Calcium/Calmodulin Kinase Inhibitors and Immunosuppressant ...

Calcium/Calmodulin Kinase Inhibitors and Immunosuppressant Macrolides Rapamycin and FK506 Inhibit Progestin- and Glucocorticosteroid ReceptorMediated Transcription in Human Breast Cancer T47D Cells

Ste´phane Le Bihan, Ve´ronique Marsaud, Christine Mercier-Bodard, Etienne-Emile Baulieu, Sylvie Mader, John H. White, and Jack-Michel Renoir URA 1218 Centre Nationale de la Recherche Scientifique (V.M., J.-M.R.) Pharmacologie Cellulaire 92296 Chatenay-Malabry Ce´dex, France INSERM U33 (S.L.B., E.-E.B.) and IFR 21 (C.M.-B.) 94276 Le Kremlin-Biceˆtre Cedex, France Departments of Physiology and Medicine (J.H.W.) McGill University Montreál, Quebec, H361Y6 Canada Department of Biochemistry (S.M.) Faculty of Medicine University of Montreal Montreál, Que´bec, H3TIJ4, Canada

progesterone receptor (PR)-, as well as glucocorticosteroid receptor (GR)- mediated transactivation are targets of immunosuppressants and CaMKs in T47D cells. Indeed, Northern analysis showed that Rap, KN62, and, to a lesser degree, FK506 inhibited progestin stimulation of Cyclin D1 mRNA levels, but not those of the non-steroidregulated glyceraldehyde 3-phosphate dehydrogenase (GAPDH) gene. Addition of Rap or KN62 after exposure of cells to progesterone agonist Org 2058 had no effect on induction of CAT activity. Taken together, these data indicate that Rap and FK506, as well as CaMK inhibitors, inhibit steroid-induced activities of exogenous, as well as of some endogenous, steroid receptor-regulated genes by a mechanism preceding hormone-induced receptor activation. Rap appeared to stabilize a 9S form of [3H]Org 2058-PR complexes isolated from T47D (GRE)5CAT cell nuclei. By contrast, the progesterone receptor (PR) was isolated from cells treated with KN62 as a 5S entity, undistinguishable from the 5S PR species extracted from cells treated with progestin only. The nuclear 9S-[3H]Org2058-PR resulting

The effects of immunosuppressants and inhibitors of specific calcium/calmodulin kinase (CaMK) of types II and IV on progestin/glucocorticosteroidinduced transcription were studied in two human stably transfected breast cancer T47D cell lines. The lines contain the chloramphenicol acetyl transferase (CAT) gene under control either of the mouse mammary tumor virus promoter (T47DMMTV-CAT), or the minimal promoter containing five glucocorticosteroid/progestin hormone response elements [T47D-(GRE)5-CAT]. Progestinand triamcinolone acetonide (TA)-induced CAT gene expression was inhibited in a dose-dependent manner in both lines by preincubation with rapamycin (Rap) and, to a lesser extent, with FK506, but not with cyclosporin A. CaMK II and/or IV inhibitors KN62 and KN93 also inhibited progestin- and TA-stimulated transcription in both lines. None of these drugs had any effect on basal transcription. The antagonist RU486 inhibited all the effects of both progestin and TA, suggesting that 0888-8809/98/$3.00/0 Molecular Endocrinology Copyright © 1998 by The Endocrine Society

986

Immunosuppressants and CaM Kinase-Modulated Steroid-Regulated Transcription

from cells exposed to Rap, contained, in addition to the heat shock proteins of 90 kDa and 70 kDa (hsp90 and hsp70), the FK506-binding immunophilin FKBP52 but not FKBP51, although the latter was part of unliganded PR heterocomplex associated with hsp90. These results suggest that Rap and KN62 act upon the PR by distinct mechanisms, with only Rap impeding progestin-induced PR transformation. FKBP51 appeared to dissociate from the receptor heterocomplex, but not from hsp90, after hormone binding to PR in vitro and in vivo, whether in the presence or not of Rap and KN62. Immunoprecipitation experiments distinguished two PR- and glucocorticosteroid (GR)associated molecular chaperone complexes, containing hsp90 and hsp70 and FKBP52 or FKBP51. Another complex identified in T47D cytosol contained hsp90 and the cyclosporin A-binding cyclophilin of 40 kDa, CYP40, but not hsp70, PR, or GR. These observations support the concept that FKBP51 and FKBP52 can act as regulators of Rap and FK506 activity upon PR and GR-mediated transcription, a mechanism that could be also regulated by type II and/or type IV CaMKs. (Molecular Endocrinology 12: 986–1001, 1998)

INTRODUCTION In their unliganded form, steroid receptors from cytosolic extracts are complexed in large inactive heterooligomeric structures with several receptor-associated proteins (see Ref. 1 for a review). Among these proteins, two have been identified as heat shock proteins (hsps) namely hsp901 (2–4), and hsp70 (5, 6). Other receptor-associated proteins belong to a class of peptidyl-prolyl isomerases, termed immunophilins (7): two FK506-binding proteins, FKBP52 (8–11), which migrates at 59 kDa in denaturing gel electrophoresis and therefore was initially called p59 (12), and FKBP51 (13, 14), previously named FKBP54 (15). In addition to these FKBPs, a cyclosporin A (CsA)-binding protein of 40 kDa, cyclophilin 40 (CYP40) (16, 17) was charac1 Abbreviations used are: hsp90, heat shock protein of molecular mass 90 kDa; hsp70, heat shock protein of molecular mass 70 kDa; FKBP, FK506 binding protein; FKBP51, FK506-binding protein of apparent molecular mass 54 kDa; FKBP52, FK506-binding protein of apparent molecular mass 59 kDa; CYP40, cyclophilin of molecular mass 40 kDa; R5020, promegestone, 17a,21-dimethyl-19-norpregna-4,9diene-3,20 one [17a methyl-3H]; [3H] Org2058, 16a-ethyl-21hydroxy-19-nor [6,7-3H]-pregn-4-ene-3,20-dione; RU486, mifepristone, 11b-(4-dimethyl-aminophenyl)-17b-hydroxy17a-(prog-1-ynyl) estra-4,9-dien-3-one; dexamethasone, (11b, 16a)-9-fluoro-11, 17, 21, trihydroxy-16-methylpregna1,4-diene 3,20 dione; triamcinolone acetonide, 9a-fluoro11b,16a,17,21-tetrahydroxypregna-1,4 dienee, 3,20 dione cyclic 16,17-acetal with acetone; KN62, 1-[N,O-bis(5isoquinoline sulfonyl)-N-methyl-L-tyrosyl]-4phenylpiperazine. KN93, (2-[N-(2-hydroxyethyl)-N-(4methoxybenzenesulfonyl)]amino-N(-chlorocinnamyl)-Nmethylbenzylamine).

987

terized. These immunophilins, FKBP51 and FKBP52, as well as CYP40, bind at the same site on hsp90 (Refs. 14, 18; see Ref. 1 for a review). CYP40, FKBP51, and FKBP52 were the first of the cyclophilin and FKBP classes of immunophilins found to possess C-terminal amino acid sequence identities (10, 11, 13–17) in the form of three tetratricopeptide repeat units required for binding to hsp90 (19). The association of FKBP52, FKBP51, and CYP40 with hsp90 was demonstrated by coimmunoprecipitation (8, 12, 14–16, 20–23), native gel electrophoresis (19), cross-linking (20, 21), and immunosuppressant-affinity chromatography (14–18, 22–24) experiments. Therefore, FKBP52, FKBP51, and CYP40 may represent the first members of a new class of peptidyl-prolyl isomerases, the heat shock protein binding immunophilins (25). They all exist in independent heterocomplexes with hsp 90 (14, 18, 24), in association with other proteins thought to participate in the folding process of the receptor (see Ref. 1 for a review). Hsp70 has also been shown to be part of heterooligomeric complexes containing hsp90 and FKBP52 that are associated with the PR in human breast cancer T47D cells (26, 27). In fact, large molecular chaperone complexes including FKBP52, FKBP51, hsp90 and, in some cases, hsp70 seem to exist in many cells. Up to now, these receptorassociated immunophilins are thought to be involved in a molecular chaperone mechanism, and although it has been recently proposed that FKBP 52 could serve as a mediator of dexamethasone (Dex)-induced glucocorticosteroid (GR) trafficking (28), their targets in the receptor complex(es) are still not well identified. Recent studies have suggested that FKBPs and cyclophilins modulate Dex-induced transcription in mouse fibroblasts since immunosuppressants potentiate steroid-induced transcription in these cells (24, 29). Similarly, FK506 potentiates progestin-induced transcription in yeast strains transfected with human PR B form (30). This potentiation was shown recently to be the consequence of the reversal activity of a membrane efflux mechanism resembling, but different, from that of phospho-glycoprotein, P-gp (31–33). Then, at least in mouse fibroblasts and in yeast, it is obvious that the targets of the effects of immunosuppressants on steroid actions could not explain the observed cellular responses. In this work, we used two lines derived from T47D cells, known to express high levels of PR (34), to study the effects of immunosuppressants on steroidinduced transcription. The lines contained stably transfected plasmids encoding CAT reporter genes under control of the mouse mammary tumor virus long terminal repeat (MMTV-CAT) (35) or the synthetic steroid-inducible minimal promoter GRE5 (GRE5tkCAT) (36). In the two T47D cell lines, Rap and FK506, but not CsA, inhibited progestin-, as well as triamcinolone acetonide (TA)-, induced transcription. This occurred at a step preceding steroid-induced receptor activa-

