This Article Abstract Full Text (PDF) All Versions of this Article: 21/6/1335    most recent Author Manuscript (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager NURSA Molecule Pages Link Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Carnevale, R. P. Articles by Elizalde, P. V. Search for Related Content PubMed PubMed Citation Articles by Carnevale, R. P. Articles by Elizalde, P. V.

Progestin Effects on Breast Cancer Cell Proliferation, Proteases Activation, and in Vivo Development of Metastatic Phenotype All Depend on Progesterone Receptor Capacity to Activate Cytoplasmic Signaling Pathways

Instituto de Biología y Medicina Experimental (IBYME) (R.P.C., C.J.P., M.S., E.H.C., R.S., P.V.E.), Consejo Nacional de Investigaciones Científicas y Técnicas, Buenos Aires C1428ADN, Argentina; Research Area (A.U., G.P., E.B.d.K.J.), Institute of Oncology Angel H. Roffo, University of Buenos Aires, C1417DTB Buenos Aires, Argentina; and Department of Molecular and Cellular Biology and Pathology (D.P.E., V.B.), Baylor College of Medicine, Houston, Texas 77030

Address all correspondence and requests for reprints to: Patricia V. Elizalde, Ph.D., Laboratory of Molecular Mechanisms of Carcinogenesis, Instituto de Biología y Medicina Experimental (IBYME), Vuelta de Obligado 2490, Buenos Aires 1428, Argentina. E-mail: elizalde{at}dna.uba.ar.

	ABSTRACT

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 
Accumulating evidence indicates that progestins are involved in controlling mammary gland tumorigenesis. Here, we assessed the molecular mechanisms of progestin action in breast cancer models with different phenotypes. We examined C4HD cells, an estrogen (ER) and progesterone (PR) receptor-positive murine breast cancer model in which progestins exert sustained proliferative response, the LM3 murine metastatic mammary tumor cell line, which lacks PR and ER expression, and human PR null T47D-Y breast cancer cells. In addition to acting as a transcription factor, PR can also function as an activator of signaling pathways. To explore which of these two functions were involved in progestin responses, reconstitution experiments in the PR-negative models were performed with wild-type PR-B, with a DNA binding mutant C587A-PR, and with mutant PR-BmPro, which lacks the ability to activate cytoplasm signaling pathways. We found that in a cell context either ER-positive or -negative, progestins induced cell growth and modulation of matrix metalloproteinases-9 (MMP-9) and -2 (MMP-2), and urokinase-type plasminogen activator (uPA) activities, via MAPK and phosphatidylinositol 3-kinase/Akt pathways, in cells expressing wild-type PR-B or DNA binding mutant C587A-PR. In contrast, in cells expressing mutant PR-BmPro, progestins did not induce growth. We also found that unliganded PR expression conferred breast cancer cells an in vitro less proliferative phenotype, as compared with cells lacking PR expression. Modulation of this behavior occurred when PR was functioning either as transcription factor or as signaling activator. Finally, we for the first time demonstrated that progestins favor development of breast tumor metastasis via PR function as activator of signaling pathways. Our present findings provide mechanistic support to the design of a novel therapeutic intervention in PR-positive breast tumors involving blockage of PR capacity to activate cytoplasmic signaling.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 
PROGESTERONE PLAYS A key role in the regulation of cell proliferation and differentiation in the mammary gland, as demonstrated, among an increasing body of evidence, by incomplete mammary gland ductal branching and failure of lobulo-alveolar development in progesterone receptor (PR) knockout mice (1, 2). Accumulated findings indicate that progestins are also involved in controlling mammary gland tumorigenesis both in women and in animal models (3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14). Progestin effects in vitro growth of breast cancer cells were found to be highly dependent on experimental culture conditions and on the presence of estrogens. Particularly, progestins were found to either support sustained growth (11) or to induce cells to progress through one round of cell division, followed by growth arrest at the G1/S phase of the second cycle (6, 15) in the human breast cancer cell line T47D, the ideal model of breast cancer cell with estrogen receptor (ER)-independent, constitutive expression of PR-A and PR-B. Seminal work by Horwitz and co-workers (6, 8, 15, 16) provided evidence that the initial pulse of progestins primes progesterone-arrested cells for the action of secondary proliferative or differentiative signaling pathways. Hence, the commitment to one path or the other would be determined by cross talk between growth factor/cytokine pathways and by progesterone/PR signaling (6, 15, 16). Several lines of evidence, including our own work, (17, 18, 19) have shown that convergence between progestins and type I receptor tyrosine kinases mediates proliferative effects of progestins in breast cancer cells (8, 15). Recently, strong evidence of progestin implication in breast cancer etiology was provided by the observation that postmenopausal women who undergo a combined estrogen and progestin hormone replacement therapy suffer a higher incidence of breast cancer than women who take estrogen alone (20, 21).

