The distribution and spatial organization of the extracellular matrix encountered by mesencephalic neural crest cells.код для вставкиСкачать
THE ANATOMICAL RECORD 21157-68 (1985) The Distribution and Spatial Organization of the Extracellular Matrix Encountered by Mesencephalic Neural Crest Cells PHILIP R. BRAUER, DAVID L. BOLENDER, AND ROGER R. MARKWALD Department of Anatomy, Medical College of Wisconsin, Milwaukee, WI 53226 ABSTRACT Cephalic neural crest (NC) cells enter a cell-free space (CFS) that contains a n abundant extracellular matrix (ECM). Numerous in vitro investigations have shown that extracellular matrices can influence cellular activities including NC cell migration. However, little is known about the actual ECM composition of the CFS in vivo, how the components are distributed, or the nature of NC cell interactions with the CFS matrix. Using ultrastructural, autoradiographic, and histochemical techniques we analyzed the composition and spatial organization of the ECM found in the CFS and its interaction with mesencephalic NC cells. We have found that a specific distribution of glycoproteins and sulfated polyanions existed within the CFS prior to the translocation of NC cells and that this ECM was modified in areas occupied by NC. The interaction between the ECM components and the NC cells was not the same for all NC cells in the population. Subpopulations of the NC cell sheet became associated with ECM of the ectoderm (basal lamina) while other NC cells became associated with the ECM of the CFS. Trailing NC cells (NC cells that emerge after the initial appearance of NC cells) encountered a modified ECM due to extensive matrix modifications by the passage of the initial NC cell population. Neural crest (NC) cells, which arise from the apex of sulfate, human plasma fibronectin) [Newgreen, 1982; Erfusing neural folds, are the major mesenchymal source for ickson and Turley, 1983; Rovasio et al., 19831. Although the primordia of the head and neck [Johnston, 1966; Wes- these studies do provide information on NC cell behavior ton, 1970; Noden, 19751. Dispersion of the pluripotential in culture, these in vitro conditions fall short of providing cephalic NC cells is not random but rather temporally and an adequate model for the actual in vivo situation, where spatially ordered. The mechanisms controlling the or- multiple matrical associations may occur. For instance, dering appears to be under influence of the local environ- the association of NC cells with a specific component (e.g., ment, but how this occurs is not understood [Weston and fibronectin) may only reflect that NC cells prefer that component to glass or denatured collagen and not that the Butler, 1966; Noden, 1978a,b]. Cephalic NC cells emerge from the neural tube and en- particular ECM component is actually utilized for conter a narrowing cell-free space (CFS) bordered by ecto- tact in situ by NC cells. Also, nothing is known about the derm, neural tube, and mesoderm [Pratt et al., 19751.The endogenous nature of sulfated GAG in this matrical comCFS contains abundant extracellular matrix (ECM), partment (i.e., are they in a proteoglycan or a free GAG which has been circumstantially linked to morphoge- form), if glycoproteins other than fibronectin occur in the netic activities in many developing systems [Bissell et al., CFS, or the actual conformational state of CFS matrix 19821. The only components of this particular ECM iden- moieties (i.e., posttranslational intermolecular associatified to date are hyaluronate, sulfated glycosaminogly- tions). Indeed, the spatial ordering and distribution of cans (GAG), and fibronectin [Pratt et al., 1975; Bolender these components within the ECM may be as important et al., 1980; Mayer et al., 1981; Newgreen and Thiery, to cell behavior in vivo, a s is the actual composition [Old19801.The basal laminae of the ectoderm and neural tube berg and Ruoslahti, 19821. We have initiated a n investigation into the role of enare contiguous with the CFS. Whether or not the basal lamina (BL) components extend into the CFS remains to dogenous ECM in the translocation of NC cells in avian embryos. In this study, we examined morphologically and be established. Previous studies have examined NC cell attachment histocheniically the cortipositioti, orderirig, arid distriand migration on substrata coated with ECM compo- bution of ECM found in the CFS and any ECM modifinents isolated from various animals and body regions cations that accompanied the translocation of NC. Auto(e.g., rat tail type I collagen, human umbilical hyaluro- radiographic, histochemical, and ultrastructural technate, bovine nasal proteoglycans, shark chondroitin niques were used with fixatives that optimally preserve ~ Received October 19, 1983; accepted J u l y 25, 1984. 