In contrast to other mammalians spontaneous endothelialization does not occur in humans. Since this lack of a functioning endothelium is thought to be a main reason for the unsatisfying performance of synthetic arterial grafts, we attempted to create an endothelium on the prosthetic surface by lining the graft with cultured endothelial cells. To provide sufficient cell-numbers for lining, a rapid and reliable harvest technique had to be developed first. In an initial series of 120 primary cultures, short segments of human saphenous veins were excised and cannulated before 0.1 % collagenase (CLS II) was instillated for 15 minutes to detach EC's. During the following growth under tissue culture conditions, we could demonstrate that a short transport time between vein excision and collagenase application is essential for a successful EC proliferation. In a subsequent study, evaluating risk factors, cell harvest and reproductive capacity of EC' s were markedly worse in smokers than in non-smokers. In order to further shorten the time span between vessel excision and collagenase application, an in situ canulation technique with immediate application of collagenase was developed. This harvest technique resulted in a 100% growth rate in 124 baboon primary cultures. Moreover, low plating densities not only achieved 43% shorter cell cycles than the usual plating densities, but also enabled mass-cultures after one single passage, thus reducing the cell damaging effect of trypsin. When the growth characteristics from 5 anatomically different sites were compared, the external jugular vein - which could be easily accessible and dispensable in each patient - proved to be an excellent source for EC cultures. The second goal of in-vitro endothelialization was the creation of shear stress resistant EC monolayers on the graft surface. A prerequisite for this is the use of a suitable underlying protein matrix, because EC do not adhere to uncoated surfaces of synthetic grafts. For that reason, primary adherence and cell attachment area of seeded human endothelial cells were determined on differently coated polytetrafluoroethylene (PTFE) grafts. Cell adherence and spreading was distinctly superior on two surface covering substrates: on fibronectin-treated type I/III collagen and on fibrinolytically inhibited fibrin glue (FG). Since FG is less thrombogenic than collagen, the shear stress resistance of cultured human endothelium was investigated on 6mm PTFE grafts, precoated with PG. EC seeding was performed in a microprocessor controlled seeding device, allowing a low inoculum of 12x10^4 EC/cm^2. A mock circulation simulated the flow patterns and the shear stress of the femoral artery. After 24 hours of perfusion, the grafts were still covered by a confluent endothelium. The in vivo persistence of cultured endothelium was then determined in the nonhuman primate model, using baboons. Since autologous in vitro lining requires a cell laboratory and is therefore confined to big centers, alogenic endothelialization with cryopreserved, pooled, blood-group identical endothelial cells was performed first. Bilateral femoral interpositions (10- 12cm) of experimental and control grafts were subsequently implanted into 18 baboons with the same blood group as that of the pooled cells. After 16 days of implantation the patency rate of grafts of the endothelialized group was similar to that of the control group (55.5 vrs 61.1 %; p> 0.1). Scanning electron microscopy revealed that 44.4% of experimental grafts were completely free of endothelium, while the remaining grafts showed a moderate cell coverage of 34.4+17.1 %. These remaining endothelial cells were found mainly as small cell islands, densely covered by leukocytes. For that reason, a similar series of baboon experiments was performed, using autologous EC from the external jugular vein. After 9 days of implantation, the patency rate of EC-lined and control grafts was 100% and 66.6%. Moreover, after both periods of observations, the endothelial coverage was shown to be intact in scanning electron microscopy and the underlying fibrin glue was well preserved in histological cross sections. An initially observed dense leukocyte adherence at the surface of the endothelium had disappeared after 4 weeks. In summary, we developed a method which results in endothelialized prosthetic vascular grafts with a significantly better patency rate than control grafts in the primate model. This method consists of the combination of in situ harvest of EC from external jugular veins, micro-grid follow-up, low density plating and microprocessor controlled seeding of endothelial cells onto fibrin glue coated PTFE grafts. We believe that these data provide a sufficient rational for initial clinical trials with autologous in vitro lining of prosthetic vascular bypass grafts.