Expression of hyaluronic acid-binding glycoprotein, hyaluronectin, in the developing rat embryo

DEVELOPMENTAL

BIOLOGY

101.391-4~

(1984)

Expression of Hyaluronic Acid-Binding Glycoprotein, Hyaluronectin, in the Developing Rat Embryo ANNIE DELPECH*

AND BERTRAND

DELPEcHt

“Laboratoll/ of Histobgg, Hop&J Charles Nicolle, rue de Germat, 76O.WRowm, and tLa.bcrratorg of Immunochemistry, Centre Henri Becquerel, rue d’Amiens, 760.?8-Row-n, France Received O&be-r 25, 1982; accepted in revised form September 6, 1983 Immunological and histological methods have been applied to the developing rat embryo to study the distribution of hyaluronectin (HN, a glycoprotein with hyaluronic acid-binding properties) previously shown to be present in the nervous system and in desmoplasias. HN was absent in the morula and the blastula and was first detected in the mesenchyme bordering the neural tube and somites on Day 10, i.e., at a time when hyaluronic acid is already widely dispersed in the mesenchyme. At this stage HN appeared to be closely associated with the basement membrane around the epithelial structures (somites, notochord, ectoderm) whereas the intercellular areas of mesenchyme were less strongly stained. The delineation of basement membranes decreased progressively, while the accumulation of HN increased in the cell-free areas of mesenchyme, giving a continuous, diffuse pattern. Differentiation of niesenchyme into vertebral cartilage and gut smooth muscle was accompanied by a progressive disappearance of HN. Even after streptomyces hyaluronidase or chondroitinase digestion the antigen was not unmasked in these tissues. The results are in agreement with the few observations made in the human. They suggest that HN could play a role, in association with fibronectin and glycosaminoglycans (hyaluronic acid), in the physiology of the embryonic extracellular matrix. HN appeared at a later stage in the embryonic nervous tissue; its distribution was extracellular in areas where both cell migration and proliferation occur. INTRODUCTION

Hyaluronectin is a glycoprotein which has been isolated from adult human brain and shown to bind specifically to hyaluronic acid in vitro (Delpech and Halavent, 1981). The specificity of binding has been further confirmed by the demonstration that mixing HN’ and HA resulted in high-molecular-weight complexes without any loss of HN antigenicity. Complexes were not formed when other GAGS were substituted for HA (Delpech, 1982). HN is a protein antigen whose antigenicity is destroyed by proteolytic enzymes such as trypsin and pronase but is unaltered by hyaluronidase and chondroitinase. Its mean molecular weight is 68,000 but polyacrylamide gel electrophoresis and gel chromatography indicate that forms exist with molecular weights ranging from 45,000 to 110,000. In the adult nervous system HN is localized at the nodes of Ranvier and in the microenvironment of a proportion of neurones (Delpech et d, 1982b). I-IN, however, is not restricted to nervous tissue, We have previously observed by immunohistological techniques that it is present in other adult organs, i.e., loose connective tissue ’ Abbreviations used: HA, hyaluronic acid, HN, hyaluronectin; FN, fibronectin; GAG, glycosaminoglycans; PBS, phosphate-buffered saline (NaCl 8 g/liter buffered at pH 7.2 with 0.01 M sodium phosphate); SDS, sodium dodecyl sulfate.

(also noted for its content of HA), the kidney papilla, the subendothelium of arteries, the intralobular connective tissue of mammary glands, and the interfascicularis network of the muscularis externa of the digestive tract (Delpech et aL, 1978). In the human fetus it has been found in much higher quantities, mainly in the upper layer of the dermis, the periannexial dermis, and the mucosal connective tissue of the gut (Delpech d c& 1978). In addition, we have investigated this antigen in human tumors and shown it to be present in cancers of mesenchymal origin (fibrosarcoma, myxoma) and in the reactive connective tissue associated with carcinomas (Delpech et a& 19’79b). Similar results have been observed in human tumors grafted into nude mice (Girard et d, 1982). These data prompted us to investigate HN in developing embryonic mesenchyme, a tissue known to have a high content of HA (Pratt et al, 1975; Solursh and Morriss 1977; Morriss and Solursh, 1978). We used the rat as an experimental model as we have found that it possesses an analogous antigen which cross-reacts with anti-human HN antiserum and which has a histological localization similar to that of human antigen (Delpech et al, 1978,1982b; Chevrier d a& 1979). The work presented here was designed to answer three questions: 391

0012-1606134 $3.00 Copyright All righte

8 19fM by Academic Press, Inc. of reproduction in any form reserved.

