The calcium pump plasma membrane Ca2+-ATPase 2 (PMCA2) regulates breast cancer cell proliferation and sensitivity to doxorubicin

Regulation of Ca2+ transport is vital in physiological processes, including lactation, proliferation and apoptosis. The plasmalemmal Ca2+ pump isoform 2 (PMCA2) a calcium ion efflux pump, was the first protein identified to be crucial in the transport of Ca2+ ions into milk during lactation in mice. In these studies we show that PMCA2 is also expressed in human epithelia undergoing lactational remodeling and also report strong PMCA2 staining on apical membranes of luminal epithelia in approximately 9% of human breast cancers we assessed. Membrane protein expression was not significantly associated with grade or hormone receptor status. However, PMCA2 mRNA levels were enriched in Basal breast cancers where it was positively correlated with survival. Silencing of PMCA2 reduced MDA-MB-231 breast cancer cell proliferation, whereas silencing of the related isoforms PMCA1 and PMCA4 had no effect. PMCA2 silencing also sensitized MDA-MB-231 cells to the cytotoxic agent doxorubicin. Targeting PMCA2 alone or in combination with cytotoxic therapy may be worthy of investigation as a therapeutic strategy in breast cancer. PMCA2 mRNA levels are also a potential tool in identifying poor responders to therapy in women with Basal breast cancer.


PMCA2 expression in human breast tissue exhibiting lactational remodeling and malignant
transformation. Elevated PMCA2 is a feature of mammary glands from lactating mice 2,22 , however, PMCA2 expression has not been assessed in human breast tissue undergoing lactational change. Therefore, we used unique tissue specimens from a breast cancer patient in the third trimester of pregnancy to investigate PMCA2 expression in histologically normal glandular tissue in the context of lactational remodeling (morphological changes and positive β -casein staining are shown in Fig. 1A,B). Positive PMCA2 staining was observed on the plasma membranes of the epithelial cells, but not the surrounding stromal cells (Fig. 1C,D). The magnified image (Fig. 1D, arrows and insert) shows elevated PMCA2 expression on the luminal membrane compared to the basal membrane, consistent with a role for PMCA2 in the direct transport of Ca 2+ into milk.
The possible pathological role of PMCA2 was assessed in human breast cancer samples. TMAs comprising 96 breast tumors in duplicate were assessed for PMCA2 expression by IHC. The tumors were mostly histological grade 3 invasive ductal carcinomas (Supplementary Table 1). Figure 1E-G show examples of the three types of staining observed: no staining, cytoplasmic staining or membranous staining. Since PMCA2 is a plasma membrane Ca 2+ -transporter 23 , tumors with cytoplasmic but not membrane staining were classified as negative in our analysis. In total, 9/96 (~9%) of invasive tumors analyzed were PMCA2 plasma membrane positive (PMCA-PM+ ). We investigated associations between PMCA2 plasmalemmal expression and key prognostic indicators (histological grade, estrogen receptor (ER), progesterone receptor (PR) and HER2 status). A significant correlation between PMCA2-PM+ staining and the common breast cancer pathological markers, tumor grade, ER, PR or HER2 status was not observed (Table 1; Fisher's exact tests, P > 0.05). However, we found a relationship with HER2-positivity (Table 1), with 8/9 of the PMCA2-PM+ cases also HER2+ according to clinical diagnostic criteria (> 6 copies of the ERBB2 gene by SISH). PMCA2 membrane expression is not particularly frequent in breast cancer and so in a cohort of 96 tumors this association did not reach statistical significance (Fisher's exact P = 0.077), nevertheless given that the same trend was observed by others in a separate cohort 19 this relationship could illuminate aspects of the biology underlying PMCA2 function and/or behavior of HER2+ breast tumors.
