Previous immunocytochemical studies indicate that receptor regulation varies in different uterine cell types. In primates, progesterone (P) suppresses estrogen receptor (ER) in glandular epithelial cells in the functionalis, but fails to suppress ER in the glandular epithelial (GE) cells of the basalis. P also fails to suppress ER in the perivascular stromal and smooth muscle cells of the spiral arteries in the functionalis. We used nonradioactive in situ hybridization to determine whether similar cell type differences occur at the ER mRNA level. We used digoxigenin-labeled oligodeoxynucleotides (oligo-DNAs; 45-mer) as probes and detected the hybrids immunocytochemically with horseradish peroxidase-labeled antidigoxigenin antibody. This technique can discriminate between positive and negative cells in closely packed histological associations. In spayed monkeys, most of the GE cells as well as endometrial stromal cells were positive for ER mRNA, while all vascular smooth muscle, endothelium, and perivascular stromal cells were negative. Estradiol treatment for 14 days markedly increased ER mRNA staining in the GE cells, most stromal cells, and the vascular smooth muscle and perivascular stromal cells of spiral arteries in the functionalis. However, in the basalis, these components of the spiral arteries were negative as were the small basal arteries of the basalis. In most positive cells, ER mRNA was not homogeneously distributed in the cytoplasm, but, rather, was concentrated in their perinuclear regions. The GE cells in the basalis had especially intense concentrations of perinuclear signal at their apical poles. After sequential estradiol plus P treatment, the signal was greatly reduced in the GE cells of the functionalis, but not in the GE cells of the basalis or in the vascular smooth muscle or perivascular stromal cells of the spiral arteries of the functionalis. In myometrium, ER mRNA was localized to the perinuclear region of smooth muscle cells, but the staining intensity was not dramatically affected by hormonal manipulation. Unexpectedly, we observed clusters of stromal cells characterized by extremely high positive signals for ER mRNA ("hot cells") at the endometrial/myometrial border and deeper in the connective tissue of the myometrium, although such cells did not express high levels of ER protein. In general, however, the cellular distribution of ER mRNA and its hormonal regulation paralleled those of ER protein.