0 1985 ALAN R. LISS. INC. 58 P.R. BRAUER, D.L. BOLENDER, AND R.R. MARKWALD TABLE 1. Matrix distribution, staining, and labeling in the cell-free space’ 3-5 nm Filaments 30-40 nm gr an u 1es Prior to NC Zone 1 BL AM Zone 2 ECM of the mesoderm - + Electron-dense material + + +2 +3 +2 - +4 - + + + Areas occupied by NC Zone 1 BL AM Zone 2 ECM of the mesoderm + CI-stain ~ - + + 3H-fucose 3H-GSA 35S-S04 + + + + + + + + + + + + + - - - - - + + + + + + + + + + ’+ means present and - means absent or displaced. Abbreviations: BL, basal lamina; AM, basal lamina-associated matrix; ECM, extracellular matrix. ‘Less abundant than Zone 1. ‘3-5 nm filaments found in less abundance ‘BL was present but thinned. the ECM with a minimal sacrifice of cell morphology. We found that a specific distribution of ECM components existedwithintheCFSpriortothe appearanceofNCcellsand that this distribution was subsequently modified in areas occupied by NC cells. Recently, the controversy as to whether NC arrive at their final destination through mechanisms of active vs. passive (differential growth) means has arisen [Gasser, 1979; Nichols, 1983; Noden, 19841. This paper does not deal directly with migratory mechanisms (active vs. passive) but is concerned with potential mechanisms that might orient forming NC to selectively develop associations with adjacent tissues (surface ectoderm or “lateral pathway,” neural tube or “ventral pathway”). Our working hypothesis is that the ECM provides the directional cues necessary for the development of such associations. METHODS Tissue Preparation Fertilized white leghorn chicken eggs obtained from the Texas A&M Poultry Science Department were incubated in a forced air incubator a t 38°C in 60% humidity. Embryos corresponding to stages 8 through 10 (26-38 hr, Hamburger and Hamilton [ 19511 were removed from the surface of the yolk and placed in warm (37°C) Tyrode’s balanced salt solution and the extraembryonic membranes were removed. Embryos were fixed at room temperature for 2 hr in 3% glutaraldehyde (in 0.1 M sodium cacodylate buffer, pH 7.3) with or without the addition of either 0.5% cetylpyridinium chloride (CPC) or 2% tannic acid (both from Sigma, St. Louis, MO). Embryos for transmission and scanning electron microscopy studies were postfixed in 1% osmium in 0.1 M sodium cacodylate buffer for 2 hr. For transmission electron microscopy (TEM), specimens were embedded in Polybed 812 after dehydration in graded alcohols, sectioned on a Reichert ultramicrotome, stained with lead citrate and uranyl acetate, and examined on a Zeiss EM 10A. Scanning electron microscopy (SEM) samples, after dehydration, were critical-point dried in COz and then dissected using tungsten needles or Scotch tape according to Tosney [ 19781. Specimens were then sputter-coated with gold-palladium alloy and examined on a Hitachi HS-500 scanning electron microscope at 20 KV. Autoradiography and Histochemistry Twenty microcuries of 6-3H-glucosamine, 50 pCi of 35S-sodium sulfate, or 20 pCi of 6-3H-L-fucose(19 Ci/ mmol, 975 mCi/mmol, 60 Cilmmol, respectively; all obtained from New England Nuclear, Boston, MA) were applied in ovo onto the vitelline membrane of stage 8 or 9 embryos in 0.1 ml Tyrode’s solution and incubated for 5 hr, allowing a sufficient amount of time for the embryos to reach stage 9 or 10. Embryos were rinsed, fixed, and embedded a s described and processed for light microscopic autoradiography. Sections, 1.5 pm thick, were dipped in NTBS nuclear track emulsion (Kodak) diluted 1:1, and incubated 2-3 weeks prior to development of the autoradiographs. Embryos were collected at the appropriate stages, fixed, and stained with colloidal iron (CI) (for light microscopy) or dialyzed iron (for TEM) a t pH 1.7 as described by Spicer et al.  and Markwald et al. . At pH 1.7, the sulfate esters are ionized and the carboxy groups suppressed, allowing the CI and dialyzed iron to bind and thereby locate, histologically, areas containing sulfate esters. Following the staining, the embryos were embedded in Polybed 812, sectioned, and the CI staining was examined with Zeiss Normarski-enhanced optics and the dialyzed iron staining with a Zeiss EM 10A. Concanavalin A (Con A) lectin was used as a probe to localize a-D-mannose-and a-D-glucose-rich moieties of the ECM. Embryos were fixed in 3% glutaraldehyde with 0.5% CPC and rinsed in 0.1 M sodium cacodylate buffer. Embryos were placed in 5% agar, tissue chopped at 50 pm using a TC-2 Sorvall tissue chopper, and sections were rinsed in phosphate buffer, pH 7.4. Sections were then placed in phosphate buffer containing 200 pgl EXTRACELLULAR MATRIX AND CEPHALIC NEURAL CREST Fig 1 . TEM of a stage 8 embryo fixed in glutaraldehyde-CPC. The contained the CFS can be divided into two zones (arrows).Zone 1 (Z1) ectodermal BL and was made of electron-dense material arranged in a looping configuration (arrowheads).Subjacent were anastomosing electron-dense clusters of matrical material. Zone 2 (22) contained electron-dense material, 3-5 nm filaments, and 30-40 nm granules. ~5,200.E, ectoderm; M, mesoderm. Fig. 2. SEM of ECM found in the CFS when fixed with glutaraldehyde-CPC. Strands of matrix spanned across the CFS (asterisk). Mesodermal cell processes often extended toward Zone 2. ~2,000.Inset, matrix observed when tannic acid replaced CPC. The BL presented a lamina lucida, a lamina densa (arrowhead), and associated interstitial bodies (asterisk). Zone 2 matrix of the CFS was virtually absent. ~22,500. Fig. 3. Higher magnification of matrical components of Zone 2 fixed with glutaraldehyde-CPC. Electron-dense material (arrows) was en- 59 meshed in a network of 3-5 nm filaments and 30-40 nm granules (arrowheads). x 46,900. Fig. 4. Dialyzed iron stained the 30-40 nm granules indicating the presence of ionized sulfate esters (arrowheads). Filaments show only background staining. ~63,000. Fig. 5. TEM demonstrating the relationship of mesodermal cell processes with the matrix of Zone 2; 3 0 4 0 nm granules were particularly numerous in areas adjacent to mesoderm cells (arrowheads). ~32,600. Fig. 6. SEM of the matrix found between mesoderm cells. Granules studded the length of intercellular matrical strands and fibrils (arrowhead). ~7,620. 