SK4 channels modulate Ca2+ signalling and cell cycle progression in murine breast cancer

Oncogenic signalling via Ca2+‐activated K+ channels of intermediate conductance (SK4, also known as KCa3.1 or IK) has been implicated in different cancer entities including breast cancer. Yet, the role of endogenous SK4 channels for tumorigenesis is unclear. Herein, we generated SK4‐negative tumours by crossing SK4‐deficient (SK4 KO) mice to the polyoma middle T‐antigen (PyMT) and epidermal growth factor receptor 2 (cNeu) breast cancer models in which oncogene expression is driven by the retroviral promoter MMTV. Survival parameters and tumour progression were studied in cancer‐prone SK4 KO in comparison with wild‐type (WT) mice and in a syngeneic orthotopic mouse model following transplantation of SK4‐negative or WT tumour cells. SK4 activity was modulated by genetic or pharmacological means using the SK4 inhibitor TRAM‐34 in order to establish the role of breast tumour SK4 for cell growth, electrophysiological signalling, and [Ca2+]i oscillations. Ablation of SK4 and TRAM‐34 treatment reduced the SK4‐generated current fraction, growth factor‐dependent Ca2+ entry, cell cycle progression and the proliferation rate of MMTV‐PyMT tumour cells. In vivo, PyMT oncogene‐driven tumorigenesis was only marginally affected by the global lack of SK4, whereas tumour progression was significantly delayed after orthotopic implantation of MMTV‐PyMT SK4 KO breast tumour cells. However, overall survival and progression‐free survival time in the MMTV‐cNeu mouse model were significantly extended in the absence of SK4. Collectively, our data from murine breast cancer models indicate that SK4 activity is crucial for cell cycle control. Thus, the modulation of this channel should be further investigated towards a potential improvement of existing antitumour strategies in human breast cancer.