MOL ENDO · 1998 988

Vol 12 No. 7

Fig. 1. Effects of Immunosuppresssants Rap, FK506, and CSA on R5020-Induced CAT Expression in T47D MMTV-CAT Cells A, Cells were exposed as indicated for 2 h to 10 mM of immunosuppressants, followed by the indicated concentrations of R5020 for 16 h. CAT activity was calculated relative to the basal activity measured in the absence of steroid. The results are the mean values for at least three determinations in duplicate. B, Cells were exposed to the indicated concentrations of immunosuppressants then to 0.1 nM (filled bars) or 1 nM (open bars) R5020. CAT activity was measured as in Fig. lA. The CAT activity calculated for each drug concentration is the result of a decreased activated level of CAT and not that of an increased background. Control experiments with only 0.1 and 1 nM R5020 alone are shown (bars, C). Similar results were obtained when drugs and steroid were added at the same time.

tion.2 Similar inhibiting activity by Rap was also observed on progestin-induced Cyclin D1 mRNA levels, a progestin-regulated gene (37, 38). Numerous works have established that immunophilins play a role in cell-signaling pathways involving phosphorylation (Ref. 39 for a review). Since we had previously found that rabbit uterus FKBP52 copurifies with either a Mg21 or a calcium-calmodulin dependent kinase (CaMK) (40), and since two CaMK consensus sequences are present in FKBP52 (10, 11) and FKBP51 (14, 41), we tested the effect of CaMK inhibitors on steroid-induced transcription. The two CaMK inhibitors, KN62 and KN93, inhibited the steroid-induced transcription in both cell lines, at a step upstream to steroid-induced receptor activation; in addition, Rap and KN62 inhibited progestin-induced expression of Cyclin D1 mRNA. This suggests that a CaMK phosphorylation/dephosphorylation-mediated mechanism regulates steroidinduced transactivation before the steroid-induced dissociation of receptor-associated hsps/FKBP complexes from the receptor, that immunosuppressant ligands for FKBPs 51 and 52 act at a step preceeding such a dissociation process, and that two exogenous and endogenous genes are similarly regulated by Rap and by CaMKs.

RESULTS Inhibition by Rap and FK506 of ProgestinInduced Transcription in T47D(MMTV-CAT) Cells The maximal induction of CAT activity in T47D (MMTVCAT) cells (; 200 fold that of the basal level) occurred at 100 nM R5020 (Fig. 1A). When cells were preincubated with a single concentration (10 mM) of Rap, before addition of a range of R5020 concentrations, an inhibition of CAT activity was observed (Fig. 1A). FK506 was a weaker inhibitor than Rap, and CsA did not seem to affect CAT gene expression. At low R5020 concentrations, a dose-dependent inhibition by Rap and FK506 was obtained (Fig. 1B). The inhibitory effect of Rap appeared more pronounced than that obtained with FK506. CsA had no significant inhibitory effect on CAT activity even when used at 10 mM (Fig. 1B). None of the drugs had any effect on CAT activity when used alone (not shown). A similar inhibitory effect of Rap was obtained, whether the immunosuppressant was added before or at the same time as the progestin (Fig. 1B). All these effects were abolished by a 100-fold excess of the antiprogestin RU486 (not shown), indicating that induction of CAT activity is mediated by the PR. Effect of Immunosuppressants on ProgestinInduced Transcription in T47D (GRE5-CAT) Cells

2 The term “activation” relates to the ligand binding-induced process by which a receptor converts from a non-DNA binding form to a DNA-binding form, while the term “transformation” identifies the dissociation process of molecular chaperone proteins from the receptor heterocomplexes.

In T47D-GRE5-CAT cells, the maximum CAT activity (;200 times that of the basal level) was obtained for ;3 nM R5020 (36) or 0.3 nM Org 2058 (not shown). Figure 2 shows that, similar to T47D MMTV CAT cells,


989

steroid-inducible CAT activity in both cell lines. Figure 4A shows that KN62 decreased R5020-induced transcription in a dose-dependent manner in both T47D cell lines. Similarly, KN62 also inhibited TA-induced transcription in T47D (GRE5)-CAT (Fig. 4B), as well as in T47D (MMTV)-CAT (not shown). KN93 inhibited both R5020- and TA-induced CAT activity in both cell lines but at a 2- to 3-fold lower concentration than KN62 iself (Fig. 4C). This difference may reflect the greater water solubility of KN93 (44). These results suggest that a CaMK activity is involved in the steroid-induced activation of either PR and GR and that its inhibition decreases steroid-induced transcription in T47D cells. Fig. 2. Inhibition by Rap and FK506 of R5020-Induced CAT Expression in T47D (GRE)5-CAT Cells T47D cells stably transfected with the reporter gene GRE5CAT were exposed for 16 h to 0.1 or 10 nM R5020 before (C) or after exposure to different concentrations (0.1 to 10 nM) of Rap, FK506, or CsA. CAT activity was measured as described in the legend of Fig. 1.

Rap strongly inhibited progestin-induced transcription, in a dose-dependent manner. Once again, Rap was more efficient than FK506 and CsA had no effect. Addition of a 100-fold molar excess of RU486 blocked progestin-inducible CAT activity (not shown), strongly suggesting that the observed gene expression in both cell lines is mediated by the PR. Effect of Immunosuppressants on Glucocorticosteroid-Induced Transcription In the cell line T47D (GRE5)-CAT, TA induced transcription in a dose-dependent manner (Fig. 3A). The maximal induction of CAT activity occurred at 10 nM. This maximum reached almost 150-fold that of the basal level and was inhibited by RU486 (not shown). Figure 3B shows that exposure of cells to Rap or FK506 provoked a decreased response to either 1 or 10 nM TA. CsA, on the other hand, did not affect TA-induced CAT activity. Effect of Ca21/Calmodulin-Dependent Protein Kinase Inhibitors on Progestin- and TA-Induced Transcription in Both T47D Cell Lines The immunosuppressants FK506 and CsA bind to a variety of immunophilins in cells and act on Calcineurin (CN) by inhibiting its phosphatase activity (for a review, see Ref. 39). However, CN is not a target of Rap, which has been shown to modify the signaling pathways mediated by the expanding family of lipid/protein kinases (review in Ref. 42). In addition, we showed that FKBP52 from rabbit uterus is phosphorylated in vitro by a copurified Ca21/calmodulin-like dependent protein kinase (CaMk) activity (40). We have therefore tested the effects of KN62 and KN93, two specific inhibitors of both types II and IV CaMks (43, 44), on

Effects of Rap and KN62 after Exposure of Cells to Steroids All the experiments described above were performed by adding drugs (immunosuppressants or CaMK inhibitors) before steroids in cell culture medium. Figure 5 shows that if Org 2058 (0.33 or 1 nM) was added to T47D (GRE5)-CAT cells before addition of either KN62 or Rap, no decrease of CAT gene expression was observed. It has been established that hormone binding induces a conformational change in receptor structure, which leads to dissociation of molecular chaperones from the hormone binding unit(s) (see Ref. 1 for a review). Thus, the data from Fig. 5 may indicate that the drugs Rap and KN62 act before this transformation process. Similar observations were made in T47D MMTV-CAT cells, as well as with TA instead of Org 2058 in both cell lines (not shown). Effects of Immunosuppressants and CaMK Inhibitors on Progestin-Induced mRNA Synthesis The mRNA encoding cyclin D1 is one of many transcripts whose levels are increased in T47D cells in the presence of progestin (37, 38). We analyzed the influence of immunosuppressants and KN62 on cyclin D1 mRNA level. Rap, FK506, or KN62 had no individual effect on the abundance of Cyclin D1 transcripts. By contrast, Rap and KN62 stongly inhibited the R5020induced increase of Cyclin D1 mRNA, while FK506 had a weaker effect (Fig. 6). In a control experiment, the mRNA level of the ubiquitous non-progestin-regulated GAPDH was not modified by any of the drugs tested under these conditions. These results would indicate that Rap, FK506, and KN62, but not CsA, down-regulate the progestininduced transcription of the endogenous Cyclin D1 gene. Effects of Rap and KN62 on the Heterooligomeric Forms of PR and GR Given the drug effects reported above, we next sought to investigate whether stabilization of an inactive nonDNA-binding form of steroid receptors could be responsible for these effects. In the absence of hor-

MOL ENDO · 1998 990

Vol 12 No. 7

Fig. 3. Effect of Immunosuppressants on Glucocorticosteroid-Induced CAT Expression in T47D (GRE)5-CAT Cells A, Cells were exposed for 16 h to different concentrations (10 pM to 1 mM) of TA, and CAT activity was measured for 50 mg of protein as described in Fig. 1. B, Cells were treated for 2 h with 5 mM of Rap or FK506 or CSA before exposure for 16 h to 1 nM (filled bars) or 10 nM (open bars). CAT activity was measured and compared with that of control experiments performed without immunosuppressants (C).