The classical pathway of progesterone inducible PR-mediated gene transcription has long been described. Upon progesterone binding, PR undergoes a conformational change, dissociates from a multiprotein chaperone complex, dimerizes and binds to specific progesterone response elements (PREs) in the promoter of target genes (22). DNA-bound receptor recruits coactivators that facilitate transcription initiation through interaction with components of the basal transcription machinery (23). In addition to their direct transcriptional effects, rapid or nongenomic biological effects of progestins have been described in breast cancer cells, evidencing that progestin treatment of human breast cancer T47D cells, activates the signal-transducing c-Src/p21^ras/MAPK pathway which results in cell proliferation (13, 14, 24). Particularly interesting is the interaction between PR and c-Src kinase because it has been demonstrated that activation of c-Src in breast cancer cells is a critical event in tumor progression (3, 14, 25, 26, 27). A series of previous studies demonstrated that transfection with dominant-negative c-Src significantly reduced cell spreading, migration and proliferation of MCF7 and MDA-MB-468 cells (28, 29). Furthermore, c-Src kinase activity is associated with tumor cell colonization in bone and lung (30) and with cell invasion and metastasis of breast cancer cells (31, 32). Hence, activation of c-Src/MAPK signaling pathway mediated by progesterone receptor may contribute to multiple aspects of breast cancer progression.

Although progestins were found to enhance growth and to act as survival factors of breast cancer cells (11), the molecular mechanisms through which progestins control breast cancer growth and metastasis still remain elusive. On the other hand, the presence of PR in breast tumors, regardless of ER expression, is associated with response to hormonal therapy and good prognosis (33, 34), underscoring the complex role of PR in mammary cancer.

In the present study, we took advantage of three breast cancer models with different phenotypes to study molecular mechanisms of progestin action in mammary tumor cell growth, proteases activity, and development of metastasis. The first model examined were primary cultures of C4HD cells, our well-described ER-, and PR-positive murine breast cancer model in which progestins exert sustained proliferative response both in vitro and in vivo (17, 18, 19, 35, 36). The second was our well-characterized model of metastasis for mammary cancer, the LM3 murine cell line (37). The LM3 mammary tumor cell line, which lacks PR and ER expression, derives from the M3 mouse mammary adenocarcinoma (37). Like its parental tumor LM3 cells exhibits a highly invasive behavior and 90–100% incidence of lung metastasis (37). Finally, we studied human PR null T47D-Y breast cancer cells (8, 38, 39). Reconstitution experiments in our PR-negative models were performed with wild-type PR-B (10, 40), DNA binding mutant C587A-PR (3), and mutant PR-BmPro devoid of the ability to activate c-Src-dependent signaling (3). Our results demonstrated that, either in an ER-positive or -negative cell context, progestins induced proliferation and regulated proteases activity and metastasis through PR ability to activate c-Src-dependent signaling pathways. In addition, in the course of the present study we were able to show that unliganded PR expression conferred breast cancer cells an in vitro less proliferative phenotype, as compared with cells lacking PR expression. Modulation of this behavior occurred through PR functioning as either transcription factor or signaling activator. Our present findings could lead to a novel therapeutic approach in PR-positive breast tumors through blockage of PR capacity to activate c-Src-dependent signaling pathways.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 
Progestins Induce Not Only MAPK But Also Phosphatidylinositol 3-Kinase (PI-3K) Activation in Breast Cancer Cells by a Rapid, Nongenomic Mechanism, Mediated by the Classical PR
Progestins have already been found to activate the signal-transducing c-Src/p21^ras/MAPK pathway in the proliferative phase of the biphasic T47D human breast cancer cell response to progestins (14, 41). We here took advantage of a well-described model of hormonal carcinogenesis, in which the response of breast cancer cells to progestins is sustainedly proliferative, to investigate progestin effect on MAPK activity. We worked with primary cultures of the C4HD murine mammary tumor line from an experimental model in which the synthetic progestin medroxyprogesterone acetate (MPA) induced mammary adenocarcinomas in female Balb/c mice (17, 18, 19, 35, 36). As we have previously described in detail (17, 18, 19, 35, 36), C4HD epithelial cells derive from a ductal, progestin-dependent mammary tumor line. C4HD cells require MPA administration to proliferate both in vitro and in vivo and express high levels of PR-B, PR-A, and ER (17, 18, 19, 35, 36). As shown in Fig. 1A, MPA was able to induce rapid phosphorylation of p42/p44 MAPK (2-fold) in C4HD cells, detected by using antisera specific for the active dually phosphorylated form of this kinase. MPA-induced activation was observed as early as 5 min after treatment and returned to basal levels after 30 min. Pretreatment of C4HD cells with 10 µM PD98059, a specific inhibitor of MEK1, or with 5 µM U0126, a MEK1/MEK2 inhibitor, suppressed activation of MAPKs by MPA (Fig. 1A). At concentrations of either 10 nM (not shown) or 100 nM, progestin antagonist RU486 did not induce MAPK activation but significantly inhibited MPA-stimulation of MAPK activity (Fig. 1A). Progestin regulation of PI-3K/Akt pathway activity in breast cancer still remains poorly explored (42). Therefore, we here investigated PI-3K/Akt activation status by performing Western blots with an antiphosphoserine 473 Akt antibody, as marker of Akt activation. As we had already found in our in vivo studies (36), MPA was able to induce strong Akt phosphorylation after 5 min treatment, with the highest phosphorylation level (3-fold) observed at 10 min (Fig. 1B). Wortmannin (1 µM) and LY294002 (2 µM), specific chemical inhibitors of PI-3K, completely blocked MPA-induced Akt phosphorylation (Fig. 1B). Treatment of C4HD cells with RU486 did not result in stimulation of Akt activity, but significantly inhibited MPA capacity to induce Akt activation (Fig. 1B). Blockage of MAPK activation with PD98059 or U0126 had no effect on MPA capacity to stimulate Akt phosphorylation and similarly inhibition of Akt activity with wortmannin or LY294002 caused no effect on MPA-induced MAPK phosphorylation (data not shown).

