PPARγ antagonist GW9662 induces functional estrogen receptor in mouse mammary organ culture: potential translational significance
- First Online:
- Cite this article as:
- Mehta, R.G., Peng, X., Roy, S. et al. Mol Cell Biochem (2013) 372: 249. doi:10.1007/s11010-012-1466-9
- 434 Views
The nuclear receptor peroxisome proliferator-activated receptor gamma (PPARγ) plays a central role in regulating metabolism, including interaction with the estrogen receptor-α (ERα). Significantly, PPARγ activity can be modulated by small molecules to control cancer both in vitro and in vivo (Yin et al., Cancer Res 69:687–694, 2009). Here, we evaluated the effects of the PPARγ agonist GW7845 and the PPARγ antagonist GW9662 on DMBA-induced mammary alveolar lesions (MAL) in a mouse mammary organ culture. The results were as follows: (a) the incidence of MAL development was significantly inhibited by GW 7845 and GW 9662; (b) GW9662 but not GW7845, in the presence of estradiol, induced ER and PR expression in mammary glands and functional ERα in MAL; (c) while GW9662 inhibited expression of adipsin and ap2, GW 7845 enhanced expression of these PPARγ-response genes; and (d) Tamoxifen caused significant inhibition of GW9662 treated MAL, suggesting that GW9662 sensitizes MAL to antiestrogen treatment, presumably through rendering functional ERα and induction of PR. The induction of ERα by GW9662, including newer analogs, may permit use of anti-ER strategies to inhibit breast cancer in ER− patients.
KeywordsOrgan cultureEstrogen receptorPPARγMammary gland
The peroxisome proliferators activated receptors (PPARs) are ligand activated transcription factors, belonging to the nuclear receptor super family, which hetero-dimerize with RXR and control multiple pathways in normal tissues, including cancer . There are three distinct PPAR subtypes; PPARα, β, and γ, with each demonstrating a particular tissue distribution and ligand specificity . PPARγ is principally expressed in adipocytes but is also found in a range of other tissues . Several natural ligands that activate PPARγ have been identified, including 15-deoxy-∆12,14-prostaglandin J2 (15d-PGJ2), linoleic acid, and lysophosphatidic acid . Synthetic ligands shown to activate PPARγ include agents of the thiazolidinedione (TZD) family such as troglitazone, pioglitazone, rosiglitazone, and ciglitazone. Significantly, members of the TZD family as well as the non-thiazolidinedione tyrosine based PPARγ agonist GW7845 have been shown to inhibit breast cancer both in vitro and in vivo . The effects by PPARγ agonists are apparently, largely, context-dependent; for example, specific tissues or cancer cell type [6, 7].
Previous studies have demonstrated close interaction between PPARγ and the estrogen receptor (ERα) during normal development and during carcinogenesis. The estrogen receptor-α (ERα) is an established predictive marker in the management of breast cancer patients [8, 9]. While an ERα that is 10 % or greater by immunostaining would qualify a tumor as ERα+ and be considered for SERM or aromatase inhibitor treatment [10, 11], these treatments are not recommended for patients expressing less than 10 % ERα. Here we show that the PPARγ antagonist GW9662 and the PPARγ agonist GW7845, both inhibit development of MMOC-derived mammary alveolar lesion (MAL) and that GW9662 can induce ERα and PR significantly in the mammary glands and PR selectively in MAL. The induction of ERα in the breast cancer patients may help control cancer in individuals, who are otherwise not eligible for hormonal treatment, for example patients with ER−, BRCA1/2, or triple-negative tumors.
