Objective: Our purpose was to identify temporal and stage-specific expression of endometrial genes during coculture with trophoblast cells. Study Design: Endometrial stromal cells were cultured to confluence in the presence of estradiol and progesterone. During these culture conditions the gene expression of 1 tissue specimen that secreted abundant prolactin (415 ng/mL culture medium at 21 days) was compared with a second specimen that did not. These 2 tissues were coincubated with trophoblast tissue in a specialized coculture flask. After 4 and 24 hours of culture messenger ribonucleic acid was extracted and reverse transcribed, and the complementary deoxyribonucleic acid products were amplified by polymerase chain reactions. The reverse transcriptase–polymerase chain reaction products were separated by electrophoresis, and potentially important complementary deoxyribonucleic acid fragments were reamplified, inserted into a plasmid vector, and sequenced after recovery. Sequences were submitted for Basic Local Alignment Search Tool searches of GenBank. Results: We observed up-regulation of 6 gene fragments in decidualized endometrium after 4 hours of coculture with choriocarcinoma-derived trophoblast BeWo cells, but only 1 gene fragment was up-regulated after 24 hours of exposure. Conversely, 2 fragments were down-regulated in decidualized stroma that was exposed to BeWo for 4 hours and 2 fragments were underexpressed after the 24-hour exposure. In the parallel experiment stromal cells that failed to secrete prolactin did not elicit the same regulation of expression. The nondecidualized endometrium overexpressed 1 gene fragment after 4 hours of BeWo exposure and overexpressed 4 gene fragments after exposure to BeWo for 24 hours. Underexpression of gene products also occurred with the nondecidualized endometrium, and we observed 2 fragments and 1 fragment to be underexpressed after 4 and 24 hours of BeWo exposure, respectively. To date, 3 of the candidate differential display fragments from these experiments have been cloned and sequenced. An up-regulated fragment (C6225J4EB-1) was 99% identical (167/168 sequences) to a reported nonredundant expressed sequence tags isolated from muscle, brain, ovary, testis, liver, and pregnant uterus tissues. A second up-regulated fragment (C4375J4EB-1) matched 100% identity (117/117) with a reported gene fragment in the expressed sequence tags database of GenBank that was derived from fetal heart and pregnant uterus. Additional characterization of these expressed sequence tags has not been reported. The third up-regulated fragment (C4250J24EB-2) was 100% identical (265/265) to human reduced nicotinamide adenine dinucleotide dehydrogenase III in the nonredundant gene database of GenBank. Conclusion: This report demonstrates the potential usefulness that endometrial-trophoblast coculture and differential display can offer for the molecular analysis of implantation phenomena. We have recognized both overexpression and underexpression of interesting gene fragments during the early phases of endometrial responses to paracrine regulators derived from BeWo trophoblast cells. These responses appear to be specific to the degree of endometrial transformation (decidualization) before challenge by the trophoblast and to the duration of the BeWo exposure. Sequence data identified 1 gene with an unidentified function, another gene with a known function, and a fragment not previously recognized. We submit that our model of endometrial-trophoblast coculture offers a novel tool to test cellular responses during implantation, and differential display represents a sensitive technique that can identify many of the important elements of genomic signaling during nidation. (Am J Obstet Gynecol 1999;180:806-14.)