Introduction
Breast cancer is the most common malignant disease among women and a major cause of cancer-related mortality in the developed world (Senkus et al., 2015;Siegel et al., 2016). Breast cancer heterogeneity resulting in variable therapy response and survival emphasizes the need for novel diagnostic, prognostic and predictive markers regulating different aspects of tumour cell behaviour in order to improve disease management (Lastraioli et al., 2015).
Members of the Ca 2+ -activated K + channel (K Ca ) family, that is BK (also known as K Ca 1.1, Maxi-K or Slo1) and SK4 (also known as K Ca 3.1 or IK), have been implicated in the progression of malignant diseases including but not limited to gliomas (D'Alessandro et al., 2013;Liu et al., 2002;Turner et al., 2014), clear cell renal cell carcinoma (Rabjerg et al., 2015) and breast cancer (Haren et al., 2010;Oeggerli et al., 2012;Thurber et al., 2017). Aberrant K + efflux via oncogenic K Ca channels may affect multiple parameters in the breast tumour cell such as membrane potential, cytosolic Ca 2+ ([Ca 2+ ] i ), pH and cell volume (Huang and Jan, 2014;Huber, 2013). Current knowledge suggests that BK (Lallet-Daher et al., 2009;Parihar et al., 2003;Stegen et al., 2015) and SK4 (Ouadid-Ahidouch et al., 2004;Parihar et al., 2003;Steinle et al., 2011) activities are required for malignant growth of several tumour-derived cell lines and of xenografts in immunocompromised mice, highlighting a general role of K Ca channels for cell cycle-specific functions (Huang and Jan, 2014;Pardo and Stuhmer, 2014). Expression of SK4 and BK in cancer cells follows a cell cycle-dependent mode (Ouadid-Ahidouch et al., 2004;Pardo et al., 1998) and the mitogen-dependent regulation of K Ca activity supports a role for both channels in malignant Lallet-Daher et al., 2009;Wang et al., 2007a) and nonmalignant cell proliferation (Grgic et al., 2005;Khanna et al., 1999;Toyama et al., 2008;Yu et al., 2013). By inducing a more negative membrane voltage, activation of K + channels provides a driving force for Ca 2+ influx into the nonexcitable tumour cell. In addition, by modifying the membrane potential, K + channels regulate voltage-dependent activity and/or conductivity of Ca 2+ entry pathways (Huang and Jan, 2014;Huber, 2013). Because Ca 2+ is an ubiquitous intracellular messenger relevant for cell survival, apoptosis and proliferation (Azimi et al., 2014;Monteith et al., 2007), it was proposed that breast tumour K Ca channels promote pro-oncogenic functions at least in part by Ca 2+dependent signalling pathways (Kunzelmann, 2005;Ouadid-Ahidouch and Ahidouch, 2013). Oscillations of [Ca 2+ ] i have been observed during exit from G 1 , during S phase, and at the end of M phase (Parkash and Asotra, 2010;Santella et al., 2005). A hyperpolarization of the breast tumour cell membrane reportedly resulted in elevated [Ca 2+ ] i (Wang et al., 1998), which in turn may activate K Ca to control spatiotemporal Ca 2+ dynamics. Accordingly, the SK4 inhibitors TRAM-34 and clotrimazole reduced Ca 2+ entry and cell cycle progression as shown for colon and endometrium cancer cells (Lai et al., 2011;Wang et al., 2007b). In the metastatic breast cancer cell line MDA-MB-231, suppression of SK4 channels with TRAM-34 or siRNA inhibits both cell proliferation and migration and induces apoptosis (Zhang et al., 2016). Moreover, overexpression of SK4 in MDA-MB-231 cells increased primary tumour growth and metastasis (Thurber et al., 2017), whereas TRAM-34 at concentrations known to inhibit plasma membrane SK4 channels stimulates the proliferation of MCF-7 cells presumably by a mechanism that involves oestrogen receptors (Roy et al., 2010). Besides direct effects on the tumour cell, TRAM-34-sensitive SK4 channels decreased the ability of natural killer cells to suppress tumour grafts in the NOD-SCID mouse model (Koshy et al., 2013). Obviously, further studies are needed to clarify the pharmacological properties of SK4 channels in nontumour/tumour cells and in breast cancer. Due to reported nonspecific effects of TRAM-34 and other compounds modulating SK4 (Agarwal et al., 2013;Wulff et al., 2000), proofof-concept studies are required to establish the oncogenic roles of endogenous SK4 channels for Ca 2+ signalling and proliferation and thereby their therapeutic potential in breast cancer.
In order to validate the relevance of SK4 for mammary tumour development in vivo, we crossed the wellestablished polyoma middle T-antigen (PyMT) and epidermal growth factor receptor 2 (cNeu) oncogenedriven mouse breast tumour models to gene-targeted mice lacking functional SK4 (Guy et al., 1992a,b;Sausbier et al., 2006). To test whether oncogenic properties of SK4 directly stem from the tumour cells, all animal studies were supported by a parallel analysis of mammary tumour cells derived from PyMT (and cNeu) transgenic tumours in the presence and absence of SK4 channels. While endogenous SK4 channels appear to be molecular classifiers of breast tumour cell growth, SK4 channels present in the tumour stroma may also contribute to the pathogenesis of breast cancer.
Animals were housed in a 12-h light/dark circle with access to water and food ad libitum. All experiments were approved by the local Ethics Committee for Animal Research (Regierungspraesidium Tuebingen). For genotyping, genomic DNA was isolated from tail-tip biopsies. Primer sets used to detect the MMTV-PyMT and MMTV-cNeu transgenes or the different SK4 alleles were as follows: MMTV-PyMT forward: GGA AGC AAG TAC TTC ACA AGG G; MMTV-PyMT reverse: GGA AAG TCA CTA GGA GCA GGG; MMTV-cNeu forward: TTT CCT GCA GCA GCC TAC GC; MMTV-cNeu reverse: CGG AAC CCA CAT CAG GCC: SK4 forward 1: TAA GTG CTT GCT GAG TCT GGA; SK4 forward 2: CAG GAA GCA CAG GCA CTG C; SK4 reverse: AGG AGA GTG ACT GTA GGT GAG.
Starting at an age of 40 or 120 days, tumour-prone MMTV-PyMT tg/+ or MMTV-cNeu tg/+ SK4 WT and SK4 KO mice, respectively, were palpated weekly until first tumours were identified. Palpation frequency was increased to three times a week after tumour detection and tumour size was measured using a digital calliper. When tumour size had reached a diameter of 15 mm, mice were euthanized with CO 2 and tissue samples were collected for the subsequent analyses.

