Direct transcriptional regulation of the progesterone receptor by retinoic acid diminishes progestin responsiveness in the breast cancer cell line T-47D.

Retinoic acid (RA) treatment of T-47D human breast cancer cells results in a rapid decrease in the concentration of progesterone receptor (PR) mRNA which causes a slow loss of cellular PR protein (Clarke, C. L., Roman, S. D., Graham, J., Koga, M., Sutherland, R. L. (1990) J. Biol. Chem. 265, 12694-12700). The mechanisms involved are unknown and this study was undertaken to determine whether the decline in PR mRNA was due to transcriptional inhibition and to evaluate the functional consequences of the RA-mediated decrease in PR. The transcription rate of the PR gene was decreased by RA, and the effect was maximal 2-3 h after treatment. Cycloheximide cotreatment was unable to relieve the inhibitory effect of RA and PR transcription suggesting that the effect was not dependent on ongoing protein synthesis. There was no effect of RA on PR mRNA half-life at the times examined (0-6 h of RA treatment). To determine the functional consequence of PR down-regulation the progestin-responsive plasmid pMSG-CAT was expressed transiently in T-47D cells which were then exposed to RA for 24 h. RA-pretreated cells were then treated with the synthetic progestin ORG 2058 and the extent of progestin stimulation of chloramphenicol acetyltransferase (CAT) activity measured. ORG 2058 treatment resulted in an induction of CAT activity which was maximal at a progestin concentration of 1 nM. Interestingly, the ability of ORG 2058 to induce CAT activity was decreased in RA-pretreated cells. The diminished progestin responsiveness of RA-pretreated cells was confirmed in separate experiments which showed that the progestin inducibility of TGF-alpha mRNA was also decreased in cells treated with ORG 2058 following pretreatment with RA for 24 h. These data demonstrate that RA decreases PR mRNA concentrations by direct transcriptional inhibition, leading to decreased cellular PR concentrations. The decreased levels of PR result in impaired responsiveness to progestins and this suggests that RA derived from dietary vitamin A may have a role in modulating cellular sensitivity to progestins.


Direct Transcriptional Regulation of the Progesterone Receptor by Retinoic Acid Diminishes Progestin Responsiveness in the Breast
Cancer Cell Line T-47D* (Received for publication, February 25, 1991) Christine L. Clarke Retinoic acid (RA) treatment of T-47D human breast cancer cells results in a rapid decrease in the concentration of progesterone receptor (PR) mRNA which causes a slow loss of cellular PR protein (Clarke, C. L., Roman, S. D., Graham, J., Koga, M., Sutherland, R. L. (1990) J. Biol. Chern. 265,[12694][12695][12696][12697][12698][12699][12700]. The mechanisms involved are unknown and this study was undertaken to determine whether the decline in PR mRNA was due to transcriptional inhibition and to evaluate the functional consequences of the RA-mediated decrease in PR. The transcription rate of the PR gene was decreased by RA, and the effect was maximal 2-3 h after treatment. Cycloheximide cotreatment was unable to relieve the inhibitory effect of RA on PR transcription suggesting that the effect was not dependent on ongoing protein synthesis. There was no effect of RA on PR mRNA half-life at the times examined (0-6 h of RA treatment). To determine the functional consequence of PR down-regulation the progestin-responsive plasmid pMSG-CAT was expressed transiently in T-47D cells which were then exposed to RA for 24 h. RA-pretreated cells were then treated with the synthetic progestin ORG 2058 and the extent of progestin stimulation of chloramphenicol acetyltransferase (CAT) activity measured. ORG 2058 treatment resulted in an induction of CAT activity which was maximal at a progestin concentration of 1 nM. Interestingly, the ability of ORG 2058 to induce CAT activity was decreased in RA-pretreated cells. The diminished progestin responsiveness of RA-pretreated cells was confirmed in separate experiments which showed that the progestin inducibility of TGF-a mRNA was also decreased in cells treated with ORG 2058 following pretreatment with RA for 24 h. These data demonstrate that RA decreases PR mRNA concentrations by direct transcriptional inhibition, leading to decreased cellular PR concentrations. The decreased levels of PR result in impaired responsiveness to progestins and this suggests that RA derived from dietary vitamin A may have a role in modulating cellular sensitivity to progestins.
