The limited neural crest cell emigration from the neural tube observed on collagen type II substrates (average, 60-65 cells) was unaffected by addition of PG-LA (data not shown). The few neural crest cells that were able to extend onto the substrates in the presence of PG-LA exhibited a rounded morphology characteristic of poorly adherent cells (Fig. 3). The migration on collagen type IV was less sensitive to inhibition by PG-LA as indicated by both the number of cells emigrated from the neural tube (Fig. 2) and the distance migrated by the cells. The overall migratory pattern on collagen type IV was, however, markedly different than that observed on other matrix substrates (Fig. 3).

VOLUME 137.1990

BS

Cal I

Cal II

Cal III

Cal IV

Cal v

Cal VI

FIG. 1. Axolotl neural crest cell migration on isolated collagens. In the native, hydrated form, collagen type I-III (Co1 I-III) gels were entirely nonpermissive for cell movement (data not shown). Therefore, Co1 I-III denote here air-dried gels with protein concentrations ranging from 250 to 600 pg/ml. Collagen types IV-VI were adsorbed directly onto plastic at lo-100 rg/ml according to data on their maximal binding to plastic (Perris and Bronner-Fraser, 198913). The bars represent mean values with corresponding standard deviations from a total of 12-20 explants examined.

bronectin, vitronectin, laminin-nidogen, and collagen types IV and VI were preincubated with PG-LA and subsequently washed prior to plating of the cells (Table 2). Preincubation of collagen type I substrates, however, caused a reduction of cell movement comparable to that seen in the continuous presence of PG-LA in the culture medium (Table 2). In an attempt to uncover the mechanism by which neural crest cell migration is affected by the presence of PG-LA in the substrate (e.g., bound to collagen fibers in the matrix), dishes were sequentially coated with PG-LA and the 105-kDa cell-binding fragment of fibronectin. We previously reported that neural crest cell migration on such substrates was progressively reduced with increasing coating concentrations PG-LA, which was assumed to result in proportionally decreasing amounts of the 105-kDa fragment incorporated into the substrate (Perris and Johansson, 1987). By quantitating Eflects of Substrate-Bound PG-LA on Cell Migration the relative amounts of PG-LA and 105-kDa fragment in the substrate, we show here that the reduction in the Since other types of galactosaminoglycan-containing extent of neural crest cell movement on these composite proteoglycans have been reported to inhibit cell adhe- substrates directly correlates with the amount of 105sion by binding to the substrate (Knox and Wells, 1979; kDa bound to the plastic (Fig. 5). Thus, substrate-bound Rich et al., 1981; Brennan et al., 1983; Lewandowska et PG-LA did not reduce the response of the cells to the al., 1987; Yamagata et al., 1989), we investigated the 105-kDa fragment by migratory-inhibiting signals, but possibility that PG-LA acted in a similar manner. inhibited cell migration by being a nonpermissive subWhen tested in a sensitive solid-phase binding assay, no strate molecule which competed with the 105-kDa fragsignificant binding of PG-LA to dishes coated with fiment for surface area on the substrate. bronectin, the 105-kDa fragment, vitronectin, lamininnidogen, and collagen types IV and VI with PG-LA Stmctural Requirements for the Inhibitory could be detected (Fig. 4). In contrast, PG-LA bound Activity of PG-LA avidly to collagen types I-III and V (Fig. 4). In agreeFor further investigation of the molecular mechament with the above results, cell migration was not affected when BSA/ovalbumin-blocked substrates of fi- nisms underlying the inhibitory effect of soluble PG-LA

PERRIS AND JOHANSSON

Inhibition

of Neural

Crest Cell 200-

by Proteoglycans

Migration

5

b 0-o A-A n-0

Cd I Cd N COIW

150--

5O%Inhlbklon

04 1.0

PG-IA

added

&g/ml)

FIG. 2. (a, b) Dose-dependent inhibition of neural crest cell migration soluble PG-LA. Dotted lines correspond to the level of 50% inhibition.

10

PG-LA

on cell motility-promoting The graphs are derived

100

added

1000

(&ml)

proteins (a) and collagens (b) by addition from 8-15 explants/data point.