60 P.R. BRAUER, D.L. BOLENDER, AND R.R. MARKWALD EXTRACELLULAR MATRIX AND CEPHALIC NEURAL CREST 61 Fig. 9A,f3. TEM of the matrix-NC interface. Direction of NC movement from the neural tube is indicated by the large black arrows. Electron-dense material more heavily coated the “leading surfaces” (hollow arrows) compared to the trailing surfaces of pioneering NC processes (arrowheads). x 12,850 (A) and ~9,900(B). ml Con A (Sigma, grade IV). Control sections were treated with 200 pg/ml Con A and 0.1M a-methyl-Dmannose (Sigma) in phosphate buffer or with buffer alone. All sections were allowed to incubate for 1 h r at 37°C on a shaker water bath. After incubation, the sections were rinsed with buffer and agitated (three changes in 1.5 hr). Localization of the bound Con A lectin was sought by adding glycopeptidyl-ferritin (150 pg/ml; Polyscience, Inc., Warrington, PA Cat. No. 8742; a glycosylated ferritin that binds only to Con A) to all specimens at 37°C for 1hr with agitation. Tissues were then rinsed in buffer (three changes in 1.5 hr), postfixed in 1% osmium for 2 hr, and processed for TEM as described above. Specimens were thin sectioned, stained with uranyl acetate only, and examined on a Zeiss EM 10A. Figs. 7,8.Interaction of pioneering NC cells with the ECM of the CFS. Leading edges of the NC cells were moving in the direction of the large arrows relative to the neural tube (N). Inset, light micrograph of a forming NC cell population (asterisk) comparable to that shown above and below. Matrix of Zone 1 and Zone 2 has been altered or displaced a t the NC front (arrowhead). x 195. 7, TEM of a leading cell process of pioneering NC. The leading NC cell processes extended into the electron-dense material above Zone 2 (bracket) and were coated with electron-dense material (small arrow). Vesicles in these processes contained material similar to the ECM (arrowheads). BL-associated matrix of the CFS was diminished behind the leading cell processes (asterisk). ~22,000.8, SEM of the matrix-NC interface. Note that the ECM of the CFS (between the arrowheads) underlying the ectoderm abruptly diminished at the NC front (between the arrows). ~3,200. RESULTS CFS Prior to the Appearance of NC Prior to the appearance of mesencephalic NC cells (stage 8 + to 9-), the CFS is bounded by the surface ectoderm and neural tube and is partitioned by the mesoderm cells into two channels: one between the surface ectoderm and underlying mesoderm and the other lateral to the neural folds [Pratt et al., 1975; Bolender et al., 19801. When glutaraldehyde with or without tannic acid was employed as the fixative the channel below the surface ectoderm contained both a BL, which has traditionally been described as representing a lamina lucida, lamina densa, and interstitial bodies, and a CFS containing fibrils (see Fig. 2, inset) [Bolender et al., 1981; Tosney, 19821. Embryos fixed in glutaraldehyde with the addition of 0.5% CPC preserved much more ECM (i.e., CPC-dependent matrix; Fig. 2). The results from this study are summarized in Table 1. The CPC-dependent matrix was arbitrarily arranged into two zones. The first zone (i.e., Zone 1)contained a n ectodermal BL comprised of sheets of electron-dense material that merged into a single continuum and periodically exhibited a looping configuration (Figs. 1, 19). Subjacent to the ectodermal BL was additional electrondense material arranged in interconnecting clusters (referred to in this study as BL-associated matrix). Zone 2, which was much more extensive than the overlying Zone 1, was characterized by electron-dense material arranged into pleomorphic strands of variable length, some extending the entire expanse of the CFS toward the mesoderm. This electron-dense material became increasingly enmeshed within a network of 3-5 nm filaments and 30-40 nm granules in areas approaching the 62 P.R. BRAUER, D.L. BOLENDER, AND R.R. MARKWALD mesoderm (Fig. 3). Mesodermal cell processes were observed extending into Zone 2 (Figs. 1,2,19). The 30-40 nm granules were particularly numerous in these areas (Fig. 5,6). At these stages, Zone 1did not stain with CI. However, in Zone 2 positive CI staining of sulfate esters was found in the same area where the 30-40 nm granules had been primarily localized (Figs. 12A,B). At the ultrastructural level, the 30-40 nm granules in Zone 2 specifically stained with dialyzed iron, demonstrating that they contained sulfate esters (Fig. 4). Autoradiographic localization of incorporated sulfated label was found in both Zone 1 and Zone 2 (Fig. 17A), both prior to and after the appearance of NC. Mesodermal intercellular spaces were also heavily labeled with sulfate as were cells of the neural tube, their associated BL, and the extracellular space surrounding the notochord. Localization of 3H-glucosamine label was observed in Zone 1, Zone 2, and intercellular spaces of the mesoderm population (Fig. 15A). Labeling was especially heavy lateral to the fusing neural folds. 