View larger version (19K): [in this window] [in a new window]   Fig. 1. MPA Induces p42/p44MAPK and PI-3K/Akt Activation in C4HD Cells Expressing Endogenous PR-B and PR-A A, Primary cultures of C4HD cells were treated with 10 nM MPA for the indicated times and with 100 nM RU486 for 5 min, or pretreated with 100 nM RU486, 10 µM PD98059, or 5 µM U0126 before MPA stimulation. Twenty-five micrograms of protein from cell lysates were electrophoresed and Western blot assay was performed with antiphospho-p42/p44MAPK antibody. Membranes were then stripped and hybridized with anti-p42/p44MAPK antibody. B, C4HD cells were treated with 10 nM MPA and 100 nM RU486 or pretreated with 100 nM RU486, 1 µM Wortmannin, or 2 µM LY294002 before adding MPA. Fifty micrograms of protein from cell lysates were electrophoresed and Western blot was performed with antiphosphoserine 473 Akt antibody. Membranes were then stripped and hybridized with anti-Akt antibody. Phospho MAPK and Akt bands underwent densitometry and values were normalized to total MAPK or Akt protein bands. Data analysis showed that increase in MAPK (5–30 min) and Akt phosphorylation (5–30 min) in cells treated with MPA, as compared with untreated cells, and inhibition of MPA-induced phosphorylation levels caused by the respective pharmacological inhibitor and by RU486 were significant (P < 0.001). Experiments shown in A and B were repeated five times with similar results.

View larger version (19K): [in this window] [in a new window]   Fig. 2. MPA Induces p42/p44MAPK and PI-3K/Akt Activation in LM3 and T47D-Y Cells Transfected with PR-B A, LM3 cells were transiently transfected with the hPR-B expression vector (LM3-PR-B), with the empty pSG5 plasmid (LM3-pSG5), or remained untreated. Cells were then stimulated for 2 min with 10 nM MPA. Twenty-five micrograms of protein from cell lysates were electrophoresed and Western blot assay was performed with antiphospho c-Src antibody. Membranes were then stripped and hybridized with anti-Src antibody. Phospho c-Src bands underwent densitometry and values were normalized to total c-Src protein bands, showing significant (P < 0.001) induction of c-Src phosphorylation by MPA in cells expressing PR-B. B and C, Wild-type LM3 cells, LM3-pSG5 cells, and LM3-PR-B cells (B), and wild-type T47D-Y, T47D-Y cells transfected with PR-B (T47D-Y-PR-B), and with the pSG5 vector (T47D-Y-pSG5) (C) were stimulated for 5 min with 10 nM MPA and 100 nM RU486, or pretreated with 5 µM U0126 or RU486 before MPA stimulation. Twenty-five (LM3 cell types) or 50 (T47D-Y cell types) µg of protein from lysates were electrophoresed and Western blot assays were performed with antiphospho-p42/p44MAPK antibody. Membranes were then stripped and hybridized with anti-p42/p44MAPK antibody. D and E, LM3 (D) and T47D-Y (E) cells transfected as described in B and C were treated for 5 min with 10 nM MPA and 100 nM RU486 or pretreated with 2 µM LY294002 or 100 nM RU486 before MPA stimulation. Western blot assays were performed using an antiphosphoserine 473 Akt antibody. Membranes were stripped and hybridized with anti-Akt antibody. Phospho MAPK and Akt bands underwent densitometry and were analyzed as described in Fig. 1, showing significant (P < 0.001) induction of MAPK and Akt phosphorylation by MPA in cells expressing PR-B and significant (P < 0.001) inhibition of MPA-induced phosphorylation levels caused by the respective pharmacological inhibitor and by RU486. Experiments shown were repeated five times for each cell type with similar results.

To further evidence progestin ability to activate cytoplasmic kinases by a nongenomic mechanism, we transfected LM3 and T47D-Y cells with a PR-B engineered to contain a point mutation in a conserved cysteine in the first zinc finger of the DNA binding domain (C587A) that was engineered and characterized by Horwitz and co-workers (43) and lacks capacity to bind to PRE. As reported in Cos-7 cells (3), MPA treatment of LM3 cells transfected with C587A-PR (LM3-C587A-PR) resulted in rapid induction of c-Src phosphorylation (Fig. 3A). In addition, MPA exerted a reproducible increase in p42/p44 MAPK phosphorylation in LM3-C587A-PR cells, which was abolished by U0126 and inhibited by RU486 (Fig. 3B). We also found significant induction of p42/p44 MAPK phosphorylation by MPA treatment of T47D-Y cells transfected with the C587A-PR (T47D-Y-C587A-PR), which was completely blocked by U0126 and significantly inhibited by RU486 (Fig. 3C). MPA treatment also induced reproducible Akt phosphorylation in LM3-C587A-PR (Fig. 3D) and T47D-Y-C587A-PR (Fig. 3E) cells, that was abrogated by LY294002 and inhibited by RU486 (Fig. 3, D and E).

