Preliminary characterization of glial-secreted factors responsible for the induction of high electrical resistances across endothelial monolayers in a blood-brain barrier model.

Factors secreted by C6 glioma cells which induce electrical resistances across endothelial monolayers in an in vitro blood-brain barrier model have been partially characterised for the first time. These transendothelial electrical resistances (TEERs) were only evident when cell-free conditioned medium derived from C6 glioma cells was applied to the basolateral surfaces of confluent ECV304 or ECV304-9 cells which are both human umbilical vein endothelial cell lines (HUVEC). Electrical resistance values as high as 600 ohm. sq cm were obtained with this blood-brain barrier model and ultrafiltration techniques suggest that any factor(s) in the conditioned medium responsible for these TEERs have molecular masses of less than 1000 Da. Enzymic proteolysis and heat treatment carried out on the conditioned medium failed to inhibit its effect on the HUVEC monolayers suggesting that these C6 cell-secreted factors are unlikely to be proteins.

Effects of overproduction of the catalytic domain of 3-hydroxy-3-methylglutaryl coenzyme A reductase on squalene synthesis in Saccharomyces cerevisiae.

The enzyme 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase (HMG-R) is the major rate-limiting enzyme of the mevalonate pathway in many organisms, including yeasts. In the yeast Saccharomyces cerevisiae, there are two isoenzymes of HMG-R (Hmg1p and Hmg2p). Both consist of an anchoring transmembrane domain and a catalytic domain. We have removed the known controlling features of HMG-R by overproducing the catalytic domain of Hmg1p. This overproduction leads to an enhancement of squalene production, implying that HMG-R has been deregulated. The enhancement is apparent under semianaerobic and aerobic conditions. Despite the increase in squalene production, the amount of ergosterol produced by the HMG-R-overproducing yeast was not increased. This result suggests the presence of another regulatory step between squalene and ergosterol formation. Squalene levels generated by cells overproducing the catalytic domain of HMG-R were estimated to be up to 10 times those produced by wild-type cells. The enhancement in squalene production coincided with a reduction in growth rate. This reduction may be a direct consequence of the buildup of high concentrations of squalene and presqualene intermediates of the pathway.

Properties of an immortalised vascular endothelial/glioma cell co-culture model of the blood-brain barrier.

In an effort to obtain a useful in vitro model possessing some of the properties of the blood-brain barrier, we have investigated the properties and interactions of immortalized cell lines. Immortalised human umbilical vein endothelial cells (HUVEC-304), in co-culture with rat C6 glioma cells in a two-chambered assembly, form tight junctional complexes, and develop a permeability barrier having a high transcellular electrical resistance. The endothelial cells generate a barrier with greatest integrity in the presence of glioma cells, or in the presence of glioma cell conditioned medium. Under these conditions, the endothelial cells also display pronounced structural changes which do not occur in the absence of glioma cells. Morphological alterations include a flattening of cell shape from a cuboidal-type to a squamous-type of appearance, and a re-organization of F-actin microfilaments. The integrity of the barrier can be reversibly disrupted by osmotic shock or by tumor necrosis factor-alpha (TNF-alpha). We interpret these observations to indicate that co-cultures of immortalized vascular endothelial and C6 glioma cells provide a model for the investigation of cell-cell interactions required for the generation of a barrier having several properties of the blood-brain barrier.

Nitric oxide-induced perturbations in a cell culture model of the blood-brain barrier.

The actions of an intracellular nitric oxide generator compound on the properties of a co-culture model of the blood-brain barrier are described. Addition of the iron-sulphur cluster nitrosyl Roussin's black salt (RBS, heptanitrosyl-tri-mu3-thioxotetraferrate (1-)) resulted in a rapid and dose-dependent (50-250 microM) decline in the electrical resistance displayed by co-cultures of vascular endothelial cells and C6 glioma cells. The breach in barrier integrity elicited by RBS (250 microM) could be prevented by either haemoglobin (100 microM), methylene blue (200 microM), or by photon-induced inactivation of RBS. In contrast, the nitric oxide synthase inhibitor nitro-L-arginine methyl ester (250 microM) caused no inhibition in the decline in resistance of RBS-exposed cultures. Addition of 8-bromo-guanosine-cyclic monophosphate (500 microM) did not mimic the actions of RBS. Exposure to intense light of co-cultures manifesting a high transcellular electrical resistance resulted in a reduction in tissue resistance which could be prevented by the presence of haemoglobin (100 microM). We conclude that nitric oxide liberated from RBS results in a reversible diminution in the integrity of the endothelial cell barrier in the co-culture system, and we suggest that light-sensitive endogenous nitric oxide generator compounds may be present in intact cells. Possible roles of nitric oxide in blood-brain-barrier function are considered.