of

on neural crest cell migration, the importance of difHABr (Table 4). To rule out that the antiserum could ferent structural units of PG-LA was analyzed. Isolated interfere nonspecifically with the inhibitory action of chondroitin sulfate chains at concentrations up to 3 PG-LA, the proteoglycan was also preincubated with mg/ml did not affect cell migration on fibronectin, or the monoclonal antibody 7-D-4 recognizing its native the 105-kDa fragment (Table 3), demonstrating that the chondroitin sulfate chains. Preincubation of PG-LA dominating glycosaminoglycan of PG-LA did not per- with this antibody had only a minor effect on the inhibiturb locomotion through polyanionic interactions. Ad- tory efficiency of the proteoglycan (Table 4). ditional evidence that the inhibitory effect of PG-LA Since hyaluronan is known to be expressed on the was not exerted by the chondroitin sulfate chains acting surface of migrating neural crest cells (Manasek and independently of the core protein was provided by the Cohen, 1977; Pintar, 1978; Glimelius and Pintar, 1981; observation that reduced and alkylated PG-LA essen- Luckenbill-Edds and Carrington, 1988; Perris et aZ., tially lost its inhibitory efficiency on fibronectin (P 1989b,c), we tested further the possibility that the hy< 10-l’ at 500 pg/ml) and completely failed to inhibit aluronan-binding activity of HABr mediated the mimigration on the isolated 105-kDa fragment (P = 0.079; gration-inhibitory effect of PG-LA. Hyaluronan oligoFig. 6). saccharides (comprising 16-30 monosaccharides) were When added to the medium, the isolated HABr of used as specific competitors of the PG-LA-hyaluronan PG-LA and a proteolytic fragment comprising its kerainteraction (Hascall and Heinegbrd, 1979; Johansson et tan sulfate-enriched region were ineffective in impairal., 1985) and were found to reverse the inhibitory effect ing neural crest cell motility on intact fibronectin or the of PG-LA on cell migration in a dose-dependent fashion 105-kDa fragment. Addition of a similar excess of the similar to that observed with the anti-HABr antiserum. core protein free of glycosaminoglycan side chains re- Oligosaccharide blockage of PG-LA restored neural duced the migratory ability of neural crest cells only crest cell migration to 79% of that occurring in the marginally (Table 3). In the presence of 0.7-l mg/ml of absence of PG-LA (Table 4; Fig. 3). Addition of the core protein, which corresponds to a molar equivalent of hyaluronan oligosaccharides without PG-LA had no efan about lo- to 20-fold higher concentration of intact fect on neural crest cell migration on fibronectin (Table PG-LA, the number of neural crest cells moving on 4; Fig. 3) or the 105-kDa fragment (data not shown). substrates of fibronectin or the 105-kDa fragment was To further ascertain that the interaction of PG-LA reduced by 25 and 15% (P < 0.005), respectively. In with cell membrane-associated hyaluronan was a precontrast, the distance of dispersion from the neural requisite for the inhibition of cell migration, the neural tube explant as well as the pattern of outgrowth was crest cells were digested with Streptmyces hyaluronilargely unaffected. dase. Incubation of the cells with the enzyme reduced Preincubation of 500 pg/ml PG-LA, a concentration the inhibitory effect of PG-LA by 46% (P = 5.8 X 10-15; causing virtually complete inhibition of neural crest cell Table 4). Heat-inactivated hyaluronidase did not have a dispersion, with an antiserum against the HABr signifsimilar effect. Enzyme treatment without addition of icantly (55%; P = 3.8 X 10-14) decreased the inhibitory PG-LA inhibited neural crest cell migration on fibroeffect of the PG-LA on cell migration (Table 4). The nectin by 17% (P = 0.0036; Table 4). Conceivably, the effect of the anti-HABr antiserum was strongly abroresidual inhibitory effect of PG-LA on hyaluronidasegated by preincubation of it with an excess of isolated treated cells could be due to incomplete degradation of

6

DEVELOPMENTALBIOLOI GY

VOLUME 137,199O

Inhibition

PERRIS AND JOHANSSON

of Neural Crest Cell Migration

I 1.0

...*

n

A-A

E

v--v n-.-u

0.6

I

0-

0

by Proteoglgcans

7

TABLE 2 NEURAL CREST CELL MIGRATION ON CELL ADHESION/MOTILITYPROMOTING PROTEINS AND COLLAGEN SUBSTRATES WITH (+) OR WITHOUT (-) PREINCUBATION OF THE SUBSTRATES WITH PG-LA

Plaatia FN 10516) VW IN-N

Substrate

-

FN FN

I

/

L

2

/I-----

T .

+

VN VN

-

LN-N LN-N

-

+ +

co1 I co1 I co1 IV co1 IV

/’

Preincubation with PG-LA

co1 VI co1 VI

+ + +

Note. FN, fibronectin; 0.1

1.0 PG-IA

10.0

100.0

1000.0

I-VI,

Number of explants

Localization zone O-4 (mean)

Number of cells emigrated (mean + SD)

10 12

3 (3.8) 4 (3.8)

426.0 + 36.9 408.7 -t 16.5

12 16

4 (3.6) 4 (3.8)

408.6 f 38.7 414.1 + 16.8

12 15

3 (2.9) 2 (2.7)

278.3 + 24.8 287.5 f 19.5

20 12

2 (2.2) 0 (0.2)

162.1 k 30.1 24.2 2 6.2

16 12

2 (2.1) 2 (2.1)

150.2 f 16.2 143.9 f 19.8

15 11

2 (2.3) 2 (2.2)

147.3 f 13.1 155.9 -+ 19.9

VN, vitronectin;

LN-N, laminin-nidogen;

Co1

collagen types I-VI.

added (&ml)