DEVELOPMENTAL BIOLOCY

392

VOLUME 101,19&p

(1) -When does HN appear during rat embryogenesis? (2) -Where is it first located? (3) -Do the results support a hypothesis of HN-HA binding in viva?

MATERIALS

AND METHODS

Purification of [email protected] Our previously described method (Delpech and Halavent, 1981) has been slightly modified. The hyaluronic acid absorbent was prepared by coupling HA to AH Sepharose using carbodiimide (Tengblad, 1979). HA was not submitted to hydrolysis and 20 mg of it was used in 20 ml deionized water/4 ml wet gel. The proportion of HA binding was estimated at 60% (3 mg/ml wet gel) using the carbazole reaction of Bitter and Muir (1962). HN was extracted from a fragment of human brain, obtained within 6 hr of death, by grinding in 0.2 M glycine HCl buffer, pH 2. After centrifugation it was neutralized, dialized against PBS, and recentrifuged at 40,OOOg for 10 min at 4°C. The supernatant was incubated with the HA absorbent (18 ml/4 ml wet gel) for 40 min at room temperature. It was washed with 0.5 liter of 0.5 MNaCl buffered at pH 7.2 by 0.01 M phosphate. HN was eluted by 0.2 M glycine HCl at pH 2.2, dialyzed against PBS, precipitated by 50% saturated ammonium sulphate, redissolved in PBS, and dialyzed against PBS in order to obtain a final concentration between 0.5 and 1 mg/ml. The purity of the HN preparation was confirmed immunologically and by acrylamide gel electrophoresis. In Ouchterlony plates, a single line of precipitation was seen between the HN sample and anti-HN antiserum or polyspecific anti-human brain antiserum. The absence of a precipitation line with anti-human plasma antiserum and an anti-human liver antiserum indicated the absence of detectable plasma or tissue antigen. Biochemical purity was assessed by slab gel electrophoresis. The 7% acrylamide gel (Cellacryl, Sebia, Paris) was buffered in Tris-glycine (Tris 7.05 g/l, glycine 11.3 g/l) without SDS. One band was obtained which was stained by Coomassie blue. The affinity of the protein for HA was demonstrated by suppression of the protein band when HN was run in the presence of HA (Fig. 1). Under this condition staining material which was composed of HN-HA complexes accumulated in the stacking gel. This blocking test could not be done in the presence of 0.1% SDS which dissociates HN-HA complexes. PurQication of antibodies. Pure antibodies were obtained as follows: an HN immunosorbent was prepared by coupling 4 mg HN to 3 ml wet AH Sepharose gel with glutaraldehyde (Cambiaso et aL, 1975). This immunosorbent was used to separate pure anti-HN an-

FIG. 1. Acrylamide gel electrophoresis of HN in the absence of SDS-anode was to the bottom. Samples were 20 ~1. HN was 10 pg (left). The addition of HA (2 pg right) blocked HN migration. Migration duration was 60 min under an 8 V/cm electric field, at room temperature. Current was 8 mA.

tibodies from the absorbed antiserum which was used as a control in every immunohistological experiment. The antiserum was obtained in rabbits injected subcutaneously once a week with 100 pg of HN in complete Freund’s adjuvant until a good titer was obtained when assayed by electrosyneresis (Delpech, 1982). Before use, the HN immunosorbent was washed with 20 ml glycine HCl buffer at pH 2.2, followed by 200 ml PBS. Antiserum (10 ml) was mixed with the immunosorbent (3 ml) for 24 hr at room temperature and gently agitated on a roller. The antiserum after absorption was recovered by squeezing the gel in a syringe. The gel was washed with 0.5 liter PBS. The antibodies were eluted by the glycine-HCl (pH 2.2) buffer and the solution immediately adjusted to pH 7 by sodium hydroxide. The gel was washed with PBS and the procedure repeated once. The gel was kept at 4°C in PBS containing 0.02% sodium azide. Purified antibodies were precipitated by ammonium sulphate at 50% saturation, dissolved in PBS, and dialyzed against PBS. Their final volume was adjusted to a 3 mg/ml protein concentration according to their absorbance at 230 nm. Their specificity was assessed by electrosyneresis and by the Ouchterlony technique.

