PDGF, TGF-beta, and heterotypic cell-cell interactions mediate endothelial cell-induced recruitment of 10T1/2 cells and their differentiation to a smooth muscle fate

We aimed to determine if and how endothelial cells (EC) recruit precursors of smooth muscle cells and pericytes and induce their differentiation during vessel formation. Multipotent embryonic 10T1/2 cells were used as presumptive mural cell precursors. In an under-agarose coculture, EC induced migration of 10T1/2 cells via platelet-derived growth factor BB. 10T1/2 cells in coculture with EC changed from polygonal to spindle-shaped, reminiscent of smooth muscle cells in culture. Immunohistochemical and Western blot analyses were used to examine the expression of smooth muscle (SM)-specific markers in 10T1/2 cells cultured in the absence and presence of EC. SM-myosin, SM22alpha, and calponin proteins were undetectable in 10T1/2 cells cultured alone; however, expression of all three SM-specific proteins was significantly induced in 10T1/2 cells cocultured with EC. Treatment of 10T1/2 cells with TGF-beta induced phenotypic changes and changes in SM markers similar to those seen in the cocultures. Neutralization of TGF-beta in the cocultures blocked expression of the SM markers and the shape change. To assess the ability of 10T1/2 cells to contribute to the developing vessel wall in vivo, prelabeled 10T1/2 cells were grown in a collagen matrix and implanted subcutaneously into mice. The fluorescently marked cells became incorporated into the medial layer of developing vessels where they expressed SM markers. These in vitro and in vivo observations shed light on the cell-cell interactions that occur during vessel development, as well as in pathologies in which developmental processes are recapitulated.

An activated form of transforming growth factor beta is produced by cocultures of endothelial cells and pericytes

Using an in vitro coculture system to mimic the interactions between the cells of the vessel wall, we have previously shown that pericytes and smooth muscle cells (SMC) inhibit the growth of capillary endothelial cells (EC). We have undertaken studies to determine the mechanism of this inhibition. Using conditioned media and affinity-purified antibodies to transforming growth factor beta (TGF-beta), we now demonstrate that activated TGF-beta produced in these cocultures mediates EC growth inhibition. No inhibitory activity was detected when media conditioned by individual cultures of EC, SMC, or pericytes were examined for their effect on EC growth. In contrast, media conditioned by cocultures of EC-SMC and EC-pericytes inhibited EC proliferation to the same degree as the coculture itself. Immunoadsorption of coculture-derived conditioned media with antibodies to TGF-beta eliminated the inhibitory activity. Acid activation of serum-free media conditioned by any of the cells cultured alone yielded inhibitory activity, whereas activation of coculture conditioned media did not increase its inhibitory activity. Addition of anti-TGF-beta neutralizing antibodies to cocultures blocked the pericyte-mediated EC growth inhibition. These results indicate that latent TGF-beta is produced by these cells and it is activated by a mechanism that requires contact between the two cell types.

Inhibition of capillary endothelial cell growth by pericytes and smooth muscle cells

Morphological studies of developing capillaries and observations of alterations in capillaries associated with pathologic neovascularization indicate that pericytes may act as suppressors of endothelial cell (EC) growth. We have developed systems that enable us to investigate this possibility in vitro. Two models were used: a co-culture system that allowed direct contact between pericytes and ECs and a co-culture system that prevented physical contact but allowed diffusion of soluble factors. For these studies, co-cultures were established between bovine capillary ECs and the following growth-arrested cells (hereafter referred to as modulating cells): pericytes, smooth muscle cells (SMCs), fibroblasts, epithelial cells, and 3T3 cells. The modulating cell type was growth arrested by treatment with mitomycin C before co-culture with ECs. In experiments where cells were co-cultured directly, the effect of co-culture on EC growth was determined by comparing the mean number of cells in the co-cultures to the mean for each cell type (EC and modulating cell) cultured separately. Since pericytes and other modulating cells were growth arrested, any cell number change in co-cultures was due to EC growth. In the co-cultures, pericytes inhibited all EC proliferation throughout the 14-d time course; similar levels of EC inhibition were observed in SMC-EC co-cultures. Co-culture of ECs with fibroblasts, epithelial cells, and 3T3 cells significantly stimulated EC growth over the same time course (30-192% as compared to EC cultured alone). To determine if cell contact was required for inhibition, cells were co-cultured using Millicell chambers (Millipore Corp., Bedford, MA), which separated the cell types by 1-2 mm but allowed the exchange of diffusible materials. There was no inhibition of EC proliferation by pericytes or SMCs in this co-culture system. The influence of the cell ratios on observed inhibition was assessed by co-culturing the cells at EC/pericyte ratios of 1:1, 2:1, 5:1, 10:1, and 20:1. Comparable levels of EC inhibition were observed at ratios from 1:1 to 10:1. When the cells were co-cultured at a ratio of 20 ECs to 1 pericyte, inhibition of EC growth at 3 d was similar to that observed at other ratios. However, at higher ratios, the inhibition diminished so that by the end of the time course the co-cultured ECs were growing at the same rate as the controls. These results suggest that pericytes and SMCs can modulate EC growth by a mechanism that requires contact or proximity. We postulate that similar interactions may operate to modulate vascular growth in vivo.