FIG. 4. Binding of soluble PG-LA to various matrix molecules. Mierowell plates were coated with the concentration of the molecules (except for collagen types I-III) resulting in maximal adsorbance to the plastic (Perris and Johansson, 1987; Perris et aL, 1989a; Perris and Bronner-Fraser, 1989b). Areas of the plastic that remained available after this first coating were saturated with BSA/ovalbumin. Substrates of collagen types I-III were prepared the same as those used for cell migration assays (see Fig. 1).

cell surface hyaluronan and/or increased hyaluronan synthesis as a response to the enzymatic treatment. DISCUSSION

While experimental data show that several glycoproteins and certain collagens are permissive or even stimulatory for neural crest cell migration, it has become increasingly evident that certain proteoglycans have restricting functions in this process. In a recent study we showed that PG-LA does not support initial migration of axolotl neural crest cells in vitro and that its inclu-

sion in the culture medium strongly perturbs dispersion of neural crest cells on fibronectin substrates (Perris and Johansson, 1987). These findings were partly corroborated in a parallel study in which it was shown that a chondroitin sulfate proteoglycan, purified from embryonic chick brain, was nonpermissive as a substrate for avian neural crest cell migration (Tan et ah, 1987). In addition, there are several indications that regional accumulations of chondroitin/keratan sulfate proteoglycans may direct neural crest cell migration in the embryo (Tucker, 1986; Tucker and Erickson, 1986; Newgreen et al, 1986; Liifberg et al., 1989a; Perris et al., 1989b,c). In this study we have extended the examination of the role of cartilage-type proteoglycans during neural crest development and show that the inhibitory effect of PG-LA on cell motility is not restricted to substrates composed of fibronectin. PG-LA also perturbed dose-dependently neural crest cell movement on vitronectin, laminin/laminin-nidogen, and collagen

FIG. 3. Phase-contrast micrographs showing the representative migration of neural crest cells on various substrates in the presence or absence of PG-LA. Addition of the proteoglycan to neural crest cells plated on laminin (LN) results in a virtually complete blockage of migration (A, B). Neural crest cells migrate to some extent on collagen type IV (IV) and PG-LA strongly affects the migratory pattern on this extracellular matrix component (C, D). Neural crest cells are capable to disperse on collagen type III (Co1 III; E), but migration on this molecule is notably less pronounced than that on motility-promoting proteins and collagen types I, IV, and VI. Addition of hyaluronan oligosaeeharides to neural crest cells migrating on fibronectin substrates does not alter their migratory behavior (F). In contrast, addition of PG-LA strongly impedes neural crest cell migration on fibronectin (G), whereas its inhibitory activity is abolished by preincubation with hyaluronan oligosaceharides (H). In D and H, the photographs were taken at the edge of the neural tube explants. nt, neural tube. X95 except in D X190.

8

DEVELOPMENTAL BIOLOGY Coating concentration z.cm-3

0.01

0.700

o-o

PodA +

E c

0.500

q - -0 b.. -A

loold) M

8 10

OJW

-

10.00

11y)

0.10

(j&ml)

10sla (10pg/n@ P

p’

” \,

IT:-

.x?0.400 e 4 0.3w B o.zm ‘xi R I o.lDO 0.m -

D

1OO.M

-

,---o-u

‘i5

J...X

.-

x. .

p”

I

:

.

~\~...~ :.\a\ - .4 A :

&...A . . . . *...b..”

\

a’ .

j 400.0 .P fi g 300.0 8 y 3 ii it

m.0 100.0

z'

VOLUME 137.1990

yolk sac chondroitin sulfate proteoglycan, named PG19, has been found to perturb binding of yolk sac tumor cells to intact fibronectin, but not to collagen type IV nor to a 120-kDa fragment comprising the cell-binding domain (Brennan et al., 1983) and lacking the predominant heparin-binding domain of fibronectin (Benecky et ah, 1988). PG19 also seems to perturb binding of the isolated fibronectin receptor to its ligand (Hautanen et al, 1989). The proteoglycan PG-M has recently been reported to modulate fibroblastic cell-substrate adhesion when immobilized on extracellular substrates (Yamagata et ah, 1989). Furthermore, the small dermatan sulfate proteoglycans from bovine articular cartilage has been suggested to prevent cell attachment to fibronectin by binding to a “cryptic” glycosaminoglycanbinding region located within the 120-kDa fibronectin fragment (Lewandowska et al., 1987; Mugnani et al., 1988). The findings of this study differ from those described above. In the case of inhibition of neural crest cell migration by PG-LA, a mechanism independent of binding to the substrate seems to operate. Preincubation of fibronectin, vitronectin, laminin-nidogen, and collagen

0.0 0.01

0.1

0.3

1

Caatlng concantmtlon PC-U (w/ml)

3

10

100

+ 105 kD (10 w/ml)

FIG. 5. Solid-phase binding assays for the determination of the relative amount of PG-LA and 105-kDa fragment bound to plastic in composite substrates with different relative proportions of the two molecules, and neural crest cell migration on equivalent substrates. The upper graph shows binding of biotinylated 105 kDa to plastic precoated with increasing concentrations PG-LA. Bindings of the 105-kDa fragment and PG-LA alone are also indicated for comparison. The lower graph shows the extent of neural crest cell migration on composite substrates containing the corresponding relative proportions of the two molecules.

