One of the promises of modem molecular biology has been the opportunity to use genetically modified human cells in a patient to permanently restore inborn errors of metabolism. Although it has been possible to introduce genes into mammalian cells and to control their expression, it has proven difficult to introduce mammalian cells as carriers of the modified genetic information into hosts. The successful implantation of selective cells cannot be achieved without adequate vascular support, an essential step toward integration and reconstitution of a new biological function. Although a partial solution to this problem has been found by inducing specific site-directed neovessel formation using heparinbinding growth factor 1 (HBGF-1) adsorbed to a collagen matrix, these implants function for only a short period (weeks). We now report the formation of organoid neovascular structures using polytetrafluoroethylene fibers coated with collagen and HBGF-1 implanted in the peritoneal cavity of the rat. The organoid structures contained readily visible vascular lumina and nonvascular structures that resemble nerve tissue. It was also possible to demonstrate that the vascular system on the implant is continuous with the vascular tree of the host. This feature was used to demonstrate that the organoid structures are capable of sustaining the biological function of implanted normal rat hepatocytes over long periods of time (months) in the homozygous Gunn rat, thereby facilitating future applications involving the delivery of new genetic information.Angiogenesis involves the orderly migration and proliferation of blood vessels and occurs during development (1,2). Although angiogenesis is an infrequent event in the adult, mainly associated with wound and fracture repair, exceptions are found in the female reproductive system, where this process occurs in the follicle during development, in the corpus luteum during ovulation, and in the placenta during pregnancy (1, 2). These physiologic, specific periods of angiogenesis are relatively brief and highly regulated in contrast to the pathologic angiogenic events associated with tumor growth, diabetic retinopathy, and other disorders. Because the endothelial cell is considered to be the primary cell involved in all forms of angiogenesis, efforts have concentrated on the identity of polypeptide factors that control endothelial cell proliferation (1,2). Indeed, the heparinbinding growth factor (HBGF) family of polypeptides has gained general acceptance as initiators of angiogenesis especially during development (1)(2)(3)(4)(5)(6). This gene family includes the prototypes HBGF-1 (acidic fibroblast growth factor), HBGF-2 (basic fibroblast growth factor), and three additional HBGF-like structures, hst/KS, int-2, and fibroblast growth factor 5 (2, 5). HBGF-1 and HBGF-2 are potent inducers of endothelial cell migration and/or proliferation in vitro (1, 2, 7) and are known to modulate the expression of endothelial cell-derived proteases (8-13). Further, the HBGF prototypes are tightl...