View larger version (16K): [in this window] [in a new window]   Fig. 3. MPA Induces p42/p44MAPK and PI-3K/Akt Activation via a Transcriptionally Impaired PR-B A, LM3 cells were transfected with a PR-B expression vector engineered to contain a point mutation in the second zinc finger of the DNA binding domain (LM3-C587A-PR) or remained untreated. Cells were then stimulated for 2 min with 10 nM MPA. Twenty-five micrograms of protein from cell lysates were electrophoresed and Western blot assay was performed with antiphospho-c-Src antibody. Membranes were stripped and hybridized with anti-c-Src antibody. The experiment shown was repeated five times with similar results. B and C, LM3 (B) and T47D-Y (C) cells were transfected with the C587A-PR mutant (LM3-C587A-PR and T47D-Y-C587A-PR), or remained untreated. Cells were then stimulated for 5 min with 10 nM MPA and 100 nM RU486 or pretreated with 5 µM U0126 or 100 nM RU486 before MPA stimulation. Twenty-five (LM3 cell types) or 50 (T47D-Y cell types) µg of protein from lysates were electrophoresed and Western blot assays were performed with antiphospho p42/p44MAPK antibody. Membranes were then stripped and hybridized with anti-p42/p44MAPK antibody. D and E, LM3 (D) and T47D-Y (E) cells transfected as described in B and C were treated for 5 min with 10 nM MPA and RU486 or pretreated with 2 µM LY294002 or 100 nM RU486 before MPA stimulation. Western blot assays were performed using an antiphosphoserine 473 Akt antibody. Membranes were stripped and hybridized with anti-Akt antibody. Phospho-MAPK and Akt bands underwent densitometry and were analyzed as described in Fig. 1 showing that MPA-induced increase in MAPK and Akt phosphorylation in cells transfected with the C587A-PR mutant and inhibition of MPA-induced phosphorylation levels caused by the respective pharmacological inhibitor and by RU486 were significant (P < 0.001). Experiments shown in B–E were repeated five times for each cell type with similar results.

View larger version (17K): [in this window] [in a new window]   Fig. 4. PR-B and C587A-PR Expression and Transcriptional Activity in LM3 and T47D-Y Transfected Cells LM3 (A) and T47D-Y (B) cells were transiently transfected with 2 and 4 µg of PR-B or C587A-PR expression vectors. LM3 cells were treated with 10 nM MPA for 48 h or remained untreated. One hundred micrograms of protein from cell lysates were electrophoresed and Western blot assays were performed with anti-PR antibody. T47D cell lysates were used as positive control. Western blot assay using antiactin antibody was carried out using identical protein lysates as control of specificity of PR expression and MPA treatment. C, LM3 cells were transiently transfected with 2 µg of a PRE-2X-TATA-luciferase reporter plasmid and with 1 µg of cytomegalovirus-ßgal expression vector as internal control. Cells were cotransfected with 40 ng of PR-B, C587A-PR, or pSG5 empty vector. After transfection, cells were treated with MPA and MPA+RU486 for 48 h, or were left untreated. Cells were then harvested and lysed. Luciferase and ß-galactosidase activities were measured as described in Materials and Methods. Results are presented as fold induction of luciferase activity with respect to cells growing in absence of MPA stimulation. The data shown represent the mean of six independent experiments ± SEM. Significance for b vs. a and c vs. b, P < 0.001. D, C587A-PR transcriptional activity on endogenous genes. Left panel, LM3-PR-B and LM3-C587A-PR cells were treated with MPA for 48 h or remained untreated. Fifty micrograms of protein from lysates were electrophoresed and Western blots were performed with anti-Bcl-xL antibody. Membrane was then stripped and hybridized with antiactin antibody. Bcl-xL bands underwent densitometry and values normalized to actin protein bands. Results are presented as fold induction of Bcl-xL expression with respect to cells growing in absence of MPA stimulation. The data shown represent the mean of four independent experiments ± SEM. Significance for b vs. a, P < 0.001. Right panel, LM3-PR-B and LM3-C587A-PR cells were treated with MPA for the indicated times or remained untreated. Fifty micrograms of protein from lysates were electrophoresed and Western blots were performed with anti-p21antibody. Membrane was then stripped and hybridized with antiactin antibody. p21 bands underwent densitometry and values normalized to actin protein bands. Results are presented as fold induction of p21 expression with respect to cells growing in absence of MPA stimulation. The data shown represent the mean of four independent experiments ± SEM.

