Development and comparison of in vivo and in vitro models for endometritis in cows and mares.

In order to investigate pathogenic mechanisms of acute endometritis in cows and mares, we established an in vivo model in both species. Based on the results of an in vitro transmigration system, human recombinant interleukin-8 (rhIL-8; 1.25 microg per mare and 5 microg per cow in 50 ml phosphate-buffered saline) was used to attract polymorphonuclear neutrophil granulocytes (PMNs) into the uteri. Peak numbers of uterine neutrophils were attracted after 6h, in both cows and mares. On average, mares responded more sensitively than cows, with 15 times higher numbers of rhIL-8-attracted uterine neutrophils (72+/-8 x 10(7)cells). In contrast to in vitro studies, in vivo migrated neutrophils (uterine neutrophils) of both species displayed a significantly reduced MHC class I expression. Expression of the CD11a molecule was significantly enhanced on equine uterine neutrophils but downregulated on bovine cells. Compared with untreated autologous peripheral neutrophils, both uterine and in vitro migrated neutrophils showed no alteration of phagocytic capacity. The ability to generate reactive oxygen species (ROS) was significantly upregulated in bovine and equine uterine neutrophils. This was also observed after in vitro migration of equine neutrophils, whereas ROS generation by bovine neutrophils was significantly depressed. In summary, the concept of inducing endometritis directly by local application of human interleukin-8 has been reliably successful in cows and mares. The model permits the analysis of PMN migration into the uterus under defined and controlled conditions. The observed differences between cows and mares with respect to phenotypical and functional characteristics of in vivo attracted uterine cells point to species-related features of neutrophil migration. In vitro transmigrated bovine and equine cells partially differ in phenotype and function from uterine neutrophils. Therefore, the in vitro transmigration assay cannot completely represent the in vivo endometritis model described here.