PMCA2 mRNA is significantly enriched in the Basal breast cancer molecular subtype where is it associated with survival. PMCA2, PMCA1 and PMCA4 mRNA levels were compared in breast cancer molecular subtypes 24 from RSEM data from the TCGA consortium ( Fig. 2A-C). Consistent with the IHC data presented above, there were individual breast cancers with relative high levels of PMCA2 (ATP2B2) in all of the molecular subtypes ( Fig. 2A), this was not seen for PMCA1 (ATP2B1) or PMCA4 (ATP2B4) (Fig. 2B,C). PMCA2 levels were, however, significantly enriched in the Basal molecular phenotype compared to HER2, Luminal A and Luminal B. Assessment of PMCA2 levels in Basal breast cancer cell lines identified PMCA2 as the minor isoform at the mRNA level in all basal breast cancer cell lines, with a trend for cell lines with higher levels of PMCA2 to be identified as Basal B (Fig. 3A). Indeed, PMCA2 levels were significantly greater in Basal B breast cancer cell lines compared to Basal A, a trend which was also observed for PMCA1 but not PMCA4 (Fig. 3B). In contrast to basal breast cancer cell line differences, PMCA2 levels were not significantly different between the recently defined triple-negative breast cancer (TNBC) molecular subtypes, BLIS, BLIA, LAR and MES (Fig. 4A) and there was no significant distribution of tumor subtypes in the low and high PMCA2 expression groups (Fig. 4B). However, assessment of patient survival in TNBC, identified PMCA2 levels as positively associated with patient survival, Scientific RepoRts | 6:25505 | DOI: 10.1038/srep25505 this stratification was more pronounced than the stratification by the defined molecular subtypes in TNBC; BLIA, BLIS and MES (Fig. 4C). The positive association between PMCA2 levels and survival in Basal breast cancers was also observed in three different patient cohorts and this positive association was not consistently seen for PMCA1 or PMCA4 (Supplementary Tables 2-4). PMCA2 mRNA was significantly elevated in specific breast cancers in both HER2 and Basal molecular subtypes, although within each subtype there was clear variation ( Supplementary  Fig. 1), this variance within Basal breast cancers may be the cause of the association between PMCA2 and survival in this subtype. Correlation analysis demonstrated a positive and significant correlation between PMCA2 and the Basal marker EGFR across all breast cancers, however, within the Basal subtype this association was negative ( Supplementary Fig. 1).  25 ) by real-time RT-PCR. Its expression was >3-fold lower than PMCA1 and PMCA4 isoforms (Fig. 5A), which contrasts with rodent models of lactation where PMCA2 is the predominant isoform 6,17 and was consistent with our RNAseq cell line data of MDA-MB-231 and other Basal breast cancer cell lines. PMCA2 mRNA expression was significantly higher in confluent compared to sub-confluent cells (Fig. 5B). PMCA2 protein analysis in this cell line, suggested low levels of PMCA2 protein or a lack of full length PMCA2 protein in MDA-MB-231 cells, despite confidence in the antibody and techniques used (e.g. PMCA2 was readily detectable in human tissue; Fig. 1D,G). This may relate to the transient expression of full length PMCA2 protein during specific cell cycle stages and/or the potential for a PMCA2 fragment to be expressed which has been reported to have biological activity.

Silencing PMCA2 inhibits proliferation of breast cancer cells. Exogenous expression of PMCA2
in luminal T47D breast cancer cells protects against ionomycin-mediated death 19 . Here, we assessed the effect of silencing endogenous PMCA2 on the proliferation of basal-like MDA-MB-231 cells, which express elevated PMCA2 mRNA compared to non-malignant breast cell lines 26 . PMCA isoform expression was silenced using    Fig. 2). Assessment of proliferating cells using EdU staining showed that PMCA2 silencing reduced the percentages of cells in S-phase by 17% and 33% (P < 0.05; Fig. 6A), with corresponding decreases in total cell number of 43% and 53%, using two separate siRNAs (Fig. 6B). In contrast, silencing PMCA1 or PMCA4 had no significant effect on cell cycle progression rate or cell number (Fig. 6).
Effects of PMCA2 silencing combined with cytotoxic chemotherapy on Ca 2+ signaling and proliferation. The use of rational combination therapies reduce the likelihood that tumors will develop therapeutic resistance 27 , and that the patient will experience toxic side effects 28 . Others have shown PMCA2 deficiency leads to increased sensitivity to Ca 2+ -induced apoptosis 19 , and we hypothesized that PMCA2 suppression may enhance the effects of cytotoxic chemotherapy on breast cancer cells. We tested this by assessing doxorubicin efficacy in the siPMCA2 MDA-MB-231 model. Consistent with the anti-proliferative effects identified in high-content analysis (

Discussion
The altered expression of specific Ca 2+ channels is a characterizing feature of many cancers [7][8][9] . These include enhanced expression of specific isoforms of transient receptor potential (TRP) [29][30][31] and Orai Ca 2+ permeable ion channels 32,33 , as well as voltage [34][35][36] and ligand gated Ca 2+ channels [37][38][39] . Although not as widely characterized, altered expression of particular isoforms of p-type Ca 2+ -ATPase family members is associated with specific cancer subtypes, for example, elevated levels of SPCA1 in basal-like breast cancers 40 and of SERCA2 in colorectal cancers 41,42 . The identification of PMCA2 mRNA in breast cancer cell lines 20,26 , PMCA2 protein in clinical breast cancer specimens 19 , and a role for PMCA2 in the transport of Ca 2+ into milk during lactation 4 , highlight the relevance of this p-type ATPase in the context of human breast cancer and the physiology of the human breast.
