Significant spontaneous fetal loss of unknown cause occurs in North American commercial swine. About 30% of conceptuses, thought to be genetically normal, are lost during the peri-attachment period. An additional 20% are lost at mid-pregnancy. Littermate endometrial and trophoblast biopsies were studied by quantitative real-time PCR for gene expression, and immunohistochemistry for protein expression at gestation day (gd)15–23 and 50. RNA analyses were also conducted on endometrial lymphocytes and arterial endothelial cells removed from biopsies by laser capture microdissection. Genes were selected for study from human literature and cloned as required. As in humans, angiogenic, cytokine, chemokine and chemokine decoy receptor gene expression occurs at the porcine maternal–fetal interface. In each tissue studied, distinct patterns of expression are found between early and mid-pregnancy, as well as between viable and arresting conceptus attachment sites. These changes involve both endometrial lymphocytes and dendritic cells. Restriction in endometrial angiogenesis, reduction in expression of the chemokine decoy receptor D6, and reduction in dendritic
cell numbers contribute to fetal arrest. In peri-attachment loss, interferon- c is more abundantly transcribed than tumor necrosis factor-a, but this ratio is reversed during midgestation failure. Further characterization of spontaneous fetal loss in pigs will identify targets for modification by hog producers and may provide a model for identification of antecedents to fetal loss in humans.

Our studies also characterize spontaneous fetal loss in pigs as an excellent animal model to assist in understanding the maternal–fetal interactions leading to pregnancy failure in humans. We have found porcine and human endometria, as well as trophoblasts, express similar cytokines, chemokines, and chemokine decoy receptors. The large amount of material available for study from porcine attachment sites, the reliability of spontaneous fetal arrest and its consistency of timing, the accessibility to timed pregnant reproductive tracts, and the ability to conduct paired sampling of littermates with different survival potential are all strong features that should complement studies in patients where these features are limiting.