DELPECH

AND DELPECH

Hyaluronedin

Tissue preparation. Wistar rat embryos at various intervals of gestation (Day 4 to birth) were used. The day of positive vaginal smear was designated Day 0 of pregnancy. All tissue specimens were treated for 3 days with a cold (+4”C) mixture of absolute ethanol and glacial acetic acid (98:2, v/v), followed by cold absolute ethanol for 15 hr. The fixed specimens were embedded in paraffin at 56°C and cut with a microtome at 5pm. Embryos, before implantation and up to the ninth day, were studied on serial sections of uterus. From Day 9, each uterine swelling corresponding to an embryo was removed and serially sectioned. From Day 15, embryos were removed from the placenta and serially sectioned. ImmunoJEucwescence. The immunofluorescence study was performed as previously described (Delpech et aL, 1982a). Paraffin was removed by xylene (3X 20-set washes) followed by absolute ethanol (3X 40 set) and rinsed in PBS, pH 7.2 (3X 10 min). Sections were then incubated for 40 min in a humid chamber with purified antibodies at a concentration of 50 pg/ml and given 5~ 5-min rinses in phosphate buffer before incubation for 30 min with FITC-labeled sheep anti-rabbit immunoglobulins at a dilution of l/50 (Institut Pasteur, Paris). After three rinses of 5 min each, the sections were stained for 5

FIG. Control

2. Morula stained by indirect with anti-HN absorbed with

immunofluorescence; purified HN provides

in

the Rat Embryo

393

min in Evans blue at l/10,000, rinsed in phosphate buffer, and mounted in buffered glycerin. In controls, anti-HN antibodies were replaced either by preimmune rabbit serum or by antiserum absorbed with purified antigen as described above. Enzyme digestion. Prior to immunofluorescent staining, sections were incubated for 3 hr (2 X 1.5 hr) at 37°C in one of the following solutions: (a) 0.1 M phosphate buffer, pH 5.3, containing NaClO.15 M and 10 TRU/ml streptomyces hyaluronidase (Calbiochem); (b) PBS with 5 U/ml chondroitinase ABC (Sigma); (c) buffers without enzyme as controls. Following incubation, the slides were rinsed in phosphate buffer, pH 7.2, and stained by immunofluorescence. Micro.scopy. We used a Leitz Orthoplan fluorescent microscope with a Ploem-type vertical illuminator. This microscope was equipped with an HBO 200 W mercury vapor lamp combined with FITC-specific filters including excitation filters BG 38, BG 12, 2KP 490, and barrier filters S 510. Photographs were taken on Agfapan 400 ASA professional film and Ektachrome 160 ASA. RESULTS

Before implantation, study of the morula

(a) For HN, both the embryo the background level. X640.

an indirect immunofluorescence showed no trace of HN (Fig. 2).

and the zona

pellucida

(Z.P.)

stain

negatively.

(b)

394

DEVELOPMENTAL BIOLOGY

Aft4 ?r implantation, HN was not detectable at the blastocy 1st stage nor at Day 9, when the tridermic stage star ‘ts . In the uterine wall, HN was seen as an inter-

VOLUME 101, 1934

cellular network in the chorion of mucosa but remain ed at a distance from the implantation zone; the zone of decidual transformation had no HN. HN fluorescen ce

FIG. 3. Parasagittal section of lo-day embryo. (a) Immunofluoresosnce; (b) histological staining of an adjacent section. Thin line of strong HN fluorescence is seen around somites (S) and loose hyaluronectin network in the heart (H). Reichert’s membrane (R) stains negatively. X35. FIG. 4. Transverse section of lo-day embryo. (a) Immunofluorescence; (b) histological staining of an adjacent section. Hyaluronectin fluorescence is visible around somites and neural tube. X135.