Materials and methods
Mouse mammary gland organ culture (MMOC)
For determining the effects of PPARγ modulators on gene expression, we modified the MMOC protocol. For accurate comparison of the induction of selected gene expression, instead of collecting a pool of glands from each treatment group, we carried out a comparison from paired glands. One gland from each mouse served as a control for the treatment of a contralateral gland from the same mouse. Following the treatment period, the mammary glands were collected individually and stored in Trizol reagent (Invitrogen, Carlsbad, CA) at −80 °C. Total RNA was extracted from each gland as per manufacturer’s instruction. qRT-PCR was carried out as previously described . Mouse ribosomal 18S RNA was used as a house keeping gene for normalization, which is expressed at relatively stable level from organ to organ. Primers used for real-time PCR are: mERα (forward: 5′-TGCAATGACTATGCCTCTGG-3′, reverse: 5′-CTCCGGTTCTTGTCAATGGT-3′), m18S (forward: 5′-CATGGCCGTTCTTAGTTGGT-3′, reverse: 5′-GAACGCCACTTGTCCCTCTA-3′), mPR (forward: 5′-ATGAAGCATCTGGCTGTCACTA-3′, reverse: 5′-AAATAGTTATGCTGCCCTTCCA-3′), mAdipsin (forward: 5′-CAAGCGATGGTATGATGTGC-3′, reverse: 5′-ATTGCAAGGTGAGGGGTCTC-3′), maP2 (forward: 5′-TGGAAGCTTGTCTCCAGTGA-3′), and mPPARγ (5′-GATGGAAGACCACTCGCATT, reverse: 5′-AACCATTGGGTCAGCTCTTG-5′).
Mammary glands cultured in Waymouth’s serum-free media with appropriate PPARγ agonist or antagonist were fixed in buffered formalin, and 4 µm-thick sections were prepared for immunostaining using standard protocol as described previously . Briefly, after quenching peroxidase activity with 3 % hydrogen peroxide and blocking non-specific binding with goat serum, a 1/200 dilution of primary antibody (ERα) was applied (Santa Cruz Biotechnology, Santa Cruz, CA) and incubated at 4 °C overnight. The sections are then washed in PBS, followed by incubation with a horse radish peroxidase labeled anti-rabbit IgG for 30 min using the DAKO EnVision + System (DAKO Corp, Carpentaria, CA). After a final wash in PBS, the polymer bound antibody is detected with liquid DAB substrate chromogen system for 3–5 min. The chromogen stained tissue is counterstained for 30 s in Gill’s modified hematoxylin.
Data were expressed as mean ± SD and analyzed through one-way ANOVA followed by pair-wise comparisons made with the Bonferroni correction method for multiple comparisons using the GraphPad Instat Statistical program (La Jolla, CA). All of the tests were two-sided, and a p value of <0.05 was considered to be significant. The Pearson’s Chi-square test was used to determine significant difference in MAL lesion multiplicity between treatments. The p value of <0.05 was considered to be significant. For qRT-PCR analysis, five pairs of glands obtained after MMOC were evaluated for gene expression resulting in five datasets. Experiments were repeated at least twice, data from individual organs per treatment were analyzed for statistical significance using Student’s t test. The p value <0.05 was considered as statistically significant.
Effects of GW9662 on the development of DMBA-induced MAL
Effects of PPARγ analogs and Tamoxifen on the development of MAL in MMOC
No. of glands
No. of glands with lesions/total number of glands (% incidence)
% Inhibition of MAL incidence
GW9662 (10 μM) (0–10 days)
GW7845 (0.1 μM) (0–10 days)
GW9662 (10 μM) + GW7845 (0.1 μM)
GW9662 10 μM (0–4 days)
GW9662 10 μM
Tamoxifen (1 μM) (0–10 days)
Tamoxifen (1 μM) (4–10 days)
GW9662 (0–4 days) (10 μM) + Tamoxifen (4–10 days) (1 μM)
Effects of GW7845 singly and in combination with GW9662 on the induction of MAL
Here we evaluated the effects of GW7845, a PPARγ agonist, alone and in combination with GW9662 on MAL development. A GW7845 dose response showed >60 % inhibition of MAL incidence at 0.1 μM of GW7845 with no further inhibition at higher concentrations (Figs. 2, 3). A head to head comparison of the effect of GW7845 and GW9662 on MAL incidence showed an 80 % inhibition at 0.1 μM GW7845 (p < 0.01) and a 71 % inhibition at 10 μM of GW9662. A combination of these two agents at these concentrations did not enhance their efficacies beyond those shown individually (Table 1; Fig. 3).
