Cell specific effects of glycosaminoglycans on the attachment and proliferation of vascular wall components

MICROVASCULAR

RESEARCH

31, 41-53 (1986)

Cell Specific Effects of Glycosaminoglycans on the Attachment and Proliferation of Vascular Wall Components ALICIA ORLIDGE AND PATRICIA A. D’AMORE* Departments of *Surgical Research and Pathology, *The Children’s Hospital and Harvard Medical School, Boston, Massachusetts 02115 Received March 4, 1985 Capillary formation has been correlated with changes in basement membrane-associated glycosaminoglycans (GAGS).’ During capillary growth when endothelid cells (EC) undergo extensive proliferation and migration and pericytes are scarce, hyaluronic acid (HA) levels are elevated. Upon capillary maturation when EC migration and proliferation cease and pericytes appear, the dominant GAG is heparan sulfate (HS). To investigate the potential role of GAGS in the angiogenic process, we studied the effect of HA, heparin, chondroitin sulfate, and dermatan sulfate on the attachment and proliferation of vascular wall cells in vitro. Cell attachment was studied by determining the number of cells attached to GAGtreated substrates. Whereas HA inhibited the attachment of both pericytes and smooth muscle cells (SMC) by nearly 80% after 8 hr, it enhanced capillary EC attachment by more than 30%. Retinal pigment epithelial cells and dermal fibroblasts were employed as controls and none of the GAGS exammed significantly altered the attachment of these cells. The effect of GAGS on cell proliferation was determined by the addition of soluble GAGS to cells cultured for the time required for three population doubhngs. Heparin addition resulted in a dose-dependent inhibition of both pericyte and SMC proliferation with maximal inhibition of 50% at 100@g/ml, whereas this concentration of heparin moderately enhanced capillary EC proliferation. These effects were not observed for any other cell type or with any other GAG and indicate that GAGS have cell-specific effects on the attachment and proliferation of cells of the vascular wah. 0 1986 Academic press, IIIC.

INTRODUCTION A number of studies have reported the effects of matrix and matrix components on cell function in vitro (for review see 19). Cell-matrix interactions have been shown to modulate growth (17,18,25), migration (29,34), and differentiation (27,29) of various cell types in experimental systems. Among the isolated matrix components that have been shown to influence cell function, glycoproteins have received much attention. The effects of various collagens on cell attachment in vitro have been thoroughly investigated (24,25,29,32). Cells that normally grow on the interstitial collagens types I and III preferentially attach to these collagens in vitro, whereas cells that associate with a basement membrane in vivo adhere to type IV collagen in vitro. Two other matrix-associated glycoproteins, fibronectin ’ The abbreviations used are: GAGS, glycosaminoglycans; EC, endotheliaf cells; HA, hyaluronic acid; HS, heparan sulfate; SMC, smooth muscle cells; DMEM, Dulbecco’s modified Eagle’s medium; PBS, phosphate-buffered saline; BSA, bovine serum albumin. 41

0026-2s62/86 $3.00 Copyright0 I!%36 by Academic Press,Inc. All rightsof reproduction in anyformreserved. Printedin U.S.A.

42

ORLIDGE

AND

D’AMORE

and laminin, have been demonstrated to mediate the adhesion of cells to various substrates including collagen and plastic (15,16,41,50). Fibronectin, which has been localized in a variety of basement membranes (40), has specific binding sites for the surfaces of many cell types as well as for various matrix components including collagen and glycosaminoglycans (GAGS) (39,51,52). Laminin, which has been localized in the lamina rara of basement membranes and is believed to have a binding site for type IV collagen (10,43), has been shown to mediate the attachment of epithelial cells to collagen substrates (41). Less-studied components of the extracellular matrix are the GAGS. The GAGS exist in vivo as complexes of proteoglycans localized at the cell surface, either in direct association with the cell membrane or in basement membranes, and in the extracellular matrix. GAGS have been shown to influence a variety of cell functions including attachment (1,3,39), motility (13,30,34), differentiation (27,4547), and proliferation (4-6,23,31,33). Several in vitro studies correlate shifts in GAG populations with modulation of cell growth. Cell surface HS appears to influence cell cycle events and is reported to be selectively shed just prior to mitosis (26). Exponentially growing cells produce higher levels of hyaluronic acid (HA) than do confluent cells (44), which synthesize sulfated GAGS, particularly heparan sulfate (HS) (7). The amount of cell surface HS in human fibroblasts varies with the density-dependent inhibition of cell proliferation, suggesting a role for HS in growth control (31). Furthermore, the addition of HS purified from plasma membranes of confluent, quiescent hepatocytes induces contact inhibition in hepatoma cultures (23). A comparison of GAG populations in parent and virally transformed 3T3 cells reveals that cell surface HS is decreased in the transformed state and does not vary with increasing cell density (7,48), thus correlating the loss of contact inhibition with reduced cell surface HS. GAGS have been similarly implicated in cell-substrate interactions. Following the analysis of the molecular composition of adhesion sites, Culp and his coworkers (8,36) proposed that cell-substrate interaction may be determined by the competitive binding of proteoglycans to fibronectin. Several investigators have correlated altered cell adherence with increased exogenous HA. A Chinese hamster cell line (CHO) variant characterized by enhanced detachment, has been shown to have a threefold increase in cell surface HA (3). Fisher and Solursh (11) observe that HA is a poor substrate for the attachment of quail embryo cells, and suggest that this decrease in cell attachment may facilitate cell migration which is known to occur in a HA-rich matrix in vivo. We now report the effects of purified GAGS on the attachment and proliferation of components of blood vessel walls: capillary EC, aortic SMC, and microvascular pericytes. Our results indicate that the GAGS, HA and heparin, have cell-specific effects on both attachment and proliferation of vascular wall components. MATERIALS AND METHODS QuantiJcation of substrate-bound GAGS. Tritium-labeled heparin and HA were employed to determine the amount of those GAGS that adhered to the tissue culture wells. [3H]HA (1.1 x lo5 cpm/pg) was biosynthetically labeled with [3H]acetate (6 Ci/mmole, New England Nuclear, Boston, Mass.) by cultures of rat fibrosarcoma cells (the generous gift of Dr. Brian Toole) using methods