The ovarian steroid hormone progesterone plays a major * This study was supported by the National Health and Medical Research Council of Australia, the NSW Cancer Council and MLC-Life Ltd. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ MLC-Life Research Fellow. To whom correspondence should be addressed. P Recipient of an Australian Postgraduate Research Award. role in regulating the proliferation and differentiation of a number of target tissues including the normal and neoplastic breast (1) and expression of the cellular receptor for progesterone is an important prerequisite for responsiveness to progestins. Regulation of the expression of the progesterone receptor (PR)' by estrogens (2)(3)(4)(5) and progestins (6-10) has been well defined, but there have been few, if any, studies examining the regulation of P R by other agents.
Studies from this group have recently documented the regulation of P R by the morphogen retinoic acid (RA), which decreased the concentration of P R mRNA in human breast cancer cells (11). RA has been shown to act via nuclear receptors which fall into three main groups: RAR-a, . PR and the RARs are members of the same gene family of ligand-activated transcriptional modulators (19) and have extensive structural features in common. There is also a potential epidemiological link between steroid hormone and retinoid-mediated effects. Several studies have shown that low dietary or serum levels of p-carotene are associated with increased risk of a number of epithelial malignancies (20-22), although the involvement of RA in modulating breast cancer risk is controversial. However, retinoids have been shown to act in combination with hormones in the chemoprevention of mammary cancer in animals (20,22,23). In addition to a possible involvement in carcinogenesis, retinoids can inhibit proliferation of human breast cancer cells both alone and in combination with antiestrogens (24-29). There is one report that breast tumors with high estrogen receptor levels occur in patients with high dietary retinoid levels (30).
The mechanisms underlying these observations are presently unknown, and there is only limited evidence that retinoids can directly affect steroid hormone-mediated events. RA decreases glucocorticoid receptor in squamous carcinoma cells resulting in resistance to glucocorticoid administration (31), and retinoic acid and glucocorticoids have been shown to have opposing effects on the expression of osteonectin and collagen (32).
Our previous data have shown that RA regulates the PR in human breast cancer cells primarily by a decrease in the cellular mRNA concentration followed by a gradual loss of receptor protein (11). The mechanism responsible for the decrease in mRNA concentration is unknown. The present study was designed to investigate whether the effect was mediated a t a transcriptional level and if so whether the effect was direct or mediated via an intermediary protein. Further-' The abbreviations used are: PR, progesterone receptor; RAR, retinoic acid receptor; FCS, fetal calf serum; ORG 2058, 16a-ethyl-21-hydroxy-19-nor-4-pregnene-3,20-dione; RA, retinoic acid CAT, chloramphenicol acetyltransferase; kb, kilobase; TGF-a, transforming growth factor a.
18969 more, it is not known whether the RA effect on PR results in a change in cell responsiveness to progestins. Functional interactions between RAR and the thyroid hormone receptor have been described (33-36), but no such interactions have yet been described between RAR and PR. Because the loss of PR protein is gradual after RA treatment, the effect of this agent may be to diminish sensitivity to progestin treament, and therefore the responsiveness to progestins of cells pretreated with RA was investigated.

