Patricia A. D'Amore

Cell-cell Interactions in the Regulation of Capillary Growth and Stability

New vessels form de novo (vasculogenesis) or from pre-existing vessels (angiogenesis) in a process that involves the interaction between 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 or SMC. We have previously demonstrated that transforming growth factor b (TGFb) induces the differentiation of C3H/10T1/2 (10T1/2) mesenchymal cells toward a SMC/pericyte lineage. We tested the hypothesis that TGFb not only induces SMC differentiation, but stabilizes capillary-like structures in a three-dimensional (3D) model of in vitro angiogenesis. 10T1/2 cells and EC in Matrigel™ were used to establish cocultures that formed cord structures that were reminiscent of new capillaries in vivo. Cord formation was initiated within 2-3 hr after plating and continued through 18 hr after plating. In longer cocultures the cord structures disassembled and formed aggregates. 10T1/2 cell expression of proteins associated with the SMC/pericyte lineage, such as smooth muscle actin (SMA) and NG2 proteoglycan, were upregulated in these 3D cocultures. Application of neutralizing reagents specific for TGFb blocked cord formation and inhibited expression of SMA and NG2 in the 10T1/2 cells. These results indicate that TGFb 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™.

To further understand the mechanism through which pericytes might stabilize vessels, we tested the hypothesis that differentiation of mesenchymal cells to pericytes/SMC is accompanied by vascular endothelial growth factor (VEGF) expression, which mediates EC survival. Coculture of EC and 10T1/2 cells (multipotent mesenchymal cells), which led to 10T1/2 differentiation to a pericyte fate, accompanied the induction of VEGF expression. The increase in VEGF depended on contact between EC and 10T1/2 cells. The VEGF induction was due to the action of TGFb, as coculture of EC smad3-/- mouse embryo fibroblasts did not yield elevated VEGF. A majority of the VEGF in the cocultures was cell- and/or matrix-associated via heparan sulfate proteoglycans; treatment of the cells with high salt, protamine, heparin or suramin led to significant VEGF release. Inhibition of VEGF in the cocultures led to a 75% increase in EC apoptosis, relative to coculture controls, indicating a role for VEGF in EC survival. These observations indicate that differentiated pericytes produce VEGF, which acts in a juxtacrine/paracrine manner as a survival and/or stabilizing factor for EC in newly formed vessels.

The capillary-like tubes formed in Matrigel, as described above, form rapidly (within 6 hr) but the structures are short-lived (18-24 hr). During the past year we worked to develop a novel 3D coculture system that will produce capillary-like structures that are stable over extended time periods, which will allow us to test the stabilizing effect of pericytes on EC forming capillaries. For these studies, we induced apoptosis in 10T1/2 cells in co-culture and looked at the consequences on the remaining bovine retinal endothelial cell (BREC)-capillary-like structures (CLS). Since a control in which EC cells alone form CLS was necessary, we tested various matrices for the basis of the 3D gel and found that in Vitrogen (collagen I and III), BREC form CLS in the absence of 10T1/2 cells if appropriate growth factors are added. In the absence of added growth factors, EC underwent cell death. Addition of VEGF induced CLS. Addition of basic fibroblast growth factor (bFGF) along with VEGF resulted in the formation of a vascular network similar to that formed in BREC-10T1/2 cells co-cultures. Thus, VEGF seems to be the differentiation factor that induces capillary-formation, whereas bFGF acts mainly as a proliferation factor, to expand a complex network.

We next determined if CLS formed in BREC-10T1/2 cell co-cultures were more stable than CLS formed from solo-cultures of BREC, and if the 10T1/2 cells/pericytes could protect the EC from pro-apoptotic stimuli. TGFb1 had been shown to induce endothelial apoptosis; the addition of exogenous TGFb1 induced cell death in BREC in solo-culture but not in co-culture with 10T1/2 cells.

