Complement and polymorphonuclear leukocytes do not determine the vascular permeability induced by intraocular LPS.

The intravitreous injection of an endotoxin of Escherichia coli 055:B5 (LPS; 0.1-0.5 microgram/50 microliters of saline) induces ocular inflammation in rabbits that is maximal 20-24 hours later and disappears by 4 days. The inflammation is characterized by an alteration in ocular vascular permeability (OVP) measured by the ocular extravasation of 125I-albumin and an outpouring of leukocytes, most of which are polymorphonuclear leukocytes (PMNs), as determined by histopathologic study. Nitrogen mustard (mechlorethamine, 1.75 mg/kg) administered 3 days prior to LPS virtually eliminates PMNs in the circulation and those infiltrating ocular tissues 20 hours after intravitreous LPS, and yet the average increase in vascular permeability is not different from that of controls. Cobra venom factor (CVF; 300-400 units) 7 hours before intravitreous LPS produces a greater than 90% decrease in both hemolytic complement activity and zymosan-inducible serum chemotactic activity; yet 20 hours after LPS, the OVP is the same in CVF-treated rabbits and controls. For comparison, an ocular passive Arthus reaction (ovalbumin-anti-ovalbumin) was significantly affected by CVF pretreatment. Chemotactic activity in the aqueous humor is found in both CVF-treated and control rabbits 20 hours after intravitreous LPS. This activity attracts rabbit, but not human, PMNs, is partially heat-sensitive, and is not inhibited when PMNs are preincubated with C5a. These results indicate that neither PMNs nor circulating complement determine the OVP following intravitreous LPS, and that the chemotactic activity present in aqueous humor at the height of the inflammatory response is not primarily C5a.