Pericytes in the microvasculature

Pericytes, also known as Rouget cells or mural cells, are associated abluminally with all vascular capillaries and post-capillary venules. Differences in pericyte morphology and distribution among vascular beds suggest tissue-specific functions. Based on their location and their complement of muscle cytoskeletal proteins, pericytes have been proposed to play a role in the regulation of blood flow. In vitro studies demonstrating the contractile ability of pericytes support this concept. Pericytes have also been suggested to be oligopotential and have been reported to differentiate into adipocytes, osteoblasts and phagocytes. The mechanisms involved in vessel formation have yet to be elucidated but observations indicate that the primordial endothelium can recruit undifferentiated mesenchymal cells and direct their differentiation into pericytes in microvessels, and smooth muscle cells in large vessels. Communication between endothelial cells and pericytes, or their precursors, may take many forms. Soluble factors such as platelet-derived growth factor and transforming growth factors-beta are likely to be involved. In addition, physical contact mediated by cell adhesion molecules, integrins and gap junctions appear to contribute to the control of vascular growth and function. Development of culture methods has allowed some functions of pericytes to be directly examined. Co-culture of pericytes with endothelial cells leads to the activation of transforming growth factor-beta, which in turn influences the growth and differentiation of the vascular cells. Finally, the pericyte has been implicated in the development of a variety of pathologies including hypertension, multiple sclerosis, diabetic microangiopathy and tumor vascularization.

Inhibition of vascular endothelial growth factor prevents retinal ischemia-associated iris neovascularization in a nonhuman primate

OBJECTIVE

To determine if the angiogenic peptide vascular endothelial growth factor (VEGF) is required for retinal ischemia-associated iris neovascularization in a nonhuman primate.

METHODS

Laser retinal vein occlusion was used to produce retinal ischemia in 16 eyes of eight animals (Macaca fascicularis). Eyes were randomized to treatment every other day with intravitreal injections of either a neutralizing anti-VEGF monoclonal antibody or a control monoclonal antibody of the same isotype. Serial iris fluorescein angiograms were assessed using a standardized grading system and masked readers. Retinal VEGF and placental growth factor expression were assessed by Northern blotting. The specificity of the antibodies was determined in capillary endothelial cell proliferation assays prior to intravitreal injection.

RESULTS

Zero of eight eyes receiving the neutralizing anti-VEGF antibodies developed iris neovascularization. Five of eight control antibody-treated eyes developed iris neovascularization. The difference was statistically significant (P = .03). Intravitreal antibody injection did not impair the ability of the ischemic retina to increase VEGF messenger RNA expression. The anti-VEGF antibodies specifically inhibited VEGF-driven capillary endothelial cell proliferation in vitro.

CONCLUSION

These data demonstrate that VEGF is required for iris neovascularization in an adult nonhuman primate eye. The inhibition of VEGF is a new potential therapeutic strategy for the treatment of ocular neovascularization.