GAGS

AND

VASCULAR

WALL

COMPONENTS

43

previously described (49). Labeled [3H]heparin (New England Nuclear, 0.2 mCi/mg) and 13H]HA were incubated in the tissue culture wells for 2 hr then rinsed with phosphate-buffered saline (PBS) as described for the attachment studies. The attached GAGS were hydrolyzed by the addition of 0.5 ml of 1 M NaOH and quadruplicate samples were counted in a Beckman Model LS 1800 scintillation counter and are reported as average micrograms GAG per square centimeter. The amount of the chondroitins that adhered to the tissue culture wells was measured using the metachromatic activity associated with GAG binding to Azure A dye. One hundred microliters of 1 N NaOH was added to wells that had been previously coated with pure populations of chondroitin sulfate C and mixed chondroitin sulfates (B and C) in order to hydrolyze the adherent GAG. Neutralization was accomplished with 100 ~1 of 1 N HCI. A standard curve of O20 pg of each chondroitin sulfate was prepared. One gram of Azure A dye was dissolved in 1.0 liter of distilled water and stored at 4” for up to 2 weeks. Immediately prior to use, the concentrated dye solution was diluted 1:50 to a final concentration of 0.02 g/liter. Triplicate IOO-~1aliquots of each standard and unknown GAG sample were added to 900 ~1 of the dilute Azure A solution and the absorbance at 620 nm was measured within 30 min. The concentration (pg/cm’) of each chondroitin was determined using the standard curve. Cell cultures. Cultures of microvascular pericytes were established from bovine retinal capillaries by the methods of Gitlin and D’Amore (14) and grown in Dulbecco’s modified Eagle’s medium (DMEM, Grand Island Biological Company, GIBCO, Grand Island, N.Y.) with 5% fetal bovine serum (Sterile Systems Inc., Logan, Utah). All culture media were supplemented with glutamine (final concentration 2 mM), penicillin and streptomycin (100 U/ml, 100 pug/ml, GIBCO). Capillary endothelial cells (EC) (provided by C. Butterfield and J. Folkman) were isolated from bovine adrenal cortex as described previously (12) and grown in DMEM with 10% calf serum (Sterile Systems) mixed 1:l with media conditioned for 48 hr by mouse sarcoma 180 cells. (The latter was omitted from media used in assays.) Smooth muscle cells (SMC), epithelial cells, and fibroblasts were obtained from explants of bovine aorta, bovine retina, and human foreskin, respectively, and were all grown in DMEM with 10% calf serum. All EC and epithelial cells were used prior to their tenth passage. Only first-passage pericytes and SMC prior to their fourth passage were used. Attachment assays. Attachment assays were performed in 16-mm wells of 24well tissue culture plates (Nunc, Vangard Inc., Neptune, N.J.) that were coated with various GAGS or bovine serum albumin (BSA). Stock solutions (GAGS, 100 pg/ml; BSA, 1%) of HA (MW 720,000, Chesapeake Biological Laboratories, Hunt Valley, Md.), sodium heparin (Hepar Industries Inc., Franklin, Ohio), chondroitin sulfate C (Miles Biologicals, Elkhart, Ind.), dermatan sulfate (chondroitin sulfate Type B, Miles, Elkhart, Ind.), and BSA (Sigma Chemical Co., St. Louis, MO.) were prepared in PBS (GIBCO) and sterilized by passage through a 0.2-pm non-protein retaining filter (Millex GV, Millipore Corp.). Wells were incubated with 1 ml of the stock solutions for 2 hr at 37”. The GAG solution was aspirated and the wells were rinsed once with PBS. The cells (lO,OOO/O.S ml) in the appropriate growth medium containing serum were added to each well and allowed to attach for time intervals of up to 8 hr. Controls consisted of cells that were plated onto uncoated tissue culture plastic or onto wells coated with

