WHY ARE THERE 2 PROGESTERONE RECEPTORS

This year represents the 20th anniversary of our first demonstration t hat human breast cancers contain progesterone receptors (PR), and are markers of hormone dependence. These receptors are now routinely measu red in tumours not only as markers of hormone dependence, but of disea se prognosis. Theoretically, their central function in breast cancers is not as markers, but as effecters of the proliferative signals of en dogenous progesterone in premenopausal women, and as targets for proge stins and antiprogestins. At present, PR are rarely measured for these functional purposes. The actions of PR are complex, and responsivenes s to progestin agonists or antagonists will depend on the gene whose a ctivity is being measured, the peculiarities of the cell and tissue un der study, and most importantly, the PR isoform that predominates in a tissue or tumour. The differential expression of PR isoforms, serves, we believe, to fine-tune responsiveness to this important reproductiv e hormone. Knowledge, not just of the PR content of a tissue, but of t he expression of B- vs A-receptors in that tissue, will be vital to un derstanding the effects of progestins therein. Recent studies with PR have forced us to revise the standard model of steroid receptor action . The conventional model, which depicts receptors as ligand-activated proteins that bind to specific DNA sequences at 'consensus' hormone re sponse elements and activate transcription, is not incorrect. It is, h owever, oversimplified, as studies with PR demonstrate. These demands include requirements for both positive and negative transcriptional re gulation; for tissue specificity of action; and for regulation of comp osite and simple gene promoters. Multiple functional domains control i ntramolecular contacts, intermolecular protein-protein interactions, a nd DNA binding. As steroid antagonists are synthetic rather than natur al hormones, their binding produces structural alterations in the rece ptors that unveil additional novel interactive capabilities. While ant iprogestins competitively inhibit agonists by forming non-productive r eceptor-DNA complexes, this is not their sole mechanism of action. Ant iprogestin effects may also be mediated by receptor interactions with coactivators whose function is in turn controlled by non-steroidal sig nals. When two different signalling pathways are activated simultaneou sly they can cooperate to produce unintended effects. Additionally, it seems clear that antagonist-occupied receptors can act without bindin g to canonical PREs, or without binding to DNA at all, relying perhaps on tethering proteins. This may be a consequence of the unusual allos teric structure imparted on the receptors by synthetic ligands. For so me of these unusual actions, the receptors may even be monomeric rathe r than dimeric. Investigators should not assume when studying antiprog estins that a specific mechanism is operating. These novel actions beg in to explain two properties of steroid antagonists that have puzzled investigators. One is the common observation that antagonists are agon ists in some normal tissues. The other, an extension of the first, is that in malignant cells, antagonists can acquire agonist-like properti es as tumours progress, leading to treatment failure. Although such tu mours are called 'resistant', they may in fact be responding quite wel l to the antagonist ! With respect to receptor protein structure, we a re only beginning to appreciate its complexity. For example, it appear ed initially that the structural independence of functional domains pe rmitted analysis of receptor fragments by fusing , them to heterologou s proteins. However, we now know that important functional domains can overlap; that other functional domains may be discontinuous; and that one domain can modulate the activity of another. This means that anal ysis of receptor fragments in chimeras is an incomplete test of domain function, and that we need innovative experimental strategies to unde rstand this intramolecular cross-talk.