Introduction
Breast cancer accounts for over one-quarter of all cancer diagnoses, with an estimated 200,000 new cases annually [
1]. Despite recent advances in diagnosis and treatment strategies, nearly 40,000 women will die of this disease in 2011 [
1]. The hormone-dependent nature of breast cancer and the important role of estrogen receptor alpha (ERα) in initiation and progression supported development of pharmacologic agents to either reduce circulating estrogen levels or modulate ERα functions [
2,
3]. Targeted endocrine therapies significantly reduce mortality in patients with hormone-responsive (ERα-positive) tumors. However, both
de novo and acquired therapy resistance limits treatment efficacy [
4].
ERα transcriptional activity is not only regulated by steroid hormones alone but also requires co-regulatory proteins [
5,
6]. Following hormone stimulation, multiprotein complexes containing ERα co-regulators and transcriptional regulators assemble to regulate gene transcription [
6]. ERα co-regulatory proteins are tightly regulated under normal conditions, with misexpression primarily reported in the literature in association with a number of disease states. Over one-third of the nearly 300 distinct co-regulators identified are overexpressed or underexpressed in human cancers; 38% of co-regulators are overexpressed in breast cancer [
7]. These findings suggest that deregulated co-regulator expression may promote carcinogenesis and/or progression of endocrine-related cancers. ERα-associated co-regulator misexpression contributes to ERα activity and often correlates with poor prognosis [
8,
9]. Consequently, co-regulator expression represents an indirect means of targeting ERα activity.
Estrogen-induced breast carcinogenesis is characterized by aberrant histone modifications [
10]. Ligand-bound ERα promotes various histone modifications on target gene promoters and such modifications are facilitated by ERα co-regulatory proteins. Regulatory effects of histone acetylation and phosphorylation have been extensively characterized. However, the role of histone methylation remains understudied. Unlike acetylation, which generally correlates with gene activation, the consequences of histone methylation are site dependent. For example, histone H3 lysine 4 dimethylation (H3K4me2) on ERα target gene promoters correlates with transcriptional activation, while lysine 9 dimethylation (H3K9me2) associates with repression [
11,
12]. Previous studies show recruitment of lysine-specific histone demethylase 1 (KDM1) to a significant fraction of ERα target genes [
13]. Unlike genetic alterations, epigenetic changes are reversible and therefore represent a promising therapeutic target.
Emerging evidence implicates a functional role of ERα co-regulator proline glutamic acid and leucine-rich protein 1 (PELP1) in the oncogenic properties of cancer cells [
14]. PELP1 deregulation occurs within several hormone-responsive malignancies including breast cancer, ovarian cancer and prostate cancer [
15]. In a subset of human breast tumors, both PELP1 expression and localization are altered [
16]; expression during breast cancer progression is associated with more invasive disease [
16,
17]. In a preclinical study of ERα-positive breast cancer patients, PELP1 expression was identified as an independent prognostic biomarker in assessing clinical outcome; elevated expression associated positively with poor prognosis [
17]. Acting as a scaffolding protein, PELP1 coordinates various signaling pathways with ERα by modulating interactions with known oncogenes and cytosolic kinases [
15]. PELP1 deregulation correlates with increased aromatase expression resulting in tumor proliferation via local estrogen synthesis [
14]. Recent studies indicated that PELP1 interaction with KDM1 plays a key role in PELP1-mediated oncogenic functions [
18]. Although such findings suggest a role for the PELP1-KDM1 axis in breast cancer progression, the therapeutic potential of targeting the PELP1-KDM1 axis is unknown.
In the present article we target the PELP1-KDM1 axis using a nanoliposomal formulation of PELP1-siRNA-1,2-dioleoyl-sn-glycero-3-phosphatidylcholine (DOPC) administered systemically and KDM1 inhibitors in xenograft-based preclinical breast tumor models. Treatment of ERα-positive tumors with PELP1-siRNA-liposomes or pargyline significantly reduced tumor volume. Further, combining KDM1 targeting drugs with current endocrine therapies substantially impeded growth and restored sensitivity of therapy-resistant breast cancer cells. Our data suggest inhibiting PELP1-KDM1-mediated histone modifications as a potential therapeutic strategy for blocking disease progression and therapy resistance among breast cancer patients.