types I, III, IV, and VI. This ability of the proteoglycan to interfere with neural crest cell migration on several motility-promoting matrix molecules suggests that the macromolecule might affect a general mechanism of cell-substrate interaction required for movement. Cartilage-type chondroitin sulfate proteoglycans have previously been reported to inhibit the fibronectin-mediated adhesion of fibroblastic cell types to interstitial collagens, as concluded by the investigators, through binding to the collagen substrate (Knox and Wells, 1979; Rich et al., 1981). However, the same proteoglycans have been found ineffective in altering the attachment of 3T3 cells (Lewandowska et al., 198’7) and human neuroblastoma cells (Mugnani et al, 1988) to fibronectin in the absence of collagen, and chick retinal neurons to fibronectin and laminin (Carri et ab, 1987). A

TABLE 3 EFFECTS OF PG-LA AND ITS ISOLATED DOMAINS ON NEURAL CREST CELL MIGRATION ON FIBRONECTIN AND ITS CELL-BINDING FRAGMENTS

Substrate

Addition to the medium

Number of explants

Localization zone O-4 (mean)

Number of cells emigrated (mean f SD)

FN FN

PG-LA

10 12

4 (3.8) 1 (0.9)

426.0 5~36.9 107.3 + 19.0

105 kDa 105 kDa

PG-LA

12 16

4 (3.9) 0 (0.1)

490.3 ?z 23.3 18.0 f 3.9

11.5 kDa 11.5 kDa

PG-LA

15 14

3 (2.7) 0 (0.2)

270.5 + 18.8 25.4 k 6.1

FN 105 kDa

cs cs

14 11

4 (3.7) 4 (3.8)

406.7 + 32.1 468.5 Y!C59.1

FN 105 kDa

HABr HABr

19 18

4 (3.8) 4 (3.9)

421.3 t 26.3 485.4 k 19.8

FN 105 kDa

CP CP

18 7

4 (3.6) 4 (3.7)

329.9 + 24.9 398.7 f 37.4

FN

KSF

9

4 (3.7)

423.6 _+22.3

Note. Isolated ehondroitin sulfate chains (CS), hyaluronan-binding fragment (HABr), the core protein free of glycosaminoglycans (CP), and the keratan sulfate-rich fragment (KSF) were added at 3.0, 0.042-0.125,0.7-1.0, and 0.55 mg/ml, respectively. The average migration distance (localization zone) of the dispersed neural crest cells is indicated by mean values within parentheses. The total number of neural crest cells that emigrated from the neural tube explant on each substrate is indicated in the right column as mean k SD.

PERRIS AND JOHANSSON a

El NoaddMon la2!lPG-u

I

I

Inhibition

Fibronectin

of Neural

1

PG-lARcdAlk

105 kDa

I 0

0.05

0.5 PG-IA

5.0

I

50

Crest Cell Migration

bg Proteoglycans

9

findings indicate that binding of the PG-LA to hyaluronan exposed on the cell surface is a prerequisite for its inhibitory action on neural crest cell migration. In addition, the intact PG-LA appeared to be required for this activity since the isolated HABr was not able to mimic the effect of the native proteoglycan, and the core protein free of glycosaminoglycan side chains affected only marginally the migratory ability of the cells. Formation of cell surface-associated hyaluronanPG-LA complexes have been described for cultured chondrocytes (Sommarin and HeinegHrd, 1983). Other studies have also shown that association of PG-LA with hyaluronan on the cell surface can alter the interaction of chondrocytes with fibronectin. When plated on this glycoprotein in the absence of surrounding PG-LA, chondrocytes acquire a fibroblastic morphology and

a

s-00

1500

added (&ml)

FIG. 6. (a, b) Comparison of the inhibitory effect of intact PG-LA and PG-LA after reductive alkylation on neural crest cell migration on substrates of fibronectin (a) or the 105-kDa fragment of fibronectin (b). A total of 9-13 explants per experimental case was analyzed. Error bars denote standard deviations of the means.

types IV and VI substrates with soluble PG-LA did not impair the locomotory ability of neural crest cells. These results lead to the conclusion that the minimal binding of PG-LA to these cell motility-promoting proteins was not sufficient to perturb cell movement. Additionally, reductive alkylation of the PG-LA essentially abolished its inhibitory activity. This observation suggests that if any interactions between the glycosaminoglycans of PG-LA and the substrate occur, they could potentiate the inhibitory effect of the proteoglycan, but are not primarily responsible for its restriction of neural crest cell migration. Further support for this conclusion derives from the observation that native PG-LA inhibited migration on the 11.5kDa cell-binding fragment of fibronectin, which presumably lacks “cryptic” glycosaminoglycan-binding the postulated, site of the 105-kDa fragment (Lewandowska et al., 1987). The inhibitory effect of soluble PG-LA on neural crest cell migration was drastically reduced by preincubation of the proteoglycan with antibodies against its HABr or hyaluronan oligosaccharides and by treatment of the cells with Streptmyces hyaluronidase. These