44

ORLIDGE AND D’AMORE

BSA. After removing the nonadherent cells by gently washing with PBS, the cells were trypsinized (0.05% trypsin, 0.02% EDTA) and the number of attached cells at each time point was determined by electronic counting using a Coulter Counter (Coulter Electronics Inc., Hialeah, Fla.). Differential cell attachment at each time point was expressed as a ratio of the number of cells attached to the treated wells over the number attached to the untreated wells. Proliferation assays. Proliferation assays were performed in 16-mm wells of 24-well tissue culture dishes. Capillary EC, fibroblasts, SMC, and epithelial cells were plated at a density of 10,000cells per well. Due to their low plating efficiency (50%) pericytes were plated at 20,000 cells per well. After allowing the cells to attach overnight, the unattached cells were removed and 1.O ml of fresh medium (containing serum as described above for each cell type) was added. The GAGS (in PBS) were added in lOO+l aliquots to final concentrations of 1 ng-1 mg/ml. Control cells received 100 ~1 of PBS. Each cell type was incubated for the time required for at least three doublings. The doubling time for pericytes is approximately 4 days and ranges from 12 to 24 hr for the other cell types used. Fibroblasts, epithelial cells, EC, and SMC were incubated for 3 days; pericytes were incubated for 2 weeks (and refed twice per week). Cell counts were determined electronically as described above. The effect of GAGS on cell proliferation was expressed as the percentage change in cell number: a ratio of the number of cells incubated in the presence of GAGS to the control cell number. Statistical analyses. Data from a representative experiment, indicative of the trends reported in this manuscript, were anaIyzed for statistical significance. Although each experiment was repeated several times, the nature of the in vitro experimentation prevents the simple averaging of all the cell numbers. Final data, expressed as percentage change from control, were determined from an average of cell numbers from 4-6 samples per test variable, from one representative experiment. Statistical analysis on these included the determination of mean, standard deviation, T interval confidence limits, and P values (where applicable); all are included in the appropriate figures. RESULTS Quuntitution of Substrate-bound GAG [3H]Heparin and [3H]HA were used to quantitate the amount of these GAGS that attached to the tissue culture plastic under the conditions utilized in this study. These experiments indicated that 1.45-2.90 pg/cm2 of heparin and 0.250.85 pg/cm2 of HA were attached to the well surface at the time when cells were plated (Table 1). Azure A binding was used to quantitate chondroitin sulfate C and mixed chondroitins. These experiments indicated that under the conditions employed in this study, 0.64-0.8 pug/cm2of each GAG was attached to the well surface (Table 1). These levels agree with the values for binding of GAGS to plastic published by Lander et al. (28). Attachment Studies HA inhibited the attachment of pericytes by more than 80% after 8 hr (Fig. 1). None of the other GAGS or the BSA control significantly inhibited the attachment of pericytes. The inhibition of pericyte attachment observed at 2 hr after plating

GAGS AND VASCULAR WALL COMPONENTS

45

TABLE 1 QUANTIFICATION OF SUBSTRATE-BOUND GLYCOSAMINOGLYCANS

Glycosaminoglycan Heparin Hyaluronic acid Chondroitin sulfate-C Chondroitin sulfate (B and C)

Concentration bdcm’) 1.45-2.90 0.25-0.81 0x54-0.88 0.64-0.75

Note. Glycosaminoglycans (100 fig/ml) were incubated in tissue culture wells for 2 hr and rinsed once with PBS. The adherent GAGS were hydrolyzed with NaOH and assayed for GAG content using either an Azure A binding assay (for chondroitins) or ‘H-labeled GAGS (for heparin and hyaluronic acid).

remained constant over 8 hr (Fig. 2). The GAG-coated substrates influenced the attachment of aortic SMC in a manner similar to that observed for pericytes. As Fig. 1 illustrates, SMC attachment to HA-treated substrate was inhibited by greater than 80% after 8 hr. This inhibition was evident at 2 hr and was consistent over 8 hr (Fig. 2). In order to determine if any reattachment occurred during PERCYTES

FIG. 1. The effect of GAG-treated substrates on the attachment of pericytes and SMC. Pericytes and SMC were plated into tissue culture wells that had been treated with 100 Fg of GAG in 1 ml PBS. Controls consisted of cells plated into untreated and BSA-treated tissue culture wells. The number of attached cells was determined after an S-hr incubation. In this representative experiment each bar corresponds to the average of quadruplicate cell counts with standard deviations.