DELPECH AND DELPECH

Hyaluronectin

appeared in the embryo at Day 10. It was localized in the mesoderm, at the neural tube periphery, and around the somites, bordering them by a fine dense fluorescent line where it appeared to be closely associated with the basement membrane. The structures thus delineated were themselves negative for HN (Figs. 3 and 4). At the end of Day 10, HN was present in the mesoderm between the primitive intestine and the coelomic cavity and the mesenchymal tissue of the primitive endocardium, but it always remained extracellular. In the endometrium the appearance was identical to that on Day 9; in addition, no HN was detectable at the Reichert membrane or the giant cells of the trophoblast. On Days 11 and 12, HN became more dense around the notochord and extended to the mesenteric root. The fluorescent stain penetrated between the ectoblast and the dermomyotome (Fig. 5) and bordered the nephrotome. Trace amounts of protein were seen between the cells of the neural tube in the cephalic region. At Day 13 fluorescence intensified, particularly around the well-formed notochord, encircled the spinal ganglia (Fig. 6) and appeared in the undifferentiated mesenchymal tissue of the primitive intestine (Fig. 10). In all these areas the protein was present between the cells and no specific intracellular staining was detected. The liver anlage was unlabeled. Other changes appeared at Day 15. HN was still visible in the mesenchyme but was beginning to disappear in the differentiating zones. For example, around the no-

in the Rat Embryo

395

tochord, where chondrogenic differentiation of the vertebra begins, HN distribution was less widespread and appeared as small isolated patches between cells undergoing cartilage differentiation (Fig. 7). In addition, in the nervous tissue accumulation of HN was visible in the telencephalon at the corpus striatum (Fig. 13); no intracellular localization was detectable. At Day 16, with perichordal differentiation more advanced, HN deposits became less dense (Fig. 8). The same was seen at the trachea where cartilage differentiation was associated with progressive reduction of the HN present in the preexisting mesenchyme. A similar phenomenon was observed in the mesenchyme of the primitive intestine; when its muscle coat differentiated, HN disappeared and remained only in the submucosal connective tissue (Fig. 11). At this stage bronchial arborizations and kidney ureteral ramifications were surrounded by HN-positive loose connective tissue. The dermis was strongly labeled with local densities suggesting a basal membrane labeling. The liver remained completely negative. In nervous tissue traces of HN appeared in the cerebral cortex at the marginal layer and the external part of the intermediate layer. In the cerebellum traces were visible in the marginal layer and the external granular layer. On Day 17, HN disappeared in the precartilaginous vertebral body (Fig. 9). Preincubation of the sections with streptomyces hyaluronidase or with chondroitinase ABC did not alter these results, showing that this was

FIG. 5. Transverse section of 11-day embryo. (a) Immunofluorescence; (b) histological staining of an adjacent section. The zone between ectoderm (E) and dermomyotome (D) as well as around the notochord (N) is seen as a brightly staining band. Faint reticular fluorescence pattern is observed in the undifferentiated mesenchyme (M) and around nephrotome (No). X85.

396

DEVELOPMENTAL

BIOLOGY

VOLUME

101.1984

FIGS. 6 -9. Evolution of hyaluronectin distribution in perinotochordal region in the course of cartilage histogenesis. FIG. 6. A 13-day embryo. Bright and diffuse hyaluronectin fluorescence is seen around notochord (N) and spinal ganglia (S.G.). X85. FIG. 7. A 15-day embryo. Hyaluronectin fluorescence is no longer diffuse but displays patchy staining. X85. FIG. 8. A 16-day embryo. (a) Immunofluorescence; (b) histological staining of an adjacent section. Hyaluronectin fluorescence disapp ears gradually around notochord (N). X85.

due to In the enzyma around

Labsence of antigen and not to a masking effeet. s;ame manner the liver remained negative after tic preincubation. Dense staining was seen cartilage (vertebral body, tracheal rings) and

between bundles of smooth and of striated must :les. Weakly positive staining also appeared between the elastic fibers of the aorta. Distribution in the brain was superimposed on the various layers which were dit Ter-

DELPECH

AND

DELPECH

Hyalurmectin

in

the Rat

Embryo

397

DISCUSSION

The essential

findings

of this work are:

-the absence of HN before implantation of the embryo; -the appearance of HN at Day 10 of gestation; -the accumulation of HN between undifferentiated mesodermic cells and its disappearance accompanying mesenchymal differentiation into muscle and cartilage; -appearance of HN in nervous tissue at Day 12 of gestation.