A

cell adhesion

n

integrin

protein

PG.LA 7 hyaluronan

anchored

to its synthase

/

FIG. ‘7. Schematic illustration of how aggregating proteoglycans are proposed to influence neural crest cell migration. (a) A neural crest cell migrating on cell adhesion/motility-promoting proteins by use of the corresponding receptors (integrins). (b) Inhibited neural crest cell migration on cell adhesion/motility-promoting molecules in the presence of soluble PG-LA. The proteoglyean, which could be secreted by either the neighboring tissues or the cells themselves (Erickson and Turley, 1987), binds to cell surface hyaluronan anchored to its synthase (Prehm, 1934; Philipson and Schwartz, 1984; Mian, 1986) or specific receptors (Underhill et al, 198’7). Accumulation of a large number of proteoglycan molecules on the cell surface would interfere with the function of cell surface receptors involved in cell movement, by either steric hindrance or specific modulatory interaction.

DEVELOPMENTALBIOLOGYVOLUME137,199O

10

TABLE4 EFFECTS OFPG-LA IN THEPRESENCE OFANTIBODIES AGAINST ITS DOMAINS, STREPTOMYCES HYALURONIDASE, ORHYALURONANOLIGOSACCHARIDESONNEURALCRESTCELLMIGRATIONONFIBRONECTIN Number to the medium

of explants

Localization zone (O-4) (mean)

Number of NC cells emigrated (mean + SD)

Substrate

Addition

FN

PG-LA/HAIGm.?,, 0.510.2 mg/ml 0.5/0.02 mg/ml 0.5/0.002 mg/ml

12 9 11

3 (2.8) 3 (2.7) 1 (1.3)

339.1 + 33.8 223.2 k 21.8 103.4 + 16.7

FN’

HA1s-no (0.2 mg/ml)

9

4 (3.7)

413.2 +_20.7

FN

PG-LA/anti-HABr 0.5/1:10 0.5/1:50 0.5/1:100 0.5/1:200

12 12 12 12

3 2 1 0

220.8 179.2 86.5 47.7

12

2 (2.1)

136.5 + 21.8

mAb

(2.7) (2.0) (1.3) (0.4)

+ 21.3 + 18.3 +_ 13.3 * 13.0

FN

PG-LA/7-D-4

FN

PG-LA/anti-HABr-HABr

11

2 (1.9)

123.6 & 16.5

FN

S-Hyase

18

4 (3.5)

355.6 + 68.9

FN

PG-LA

+ S-Hyase

16

2 (2.4)

191.9 f 57.7

FN

PG-LA

+ S-Hyase heat-inactivated

14

1 (0.9)

48.5 f 12.3

Note. A concentration of 500 pg/ml PG-LA was preincubated with the given concentrations of hyaluronan oligosaccharides (HA,6&, the antiserum against the hyaluronan-binding region (anti-HABr, 1:lO) of the proteoglycan, the monoclonal antibody 7-D-4 (1:lOO dilution) against the native chondroitin sulfate chains, or the anti-HABr antiserum (1:lO dilution) previously incubated with an excess of HABr (25 pg/ml), and the mixture was then added to the cells. ” Addition of HAu-aO to neural crest cell migrating on the 105-kDa fragment was similarly without noticeable effect on cell movement. Digestion of freshly isolated neural tube-neural crest explants with Streptmnyces hyaluronidase (S-Hyase) was performed as described under Material and Methods. Explants were then rinsed once and plated onto fibronectin substrates in the presence of 500 pg/ml PG-LA and 210 U/ml hyaluronidase.

loose their phenotypic characters. In contrast, a pericellular accumulation of PG-LA (bound to cell surface hyaluronan) causes the chondrocytes to round up and preserve their phenotypic state, characterized by synthesis of PG-LA and collagen type II (von der Mark et ab, 1977; Pennypacker et al., 1979). Our results may relate to the well-documented connection between hyaluronan synthesis and cell migration within tissues (Weston et al., 19’78;Toole et al., 1987; Perris et al., 1989c). Several motile cell types have been shown to express both hyaluronan and hyaluronanbinding proteins (Turley and Torrance, 1984; Underhill et ah, 1987). Neural crest cells migrating in vitro have been reported to synthesize hyaluronan and to accumulate the glycosaminoglycan on their surfaces (Manasek and Cohen, 1977; Pintar, 1978; Glimelius and Pintar, 1981). Accumulation of hyaluronan around premigratory neural crest cells seems also to be critical for their initiation of movement from the neural tube (Luckenbill-Edds and Carrington, 1988), and this observation may explain the partial inhibition of movement on fibronectin caused by hyaluronidase treatment. Taking

together our present and previously published results we suggest the following mechanisms to explain how aggregating chondroitin sulfate proteoglycans of cartilage-type may influence neural crest cell migration in vivo. It is proposed that when associated with hyaluronan and/or certain collagens of the extracellular matrix, proteoglycans can inhibit neural crest cell migration when their local abundance surmounts that of the motility-promoting molecules. In addition, proteoglycans may bind to cell surface hyaluronan and thereby alter the response to the microenvironment. Accumulation of a great number of such large and negatively charged macromolecules around the cells could interfere with the activity of receptors for the motilitypromoting components of the extracellular matrix (Fig. 7). On the other hand, an elevated rate of hyaluronan synthesis in neural crest cells may favor their migration by causing transfer or shedding of hyaluronanbinding proteoglycans away from the cells. If aggregating proteoglycans become linked to the cell surface and thereby modulate cell surface receptors in vivo, it would provide a specific means for the embryo