Effects of GW7845 and GW9662 singly and in combination on the ER and PR expression
In order to determine whether induction of ER by GW9662 in the presence of estradiol has functional consequences, we examined the effects of GW9662 and GW7845 on the expression of progesterone receptor (PR). As shown in Fig. 4c, GW9662 increased expression of PR in the presence of estradiol (p < 0.005). However, we have not detected a significant increase in the PR expression by GW9662 in the absence of estrogen in the mammary glands. These results suggest that the newly induced ERα required the presence of estradiol in the medium to affect PR expression. On the other hand, GW7845, which is a potent activator of PPARγ, suppressed expression of PR in the presence of estradiol (p < 0.001). Once again when both GW9662 and GW7845 were present in the medium, the effect of GW9662 on PR expression was suppressed by GW7845.
Effects GW9662 on the ERα and PR expression in MAL
The translational significance of the induction of ERα by GW9662 can be the opportunity for ER− or ER+/− breast cancer patients to be able to receive SERM treatment. We, therefore, measured ER protein expression in MAL. The MAL were induced by incubating mammary glands for 10 days with growth promoting hormones in the absence of estradiol. The glands were incubated with DMBA for 24 h on day 3 as described in the “Materials and methods” section. One group of glands served as DMBA control whereas glands from another group were treated with 1 μM GW9662 for the entire culture period of 24 days. The glands containing MAL were processed for histopathology, and expression of ERα and PR was determined in MAL by immunohistochemistry. Results showed that the MALs from both control and GW9662 treatment groups expressed ERα. As expected, there was no apparent difference in the ERα expression between these two groups since there was no estradiol in the medium.
In order to further examine whether the ERα in the MAL is functional, the expression of PR protein was also measured by immunohistochemistry in the MAL and compared between control and GW9662 treatment groups. Results are shown in Fig. 5e–h. While there was very little expression of PR in control MAL, extensive nuclear PR expression in the GW9662 treated glands containing MAL was observed. These results suggested that although ERα is present in MAL in the absence of estradiol, the ER present in MAL may be non-functional since there was no PR expression, which is an ER responsive gene. On the other hand, there was significant induction of PR in GW9662 treated glands, which indicated that GW9662 may have rendered ERα in MAL functional.
Effects of Tamoxifen singly or in combination with GW9662 on MAL incidence in MMOC
As shown here and elsewhere , we were unable to show an effect by Tamoxifen on MAL incidence following the treatment of the glands with Tamoxifen during the first 4 days (0–4 days) or during 10 days of incubation (Fig. 3; Table 1). However, when we treated the glands with GW9662 for 4 days followed by Tamoxifen in the absence of GW9662 for 6 days (4–10 days) we saw a 50 % reduction in MAL incidence (Table 1). These results indicate that treatment with GW9662 induced expression of functional ERα, which in turn induced PR and rendered the MAL lesions sensitive to inhibition by Tamoxifen.