Our finding that PMCA2 is expressed at the apical membrane of luminal epithelia in the pre-lactational human breast is consistent with data from rodent models demonstrating that PMCA2 is a key pump responsible for the efflux of Ca 2+ from the maternal compartment into milk. In the breast cancer cohort assessed in this study, 9/96 tumors expressed PMCA2 in the tumor cell plasma membrane. Consistent with VanHouten et al. 19 , we found a positive association between PMCA2 expression and HER2 status, with eight out of the nine PMCA2 membrane-positive cases classified HER2+ . This relationship did not reach statistical significance, owing largely to the size of the PMCA2+ subgroup in our study (n = 9 cases). A key difference between our study and VanHouten's was that we assessed the subcellular localization of PMCA2. We analyzed membrane-associated PMCA2 as a categorical variable, whereas the previous study used digital scoring to quantify overall tumor cell positivity as a continuous variable. This is an important distinction given that the membrane residence time of ion pumps is dynamic and often tightly regulated, and subcellular compartment-specific expression of PMCA2 alternative splice isoforms 43 has not been thoroughly investigated.
Our investigation of PMCA2 levels in molecular breast cancer subtypes supported our IHC data of high levels of PMCA2 across different breast cancer subtypes. Specific breast cancers of the Basal, Luminal A, Luminal B and HER2 molecular subtypes had high levels of PMCA2, this was not as obvious for PMCA1 and PMCA4. However, this large cohort identified that PMCA2 mRNA levels were significantly higher in basal breast cancers overall compared to Luminal A, Luminal B and HER2 subtypes. The absence of any differences in PMCA1 and PMCA4 in the different molecular subtypes reinforces the potential unique roles of the PMCA2 isoform in the breast in both lactation and in breast cancer. Although PMCA2 levels were higher in Basal B vs Basal A breast cancer cell lines, PMCA2 levels were not different amongst the recently identified TNBC molecular subtypes BLIS, BLIA, LAR and MES. However, PMCA2 levels were highly correlated with patient survival in triple negative breast cancers and basal breast cancers, which was seen across multiple cohorts. In these cases, high levels of PMCA2 were associated with better patient survival. This is in contrast to a previous report of Oncomine cDNA microarray data in only patients under the age of 50, which found a negative association between high PMCA2 levels and survival 19 . Our identified relationship between high levels of PMCA2 mRNA and patient survival in basal breast cancers, may represent an ability for PMCA2 to identify less aggressive basal breast cancers and signify that PMCA2 overexpression is a not a driver in breast cancer. The potential dichotomy in PMCA2 levels between subtypes and its correlation with survival in the basal subtype is exemplified by the very different association between EGFR and PMCA2 in all breast cancers (positive correlation) versus the basal subtype (negative correlation). Hence, within basal breast cancers, PMCA2 may associate with characteristics of better prognosis which may make it a biomarker for good survival. This association does not exclude the potential of PMCA2 as a drug target in some breast cancer cells, either via the previously proposed mechanism of promotion of apoptosis through PMCA2 inhibition 19 and/or the anti-proliferative effects of PMCA2 inhibition identified in MDA-MB-231 cells Previous studies of PMCA2 in human breast cancer cells have focused on the role of this Ca 2+ pump in protection against cell death mediated by agents that produce sustained increases in [Ca 2+ ] CYT 19 or its role in apoptosis regulation through interactions with calcineurin 21 . However, calcium signaling also plays a vital role in cell cycle regulation in cancer cells 7 and global inhibition of PMCA expression reduces the proliferation of ER+ , luminal-like MCF-7 breast cancer cells 44 . In these studies, PMCA2 was the only PMCA isoform associated with proliferation of basal-like MDA-MB-231 cells, despite the lower expression of this pump relative to other isoforms. The lack of any effect of PMCA1 silencing on cellular proliferation may seem surprising given that this isoform is the predominant regulator of global [Ca 2+ ] CYT in MDA-MB-231 cells 45 . However, Curry et al. demonstrated that despite only modest effects on global [Ca 2+ ] CYT with PMCA4 silencing, PMCA4 but not PMCA1 silencing augmented apoptosis mediated by the Bcl-2 inhibitor, ABT-263, likely through selective effects on Ca 2+ dependent NFkB activity 45 . The ability of PMCA2 to regulate MDA-MB-231 proliferation may also be due to localized specific regulation of Ca 2+ dependent transcription factors involved in proliferation. However, other mechanisms are also possible. For example, the pronounced increase of PMCA2 mRNA in MDA-MB-231 cells with increasing confluence in vitro may reflect dynamic expression of PMCA2 at critical stages of the cell cycle, such that silencing of PMCA2 has pronounced effects only at specific stages of cell division. Moreover, the ability of specific regions of PMCA2 to interact with calcineurin 21 suggests that the entire intact PMCA2 protein may not be required to have a functional consequence in at least some breast cancer cells.
