TUMOR-NECROSIS-FACTOR-ALPHA MESSENGER-RNA-POSITIVE CELLS IN SPONTANEOUS RESORPTION IN RODENTS

PROBLEM: It has been proposed that high rates of resorption/spontaneou s abortion may result from interaction in the decidua of gamma-interfe ron-producing natural killer (NK) cells and tumor necrosis factor (TNF )-alpha-producing macrophages. An increased release of TNF-alpha from placental tissue of resorptions has been reported, but macrophages pro ducing TNF-alpha have so far not been demonstrated at the fete-materna l interface. Therefore, we have sought to identify TNF-alpha-producing cells by in situ hybridization at the fete-maternal interface in two inbred, well-characterized, and stable strains of laboratory rodents w ith high and low resorption rates. METHOD OF STUDY: Pregnant DBA/2-mat ed CBA/J mice with a resorption rate of 20% to 30% (dependent on NK ce lls and macrophages) and diabetes-resistant Bio-Breeding/Edinburgh (DR -BB/E) rats with low resorption rates (presumed to result from chromos omal abnormalities) were studied. AsialoGM1(+) cells were detected by immunohistochemistry, and TNF-alpha mRNA(+) cells were detected by in situ hybridization. RESULTS: TNF-alpha mRNA(+) cells were detected in DBA/2-mated CBA/J mice at the time of resorption but only at the troph oblast-decidual junction. AsialoGM1(+) cells were present in decidua, as assessed by immunohistochemistry, but few if any gave a positive si gnal for TNF-alpha. In rat resorptions, TNF-alpha mRNA-positive cells were present within the yolk sac and in contact with the trophoblast, but not at the trophoblast-decidual junction. In neither species did a significant accumulation of detectable TNF-alpha mRNA(+) cells occur before the usual time of onset of resorption. CONCLUSIONS: In the DBA/ 2-mated CBA/J mouse, the removal of the placenta is associated with re moval of a thin rim of adherent decidua similar to the location of the TNF-alpha mRNA(+) cells detected in this study. Our data suggest that increased TNF-alpha in tissues associated with failing fete-placental units may arise from infiltration/activation of scavenger cells from decidua that are likely to be macrophages. Local TNF-alpha production in decidua, which occurs as a prelude to resorption in the CBA x DBA/2 model, could not be detected due to the insensitivity of the TNF-alph a probe we used; the release of TNF-alpha from decidual tissue left af ter the removal of the placenta does not differ between resorbing and healthy implant sites. AsialoGM1(+) cells did not seem to be major pro ducers of TNF-alpha. TNF-alpha mRNA(+) cells in a low rate of resorpti on (rat) model were only found on the fetal side of the trophoblast, a nd they may also represent a macrophage response (to dying embryo tiss ue) derived from a nondecidual source. The location of TNF-alpha mRNA( +) cells may identify distinct and different mechanisms of resorption in rodents.