INTERCELLULAR GAP FORMATION INDUCED BY THROMBIN IN CONFLUENT CULTUREDBOVINE RETINAL-PIGMENT EPITHELIAL-CELLS

Purpose. Thrombin is formed at the site of intraocular hemorrhage and may be important in the development of progressive retinal damage. The authors observed that thrombin-treated bovine retinal pigment epithel ial (RPE) cell cultures develop intercellular gaps and initiated this study to examine in detail the effects of thrombin on RPE cell morphol ogy, adhesion, and cytoskeleton. Methods. Confluent cultures of bovine RPE cells were incubated for various times (0 to 24 hours) with alpha -thrombin (0.1 to 100 U/ml) or enzymatically inactive thrombin. Interc ellular gaps were quantitated by light microscopy in ten representativ e fields (magnification X400) as number of gaps per field (gaps/f). RP E cytoskeleton was studied using immunofluorescent staining for vincul in and F-actin. The mechanism of thrombin-induced RPE cell gap formati on was studied by preincubation with specific drugs, including a prote in kinase inhibitor (stauro-sporine), protein kinase C inhibitors (H-7 and calphostin C), cyclic adenosine monophosphate (cAMP) inducer (for skolin), and cytoskeleton-disrupting agents (cytochalasin B or colchic ine). Results. Intercellular gaps (20 to 80 mu m in diameter) were mar kedly increased in number in thrombin-treated cultures in a dose-depen dent and time-dependent manner and were associated with an alteration in the distribution of F-actin and vinculin. Whereas control cultures showed 3.3 +/- 2.4 gaps/f, incubation with 8 U/ml of alpha-thrombin fo r 3 hours resulted in 44.8 +/- 15.3 gaps/f. These changes were most pr ominent shortly after the 3-hour coincubation, but the cultures did re turn to their original confluent state within 24 hours. Cultures treat ed with an enzymatically inactive thrombin showed fewer intercellular gaps than those treated with enzymatically active thrombin but had sig nificantly more intercellular gaps than control cultures. Thrombin-ind uced intercellular gap formation was blocked by preincubation with for skolin (14.6 +/- 7.1 gaps/f), staurosporine (10.2 +/- 5.0 gaps/f), or H-7 (24.5 +/- 9.8 gaps/f). Conclusions. Exposure to an enzymatically a ctive thrombin results in formation of intercellular gaps between cult ured RPE cells. Inhibition of this phenomenon by protein kinase inhibi tors and by a cAMP inducer suggests that this effect is mediated, at l east in part, through protein kinase C- and cAMP-dependent pathways. T hrombin generation associated with intraocular hemorrhage may thus res ult in direct damage to the RPE monolayer, possibly via the same pathw ay(s).