View larger version (40K): [in this window] [in a new window]   Fig. 5. Blockage of p42/p44 MAPK and PI-3K/Akt Activities Inhibits MPA-Induced Proliferation of Mammary Tumor Cells A, C4HD cells were treated for 48 h with 10 nM MPA, 10 nM RU486, 10 µM PD98059, 5 µM U0126, 1 µM Wortmannin, and 2 µM LY294002, or pretreated with these inhibitors before MPA stimulation. Significance for b vs. a, c vs. b, and d vs. b, P < 0.001. B, LM3 cells were transfected with 40 ng of pSG5 empty vector and with 40 ng of PR-B, and in C with pSG5 vector, 40 ng of C587A-PR, and 40 ng of PR-BmPro. After transfection, cells were treated for 24 h with 10 nM MPA, 100 nM RU486, 5 µM U0126, 2 µM LY294002, or pretreated with these inhibitors before MPA stimulation. Significance for b and e vs. a, c vs. b, d vs. c, P < 0.001. A–C, Incorporation of [3H] thymidine was used as a measure of DNA synthesis and data are presented as mean ± SD. Experiments shown in A–C are representative of a total of five. D, Left panel, LM3 cells were transiently transfected with 2 µg of PR-B or PR-BmPro expression vectors and were then treated with 10 nM MPA for 48 h, or remained untreated. One hundred micrograms of protein from cell lysates were electrophoresed and Western blot assays were performed with anti-PR antibody. T47D cell lysates were used as positive control. Western blot assay using antiactin antibody was carried out using identical protein lysates as control of specificity of PR expression and MPA treatment. The experiment shown was repeated four times with similar results. Right panel, LM3-PR-B and LM3-PR-BmPro were treated with MPA for the indicated times or remained untreated. Fifty micrograms of protein from cell lysates were electrophoresed and Western blot assays were performed with anti-p21 antibody. Membrane was then stripped and hybridized with antiactin antibody. p21 bands underwent densitometry and values normalized to actin protein bands. Results are presented as fold induction of p21 expression with respect to cells growing in absence of MPA stimulation. The data shown represent the mean of five independent experiments ± the SEM. Significance for b vs. a, P < 0.001. E, LM3-PR-B and LM3-PR-BmPro were stimulated for 5 min with 10 nM MPA or pretreated with 5 µM U0126 or 2 µM Ly294002 before MPA stimulation. Twenty-five micrograms of protein from lysates were electrophoresed and Western blot assays were performed with antiphospho-p42/p44MAPK (upper panel) or antiphosphoserine 473 Akt antibody (lower panel). Membranes were then stripped and hybridized with anti-p42/p44MAPK or anti Akt antibody, respectively. Experiments shown were repeated four times with similar results. F, LM3, LM3-pSG5, LM3-PR-B, LM3-C587A-PR, and LM3-PR-BmPro cells were cultured at very low density (500 cells/well) and treated for 5 d with 10 nM MPA, or left untreated. Cells were then fixed, dyed, and number of colonies containing more than 20 cells was counted. Data are presented as mean ± SD. Significance for b vs. a and c vs. b, P < 0.001. G and H, T47D-Y cells were transfected with 40 ng of pSG5 empty vector and with 40 ng of PR-B (G), and with pSG5 vector, with C587A-PR and PR-BmPro (H). After transfection, cells were treated and proliferation was measured. Data are presented as described in A–C. Significance for b and e vs. a, c vs. b, and d vs. c, P < 0.001. Experiments shown in F–H are representative of a total of five for each cell type.

View larger version (48K): [in this window] [in a new window]   Fig. 6. MPA Inhibits MMP-2, MMP-9, and uPA Activities and Expression in C4HD Cells A, Left panel, Primary cultures of C4HD cells were incubated in serum-free medium for 24 h with 10 nM MPA, 10 nM RU486, or pretreated with RU486 before MPA stimulation. Conditioned media were collected and cell monolayers were lysed to measure protein content. MMPs (upper panel) and uPA (lower panel) activities were evaluated by gelatin or casein and plasminogen zymography, respectively. Zymograms incubated with EDTA for MMPs or without plasminogen for uPA indicated the specificity of the reaction. Shown are representative experiments of a total of six for MMPs and uPA. Right panel, MMPs and uPA bands from all six C4HD cell cultures growing in presence and absence of MPA or MPA plus RU486 underwent densitometry and values were normalized to protein content of the cell monolayer. Activities are expressed as percentages of those of control cells growing in absence of MPA stimulation. Significance for b vs. a and c vs. b, P < 0.001. B, Upper panel, Primary cultures of C4HD cells were incubated in serum-free medium for 24 h with or without 10 nM MPA stimulation and conditioned media were collected and concentrated. Samples were electrophoresed and Western blot assays were performed with MMP-2, MMP-9, and uPA antibodies. The experiment shown is representative of a total of three with similar results. Lower panel, MMPs and uPA bands from all six C4HD cell cultures growing in presence and absence of MPA underwent densitometry and values were normalized to protein content of the cell monolayer. Protein levels present in conditioned media are expressed as percentages of those of control cells growing in absence of MPA stimulation. Significance for b vs. a, P < 0.001. C, Primary cultures of C4HD cells were incubated in serum-free medium for 24 h with or without 10 nM MPA stimulation. Eighty micrograms of protein from cell lysates were electrophoresed and Western blot assay was performed with anti-MMP-9, MMP-2, and uPA antibodies. Membranes were then stripped and hybridized with antiactin antibody. The experiment shown is representative of a total of three with similar results. D, Upper panels, primary cultures of C4HD cells were incubated in a serum-free medium for 24 h with 10 nM MPA, 5 µM U0126, 2 µM LY294002, or pretreated with these inhibitors before MPA stimulation. Conditioned media were collected and cell monolayers were lysed to measure protein content. MMPs and uPA activities were evaluated as described above. Shown is a representative experiment of a total of five. Lower panels, MMPs and uPA bands from all five C4HD cell cultures growing as indicated underwent densitometry and values were normalized to protein content of the cell monolayer. Activities are expressed as percentages of those of control cells growing in absence of MPA stimulation. Significance for b vs. a and c vs. b, P < 0.001.