Growth factors are released by mechanically wounded endothelial cells

Growth factors may be required at sites of mechanical injury and normal wear and tear in vivo, suggesting that the direct action of mechanical forces on cells could lead to growth factor release. Scraping of cells from the tissue culture substratum at 37 degrees C was used to test this possibility. We show that scraping closely mimics in vitro both the transient plasma membrane wounds observed in cells subject to mechanical forces in vivo (McNeil, P. L., and S. Ito. 1989. Gastroenterology. 96:1238-1248) and the transient plasma membrane wounds shown here to occur in endothelial cells under normal culturing conditions. Scraping of endothelial cells from the culturing substratum released into the culture medium a potent growth-promoting activity for Swiss 3T3 fibroblasts. Growth-promoting activity was released rapidly (within 5 min) after scraping but was not subsequently degraded by the endothelial cells for at least 24 h thereafter. A greater quantity of growth-promoting activity was released by cells scraped 4 h after plating than by those scraped 4 or 7 d afterwards. Thus release is not due to scraping-induced disruption of extracellular matrix. Release was only partially cold inhibitable, was poorly correlated with the level of cell death induced by scraping, and did not occur when cells were killed with metabolic poisons. These results suggest that mechanical disruption of plasma membrane, either transient or permanent, is the essential event leading to release. A basic fibroblast growth factor-like molecule and not platelet-derived growth factor appears to be partially responsible for the growth-promoting activity. We conclude that one biologically relevant route of release of basic fibroblast growth factor, a molecule which lacks the signal peptide sequence for transport into the endoplasmic reticulum, could be directly through mechanically induced membrane disruptions of endothelial cells growing in vivo and in vitro.

Vascular endothelial growth factor and its receptors

Vascular endothelial growth factor (VEGF) is a prime regulator of endothelial cell proliferation, angiogenesis, vasculogenesis and vascular permeability. Its activity is mediated by the high affinity tyrosine kinase receptors, KDR/Fik-1 and Fit-1. In this article, recently discovered structural, molecular and biological properties of VEGF are described. Among the topics discussed are VEGF and VEGF receptor structure and bioactivity, the regulation of VEGF expression, the role of VEGF and its receptors in vascular development, and the involvement of VEGF and its receptors in normal and pathological (ocular and tumor) angiogenesis.

Synthesis and secretion of vascular permeability factor/vascular endothelial growth factor by human retinal pigment epithelial cells

The presence of the secreted angiogenic molecule, vascular permeability factor/vascular endothelial growth factor, was investigated in human retinal pigment epithelial cells. Northern and in situ hybridization analysis of cultured cells identified vascular permeability factor/vascular endothelial growth factor mRNA. Western analysis of cell lysates identified a 42 kD protein that comigrated with human recombinant vascular permeability factor/vascular endothelial growth factor. Immunoassay measurements detected vascular permeability factor/vascular endothelial growth factor protein in cell lysates and conditioned media in vitro and in cell lysates isolated directly from post-mortem eyes. These data demonstrate that human retinal pigment epithelial cells can synthesize the secreted angiogenic peptide vascular permeability factor/vascular endothelial growth factor in vitro and in situ. The production and secretion of this factor by human retinal pigment epithelial cells may be important in the pathogenesis of ocular neovascularization.

Pericyte production of cell-associated VEGF is differentiation-dependent and is associated with endothelial survival

Pericytes have been suggested to play a role in regulation of vessel stability; one mechanism for this stabilization may be via pericyte-derived vascular endothelial growth factor (VEGF). To test the hypothesis that differentiation of mesenchymal cells to pericytes/smooth muscle cells (SMC) is accompanied by VEGF expression, we used endothelial cell (EC) and mesenchymal cell cocultures to model cell-cell interactions that occur during vessel development. Coculture of EC and 10T1/2 cells, multipotent mesenchymal cells, led to induction of VEGF expression by 10T1/2 cells. Increased VEGF expression was dependent on contact between EC-10T1/2 and was mediated by transforming growth factorbeta (TGFbeta). A majority of VEGF produced in coculture was cell- and/or matrix-associated. Treatment of cells with high salt, protamine, heparin, or suramin released significant VEGF, suggesting that heparan sulfate proteoglycan might be sequestering some of the VEGF. Inhibition of VEGF in cocultures led to a 75% increase in EC apoptosis, indicating that EC survival in cocultures is dependent on 10T1/2-derived VEGF. VEGF gene expression in developing retinal vasculature was observed in pericytes contacting newly formed microvessels. Our observations indicate that differentiated pericytes produce VEGF that may act in a juxtacrine/paracrine manner as a survival and/or stabilizing factor for EC in microvessels.