Influences of silicates and carnitine-silicate mixtures on the inhibition of aggregation of erythrocytes elicited by the presence of fibrinogen.

Carnitine and acylcarnitine derivatives have been reported to inhibit cell aggregation (Fritz and Burdzy, 1989, J. Cell. Physiol., 140:18-28). A follow-up of these observations showed that whereas the previously described effects of long-chain acylcarnitines were well replicated, those of carnitine on erythrocytes showed marked variability. The latter phenomenon was traced to the presence of silicates in carnitine solutions derived from the use of sodium hydroxide solutions stored in glass containers for the neutralization of carnitine. The present experiments have led to the discovery that oligomeric forms of silicates are powerful inhibitors of red blood cell aggregation which otherwise occurs in the presence of fibrinogen alone. The active form(s) of silicates in this assay, which appear to be generated by polymerization of silicates in metasilicate solutions on neutralization, are unstable and therefore transient under usual conditions. We estimate that the active oligomeric forms contain between 4 to 18 silicon atoms per molecule. When maintained at -18 degrees C in the presence of carnitine, but not in its absence, the active forms of oligomeric silicates remained stable for months, judging from their ability to inhibit cell aggregation. We conclude that carnitine stabilized the oligomeric form(s) of silicate, or that the species stabilized is an oligomeric silicate-carnitine complex. Comparable concentrations of choline, deoxycarnitine, or gamma-aminobutyrate were less effective in stabilizing the active silicate oligomers. The active forms of the silicate oligomers had Ki values of about 10 microM, calculated as the monomeric form, in inhibiting red blood cell aggregation. The data indicate that free carnitine does not directly inhibit erythrocyte inhibition, as previously interpreted, whereas long-chain acylcarnitine derivatives are active in the absence of silicates. Possible mechanism of actions of silicate oligomers on membranes are discussed.

Role of laminin in the morphogenetic cascade during coculture of Sertoli cells with peritubular cells.

Observations summarized in this article demonstrate an essential role of laminin during the restructuring processes that occur during coculture of Sertoli cells with testicular peritubular cells. The data presented indicate that laminin becomes detectable on the free surfaces of Sertoli cells only after reaggregation of Sertoli cells begins, coincident with the initiation of repolarization at a specific stage of the morphogenetic cascade. We infer that laminin deposited at this time serves as a cohesion molecule that permits peritubular cells to come into close contact with Sertoli cells and subsequently to spread along the free surfaces of Sertoli cells. These conclusions and inferences are based on the following experiments. Cycloheximide-treated peritubular cells in culture in MEM containing cycloheximide readily attach to laminin-coated polystyrene surfaces. By contrast, added peritubular cells do not attach onto monolayers of Sertoli cells in monoculture or onto Sertoli cells plated on top of peritubular cells and maintained in coculture for periods of up to 48 h. In cocultures maintained for 6 days, however, labeled peritubular cells readily adhere to the free surfaces of reaggregated Sertoli cells. Laminin, but not fibronectin, appears on the free surfaces of the reaggregated Sertoli cells at this time, coinciding with the period of initial mound formation. The addition of antilaminin IgG, but not antifibronectin IgG, blocks the attachment of cycloheximide-treated peritubular cells to laminin-coated plates and also blocks the subsequent migration of peritubular cells required to form a monolayer. Similarly, anti-laminin IgG inhibits the attachment and spreading of labeled peritubular cells seeded on the free surfaces of reaggregated Sertoli cells in mounds generated during the morphogenetic cascade. We interpret the combined data to indicate that the appearance of laminin on the free surfaces of Sertoli cells is required to permit peritubular cells to adhere and subsequently to migrate on Sertoli cell surfaces, resulting in the formation of a tubule-like structure.

Somatic cell-germ cell relationships in mammalian testes during development and spermatogenesis.