View larger version (23K): [in this window] [in a new window]   Fig. 7. MPA Inhibits uPA Activity by a Nongenomic Mechanism LM3, LM3-PR-B, LM3-C587A-PR, and LM3-pSG5 cells were incubated in serum-free medium for 24 h with 10 nM MPA or pretreated with 100 nM RU486 (A, upper panel), 5 µM U0126 (B, upper panel), or 2 µM LY294002 (C, upper panel) before MPA stimulation. uPA activity in conditioned media was studied as described in Fig. 6. Shown is a representative experiment of a total of four performed to study RU486, U0126, or LY294002 effects. A–C, Lower panels, Data analysis was performed as described in Fig. 6. uPA activities are expressed as percentages of those of control cells growing in absence of MPA stimulation. Significance for b vs. a and c vs. b, P < 0.001.

View larger version (32K): [in this window] [in a new window]   Fig. 8. Time Course of PR-B, C587A-PR, and PR-BmPro Expression in LM3 Cells LM3 cells were transiently transfected with wild-type PR-B (A), C587A-PR-B mutant (B) or PR-BmPro mutant (C) and were harvested at the indicated times. One hundred micrograms of protein from cell lysates were electrophoresed and Western blot assays were performed with anti-PR antibody. Wild-type LM3 cell lysates were used as negative control and T47D cell lysates were used as positive control. Membranes were then stripped and hybridized with antiactin antibody. The experiment shown is representative of a total of four with similar results.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

Progestins Induce Nongenomic Activation of Both p42/p44 MAPKs and PI-3K/Akt Signaling Pathways in Breast Cancer Acting through the Classical PR, Independently of ER Expression

Progestin regulation of PI-3K/Akt pathway activity in breast cancer remains poorly explored (42). To our knowledge, we were the first to demonstrate that in breast tumors MPA activates the PI-3K/Akt signaling pathway, which participates in progestin-induced in vivo growth (36). Our present in vitro findings demonstrated that MPA induced Akt phosphorylation in cells expressing both wild-type PR-B and C587A-PR. These results are in agreement with recent findings showing progesterone and MPA capacity to induce rapid phosphorylation of Ser473 Akt in B-474, T47Dco cells and T47D-Y cells stably transfected with PR-B (42). Little is known about RU486 capacity to modulate signaling pathways in breast cancer cells. We here found that when administered alone, RU486 was unable to induce either p42/p44 MAPK or PI-3K/Akt activation in any of the cell types studied. As previously reported in several breast cancer cell lines (14, 41), RU486 antagonized MPA-induced MAPK activation in all three model systems used in the present work. Notably, Lange and co-workers (41) recently reported that RU486 activated p42/p44 MAPK in T47D cells stably expressing either PR-B or a transcriptionally impaired PR-B harboring a point mutation at Ser294, a ligand-dependent and MAPK consensus phosphorylation site. RU486 was also found to stimulate p42/p44 MAPK activation in PR-transfected MDA-MB-231 breast cancer cells (56). Differences between endogenous levels of PR expression in C4HD cells or transient PR expression in LM3 or T47D-Y cells used in our study, and stable expression of PR from a constitutive promoter in T47D-YB (41) or in MDA-MB-231A cells (56) could account for the discrepancies observed in RU486 modulation of p42/p44 MAPK activation. We also found that RU486 antagonized MPA effect on PI-3K/Akt activity in C4HD and LM3 or T47D-Y cells transfected with either wild-type PR-B or C587A-PR. To the best of our knowledge, ours is the first description of the role of antiprogestins in progestin-induced PI-3K/Akt activity in breast cancer. Physical association between Xenopus PR and active PI-3K was found in progesterone-stimulated Xenopus laevis oocyte maturation (57). However, no association was found in oocytes induced to mature with RU486 (57), in accordance with our present findings, showing that RU486 did not induce PI-3K/Akt activity in breast cancer. In the above early report, RU486 did not modulate progesterone stimulated PI-3K activity in XPR immunoprecipitates (57).

Progestin Regulation of Breast Cancer Growth and Metastatic Phenotype Depends on PR Function as Either Transcription Factor or as Activator of Signaling Pathways
Our present findings have provided new insight into the complex role of PR in breast cancer biology. Firstly, our results in C4HD cells clearly evidenced that progestins are able to induce sustained breast cancer cell proliferation acting through the classical PR via nongenomic mechanisms involving activation not only of MAPK, as has been well documented (3, 14, 41), but also through the PI-3K/Akt pathway. Furthermore, our reconstitution studies using wild-type PR-B and DNA binding mutant C587A-PR, clearly demonstrated that once PR with full capacity to activate signaling pathways is expressed in either an ER-negative (LM3 cells) or -positive (T47D-Y cells) cell context, progestins are strongly proliferative. Notably, when LM3 cells expressed mutant PR-BmPro, which lacks the ability to activate cytoplasmic signaling pathways, the proliferative response to MPA was lost, in accordance with recent findings in MCF-7 cells stably expressing PR-BmPro (41). Startling findings have also shown that PR-BmPro caused delayed progestin-mediated initiation of Xenopus oocyte germinal vesicle breakdown and delayed progestin-induced growth arrest in normal breast epithelial MCF-12 cells, as compared with effects exerted by wild-type PR-B (3). On one hand, our findings are in accordance with accumulated evidence demonstrating progestins’ positive role in breast cancer development (11, 12, 14, 41), mainly through PR capacity to activate cytoplasmic kinases (14, 41). On the other, they clearly show that cell response to progestins will rely on whether PR functions as a transcription factor or as an activator of signaling cascades. It is to be noted that DNA binding mutant C587A accumulates in the nucleus, but unlike endogenously expressed PR in C4HD or wild-type PR-B transfected in LM3 cells, C587A does not patch at foci (data not shown), which are considered active transcription sites (58, 59). Furthermore, C587A-PR capacity to induce Bcl-x_L protein expression in response to MPA was strikingly lower as compared with wild-type PR-B. Here, we found that RU486 alone did not regulate growth of C4HD cells or of LM3 and T47D-Y cells transfected with either wild-type PR-B or mutant C587A-PR, in accordance with our findings showing lack of RU486 capacity to activate p42/p44 MAPK and PI-3K/Akt signaling pathways in these cells. Contrastingly, RU486 has been found to induce proliferation in T47D cells (60) and in T47D-Y cells stably expressing either PR-B or PR-B mutant Ser294A-PR (41), in line with RU486 activation of p42/p44 MAPK in these latter cells. Differences in PR levels and availability of cell signaling molecules between cells used in the above mentioned studies and in our present work could account for these contrasting results. It is notable that our current results with RU486 further demonstrate that only a PR ligand capable to stimulate PR function as activator of cytoplasmic signaling will drive proliferation of breast cancer cells.