It seemed clear that the actions of TGFb depend on the context in which the factors act. TGFb1 was known to be produced and activated in co-cultures where we suspected it mediated aspects of vessel stabilization. On the other hand, we showed that addition of TGFb1 to BREC (in the absence of 10T1/2 cells-pericytes) induced EC apoptosis. Thus, it appears that for TGFb1 to be involved in the dynamic process of capillary-stabilization it has to be balanced by another factor(s). Current studies are aimed at elucidating what other factors, if any, are involved in vessel stabilization

Role of Vascular Endothelial Growth Factor (VEGF) Isoforms in Vascular Development and Growth

VEGF is critical for normal development of the vascular system in mouse embryos. Murine VEGF exists as at least three homodimeric isoforms as a result of mRNA alternative splicing; they include VEGF120, VEGF164 and VEGF188. These three isoforms have different affinities for heparan sulfate as well as for the three known VEGF receptors: Flk-1, Flt-1, and neuropilin-1, suggesting that differential receptor binding and/or extracellular localization may allow the different VEGF isoforms to play distinct roles in vascular development. To address the possibility of distinct functions for the individual isoforms, we have generated lines of mice that each expressed single VEGF isoforms. Mice that expressed only VEGF120 had the most dramatic phenotype. Though they survived to term, they died immediately after birth, presumably due to pulmonary dysfunction. VEGF188 mice lived to term and survived through adulthood but had a variety of non-lethal vascular defects. We used these animals to study the role of various VEGF isoforms in tissue and organ development.

Retina

Retinal vascular development was studied in mice expressing VEGF120, VEGF164, or VEGF188. VEGF164/164 mice appeared normal and healthy, and had normal retinal angiogenesis. In contrast, VEGF 120/120 mice exhibited severe defects in outgrowth and patterning of retinal vessels. VEGF188/188 mice displayed normal outgrowth of retinal veins, but impaired retinal arterial development. Neuropilin-1, a receptor for VEGF164, was predominantly expressed in developing retinal arterioles.

Bone

As VEGF120/20 mice survived until after birth, they offered an attractive opportunity to study the role of VEGF during bone development. Analysis of cartilage vascularization of wild-type and VEGF120/120 mice identified two key differences. First, at embryonic day 14.5 there was a lack of blood vessels surrounding the cartilage, and second, at E15.5 there was a lack of vascular invasion into hypertrophic cartilage. At E13.5, strong VEGF expression was detected in the perichondrium and the surrounding tissues. This expression, in combination with the known VEGF expression in hypertrophic chondrocytes at E15.5, provided a new understanding of cartilage vascularization and defined a role for VEGF in the developmentally regulated patterning of skeletal vascularity. The sequence of appearance of cartilage differentiation markers identified a delay at E14.5 of cartilage maturation in VEGF120/120 mice, further suggesting a role for VEGF in maintaining the normal chondrocyte differentiation program. Finally, reduction in both endochondral and membranous bone formation in VEGF120/120 mice suggested a role for VEGF in osteoblast function. Enhanced bone formation in calvarial organ culture after treatment with VEGF demonstrated that VEGF has a stimulatory effect on osteoblastic activity during skeletal development. Thus, we have described a novel two-step model for VEGF controlled vascularization of skeletal elements and have provided evidence for a role of VEGF in osteoblastic regulation.

Lung

Lung development was also defective in VEGF120 mice. Lung vessel development was studied by scanning electron microscopy of Mercox casts of lung vasculature. Airway and air-blood barrier development was analyzed by light microscopy, transmission electron microscopy, immunohistochemistry and morphometry. In all VEGF120/120 fetuses and pups, lung vascular casts were smaller and less dense compared to those of 120/+ and wild type (wt) littermates. Although in all three genotypes the generation count of preacinar vessels was similar, the most peripheral vessels had thicker profiles and were more widely separated in 120/120 fetuses of all ages, compared to 120/+ and wt littermates. In addition, 120/120 animals had fewer air-blood barriers and a decreased air-parenchyma ratio compared to 120/+ and wt littermates. These data indicate that the absence of VEGF164 and 188 isoforms impaired lung microvascular development and delayed airspace maturation, and indicate an essential role for heparin-binding VEGF isoforms in normal lung development.