Materials and methods
Cell lines and reagents
Human breast cancer MCF-7 cells were obtained from American Type Culture Collection (Manassas, VA, USA). All of the proposed cells were passaged in the user's laboratory for fewer than 6 months after receipt or resuscitation. MCF-7-PELP1 cells [
19], MCF-7-HER2 [
20], MCF-7-TamR cells [
20] and MCF-7-LTLTca cells [
21] have been described earlier. MCF-7 cells transfected with control vector were used as controls. MCF-7-LTLTca and MCF-7-TamR cells were cultured in Phenol red-free RPMI medium containing 5% dextran charcoal-treated serum supplemented with either 1 μmol/l letrozole or 1 μmol/l tamoxifen, respectively.
Estradiol (catalogue number E2257), tamoxifen (catalogue number H7904), androstenedione (catalogue number A9630) and pargyline (catalogue number P8013) were purchased from Sigma (St Louis, MO, USA).
N-((1S)-3-(3-(trans-2-aminocyclopropyl)phenoxy)-1-(benzylcarbamoyl)propyl)benzamide (NCL-1) was synthesized as previously described [
22].
The anti-PELP1 (catalogue number 300-180A) and anti-KDM1 (catalogue number A300-215A) antibodies were purchased from Bethyl Laboratories (Montgomery, TX, USA). Anti-GFP antibody (catalogue number 632381) was purchased from Clontech (Mountain View, CA, USA). Anti-dimethyl-H3K4 (catalogue number 07-030) and anti-H3K9 (catalogue number 07-441) antibodies were purchased from Upstate (Chicago, IL, USA). Anti-acetyl-histone H3 (lys9; catalogue number 9671) was purchased from Cell Signaling (Danvers, MA, USA).
The terminal deoxynucleotidyl transferase dUTP nick end-labeling (TUNEL) kit (catalogue number 11684795910) for apoptosis detection was purchased from Roche (Mannheim, Germany) and Ki-67 anti human Clone MiB-1 antibody (catalogue number M7240) was purchased from Dako (Carpinteria, CA, USA). The PELP1 (catalogue number L004463-00-0050) and KDM1 (catalogue number L-009223-00-0005) nontargeting control (catalogue number D-001810-01-05) SMARTpool siRNA duplexes were purchased from Dharmacon (Lafayette, CO, USA).
Real-time PCR
Cells were harvested with Trizol Reagent (Invitrogen, Carlsbad, CA, USA) and total RNA was isolated according to the manufacturer's instructions. cDNA synthesis was performed using the Superscript III RT-PCR kit (Invitrogen). Real-time PCR was carried out using a Cepheid SmartCycler II (Sunnyvale, CA, USA) with gene-specific real-time PCR primers. Results were normalized to actin transcript levels and the difference in fold expression was calculated using the ΔΔCT method. The primers for aromatase were 5'-AAATCCAGACTGTTATTGGTGAGAG-3' (sense) and 5'-GTAGCCATCGATTACATCATCTTCT-3' (antisense), and the primers for GREB1C were 5'-GGCAGGACCAGCTTCTGA-3' (sense) and 5'-CTGTTCCCACCACCTTGG-3' (antisense).
Cell proliferation assay
The cell proliferation rate was measured using a 96-well format with Cell Titer-Glo Luminescent Cell Viability Assay (G7572; Promega). Cells (5×103) were plated in each well of Corning® 96-well, clear, flat-bottom, opaque wall microplates and cultured in RPMI media containing 2.5% Dextran Charcoal (DCC) treated serum for 24 hours and followed by treatment with or without estradiol (100 nM/well) for an additional 72 hours. Luminescence was recorded using automatic Fluoroskan Luminometer as per the manufacturer's recommendations. All experiments are carried out using three biological replicates.
Chromatin immunoprecipitation
The chromatin immunoprecipitation analysis was performed as described previously [
18]. Briefly, MCF-7, MCF-7-HER2 and MCF-7-PELP1 cells were cross-linked using 1% formaldehyde, and the chromatin was subjected to immunoprecipitation using the indicated antibodies. Isotype-specific IgG was used as a control. DNA was re-suspended in 50 µl Tris-EDTA buffer (TE) buffer and used for real-time PCR amplification using the gene-specific primers. The primers for aromatase PI.3/11 promoter regions were 5'-CAAGGTCAGAAATGCTGCAA-3' (sense) and 5'-AGCTCCTGTTGCTTCAGGAGG-3' (antisense), and the primers for GREB1C promoter were 5'-TTGTTGTAGCTCTGGGAGCA-3' (sense) and 5'-CAACCAGCCAAGAGGCTAAG-3' (antisense).