Secondly, one of the most exciting findings of the present study was that unliganded PR modulated the phenotype of both ER-negative (LM3) and ER-positive (T47D-Y) breast cancer cells. Thus, PR-B was able to inhibit basal proliferation of LM3-PR-B and T47D-Y-PR-B cells. Interestingly, several pieces of evidence, including our own work with PR and heregulin (17), have demonstrated that unliganded steroid receptors could indeed be transcriptionally activated through cross-talk with growth factors signaling pathways (40, 61). Similarly, expression of unliganded DNA binding mutant C587A-PR resulted in inhibition of basal LM3-C587A-PR and T47D-Y-C587A-PR cell growth. We speculate that growth effects of C587A-PR are due to its ability to activate signaling pathways that, as with PR transcriptional activity, might be stimulated in a ligand-independent manner by growth factors. These signaling pathways could in turn regulate expression of genes involved in proliferation control. Expression of PR-BmPro in absence of progestin stimulation also caused cell growth inhibition. Although PR-BmPro is unable to mediate c-Src-dependent signaling activation, it retains transcriptional functions and key phosphorylable residues that might, as in wild-type PR-B, be activated in a ligand-independent manner by cytoplasmic signaling cascades. Our present results are in accordance with pioneering findings of Horwitz and co-workers (38, 39) who recently engineered ER-positive T47D-Y breast cancer cells that express each PR isoform under the control of an inducible promoter. These models allowed them to show that PRs regulate transcript levels in the absence of progesterone and to find extensive ligand-independent regulation of genes involved in regulation of cell-to-cell signaling and adhesion, extracellular matrix, motility and invasion. They also found modulation of cell signaling molecules as well as of genes involved in cell growth and apoptosis.

Thirdly, we found that MPA induced significant reduction of MMP-9, MMP-2, and uPA production in C4HD cells, and of uPA production in LM3-PR-B and LM3-C587A-PR cells. Interestingly, in C4HD cells endogenously expressing both PR-A and PR-B, only MAPK participated in MPA-induced down-regulation of uPA, with no involvement of PI-3K/Akt. Conversely, MPA effect of MMPs activity required functional PI-3K/Akt signaling but no MAPK activation. Contrastingly, in LM3 cells transfected with wild-type PR-B and C587A-PR, both signaling pathways participated in MPA inhibition of uPA activity. To the best of our knowledge, ours is the first description of progestin regulation of MMPs and uPA via nongenomic mechanisms involving activation of MAPK and PI-3K/Akt in breast cancer cells. In addition, these findings indicate that specific signaling pathways, acting in response to progestins to trigger a certain biological response, will likely depend on PR isoforms expressed in a given cell type. Our results agree with recent findings showing differential modulation of cell signaling molecules and genes that participate in motility and invasion by PR-B or PR-A (38, 39). The presence of ER in C4HD cells could also account for the difference between signaling pathways involved in MPA inhibition of uPA activity because we have found that ER expression could modify MPA capacity to activate signaling pathways. Progesterone has been found to inhibit uPA secretion and both MMP-9 and MMP-2 activities in human endometrial and ovarian carcinoma cells (46, 47, 48). However, effects of progestins on invasion and metastasis in breast cancer remain poorly studied. Particularly, it has been demonstrated that increased uPA activity levels are positively associated with poor prognosis and that they correlated inversely with estrogen and progesterone receptor status (62). In a series of recent studies assessing the effect of progesterone on invasive properties and tumor growth in PR-transfected human breast cancer MDA-MB-231 cells, it was found that progesterone strongly down-regulated expression of uPA (63, 64) and MMP-9 (65).

Finally, our present findings for the first time demonstrated that progestins favor development of breast tumor metastasis via PR function as activator of signaling pathways. Our results with PR-BmPro mutant further demonstrate that progestin-activation of c-Src kinase is necessary for MPA induction of metastatic capacity in breast cancer cells.