3;3-fucose (incorporated into glycoproteins) [Coffey et al., 19641 labeling coincided with the distribution of the electron-dense material seen ultrastructurally in both zones (Fig. 16A). Incorporation was heaviest in Zone 1 with a decreasing concentration of silver grains found throughout Zone 2 as the strands spanned the CFS toward the mesoderm. Con A binding sites were primarily localized in Zone 1 on the external surfaces of the electron-dense material (Fig. 18). The addition of a competing saccharride or omission of the lectin, consistently diminished the labeling. Initial Appearance of NC Chick NC cells emerge from the apex of the fusing mesencephalic neural folds at stage 9+ to 10- as a sheet of cells (Inset Figs. 7 3 ) and enter the CFS. The pioneering NC cells (i.e., those at the leading edge of the NC sheet) were found in Zone 1 dorsal to the interface between the CI-negative matrix of Zone 1 and the CIpositive matrix of Zone 2 (Figs. 13AJ3B). Concomitant with the appearance of NC cells, the majority of the CPC-dependent matrix found in Zone 1 and Zone 2 was abruptly diminished or rearranged at the leading edge of the NC cell population (Figs. 7,8). In areas occupied by the NC population, the BL of Zone 1 appeared greatly thinned (Fig. 111, whereas much of the BL-associated matrix became affiliated with pioneering NC cell surfaces resulting in its depletion from the extracellular space (Figs. 7,9). Cell processes of pioneering NC often exhibited cell surface invaginations with associated microfilaments, microtubules, and vesicles containing material resembling the BL electron-dense material (Figs. 7,101. Autoradiographic studies also indicated that Zone 1 matrix was modified or displaced by the NC cells. Both 3H-fucose-and 3H-glucosamine-labeled matrix accumulated on NC cell pericellular domains (Figs. 15B,16B). This rcduction or modification was particularly striking with 3H-glucosamine-labeled tissue. Intracellular labeling was not heavy for either precursor. Zone 2 was less modified except for that portion in direct contact with NC cells. 35S-sulfate, unlike the other labeled precur- Fig. 10. TEM of an NC cell. Cell surface invaginations contained matrical material (arrow) and associated intracellular microtubules and microfilaments (arrowheads). X43,900. Fig. 11. TEM of the BL and associated matrix in areas occupied by NC cells. BL and associated loops (arrowheads) were greatly thinned when compared to those before NC interaction (cf. Fig. 7). Also, the electron-dense material of BL-associated matrix is almost completely absent in extracellular regions occupied by the NC cell population (asterisk). x 10,600. sors, was not extensively associated with NC cell surfaces or diminished from the extracellular space (Fig. 17B). In areas occupied by NC cells, the BL of Zone 1, which had not previously stained with CI, did so after encountering the NC cells (Figs. 14A,B). This newly acquired positive CI staining of the BL was found within two or three cell lengths behind the leading edge of pioneering EXTRACELLULAR MATRIX AND CEPHALIC NEURAL CREST Figs. 12-14. CI staining a t p€I 1.7 prior to and after the initial appearance of NC cells. 12(A,B), prior to the appearance of NC, CI staining is localized on the mesodermal cell surfaces (arrows), the ECM of the mesoderm, and Zone 2 (asterisk). Zone 1 (between the arrowheads) did not stain with CI. Both x 3,600. 13(A,B), arrow indicates direction of NC movement from the neural tube. Note that cells are 63 situated between the CI-negative matrix of Zone 1 and CI-positive matrix of Zone 2. CI staining was also found along the ventral surface of the NC sheet (arrowheads). ~ 3 , 6 0 0(A) and ~ 2 , 0 5 003).14(A,B), modifications in CI staining of the matrix in areas occupied by NC. Note the BL of Zone 1, which prior to NC interaction did not stain with CI (cf. Fig. 13A,B),is now reactive (arrowheads).Both ~ 3 , 6 0 0 . 64 P.R. BRAUER, D.L. BOLENDER, AND R.R. MARKWALD Figs. 15-17. Autoradiographs prior to and after the appearance of appearance of NC cells is greatest in Zone 1 (arrowhead) with fewer . the outline of fucose NC cells, when incubated 5 hr in ovo with 3H-glucosamine, .%L- grains localized in Zone 2 (arrows). ~ 3 2 0 16(B), . fucose, or 35S-sodium sulfate precursors, respectively. 15(A), glucosa- labeling found in Zone 1 is disrupted by NC (arrowheads). ~ 6 5 017(A), mine label is primarily localized in the ECM of both Zone 1 and Zone sulfate incorporation occurred in both Zone 1 and Zone 2 in areas not 2 prior to the appearance of NC cells (arrowheads). ~ 4 3 0 15(B), . glucos- yet occupied by NC cells. Mesoderm cells and their intercellular spaces amine labeling of Zone 1 and Zone 2 in areas occupied by NC was were also labeled as well as the neural tube. X330. 