EXPERIMENTAL PROCEDURES
Materials-Materials were obtained from the sources previously listed (11). ["S]Methionine, [a-"PIUTP, and ["]uridine were Amersham Australia (North Ryde, Sydney) products. Plasmids bearing the progestin-responsive mouse mammary tumor virus long terminal repeat linked to chloramphenicol acetyltransferase (pMSG-CAT) and bacterial @-galactosidase sequences (pCH110) were obtained from Pharmacia LKB Biotechnology (North Ryde, Sydney, Australia); pSG5-hPR1 was a gift from Professor Pierre Chambon, Strasbourg, France; hRAR-a and -7 were gifts from Dr. Ronald Evans, Salk Institute, La Jolla, CA, and hRAR-p was a gift from Dr. Magnus Pfahl, La Jolla Cancer Research Foundation, La Jolla, CA. Human fibroblast @-actin was encoded in a 2.9-kb Okayama-Berg vector (37) and human TGF-a cDNA was a 3.1-kb fragment encompassing the entire coding region for the 160 amino acid precursor of TGF-cu (38), obtained from Genentech Inc. (San Francisco, CA). pSG5 was obtained from pSG5-hPR1 by excision of the hPR insert, religation of the EcoRI ends, and selection of hPR negative clones. T-47D (39) cells were supplied by E. G. and G. Mason Research Institute, Worcester, MA and were cultured as previously described (11,40,41). Cells were negative for mycoplasma contamination as determined using the Gen-Probe rapid detection system (Gen-Probe lnc., San Diego, CA).
Isolation and Analysis of RNA-RNA was isolated by the guanidiniurn isothiocyanate-cesium chloride method, and Northern analysis carried out as described (11) except that cDNA probes were labeled by random priming using the Amersham multiprime DNA labeling system. To control for variation in RNA loading and transfer, membranes were stored at -20 "C until the probe had decayed and were then reprobed with 0-actin cDNA. Autoradiograms were analyzed as described previously (11).
Immunoblot Analysis of PR-High salt cytosolic extracts of harvested cells were prepared, separated by electrophoresis, and blotted onto nitrocellulose as described previously (11). Blots were incubated with hPRa 7, a monoclonal antibody against human PR ((42) immunoglobulin fraction of hybridoma supernatant), at a final dilution of 1:lOOO and immunoreactivity was revealed colorimetrically as described (11).
Transcriptional Analysis-PR gene transcription was measured using the method of Greenberg and Ziff (43) as previously described (6), with modifications. Cells were plated in RPMI 1640 medium containing 5% fetal calf serum (FCS) and supplemented as described (40,41). The medium was replaced with medium containing 1% charcoal-stripped FCS 4 days after plating and 16-24 h after medium change cells were treated with retinoic acid M) or dimethyl sulfoxide vehicle and harvested various times thereafter. Where stated cycloheximide (20 pg/ml medium) was added at the start of treatment: control experiments showed that this concentration of cycloheximide resulted in greater than 90% inhibition of protein synthesis by 1 h after treatment (measured by ["S]methionine incorporation into trichloroacetic acid-precipitable material). Nuclei were prepared by lysis of harvested cells (43) and frozen in liquid nitrogen until required. Initiated transcripts were labeled with [a-"PIUTP as described (43) except that incubation was for 45 min. RNA was isolated by phenol/chloroform extraction, precipitated (three to four times) with 10% trichloroacetic acid until the washes were free of radioactivity, then treated with 200 mM NaOH, and run-on transcripts were quantitated using immobilized cDNA. Labeled RNA consisting of equal radioactivity/sample was hybridized at 65 "C for 36 h (43) to nitrocellulose filters onto which 5 pg each in duplicate of pSG5-hPR1, @-actin, and pSG5 cDNAs had been blotted and fixed by UV radiation. Filters were washed (43), dried, and exposed to xray film. Autoradiograms were analyzed densitometrically, and the mean heights of duplicate PR RNA signals on filters bearing RAtreated samples were normalized for the mean heights of the duplicate @-actin signals on the same filter, then expressed as a percentage of the similarly normalized PR RNA signal on control filters.