46

ORLIDGE

AND

D’AMORE

FIG. 2. An 8-hr time course of the effect of HA-coated plastic on the attachment of pericytes and SMC. Pericytes and SMC were plated into untreated tissue culture wells (0) or into wells that had been treated with 100 pg HA in 1 ml PBS (0). The number of attached cells was determined at 2-hr intervals over an 8-hr period. In this representative experiment each point corresponds to an average of quadruplicate cell counts.

60 60 40 20 &

0

b+

BOVIMF CAP/U

AUY

ENOOKWWAL

CELLS

80-

40

FIG. 3. The effect of GAG-treated substrates on the attachment of fibroblasts, pigment epithelial cells, and capillary EC. Individual cell types were plated into tissue culture wells that had been treated with 100 pg of a GAG in 1 ml PBS. Controls consisted of cells plated into untreated and BSA-treated tissue culture wells. The number of attached cells was determined after an 8-hr incubation. In this representative experiment each bar corresponds to the average of quadruplicate cell counts with standard deviations.

GAGS

AND

VASCULAR

WALL

COMPONENTS

47

extended times, pericyte and SMC attachment was measured after 24 hr of incubation on HA-coated substrates. The level of pericyte and SMC attachment remained constant over this extended time (data not shown). The effect of HA on SMC attachment varied as a function of the passage number of the cells; with increasing passage numbers the SMC were less inhibited by the HA coating. SMC attachment measured prior to the cells’ fourth passage was inhibited 7080% by HA, whereas with eighth and twelfth passagecells the inhibition decreased to 47 and 36%, respectively. To investigate if the HA inhibition of attachment was a general or cell-specific response, the effect of GAG-treated substrates on the attachment of other cell types was examined (Fig. 3). HA did not inhibit capillary EC attachment but, on the contrary, enhanced their attachment by 30 + 8%. Furthermore, plating the EC on HA-treated plastic consistently induced an alteration in cell shape from a polygonal, spread morphology to an elongated, spindle shape. The attachment of EC was not influenced by treating the substrate with any of the other GAGS or with BSA. The attachment of human fibroblasts and retinal pigment epithelial cells was unaffected by any of the GAG- or BSA-coated substrates. Proliferation

Studies

To determine the effect of the GAGS on proliferation of vascular wall and control cells, the change in cell number was measured in response to increasing concentrations of each GAG (1 ng-1 mg/ml). The most significant effect of a GAG on cell proliferation was the inhibition of pericytes and SMC proliferation by heparin (Fig. 4). The inhibition of pericyte proliferation by heparin was dose dependent (Fig. 5). Slight inhibition was evident at heparin concentrations as

FIG. 4. The effect of soluble GAGS on the proliferation of pericytes and SMC. Pericytes and SMC were plated into tissue culture wells at densities determined by the differential cell plating efficiency (pericytes at 20,00O/well; SMC at lO,OOO/well) and allowed to attach overnight. Each GAG (1 ng-I mg/ml PBS) was added to the cells in fresh media and incubated for the time required for three population doublings (2 weeks for pericytes; 3 days for SMC). Control cells received PBS with fresh media. The change in proliferation is expressed as the percentage change in cell number: the ratio of the number of GAG-treated cells to the control cell number. In this representative experiment each bar corresponds to the average of quadruplicate cell numbers with standard deviations.

48

ORLIDGE

AND

D’AMORE

FIG. 5. The effect of increasing concentrations of heparin on pericyte and SMC proliferation. Pericytes and SMC were plated into tissue culture wells at densities determined by the differential cell plating efficiency (pericytes at 20,00O/well; SMC at 10,000/we11)and allowed to attach overnight. Heparin (1 ng-1 mg/ml PBS) was added to the cells with fresh media and incubated for the time required for three population doublings (2 weeks for pericytes; 3 days for SMC). Control cells received PBS with fresh media. In this representative experiment each point corresponds to the average of quadruplicate cell counts.