mone, PR is located in the cell nuclei, from which it can be recovered in the low-salt cytosoluble fraction of cellular homogenates. Under these conditions, PR can be postlabeled with 3H-labeled ligands, and ligand/PR complexes migrate at 9S in density gradients. They contain PR hormone-binding units, hsp(s), and immunophilin(s). In vivo, after exposure to hormone, the PR is tightly bound to nuclei and can be extracted by high salt-containing buffers (.0.4 M NaCl), giving rise to 4–5S complexes that are free of receptor-associated proteins (see Ref. 1 for a review). Molybdate ions have been widely shown to protect 9S species from the in vitro transformation into 4–5S species induced by high salt. We have found that tungstate oxyanions are more efficient than molybdate in preventing such a saltinduced transformation (20, 45). In the presence of tungstate, 9S nuclear [3H]RU 486-PR species were recovered in which hsp90 and FKBP52 still remained associated with the PR units (45). Here, we analyzed tungstate-stabilized complexes extracted from cells treated with Rap. T47D cells were exposed to 10 mM Rap in the presence of 10 nM [3H]Org 2058. The tungstate-stabilized nuclear form of [3H]Org 2058-PR complexes extracted in high salt separated in two unequal species in density gradients (Fig. 7C), one sedimenting at approximately 9S, the other at approximately 4–5S. The former represents approximately 40% of the total [3H]Org 2058-PR complexes. In contrast, in the absence of steroid and of any drug, identical 9S sedimentation profiles for both cytosolic and nuclear PR were obtained after postlabeling with the 3H-labeled ligand (Fig. 7A). However, a dramatic loss of hormone binding capacity occurred, as a consequence of the instability of unoccupied PR during cellular extraction and homogenization procedures. In the absence of Rap, and in the presence of [3H]Org 2058, only the ;5S form of nuclear PR was obtained with a small shoulder at 9S position (Fig. 7B). In contrast to treatment with Rap, treatment with KN93 or CsA (Fig. 7, D

and E) had no effect on the density gradient profile of PR-[3H]Org 2058 complexes extracted by high salt. Under these conditions, there was still a small portion of the PR that was recovered in the cytosolic extracts (Fig. 7, B, D, and E). Identical results were obtained with [3H]TA-GR complexes extracted from nuclei of T47D(GRE5)-CAT cells exposed to the same drugs (not shown). These results indicate that Rap would act as a stabilizer of the 9S species (Fig. 7C). Numerous studies from many laboratories have established that the 9S form of PR represents a combination of complexes of the hormone binding B and A PR forms associated with the molecular chaperone hsp90 and FKBPs (review in Ref. 1). In addition, the unliganded rabbit uterus 9S-PR has been shown to be stabilized in vitro by FK506 and Rap (22). Similarly, the untransformed GR complex from S49 lymphocytes was also stabilized in vitro by FK506 (46) and did not interact with nonspecific DNA, consistent with the behavior of the GR compatible with the presence of GR in complexes with associated hsp(s) and immunophilin(s). We conclude from experiments of Fig. 7 that Rap may act in T47D cells by impairing the agonist-induced release of molecular chaperones from the PR. In contrast, KN93 itself does not stabilize the PR and seems to act differently than Rap (Fig. 7D). Protein Composition of PR Heterocomplexes from T47D Cells Immunoprecipitation experiments were used to identify the different proteins present in PR heterocomplexes. Cytosol from T47D-(GRE5) CAT cells was incubated with [3H]Org 2058. The [3H]Org2058-PR B complexes were immunoprecipitated with the MPR1 antibody, and the radioactivity in the immunoprecipitate was measured by the dextran-coated charcoal technique described previously (47). Specifically bound [3H]Org 2058 was found in the immunopellet


991

Fig. 4. Effect of CaMK Inhibitors on R5020- and TA-Induced CAT Gene Expression in T47D Cells Different concentrations of KN62 were added for 2 h to the culture medium of both cell lines before exposure for 16 h to 1 nM R5020 (panel A), or TA (panel B). In panel C, different concentrations (0.1 to 1 mM) of KN93 were incorporated as indicated for 2 h in the culture medium of T47D-(GRE5)-CAT cells, followed by addition for 16 h of either 1 nM R5020 (left) or TA (right). CAT activity was measured as in Fig. 1.

(Fig. 8A). When immunoprecipitation was performed with the anti-FKBP52 EC1 antibody, 20% of the initial PR binding was immunoprecipitated. No radioactivity was recovered in immunoprecipitates of similar experiments performed with either the anti-CYP40 638W antibody or with the anti-FKBP51 HI51 antibody, or with control nonimmune rabbit or mouse antibodies. This suggested that PR-B was absent from immunopellets. This was confirmed by Western blot experiments (Fig. 8B). The PR-B form was detected in the mPRI and in the EC1 immunopellets, along with hsp90 and hsp70, arguing for the presence of cytosolic complexes including PR-B, hsp90, hsp70, and FKBP52 in T47D cells. Neither the PR-B form nor hsp70 was detected in the 638W immunopellets, whereas CYP40 and hsp90 were readily detected. This indicates that CYP40 is neither associated with PR-B nor with hsp70 in T47D cells cytosol. Western blot analysis with MA1–

140 antibody of 638 W immunopellets did not detect PR-B or PR-A (not shown), confirming that no PR is associated with CYP40 in cytosol of T47D cells. Similarly, FKBP51 was not associated with PR either in mPRI or HI51 immunopellets, but it coprecipitated with hsp90. FK506- and CsA-Affigel Affinity Chromatography Data from Fig. 8 suggested that PR from T47D associates with only one immunophilin-hsp complex. This may appear to contrast with early published results (14), indicating a preferential association of PR with FKBP51 in a cell-free reconstitution system (14). Therefore, we checked for the protein composition of PR from T47D cells exposed to immunosuppressants or to PR ligands or to both, by FK506 and CsA affinity chromatography, performed as described in Materials

MOL ENDO · 1998 992

Vol 12 No. 7

Fig. 5. Effect of Immunosuppressive Drugs after Steroid Incorporation in Cell Culture Medium T47D (GRE5) CAT cells were grown as indicated for 4 h in culture medium in the presence of 0.33 or 1 nM Org 2058. Then the medium was replaced for 2 h by steroid-free medium (2 h) or medium containing either 10 mM KN62 (KN62 2 h) or 10 mM Rap (Rap 2 h). Control experiments in which 10 mM KN62 and 10 mM Rap were added 2 h before 4 h of incubation of cells with 0.33 and 1 nM Org 2058 are shown (black bars). CAT activity was measured as in Fig. 1.

and Methods with T47D cytosolic extracts in which the PR had been labeled with 10 nM [3H]RU486. Specific [3H] RU486 binding was determined by the dextrancoated charcoal technique (47). Complexes of [3H]RU486-PR were retained by and eluted from the FK506-Affigel column (Fig. 9, upper panel), but not from the CsA-Affigel or the non-grafted control Affigel 10 columns. Western blot experiments (Fig. 9, A–E) confirmed these results, as no PR-A and PR-B forms were detected in the CsA eluate from the CsA-Affigel column. In contrast, both PR forms were detectable in the FK506 eluate from FK506-Affigel (see lane 4 in Fig. 9A). In addition, FKBP52 was almost totally retained on the FK506-Affigel resin and eluted as a doublet, concomitantly with hsp90 and hsp70, as revealed by Western blot using EC1 and 174 antibodies, respectively (Fig. 9, A and C, left panels, lane 4). Similarly, CYP40 and hsp90 coeluted from the CsAAffigel column (Fig. 9C, right panel, lane 4). Since no radioactivity or immunoreactive PR were found in the CsA eluates, we concluded that PR from T47D cells is not associated with CYP40, or that high-salt exposure provoked dissociation of CYP40 from the PR-hsp90CYP40 complex. However, both FKBP52 and CYP40 were found associated with hsp90, and exposure to immunosuppressant did not disrupt this association. Confirming the results of Fig. 8, hsp70 was not found to be associated with CYP40, (Fig. 9, panel C, right, lane 4). FKBP52, PR-A, PR-B, hsp90, and CYP40 were all found in the flowthrough of the control Affigel-10 column (Fig. 9, B and D, lane 1), none of them being detectable in the eluates (Fig. 9, B and D, lane 4).

Fig. 6. Northern Blot of GAPDH and CYCLIN D1 T47D (GRE5) CAT cells were grown to 50% confluence in T150 flasks in DMEM supplemented with 10% FCS. The medium was then replaced by DMEM without serum for 48 h, followed by DMEM supplemented with 10% stripped and heat-inactivated serum containing or not (control) 10 mM of the indicated drugs for 2 h; then 10 nM R5020 was added for 16 h to the culture medium. RNA was isolated with Trizol reagent, and Northern analysis was performed to determine Cyclin D1 and GAPDH expression as described in Materials and Methods. Typical GAPDH signals were achieved after 24 h of autoradiography, and Cyclin D1 signals were achieved after 72 h of autoradiography (panel A). Using densitometric analysis, GAPDH was used to normalize for variations in signal intensity. The autoradiogram shown was typical of three Northern blots.