In the mammalian testis, somatic cells under hormonal regulation greatly influence the different stages of spermatogenesis, both in intermittent breeders and in animals which produce sperm continuously. In turn, specific populations of germinal cells modulate the function of Sertoli cells, the chief somatic cells within mammalian seminiferous tubules. Tubule formation can take place in the absence of germinal cells. Unlike homologous granulosa cells in the ovary, Sertoli cells retain many of their usual functions in germ cell-free animals. Some of the properties of Sertoli cells and their responses to stimulation by androgens or follicle-stimulating hormone are dependent upon information transmitted from neighbouring germinal cells at specific stages of the cycle of the seminiferous epithelium. We review the roles of some of the growth factors and paracrine agents synthesized and secreted by different classes of testicular cells. The potential roles of some of the known factors secreted by Sertoli cells (e.g. activin, inhibin, anti-Müllerian hormones, TGF-beta and somatomedin C) are considered in relation to the control of tubule formation, spermatogonial proliferation and cytodifferentiation, meiosis and the subsequent stages of spermatogenesis. We stress the importance of the unique tubule cytoarchitecture within which cell interactions take place and the changing nature of this cytoarchitecture at different stages of gonadal maturation.

Sites of action of carnitine and its derivatives on the cardiovascular system: interactions with membranes.

Carnitine plays an essential role in the regulation of long-chain fatty acid metabolism in skeletal and cardiac muscle, a process that is mediated by well-characterized enzymatic mechanisms. Here, Irving Fritz and Edoardo Arrigoni-Martelli review the evidence that carnitine and its O-acyl derivatives also influence membrane fluidity, ion channel functions, smooth muscle contractility, membrane stability and cardiac functions. The authors present the view that direct interactions of carnitine derivatives with cell membranes are independent of reactions catalysed by carnitine acyltransferases. They propose that the novel actions discussed are implicated in the mechanisms by which carnitine and its derivatives protect perfused hearts subjected to ischaemia or to oxidative stress, and help people suffering from certain types of myocardial ischaemia or peripheral arterial disease.

Interactions of Sertoli cells with laminin are essential to maintain integrity of the cytoskeleton and barrier functions of cells in culture in the two-chambered assembly.

The addition of anti-laminin IgG to the basal surfaces of rat Sertoli cells in culture in a two-chambered assembly results in a perturbation of F-actin arrangements, including disruption of the pericellular circumferal rings, impairments of the Sertoli cell permeability barrier, and subsequently focal defoliation, followed by cell reaggregation. The pentapeptide YIGSR, which competes with the laminin receptor for laminin (Kleinman and Weeks: Curr. Oph. Cell Biol., 1:964-967, 1989; Graf et al.: Biochemistry, 26:6896-6900, 1987) also elicited focal defoliation of Sertoli cells from the extracellular matrix-coated filter in the two-chambered assembly. Addition of YIGSR to Sertoli cell cultures resulted in cell detachment within 2 to 3 h. In contrast, the irrelevant peptide YIGSE had no detectable effects. The anti-laminin IgG was effective only when added to the chamber in which access was readily available to the basal surfaces of Sertoli cells, but YIGSR was effective when added either to the outer chamber or to the inner chamber. These data were interpreted to indicate that the Sertoli cell barrier generated in the two-chambered assembly allowed a relatively rapid diffusion of YIGSR between chambers, but prevented the rapid equilibration of anti-laminin IgG between compartments. Addition of anti-laminin IgG to the basal, but not to the apical surfaces of Sertoli cells, resulted in more rapid rates of equilibration of [3H]-methoxyinulin and [86Rb]Cl across the Sertoli cell monolayer. This evidence of impairment to the integrity of the barrier was detected prior to the disruption of stress fibers and focal defoliation, but after evidence of dissolution of the circumferal F-actin ring, which occurred within 1 h after addition of anti-laminin IgG. We consider the possibility that a transmembrane link exists between extracellular laminin and cytoskeletal elements which modulates the circumferal F-actin ring. We further postulate that this linkage can influence the nature of tight junctional complexes, and thereby the integrity of the Sertoli cell barrier.

Proteases are implicated in the changes in the Sertoli cell cytoskeleton elicited by follicle-stimulating hormone or by dibutyryl cyclic AMP.