PERRIS AND JOHANSSON

Inhibitimz of Neural Crest Cell Migration by Proteoglycans

to regulate neural crest cell migration. Particularly, the spatial and temporal biosynthesis of proteoglycans/ hyaluronan in neural crest cells and their surrounding tissues could contribute to the control of the onset, directionality, and cessation of neural crest cell migration. We are grateful to Drs. Bruce Caterson, Ulf Lindahl, Mats Paulsson, Kristofer Rubin, Erkki Ruoslahti, Rupert Timpl, and Jiirgen Wieslander for most generously providing various purified extracellular matrix molecules. We are also indebted to Drs. Dick Heineglrd for providing various purified proteoglycan domains and valuable suggestions; Bruce Caterson for characterizing the anti-HABr antiserum; Marianne Bronner-Fraser for her financial support during the final phases of the work; Susan Bryant for providing some of the axolotl embryos; Michael Sorrel1 for donation of the 7-D-4 monoclonal antibody; and Mats Paulsson for critical reading of the manuscript. This study was primarily carried out at the Department of Zoology, Uppsala University, Uppsala, Sweden, under the financial support of grants from the Swedish Natural Science Research Council (Grant B-BU 3910-100) to Dr. Jan Lofberg, the Swedish Medical Research Council (Grant 7147), the King Gustav V 80th Birthday Fund, and the Torsten & Ragnar Siiderberg Foundation to S.J. REFERENCES BAKER, J. R., RODEN, L., and STOOLMILLER, A. C. (1972). Biosynthesis of chondroitin sulfate proteoglycans. J. BioL Chem. 247,3838-3847. BENECKY, M. J., KOLVENBACH, C. G., AMRANI, D. L., and MOSESSON, M. W. (1988). Evidence that binding to the carboxyl terminal heparin-binding domain (Hep II) dominates the interaction between plasma fibronectin and heparin. Biochemistry 27,7565-7571. BOUCAUT, J.-C., DARRIBERE, T., POOLE, T. J., AKIYAMA, H., YAMADA, K. M., and THIERY, J.-P. (1984). Biologically active synthetic peptides as probes of embryonic development: A competitive peptide inhibition of fibroneetin function inhibits gastrulation in amphibian embryos and neural crest cell migration in avian embryos. J. Cell Biol. 99, 1822-1830. BRENNAN, M. J., OLDBERG, A., HAYMAN, E. G., and RUOSLAHTI, E. (1983). Effect of a proteoglycan produced by rat tumor cells on their adhesion to fibronectin-collagen substrata. Cancer Res. 43, 4302-4307. BRONNER-FRASER, M. (1986). An antibody to a receptor for fibronectin and laminin perturbs cranial neural crest development in uivo. Dev. BioL 117,528-536. CARRI, N. G., PERRIS, R., JOHANSSON, J., and EBENDAL, T. (1987). Differential outgrowth of retinal neurites on purified extracellular matrix molecules. J. Neurosci Res. 19, 428-439. CATERSON, B., CHRISTNER, J. E., and BAKER, J. (1983). Identification of a monoclonal antibody that specifically recognizes cornea1 and skeletal keratan sulfate. J. BioL Chem. 8848-8854. DUBAND, J.-L., and THIERY, J.-P. (1987). Distribution of laminin and collagens during avian neural crest development. Development 101, 461-478. DUFOUR, S., DUBAND, J.-L., HUMPHRIES, M. J., OBARA, M., YAMADA, K. M., and THIERY, J.-P. (1988). Attachment, spreading and locomotion of avian neural crest cells are mediated by multiple adhesion sites on fibronectin molecules. EMBO J. 7,2661-2671. ERICKSON, C. A., and TURLEY, E. A. (1987). The effects of epidermal growth factor on neural crest cells in tissue culture. Exp. Cell Res. 169,267-279. GLIMELIUS, B., and PINTAR, J. E. (1981). Analysis of developmentally homogeneous neural crest cell populations in vitro. IV. Cell prolif-