Our data raise the possibility that PMCA2 depletion or inhibition could be a chemo-sensitizing strategy in some breast cancer cells 21 . On the background of PMCA2 deficiency, we found that a low dose of doxorubicin was sufficient for pronounced inhibition of MDA-MB-231 proliferation in vitro, suggesting tumor-targeted PMCA2 depletion or inhibition may allow the use of doxorubicin doses associated with a better side effect profile in these cells.

Conclusions
These data provide further evidence for an important role of PMCA2 in calcium transport during human lactation, and the expression of PMCA2 in a significant percentage of breast cancers. PMCA2 function does not appear to be restricted to the regulation of cell death pathways in breast cancer cell lines, and may regulate other hallmarks of cancer including sustained cellular proliferation 46 . Targeting PMCA2 to reduce breast tumor cell proliferation and increase sensitivity to cytotoxic chemotherapy is a strategy worth further investigation.

Methods
Human clinical samples. Formalin-fixed, paraffin-embedded (FFPE) samples of histologically normal human breast tissue exhibiting pregnancy-induced lactational change were obtained from Pathology Queensland. This patient presented with breast cancer in the third trimester of pregnancy, and underwent a wide local excision procedure. Histopathologic diagnostic assessment revealed a grade 3 invasive ductal carcinoma of no special type  (negative for estrogen, progesterone and human epidermal growth factor receptors ('triple-negative')), associated with high-grade ductal carcinoma in situ and lymph node metastases. Immunohistochemistry (IHC) was performed on glandular tissue in the specimen that exhibited hyperplasia, but no evidence of in situ or invasive disease. Physiological hyperplasia is an expected feature in the pre-lactation breast.
For breast cancer studies, tissue microarray (TMA) sections containing duplicate tissue cores (0.6 mm) from 96 tumors (supplementary data, Table 1) were constructed using archival FFPE blocks from Pathology Queensland 47 . The retrospective analysis of archival human clinical samples in this study was approved the human research ethics committees at the Royal Brisbane and Women's Hospital and The University of Queensland (UQ 2005000785; RBHW 2005/022). All studies were conducted in accordance with institutional approved guidelines.
Immunohistochemistry. Tissue sections (4 μ m) were deparaffinized, rehydrated, washed and heated in 0.01 M citrate buffer (pH 6) at 125 °C for 5 min and at 90 °C for 10 min in a decloaking chamber (Biocare Medical). Sections were stained using the rabbit anti-PMCA2 ATPase polyclonal antibody (1:300; PA1-915 Thermo-Fisher Scientific) or β -casein monoclonal antibody (1:100; sc-53189, Santa Cruz), and the MACH-1 Universal HRP-Polymer Detection Kit (Biocare Medical) according to the manufacturer's instructions. Nuclei were counterstained with hematoxylin using a Varistain Gemini ES Automated Slide Stainer (Thermo Fisher Scientific). The negative and positive controls were no primary antibody and cerebellar tissue, respectively. Stained tissue sections were scanned at 20× magnification using a ScanScope XT Digital Slide Scanner (Aperio), and evaluated by a blinded pathologist (LdS) using the following criteria: (1) positive: intense plasma membrane staining with or without cytoplasmic staining; (2) negative: cytoplasmic or no staining (since PMCA2 is a plasma membrane Ca 2+ -transporter). Assignments of PMCA expression groups. Both the UNC 51 and Veridex (VDX) 54 cohorts utilize microarray expression. Affymetrix probe sets for VDX genes were combined and averaged and used in downstream analysis (PMCA2/ATP2B2 204685_s_at, 211586_s_at and 216120_s_at, PMCA1/ATP2B1 209281_s_at, 212930_at and 215716_s_at and PMCA4/ATP2B4 205410_s_at, 212135_s_at and 212136_at). For high and low expression groups in the UNC and VDX cohorts of Basal-like and/or Claudin-low tumors, receiver-operator characteristic (ROC) curves were produced for PMCA2, PMCA1 and PMCA4 expression against survival outcome (relapse-free survival (RFS) and distant-metastasis-free survival (DMFS)). These curves were produced using the software MedCalc (www.medcalc.org), with optimal values used to call expression cutoff points. Percentile cutoffs are reported in supplementary Tables.