Labeling of RNA with PHIUridine-PR mRNA half-life was determined using the pulse-chase technique (44) with modifications. Cells were incubated in the presence of 200 mCi/ml ["Hluridine (39 Ci/mmol) for 5 h and the medium replaced with medium containing 5 mM unlabeled uridine and 2.5 mM cytidine in the presence of 1O"j M RA or vehicle. Cells were harvested 1, 2, 4, and 6 h following the medium change. Total cellular RNA was isolated as described above for transcriptional analysis. The yield of RNA was similar in treated and control groups at all times. Tritiated PR mRNA levels in each RNA sample were determined using the membrane-bound excess cDNA technique (45), and filters were prepared as described for transcriptional analysis. Membranes were prehybridized overnight at 50 "C, and then equal amounts of RNA/time point were added and filters hybridized as described previously (6). Membranes were washed and treated with ribonuclease A (43). Radioactivity in the excised membrane slots was determined by liquid scintillation. Specific radioactivity was determined by subtracting the radioactivity in pSG5 slots from the radioactivity in pSG5-hPR1 or @-actin slots.
Gene Transfection into T-47D Cells-Cells were plated into 150-cm2 flasks (2.5-3.0 X lo6 cells/flask) in RPMl 1640 medium containing 5% FCS 3 days prior to transfection. On the morning of transfection medium was changed to Dulbecco's modified Eagle's medium, pH 7.3-7.4, containing 5% FCS, and cells were cotransfected, using the calcium phosphate precipitation method (46), with pMSG-CAT (40 pg/flask) and pCHllO (10-20 pg/flask). Cells were subjected to osmotic shock using 15% glycerol (45 s) 4 h after transfection and exposure of cells to the DNA continued for 18 h thereafter. Cells were harvested, mixed, and replated into multiwell plates (1-3 X 10' cells/ well in 6-well plates) to reduce variations due to interflask differences in transfection efficiencies. When the cells had adhered to the substratum (3-6 h after plating) RA M) or vehicle was added and cells were further treated 3, 6, or 24 h thereafter with ORG 2058 at the indicated concentrations. Cells were harvested, counted, and lysed in 0.25 M Tris/HCl buffer, pH 7.8, by freeze thawing (three times). Aliquots were removed from the lysate for measurement of cell-free @-galactosidase activity (46) and the lysates heated at 65 "C for 10 min. After centrifugation, an aliquot was removed for measurement of CAT activity by a non-chromatographic method (47). Protein determination was by the Bio-Rad method. CAT activity was normalized for cell-free @-galactosidase activity. Transfection efficiencies and the effect of treatment on pCHllO expression were estimated by counting @-galactosidase-positive cells revealed on cell monolayers using the histochemical indolyl method (48,49). Transfection efficiencies between experiments typically varied between 0.1 and 0.6%.

Retinoic Acid Decreased the Transcription Rate of the PR
Gene-Treatment of T-47D cells with 10" M RA resulted in a decrease in the transcription rate of the PR gene, which was seen 1 h after and persisted for at least 3 h after treatment ( Fig. 1). This decrease in transcription rate preceded the previously demonstrated fall in PR mRNA levels ( Fig. 1 and Ref. 11). Given the known half-life of 2-2.5 h for PR mRNA (6), the time between the RA inhibition of transcription and the decrease in PR mRNA levels is consistent with the effect resulting primarily from a transcriptional event. The transcription rate recovered partially at 6 h ( Fig. 1) although PR mRNA levels remained depressed for at least 36-48 h (11). The effect of RA on PR gene transcription persisted in the presence of cycloheximide, suggesting that it was independent of ongoing protein synthesis (Fig. 2). The increase in @-actin transcription rate upon cycloheximide exposure has been described previously (50).
PR mRNA Half-life Was Unaffected by RA Treatment--In order to verify whether the inhibitory effect of RA on PR mRNA was due primarily to inhibition of transcription or whether RA also affected PR mRNA half-life, total cellular RNA was isolated from cells which had been pulse-labeled using [3H]uridine, then treated with M RA or vehicle for 6 h. The radioactivity in PR and actin RNA was measured over time. PR mRNA decreased rapidly and reached background levels within 6 h (Fig. 3). The rate of the decrease in Autoradiograms were scanned densitometrically, I' R was normalized a s described under "Experimental Procedures" and expressed as a percentage of vehicle-treated controls. Opcm bars, control: hntrhcd hnrs, 11.4 treated.