17(B), during the diminished when compared to the matrix not yet occupied by NC translocation of NC cells, sulfate label did not appear to be altered by (arrowhead). Little label was found in the extracellular space among the NC. x320. NC cell populations. x300. 16(A), incorporation of fucose prior to the EXTRACELLULAR MATRIX AND CEPHALIC NEURAL CREST 65 Fig. 18A,B. Localization of Con A binding. A, using glutaraldehydeCPC fixation, Con A binding was found to preferentially bind to the electron-dense material of the BL (arrowheads) and of the associated matrix (arrow). ~61,000.B, control. Competive saccharide was added along with Con A lectin during the first step of indirect labeling (see Methods). Note the reduction in labeling of electron dense material in Zone 1. ~61,000 NC cells. Matrix located ventral to the NC sheet that persisted after encountering NC cells continued to stain with CI and incorporate labeled precursors (sulfate, fucose, and glucosamine). rolyticus hyaluronidase digestion [Bolender et al., 1981J, a n enzyme specific for hyaluronate [Yamada, 19731. Purified hyaluronate, when fixed in a manner similar to the one used in this study and processed for TEM, takes the form of 3-5 nm filaments [Markwald et al., 19791. Thus, these 3-5 nm filaments identified in Zone 2 may constitute an ultrastructural representation of hyaluronate since hyaluronate has been previously identified in the CFS autoradiographically [Pratt et al., 19751 and histochemically [Bolender et al., 19801. Another GAG previously identified in the CFS is chondroitin sulfate [Bolender et al., 19801. In this study, chondroitin sulfate appeared to be localized to 30-40 nm granules identified by their sensitivity to testicular hyaluronidase [Bolender et al., 19811 and capacity to bind dialyzed iron a t pH 1.7. In the developing heart [Markwald et al., 19781 and blastula [Solursh and Katow, 19821 such granules were identified a s proteoglycans. Indeed, chrondroitin sulfate has generally been reported, whenever studied, to be in the form of a proteoglycan [Hascall and Sajdera, 1970; Norling et al., 1978; Ehrlich, 1981; Vogel and Petersen, 19811. Whether or not the 3040 nm granules reported in this study con- DISCUSSION In this study, CPC, a quaternary salt of pyridine and cetyl chloride, was used to retain maximally the ECM in a consistent, reproducible manner [Kvist and Finnegan, 19701. Preliminary experiments using a n alternative form of fixation (i.e., cryopreservation) [Kitten et al., 19811 indicated a close correlation between CPC fixation and nonchemically preserved matrix. Both of these methods retained more material in the CFS than is seen with other conventional fixation procedures (e.g., aldehyde fixes with or without tannic acids) [Singley and Solursh, 19801. Figure 19 represents a summary of the ECM distribution found prior to and after the initial appearance of NC using CPC. The present results indicate zonal differences in the distribution of the matrical components found in the mesencephalon. The 3-5 nm filaments of the CFS were previously shown to be sensitive to Streptomyces hyalu- 66 P.R. BRAUER, D.L. BOLENDER, AND R.R. MARKWALD I Fig. 19. Schematic representation of the distribution of matrical components found in the CFS and their relationship to NC. Zone 1, which did not stain with CI, contained the BL and associated matrix (AM) and consisted of electron-dense material (represented by the black clusters). Zone 2 stained with CI and consisted of electron-dense material enmeshed in a network of 3-5 nm filaments (represented by fine interdigitating lines) and surrounded by 30-40 nm granules (represented by the open circles). The NC occupied Zone 1; electron-dense material was found on the leading edge of pioneering NC cell processes with little or no matrix found in the extracellular spaces of trailing cells. Behind the leading NC cell front, the BL was thinned and stained with CI broken lines). tain protein has not yet been determined. However, any future in vitro studies on the biological effects of chondroitin sulfate should take into consideration chondroitin sulfate’s aggregation into larger molecular units. The distribution of 30-40 nm granules and CI staining reported here suggests that mesencephalic NC cells do not associate with areas that are high in sulfated polyanions (e.g., Zone 2 and the intercellular spaces of the mesoderm). In the trunk region, the majority of the NC cells utilize the ventral pathway (i.e., along the BL of the neural tube) [Pintar, 19781. Pintar  showed that GAG resistant to Streptomyces hyalurolyticus hyaluronidase but sensitive to chondroitinase ABC occurs at higher levels in the CFS under the ectoderm (lateral pathway). Using ruthenium red to preserve matrix, Newgreen et al.  found that NC do not initially migrate into trunk areas rich in granules, which appear similar to those described in this study. Conversely, the NC pathway in the head region (i.e., under the ectoderm) has been