Therefore, to integrate our present in vitro and in vivo findings, we postulate that unliganded PR expression confers breast cancer cells, both ER-negative and positive, an in vitro less proliferative phenotype, as compared with cells lacking PR expression. Modulation of this behavior occurs by PR function as either both transcription factor and signaling activator (wild-type PR-B), as transcription factor (PR-BmPro) or as activator of p42/p44 MAPK and PI-3K/Akt pathways (C587A-PR). When cells that we have engineered to express PR are exposed to physiological levels of circulating progesterone in the intact animal, different metastatic behaviors will result depending on PR capacity to act as transcription factor or as activator of signaling pathways. On the one hand, cells expressing either wild-type PR-B or C587A-PR, will respond to progesterone with proliferative rates that, according to our in vitro results, will be approximately equal to those of PR-negative cells. In addition, we found that progestins stimulated breast cancer cell survival, as first reported by Moore and co-workers (12). However, progesterone will significantly reduce uPA activity in PR-transfected cells, as compared with their PR-negative counterparts. The balance between these responses will be a tumor cell with lower metastatic capacity than a PR-negative cell. On the other hand, in cells expressing the mutant PR-BmPro, basal in vitro proliferation was lower as compared with wild-type LM3 cells. However, taking our in vitro results into account, no proliferative response to endogenous progesterone will occur when LM3-PR-BmPro cells are injected in the intact animal. As a result, LM3-PR-BmPro cells showed a metastatic capacity lower not only than cells expressing wild type-PR-B but also than cells transfected with C587A-PR. Therefore, our present findings for the first time provide a mechanistic explanation to the clinical data showing that PR expression in breast tumors, independently of ER status, is a prognostic marker of low aggressiveness with favorable disease-free survival (33, 34). Interestingly, our previous finding demonstrating that inhibition of uPA activity reduced the incidence and number of experimental and spontaneous lung metastasis in the LM3 model (66), provides a causal relationship between progestin capacity to inhibit uPA activity and reduced metastatic capacity in LM3 cells expressing PR-B as compared with wild type LM3 cells. However, when PR-B or C587A-PR-positive cells are exposed to exogenous progestin administration, the forcefully proliferative and prosurvival stimulus of progestins, defeats progestin capacity to inhibit protease systems with a final outcome of higher metastasis as compared with cells growing under physiological levels of progesterone. On the contrary, LM3-PR-BmPro will not proliferate in response to exogenously administered MPA with consequent no regulation of metastatic behavior in animals treated with MPA. We find these results particularly interesting not only in the light of the recent epidemiological data from the hormonal replacement therapy study in postmenopausal women sponsored by the Women’s Health Initiative, evidencing that a combined estrogen plus progestin regimen led to higher rates of breast cancer over the estrogen alone (20, 21), but also because advanced endometrial and breast cancers are treated with MPA therapeutically, using doses of progestins comparable to those used in our in vivo experiments (67, 68). It is to note that we have successfully detected different outcomes in the metastatic capacity between cells transiently transfected with PRs and wild-type LM3 cells, because we have employed an in vivo experimental metastasis assay. In our present model, the efficient and rapid steps of arrest and extravasation are completed when the injected cells still express the transfected PRs because we have found that LM3 transiently transfected with wild-type or mutant PRs, express high levels of PR protein 1 wk after transfection. Furthermore, it has long been acknowledged that proteases have a primary effect in the metastatic process facilitating extravasation. Because we have here demonstrated that MPA induced significant reduction of uPA production in LM3-PR-B and LM3-C587A-PR cells, it is likely that when injected into mice with physiological levels of circulating progesterone or into mice treated with MPA, LM3-PR-B and LM3-PR-C587A cells would be less efficient to extravasate to lung than wild-type LM3 cells. Notably, it has recently been shown that proteases not only have a primary effect mediating the extravasation, but also play a crucial role in regulating cell proliferation at the secondary site (reviewed in Ref. 51). Because we found that progestins significantly reduce uPA activity in PR-transfected cells, as compared with their PR-negative counterparts, this could result also in early reduced proliferation of PR-transfected cells. As expected, when we examined lung metastasis in mice injected with PR-expressing LM3 cells, using immunohistochemistry, we could not detect PR expression. It is highly probable that through these 28 d, cells have lost PR expression. This finding, however does not invalidate our result because the metastatic capacity of the different LM3 cell types is defined during the first days of the study in which cells retain PR expression. Moreover, similar results regarding PR reduction of lung metastasis were observed at 7 d, when transfected cells still express high levels of PR. Our results with wild-type PR-B or C587A-PR-transfected cells help to explain epidemiological data from the Women’s Health Initiative study of the hormonal replacement therapy with a combined estrogen and progestin regimen (20, 21). We speculate that either the response to progestins of normal epithelial breast cells may be proliferative, as we demonstrated in breast cancer cells, and will finally result in breast tumor development, or that a population of preexisting and perhaps dormant tumor cells in the mammary gland of treated women are stimulated, giving rise to tumor development. Earlier epidemiologic work had also shown that progesterone is not protective in the breast (69).

Comprehensively, our present findings have unraveled molecular mechanisms of PR ligand-dependent and -independent action in breast cancer. On the one hand, we have for the first time provided a mechanistic support to a recently suggested proposal (41) of therapeutic intervention in PR-positive breast tumors, involving the specific blockage of PR function as activator of signaling to inhibit growth and metastasis. On the other, our findings raise the possibility of considering a gene therapy protocol for PR-negative breast tumors consisting of reexpression in breast cancer cells of a mutant PR that lacks the