Follicle-stimulating hormone (FSH) or dibutyryl cyclic AMP (dbcAMP) elicits striking morphological changes in Sertoli cells in culture in serum-free medium, resulting in a transition from an epithelial type of cell association pattern to that of an astrocytic or fibroblast-like cell, with attenuated cytoplasmic extensions between cells, and with diminished F-actin stained stress fibers. These responses of Sertoli cells do not occur in the presence of normal untreated serum, but they do take place in the presence of acid-treated serum which is depleted of antiproteases. The addition of alpha 2-macroglobulin to serum-free medium or to antiprotease-depleted serum resulted in the blockage of morphological responses of Sertoli cells to FSH or to dbcAMP. Changes in pattern of arrangements of F-actin in Sertoli cells in culture, which occur in response to FSH or to dbcAMP, were also prevented by the presence of alpha 2-macroglobulin. Thus, the diminution in bundles of F-actin containing stress fibers, which otherwise takes place in Sertoli cells stimulated by FSH or by dbcAMP, did not occur in cells in culture in the presence of alpha 2-macroglobulin, in the presence or absence of acid-treated serum. The inhibiting effects of dbcAMP on the migration of Sertoli cells in serum-free medium became nondetectable in medium containing normal untreated serum, but remained evident in Sertoli cells in culture in medium containing acid-treated serum depleted of antiproteases. Addition of alpha 2-macroglobulin blocked the inhibitory effects of dbcAMP on Sertoli cell migration. Similarly, the presence of alpha 2-macroglobulin prevented the inhibitory effects of dbcAMP on the contractility of TM4 cells which had been embedded in collagen type-I and incubated in serum-free medium. We discuss the possibility that cellular proteases may be implicated in the disintegration of microfilament bundles, either by favoring depolymerization of actin filaments; by facilitating breakage of the link of the transmembrane molecular assembly between cytoskeletal extracellular matrix components; or by catalyzing a disruption of the modular organization of one or more of the actin cross-linking proteins. By inference, we postulate that morphological responses of Sertoli cells to FSH require the activation of cellular proteases for one or more of these reactions, and that alpha 2-macroglobulin blocks the Sertoli cell morphological responses to FSH by inhibiting the proteases involved.

Competition between cell-substratum interactions and cell-cell interactions.

Clusterin, a glycoprotein which elicits the aggregation of a wide variety of cells (Fritz, I. B., and Burdy, K.:J. Cell Physiol., 140:18-28, 1989), has been utilized to investigate some of the factors modulating the competition between cell-substratum interactions and cell-cell interactions. We compared the responses to clusterin by anchorage-independent cells (erythrocytes) with those by anchorage-dependent TM4 cells (a cell line derived from neonatal mouse testis cells). Cells were maintained in culture in the presence of various substrata chosen to enhance cell-substratum interactions (laminin-coated wells), or to diminish cell-substratum interactions (agarose-coated wells). Results obtained showed that the aggregation of erythrocytes elicited by clusterin was independent of the nature of the substratum. In contrast, clusterin addition resulted in aggregation of anchorage-dependent TM4 cells only when TM4 cell-substratum interactions were weak. Thus, clusterin did not aggregate TM4 cells plated upon a laminin substratum, but readily aggregated TM4 cells plated upon an agarose-coated substratum, independent of the sequence of addition of cells and clusterin to the culture dish. We utilized YIGSR, a peptide which competes with laminin for laminin receptors, to determine the possible role of laminin receptors on TM4 cells in the competition between cell-substratum interactions and cell-cell interactions. The presence of YIGSR did not alter responses of erythrocytes to clusterin under all conditions examined. In contrast, the responses of TM4 cells to clusterin were greatly changed. YIGSR addition resulted in the inhibition of aggregation of TM4 cells otherwise elicited by clusterin. YIGSR also prevented attachment of TM4 cells to a laminin-coated surface, but this was reversed by the presence of clusterin. We discuss the possible roles of clusterin and laminin in altering the balance in the competition between cell to cell interactions and cell to substratum interactions.

Inhibition of cell-cell adhesion and morphogenesis of Dictyostelium by carnitine.