PR mRNA concentration was unaffected by RA, and the radioactivity in P-actin mRNA did not change appreciably over the time course examined. detectable as a single mRNA species of 3.4 kh (Fig. 4). Treatment of T-47D cells with increasing concentrations of RA for either 6 h (not shown) or 48 h (Fig. 4) had no effect on the mRNA levels of either receptor. RAR-/j mRNA was barely detectahle in T-47D cells and was increased to low levels after 6 or 48 h of treatment with 10'; M or IO"' M HA (not shown).

RA Treatment Causm Diminkhed Prozestin Rrspmsiucnrw
of T-47D Cells-In order to examine the functional consequences of RA-mediated P R loss, the effect of RA on endogenous ORG 2058-inducible events was examined. TCF-tr was chosen as an endogenous progestin-sensitive end point because its mRNA levels are increased hv progestins in breast cancer cells (51-53). T-47D cells were pretreated with RA for 24 h a t which time PR protein levels were around 50'';. (Fig.  5, lane 2 ) of vehicle-pretreated controls (Fig. 5 , ion(. I ) a s previously described (11). Cells were then treated for 9 h with ORG 2058, which increased the size of the receptor proteins as previously described (54)(55)(56)(57)(58)(59) in hoth vehicle-pretreated and RA-pretreated samples (Fig. 5 , lanm 4 and 6 ) . In samples which had not been treated with ORG 2058, the concent rat ion of P R in the RA-pretreated sample at 9 h (Fig. 5 , lonp .5  further diminished relative to zero time (Fig. 5,  l a w 2 ) , consistent with a continuing slow decline in immunoreactive P R levels upon prolonged RA treatment (ll), whereas PR levels in vehicle-pretreated controls remained the same (Fig.  5, compare lane 3 with lane 1 ). The concentration of TGF-(U RNA was estimated 9 h after progestin treatment. ORG 2058 treatment caused a 2-2.5-fold increase in the abundance of this mRNA species (Fig. 6), and the ORG 2058 effect was dose-dependent as previously shown (51). RA itself was ineffective in modulating TGF-n mRNA levels, but the magnitude of the ORG 2058 mediated increase was blunted in cells which had been pretreated with RA for 24 h (Fig. 6). The RAmediated decrease in ORG 2058 induction of TGF-n mRNA was seen at all the ORG 2058 concentrations tested, not only a t 9 h (Fig. 6) but also at 3 and 6 h (not shown) after treatment. RA pretreatment also decreased the ahility of ORG 2058 to induce c-myc mRNA (not shown). Overall, however, KA had only modest effects in decreasing progestin sensitivity of endogenous progestin-sensitive mRNAs. This could be due to a number of factors including the still unknown mechanism through which ORG 2058 induces these mRNAs. If the OHCI 2058 effect is indirect, the PR-mediated component of the response may be minor and RA-mediated decreases in PH have little effect on the magnitude of the outcome.
In order to address more directly the functional consequences of RA-mediated decreases in PR, the progestin-responsive plasmid pMSG-CAT was transfected into T-471) cells. This plasmid contains the long terminal repeat of the mouse mammary tumor virus, and this has several progestinresponsive elements with which PR has been shown to interact directly (60)(61)(62)(63)(64)(65)(66)(67)(68). ORG 2058 treatment of T-4'iD cells transiently expressing pMSG-CAT resulted in induction of CAT activity which was essentially maximal 24-48 h after ( Fig. 7, insset) and persisted at least until 68 h after ORG 2058 treatment, which was equivalent to a total time of 96-h posttransfection. The maximal induction achieved was proportional to the quantity of plasmid transfected (not shown).