shown to be hyaluronate-rich and sulfate-poor [Pratt et al., 1975), whereas the areas that NC do not enter (i.e., the cranial mesoderm, the area adjacent to the notochord, and the ventral portion of the neural tube) are sulfate-rich [Pratt et al., 1975; Bolender et al., 19801. The electron-dense material observed in this study resembled the interstitial bodies found in the pathway of trunk NC described by Mayer et al. [ 19811 as containing fibronectin. Comparable matrical structures found in the developing heart were trypsin-sensitive and labeled with 3H-fucose [Hay and Markwald, 19811. Con A binding to the electron-dense material is suggestive that this material contained glycoproteins. The codistribution of 3H-fucoseand 3H-glucosarnine incorporation with the electron-dense material is consistent with a possible glycoprotein identification. The only extracellular glycoprotein of the CFS identified at present in avian embryos is one that immunologically cross-reacts with plasma fibronectin antibody [Newgreen and Thiery, 1980;Mayer et al., 1981; Duband and Thiery, 19821. However, confirmation of a 180-220 KD glycoprotein isolated from the CFS has not been EXTRACELLULAR MATRIX AND CEPHALIC NEURAL CREST established. Although plasma fibronectin has been shown to enhance NC cell attachment to collagen and increase cell locomotion in vitro [Greenburg et al., 1981; Erickson and Turley, 1983; Rovasio et al., 19831. NC cells themselves do not synthesize or secrete fibronectin [Loring et al., 1977; Newgreen and Thiery, 1980; SieberBlum et al., 19811. NC cells may become associated with the BL not so much because of a repulsion to sulfated macromolecules but rather because fibronectin or other glycoproteins (e.g., laminin) may be included within the electron-dense material. The observations that cranial NC pathways appear rich in interstitial bodies [Tosney, 19821, the occurrence of frequent contacts between the BL and NC cell processes, and the accumulations of electron-dense material on NC cell surfaces observed in this study support this possibility. Since the interaction of fibronectin with umbilical hyaluronate, chondroitin sulfate, or cartilage proteoglycan strongly inhibited NC cell adhesion and migration in culture [Newgreen, 1982; Erickson and Turley, 19831, we suggest that the most likely basis for the preferential association of NC with cranial ectoderm is that the ectodermal BL components facilitate attachment. Conversely, the mesodermally associated matrix, rich in sulfated GAG and hyaluronate, may inhibit attachment or promote detachment. Investigation of this hypothesis will require further characterization of the matrical composition and future in vitro studies utilizing isolated endogenous matrices and tissues. One of the most significant observations of the study was that CI staining of the BL and immediate subjacent matrix (i.e., Zone 1)prior to the appearance of NC was not observed even though such structures labeled with 35S-sulfate. It seems unlikely that this was a n artefact due to a penetration problem since the embryo remained intact during the processing and staining procedures and therefore the CI must have first passed through Zone 1 to reach inner areas of the embryo that stained positively. The explanation for the absence of CI staining in Zone 1 where sulfate had been incorporated may be that 35S-sulfate incorporation is a more sensitive method of detection or, more likely, that the sulfate groups were masked from CI dye molecules (i.e., sulfate groups were ionically linked or complexed with other matrical components). One such component that could have masked sulfated matrix was the ectodermal electron-dense material in Zone 1 tentatively identified as glycoprotein. Identification of the sulfated polyanion that stained with CI in areas occupied by NC cells was not established; possibilities include heparan sulfate proteoglycan (identified in other BL [Kanwar and Farquhar, 1979; Hassell et al., 19801or sulfated glycoproteins such as entactin [Carlin et al., 19811. In prior light microscopic histochemistry studies [Bolender et al., 19801, sulfated polyanions were found beneath the ectoderm that were testicular hyaluronidase-resistant and stained with alcian blue (pH. 5.7) at high MgCl2 concentrations (0.6 M)indicating the presence of polyanions more highly sulfated than chondroitin sulfate. Therefore, the acquired CI staining of sulfated polyanions in Zone 1 is probably not chondroitin sulfate. Alternatively, these modifications i n sulfate staining may reflect synthesis of new matrical components by NC [Greenburg and Pratt, 1977; Pintar, 19781. 