Tumour preparation and immunohistochemistry
To guarantee optimal fixation, tumour biopsies were carefully adjusted to a standard size of approx. 7.5 9 7.5 mm (length x width) (Tomayko and Reynolds, 1989) using a scalpel. Fixation was performed for 3 h in 4% paraformaldehyde/PBS (Carl Roth, Karlsruhe, Germany) at room temperature.
After several washing steps, standardized tumour samples were cryoprotected using a sucrose gradient starting with 5% sucrose (Carl Roth) for 1 h, followed by 10% sucrose overnight and finally 20% sucrose for 24 h (all diluted in PBS). Upon cryoprotection, tumours were stored at À80°C in Neg-50 TM frozen section medium (Thermo Fisher Scientific, Waltham, MA) for at least 24 h. Prior to the subsequent experiments, tumour tissues were adjusted to À18°C (cutter temperature À21°C), sectioned at 10 lm with a Microm HM 560 (Thermo Fisher Scientific) and mounted on glass slides (Carl Roth). In general, tumour sections derived from SK4 WT and SK4 KO MMTV-PyMT transgenic tumours, respectively, were mounted together on the same glass slides for further processing. Haematoxylin and eosin staining was performed on sections rehydrated by a series of ethanol according to a previously published protocol (Leiss et al., 2011) using Harris haematoxylin solution (Carl Roth) and eosin-g solution (Carl Roth). After removing excessive dye with ethanol 80% (v/v), tumour sections were dehydrated again with 100% ethanol (v/v) and toluol and finally embedded in DePeX (VWR, Darmstadt, Germany). To identify SK4 channels in tumour biopsies, we adopted an established immunohistochemical method (Bausch et al., 2015). In brief, sections were encircled with ImmEdge Hydrophobic Barrier Pen (Vector Laboratories, Burlingame, CA) and then fixed again for 15 min in 4% paraformaldehyde/PBS solution. Unspecific epitopes were blocked for 1 h in blocking solution containing 5% normal goat serum (Vector Laboratories, Burlingame), 1% BSA (Carl Roth), 0.2% glycine (Carl Roth), 0.2% lysine (Carl Roth) and 0.3% Triton X-100 (Carl Roth) in PBS. Sections were then incubated in primary anti-SK4 antibody solution (1 : 1000 dilution in blocking solution) overnight at 4°C. The next day, biotinylated anti-rabbit secondary antibody (1 : 1000 in blocking solution, Vector Laboratories, Burlingame) was applied for 1 h in the dark at room temperature. ABC alkaline phosphatase (AP) solution (Vectastain ABC-AP-Kit Standard Alkaline Phosphatase, Vector Laboratories, Burlingame) was used with Vector Blue AP-substrate (Vector Blue Alkaline Phosphatase Substrate Kit III Vector Laboratories, Burlingame) and levamisole solution (Vector Laboratories, Burlingame) to visualize the antigen-antibody complexes. Stained sections were covered with AquatexÒ (Merck Millipore, Darmstadt, Germany), and digital image acquisition was performed using an Axiocam MRcRev.3 camera (Carl Zeiss GmbH, Jena, Germany) attached to an inverse Axiovert 200 M microscope (Carl Zeiss GmbH).

Cell culture of primary MMTV-cNeu tg/+ and MMTV-PyMT tg/+ tumour cells
Tumour biopsies were obtained from SK4-proficient and SK4-deficient MMTV-PyMT tg/+ tumours and from MMTV-cNeu tg/+ SK4 WT tumours (diameter of approx. 5 mm) and cut into small pieces with a scalpel and then immediately incubated in a falcon tube containing culture medium [IMEM, 5% fetal bovine serum (FBS), 1% penicillin/streptomycin (Life Technologies, Darmstadt, Germany)] on ice. Under sterile conditions, tumour pieces were then transferred into IMEM (Life Technologies) with 1 mgÁmL À1 collagenase D (Roche, Mannheim, Germany) at 37°C for 11 min. To achieve optimal dissociation of the tumour tissue, several titration steps supported enzymatic digestion. Cells were then transferred through a 40-lm cell strainer (BD Biosciences, Heidelberg, Germany) to a petri dish and were grown in monolayer in culture medium [IMEM + 5% FBS and 1% penicillin/streptomycin (Life Technologies)] at 37°C with 5% CO 2 and 95% relative humidity. Cell concentrations were adjusted, depending on experiment set-up, from 80 000 to 5 9 10 6 cells per mL by using 10 lL of the cell suspension for counting in a disposable haemocytometer (Biochrom, Berlin, Germany). We used tumour cells between passages 4 and 13 for all subsequent analyses.