Carnitine (gamma-trimethylammonium beta-hydroxy-butyric acid) possesses the novel property of preventing cell aggregation elicited by clusterin or by fibrinogen (I.B. Fritz and K. Burdzy, J. Cell. Physiol., 140:18-28 [1989]). In investigations reported here, we show that carnitine also affects cell-cell adhesion in Dictyostelium discoideum, a cellular slime mold whose cells interact in specific and complex manners during discrete stages of development. Two types of cell adhesion systems sequentially appear on the surface of developing Dictyostelium cells, involving the surface glycoprotein gp24 which mediates EDTA-sensitive binding sites, and the surface glycoprotein gp80 which mediates the EDTA-resistant binding sites. Addition of increasing concentrations of D(+)-carnitine and L(-)-carnitine resulted in a progressive inhibition of both the EDTA-sensitive binding sites and the EDTA-resistant binding sites of Dictyostelium cells at different stages of development. In contrast, comparable or higher concentrations of choline, acetyl-beta-methylcholine, or deoxycarnitine had no detectable effects on cell aggregation. Concentrations of carnitine required for 50% inhibition of EDTA-resistant adhesion sites were found to be dependent upon levels of gp80 expressed by Dictyostelium, with greatest inhibition by carnitine of reassociation of cells containing the lowest levels of gp80. Removal of carnitine from cells by washing resulted in the rapid restoration of the ability of Dictyostelium to form aggregates and to resume normal development. We discuss possible mechanisms by which carnitine inhibits the aggregation of cells.

Clustering of erythrocytes by fibrinogen is inhibited by carnitine: evidence that sulfhydryl groups on red blood cell membranes are involved in carnitine actions.

Carnitine is bound by intact red blood cells, by red blood cell ghosts, and by glutaraldehyde-fixed human erythrocytes in a non-saturable, temperature-dependent manner. Binding of carnitine by these preparations is blocked by sulfhydryl reagents. Incubation or preincubation of red blood cell preparations with carnitine inhibits the aggregation of erythrocytes otherwise elicited by fibrinogen. Identical effects are obtained with red blood cell ghosts. In contrast, choline, even at high concentrations, is inactive in preventing the aggregation of erythrocytes. We discuss possible mechanisms by which carnitine favors the dispersion of red blood cells, and we present data indicating that sulfhydryl groups on erythrocyte membranes are required to permit these carnitine actions to be manifested.

Secretion of latent type IV procollagenase and active type IV collagenase by testicular cells in culture.

Testicular peritubular myoid cells, which have properties similar to those of vascular smooth-muscle cells, secrete a variety of metalloproteinases when maintained in culture in a chemically defined medium. The predominant metalloproteinases secreted were identified as latent type IV procollagenases having molecular masses of 72 kDa and 75 kDa, as detected in Western immunoblots with specific antibodies against type IV procollagenase. When peritubular cells were stimulated by dibutyryl cyclic AMP, forskolin or cholera toxin, they secreted increased amounts of type IV procollagenase. However, little if any of the active type IV collagenase, having a lower molecular mass of 66 kDa, could be detected under these conditions. Addition of low concentrations of cytochalasin D to peritubular cells in monoculture resulted in conversion of the latent type IV collagenase into its active form, assessed with antibody-specificity studies and by the appearance of the 66 kDa protein. In contrast, Sertoli cells in culture did not manifest an increased conversion of type IV procollagenase into type IV collagenase in the presence of cytochalasin D, even though cytochalasin D addition invariably resulted in a disruption of the microfilament assembly in each of these gonadal somatic cell populations. When peritubular cells were co-cultured with Sertoli cells, addition of cytochalasin D no longer resulted in formation of increased amounts of the active form of type IV collagenase. Sertoli cells and peritubular cells each secreted a tissue inhibitor of metalloproteinase type 2, detected with a specific antibody in a Western immunoblot to have a molecular mass of 21 kDa. We conclude that cytochalasin D acts on mesenchymal-type peritubular cells, but not on epithelial-type Sertoli cells, to enhance the conversion of latent type IV procollagenase into active type IV collagenase. This conversion of type IV procollagenase into type IV collagenase by peritubular cells was inhibited by factor(s) secreted by Sertoli cells. Interactions between Sertoli cells and peritubular cells are postulated to modulate net proteinase activities in discrete regions of the testis.

High-resolution scanning electron micrographs of freeze-cracked cells in testes from normal and irradiated rats at different stages of gonadal development.