low as IO rig/ml and concentrations above 100 pg/ml did not result in greater inhibition (48 k 5%). The response of vascular SMC to soluble GAGS was similar to that described for pericytes (Fig. 4). Heparin inhibited SMC proliferation in a dose-dependent manner and resulted in a 53 + 7% decrease in cell number at 100 pg/ml. No other GAG significantly altered the proliferation of SMC or pericytes (Fig. 4). Since the maximum effect of heparin on cell proliferation was observed at 100 pg/rnl, the results of experiments measuring the effect of these GAG concentrations are shown for epithelial cells, fibroblasts, and EC (Fig. 6). The only effect of a GAG on cell proliferation was a small (25%) but significant (P < 0.02) stimulation of capillary EC proliferation by heparin. Interestingly, the proliferative effect of heparin on capillary EC was found to be variable. When the response to soluble heparin was measured in seven separate experiments, proliferation of capillary EC was stimulated four times and unaffected three times. In order to quantitate the influence of heparin on the mitogenic activity of an

GAGS

AND

VASCULAR

WALL

COMPONENTS

49

UGMFNT EFYTHELIAL CELLS I50

BObWE CdULLARY EM%XVEUAL 150-

CELLS

PCO.02

loo-

SO-

.u

Modlo Control

bly-

Her&n Add

s C LclwndmitIn_1 sullat~

FIG. 6. The effect of soluble GAGS on the proliferation of fibroblasts, pigment epithelial cells, and capillary EC. Each cell type was plated into tissue culture wells (10,000 cells/well) and allowed to attach overnight. The various GAGS (I ng-1 mg/ml PBS) were added to the cells with fresh media and incubated for 3 days. Control cells received PBS with fresh media. The change in proliferation is expressed as the percentage change in cell number: the ratio of the number of GAGtreated cells to control cell number. In this representative experiment each bar corresponds to the average of quadruplicate cell numbers with standard deviations.

EC growth factor that is known to bind heparin (9), the effect of the simultaneous addition of heparin and retina-derived growth factor was measured. These studies revealed that the simultaneous addition of heparin (100 pg/ml) and the mitogen resulted in a significant potentiation of the proliferative response of the capillary EC to this growth factor (Fig. 7). The synergism was most obvious at lower mitogen concentrations and became less significant as the mitogen concentration was increased. DISCUSSION The GAGS, HA, and heparin have cell-specific effects on the attachment and growth of the cellular components of the blood vessel wall. These effects are antagonistic: whereas heparin inhibits the proliferation of pericytes and SMC and potentiates capillary EC growth, HA inhibits pericyte and SMC attachment and enhances capillary EC attachment. Other workers have implicated HA and heparin-like molecules in the modulation

SO

ORLIDCE

AND

D’AMORE

LWV~~NECAPUARYENW??%EL/AL G5LLS loo-

FIG. 7. The effect of heparin on the response of capillary EC to increasing concentrations of retina-derived growth factor. Capillary EC were plated into tissue culture wells (10,000 cells/well) and allowed to attach overnight. Heparin (100 pg/ml in PBS) was added with varying amounts (530 ~1) of crude retina-derived growth factor (6 mg/ml) and incubated for 3 days. In this representative experiment each point corresponds to the average of quadruplicate cell counts.

of cell attachment. In a study of a L-6 myoblast variant selected for an adhesion deficiency, Schubert and LaCorbiere (38) report that alterations in GAG synthesis by variant cell types account for the cells’ impaired adhesion and propose that GAGS secreted by these cells are mediators of cell adhesion. They measured adhesion of the variants to “substrate-attached material” prepared from parent and variant cells. They observe that the addition of purified GAGS inhibits the adhesive interaction of the variant in a manner specific to the type and concentration of GAG; HA and chrondroitin sulfate prevent variant attachment to parent “substrate-attached material” whereas HA, heparin, and HS inhibit variant adhesion to variant “substrate-attached material.” The complexity of GAG-dependent interactions has, thus far, prevented elucidation of mechanisms involved in the modulations of attachment. It is clear, however, that specific matrix and cell surface GAGS are involved in both receptordependent and concentration-dependent events. Studies of the specific interactions of GAGS with other glycoproteins known to mediate attachment have provided some insight into potential pathways. The binding of HS and heparin to rat hepatocytes is reported to modify their attachment to fibronectin-coated substrates and to stabilize fibronectin-dependent binding to matrix collagen in vitro (21,51,.52). Such interactions are thought to occur via conformational changes or “activation” of fibronectin and have led to the suggestion that fibronectin may form different macromolecular complexes depending on which GAG binds to it (21,22). Our results indicate that heparin significantly inhibits the proliferation of pericytes and SMC. This finding is in agreement with the work of Clowes and Karnovsky (6) and Castellot and his coworkers (4,5) who report the inhibition of SMC growth by heparin and have suggested a role for EC in the regulation of SMC function. Heparin-like molecules have been shown to be enzymatically released from the surface of EC both in vivo and in vitro and are postulated to suppress both proliferation (4-6) and migration (30) of adjacent SMC. Physiologically, this heparin may function to prevent myointimal thickening of muscular arteries following EC injury. In light of our data, which indicate that pericyte proliferation