FIG. 9. A 17-day embryo. No detectable is seen in the vertebral cartilage. X85.

hyaluronectin

fluorescence

entiating, with predominance in the marginal zone, absence in the highly cellular cortical layer, presence in the fibrous external part of the intermediate zone, and absence in the ependymal zone (Fig 14). At Day 18, at the intestinal loops the distinction between the submucosa (HN positive) and the muscular coat (HN negative) was very clear (Fig. 12). Labeling was found in mesenchyme situated between epithelial structures of digestive gland anlages and between thymic lobules. The staining remained intense around the tracheal rings and in the dermis. At Day 19, in the cerebral cortex, HN increased with the thickening of the intermediate zone (Fig. 15). Areas where cartilage differentiation was not complete were still weakly positive. The connective tissue between muscular bundles and epithelial structures of digestive glands were also labeled, but the thymus was completely negative. In the 1-hr-old newborn rat as the cerebral cortex thickened, HN staining moved inward but always remained at a distance from the ependymal zone. It was present in the thalamus but seemed to diminish in the corpus striatum. In the cerebellum it was localized in the central white matter and encircled the nuclei of the cerebellum (Fig. 16). At this stage the spinal cord showed very slight reactivity with the antiserum in the vicinity of the dorsal fissure. In general, results were identical to those recorded on Day 19 with the exception that in the kidney, traces of HN were visible only between the tubes of the papilla.

These results are in support of the suggested HNHA linkage in embryonic tissues. We have shown that HN is included in the extracellular mesodermal matrix at the predifferentiation stage. It is abundant in the undifferentiated mesoderm, progressively disappears during mesodermal differentiation and is no longer found in totally differentiated tissue such as vertebral cartilage and smooth muscle tissue. It could be suggested that HN determinants are masked during differentiation by high-molecular-weight glycosaminoglycans or proteoglycans but the failure to obtain staining after hyaluronidase and chondroitinase treatment indicated that this unlikely to be the case or that at least if masking does occur it does not involve hyaluronic acid or chondroitin sulphate. Similar results were recorded by Duband and Thierry (1982) for fibronectin in the chick embryo. HN appeared at a later stage in the embryonic nervous system. Its distribution was extracellular, in areas where both cell migration and proliferation occur (Sidman and Rakic, 19’73), i.e., in the telencephalon at the corpus striatum, then in the thalamus and later in the cerebral cortex and cerebellum. Conversely in the spinal cord where cellular migration is less marked, HN was hardly detectable at birth. These results, when compared with those obtained in the adult rat nervous system where HN was found only at precise points such as the nodes of Ranvier and the neuronal microenvironment (Delpech et al, 19’79a, 1982b), indicate that considerable modification will still occur after birth when the nervous system reaches complete maturation. A similar conclusion may be drawn about the mesenchymal HN of other organs, which is abundant in embryonic mesenchyme and is present in the adult in only a small number of locations such as the kidney papilla, endometrium, and hair follicles. Observations made in the human indicated that HN is also associated with normal structures in the process of continual renewal (hair follicle, uterine mucosa) and with tumors. In these structures the protein is present exclusively in connective tissue and never in the epithelial component. These observations are relevant to the results presented here which show a close association

DEVELOPMENTAL

BIOLOGY

VOLUME

101, 1934

of hyaluronectin in the wall of the small intestine during smooth muscle histogenesis. FIGS 3. 10-12. Distribution FIG. 10. A 13-day embryo. Epithelial lining stains negatively; the layer of undifferentiated mesenchyme stains positively. fluoret scence is seen only in extracellular location. X340. FIG. 11. A 16-day embryo. Hyaluronectin disappears in the peripheral layer where the tunica muscularis differentiates (from adjace nt connective tissue of the submucosa is positive and the epithelial lining is still negative for hyaluronectin. X135. FIG. 12. An l&day embryo. Between mucosa and tunica muscularis, the submucosa is brightly fluorescent. X105.