Microvascular pericytes contain muscle and nonmuscle actins

We have affinity-fractionated rabbit antiactin immunoglobulins (IgG) into classes that bind preferentially to either muscle or nonmuscle actins. The pools of muscle- and nonmuscle-specific actin antibodies were used in conjunction with fluorescence microscopy to characterize the actin in vascular pericytes, endothelial cells (EC), and smooth muscle cells (SMC) in vitro and in situ. Nonmuscle-specific antiactin IgG stained the stress fibers of cultured EC and pericytes but did not stain the stress fibers of cultured SMC, although the cortical cytoplasm associated with the plasma membrane of SMC did react with nonmuscle-specific antiactin. Whereas the muscle-specific antiactin IgG failed to stain EC stress fibers and only faintly stained their cortical cytoplasm, these antibodies reacted strongly with the fiber bundles of cultured SMC and pericytes. Similar results were obtained in situ. The muscle-specific antiactin reacted strongly with the vascular SMC of arteries and arterioles as well as with the perivascular cells (pericytes) associated with capillaries and post-capillary venules. The non-muscle-specific antiactin stained the endothelium and the pericytes but did not react with SMC. These findings indicate that pericytes in culture and in situ possess both muscle and nonmuscle isoactins and support the hypothesis that the pericyte may represent the capillary and venular correlate of the SMC.

Differential expression of VEGF isoforms in mouse during development and in the adult

Vascular endothelial growth factor (VEGF), a factor that is critical for development of the vascular system in mouse embryos, exists as at least three isoforms, VEGF120, VEGF164, and VEGF188. The isoforms have different affinities for heparan sulfate as well as for the three known VEGF receptors, VEGFR-1 (Flt-1), VEGFR-2 (Flk-1), and neuropilin-1, suggesting that different VEGF isoforms may play distinct roles in vascular development. To determine whether there are differences in the organ-specific expression patterns that would support this concept, we used a quantitative RNase protection assay (RPA) to determine the distribution of different VEGF isoform mRNA in developing and adult mouse organs. Results revealed that the ratios of the three VEGF isoforms changed during organ development and that adult organs expressed different levels of the three VEGF isoforms. Because the lung expressed the highest levels of VEGF188 isoform, we used VEGF isoform-specific in situ hybridization in the developing lung and determined that type II alveolar epithelial cells were expressing high levels of VEGF188 mRNA. Finally, targeted exon deletion of the VEGF gene revealed that mice that developed in the absence of the heparan sulfate binding isoforms VEGF164 and VEGF188, displayed a variety of vascular defects, including abnormal pulmonary vascular development. Our results support the concept that different VEGF isoforms have distinct functions in vascular development.

Endothelial cells modulate the proliferation of mural cell precursors via platelet-derived growth factor-BB and heterotypic cell contact

Embryological data suggest that endothelial cells (ECs) direct the recruitment and differentiation of mural cell precursors. We have developed in vitro coculture systems to model some of these events and have shown that ECs direct the migration of undifferentiated mesenchymal cells (10T1/2 cells) and induce their differentiation toward a smooth muscle cell/pericyte lineage. The present study was undertaken to investigate cell proliferation in these cocultures. ECs and 10T1/2 cells were cocultured in an underagarose assay in the absence of contact. There was a 2-fold increase in bromodeoxyuridine labeling of 10T1/2 cells in response to ECs, which was completely inhibited by the inclusion of neutralizing antiserum against platelet-derived growth factor (PDGF)-B. Antisera against PDGF-A, basic fibroblast growth factor, or transforming growth factor (TGF)-beta had no effect on EC-stimulated 10T1/2 cell proliferation. EC proliferation was not influenced by coculture with 10T1/2 cells in the absence of contact. The cells were then cocultured so that contact was permitted. Double labeling and fluorescence-activated cell sorter analysis revealed that ECs and 10T1/2 cells were growth-inhibited by 43% and 47%, respectively. Conditioned media from contacting EC-10T1/2 cell cocultures inhibited the growth of both cell types by 61% and 48%, respectively. Although we have previously shown a role for TGF-beta in coculture-induced mural cell differentiation, growth inhibition resulting from contacting cocultures or conditioned media was not suppressed by the presence of neutralizing antiserum against TGF-beta. Furthermore, the decreased proliferation of 10T1/2 cells in the direct cocultures could not be attributed to downregulation of the PDGF-B in ECs or the PDGF receptor-beta in the 10T1/2 cells. Our data suggest that modulation of proliferation occurs during EC recruitment of mesenchymal cells and that heterotypic cell-cell contact and soluble factors play a role in growth control during vessel assembly.