Orthotopic transplantation of murine breast tumour cells
MMTV-PyMT tg/+ SK4 WT or SK4 KO breast cancer cells were orthotopically transplanted into the fourth mammary gland of female FVB/N WT mice (12 weeks of age) (DeNardo et al., 2011;Kocaturk and Versteeg, 2015). Briefly, mice were anaesthetized by i.p. injection with 10 lLÁg À1 body weight of ketamine/xylazine solution (ketamine (10 mgÁmL À1 ) and xylazine (1 mgÁmL À1 ) in NaCl 0.9%). When sufficient deepness of anaesthesia was confirmed by the loss of pedal reflexes, the experimental mice were fixed on a heating plate (37°C). The area around the fourth mammary gland was shaved and cleaned with ethanol. A short cut of approx. 10 mm was made directly beneath the mammary gland to make it accessible. The gland was cleaned with PBS and fixed with a surgical clamp for the injection of 10 6 breast cancer cells in PBS (50 lL). Postoperative analgesia was performed with metamizole (200 mgÁkg À1 , s.c.). One week after the procedure, tumour onset was detected by palpation three times a week as described above.

SK4 modulators and chemicals
TRAM-34 (Wulff et al., 2000) was synthesized in-house and dissolved in ethanol (Sigma Aldrich, Taufkirchen, Germany) to a stock solution of 5 mM and then diluted to a final concentration of 0.1, 1, or 10 lM. Apoptosis initiator staurosporine (Cell Signaling/New England Biolabs, Frankfurt, Germany) was dissolved in DMSO to a stock concentration of 1 mM and then diluted to a final concentration of 1 lM in cell culture medium.
We determined the effect of TRAM-34 in unpaired experiments by comparing macroscopic on-cell currents recorded with TRAM-34 in the pipette with those obtained with vehicle alone. A supraeffectively high TRAM-34 (10 lM) concentration was chosen in order to guarantee a quantitative and instantaneous blockage of SK4 channels. The latter was required because of the high fragility of the giga ohm seal in our experiments. At this high concentration, TRAM-34 may inhibit other ion channels. The SK4 specificity of the observed TRAM-34-sensitive current fraction was therefore estimated by comparing the TRAM-34 effect between SK4-proficient and SK4-deficient cells.
Macroscopic on-cell currents were analysed by averaging the currents between 100 and 700 ms of each square pulse. Applied voltages refer to the cytoplasmic face of the membrane with respect to the extracellular space. Outward currents, defined as flow of positive charge from the cytoplasmic to the extracellular membrane face, are positive currents and depicted as upward deflections of the original current traces.

mRNA isolation and quantitative real-time polymerase chain reaction
Isolation of mRNA was performed under RNase-free conditions based on acid guanidinium thiocyanate/ phenol/chloroform extraction as described before (Leiss et al., 2011). Tumour samples were homogenized in 1 mL PeqGold RNA pure TM (PeqLab, VWR International, Erlangen, Germany) using an ultraturrax (Dispergieraggregat 1130, Kinematics). Breast tumour cells (5 9 10 6 ) at passages 4 to 13 were plated in 55-cm 2 petri dishes (Corning, Amsterdam, the Netherlands) and then cultivated at 37°C and 5% CO 2 for 24 h before G 1 arrest was induced by serum deprivation (72 h) using IMEM without phenol red (Life Technologies). Arrested cells were either left untreated (t 0 ) or were restimulated with culture medium containing 5% FBS in the absence or presence of TRAM-34 (10 lM) for 0.5, 1.5 or 24 h, respectively. mRNA was isolated from PBS-washed cells with 1 mL of PeqGold RNA pure TM directly applied to the monolayer. Both tissueand cell-derived mRNA samples were transferred to 1.5-mL test tubes and then centrifuged at 13 000 rpm for 10 min at 4°C to remove cell debris. Supernatants were transferred to fresh tubes and then processed according to the manufacturer's recommendations. mRNA concentrations were determined using a nanophotometer (Implen, M€ unchen, Germany). DNA contaminations were removed by a DNase digestion step by adding 5 lL of DNase (Roche) and 6 lL of DEPC-treated water per 50 lL sample. Digestion was allowed for 30 min at 37°C and followed by enzyme inactivation for 5 min at 80°C. For subsequent analyses, mRNA samples were adjusted to a final concentration of 0.1 lgÁlL À1 in DEPC-treated water. Tumour tissue-or tumour cell-derived mRNA (5 lL) was mixed with iScript TM (4 lL; Bio-Rad, M€ unchen, Germany) and DEPC-treated water (10 lL) in the absence or presence of reverse transcriptase (1 lL) for cDNA synthesis for 30 min at 42°C in a MasterCycler (Eppendorf, Hamburg, Germany). After heat inactivation of the reverse transcriptase (RT), cDNA was diluted with 180 lL DEPC-treated water to a final volume of 200 lL. Quantitative real-time polymerase chain reaction (qPCR) was performed using 2 lL of the diluted cDNA with SYBR Ò green Kit (Bio-Rad, M€ unchen, Germany) and specific primer pairs. SYBR Ò green fluorescence was detected with Opticon TM (MJ Research, St. Bruno). Cycle of time values (C t ) were determined with Opticon Monitor TM analysis software version 1.4 from triplicate. C t values of genes of interest were then subtracted from C t values of the simultaneously amplified housekeeping gene (b-actin) to obtain DC t values. 2 raised by the exponent of negative DC t value represents the relative mRNA level to the corresponding housekeeper b-actin. Primers were designed with Primer3web