Preparation of liposomal siRNA
For
in vivo delivery, PELP1 siRNA was incorporated into DOPC nanoliposomes as described previously [
23,
24]. Briefly, siRNA and DOPC were mixed in excess tertiary butanol at a ratio of 1:10 (w/w) respectively. Subsequently, Tween 20 was added to the mixture at the ratio of 1:19 (Tween 20:siRNA/DOPC). The mixture was vortexed and frozen in an acetone/dry-ice bath and lyophilized. For
in vivo administration, the mixture was hydrated with 0.9% saline to a concentration of 15 μg/ml, and 200 to 250 μl of the mixture was used for each injection. siRNA for preparation of liposomes were purchased from Sigma. The targeted sequences used were 5'-CCACAGAGCCUGACUCCUA-3' for PELP1 and 5'-UUCUCCGAACGUGUCACGU-3' for control.
Tumorigenesis assays
All animal experiments were performed after obtaining University of Texas Health Science Center, San Antonio Institutional Animal Care and Use Committee approval, and animals were housed in accordance with the University of Texas Health Science Center, San Antonio institutional protocol for animal experiments.
For tumorigenesis studies, model cells (5×10
6) were injected into the mammary fatpad of 6-week-old to 7-week-old female nude mice (
n = 10 per group) as described elsewhere [
16]. Athymic nude mice (
nu/nu) were injected with control MCF-7 cells or with MCF-7 cells that overexpress PELP1 by mixing them with an equal volume of Matrigel
™ Matrix (BD Biosciences San Jose, CA, USA). In the premenopausal model, mice received one 60-day release pellet containing 0.72 mg 17β-estradiol (Innovative Research of America, Sarasota, FL, USA) 1 week before implantation of cells. For the postmenopausal model, mice were subjected to ovariectomy 1 week prior to tumor cell inoculation. Owing to the deficiency of adrenal androgens in this model, athymic mice were supplemented with subcutaneous injections of the aromatase substrate androstenedione (100 µg/day) for the duration of the experiment as described for the postmenopausal model [
25].
To examine the effects of PELP1 siRNA therapy on tumor growth, treatment was initiated 1 week after intraperitoneal injection of tumor cells. Mice were randomly assigned to two groups (
n = 10 mice per group): control siRNA-DOPC (150 μg/kg intraperitoneally twice weekly), and PELP1 siRNA-DOPC (150 μg/kg intraperitoneally twice weekly). The mice were monitored daily for adverse toxic effects. Tumor growth was measured with a caliper at weekly intervals, and the volume was calculated using a modified ellipsoidal formula:
where L is the longitudinal diameter and W is the transverse diameter. At the end of each experiment, the mice were euthanized, and the tumors were removed, weighed and processed for immunohistochemistry (IHC) staining.
Immunohistochemistry
Immunohistochemical analysis was performed as described elsewhere [
14]. Briefly, tumor sections were incubated overnight with the primary antibodies PELP1 (1:750), H3K9me2 (1:50), H3K4me2 (1:50), H3K9ac (1:50) and Ki-67 (1:150) in conjunction with proper controls. The sections were then washed three times with 0.05% Tween, incubated with secondary antibody for 1 hour, washed three times with 0.05% Tween in PBS, visualized by 3,3'-diaminobenzidine (DAB) substrateand counterstained with hematoxylin QS (Vector Lab, Burlingame, CA, USA). The proliferative index was calculated as the percentage of Ki-67-positive cells in 10 randomly selected microscopic fields at 40× per slide. TUNEL analysis was performed using the
in situ Cell Death Detection Kit (Roche, Indianapolis, IN, USA) as per the manufacturer's protocol, and 10 randomly selected microscopic fields in each group were used to calculate the relative ratio of TUNEL-positive cells. The H3K9me2 and H3K4me2 expression of tumors was quantified as 100× the number of positive cells divided by the total number of cells counted under 40× magnification in 10 randomly selected areas in each tumor sample.
Statistical analysis
Statistical differences among groups were analyzed with either the t test or analysis of variance when appropriate using Prism software (Irvine, CA, USA).