GAGS

AND

VASCULAR

WALL

COMPONENTS

51

is also inhibited by heparin, we postulate that similar interactions between the EC and pericytes may occur at the microvascular level. Pericytes have been suggested to be a less-differentiated SMC capable of performing SMC functions at the microvascular level (35). The observation that SMC and pericyte respond similarly to the GAGS provides additional evidence for functional similarities between these cell types. Our results further indicate that capillary EC proliferation is enhanced by heparin. The variability of this observation may reflect a potentiation by heparin of some unknown component of the growth medium. This possibility is supported by our finding that the simultaneous addition of heparin and retina-derived growth factor, a known EC mitogen, results in a synergistic proliferative response. This is in agreement with the work of Thornton and her coworkers (42) who report that heparin potentiates the response of human large vessel EC to an EC mitogen. Our observations indicate that a similar relationship exists at the capillary level and provides evidence that heparin may influence vascular growth control in a cell-specific manner unrelated to blood vessel size. The antagonistic character of the GAG effects noted in this study appears to parallel the shift in GAGS observed during growth and differentiation in vivo. Ausprunk and her coworkers (2) have studied blood vessel-associated GAGS in the developing chick chorioallantoic membrane. During the initial phases of vasculogenesis, which are characterized by migration and proliferation of EC, HA concentrations are high. As the vessels mature, and migration and proliferation cease, there is a marked decrease in HA and an increase in the synthesis of the sulfated GAGS, particularly heparin-like molecules. Our in vitro results are consistent with these developmental observations. We find that capillary EC preferentially attach to the HA-coated substrate which may mimic the HA-rich matrix known to be associated with capillary EC development in vivo. Furthermore, this HA-coated substrate induces a change in EC morphology from a spread, cobblestone shape characteristic of quiescent monolayers to an elongated, spindleshape associated with cell migration (20), another function associated with capillary proliferation. Conversely, pericytes which are not normally associated with the HA-rich matrix of the growing capillary in vivo, demonstrated a significant reduction in attachment to HA-coated substrates in vitro. In development, the pericytes arrive at the capillary at a time when HA levels are low and HS levels are elevated (2). In addition, pericytes exist in vivo as a single layer which is wrapped around and intimately associated with the EC (35). Since EC synthesize large amounts of heparin-like molecules which are retained at the cell surface, the observed inhibition of pericyte proliferation by heparin in vitro may reflect an in vivo mechanism for their growth control. ACKNOWLEDGMENTS We gratefully acknowledge Dr. Bob Mello and Dr. Bill Tew at Chesapeake Biological Labs for their generous contribution of purified HA, and Dr. J. Folkman and C. Butterfield for the capillary EC. We thank Dr. John Voyta and Dr. Bruce Zetter for their critique of the manuscript, Dr. Betty Hay for her helpful discussion, Elisabeth Sproul for her excellent technical assistance, and Bonnie Troped and Vicki Elms for the expert typing of this manuscript. This work was supported by EY05318 (P,D.) and NIH postdoctoral fellowship (A.O.).

52

ORLIDGE

AND

D’AMORE

REFERENCES 1. ABATANGELO, G., CORTIVO, R., MARTELLI, M., AND VECCHIA, P. (1982). Cell detachment mediated by hyaluronic acid. Exp. Cell Res. 137, 13-78. 2. AUSPRUNK, D. H. (1982). Synthesis of glycoproteins by endothelial cells in embryonic blood vessels. Dev. Biol. 90, 79-90. 3. BARNHART, B. J., Cox, S. H., AND KRAEMER, P. M. (1979). Detachment variants of Chinese hamster cells. Hyaluronic acid as a modulator of cell detachment. Exp. Cell. Res. 119, 327-332. 4. CASTELLOT, J. J., JR., ADWNIZIO, M. L., ROSENBERG,R., AND KARNOVSKY, M. J. (1981). Cultured