The mouse gene for vascular endothelial growth factor. Genomic structure, definition of the transcriptional unit, and characterization of transcriptional and post-transcriptional regulatory sequences

We describe the genomic organization and functional characterization of the mouse gene encoding vascular endothelial growth factor (VEGF), a polypeptide implicated in embryonic vascular development and postnatal angiogenesis. The coding region for mouse VEGF is interrupted by seven introns and encompasses approximately 14 kilobases. Organization of exons suggests that, similar to the human VEGF gene, alternative splicing generates the 120-, 164-, and 188-amino acid isoforms, but does not predict a fourth VEGF isoform corresponding to human VEGF206. Approximately 1. 2 kilobases of 5'-flanking region have been sequenced, and primer extension analysis identified a single major transcription initiation site, notably lacking TATA or CCAT consensus sequences. The 5'-flanking region is sufficient to promote a 7-fold induction of basal transcription. The genomic region encoding the 3'-untranslated region was determined by Northern and nuclease mapping analysis. Investigation of mRNA sequences responsible for the rapid turnover of VEGF mRNA (mRNA half-life, <1 h) (Shima, D. T. , Deutsch, U., and D'Amore, P. A. (1995) FEBS Lett. 370, 203-208) revealed that the 3'-untranslated region was sufficient to trigger the rapid turnover of a normally long-lived reporter mRNA in vitro. These data and reagents will allow the molecular and genetic analysis of mechanisms that control the developmental and pathological expression of VEGF.

Hypoxic induction of vascular endothelial growth factor (VEGF) in human epithelial cells is mediated by increases in mRNA stability

Vessel growth is often associated with ischemia. VEGF, a potent angiogenic factor, has been shown to be induced by low oxygen concentrations. These studies were conducted to investigate the molecular basis of the hypoxia-induced increase in VEGF mRNA. Run-on analysis of VEGF revealed a minimal increase in the rate of gene transcription in a human retinal epithelial cell line grown under hypoxic conditions. Examination of VEGF mRNA stability revealed that the half-life of VEGF transcripts under normoxia was short, 30-45 min, but was dramatically increased to 6-8 h in cells grown under hypoxia. Cobalt chloride, which elevates VEGF and has been suggested to be similar to hypoxia in its mechanism of action, had only a slight effect on decay rate. We postulate that hypoxia-induced increases in mRNA stability provide the sustained increases in VEGF mRNA levels necessary to support a neovascular response.

Demonstration of vasoproliferative activity from mammalian retina

Vasoproliferative activity has been demonstrated in extracts of retinas from human, bovine, and feline sources. These retinal extracts are capable of stimulating (a) proliferation and thymidine uptake of bovine vascular endothelial cells in culture and (b) neovascularization on the chick chorioallantoic membrane. Extracts of skeletal muscle, cardiac muscle, and liver lack similar stimulatory activity. The activity is nondialyzable, stable at 56 degrees C, and inactivated at 100 degrees C. Retinal extracts stimulate the proliferation of corneal fibroblasts but have no effect on the proliferation of vascular smooth muscle cells. Indirect evidence suggests the liberation of a vasoproliferative factor from retina in several ocular disorders. The data in this report represent the first direct demonstration of vasoproliferative activity from mammalian retina.

Endothelial cell mitogens derived from retina and hypothalamus: biochemical and biological similarities

Bovine retina and hypothalamus contain anionic endothelial cell mitogens that display unusual affinities for the negatively charged glycosaminoglycan heparin. Both growth factor activities are acidic polypeptides (pl's of 5.0) as determined by isoelectric focusing and DEAE-affinity chromatography. In spite of their anionic nature, the factors bound to heparin-Sepharose columns with high affinity and could be eluted only at high salt concentrations (0.9-1.1 M NaCl). The affinity of the retina-derived growth factor (RDGF) for heparin permitted a 15,000-fold purification of the mitogen in two steps: heparin-affinity chromatography and size exclusion high-performance liquid chromatography. RDGF and the anionic hypothalamus-derived factor (aHDGF) exhibit three major biochemical similarities including isoelectric point, (pl's of 5.0), heparin affinity (elution at 0.9-1.1 M NaCl) and molecular weight (18,000). Additionally, the two factors display similar biological activities, stimulating the proliferation of capillary and human umbilical vein endothelial and 3T3 cells but not vascular smooth muscle cells. We suggest that RDGF and aHDGF are related if not identical growth factor molecules.

