No abstract

The development of techniques for isolating hepatic lipocytes (Ito, stellate or fat-storing cells) from rodents has been instrumental in defining their role in hepatic vitamin A storage and fibrogenesis. In this study, we developed a method for the purification of lipocytes and Kupffer cell from wedge sections of normal human liver and examined their properties in primary culture. Sections of donor liver (400 to 600 gm) harvested but not used for transplantation were perfused in situ with University of Wisconsin solution and used for lipocyte isolation within 48 hr. Cells were isolated by catheter perfusion of the wedge through several large vessels with L-15 salts, Pronase and collagenase, followed by Larex density gradient centrifugation. Lipocytes were plated on either uncoated plastic or a basement membrane-like gel. Lipocyte and Kupffer cell yields were 2.3 +/- 0.6 x 10(5) and 8.6 +/- 1.4 x 10(5) cells, respectively, per gram of liver (n = 5). Lipocyte purity was 91% as assessed by vitamin A autofluorescence, and Kupffer cell purity was 83% as determined by uptake of fluorescinated staphylococci. Lipocytes cultured on the plastic spread within 48 to 72 hr, displaying slightly more heterogeneous retinoid droplet size than comparable rat cells; on a basement-membrane gel, the cells remained aggregated and spherical with occasional spindlelike extensions. Lipocytes on plastic expressed procollagens I and III, collagen IV and laminin by immunocytochemistry, and types I, III and IV procollagen messenger RNAs by RNAse protection. Northern blot and polymerase chain reaction, respectively. Transmission electron microscopy of lipocytes at 7 days demonstrated a prominent rough endoplasmic reticulum and contractile filaments. Scanning electron microscopy revealed a smooth cell surface with perinuclear droplets beneath the cell membrane. With continued primary culture on plastic (more than 7 days), cells appeared "activated" (i.e., increased spreading and diminished retinoid droplets) and began proliferating as assessed by nuclear autoradiography and [3H]thymidine incorporation. Kupffer cells observed by scanning electron microscopy in early primary culture displayed prominent membrane ruffling and lamellipodia. In summary, we have established a reproducible method for the isolation and primary culture of human lipocytes and Kupffer cells.

Open fenestrations are a conspicuous feature of sinusoidal endothelial cells and allow free movement of plasma into the space of Disse. In hepatic fibrosis, the number of fenestrations decreases as interstitial collagen increases in the liver, a change that correlates with deposition of extracellular matrix in the space of Disse. In this study, the possibility of a causal relationship between altered fenestral morphology and perisinusoidal matrix has been examined by culturing rat sinusoidal endothelial cells on individual matrix proteins or on a native matrix consisting of human amniotic membrane with interstitial collagen (types I and III) on one side and basement membrane proteins (collagen types IV and V and laminin) on the other. Under culture conditions, individual components of the extracellular matrix failed to maintain fenestrations. A basement-membranelike gel matrix derived from the Engelbreth-Holm-Swarm tumor war similarly ineffective. Fenestral density and porosity (percentage of cell surface occupied by fenestrations) were significantly enhanced, however, when endothelial cells were cultured on the basement-membrane side of human amnion. These data suggest that support of endothelial fenestrations requires a complex matrix. In particular, physiologically derived basement membrane maintains fenestrations, whereas interstitial collagen matrix does not. The loss of fenestrations associated with hepatic fibrosis may be related in part to an accumulation of interstitial collagens in the space of Disse.