Survival Analysis. Survival analysis was performed in both UNC and VDX cohorts and with the online tool
Kaplan-Meier Plotter 55 . For the UNC and VDX cohorts, RFS and DMFS, respectively, were stratified on the basis of PMCA2, PMCA1 and PMCA4 expression groups as described above. Univariate Cox proportional-hazards regression was carried out using MedCalc with results reported in the supplementary data. The Kaplan-Meier curve was produced using Prism software with Log-rank hazards ratios and P-values reported with each figure. Survival analysis from the Kaplan-Meier Plotter cohort of breast tumors was done using the ' Auto-select best cutoff, ' feature on the website, which analyses the median, tertile and quartile cutoffs for the more significant P-value.
Real-time RT-PCR. RNA was isolated as previously described 26 and reverse transcribed using the Omniscript RT kit (Qiagen). Real-time RT-PCR was performed using Taqman Fast Universal PCR Master Mix and gene expression assays: PMCA1 (Hs00155949_m1), PMCA2 (Hs00155975_m1), PMCA4 (Hs00608066_m1) with 18S rRNA as an input control (4319413E). Reactions were performed using StepOnePlus system (Applied Biosystems) with universal cycling conditions. Relative mRNA expression levels were determined using the comparative C T method 56 .
EdU incorporation assays: cell proliferation and cell cycle. Total MDA-MB-231 cell numbers and the proportion in S-phase of the cell cycle were assessed as previously described 57 . Briefly, 120 h after siRNA transfection, cells were treated with EdU (10 mmol/L), fixed with 3.7% formaldehyde, and permeabilized with 0.5% Triton X-100. The Click-iT reaction cocktail (Alexa Fluor 555; Life Technologies) was incubated with the cells, followed by DAPI (4′ 6-diamidino-2-phenylindole; 400 nmol/L). The cells were imaged with the ImageXpress ® Micro (Molecular Devices) automated epiflourescent microscope (10× objective). The DAPI and EdU stained cells were detected as described previously 57 , and analysis was performed using the multiwavelength cell scoring application module (MetaXpress).

Treatment with a cytotoxic to assess cell proliferation and intracellular-free Ca 2+
[Ca 2+ ] CYT . MDA-MB-231 cells were transfected with siRNAs for 48 h as described above. Cells were pulse treated with Doxorubicin (Doxo, 20 nM) for 24 h, cells were washed twice with Phosphate Buffered Saline (PBS), and the media was replaced with standard growth media. To assess cell proliferation, the total area of the cells was assessed for a period of 59 h using a kinetic imaging system, IncuCyte ZOOM (Essen Bioscience). Intracellular-free Ca 2+ [Ca 2+ ] CYT was assessed 48 h after doxorubicin treatment using the BD PBX Calcium Assay Kit (BD Biosciences 58 ) as described previously 57 with minor modifications. Briefly, cells were loaded with the Calcium Indicator, 5% PBX Signal Enhancer and probenecid (500 μ mol/L) in physical salt solution (PSS; with 1.8 mmol/L CaCl 2 ) for 1 h at 37 °C. The loading solution was replaced with PSS containing nominal Ca 2+ , 5% PBX Signal Enhancer and probenecid (500 μ mol/L). Fluorescence was assessed with an excitation intensity of 470-495 nm and a 515-575 nm emission filter using a Fluorescence Imaging Plate Reader (FLIPR) TETRA (Molecular Devices). Fluorescence was normalized to the baseline fluorescence and expressed as 'relative [Ca 2+ ] CYT ' . Statistical Analysis. Statistical associations between PMCA2 and breast cancer prognostic indicators were evaluated using the Fisher's exact test. Statistical significance for the remaining data was assessed as described in individual figure legends. All statistical analyses were performed using GraphPad Prism (version 6.04 for Windows and version 6.0f for Mac OS X, GraphPad Software, Inc.).