Impedance and grid-based proliferation assays
Proliferation was assessed with l-dishes grid-500 (ibidi Ò , Planegg/Martinsried, Germany) and xCelligence system allowing for continuous measurement by impedance readout (Steinle et al., 2011). For the grid assay, cells were passaged and counted as described. Cell suspension was adjusted to 80 000 cells per grid and cells were allowed to attach to the bottom of the petri dish for 24 h at 37°C, 5% CO 2 before serum was withdrawn to induce G 1 arrest. Restimulation with culture medium (IMEM + 5% FBS and 1% Pen/Strep, Life Technologies) was performed after 72 h using either vehicle or compounds as indicated. Every 24 h, digital pictures of previously defined areas were taken for four consecutive days. Cell number at every time point was determined using ImageJ software (IMAGEJ 1.46).

Ki-67 proliferation analysis
Cells were seeded in 12-well chambers (Ibidi, Martinsried, Germany) in duplicate at densities of 80 000, 50 000 or 30 000 cells per well and grown for 24, 48 or 72 h, respectively. For assessing proliferation with immunofluorescence, cells were fixed with 70% ethanol at À20°C for 10 min, washed three times with PBS and incubated in PBS with 0.1% Triton X-100 for 15 min. Then, cells were washed three times with PBS and blocked with 10% normal donkey serum in PBS (Dianova, Hamburg, Germany) for 1 h. Cells were then incubated in a 1 : 1000 diluted anti-Ki-67 rabbit primary antibody solution (Cell Signaling Technologies/New England Biolabs, Frankfurt, Germany) in 1.5% normal donkey serum in PBS for 2 h. After washing three times with PBS, cells were incubated in a 1 : 800 diluted Alexa Fluor Ò 555 donkey anti-rabbit secondary antibody solution (Invitrogen, Carlsbad, CA) in 1.5% normal donkey serum in PBS for 1 h. Cells were washed another three times with PBS and coverslipped with DAPI-containing Vectashield antifade mounting medium (Vector laboratories, Lorrach, Germany). For each chamber, four microscopy sections were taken and Ki-67-positive cell count was related to the total number of cells in the section.