FKBP51 was eluted from FK506-Affigel resin, loaded with unlabeled-PR containing cytosol concomitantly with both PR-A and PR-B (Fig. 9, right part of panel E). Similar experiment performed with cytosol containing [3H]RU486-PR complexes, revealed that FKBP51 did not coeluted with PR from the FK506 column; both forms of PR were found in the flowthrough fraction (Fig. 9, left panel). Some FKBP51 did not bind to the resin possibly due to overloading of the column and/or to loading conditions using buffer containing low salt concentration; FKBP51 binds in fact, better to FK506 in high salt rather than in low salt buffers (15), but to maintain as much as possible the integrity of the heterocomplex(es) which dissociate in high salt (1, 5, 6, 20), we chose to add 0.15 M NaCl in loading buffer as a compromise. These data, along


993

Fig. 7. Effect of Rap, CSA, and KN62 upon the Size of Nuclear PR Complexes from T47D Cells T47D (GRE5-CAT) cells were exposed for 2 h (panels C–E) or not (panel A and B) to 10 mM of Rap (panel C) or 10 mM of KN93 (panel D), or 10 mM CsA, (panel E), before addition of 10 nM [3H]Org 2058 for 16 h (panels B–E). After cells were harvested, cytosols and nuclear extracts were prepared as described in Materials and Methods in a buffer containing always 20 mM sodium tungstate ions to maintain the heterooligomeric PR structure. Samples from cells not exposed to drugs and steroid (panel A) were incubated for 4 h at 0 C with 10 nM [3H]Org2058 and treated with a dextran-coated charcoal suspension as described in Ref. 47, before loading on top of preformed gradients. Aliquots (100 ml) of both cytosol (Œ O Œ ) and nuclear extracts (F O F) were loaded on 5–20% sucrose preformed gradients in buffer A containing 20 mM tungstate ions and ultracentrifuged as described in Materials and Methods. Positions of internal standards GO (7.9)s and PO (3.6)s are indicated.

with those shown in Fig. 7, support the concept that hormone-binding to PR induces FKBP51 release from the heterocomplex, in agreement with previous report from Smith et al. (15). Identical experiments were performed with cytosol from T47D (GRE5) CAT cells, in which the GR had been labeled with [3H]TA. In this case, [3H]TAGR-hsp90FKBP52 and unlabeled GR-hsp90-FKBP51 complexes were characterized after the same western blot technique as that described in Fig. 9. TA binding to the GR, like progestin binding to the PR, induced a release of FKBP51 from the receptor heterocomplex (not shown). Similar to the PR, the GR was not found associated with the CYP40-hs90 complex (not shown). Altogether, these results indicate the presence of PR and GR in different molecular chaperone complexes, including always hsps plus cochaperones such as FKBP51 and FKBP52 but not CYP40. Stabilization of FKBP52-(hsp90-PR) Complex by Rapamycin in Cells Exposed to PR Ligands [3H]Org2058-PR complexes extracted from T47D cells exposed or not to Rap and KN62 (instead of KN93), were immunoprecipitated with the anti PR antibody

MA1 140. After treatment of cells with Rap and steroid, FKBP52 coimmunoprecipitated with PR B form (Fig. 10, panels A and C), along with hsp90 and hsp70 (panel B), but not with FKBP51 (panel D) or CYP40 (panel E). By contrast, no FKBP and no hsp coimmunoprecipitated with PR when cells were exposed to progestin alone, to CsA plus progestin or to KN62 plus progestin. These results supported the results obtained in Fig. 7, which suggested that Rap, but not KN62, may stabilize the FKBP52-hsp90-PR complex. In contrast, Rap did not stabilize a FKBP51-hsp90-PR complex which appeared to dissociate after steroid binding to the PR.

DISCUSSION In previous reports, immunosuppressive drugs such as FKBP’s (FK506 and rapamycin) and cyclophilin ligands (CsA and analogs) were shown to potentiate the Dex-induced expression of the MMTV-CAT reporter gene (24, 29) in a stably transfected mouse fibroblast LMCAT cell line, (29). This potentiation was demonstrated to be the consequence of the reversal by im-

MOL ENDO · 1998 994

Vol 12 No. 7

Fig. 8. Immunoprecipitation of PR-Associated Proteins with anti PR-B and anti- Immunophilin Antibodies Aliquots (500 ml) of [3H]Org2058-PR complexes were incubated with pellets of preformed protein G-Sepharose IgG complexes as described in Materials and Methods: mPR1, protein G-Sepharose, mPR1 (10 mg/100 ml); mouse n.i., protein G-Sepharose mouse nonimmune IgGs (10 mg/100 ml); 638W, protein G-Sepharose anti-CYP40 638W antibody (20 mg/100 ml); rabbit n.i., protein G-Sepharose rabbit nonimmune IgGs (10 mg/100 ml); EC1, protein G-Sepharose anti-FKBP52 EC1 antibody (10 mg/100 ml); HI51, 205 protein G-Sepharose anti-FKBP51 antibody (20 mg/100 ml). After washing of the immunopellets according to the protocol described in Materials and Methods, immunopellets were counted for determining the amount of bound [3H]Org2058 (panel A), boiled in Laemmli sample buffer, and analyzed on 10% or 12% SDS-PAGE (panel B) before Western blotting. A, Aliquots of initial labeled cytosol, supernatants (1), high salt (2), low salt (3) washes of the immunopellets and immunopellets (4) were treated with dextran-coated charcoal for measurement of [3H]Org2058 binding. The bars represent the total amount of bound [3H]Org2058 in each sample. B, Western blotting of PR B and coimmunoprecipitated proteins after 12% SDS-PAGE. PR B was revealed by mPR1 (1 ml/ml); hsp90 and hsp70 were revealed by 174 (1:6000); FKBP52 was revealed by 173 (1:10000); the rabbit anti C-terminal anti-FKBP52 was used to avoid recognition of the heavy chains of mouse IgGs that occurs when MPRI, mouse nonimmune (mouse n.i.), and EC1 antibodies are used. FF1 (2 mg/ml) was used to detect FKBP51, and 638W (1:6000) was used to detect CYP40. The ECL chemiluminescent technique was employed.

munosuppressants of an efflux membrane mechanism resembling P-gp(s) which led to increased intracellular steroid concentrations in yeast (31, 32), and in mouse fibroblasts (33). These observations called into question any role of receptor-bound immunophilins in steroid-induced gene transactivation. In this study, we observed that two human T47D breast cancer cell lines, stably transfected with the same promoter/ reporter construct MMTV-CAT (35) or a GRE5-CAT reporter (36), respond differently from LMCAT cells when exposed to immunosuppressants. In fact, instead of a potentiation of the steroid-induced transcription, an inhibition of glucocorticosteroid and progestin-induced CAT activity was observed after

exposure of T47D cells to Rap or FK506 (Figs. 1–3). This effect was also observed with the endogenous Cyclin D1 gene. All glucocorticosteroid- and progestin-inducible gene expression observed in these lines was inhibited by RU486, supporting a direct role of the GR and the PR. By contrast, no change in the expression level of GAPDH transcripts, a non-steroid-regulated gene was observed in the presence of Rap, FK506 and KN62. The effect of Rap correlated with a stabilization of 9S receptor-associated complex(es), strong suggesting the observed inhibition is not due to modulation of P-gp activity. In addition, results obtained with T47D GRE5CAT cells led us to conclude that the inhibition of steroid-induced transcription by


995

Fig. 9. Identification of Immunophilins and Immunophilin-Associated Proteins Eluted from Immunosuppressant-Affinity Chromatography by Western Blot Aliquots (500 ml) of [3H]RU486-labeled cytosolic PR from T47D MMTV-CAT cells, in the presence of 1 mM cortisol, were incubated with 0.2 ml FK506-Affigel, or 0.2 ml CsA-Affigel, or 0.2 ml Affigel control; the chromatography was performed as described in Materials and Methods. In the diagram of the figure, the total [3H]RU486 binding was measured with the dextran-coated charcoal technique. C, Control of initial cytosol: 1, flow through of each column; 2, high-salt wash; 3, final low-salt wash; and 4, pooled FK506 or CsA-eluted fractions. Panels A–E, After 10% SDS-PAGE and blotting, fractions (20 ml) 1 to 4 from FK506- and CsA-Affigel columns were probed (panels A and B) with the mixed mouse monoclonal antibodies EC1 (to identify FKBP52) and MA1–140 for PR B and A identification. In panels C and D, the rabbit antibodies 638W (to identify CYP40) and 174 antibody (to identify hsp90 and hsp70) were used. In panels B and D, samples from the control experiment performed with non-grafted Affigel 10 were analyzed. In panel E, [3H]RU486-PR containing cytosol (left part) and nonlabeled T47D cell cytosol (right part) were chromatographed onto FK506-Affigel; both PR B and A forms were probed, concomitantly with FKBP51, in each fraction, with MA1 140 and FF1 mouse antibodies, respectively. The procedure described for detection of antigen-antibody complexes in the Vectastain manufacturer protocol was used. Positions of Mr markers are indicated on the left of panels A, C, and E and on the right of panels B and D.