Effects of GW7845 and GW9662 singly and in combination on the expression of PPARγ-responsive genes
There is substantial evidence demonstrating that PPARγ and ERα closely interact during normal development and cancer progression . Here, we evaluated the effects of the PPARγ agonist GW7845 and the PPARγ antagonist GW9662 on DMBA-induced MALs in a mouse mammary organ culture (MMOC) model [13, 16, 17]. MAL occur in DMBA-treated MMOC incubated with a mixture containing insulin, prolactin, aldosterone, and hydrocortisone (IPAF). Although the ERα expression in MAL remains intact, they do not respond to anti-estrogen. We show that GW9662, a PPARγ antagonist, induced expression of ERα in the epithelial cells of mammary glands which are ER− or ER+/− in MMOC. Moreover, the expression of ER was significantly enhanced in the presence of estradiol. However, induction of ERα by GW9662 was suppressed upon co-incubation with GW7845. We then sought to determine whether estrogen-responsive genes are also affected by GW9662. We show that GW9662 also increased expression of the PR, as determined by qRT-PCR and protein by immunohistochemistry. There was no expression of PR in the control MAL. These results indicated that the ERα in MAL in the absence of estradiol is present but non-functional and is unable to induce PR. That may be the reason that MAL incidence is not reduced by Tamoxifen in the absence of estradiol. However, PR expression was dramatically increased in the MAL by GW9662, suggesting that GW9662 was able to make ERα functional. If this indeed is the case then one would expect that the MAL induced in the glands in the presence of GW9662 ought to respond to Tamoxifen. Our results showed that the glands treated with GW9662 responded to Tamoxifen, and the MAL incidence was reduced. This is consistent with a recently published report, where it was shown that treatment of mice with GW9662 made the tumors sensitive to antiestrogen, Fulvestrant, treatment .
Since GW9662 is a PPARγ antagonist we sought to understand the connection between its ability to induce ERα and its direct effect on PPARγ. Importantly, our results showed that neither antagonist nor agonist modulated expression of PPARγ mRNA or protein, suggesting that these two ligands actually affect the catalytic properties of PPARγ. In order to further validate the specificity of GW7845 and GW9662, we examined the expression of PPARγ-responsive genes in these glands. Several target tissue specific PPARγ-responsive genes have been reported, including adipsin and aP2 a fatty acid binding protein . The results showed that expression of both adipsin and aP2 was downregulated by GW9662 and that it was enhanced by GW7845. These results, once again, confirm that the effects of GW9662 and GW7845 are specific, leading to the inhibition or activation of PPARγ, respectively. However, we observed that the transient transfection of MCF-7 breast cancer cells with PPARγ expression plasmid did not respond to GW9662 and did not enhance ERE activity in the cells co-transfected with ERE-luc reporter plasmid. Similarly incubation of PPARγ expression plasmid transfected cells with GW9662 did not exhibit expression of PR mRNA as measured by qRT-PCR (data not shown). These results suggest that the action of GW9662 may be mediated not by the direct association of GW9662 with PPARγ but through a different mechanism.
It has been previously demonstrated that inhibition of PPARγ using either a dominant negative trans-gene (Pax-8) or pharmacologic intervention with GW9662 in vivo [1, 19–21] induced expression of ERα in mammary tumors and, furthermore, that addition of faslodex, an ERα inhibitor, completely inhibited the appearance of these lesions. Thus, the Induction of ERα and PR in glands that are ERα− by the PPARγ antagonist GW9662 may have translational significance whereby SERM treatment such as Faslodex (Fulvestrant) or Tamoxifen may inhibit breast lesions in these individuals (21]. Furthermore, we showed that the ERα present in MAL induced by DMBA in the absence of estrogen may be non-functional since these MAL do not express any PR. However, treatment with GW9662 induced PR significantly indicating that one of the roles of GW9662 may be to make ERα functional in these lesions.
In summary, we demonstrated that both agonists and antagonists of PPARγ can suppress the incidence of carcinogen-induced MAL in the MMOC model. Furthermore, we showed that GW9662 an antagonist but not GW7845 an agonist of PPARγ can induce ERα and PR in the presence of estradiol in the mammary glands and facilitate non-functional ERα in the MAL to induce PR. Experimentally, we showed that induction of ERα by GW9662 is associated with a functional receptor since a sequential incubation of the glands with GW9662 for 4 days (0–4 days) followed by incubation with Tamoxifen in the absence of GW9662 (4–10 days) resulted in the suppression of MAL in these glands. This finding could provide a strategy to induce ERα with GW9662 in ER and PR− breast cancer patients, making them potentially responsive to anti-estrogen intervention.
This work was supported by the National Cancer Institute contract N0-CN-43303. We thank Dr. Nishant Tiwari for his help during the early part of the project.
Conflict of interest
There is no conflict of interest for any of the authors listed on this manuscript.