T h e effect of ORG 2058 on CAT induction was maximal at a concentration of 1 nM (Fig. 7). The progestin R.50'20 and the synthetic androgen R1881 (each 10 nM) also induced CAT activity maximally whereas 17[$estradiol and dexamethasone were without effect (not shown). The inability of glucocorticoids to induce MMTV-CAT activity in T-4'iD cells has been shown previously (60,62,63) and is attributable to the low glucocorticoid receptor concentration in these cells (64, 6 5 ) . No effect of RA or ORG 2058 at any concentration was noted on the expression of pCHll0 (not shown).
T-47D cells were pretreated with RA prior to treatment with ORG 2058 in order to determine whether RA-mediated loss of PR protein would result in a diminished ability of ORG 2058 to induce CAT activity. RA pretreatment was for 24 h, which corresponded to the time needed for RA to cause essentially maximal loss of receptor protein (1 1 ). In cells pretreated with vehicle for 24 h, ORCI 2058 treatment increased CAT activity 7.3-fold a t 0.1 nM and 9.5-fold at 10 nM ( Fig. 8 A ) . However, in cells which had been exposed to HA for 24 h and consequently had lower PR levels (1 11, the induction of CAT activity was diminished to 3.1-fold at 0.1 nM and 6.4-fold at 10 nM OR(> 2058 (Fig. 8A ). The decreased ability of ORG 2058 to induce CAT activity was more nounced at 0.1 nM, a concentration which was submaximal for CAT induction (Fig. 7).
The ability of ORG 2058 to induce CAT activity was examined also in samples which had been RA pretreated for either 3 or 6 h. This time frame was chosen as it precedes the fall in P R protein seen upon RA exposure (11). There was no difference in the induction of CAT in the 3/6-h vehicle (9fold) and RA (8.9-fold) pretreated samples by 10 nM (Fig. 8B) or at lower ORG 2058 concentrations (0.1, 1 nM, not shown).

DISCUSSION
This study has shown that RA was able to decrease the transcription rate of the PR gene. The decrease in transcription rate was transient, being maximally decreased a short time (3 h) after treatment but recovering to close to control levels at longer times. This effect on transcription is comparable with the known effect of RA on PR mRNA levels. RAmediated P R mRNA loss is maximal 3-6 h after treatment, and the decreased level is sustained at least until 24-48 h thereafter (11). It is likely that transcriptional inhibition was sufficient to account for the decline in mRNA levels up to 6 h after treatment, as no effect of RA on the half-life of P R mRNA was noted during this time. However, the sustained lowering in PR mRNA levels beyond 6 h of treatment takes place in the face of an apparent recovery in transcription rate documented in this study and suggests that although the rapid fall in P R mRNA levels can be attributed to a transcriptional event in the acute phase (0-6 h), it is possible that posttranscriptional mechanisms come into play thereafter. Posttranscriptional regulation of RNA processing by retinoids has been described (reviewed in Ref. 20). Furthermore, posttranscriptional events have been invoked to explain the decrease in estrogen receptor mRNA levels after estrogen treatment, which causes only transient decreases in the transcription rate of the estrogen receptor gene, in MCF-7 cells (66).
The rapidity of the effect of RA on transcription suggested the possibility that it was direct. Coincubation of cycloheximide and RA under conditions where new protein synthesis had been inhibited by over 90% failed to abrogate the transcriptional inhibition, evidence that no intermediate protein synthesis was required. This suggests that the effects of RA on PR may be mediated by direct interaction with P R promoter sequences. It is intriguing to note that the decrease in P R mRNA levels mediated by RA never exceeds 50%, in contrast with the profound decrease caused by progestin treatment (6-10). One explanation could be that RA affects only one of the two known P R promoters (67). Alternate transcription start sites and polyadenylation signals have been shown (67) to result in the five PR mRNA species described previously (8-10,68) in which case preferential loss of one or more P R mRNA species may be detectable. Presently there is no evidence for such preferential loss upon RA treatment, although analysis of total RNA, in which the PR mRNA species smaller than 6 kb fail to be fully resolved around the 18 S and 28 S ribosomal RNA subunits, probably precludes clear demonstration of such a phenomenon. However, there is evidence that the ratio of the A and B P R proteins remains the same at all times after RA treatment (not shown).