MOL ENDO · 1998 996

Vol 12 No. 7

Fig. 10. Protein Composition of PR from T47D Cells Exposed to Immunosuppressants or CaMK Inhibitors An experiment identical to that described in Fig. 7 was performed, and whole cell extracts (250 ml) were immunoprecipitated with 100 ml protein G Sepharose to which 20 mg MA1 140 anti-PR antibody had been covalently cross-linked by the method described in Ref. 74; immunoprecipitated proteins were eluted at elevated pH as described in Materials and Methods. After 10% SDS-PAGE, antigens were probed with the following antibodies: mPR1, for detection of PR B (panel A), 174 for detection of hsp90 and hsp70 (panel B), EC1 for detection of FKBP52 (panel C), FFI for detection of FKBP51 (panel D), and 638W for detection of CYP40 (panel E). Each lane in each panel contains 40 ml of eluted proteins: 1, immunoprecipitate from control cells not exposed to steroid and drug; 2, immunoprecipitate from cells exposed to 10 nM [3H]Org2058; 3, immunoprecipitate from cells exposed to 10 mM Rap and then to radioactive progestin; 4, immunoprecipitate from cells exposed to 10 mM KN62 and then to radioactive progestin; 5, immunoprecipitate from cells exposed to 10 mM CsA and then to 10 nM [3H]Org 2058. Antigen-antibody complexes were detected with the biotinylated second antibody according to the Vectastain protocol. The figure represents a summary of two typical experiments.

immunosuppressants involves recruitment of transcription factors specific for GRE/PRE sequences and not those specific to other sequences present in the MMTV promoter. Moreover, no changes in levels of PR or GR, or their affinities for hormone was observed in drug-treated cells (S. LeBihan, unpublished results). In T lymphocytes, a number of genes was reported to be inhibited by immunosuppressants (39, 42). In contrast, in non lymphoid cells, only a few genes were affected, such as a cAMP-responsive element (CRE)regulated gene (48), the PRL promoter (49) and the gene coding for the orphan steroid receptor NUR77 (50). In these cases, an inhibition of transactivation by immunosuppressive drugs was observed which was suggested to involve the smaller forms of immunophilins, FKBP12 and CYP18 (51, 52). The inhibition occurred at drug concentration at least 100-fold smaller than those necessary to inhibit CAT activity in the studies presented here. In all cells, the common feature of CYP18-CsA and FKBP12-FK506 complexes is binding and inhibiting CN protein phosphatase activity (53, 54). In contrast, the FKBP12-rapamycin complex does not block CN activity (Ref. 42, for a review) but interrupts, although indirectly, the activation of the 70-kDa S6 kinase (55, 56) and p34cdc2 kinase activity in vivo (57). These observations provide a mechanism for cell cycle arrest by Rap. More recently, the mam-

malian target of Rap-FKBP12 complexes was identified as a 280-kDa molecular mass protein (58–60) homologous to the yeast product of TOR2 gene, which has phosphatidylinositol kinase activity (61). However, this kinase activity is not directly inhibited by binding of Rap to FKBP12. Rather, it was suggested that Rap interferes with the association or phosphorylation of a G1 effector (62). All these findings indicate that immunosuppressant-immunophilin complexes play a role in cell-signaling pathways involving phosphorylation of a number of substrates. It is therefore tempting to speculate that, by analogy, receptor-associated immunophilins could influence the phophorylation of one or several proteins of the receptor heterocomplexes or may act upon effectors of some phosphorylation/dephosphorylarion process involved in steroid receptormediated signaling. To test this hypothesis, we analyzed the effects of different kinase inhibitors on receptor signaling. We found that wortmannin, an inhibitor of PI3 and PI4 kinases, as well as the protein kinase A inhibitor H89, did not modify either progestin- and glucocorticosteroid-induced CAT gene activity in the two cell lines studied (not shown). However, KN62 and KN93, specific inhibitors of Ca21/CaM-dependent kinases of type II and of type IV, inhibited both progestin- and glucocorticosteroid-induced transcription in a dose-


dependent manner (Fig. 4). This may indicate that phosphorylation by a CaMK is a process that can modulate steroid receptor-mediated gene expression. This is compatible with the observed generation of GR-hsp 90 complexes incapable of binding hormone after in vivo treatment of LMCAT cells with Calmodulin antagonists (63). It is not known whether Rap or FK506, which also inhibits steroid-induced transcription in T47D cells, block the activity of CaMKs of type II and of type IV. It is interesting to note that these CaMK inhibitors potentiated Dex-induced expression of the CAT gene in LMCAT cells similarly to the effects of immunosuppressants in these cells (64). Thus, the effects of CaMK inhibitors and immunosuppressants with regard to steroid-induced transcription appeared strictly parallel in different cell types. CaMK inhibitors were shown to affect the Ca21 regulation of several immediate early genes such as NUR77 (50), supporting the concept of a broad role for CaMKs in transcriptional regulation. We have also shown that FKBP52 is a delayed early gene (65). It is therefore possible that a decrease of FKBP52 gene expression, which could result from CaMK inhibition, could induce a repression of steroidinduced gene transcription. Whether steroid-induced gene expression inhibited by Rap or other immunosuppressants or whether effects of CaMK inhibitors are interrelated remains to be established. But it is likely that both kinds of drugs affect pathways modulating the activity of (a) common factor(s). Such a hypothesis is consistent with the experiments of Figs. 5 and 7, which indicated that immunosuppressants and CaMK inhibitors act 1) before steroid-induced receptor activation and 2) by apparently two different pathways, since only Rap but not KN62 stabilizes the PR (and the GR) heterooligomer. In fact, unliganded PR and GR from T47D cells are contained in different heterocomplexes including immunophilins and hsps like PR and GR from other cells and/or tissues (Ref. 1 for review). Obviously, from the data of Fig. 9, the PR from T47D cells eluted from FK506-affinity chromatography and coimmunoprecipitated with FKBP51 anf FKBP52, both associated with hsp90 in separate complexes, but not with CYP40. However,CYP40 is present in T47D cell cytosol, associated with hsp90 similar to other cell types such as mouse fibroblasts (24) and lymphoids (63). Such an observation had been already made by Milad et al. (66) in T47D B11 subline stably transfected with the MMTV-CAT construct (67). In this subline, CYP40 did not coimmunopurify with PR, whereas CsA potentiated R5020-induced CAT activity. Surprisingly, Rap and FK506 had no effect and FKBP52 coimmunoprecipitated with PR (66), results that are the opposite of those obtained here. Thus, it is possible that specific cellular factors are recruited in hormone-induced transactivation of transfected exogenous genes, the level of expression of which may be quite different, as well as their integration at the chromatin level from one cell line to another one.

997

The targets of CaMK inhibitors in steroid-regulated gene transcription are not known. Like immunosuppressants Rap and FK506, they act before the ligand binding-induced conformational change of the receptors, which conducts to the release of receptor-associated hsp-90/immunophilins complexes. PR- and GR-FKBP52-hsp90 complexes have been shown to be stabilized by immunosuppressant binding (22, 46). In addition, the phosphorylation of FKBP52 by casein kinase II inhibits its binding to hsp90 (68). This suggests that interactions between the components of molecular chaperone complexes involved in steroid receptor regulation are likely controlled by a phosphorylation/dephosphorylation mechanism. It is quite possible that both CKII and CaMKs II or IV may intervene directly or not, in both the maturation of steroid receptors and the steroid receptor-mediated transactivation processes. Steroid receptors are known to be influenced by different signaling pathways such as cAMP, insulin-like growth factor-I, and kinase(s) and/or phosphatase(s) inhibitors (69–72). In addition to their immunosuppressive activity, FK506, Rap, and CsA (as well as some of their analogs), cause a variety of effects not only at the membrane level but also in calcium mobilization and cell-signaling cascades (41, 42, 51, 52). Their effects are different from one cell type to another and likely depend on the expression level of certain type(s) of immunophilin(s) and membrane transporter(s). The data from the present work indicate that CaMK inhibitors and immunosuppressants may serve as useful tools to elucidate some of the steps by which steroid receptors modulate gene transcription.

MATERIALS AND METHODS Drug Treatment of Cells Two T47D human breast carcinoma cell lines were used. One cell line was stably transfected with a MMTV-CAT reporter gene (35), while the other contained a synthetic steroidinducible promoter controlling CAT gene expression (GRE5tkCAT) (36). Cells were grown in 10-cm Petri dishes in DMEM containing 10% charcoal-treated and heat-inactivated (30 min at 55 C) FBS and 0.2 mg/ml geneticin (G418, Sigma), 24 h before drug and steroid treatment. Hygromycin B (75 mg/ml, Boehringer Mannheim, Indianapolis, IN) was used for T47D-GRE5 cells (36). Immunosuppressive drugs or kinase inhibitors were added for 2 h to the culture medium before steroid exposure when cells were at 50% confluence. Cells were harvested 16 h later, except when noted. CAT enzyme activity was measured in triplicate samples of the cell extracts containing 50 mg total protein, as previously described (24, 29) using [14C]chloramphenicol (98 mCi/mmol, ICN, Irvine, CA) as substrate and acetyl coenzyme A as cofactor. All experiments were repeated at least three times. In some experiments, steroids were added 4 h before or at the same time as immunosuppressants or inhibitors. Specific calcium/calmodulin kinase inhibitors KN62 and KN93 as well as H89 and wortmannin (Calbiochem, La Jolla, CA), inhibitors of protein kinase A, and PI3 kinase, respectively, were also used in other experiments.