Hya me8 lenchynre);

DELPECH

AND

DELPECH

Hvaluronectin

in the Rat Embryo

399

FIG. 13. Brain of 15-day embryo. Corpus striatum stains positively; cortical areas stain negatively. ~85. 14. Brain of 1%day embryo. Hyaluronectin fluorescence is seen in the cortex at the sites of the marginal zone (M) and of the external part of the intermediate layer (I) and is visible in the thalamus (T). X85. FIG. 15. Brain of Is-day embryo. In the cortex, hyaluronectin spreads within the intermediate layer (I) which is developing. The long arrow indicates the intermediate layer which extends beyond the represented area. X215. FIG. 16. Cerebellum of 1-hr newborn. Hyaluronectin fluorescence of the medulla (white) disappears at the sites of cerebellum nuclei (Nu). X85. FIG.

of HN with embryonic mesenchyme. They suggest that connective tissue may retain some embryonic features in certain physiological or pathological situations char-

acterized by the presence of proliferating epithelial cells. There is a striking similarity in this regard between these situations and embryonic development, indicating

400

DEVELOPMENTAL BIOLOGY

that a similar interaction between epithelial cells (either normal, embryonic, or cancerous) and mesenchymal cells could occur in all these cases. However, there is no evidence to resolve the question whether HN is one of the factors favoring proliferation or whether it is induced secondarily by proliferation of epithelial cells. Our results concerning HN may be compared on the one hand with those obtained with fibronectin (FN) and on the other hand with what is known about HA. If HN is compared to FN, another glycoprotein studied in the chicken embryo (Linder et a& 1975; Newgreen and Thierry, 1980) and in the young mouse embryo (Wartiovaara et a& 1979), it is seen that they are both extracellular in all mesodermal areas and are often concentrated at the periphery of tissues undergoing organization. However, the appearance of FN is earlier, with traces detected at the blastocyst stage in the endoderm precursor cells. In addition, unlike HN, FN is present in the basement membrane of the uterine mucosa, in the giant cells of the trophoblast in the Reichert membrane and in the embryonic liver, which were always negative for HN. Thus, there are some histological differences between the two proteins, as we showed earlier in adult human skin (Delpech et al, 1982b). With regard to HA, to which HN can be linked, its important role in embryology has been studied by many authors (Toole, 1973; Pratt et a& 1975; Solursh and Morriss, 1977) who have shown that HA appears earlier than we have found to be the case with HN. The later appearance of HN raises the possibility that it may modify the biochemical activity of HA. We are currently studying these two components during development and in adult tissues. It appears to us essential to determine whether HN is synthesized by the embryo or is taken up by the embryo’s HA from maternal sources. Although HN was seen in newborn rat oligodendrocytes (Asou et CL&1983) it has never been detected inside cells in the embryonic tissue. Like HA with which it is associated in the intercellular spaces, it seems possible that it may have an inhibitory effect on differentiation and a facilitatory effect on cellular migration (Brunngraber, 1979; Pratt et a& 1975; Toole, 1973, 1981). We thank Mrs. Abdelouhab and Mrs. Maingonnat for skillful technical assistance. We gratefully acknowledge Dr. Derek McCormick for revising the English text of the manuscript. This work was supported by the University of Rouen. REFERENCES Asou, H., BRUNNGRABER,E. G., and DELPECH, B. (1983). Localization of hyaluronectin in oligodendrogial cells. J. A&roc/r.em 46, 589591. BITTER, T., and MUIR, H. M. (1962). A modified uranic acid carbazole reaction. And Biock 4.330-334. BRUNNGRABER,E. G. (1979). Glycosaminoglycans. In “Neurochemistry

