The effect that topical administration of cyclosporine would have on the number and type of microorganisms isolated from the corneal surface of dogs with keratoconjunctivitis sicca was studied. Schirmer tear tests were performed on and corneal swab specimens were collected from 61 eyes of 31 dogs with keratoconjunctivitis sicca prior to and after 3, 6, and 12 months of treatment with cyclosporine. In eyes that responded to cyclosporine treatment (Schirmer tear test value increased by greater than or equal to 5 mm/min, compared with pretreatment value), the percentage of eyes from which bacteria were isolated after 3, 6, and 12 months of treatment was significantly (P < 0.001) less than the percentage from which bacteria were isolated prior to treatment. However, among eyes that did not respond to treatment, we did not detect a significant change over time in prevalence of bacteria or type of bacteria isolated. The percentage of eyes from which fungi were isolated decreased during treatment; however, the small number of eyes in which fungal culture results were initially positive precluded demonstration of a significant change. For all eyes, we did not detect any significant differences over time in the frequency with which specific bacterial genera were isolated, with the exception of beta-hemolytic Streptococcus spp. Opportunistic corneal infections were not detected even though none of the dogs received antibiotics. An increase in production of tears, which contain anti-infection proteins, was believed to be the primary factor responsible for the decrease in the percentage of eyes from which microorganisms could be isolated.

Pilus-mediated adherence is a virulence attribute of Moraxella bovis. Several pilus types have been shown to exist among strains of this bacterium, but correlation between pilus type and specific field cases of the disease has not been done. During the summer of 1987, an epizootic of infectious bovine keratoconjunctivitis was reported in 7 Iowa counties. Eight isolates of M bovis were secured from 12 episodes studied. All 8 of the isolates were nearly homogeneous in biochemical properties and had the same plasmid biotype. Pilus typing performed by immunofluorescence and immunogold electron microscopy identified a single new pilus type among 5 of the 8 isolates. This pilus type was identified in field cases that developed within a narrow time frame and over large distances. The implication of these findings is that infectious bovine keratoconjunctivitis epizootics may be associated with emergence of a novel pilus type, and that rapid dissemination over wide distances can occur, presumably by transportation of carrier cattle.

The cytotoxic effect of Moraxella bovis 118F on bovine neutrophils was evaluated and characterized by use of a 51Cr release assay. Neutrophils harvested from healthy adult cattle were labeled with 51Cr. The leukocidic activity produced by M bovis 118F, a hemolytic strain of M bovis, was heat-labile. A live culture of strain 118F, at a ratio of 100 bacteria/neutrophil, released 97.7% of the 51Cr from labeled neutrophils. Neither a heat-killed preparation of M bovis 118F nor a live or heat-killed preparation of M bovis IBH63 (a nonhemolytic and nonpathogenic strain) induced significant (P > 0.05) release of 51Cr. Moraxella bovis 118F broth culture filtrates prepared for evaluation of leukocidic activity also were evaluated for hemolytic activity. These 2 toxic activities had several characteristics in common. Both were filterable, heat-labile, produced by a hemolytic strain, and were released during early logarithmic phase growth from broth cultures. Leukocidic and hemolytic activities were protected from degradation by phenylmethyl sulfonyl fluoride, a serine protease inhibitor. Leukocidic and hemolytic activities were dependent on calcium ions. Filtrate resulted in 54.1% 51Cr release from labeled neutrophils and contained 646.7 hemolytic U/ml, respectively, when saline (0.85% NaCl) + 10 mM CaCl2 solution was used as diluent. Neither saline solution nor saline + 10 mM MgCl2 solution supported leukocidic or hemolytic activity. Serum, obtained from several calves 10 to 38 days after M bovis inoculation, substantially neutralized leukocidic and hemolytic activities, compared with paired preinoculation serum samples. In addition, no significant difference (P > 0.05) was detected when the ability of each calf's postinfection serum to neutralize leukocidic activity was compared with the ability of the serum to neutralize hemolytic activity.

An experimental model of keratoconjunctivitis sicca (KCS) was produced by removing the lacrimal gland and the gland of the third eyelid from the left eye of 6 cats. The right eye of each cat was left intact and used as a control. After 2 weeks, cats were euthanatized and the central portion of the upper eyelid from both eyes of each cat was excised. Histologic sections were stained with either hematoxylin and eosin or with a battery of biotinylated lectins including concanavalin A (conA), soybean agglutinin (SBA), wheat germ agglutinin (WGA), succinylated wheat germ agglutinin (S-WGA), Ulex europaeus agglutinin I (UEA), Dolichos biflorus agglutinin (DBA), Ricinus communis agglutinin (RCA), peanut agglutinin (PNA), and PNA pretreated with neuraminidase. Consistent differences in histologic features were not observed between conjunctivas with KCS and control conjunctivas. A variable degree of mononuclear cell infiltration of the substantia propria was observed in control conjunctivas and those with KCS. In both groups, conjunctival goblet cell density decreased and epithelial stratification increased as the degree of submucosal inflammatory cell infiltration increased. Lectin binding sites for DBA, WGA, S-WGA, UEA, PNA, and PNA pretreated with neuraminidase were detected on conjunctival goblet cells of conjunctivas with KCS and control conjunctivas. The mucus/glycocalyx layer of conjunctival epithelial cells in both groups of conjunctivas bound lectins RCA, WGA, UEA, and conA, but inconsistently bound S-WGA. In both groups, DBA principally bound to the mucus layer overlying normal epithelium, whereas PNA pretreated with neuraminidase consistently bound to the mucus layer of stratified epithelial surfaces free of goblet cells. Binding of SBA to goblet cells and the mucus/glycocalyx layer was variable.