Newly developed techniques in high-resolution scanning electron microscopy (SEM) and for tissue-processing procedures have been applied to an investigation of structures of various cells in rat testes at different stages of gonadal maturation. A series of high-resolution SEM micrographs are presented which survey the surfaces of different types of testis cells during normal development, and which also illustrate ultrastructural features of some of their intracellular organelles. In addition, a series of high-resolution SEM micrographs are presented which compare the structural features of Sertoli cells in normal testes with those in germ-cell-depleted testes obtained from rats killed at varying times after having been irradiated in utero. We describe our observations on the structural properties of surfaces and intracellular organelles in Sertoli cells, Leydig cells, peritubular myoid cells, and some classes of germinal cells. We also consider the possible role of Sertoli cell apical cytoplasmic processes in lumen formation. Similarities are pointed out between the structure of germ-cell-depleted testes, resulting from irradiation in utero, and the structure of germ-cell-depleted testes in seasonal breeders during periods of involution. Finally, we discuss advantages and disadvantages of methods employed to reveal the fine structure of intracellular organelles in cells of the testis.

Sertoli cells in culture secrete paracrine factor(s) that inhibit peritubular myoid cell proliferation: identification of heparinoids as likely candidates.

Conditioned medium from Sertoli cells, prepared from testes of 20-day-old rats, contains component(s) that inhibit the incorporation of [3H]-thymidine into DNA of peritubular myoid cells (PMC) and inhibit the proliferation of PMC. These components are trypsin-resistant, heat-stable compounds having a molecular weight less than 30,000. The active inhibitory components in Sertoli cell conditioned medium are inactivated by treatment with heparinase, but not by treatment with hyaluronidase or chondroitin sulfate lyases. Addition of heparin or heparan sulfate results in inhibition of DNA synthesis by PMC in a dose-dependent manner, whereas other glycosaminoglycans (GAGs) examined (hyaluronic acid, keratan sulfate, and chondroitin sulfate) have no detectable effects. Heparin and heparan sulfate are unique among GAGs tested in inhibiting the characteristic multilayer growth pattern of PMC following the attainment of confluence in serum-rich medium. On the basis of these and other data presented, it is concluded that heparin and other heparin-like GAGs synthesized by Sertoli cells are implicated in the modulation of growth of PMC in vitro during co-culture. It is postulated that heparin may play a similar role in maintaining the quiescent peritubular myoid cell phenotype in vivo.

Transforming growth factor-beta and platelet-derived growth factor synergistically stimulate contraction by testicular peritubular cells in culture in serum-free medium.

We report investigations on factors influencing contractility by testicular peritubular cells (PC) maintained in culture in a three-dimensional collagen gel system, and the behavior of PC in culture on a two-dimensional system. At low and moderate cell densities, PC embedded in collagen gels in serum-free Eagle's minimal essential medium (MEM) have a lesser degree of contractility than PC in culture in MEM containing calf serum. The contractility by PC, measured by determining changes in diameter of the collagen gel, was increased by addition of transforming growth factor-beta (TGF-beta) to serum-free MEM, and this was further enhanced by supplementing the medium with platelet-derived growth factor (PDGF). In the absence of TGF-beta, however, PDGF had no detectable effects on PC contractility. Other growth factors examined (epidermal growth factor, insulin, and fibroblast growth factor) did not influence the degree of contractility of PC in serum-free MEM in the presence or absence of TGF-beta. PC maintained in MEM supplemented with platelet-poor serum (PPS) have a lesser degree of contractility than their counterparts in MEM containing 2.5% calf serum. The addition of TGF-beta and PDGF to PPS-supplemented MEM restored contractility by PC to a level comparable to that observed by PC in MEM containing complete serum. The addition of nonpurified bovine serum albumin (BSA) to MEM greatly increased PC contractility. By contrast, highly purified BSA had no such effect, suggesting that one or more components adsorbed to the impure BSA was implicated. Polyclonal antibody against fibronectin did not influence the contractility of PC in collagen gels in the presence or absence of serum. Antiserum against TGF-beta partially blocked the enhancement of contractility of PC in MEM containing non-purified BSA. In PC plated on top of a collagen gel lattice, the attachment, spreading, and cell shape were greatly influenced by the presence of TGF-beta and PDGF, both singly and together. Data presented are interpreted to indicate that effects elicited by serum on the properties of PC in culture, and on the contractility of PC, can be attributed in part to the combined influences of TGF-beta and PDGF in serum.