endothelial cells produce a heparinlike inhibitor of smooth muscle cell growth. .I. Cell Viol. 90, 372-379. 5. CASTELLOT, J. J., JR., FAVREAU, L. V., KARNOVSKY, M. J., AND ROSENBERG,R. D. (1982). Inhibition of vascular smooth muscle cell growth by endothelial cell-derived heparin. Possible role of a platelet endoglycosidase. J. Biol. Chem. 257, 11256-I 1260. 6. CLOWES, A. W., AND KARNOVSKY, M. .I. (1977). Suppression by heparin of smooth muscle cell proliferation in injured arteries. Nature (London) 265, 625-626. 7. COHN, R. H., CASSIMAN, J. J., AND BERNFIELD, M. R. (1976). Relationship of transformation, cell density and growth control to the cellular distribution of newly synthesized glycosaminoglycans. J. Cell Biol. 71, 280-294. 8. CULP, L. A., MURRAY B. A., AND ROLLINS, B. J. (1979). Fibronectin and proteoglycans as determinants of cell-substratum adhesion. J. Supramol Strucr. 11, 401-427. 9. D’AMORE, P. A., AND KLAGSBRUN, M. (1984). Endothelial cell mitogens derived from retina and hypothalamus: Biochemical and biological similarities. J. Cell Biol. 99, 1545-1.549. 10. DEL Rosso, M., CAPPELLETTI, R., VITI, M., VANNUCCHI, S., AND CHIARUCI, V. (1981). Binding of the basement membrane glycoprotein laminin to glycosaminoglycans. Biochem. J. 199, 699-

704. 11. FISHER,M., ANDSOLURSH,M. (1979). The influence of the substratum on mesenchymal spreading in vitro. Exp. Cell Res. 123, l-14. 12. FOLKMAN, J., HAUDENSCHILD, C., AND ZETTER,B. R. (1979). Long-term culture of capillary endothelial cells. Proc. Nat/. Acad. Sci. USA 76, 5217-5221. 13. FORRESTER, J. V., ANDWILKINSON,P. C. (1981). Inhibition of leukocyte locomotion by hyaluronic acid. J. Cell Sci. 48, 315-331. 14. GITLIN, J. D., AND D’AMORE, P. A. (1983). Retinal capillary endothelial cells: Long-term culture using selective growth media. Micrvasc. Res. 26, 74-80. 15. GRINNELL, F. (1978). Cellular adhesiveness and extracellular substrata. In?. Rev. Cytol. 53, 6% 144. 16. GRINNELL, F., FELD, M., AND MINTER, D. (1980). Fibroblast adhesion to fibrinogen and fibrin substrata: Requirement for cold insoluble globulin. Cell 19, 517-525. 17. GOSPODAROWICZ,D., GREENBERG, G., AND BIRDWELL, C. R. (1978). Determination of cellular shape by the extracellular matrix and its correlation with the control of cellular growth. Cancer Res. 38, 4155-4171. 18. GOSPODAROWICZ,D., DELGADO, D., AND VLODAVSKY, 1. (1980). Permissive effect of the extracellular matrix on cell proliferation in vitro. Proc. Nat/. Acad. Sci. USA 77, 4094-4098. 19. HAY, E. D. (1981). Extracellular Matrix. J. Cell Biol. 91, 205S-223s. 20. HERMAN, I. M., AND D’AMORE, P. A. (1984). Capillary endothelial cell migration: Loss of stress fibers in response to retina-derived growth factor. J. Muscle Res. Ce// Motil., 5, 697-709. 21. JOHANSSON,S., AND HOOK, M. (1980). Heparin enhances the binding of fibronectin to collagen. Biochem.

J. 187, 521-524.

22. JOHANSSON,S., AND HOOK, M. (1984). Substrate adhesion of rat hepatocytes: On the mechanism of attachment to fibronectin. J. Cell Biol. 98, 810-817. 23. KAWAKAMI, H., AND TERAYAMA, H. (1981). Liver plasma membranes and proteoglycans prepared therefrom inhibit the growth of hepatoma cells in virro. Eiochim. Biophys. Acta 646, 161-168. 24. KLEBE, R. J. (1974). Isolation of a collagen-dependent cell attachment factor. Nature (London) 250, 248-251. 25. KLEINMAN, H. K., KLEBE, R. J,, AND MARTIN, G. R. (1981). Role of collagenous matrices in the adhesion and growth of cells. J. Ce// Biol. 88, 473-485. 26. KRAEMER, P. M., AND TOBEY, R. A. (1972). Cell-cycle dependent desquamation of heparan sulfate from the cell surface. J. Cell Bio/. 55, 713-717.

GAGS

AND

VASCULAR

WALL

COMPONENTS

53

27. KUJAWA, M. J., AND TEPPERMAN, K. (1983). Culturing chick muscle cells on glycosaminoglycan

substrates: Attachment and differentation. Dev. Biol. 99, 277-286. 28. LANDER, A. D., Furn, D. K., GOS~~DARO~ICZ, D., AND REICHARDT, L. F. (1982). Characterization

29. 30. 31. 32. 33.

34. 35. 36. 37. 38. 39.

40. 41. 42.

43. 44.

45. 46. 47. 48. 49. 50.

