Holger Zinke

Expression of apolipoprotein A-I in porcine brain endothelium in vitro

Abstract: Apolipoprotein (apo) A‐I is the major protein component of high‐density lipoproteins (HDLs), which are responsible for reverse cholesterol transport from peripheral tissues to the liver. A low level of plasma HDL is correlated with susceptibility to atherosclerosis and coronary heart disease. Mammalian apo A‐I synthesis has been attributed mainly to liver and intestine. Recently, apo A‐I expression has been shown in porcine brain capillaries, suggesting an independent lipid metabolism within the brain. In this study, protein synthesis and secretion were investigated in primary cultures of porcine brain microvascular endothelial cells and compared with those in large vessel endothelium. Active protein synthesis in vitro was demonstrated by metabolic labeling. Cerebral endothelial cells were shown to secrete apo A‐I into the culture supernatant, whereas aortic endothelial cells were negative for apo A‐I expression. Further studies of transcriptional regulation showed that cerebral endothelium was responsive to apo A‐I‐inducing agents, such as cholesterol, insulin, and retinoic acid, as previously shown in human hepatoma HepG2 cells. Thus, cultures of porcine cerebral endothelial cells may represent a suitable model for physiological studies of apo A‐I‐regulation with regard to brain lipid metabolism and blood‐brain barrier function. To investigate the interspecies conservation of regulatory elements, 178 bp of the 5′ flanking region of the porcine apo A‐I gene was cloned using PCR techniques. Alignments of the cDNA, of the deduced apo A‐I protein sequence, and of the 5′ promoter region with the corresponding genomic sequences of different species show a high degree of similarity between the porcine and the primate apo A‐I genes, thus indicating a similar function and possibly common regulatory mechanisms in those species. In contrast, the rodent and avian apolipoprotein A‐I promoter sequences differed significantly.

Chapter 17: Blood-brain barrier: a molecular approach to its structural and functional characterization

Our approach to analyze molecular components of the blood-brain barrier led to the identification of additional transcripts which can be regarded as “BBB markers”. Other candidates are presently analyzed in order to find hitherto unknown cell type-specific transcripts. We investigated the expression of these marker-genes in cell culture and found all genes still being transcribed after 10 days in primary cultures, although at a lower level. This is surprising, since other authors report the disappearance of BBB characteristics under such conditions. Moreover, the BBB marker γ-GT is found to be not only expressed in BMEC, but also in the closely associated pericytes. The hitherto unknown physiological function of the enzyme, especially the abundance in pericytes is still under investigation. Since the method of subtractive cloning has been proven as a fruitful approach, we consider to establish further subtractive cDNA libraries, using different subtraction parameters. The PCR method is applicable for amplification of subtracted cDNA (Timblin et al., 1990) and we expect to find additional clones, mainly of lower abundance which are of functional importance for the BBB phenomenon. The described characterization of cultured BMEC now allows to proceed to study BBB-specific gene expression with special regard to regulatory elements. We will perform these experiments by use of enhancer trap vectors transfected into BMEC. The isolation of the corresponding genomic DNA fragments of the BBB markers is in progress.

Cytotoxic activity of recombinant bFGF-rViscumin fusion proteins

A fusion protein (bFGF-rMLA), containing the mitogen basic fibroblast growth factor (bFGF) and the cytotoxic component of rViscumin (recombinant mistletoe lectin), the enzymatic A-chain (rMLA), was expressed in Escherichia coli, purified, and functionally characterized. bFGF-rMLA is cytotoxic for mouse B16 melanoma cells expressing the FGF receptor with an IC(50) value of approximately 1 nM. rMLA shows no significant effect on the viability of the B16 cells up to a concentration of 141 nM. Additionally, bFGF-rMLA was associated with the rViscumin B-chain (rMLB) in an in vitro folding procedure. The IC(50) value of bFGF-rMLA/rMLB to B16 cells in the presence of lactose-to block rMLB lectin activity-was 134 pM. Thus, it was possible to enhance the efficacy of a rViscumin A-chain mitotoxin through addition of rMLB. We conclude that rViscumin fusion proteins may be generally applicable for the receptor-specific inactivation of target cells and point out their potential in drug development.