67 Other investigators have noted changes in ECM with the passage of NC cells. Tosney  reported a decrease in the number of interstitial bodies in areas occupied by NC in situ. However, the nature of the interstitial bodies in this region was not determined. In our studies, the primary ECM component modified in areas occupied by NC cells was the electron-dense material tentatively identified as glycoprotein. Much of this material appeared to become associated with pericellular surfaces of the NC cells suggesting that, if ECM influences the pathways taken by NC or their differentiation, glycoproteins may play an active role. The ultimate significance of the matrix modifications observed in this study is as yet unclear. Weston and co-workers [Weston, 1981; Weston et al., 19781 have suggested that the environment may change in its ability to support NC motility during embryogenesis. By modifying matrix, pioneering NC themselves may limit the translocation of trailing NC cells (Fig. 19)and thus ultimately provide the mechanism terminating the initial dispersion of NC cells. The data indicate that the NC cell surfaces within the NC population may differ in their ECM associations (i.e., the dorsal layer of NC interacting with the ectoderma1 matrix and the ventral layer with the matrix of Zone 2; Fig. 19). Again, any studies on ECM interaction with NC should allow for this possibility. Evidence has accumulated suggesting that different subpopulations of NC exist at the time the NC leave the neural tube [Cohen, 1977; Cochard and Cotley, 19831. Hence, in vitro observations of migrating NC on flat substrates previously conditioned with a single or exogenous matrix component may artificially select out a specific subpopulation of NC cells. This serves as a caveat when correlating isolated in vitro observations to the more complex, heterogenous in vivo situation. The data presented suggest that through the specific distribution of ECM components NC become associated with particular surrounding tissues (surface ectoderm vs. neural tube). Once NC are associated with these tissues, they may influence NC movement. Whether NC move actively or passively has not been resolved, but experiments are in progress that will determine this. ACKNOWLEDGMENTS The authors gratefully acknowledge Stephen Sickerman for the preparation of Figure 19, Tony Frisbie for his photographic assistance, Makiko Hartman for her technical assistance, and Carol McLain for her clerical assistance. This research was supported by NIDR grant No. DE04603. LITERATURE CITED Bissell, M.J., H.G. Hall, and G. Parry. (1982) How does extracellular matrix direct gene expression? J. Theor. Biol., 99:31-68. Bolender, D.L., W.G. Seliger, and R.R. Markwald (1980) A histochemical analysis of polyanionic compounds found in extracellular matrix encountered by migrating neural crest cells. Anat. Rec., 196:401-412. Bolender, D.L., W.G. Selinger, R.R. MarkwaId, and P.R. Brauer (1981) Structural analysis of extracellular matrix prior to the migration of cephalic neural crest cells. In: Scanning Electron Microscopy/ 1981.0.M.Johari, ed. SEM, Inc., AMF: O’Hare, Illinois, Vol 11, pp. 285-296. Carlin, B., R. Jaffe, B. Bender, and A.E. Chung (1981) Entactin, a novel basal lamina-associated sulfated glycoprotein. J. Biol Chem., 256.5209-5214, 68 P.R. BRAUER, D.L. BOLENDER, AND R.R. MARKWALD Cochard, P., and P. Coltey (1983)Cholinergic traits in the neural crest: acetylcholinesterase in crest cells of the chick embryo. Dev. Biol., 98t221-238. Coffey, J.W., 0. Neal Miller, and 0.2. Sellinger (1964)The metabolism of L-fucose in the rat. J. Biol. Chem., 239:4011-4017. Cohen. A.M. (1977) Indenendent exmession of the adrenereic nhenotype by neural trek cells in dtro. Proc Natl. Acad. &i.&USA, 79t2899-2903. Duband, J.L., and J.P. Thiery (1982) Distribution of fibronectin in the early phase of avian cephalic neural crest cell migration. Dev. Biol., 93t308-323. Erickson, CA., and E.A. Turley (1983) Substrate formed by combinations .of extracellular matiix components alter neurai crest cell motility in uitro. J. Cell Sci., 61t299-323. Ehrlich, K.C. (1981)Proteoglycan synthesis by rat lung cells culture in uitro. J. Biol. Chem., 256t73-80. Gasser, R.F. (1979) Evidence that sclerotomal cells do not migrate medially during normal embryonic development of the rat. Am. J. Anat., 154t5091524. Greenburg, J.H., and R.M. Pratt (1977) Glycosaminoglycans and glycoprotein synthesis by cranial neural crest cells in uitro. Cell Diffe;., 6:iig-i32. Greenburg, J.H., S. Seppa, H. Seppa, and A.T. Hewitt (1981) Role of collagen and fibronectin in neural crest cell adhesion and migration. Dev. Biol., 87t259-266. Hamburger, V., and H.L. Hamilton (1951) A series of normal stages in the development of the chick embryo. J. Mxphol., 88~49-92. Hascall, V.C., and S.W. Sajdera (1970)Physical properties and polydispersity of proteoglycans from bovine nasal cartilage. J. Biol. Chem., 245t4920-4930. Hassell, J.R., P.G. Robey, H.J. Barrach, J. Wilczek, S.I. Rennard, and G.R. Martin (1980)Isolation of a heparan sulfate-containing proteoglycan from basement membrane Proc. Natl. Acad. Sci. USA, 77t4494-4498. Hay, D.A., and R.R. Markwald (1981)Localization of fucose-containing substances in developing atrioventricular cushion tissue. In: Perspectives in Cardiovascular Research, Vol. 5, Mechanisms of Cardiac Morphogenesis and Teratogenesis. T. Pexieder, ed. Raven Press, New York, pp. 197-211. Johnston, M.C. (1966)A radioautographic study of the migration and fate of cranial neural crest cells in the chick embryo. Anat. Rec., 156:143- 156. Kanwar, Y.S.. and M.G. Farquhar (1979) Isolation of glycosaminoglycans.