Vascular endothelial growth factor-induced migration of vascular smooth muscle cells in vitro

Angiogenesis is a complex process that includes recruitment and proliferation of mural cells-smooth muscle cells (SMC) and pericytes. Vascular endothelial growth factor (VEGF) has been shown to play an important role in angiogenesis and is an endothelial cell chemoattractant. In addition, certain VEGF isoforms have been implicated in the normal formation of smooth muscle cell-surrounded arteries. Because VEGF's role as a mural cell chemoattractant had not been explored, we examined the ability of VEGF to influence vascular SMC migration in vitro. A Boyden chamber migration assay demonstrated that VEGF (0-100 ng/ml) caused a dose-dependent migration of SMC. VEGF did not cause proliferation of SMC. Reverse transcriptase-polymerase chain reaction analysis demonstrated the presence of both KDR and flt mRNA, two known VEGF receptors, in SMC cultures. Western blot analysis of SMC lysates confirmed these data, revealing bands migrating at approximately 200 kDa and slightly below 200 kDa consistent with KDR and flt. These observations demonstrate that VEGF receptors are present on SMC, and that VEGF can act as an SMC chemoattractant.

Heparin potentiates the action of acidic fibroblast growth factor by prolonging its biological half-life

The mechanism(s) by which heparin influences the biological activities of acidic and basic fibroblast growth factors (aFGF and bFGF) is not completely understood. One mechanism by which heparin could alter the biological activities of aFGF and bFGF is by altering their biological half-lives. We investigated the possibility that heparin potentiates aFGF-induced neurite outgrowth from PC12 cells by prolonging its biological half-life. Under conditions where heparin potentiated aFGF-induced neurite outgrowth, we observed that heparin increased the biological half-life of aFGF from 7 to 39 hr. We determined that greater than 25 hr of exposure to active aFGF was required for induction of neurite outgrowth. If aFGF activity was maintained for greater than 25 hr by periodic readdition of factor, heparin no longer potentiated aFGF-induced neurite outgrowth. These observations strongly suggest that heparin potentiates the activity of aFGF by prolonging its biological half-life. The protease inhibitors hirudin, leupeptin, and pepstatin A did not potentiate aFGF-induced neurite outgrowth, indicating that proteases inhibited by these inhibitors are not responsible for the loss of aFGF activity that we observed. However, aprotinin potentiated aFGF neurite-promoting activity approximately sevenfold, indicating that proteases that are inhibited by aprotinin are at least partially responsible for aFGF inactivation. These observations suggest that heparin regulates the activity of aFGF by regulating its proteolytic degradation, thereby regulating its biological half-life.

Culture of retinal capillary cells using selective growth media

Capillary fragments are isolated from the microvasculature of the bovine retina using limited collagenase digestion and sieving. The endothelial cells obtained from these capillary fragments are cultured in media containing platelet-poor plasma supplemented with retinal extract, which contains an endothelial cell mitogen. These cells form typical monolayers in vitro, contain Factor VIII antigen, and can be passaged serially. The pericytes obtained from these capillary fragments are cultured in media containing calf serum. These cells can also be passaged serially, are not contact-inhibited, and do not stain for Factor VIII antigen. Cultures of capillary cells such as these will allow the study of the individual cells from specific microvascular beds involved in various normal and pathological processes.

TGF beta is required for the formation of capillary-like structures in three-dimensional cocultures of 10T1/2 and endothelial cells

New vessels form de novo (vasculogenesis) or from pre-existing vessels (angiogenesis) in a process that involves the interaction of endothelial cells (EC) and pericytes/smooth muscle cells (SMC). One basic component of this interaction is the endothelial-induced recruitment, proliferation and subsequent differentiation of pericytes and SMC. We have previously demonstrated that TGF beta induces the differentiation of C3H/10T1/2 (10T1/2) mesenchymal cells toward a SMC/pericyte lineage. The current study tests the hypothesis that TGF beta not only induces SMC differentiation but stabilizes capillary-like structures in a three-dimensional (3D) model of in vitro angiogenesis. 10T1/2 and EC in Matrigel were used to establish cocultures that form cord structures that are reminiscent of new capillaries in vivo. Cord formation is initiated within 2-3 h after plating and continues through 18 h after plating. In longer cocultures the cord structures disassemble and form aggregates. 10T1/2 expression of proteins associated with the SMC/pericyte lineage, such as smooth muscle alpha-actin (SMA) and NG2 proteoglycan, are upregulated in these 3D cocultures. Application of neutralizing reagents specific for TGF beta blocks cord formation and inhibits expression of SMA and NG2 in the 10T1/2 cells. We conclude that TGF beta mediates 10T1/2 differentiation to SMC/pericytes in the 3D cocultures and that association with differentiated mural cells is required for formation of capillary-like structures in Matrigel.