51. 52.

of a factor that promotes neurite outgrowth: Evidence linking activity to a heparan sulfate proteoglycan. .I. Cell Biol. 94, 574-585. MADRI, J., AND WILLIAMS, S. K. (1983). Capillary endothelial cell cultures: Phenotypic modulation by matrix components. J. Cell Biol. 97, 153-165. MAJACK, R. A., AND CLOWES, A. W. (1984). Inhibition of vascular smooth muscle cell migration by heparin-like glycosaminoglycans. J. Cell Physiol. 118, 253-256. MATUOKA, K., AND MITSUI, Y. (1981). Involvement of cell surface heparan sulfate in the densitydependent inhibition of cell proliferation. Cell Struct. Fancr. 6, 23-33. MURRAY, J. C., STINGL, G., KLEINMAN, H. K., MARTIN, G. R., AND KATZ S. I. (1979). Epidermal cells adhere preferentially to Type IV (basement membrane) collagen. J. Cell Biol. 80, 197-202. OHNISHI, T., OHSHIMA, E., AND OHTSUKA, M. (1975).Effect of liver cell coat acid mucopolysaccharides on the appearance of density-dependent inhibition in hepatoma cell growth. Exp. Cell Res. 93, 136-142. 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. Biol. 44, 298-305. RHODIN, J. A. (1968). Ultrastructure of mammalian venous capillaries, venules and small collecting veins. .I. Ultrastruct. Res. 25, 452-500. ROLLINS, B. J., AND CULP, L. A. (1979). Glycosaminoglycans in the substrate adhesion sites of normal and virus-transformed murine cells. Biochemistry 18, 141-148. RUOSLAHTI, E., AND ENGVALL, E. (1980). Complexing of fibronectin, glycosaminoglycans and collagen. Biochim. Biophys. Acta 631, 350-358. SCHUBERT, D., AND LACORBIERE, M. (1980). A role of secreted glycosaminoglycans in cellsubstratum adhesion. J. Biol. Chem. 25.5, 11564-11569. STAMATOGLOU, S., AND KELLER, J. M. (1983). Correlation between cell substrate attachment in vitro and cell surface heparan sulfate affinity for fibronectin and collagen. J. Cell Biol. 96, 18201823. STENMAN, S., AND VAHERI, A. (1978). Distribution of a major connective tissue protein, tibronectin, in normal human tissues. J. Exp. Med. 147, 1054-1064. TERRANOVA, V. P., ROHRBACH, D. H., AND MARTIN, G. R. (1980). Role of laminin in the attachment of Pam 212 (epithelial) cells to basement membrane collagen. Cell 22, 719-726. THORNTON, S. C., MUELLER, S. N., AND LEVINE, E. M. (1983). Human endothelial cells: Use of heparin in cloning and long-term serial cultivation. Science (Washingron, D. C.) 222, 623-625. TIMPL, R., ROHDE,H., GEHRON-ROBEY, P., RENNARD,S., FOIDART,J. M., AND MARTIN, G. R. (1979). Laminin-a glycoprotein from basement membrane. J. Biol. Chem. 254, 9933-9937. TOMIDA, M., KOYAMA, H., AND ONO, T. (1974). Hyaluronic acid synthetase in cultured mammalian cells producing hyaluronic acid. Oscillating changes during the growth phase and suppression by 5-bromodeoxyuridine. Biochim. Biophys. Acta 338, 352-363. TOOLE, B. P., AND GROSS, J. (1971). The extracellular matrix of the regenerating newt limb: Synthesis and removal of hyaluronate prior to differentiation. Dev. Biol. 25, 57-77. TOOLE,B. P., AND TRELSTAD,R. L. (1971). Hyaluronate production and removal during cornea1 development in the chick. Dev. Biol. 26, 28-35. TOOLE, B. P. (1981). Glycosaminoglycans in morphogenesis. In “Cell Biology of Extracellular Matrix” (E. D. Hay, ed.), pp. 259-294. Plenum, New York. UNDERHILL. C. B., AND KELLER, J. M. (1976). Density-dependent changes in the amount of sulfated glycosaminoglycans associated with mouse 3T3 cells. J. Cell. Physiol. 89, 53-64. UNDERHILL, C. B., AND TOOLE, B. P. (1979). Binding of hyaluronate to the surface of cultured cells. J. Cell Biol. 82, 475-484. YAMADA, K. M., AND OLDEN, K. (1978). Fibronectins-adhesive glycoproteins of cell surface and blood. Nature (London) 275, 179-184. YAMADA, K. M., AI(IYAMA, S. K., AND HAYASHI, M. (1980). Fibronectin structure and function, and its interaction with glycosaminoglycans. Biochem. Sot. Trans. 9, 506-508. YAMADA, K. M., KENNEDY, D. W., KIMATA, K., AND PRATT, R. M. (1980). Characterization of fibronectin interactions with glycosaminoglycans and identification of active proteolytic fragments. J. Biol. Chem. 255, 6055-6063.