Discussion
Human ERα is implicated in breast cancer initiation and progression. Despite the positive effects of hormonal therapy using anti-estrogens and aromatase inhibitors, de novo and/or acquired resistance to endocrine therapies frequently occurs. Alternate therapies are urgently needed to address this major clinical problem. In this study, we tested the hypothesis that deregulation of PELP1 promotes activation of KDM1-driven epigenetic modifications at ERα target genes contributing to cancer proliferation/therapy resistance by testing the therapeutic effect of targeting the PELP1-KDM1 axis. We found that downregulation of PELP1 in vivo by PELP1 siRNA liposomes significantly reduced estrogen-mediated breast tumor progression in xenograft model; that drugs targeting KDM1 efficiently reduced PELP1-mediated increases in breast cancer cell proliferation; that KDM1 blockers sensitized therapy-resistant cells to hormonal therapy; that treatment of PELP1 siRNA or KDM1 blockers substantially increased inhibitory histone methyl markers at ER target promoters; and that KDM1 blockers also efficiently reduced breast tumor growth in postmenopausal xenograft models, where tumor growth is driven by local estrogen. Collectively, our results implicate the PELP1-KDM1 axis as a potential therapeutic target for breast cancer.
Changes in ERα co-regulator expression have been demonstrated to substantially contribute to ERα activity and often correlate with poor prognosis. For example, deregulation of the ERα co-regulators SRC3 (AIB1), SRC2 and MTA1 was reported in breast tumors [
7]. SRC3 knockout mice studies demonstrated that normal expression of coactivator SRC3 is required for initiation of tumorigenesis by carcinogens and oncogenes [
31,
32], and overexpression of AIB1 in mouse mammary gland promoted tumorigenesis [
33]. Co-regulator PELP1 is shown to function as a proto-oncogene [
14] and was recently demonstrated to be an independent prognostic marker for poor breast cancer survival [
17]. A recent study suggests that PELP1 mediates androgen receptor activation in the absence of androgens in PCa cells and that disruption of the complex between androgen receptor and PELP1 may be a viable therapeutic strategy in advanced prostate cancer [
34]. We found that systemic administration of PELP1 siRNA in a nanoliposomal formulation (PELP1-siRNA-DOPC) significantly reduced breast tumor growth in a xenograft model. IHC analysis of excised xenograft-based tumors revealed a combined decrease in cellular proliferation, induction of apoptosis and upregulation of inhibitory epigenetic modifications of histone H3K9me2 compared with nontreated control groups. These findings suggest that blocking PELP1 expression and/or actions represent an indirect means of targeting the activity of ERα and that blocking the PELP1 axis could have therapeutic implications for reducing breast cancer growth.
Histone methylation plays a vital role in many neoplastic processes and thus represents a valuable therapeutic target [
35‐
37]. Recent evidence suggests activation or repression of estrogen-induced genes depends on the modulation of histone methyl markers on target gene promoters [
38]. Histone demethylase KDM1 belongs to a growing number of transcriptional complexes that are implicated in tumorigenesis [
39] and is recruited to a significant fraction of ERα target genes [
13]. Our previous studies indicate that PELP1 is a novel co-regulator that participates in ERα-mediated chromatin remodeling events via its interactions with KDM1 [
18]. Pargyline (Eutonyl, Supirdyl, Eutron) is a US Food and Drug Administration-approved drug to treat depression and vascular hypertension. Several recent studies demonstrated that pargyline has the potential to inhibit KDM1. Here we utilized pargyline to examine whether it has the potential to restore altered epigenetic changes in PELP1-driven breast cancer. Our results showed a significant effect of pargyline in reducing PELP1-driven proliferation. Further, pargyline-treated xenograft tissues showed inhibition of
in vivo KDM1 activity as can be seen by increased levels of H3K4me2 and H3K9me2, known substrates of KDM1. This proof-of-principle study demonstrated the significance of the PELP1-KDM1 axis in curbing breast cancer progression. However, an extended period of pargyline use at millimolar concentrations may cause side effects. To overcome this possibility, we are currently developing better inhibitors of KDM1 that work efficiently at lower doses with high specificity - we have developed the compound NCL-1, which showed significant activity in the 5 to 10 μM range. Pargyline-mediated inhibition of breast cancer cell growth was independently validated using the more potent KDM1 inhibitor (NCL-1) and also by using PELP1 and KDM1 siRNAs. Recent studies demonstrating the efficacy of KDM1 inhibitors on reducing growth of neuroblastoma [
40] and cancer stem cells [
41] also corroborate our findings using breast cancer models.