11

eration and synthesis of glycosaminoglycans. Cell Difer. 10, 1’73-182. HASCALL, V. C., and HEINEG~RD, D. (1979). Aggregation of cartilage proteoglycans. II. Oligosaccharide competitors of the proteoglycan-hyaluronic acid interaction. J. Biol Chem. 249,4242-4249. HAUTANEN, A., GAILIT, J., MANN, D. M., and RUOSLAHTI, E. (1989). Effects of modifications of the RGD sequence and its context on recognition by fibronectin receptor. J. BioL Chem. 264,1437-1442. HEINEG~RD, D., and SOMMARIN, Y. (1987). Isolation and characterization of proteoglycans. In “Methods in Enzymology” (L. W. Cunningham, Ed.), Vol. 144, pp. 319-372. Academic Press, Orlando. JOHANSSON, S., HEDMAN, K., KJELLEN, L., CHRISTNER, J., VAHERI, A., and HOOK, M. (1985). Structure and interactions of proteoglycans in the extracellular matrix produced by cultured human fibroblasts. Biochem J. 232,161-168. KNOX, P., and WELLS, P. (1979). Cell adhesion and proteoglycans. I. The effect of exogenous proteoglycans on the attachment of chick embryo fibroblasts to tissue culture plastic and collagen. J. Cell Sci. 40,77-88. LEWANDOWSKA, K., CHOI, H. U., ROSENBERG, L. C., ZARDI, L., and CULP, L. A. (1987). Fibronectin-mediated adhesion of fibroblasts: Inhibition by dermatan sulfate proteoglycan and evidence for a cryptic glycosaminoglycan-binding domain. J. Cell Biol. 105, 1443-1454. LBFBERG, J., AHLFORS, K., and FALLSTROM, C. (1980). Neural crest cell migration in relation to extracellular matrix organization in the embryonic axolotl trunk. Dev. BioL 75,148-167. L~~FBERG,J., NYN&-MCCOY, A., OLSSON, C., JBNSSON, L., and PERRIS, R. (1985). Stimulation of initial neural crest cell migration in the axolotl embryo by tissue grafts and extracellular matrix transplanted on microcarriers. Dev. BioL 107,442-459. L~)FBERG, J., EPPERLEIN, H. H., PERRIS, R., and STIGSON, M. (1989a). Neural crest cell migration in the axolotl embryo: a pictorial essay. In “The Developmental Biology of the Axolotl” (J. B. Armstrong and G. M. Malacinski, Eds.), pp. 83-101. Oxford Univ. Press, Oxford. L~FBERG, J., PERRIS, R., and EPPERLEIN, H. H. (1989b). Timing in the regulation of neural crest cell migration: Retarded “maturation” of regional extracellular matrix inhibits pigment cell migration in embryos of the white axolotl mutant. Dev. BioL 131,168-181. LUCKENBILL-EDDS, L., and CARRINGTON, J. L. (1988). Effect of hyaluranic acid on the emergence of neural crest cells from the neural tube of quail, Coturnix coturnix japonica. Cell Tissue Res. 252,

573-579. MIAN, N. (1986). Characterization of a high-M, plasma-membranebound protein and assessment of its role as a constituent of hyaluronate synthase complex. Biochem. J. 237,343-357. MANASEK, F. J., and COHEN, A. M. (1977). Anionic glycopeptides and glycosaminoglycans synthesized by embryonic neural tube and neural crest. Proc. NatL Acad Sci. USA 74,1057-1061. MUGNANI, G., LEWANDOWSKA, K., CHOI, H. U., ROSENBERG, L. C., and CULP, L. A. (1988). Ganglioside-dependent adhesion events of human neuroblastoma cells regulated by the RGDS-dependent fibronectin receptor and proteoglycans. Exp. Cell Res. 175,239-247. NEWGREEN, D. F., and ERICKSON, C. A. (1986). The migration of neural crest cells. Int. Rev. Cytol. 10, 89-145. NEWGREEN, D. F., SCHEEL, M., and KASTNER, V. (1986). Morphogenesis of sclerotome and neural crest in avian embryos. In vivo and in vitro studies on the role of notochordal extracellular material. Cell

Tissue Res. 244,299-313. PENNYPACKER, J. P., HASSELL, J. R., YAMADA, K. M., and BRATT, R. M. (1979). The influence of an adhesion cell protein on chondrogenic expression in vitro. Exp. Cell Res. 121,411-415. PERRIS, R., and JOHANSSON, S. (1987). Amphibian neural crest cell migration on purified extracellular matrix components: A chon-