MOL ENDO · 1998 998

Immunosuppressant-Affinity Chromatography Cytosoluble fractions of T47D cells were prepared by disrupting cells (5.109 cells) at 2 C with a glass/glass homogenizer in 1.75 ml buffer A (10 mM HEPES, 1 mM EDTA, 10% glycerol, 1 mM dithiothreitol, 20 mM Na2MoO4, pH 7.5), followed by centrifugation at 2 C in a Ti50 rotor [45 min at 45,000 rpm in a Beckman (Fullerton, CA) L2 68B ultracentrifuge]. The supernatant was incubated or not overnight at 2 C with 10 nM of the antagonist [3H]RU486 (38.4 Ci/mmol, ROUSSEL-UCLAF, Romainville, France) or [3H]TA (48 Ci/ mmol, Amersham, Bucks, England), or [3H]Org 2058 (51.9 Ci/mmol, Amersham). A 100-fold excess of radioinert cortisol or Org2058 was added to the incubates to block either glucocorticosteroid or progestin receptor binding in the case of progestins or glucocorticoids labeled cytosol samples, respectively. Portions of cytosol (500 ml) were chromatographed on 200 ml of packed dihydro-FK506-Affigel 10 or CsA-Affigel 10 columns (1 mg drug/ml gel). To improve FKBP51 binding to FK506-Affigel, 0.15 M NaCl was added to the cytosol sample. Control experiments used non-grafted Affigel-10 resins (Bio-Rad, Richmond, CA). The purification protocol described in Ref. 24 was employed and affinitybound proteins were eluted by exchange (3 3 30 min at 4 C) with 100 ml of the appropriate immunosuppressant solution (1 mM) in buffer A. Immunoprecipitation Experiments Cytosol from T47D cells in buffer A, incubated with 3H-labeled steroid as indicated above was treated with dextrancoated (vol/vol) charcoal (0.5, 5 g/100 ml) for measurement of ligand binding, and 0.5 ml aliquots were incubated (4 h, 4 C) with 10 mg of anti-human PR monoclonal antibody mPR1 also named LET126 (73) (Transbio, Paris, France) or MAI140 (Affinity Bioreagents, Neshanic Station, NJ), covalently crosslinked to protein G-Sepharose beads (Pharmacia, Piscataway, NJ), according to the method described in Ref. 74. Immunopellets containing the same amount of nonimmune IgGs were used as control. For CYP40 and FKBP51 immunoprecipitation, 500 ml of cytosol were incubated under the same conditions with 25 ml of 638W, a rabbit antiserum against the C-terminal peptide of human CYP40 bound to protein G-Sepharose immunoadsorbent, or 20 mg of HI51, a mouse monoclonal anti-FKBP51 efficient in native conditions (14). FKBP52 from T47D cell cytosol was immunoprecipitated by the same technique with EC1, a mouse monoclonal antibody specific for mammalian FKBP52 (75). After centrifugation (700 3 g for 5 min at 0 C), the pellets were washed by two successive centrifugations with 250 ml of buffer B (0.1 M potassium phosphate, 10% glycerol, 20 mM sodium tungstate, pH 7.5) and then three times with 250 ml of buffer A. The pellets were resuspended in 200 ml of Laemmli (76) sample buffer and boiled before denaturing SDS-PAGE and Western blotting experiments. In some experiments, immunoadsorbed proteins were eluted from the immunomatrix by a pH 10.2 buffer two times for 1 h and immediately neutralized by NaH2PO4. SDS-PAGE and Western Blotting Samples were resolved on either 10% or 12% SDS-PAGE, according to Laemmli (76). Western blots were performed as previously described (24). The nitrocellulose papers were probed with the following antibodies: the mouse monoclonal mPR1 for the detection of the human B form of PR (73), the mouse monoclonal MA1–140 for the detection of PR-A and PR-B isoforms (77). EC1 (75) and/or 173, a rabbit polyclonal antipeptide antibody (10), were both used for the detection of FKBP52, 638W (24), a rabbit polyclonal antipeptide (Asp356Asp370) antibody, for the detection of CYP 40. Antibody 174

Vol 12 No. 7

(24), a rabbit polyclonal antipeptide antibody raised against the C-terminal sequence (Met807-Glu823) of human hsp90. (78), recognizes both human hsp90 and hsp70 in Western blots due to sequence homologies (24). FKBP51 was identified with FF1 (a mouse monoclonal antibody working in denaturing conditions). All the antibodies were used at 10 mg Ig/ml, 174 and 638W antisera were used at 1:500 dilution. The detection of the antigen-antibody complexes was performed with a second biotinylated antibody (Vectastain ABC Kit, Vector Laboratories, Burlingame, CA). When the chemiluminescent detection system (ECL, Amersham Bucks, England) was employed, antibodies were used at 2 mg/ml or 1:6000 serum dilution. Nuclear Receptor Extraction and Sucrose Gradient Analysis The effects of immunosuppressants upon the size of the heterooligomeric PR were tested after exposure of T47D cells to Rap (10 mM) for 2 h before addition of 10 nM [3H]Org 2058. Cells were then harvested and washed three times in PBS, before being homogenized in buffer A (1 vol of buffer A for 1 vol of pelleted cells). The cytosol was obtained as mentioned above. The pellet remaining after cytosol preparation was washed three times with 5 vol of buffer A, before being homogenized in buffer A containing 0.4 M KCl plus 20 mM Na2WO4 (buffer B), as described previously (45). In these conditions no cytosol contamination is obtained since no lactate dihydrogenase activity can be detected (45). In particular experiments, whole cells extracts were obtained in buffer B after four freezing steps in liquid nitrogen and subsequent thawing at 20 C. The samples containing [3H]Org 2058-PR complexes were centrifuged for 2 h at 70, 000 rpm in 5–20% sucrose gradients in buffer A containing 20 mM Na2WO4 in a Vti 80 rotor (Beckman) with internal markers GO (glucose oxidase, 7.9S) and PO (peroxidase, 3.6S). Northern Blot Analysis RNA from T47D cells exposed to the different drugs was extracted using Trizol reagent (Life Technologies, Gaithersburg, MD) according to the manufacturer’s instructions. Total RNA (20 mg) of each sample was electrophoresed through a 1% denaturing agarose-formaldehyde gel and transferred to a nylon membrane (Hybond-NX, Amersham). RNAs were cross-linked to the membrane using a X-link apparatus (BioRad, Ilkirch, France) at 0.7 J/cm2. The membrane was prehybridized for 6 h in 50% formamide and then hybridized with probes labeled to high specific activity ('2.108 cpm/mg) using [a32P]dCTP by the random primer-labeling procedure (Amersham). Hybridization was carried out at 42 C in 50% formamide containing 53 Denhardts, 1% (wt/vol) SDS, 23 SSPE (225 mM NaCl, 16 mM NaH2PO4, 1.5 mM EDTA, pH 7.4) 0.25 mg/ml salmon sperm DNA (Sigma, St. Louis, MO). The membranes were washed two times in 13 SSC (15 mM NaCl, 1.5 mM sodium citrate, pH 7.0), 0.1% (wt/vol) SDS at room temperature for 20 min, followed by two 20-min washes in 0.13 SSC, 0.1% SDS at 55 C. The dry blot were exposed for 1–3 days at 280 C to Kodak X-OMAT AR-5 films, using intensifying screens. mRNA was quantitated by densitometry scanning using a Storm apparatus (Molecular Dynamics, Sunnyvale, CA). For Northern blots using Cyclin D1 probe, T47D cells were grown in serum-free DMEM for 48 h, to synchronize them after which the medium was replaced by DMEM containing stripped serum to which drugs had been added. Radioactivity Counting and Protein Determination Radioactivity was measured in a Minaxi tricarb (Packard, Downers Grove, IL) spectrometer, and proteins were determined with the BIORAD Kit using BSA as the standard.


Acknowledgments We are indebted to Fujisawa, Wyeth Ayerst and Sandoz Laboratories for the gift of immunosuppressives drugs. We thank Dr. A Cato for (MMTV-CAT)-T47D cells and Drs. D. O. Toft, D. Edwards and S. Nordeen for helpful discussion and communication of manuscripts before publication. We thank also Drs. D. Smith and R. Rimmerman for the gift of HI51 and FF1 antibodies, and L. Faber, R. Handshumacher, and K. Hoffman for the gift of antibodies EC1 and 63 8W, respectively. Dr. D. Philibert provided us with R5020 and RU486 and is acknowledged as well as Dr. D. Chalbos for the gift of the GADPH probe, and Dr E. Wientjens for the gift of the Cyclin D1 probe. P. Delangle, R. Villerot, and S. Oboukhoff are acknowledged for art work.

10.

11.

12.

13. Received June 9, 1997. Revision received February 11, 1998. Accepted March 13, 1998. Address requests for reprints to: Jack-Michel Renoir, faculte´ de Pharmacie URA 1218 CNRS, Pharmacologie Cellulaire, 5 Rue Jean-Baptiste Clement, Chatenay-Malabry France 92296. E-mail: [email protected]. This work was supported by INSERM, Centre National de la Recherche Scientifique, and Association pour la Recherche sur le Cancer, Grant 6510 (to J.-M.R.) and Grants MT11704 and MT-13147 from the Medical Research Council of Canada (to J.H.W. and S.M.), respectively. Partial financial support was obtained from the DGA Grant 94/064 (to J.-M.R.). S.L.B. and V.M. had fellowships from the Direction Generale aux Armeés and from the Ligue Nationale contre le Cancer.

14.

15.

16.

17.