The RA effects on PR are likely to be mediated by one of the family of recently described RAR (12-18). Two classes of RAR, the a and y, were detectable in T-47D cells as previously described (16). There was no effect of RA on RAR-a at any concentration tested which is consistent with what has been described in human hepatoma cells and for the murine receptor in F9 cells (18,69). There was also no effect of RA on RAR-7 in contrast with the murine homologue which is decreased by RA treatment (18). RAR-P was barely detectable in T-47D cells as previously shown (16) but was inducible to a low level a t high RA concentrations, consistent with the presence of RA-responsive elements in the upstream region of the receptor gene (70, 71). It is not yet known which of the RA receptors is responsible for regulating P R gene transcription. However, the fact that the a and y species fail to be regulated by RA treatment suggests that their continued presence in the face of high RA levels may allow a sustained effect of RA acting through either RAR on lowering PR levels. The consequences of RA treatment on progestin responsiveness were evaluated to assess the potential physiological significance of the data presented in this study. Transfection into T-47D cells of a reporter plasmid containing the progestin-responsive long terminal repeat of the mouse mammary tumor virus (60) resulted in rapid and sensitive induction of CAT activity upon progestin treatment as previously described (60)(61)(62)(63). Pretreatment of transfected cells with RA for short periods of time during which P R protein levels were unaffected by RA had no effect on the ability of progestin to induce CAT activity. However, when the RA pretreatment was prolonged, producing a maximal decrease in PR protein levels, the cell responsiveness to progestin was markedly reduced. The induction of CAT activity by 0.1 nM ORG 2058 in RA-pretreated cells was 40-50% of that noted in vehiclepretreated cells. The magnitude of this effect is of the same order as the magnitude of the RA-mediated loss of PR protein (11). The effect of RA pretreatment was more evident at concentrations of ORG 2058 which were submaximal for CAT induction and therefore less likely to be blunted by transcription factor limitations which may occur at high activated receptor concentrations (72).
This study raises the question of whether vitamin A-derived retinoids may play a part in the modulation of steroid hormone-mediated action in breast cancer. Retinoids are known antiproliferative and differentiating agents in a wide range of tissue and cell types (20-22) and can inhibit proliferation of human breast cancer cells at similar concentrations to those used in this study (24)(25)(26)(27)(28)(29). However, the effects of RA on PR documented in this study are unlikely to be simply secondary to effects of RA on proliferation, as the proliferation status of T-47D cells has little or no effect on cellular PR concentrations (not shown).
A relationship between retinoid intake or serum levels of retinoids and breast cancer incidence has been proposed but has been difficult to establish. Case control and prospective studies have both supported (73, 74) and refuted (75)(76)(77) the hypothesis that low dietary or serum /?-carotene levels are associated with an increased risk of breast cancer. Furthermore, despite the link postulated in some other cancers between low /?-carotene levels and increased risk (20)(21)(22), few if any studies have found such a link with retinol, perhaps because serum levels of carotenoids vary with nutritional intake, whereas homeostatic mechanisms maintain serum retinol levels within a defined range (21). It has yet to be demonstrated, in addition, whether the protective effects of p-carotene are due to their antioxidative properties or to local metabolism to retinol or retinoic acid (21), nor has the extent of tissue retinol or retinoic acid conversion from p-carotene been determined.
Although the role of retinoids in breast cancer incidence and/or progression has yet to be clarified, this study and in uitro studies cited above, which have examined the effects of retinoids directly on molecular events, have provided evidence for a functional association between retinoid and steroid hormone-mediated events in breast cancer cells. This study has demonstrated that RA treatment decreases PR gene transcription and results in reduced cell sensitivity to progestins. Whether this reduced sensitivity could result in diminished ability of breast cancers to respond to progestin therapy, or whether loss of sensitivity to progestins during a critical period of breast differentiation during puberty may alter normal breast cell susceptibility to carcinogenic insult is presently unknown. 14.