12

DEVELOPMENTALBIOLOGY

VOLUME 137,199O

droitin sulfate proteoglycan inhibits locomotion on fibronectin SORRELL,M. J., MAHMOODIAN,F., SCHAFER,I. A., DAVIS, B., and CATERSON,B. (1989). Identification of monoclonal antibodies that recsubstrates. J. Cell Biol. 105,2511-2521. ognize novel epitopes in native chondroitin/dermatan sulfate glyPERRIS, R., and BRONNER-FRASER,M. (1989a). Recent advances in cosaminoglycan chains: Their use in mapping functionally distinct defining the role of the extracellular matrix in neural crest development. Comm. Dev. NeurobioL 1,61-83. domains of the human skin. Biochem. J., in press. PERRIS, R., and BRONNER-FRASER,M. (198913).Differential neural TAN, S.-S., CROSSIN,K. L., HOFFMAN,S., and EDELMAN, G. M. (1987). Asymmetric expression in somites of cytotactin and its proteoglycrest cell migration on isolated collagens. Importance of RGD as a can ligand is correlated with neural crest cell distribution. Proc. potential recognition site. Submitted for publication. PERRIS,R., PAULSSON,M., and BRONNER-FRASER, M. (1989a). MolecuNat1 Acad. Sci. USA 84,7977-‘7981. lar mechanisms of neural crest cell migration on fibronectin and TOOLE,B. P., KNUDSON,C. B., KNUDSON,W., GOLDBERG,R. L., CHIlaminin. Dev. Biol. 136,222-238. Rosso, G., and BISWAS,C. (1987). Hyaluronate-cell interactions in PERRIS, R., KROTOSKI, D., LALLIER, T., and BRONNER-FRASER,M. morphogenesis and tumorigenesis. In “Mesenchymal-Epithelial (1989b). Distribution of chondroitin/dermatan/keratan sulfate Interactions in Neural Development” (J. R. Wolff, J. Silver, and M. proteoglycans and interstitial collagens correlates spatiotempoBerry, Eds.), pp. 267-278. Academic Press, New York. rally with neural crest cell migration in the chick embryo. Submit- TUCKER,R. P. (1986). The role of glycosaminoglycans in anuran pigted for publication. ment cell migration. J. Embrgol. Exp. Morphol. 92, 145-164. PERRIS,R., L~FBERG,J., FALLSTROM,C., OLSSON,L., VONBOXBERG,Y., TUCKER,R. P., and ERICKSON,C. A. (1984). Morphology and behavior and NEWGREEN,D. F. (1989c). Ultrastructural, immunohistochemiof quail neural crest cells in artificial three-dimensional extracelcal and biochemical differences in the extracellular matrix encounlular matrices. Dev. Biol. 104,390-405. tered by neural crest cells in the white mutant embryo. Daelop TUCKER,R. P., and ERICKSON,C. A. (1986). Pigment cell pattern formation in Taricha torosa: The role of the extracellular matrix in ment, in press. PHILIPSON,L. H., and SCHWARTZ,N. B. (1984). Subcellular localization controlling pigment cell migration and differentiation. Dev. Biol. of hyaluronate synthetase in oligodendroglioma cells. J. Biol. Chem 118,268-285. 259,5017-5023. TURLEY, E. A., and TORRANCE,J. (1984). Localization of hyaluronate PIERSCHBACHER,M. D., ROUSLAHTI,E., SUNDELIN,J., LIND, P., and and hyaluronate-binding protein on motile and non-motile fibroPETERSON,P. A. (1982). The cell attachment domain of fibronectin. blasts. Exp. Cell Res. 161, 17-28. Determination of the primary structure. J. Biol. Chem. 257, UNDERHILL, C. B., GREEN, S. J., COMOGLIO,P. M., and TARONE, G. 9593-9597. (1987). The hyaluronate receptor is identical to a glycoprotein of M, PINTAR, J. E. (1978). Distribution and synthesis of glycosaminogly85,000 (gp85) as shown by a monoclonal antibody that interferes with binding activity. J. Biol. Chem. 262, 13,142-13,146. cans during quail neural crest morphogenesis. Dew. Biol. 67, VON DER MARK, K., GAUSS,V., VON DER MARK, H., and MULLER, P. 444-464. PREHM, P. (1984). Hyaluronate is synthesized at the plasma mem(1977). Relationship between cell shape and type of collagen synthebrane. B&hem. J. 220,597-600. sized as chondrocytes loose their cartilage phenotype in culture. RICH, A. M., PEARLSTEIN,E., WEISSMANN,G., and HOFFSTEIN,S. T. Nature (London) 267,531-532. (1981). Cartilage proteoglycans inhibit fibronectin-mediated adhe- WESTON,J. A., DERBY,M. A., and PINTAR, J. E. (1978). Changes in the extracellular environment of neural crest cells during their early sion. Nature (London) 293,224-226. SOMMARIN,Y., and HEINEG~RD, D. (1983). Specific interaction bemigration. Zoon 6,103-113. tween cartilage proteoglycans and hyaluronic acid at the chondro- WOODS,A., COUCHMAN,J. R., JOHANSSON,S., and HOOK, M. (1986). Adhesion and cytoskeletal organization of fibroblasts in response cyte cell surface. Biochem. J. 214,777-784. SORRELL,M. J., LINTALA, A. M., MAHMOODIAN,F., and CATERSON,B. to fibronectin fragments. EMBO J. 5,665-670. (1988). Epitope-specific changes in chondroitin sulfate/dermatan YAMAGATA, M., SUSUKI, S., AKIYAMA, S. K., YAMADA, K. M., and sulfate proteoglycans as markers in the lymphopoietic and granuKIMATA, K. (1989). Regulation of cell-substrate adhesion by prolopoietic compartments of developing bursae of Fabricius. J. Imteoglycans immobilized on extracellular substrates. J. BioL Chem. munol. 140,4263-4270. 264.8012~8018.