(hep&an sulfate) from glomerular basement membranes. &&. Natl. Acad. Sci. USA, 76t4493-4497. Kitten, G.T., R.B. Runyan, and R.R. Markwald (1981)Structural characterization of a hyaluronate rich matrix. Anat. Rec., 199r143A. Kvist, T.N., and C.V. Finnegan (1970) The distribution of glycosaminoglycans in the axial region of the developing chick embryo. I. Histochemical Analysis. J. Exp. Zool., 175t221-240. Loring, J., C. Erickson, and J.A. Weston (1977) Surface proteins of neural crest, crest-derived and somite cells in uitro. J. Cell Biol., 75t71a (abstract). Markwald, R.R., T.P. Fitzharris, H. Bank, and D.H. Bernanke (1978) Structural analyses on the matrical organization of glycosaminoglycans in developing endocardial cushions. Dev. Biol., 62t292316. Markwald, R.R., F.M. Funderburg, and D.H. Bernanke (1979) Glycosaminoglycans: Potential determinants in cardiac morphogenesis. Tex. Rep. Biol. Med., 39253-270. Mayer, B.W. Jr., E.D. Hay, and R.O. Hynes (1981) Immunocytochemical localization of fibronectin in embryonic chick trunk and area vasculosa Dev. Biol., 93t308-323. Newgreen, D.F. (1982) Adhesion to extracellular materials by neural crest cells at the stage of initial migration. Cell Tissue Res., 227t297-317. Newgreen, D.F., I.L. Gibbins, J. Sauter, B. Wallenfels, and R. Wurtz (1982) Ultrastructural and tissue culture studies on the role of fibronectin, collagen, and glycosaminoglycans in the migration of neural crest cells in the fowl embryo. Cell Tissue Res., 221t521549. Newgreen, D.F., and J.P. Thiery (1980) Fibronectin in early avian embryos: Synthesis and distribution along the migratory pathways of neural crest cells. Cell Tissue Res., 211t269-291. Nichols, D.H. (1983) Neural crest formation and the origin of the first branchial arch in the mouse embryo. Anat. Rec. 205t142A. Noden, D.M. (1975) An analysis of the migratory behaviour of avian cranial neural crest cells in the chick embryo. Dev. Biol., 42t106130. Noden, D.M. (1978a) The control of avian cephalic neural crest cytodifferentiation. I. Skeletal and connective tissues. Dev. Biol., 67t296312. Noden, D.M. (1978b)The control of avian cephalic neural crest cytodifferentiation. 11. Neural tissues. Dev. Biol., 67t313-329. Noden, D.M. (1984)Craniofacial development: New views on old problems. Anat. Rec., 208:l-13. Norling, B., B. Glimelius, B. Westermark, and A. Wasteson (1978) A cho&oitin sulfate proteoglycan from cultured glial cells aggregates with hyaluronic acid. Biochem. Biophys. Res. Commun., 84t914921. Oldberg, A., and E. Ruoslahti (1982) Interactions between chondroitin sulfate proteoglycan, fibronectin, and collagen. J. Biol. Chem., 257t4859-4863. Pintar, J.E. (1978) Distribution and synthesis of glycosaminoglycans during quail neural crest morphogenesis. Dev. Biol., 67t444-464. Pratt, R.M., M.A. Larsen, and M.C. Johnston (1975) Migration of cranial neural crest cells in a cell-free matrix. Dev. Biol., 67:298-305. Rovasio, R.A., A. DeLouvee, K.M. Yamada, R. Timpl, and J.P. Thiery (1983) Neural crest cell migration: Requirements for exogenous fibronectin and high cell density. J. Cell Biol., 961462-473. Sieber-Blum, M., F. Sieber, and K.M. Yamada (1981)Cellular fibronectin promotes adrenergic differentiation of quail neural crest cells in uitro. Exp. Cell Res., 133t285-295. Singley, C.T., and M. Solursh (1980) The use of tannic acid for the ultrastructural visualization of hyaluronic acid. Histochemistry, 65: 93-102. Solursh, M., and H. Katow (1982) Initial characterization of sulfated macromolecules in the blastocoels of mesenchyme blastulae of Strongylocentrotus purpuratus and Lytechinus pictus. Dev. Biol., 94~326-336. Spicer, S.S., R.G. Horn, and T.J. Leppa (1967) Histochemistry of connective tissue mucopolysaccharides. In: The Connective Tissue. B.M. Wagner and D.E. Smith, eds. (International Academy of Pathology Monograph 7.) Williams and Wilkins, Baltimore, pp. 251303. Tosney, K.W. (1978) The early migration of neural crest cells in the trunk region of avian embryo: An electron microscopic study. Dev. Biol., 62317-333. Tosney, K.W. (1982) The segregation and early migration of cranial neural crest cells in the avian embryo. Dev. Biol., 89t13-24. Vogel, K.G., and D.W. Peterson (1981)Extracellular, surface, and intracellular proteoglycans produced by human embryo lung fibroblasts in culture ( I n - 9 0 ) . J. Biol. Chem., 256:13235-13242. Weston, J.A. (1970) The migration and differentiation of neural crest cells. Adv. Morphogen., 8:41-114. Weston, J.A. (1981) The regulation of normal and abnormal neural crest cell development. In: Neurofibromatosis. J. Mulivihill and V. Riccardi, eds. (Adv. Neurol. Vol. 29.) Raven Press, New York, pp. 77-94. Weston, J.A. and S.L. Butler (1966) Temporal factors affecting localization of neural crest cells in the chicken embryo. Dev. Biol., 14:246-266. Weston, J.A., M.A. Derby, and J.E. Pintar (1978) Changes in extracellular environment oi neural crest cells during their early migration. Zoon, 6103-113. Yamada, K.M. (1973)The effect of digestion with Streptomyces hyaluronidase upon certain histochemical reactions of hyaluronic acidcontaining tissues. J. Histochem. Cytochem., 21:794-803.