RETROVIRAL GENE-TRANSFER - EFFECTS ON ENDOTHELIAL-CELL PHENOTYPE

Background. Endothelial cells (EC) are an attractive target for somati c cell gene therapy, both for the treatment of cardiovascular disease and for the systemic delivery of recombinant gene products directly in to the circulation. Recent evidence, however, suggests that viral tran sduction may induce unfavorable changes in EC phenotype. We examined t he proliferative capacity and cell adhesion molecule (CAM) pro file of EC after retroviral gene transfer (GT), employing a clinically releva nt ex vivo GT protocol. Methods. Human umbilical vein EC (HUVEC, N = 1 4 isolates) were exposed to supernatants containing the MFG.nlsLACZ ve ctor, which codes for a nuclear localized P-galactosidase, Control HUV EC were exposed to empty virus (CRIP) or no virus (NT). Efficiency of GT was quantitated by direct counting of beta-galactosidase-stained ce lls on a grid. Proliferation was quantitated by a 1-week assay of viab le cell counts. Expression of EC activation molecules (Class II major histocompatibility antigen [MHC II], E-selectin, intercellular adhesio n molecule-1 [ICAM-1], and vascular cell adhesion molecule-1 [VCAM-1]) was examined using fluorescent cytometry (FACS) at rest and after cyt okine stimulation. Results. GT was reproducibly efficient (mean 57%, r ange 40-77%) using sequential viral exposures without selection. NT, C RIP, and LACZ-transduced HUVEC exhibited identical FAGS profiles for E -selectin, ICAM-1, VCAM-1, and MHC II at rest, consistent with a nonac tivated state. Upregulation of expression by cytokine was quantitative ly similar for all groups. Growth rates were likewise not different be tween groups. Conclusions. Retroviral vectors may be employed to achie ve high percentages of transduced EC for ex vivo GT without the use of selection. Transduced EC generated in this fashion are not activated, demonstrate an unaltered pattern of inducible CAM expression, and exh ibit normal cell. growth. The effects of GT on target cell phenotype a re likely to be both vector and protocol specific and should be carefu lly assessed in each case prior to in vivo applications. (C) 1998 Acad emic Press.