The human progesterone receptor (PR) is a ligand-activated nuclear transcription factor which mediates progesterone action in target tissues. Two PR proteins, PR A (81-83 kDa) and PR B (116-120 kDa), have been described and different physiological activities ascribed to each on the basis of in vitro studies, suggesting that their ratio of expression may control progesterone responsiveness in target cells. Presence of PR in breast tumors is an important indicator of likely responsiveness to endocrine agents. However, the relative expression of PR A and B in breast cancer has not been described and its clinical significance has not been addressed. We have examined the expression of PR A and B in PR-positive breast tumors and found that while in most tumors PR A and B were expressed in similar amounts there was a broad overall distribution of PR A:B ratio which deviated significantly from a normal log distribution with tumors containing a PR A:B ration greater than 4 being over-represented in the group. Linear regression analysis revealed that high PR A:B ratios, in general, derived from a low concentration of PR B rather than high expression of PR A. PR A:B protein ratios were not correlated with the age of the patient or with total PR concentration. A third PR protein band (PR 78 kDa) was detected which comprised greater than 20% of total PR protein in a quarter of the tumor samples examined. The characteristics of tumors containing PR 78 kDa were not different from the overall group. In summary, in PR-positive breast tumors the ratio of expression of PR A and B proteins is close to unity as is seen in a number of other progestin target tissues. However, a significant proportion of tumors expressed very low levels of PR B and a consequently high PR A:B ration. Although the clinical consequence of this observation is not known, the in vitro findings that PR A may act as a repressor for PR B suggests that tumors containing primarily PR A may identify a subset of patients with low or aberrant response to endocrine agents.