KDM1 can potentially function as a coactivator or co-repressor by demethylating H3K9 or H3K4, respectively, and co-regulators such as PELP1 in conjunction with ERα modulate KDM1 specificity from H3K4me2 to H3K9me2, leading to enhanced ERα target gene activation [
42]. As expected, blockage of KDM1 via pargyline or NCL-1 increased both H3K4me2 and H3K9me2 methylation in MCF-7-PELP1 cells. A significant increase in H3K4 methylation in MCF-7-PELP1 model cells also suggests that lack of KDM1 activity facilitated unopposed action of methylases and that PELP1 has the potential to enhance H3K4 methylation via facilitation of H3K4me2 methylases. The latter potential is supported by emerging findings that PELP1 associates with histone methylases [
15] and due to the fact that PELP1 knockdown significantly reduced H3K4 methylation (Figure
1C). Pargyline-mediated blockage of KDM1 functions may also strongly increase the repressive H3K9me2 marker, and its subsequent conversion to H3K9me3 marker may lead to reduced recruitment of H3K9 acetyltransferases at specific gene promoters. In support of the second possibility, our ongoing studies identified SETDB1 (an enzyme that facilitates conversion of H3K9me2 to H3K9me3) as a PELP1 interacting protein and showed a reduction of activation marker H3K9Ac in pargyline or NCL-1-treated cells, and earlier studies reported the existence of SETDB1, KDM1 and PELP1 complexes [
15]. Blockage of KDM1 functions may provide a favorable environment for SETDB1 to convert H3K9me2 to H3K9me3 under conditions of pargyline treatment. However, future studies are needed to discern these possibilities.
Deregulation of HER2 expression and downstream signaling has emerged as a significant factor in the development of hormonal resistance, and crosstalk with ERα has been shown to promote endocrine therapy resistance [
43]. PELP1 interacts with HER2 and is implicated in facilitating ER crosstalk with HER2 signaling pathways [
9]. Deregulated PELP1 expression during breast cancer progression is associated with more invasive disease [
16,
17]. Additionally, PELP1 is shown to contribute to HER2-mediated local estrogen synthesis via increased aromatase expression [
14]. Our findings suggest that KDM1 targeting inhibitors (pargyline and NCL-1) are efficient in reducing PELP1-mediated HER2-ERα crosstalk. KDM1 inhibitors efficiently reversed HER2-mediated epigenetic changes and promoted inhibitory histone methyl markers at ERα target genes.
Although mechanisms for hormonal therapy resistance remain elusive, emerging data implicate ERα crosstalk with growth factor pathways and deregulation of co-regulators as major causes of resistance [
44]. Earlier studies showed that PELP1 deregulation contributes to therapy resistance [
15,
45]. Because PELP1 interacts with epigenetic modifier KDM1 [
18], in this study we tested whether inhibition of KDM1 by inhibitor pargyline reduced the viability of resistant cells. Combinatorial therapy of anti-estrogen (tamoxifen) with pargyline or NCL-1 showed the most promising therapeutic effect compared with single-agent therapy to inhibit growth of therapy-resistant cells. Results suggest that targeting of the PELP1-KDM1 axis in combination with current endocrine therapies increases therapeutic efficacy and may inhibit or delay development of aromatase inhibitor resistance by promoting favorable epigenetic modifications.
Local estrogen production via deregulated expression of aromatase (Cyp19), the key enzyme in the biosynthesis of estrogen, contributes to tumor progression in postmenopausal women [
46]. Aromatase inhibitors are effective in enhancing patient survival although long-term use is limited by systemic side effects and therapy resistance [
47]. Recent studies from our laboratory demonstrated PELP1 cooperates with HER2 and modulates epigenetic changes at the aromatase promoter by interacting with lysine-specific demethylase (KDM1), leading to local estrogen synthesis [
48]. In this study, we found that KDM1 inhibitors substantially inhibited growth of local estrogen-producing cells (MCF-7-PELP1 and MCF-7-HER2). In the postmenopausal xenograft-based model, treatment with pargyline significantly inhibited the growth of local estrogen-producing PELP1 tumor cells. Our results suggest that drugs targeting the PELP1-KDM1 axis are effective in reversing the methyl modifications at the aromatase promoter that are affected by proto-oncogenes such as PELP1 and HER2 and for blocking growth of local estrogen-producing cells.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
VC performed the majority of the in vitro experiments and xenograft studies, coordinated with all of the team members and drafted the manuscript. MM participated in the KDM1 siRNA studies, demethylase assays and xenograft studies. ST participated in the IHC studies. TS and NM participated in the design and analysis of NCL-1 inhibitor studies. CR-A, GL-B and AKS participated in the design and analysis of siRNA liposome studies. RKV conceived the study, and participated in the design of the project and preparation of final manuscript. All authors read and approved the final manuscript.