VOLUME

101,1%‘i

of Aminosugars” (E. G. Brunngraber, ed.), pp. 128-177. Thomas, Springfield, Ill. CAMBIASO, C. L., GOFFINET, A., VAERMAN, J. P., and HEREMANS, J. F. (1975). Glutaraldehyde activated aminohexyl derivative of Sepharose 4B as a new versatile immunoabsorbent. Immunochemistry 12,273278. CHEVRIER, A., DELPECH, B., GIRARD, N., NOUEL, J. P. (1979). Isolation and characterization of mesenchyme associated antigen from dimethylbenzantracene induced rat fibrosarcoma. Biomeatiw 30,219222. DELPECH, A., DELPECH, B., GIRARD, N., and VIDARD, M. N. (1978). Localisation immunohistologique de trois antigenes associes au tissu nerveux (GFA, ANSa, brain glycoprotein). Bid Cell 32,207-214. DELPECH, A., DELPECH, B., and GIRARD, N. (1979a) Association de la glycoproteine cerebrale AN& a la membrane du neurone et aux Qtranglements de Ranvier. CR Acad Sti Ser. D 283,1323-1326. DELPECH, B., DELPECH, A., GIRARD, N., CHAUZY, C., and LAUMONIER, R. (1979b). An antigen associated with mesenchyme in human tumors that cross-reacts with brain glycoprotein. Brit J. Cancer 40, 123133. DELPECH, B., and HALAVENT, C. (1981). Characterization and purification from human brain of a hyaluronic acid binding glycoprotein: Hyaluronectin. J. Neurochem 36, 855-859. DELPECH, B. (1982). Immunochemical characterization of the hyaluranic acid-hyaluronectin interaction. J. Neurochem. 38,978-984. DELPECH, A., DELPECH, B., GIRARD, N., BOUILLIE, M. C., and LAURET, P. (1982a). Hyaluronectin in normal human skin and in basal cell carcinoma. Brit. J. DevmaioL 106,561~568. DELPECH, A., GIRARD, N., and DELPECH, B. (1982b). Localization of hyaluronectin in the nervous system. Brain Rea 245,251-257. DUBAND, J. L., and THIERRY, P. (1982). Distribution of fibronectin in the early phase of avian cephalic neural crest cell migration. Deu. Bid

93,308-323.

GIRARD, N., CHAUZY, C., OLIVIER, A., and DELPECH, B. (1982). Characterization of hyaluronectin in human tumor heterografts in the nude mouse. ~~UX&ZV. Biol Med 3,325~334. LINDER, E., VAHERI, A., RUOSLATHI, E., and WARTIOVAARA, J. (1975). Distribution of fibroblast surface antigen in the developing chick embryo. .I. Exp. Meat 142,41-49. MORRISS, M., and SOLURSH, M. (1978). Regional differences in mesenchymal cell morphology and glycosaminoglycans in early neural fold stage rat embryos. J. Embyol Exp. Morphd 46,37-52. NEWGREEN, D., and THIERRY, J. P. (1980). Fibronectin in early avian embryos: Synthesis and distribution along the migration pathways of neural crest cells. Cell T&sue Ra 211.269-291. PRATT, R. M., LARSEN, M. A., and JOHNSTON,M. C. (1975). Migration of cranial neural crest cells in a cell-free hyaluronate-rich matrix. Dev.

Bid

44,298-305.

SIDMAN, R. L., and RAKIC, P. (1973). Neuronal migration, with special reference to developing human brain: A review. Brain Ran 62, l35. SOLURSH,M., and MORRISS, M. (1977). Glycosaminoglycans synthesis in rat embryos during the formation of the primary mesenchyme and neural folds. Dev. Bid 67,75-86. TENGBLAD, A. (1979). Affinity chromatography on immobilized hyaluronate and its application to the isolation of hyaluronate binding proteins from cartilage. Biochim. Biophys. Actu 578.281289. TOOLE, B. P. (1973). Hyaluronate and hyaluronidase in morphogenesis and differentiation. Amer. Zool 13,1061-1065. TOOLE, B. P. (1981). Glycosaminoglycans in morphogenesis. In “Cell Biology of extra cellular matrix” (E. D. Hay, ed.), pp. 259-294 Plenum, New York. WARTIOVAARA, J., LEIVO, I., and VAHERI, A. (1979). Expression of the cell surface associated glycoprotein, Fibronectin, in the early mouse embryo. Dev, Bid 69,247-257.