REFERENCES 1. Pratt WB, Toft DO 1997 Steroid receptor interactions with heat shock protein and immunophilin chaperones. Endocr Rev 18:306–360 2. Joab I, Radanyi C, Renoir M, Buchou T, Catelli MG, Binart N, Mester J, Baulieu EE 1984 Common nonhormone binding component in non-transformed chick oviduct receptors of four steroid hormones. Nature 308:850–853 3. Catelli MG, Binart N, Jung-Testas I, Renoir JM, Baulieu EE, Feramisco JR, Welch WJ 1985 The common 90-kd protein component of non-transformed ‘8S’ steroid receptors is a heat-shock protein. EMBO J 4:3131–3135 4. Onate SA, Estes PA, Welsh WJ, Nordeen SK, Edwards DP 1991 Evidence that heat shock protein-70 associated with progesterone receptors is not involved in receptorDNA binding. Mol Endocrinol 5:1993–2004 5. Sanchez ER, Hirst M, Scherrer LC, Tang HY, Welsh MJ, Harmon JM, Simons SS, Ringold GM, Pratt WB 1990 Hormone-free mouse glucocorticoid receptors overexpressed in Chinese hamster ovary cells are localized to the nucleus and are associated with both Hsp70 and Hsp 90. J Biol Chem 265:20123–20130 6. Rehberger P, Rexin M, Gehring U 1992 Heterotetrameric structure of the human progesterone receptor. Proc Natl Acad Sci USA 89:8001–8005 7. Galat AM, Metcalfe SM 1995 Peptidyl prolyl cis/trans isomerase. Int Prog Biophys Mol Biol 63:67–118 8. Tai PKK, Albers MN, Chang H, Faber LE, Schreiber SL 1992 Association of a 59-kilodalton immunophilin with the glucocorticoid receptor complex. Science 256:1315– 1318 9. Yem AW, Tomasselli AG, Heinrikson RL, Zurcher-Neely H, Ruff VA, Johnson RA, Deibel MR 1992 The hsp56 component of steroid receptor complexes bind to immo-

18.

19.

20.

21. 22.

23. 24.

999

bilized FK506 and shows homology to FKBP12 and FKBP13. J Biol Chem 267:2868–2871 Lebeau MC, Massol N, Herrick J, Faber LE, Renoir JM, Radanyi C, Baulieu EE 1992 P59, an hsp 90-binding protein. Cloning and sequencing of its cDNA and preparation of a peptide-directed polyclonal antibody. J Biol Chem 267:4281–4284 Peattie DA, Harding MW, Fleming MA, De Cenzo MT, Lippke JA, Livingston DJ, Benasutti M 1992 Expression and characterization of human FKBP52, an immunophilin that associates with the 90 kDa heat shock protein and is a component of steroid receptor complexes. Proc Natl Acad Sci USA 89:10974–10978 Tai PKK, Maeda Y, Nakao K, Wakim NG, Duhring JL, Faber LE 1985 A 59-kilodalton protein associated with progestin, estrogen, androgen and glucocorticoidreceptor. Biochemistry 25:5269–5275 Baughman G, Wiederrecht GJ, Campbell NF, Martin MM, Bourgeois S 1995 FKBP51, a novel T-cell specific immunophilin capable of calcineurin inhibition. Mol Cell Biol 15:4395–4402 Nair SC, Rimerman RA, Toran EJ, Chen S, Prapapanich V, Butts RN, Smith DF 1997 Molecular cloning of human FKB51 and comparisons of immunophilin interactions with Hsp90 and progesterone receptor. Mol Cell Biol 17:594–603 Smith DF, Albers MW, Schreiber SL, Leach KL, Deibel MR 1993 FKBP54, a novel FK506 binding protein in avian progesterone receptor complexes and HeLa extracts. J Biol Chem 268:24270–24273 Kieffer LJ, Seng TW, Li W, Osterman DG, Handschumacher RE, Bayney RM 1993 Cyclophilin 40, a protein with homology to the p59 component of the steroid receptor complex. Cloning of the cDNA and further characterization. J Biol Chem 268:12303–12310 Ratajczak T, Carello A, Mark PJ, Warner BJ, Simpson RJ, Moritz RL, House AK 1993 The cyclophilin component of the unactivated estrogen receptor contains a tetratricopeptide repeat domain and shares identity with p59 (FKBP59). J Biol Chem 268:13187–13192 Owens-Grillo JK, Hoffman K, Hutchison KA, Yem AW, Deibel MR, Handschumacher RE, Pratt WB 1995 The cyclosporin A binding immunophilin CYP40 and the FK506-binding immunophilin hsp56 bind to a common site on hsp90 and exist in independent cytosolic heterocomplexes with the untransformed glucocorticoid receptor. J Biol Chem 270:20479–20484 Radanyi C, Chambraud B, Baulieu EE 1994 The ability of the immunophilin FKBP59-HBI to interact with the 90kDa heat shock protein is encoded by its tetratricopeptide repeat domain. Proc Natl Acad Sci USA 91:11197–11201 Renoir JM, Radanyi C, Faber LE, Baulieu EE 1990 The non-DNA-binding heterooligomeric form of mammalian steroid hormone receptors contains a hsp90-bound 59kilodalton protein. J Biol Chem 265:10740–10745 Rexin M, Busch W, Segnitz B, Gehring U 1992 Structure of the glucocorticoid receptor in intact cells in the absence of hormone. J Biol Chem 267:9619–9621 Renoir JM, Le Bihan S, Mercier-Bodard C, Gold A, Arjomandi M, Radanyi C, Baulieu EE 1994 Effects of immunosuppressants FK506 and rapamycin on the heterooligomeric form of the progesterone receptor. J Steroid Biochem Mol Biol 48:101–110 Hoffmann K, Handschumacher RE 1995 Cyclophilin 40: evidence for a dimeric complex with hsp90. Biochem J 307:5–8 Renoir JM, Mercier-Bodard C, Hoffmann K, Le Bihan S, Ning YM, Sanchez ER, Handschumacher RE, Baulieu EE 1995 Cyclosporin A potentiates the dexamethasone-induced mouse mammary tumor virus-chloramphenicol acetyltransferase activity in LMCAT cells: a possible role for different heat shock protein-binding immunophilins in

MOL ENDO · 1998 1000

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36. 37.

38.

39. 40.

glucocorticosteroid receptor-mediated gene expression. Proc Natl Acad Sci USA 92:4977–4981 Callebaut I, Renoir JM, Lebeau MC, Massol N, Burny A, Baulieu EE, Mornon JP 1992 An immunophilin that binds Mr 90,000 heat shock protein: main structural features of a mammalian p59 protein. Proc Natl Acad Sci USA 89:6270–6274 Estes PA, Suba EJ, Lauber-Heavner J, Elashry-Stowers D, Wei LL, Toft DO, Horwitz KB, Edwards DP 1987 Immunologic analysis of human breast cancer progesterone receptors. 1. Immunoaffinity purification of transformed receptors and production of monoclonal antibodies. Biochemistry 26:6250–6262 Wei LL, Sheridan PL, Krett NL, Francis MD, Toft DO, Edwards DP, Horwitz KB 1987 Immunological analysis of human breast cancer progesterone receptors. 2. Structure, phosphorylation and processing. Biochemistry 26:6262–6272 Czar M, Lyons R, Welsh M, Renoir J-M, Pratt W 1995 Evidence that the FK506-binding immunophilin heat shock protein 56 is required for trafficking of the glucocorticoid receptor from the cytoplasm to the nucleus. Mol Endocrinol 9:1549–1560 Ning YM, Sanchez ER 1993 Potentiation of glucocorticoid-mediated gene expression by the immunophilin ligands FK506 and rapamycin. J Biol Chem 268:6073– 6076 Tai PKK, Albers MW, McDonnell DP, Chang H, Schreiber SL, Faber LE 1994 Potentiation of progesterone receptor-mediated transcription by the immunosuppressant FK506. Biochemistry 33:10666–10671 Kralli A, Bohen SP, Yamamoto KR 1995 LEM1, an ATPbinding-cassette transporter, selectively modulates the biological potency of steroid hormones. Proc Natl Acad Sci USA 92:4701–4705 Kralli A, Yamamoto KR 1996 An FK506-sensitive transporter selectively decreases intracellular levels and potency of steroid hormones. J Biol Chem 271:17152– 17156 Marsaud V, Mercier-Bodard C, Fortin D, Le Bihan S, Renoir J-M, Dexamethasone, triamcinolone acetonide accumulation in mouse fibroblasts is differently modulated by the immunosuppressants Cyclosporin A, FK506, Rapamycin, their analogues, as well as by other P-Glycoprotein ligands. J Steroid Biochem Mol Biol, in press Horwitz KB, Mockus MB, Lessey BA 1982 Variant T47D human breast cancer cells with high progesteronereceptor levels despite estrogen and antiestrogen resistance. Cell 28:633–642 Cato ACB, Henderson D, Ponta H 1987 The hormone response element of the mouse mammary tumor virus mediates the progestin and androgen induction of transcription in the proviral long terminal repeat region. EMBO J 6:363–368 White JH, McCuaig KA, Mader S 1994 A simple and sensitive high-throughput assay for steroid agonists and antagonists. Biotechnology 12:1003–1007 Said TK, Conneely OM, Medina D, O’Malley, Lydon J 1997 Progesterone, in addition to estrogen, induces Cyclin D1 expression in the murine mammary epithelial cell, in vivo. Endocrinology 138:3933–3939 Musgrove EA, Hamilton JA, Lee CSL, Sweeney KJE, Watts CKW, Sutherland RL 1993 Growth factor, Steroid, and steroid antagonist regulation of cyclin gene expression associated with changes in T-47D human breast cancer cell cycle progression. Mol Cell Biol 13:3577– 3587 Wiederrecht G, Etzkorn F 1994 The immunophilins. Perspect Drug Discovery Design 2:57–84 Le Bihan S, Renoir JM, Radanyi C, Chambraud B, Joulin V, Catelli MG, Baulieu EE 1993 The mammalian heat shock protein binding immunophilin (p59/HBI) is an ATP and GTP binding protein. Biochem Biophys Res Com-

