Thrombospondin: biosynthesis, distribution, and changes associated with wound repair in corneal endothelium.

Thrombospondin is a cell adhesion molecule which interacts via specific domains with a wide array of extracellular matrix components, including fibrinogen, fibrin, fibronectin, collagen, and heparan sulfate proteoglycan. Although this protein has been localized in several human tissues, its presence in corneal tissues had not been previously established. In the present study, we have demonstrated that cultured bovine corneal endothelial cells synthesize thrombospondin and incorporate it into their extracellular matrix. We have also shown immunofluorescently the presence and distribution of thrombospondin in these cultured cells and in the noninjured and injured corneal endothelium in situ. Ultrastructural immunoperoxidase cytochemistry revealed that thrombospondin could be displaced from the cell surface by heparin, but not by keratan sulfate. Confluent cultures of corneal endothelium synthesize and secrete the three cell adhesion proteins laminin, thrombospondin, and fibronectin in the ratios 1:8.2:51.8. Only the laminin B chains were detected in immunoprecipitates. Immunofluorescent studies of these cultured cells, using a polyclonal antiserum raised against purified thrombospondin, revealed a low level of fluorescence associated with the cell layer but a punctate fluorescent pattern at the level of the extracellular matrix. Noninjured corneal endothelium in situ also demonstrated a low level of fluorescence throughout the cell layer. However, this dramatically changed after a circular freeze injury to the tissue. By 24 h after wounding, cells surrounding the injury zone displayed a prominent fluorescence that was still observed at 48 h post-injury. In addition to its increased intracellular fluorescence, thrombospondin was also localized as migration tracks, oriented in the direction of cellular migration into the wound site. Thus, in corneal endothelium, thrombospondin appears to play a major role in injury-induced cell migration in situ along a natural basement membrane.

Ultrastructural immunocytochemical localization of fibronectin deposition during corneal endothelial wound repair. Evidence for cytoskeletal involvement.

The distribution of the extracellular matrix (ECM) protein, fibronectin (FN), has been examined ultrastructurally in noninjured and injured rat corneal endothelium in vivo and in vitro by immunoperoxidase cytochemistry. In noninjured endothelia, FN was observed within the rough endoplasmic reticulum (RER) cisternae but not along the cell-Descemet's membrane (DM) interface. Twenty-four and 48 h after a circular freeze injury, immunoperoxidase reaction product was detected at the cell-DM interface as well as within cytoplasmic vesicles and intercellular spaces. By 1 and 2 wk post-injury, a line of reaction product could still be demonstrated at the cell-DM interface and evidence for newly deposited basement membrane material was observed in this region. In order to understand whether fibronectin deposition during wound repair was dependent on cytoskeletal influences, organ culture experiments were performed in which the media was supplemented with either 10(-8) M colchicine or 2.5 X 10(-3) M cytochalasin B. Without inhibitors, injured corneas cultured for 24 h had FN deposition at the cell-DM interface similar to the in vivo results. Corneas cultured in the presence of cytochalasin B also showed FN deposition at the cell-DM interface. However, when injured endothelia were cultured in the presence of colchicine, no reaction product was observed at the cell-DM interface, although it could be detected intracellularly within RER. Incubating the tissues in the presence of puromycin abolished all extracellular and intracellular staining. These results indicate that during wound repair, corneal endothelial cells produce fibronectin and deposit it upon Descemet's membrane by a mechanism that may be mediated by microtubules.

Adipose tissue growth patterns during human gestation: a histometric comparison of buccal and gluteal fat depots.

Morphometric analysis of fat lobule size and number, and fat cell number in middle buccal and gluteal fat depots during the prenatal period was carried out using histological sections from 88 typical-for-age or normal human prenates of both sexes. The sample ranged from 110 to 385 mm Crown-Rump length (or from 14 through 42 gestational weeks). Compared with the buccal fat pad, the gluteal fat was one to four weeks delayed in lobule maturation. In addition to fat maturation differences between buccal and gluteal fat sites, gluteal fat characteristically showed fewer but larger fat lobules than did the buccal fat pad. Conversely there appeared a larger number of fat cells per unit area in the buccal fat than in the gluteal fat. Fat accumulation in human buccal and gluteal fat depots include differences in growth timing and magnitude, but also different developmental patterns. These patterns suggests our hypothesis that while lobule hyperplasia and hypertrophy occur at both sites, gluteal fat lobules grow primarily through lobule hypertrophy whereas the buccal fetal pad grows through lobule hyperplasia.