Vol 12 No. 7

mun 195:600–607 41. Baughman G, Wiederrecht GJ, Chang F, Martin MM, Bourgeois S 1997 Tissue distribution and abundance of human FKBP51, an FK506-binding protein that can mediate calcineurin inhibition. Biochem Biophys Res Commun 232:437–443 42. Dumont FJ, Su Q 1996 Mechanism of action of the immunosuppressant rapamycin. Life Sci 58:373–395 43. Tokimitsu H, Chijiwa T, Hagiuwara M, Mitzutani A, Terasawa K, Hidaka H 1990 KN62, 1-[N,O-bis (5-isoquinolinesulfonyl)-N-methyl-1 tyrosyl] 4-phenylpiperazine, a specific inhibitor of Ca11/calmodulin-dependant protein kinase II. J Biol Chem 265:4315–4323 44. Sumi M, Kiuchi K, Ishikawa T, Ishii A, Hagiwara M, Nagatsu T, Hidaka H 1991 The newly synthetized selective Ca11/calmodulin dependent protein kinase II inhibitor KN93 reduces dopamine content in PC12h cells. Biochem Biophys Res Commun 181:968–975 45. Renoir JM, Radanyi C, Jung-Testas I, Faber LE, Baulieu EE 1990 The nonactivated progesterone receptor is a nuclear heterooligomer. J Biol Chem 265:14402–14406 46. Ning YM, Sanchez ER 1995 Stabilization in vitro of the untransformed glucocorticoid-receptor complex of S49 lymphocytes by the immunophilin ligand FK506. J Steroid Biochem Mol Biol 52:187–194 47. Moudgil VK, Eessalu TE, Buchou T, Renoir JM, Mester J, Baulieu EE 1985 Transformation of chick oviduct progesterone receptor in vitro: effects of hormone, salt, heat, and adenosine triphosphate. Endocrinology 116:1267– 1274 48. Schwaninger M, Blume R, Oetjen E, Lux G, Knepel W 1993 Inhibition of cAMP-responsive element-mediated gene transcription by cyclosporin A and FK506 after membrane depolarization. J Biol Chem 268:23111– 23115 49. Wera S, Belayew A, Martial JA 1995 Rapamycin, FK506 and cyclosporin A inhibit human prolactin gene expression. FEBS Lett 358:158–160 50. Enslen H, Soderling TR 1994 Roles of calmodulin-dependent protein kinase and phosphatase in calcium-dependent transcription of immediate early genes. J Biol Chem 269:20872–20877 51. Kunz J, Hall MN 1993 Cylosporin A, FK506 and rapamycin: more than just immunosuppression. Trends Pharmacol Sci 18:334–338 52. Liu J 1993 FK506 and cyclosporine: molecular probes for studying intracellular signal transduction. Trends Pharmacol Sci 14:182–188 53. Liu J, Farmer JJD, Lane WS, Friedman J, Weissman I, Schreiber SL 1991 Calcineurin is a common target of cyclophilin-cyclosporin A and FKBP-FK506 complexes. Cell 66:807–815 54. O’Keefe SJ, Tamura J, Kincaid RL, Tocci MJ, O’Neill EA 1992 FK506- and CsA-sensitive activation of the interleukin-2 promoter by calcineurin. Nature 357:692–694 55. Chung J, Kuo CJ, Crabtree GR, Blenis J 1997 Rapamycin-FKBP specifically blocks growth-dependent activation of and signalling by the 70 kDa S6 protein kinase. Cell 69:1227–1236 56. Calvo V, Crews CM, Vik TA, Bierer BE 1992 Interleukin-2 stimulation of p70 S6 kinase activity is inhibited by the immunosuppressant rapamycin. Proc Natl Acad Sci USA 89:7571–7575 57. Morice WG, Brunn GJ, Wiederrecht G, Siekerka JJ, Abraham RT 1993 Rapamycin-induced inhibition of p34cdc2 kinase activation is associated with G2/S-phase growth arrest in T lymphocytes. J Biol Chem 268:3734–3738 58. Brown EJ, Albers MW, Shiu TB, Ishikawa K, Keith CT, Lane WS, Schreiber SL 1994 A mammalian protein targeted by G1-arresting rapamycin-receptor complex. Nature 369:756–758 59. Sabatini DM, Pierchala BA, Barrow RK, Schell MJ, Sny-


60.

61.

62.

63.

64.

65.

66.

67.

68.

der SH 1995 The rapamycin and FKBP12 target (RAFT) displays phosphatidylinositol 4-kinase activity. J Biol Chem 270:20875–20878 Sabers CJ, Martin MM, Brunn GJ, Williams JM, Dumont FJ, Wiederrecht G, Abraham RT 1995 Isolation of a protein target of the FKBP12-rapamycin complex in mammalian cells. J Biol Chem 270:815–819 Kunz I, Henriquez R, Schneider U, Deuter-Reinhard M, Movva NR, Hall MN 1995 Target of rapamycin in yeast, TOR2 is an essential phosphatidylinositol kinase homolog required for G1 progression. Cell 82:121–130 Zhen XF, Fiorentino D, Chen J, Crabree GR, Schreiber SL 1995 TOR kinase domains are required for two distinct functions, only one of which is inhibited by rapamycin. Cell 82:121–130 Ning YM, Sanchez ER 1996 In vivo evidence for the generation of a glucocorticoid-receptor-heat shock protein-90 complex incapable of binding hormone by the calmodulin antagonist phenoxybenzamine. Mol Endocrinol 10:14–23 Renoir JM, Marsaud V, Mercier-Bodard C, Le Bihan S, Radanyi C, Baulieu EE 1996 Modulation of steroid-induced transcription by cyclophilin and FKBPs. In Workshop on Cyclophilin Structure and Function, Oct 10–12 1996, Lille, France, Centre National de la Recherche Scientifique and Universite´ des Sciences et Technologies de Lille, France, pp 117–127 Doucet-Brutin S, Renoir M, Le Gallic L, Vincent S, Marty L, Fort P 1995 Growth-regulated expression of FKBP-59 immunophilin in normal and transformed fibroblastic cells. Exp Cell Res 220:152–160 Milad M, Sullivan W, Diehl E, Altman M, Nordeen S, Edwards DP, Toft DO 1995 Interaction of the progesterone receptor with binding proteins for FK506 and cyclosporin A. Mol Endocrinol 9:838–847 El-Ashry D, Onate S, Nordeen SK, Edwards DP 1987 Human progesterone receptor complexed with the antagonist RU486 binds to hormone response element in a structurally altered form. Mol Endocrinol 3:1545–1548 Miyata Y, Chambraud B, Radanyi C, Lebeau M-C, Renoir J-M, Leclerc J, Shirai R, Catelli M-G, Yahara I, Baulieu EE (1997) Phosphorylation of immunosuppressant FK 506 binding protein, FKBP52 by casein kinase II (CKII): reg-

69.

70.

71.

72. 73.

74.

75. 76. 77.

78.

1001

ulation of hsp90-binding activity of FKBP-52. Proc Natl Acad Sci USA 94:14500–14505 Aronica SM, Katzenellenbogen BS 1991 Progesterone receptor regulation in uterine cells: stimulation by estrogen, cyclic adenosine 39,59-monophosphate and insulinlike growth factor 1 and suppression by antiestrogen and protein kinase inhibitors. Endocrinology 128:2045–2052 Beck CA, Weigel NL, Moyer ML, Nordeen SK, Edwards DP 1993 The progesterone antagonist RU486 acquire agonist activity upon stimulation of cAMP signalling pathways. Proc Natl Acad Sci USA 90:4441–4445 Moyer ML, Borror KC, Bona ZJ, DeFranco DB, Nordeen SK 1993 Modulation of cell signalling pathways can enhance or impair glucocorticoid-induced gene expression without altering the phosphorylation state of the receptor. J Biol Chem 268:22933–22940 Zhang Y, Bai W, Allgood VE Weigel NL 1994 Multiple signalling pathways activate the chicken progesterone receptor. Mol Endocrinol 8:577–584 Lorenzo F, Jolivet A, Loosfelt H, Vu Hai MT, Brailly S, Perrot-Applanat M, Milgrom E 1988 A rapid method of epitope mapping. Application to the study of immunogenic domains and to the characterization of various forms of rabbit progesterone receptors. Eur J Biochem 176:53–60 Schneider C, Newman RA, Sutherland DR, Asser U, Greaves MF 1982 A one step purification of membrane proteins using a high efficiency immunomatrix. J Biol Chem 257:10766–10769 Laemmli UK 1970 Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680–685 Nakao K, Myers JE, Faber LE 1985 Development of a monoclonal antibody to the rabbit 8.5S uterine progestine receptor. Can J Biochem Cell Biol 63:33–40 Traish A, Wotiz H 1990 Monocloncal and polyclonal antibodies to human progesterone receptor peptide (533– 547) recognize a specific site in unactivated (8S) and activated (4S) progesterone receptor and distinguish between intact and proteolysed receptors. Endocrinology 127:1167–1175 Nover L, Sharf KD, Nover L (eds) 1991 Heat Shock Response. CRC Press, Boca Raton, FL, pp 41–128