Cell cycle analysis by flow cytometry
For flow cytometry, a well-established protocol to stain DNA with propidium iodide (PI) was used (Pozarowski and Darzynkiewicz, 2004). 1.5 9 10 6 cells were seeded into a petri dish of 55-cm 2 area (triplicate for every condition) and cultivated for 24 h at 37°C, 5% CO 2 . To induce G 1 arrest, cells were washed twice with PBS and serum was withdrawn (Life Technologies) for 72 h at 37°C, 5% CO 2 . Restimulation was performed with culture medium for 24 h. After trypsinization and counting, cell number was adjusted with PBS to 1 9 10 6 cells per 0.5 mL. 0.5 mL of the cell suspension was mixed with 4.5 mL of 70% ethanol (v/v) for fixation and samples were stored at 4°C. During cell counting, we did not find evidence for cells stacked or grouped together. PI staining solution (10 lgÁmL À1 ) with RNase (100 lgÁmL À1 ) to avoid false RNA staining was added to the singularized cells at room temperature for 30 min. Samples that were incubated in staining solution without PI were included as negative controls. A FACSCanto flow cytometer (BD Biosciences) and the FACSDiva software package (BD Biosciences) were used for data acquisition. Disproportions between height, width and area as indicators of doublets, cell clumps and/or debris were excluded from these analyses by using respective scatter gates. Evaluation of DNA content was carried out with FlowJo 7.6.1 (TreeStar) using Watson curve-fitting algorithm to estimate the number of cells in G 1 , S and G 2 /M phases of the cell cycle. 2.12. Isolation of cytosolic and mitochondrial fractions from breast tumour cells Cells (5 9 10 6 ) were maintained at 37°C and 5% CO 2 for 24 h before treatment with vehicle, staurosporine (1 lM) (STS, Cell Signaling Technologies/New England Biolabs) or SK4 inhibitor TRAM-34 (10 lM), respectively. Ten cell culture dishes of 55-cm 2 area were necessary per condition. After several time points, cells were washed, trypsinized and washed in ice-cold PBS with 1 lM Na 3 VO 4 (in demineralized water, Sigma Aldrich) twice. After a centrifugation step at 1000 rpm, the cell pellet was resuspended in a fivefold amount of isolation buffer containing 1 tablet of cOmplete Mini, EDTA free (Roche). For homogenization, a combination of manual (30 s) and electric homogenization steps with an ultraturrax (30 s) was used as this guaranteed best separation of mitochondrial and cytosolic fraction. To isolate the purified mitochondria, a series of centrifugation steps was performed: Initial centrifugation for 10 min at 1000 g removed the nuclei and cell residues, whereas mitochondrial and cytosolic fractions in the supernatant were transferred to a fresh microtube and centrifuged again for 15 min at 14 000 g resulting in the mitochondrial pellet. Supernatant containing cytosolic fraction was carefully transferred into a fresh microtube for protein quantification, whereas the mitochondrial fraction was washed twice with 250 lL of isolation buffer and finally dissolved in 25 lL of isolation buffer. Protein determination was performed using Lowry's method with Total Protein Kit, Micro-Lowry, Peterson's Modification (Sigma Aldrich). Samples were used immediately or stored at À80°C until SDS/PAGE and western blot analyses were performed.

SDS/PAGE and western blot analyses
SDS/PAGE gels containing 17.5% acrylamide (Carl Roth) were used for separating 40 lg of the mitochondrial or cytosolic protein fractions. Before loading, protein samples were denatured in 4 9 Laemmli protein buffer for 10 min at 95°C. Protein ladder IV (Peq-Lab, VWR, Darmstadt, Germany) was used as protein standard allowing for an accurate determination of protein sizes. Electrophoresis was performed at 80-120 V for 90 min. For western blot, proteins of both fractions were transferred onto Membrane Immobilon Ò Millipore PVDF (Carl Roth) by using a semidry blotting system (Carl Roth) at 80 mA for 1 h followed by 150 mA for 15 min per membrane.

Statistics
Statistical analyses were performed with Microsoft Excel and GraphPad Prism. Graphs show mean AE SE for every condition. Two groups were compared by an unpaired t-test; if three or more groups were compared, a one-way ANOVA with following Bonferroni correction was performed. For Kaplan-Meier analyses, a log-rank test was used and P-values were calculated from v 2 (one degree of freedom).

Murine SK4 channels in breast tumours and tumour cells derived thereof
We investigated the PyMT-and cNeu-driven in vivo mouse models of breast tumour development for their SK4 expression status ( Fig. 1A and Fig. S1A). We found significantly higher levels of SK4 mRNA (34.5fold) in tumour biopsies of MMTV-PyMT tg/+ mice compared to biopsies derived from nonmalignant mammary glands of FVB/N wild-type animals. (Fig. S1B), suggesting that spontaneous formation and progression of breast tumours via oncogenic PyMT pathways may require functional SK4 channels. We also compared SK4 mRNA expression to that of phylogenetically related SK channels in breast tumour samples revealing the highest levels for SK4 in this model (Fig. 1C,D). In order to establish the endogenous role of SK4 in the MMTV-driven transgenic mouse models, we crossed the MMTV-PyMT tg/+ and MMTV-cNeu tg/+ animals, respectively, to a KCNN4 gene-targeted SK4 knockout mouse line (SK4 KO) on a FVB/N background. SK4-deficient MMTV-PyMT tg/+ and MMTV-cNeu tg/+ animals were born at the expected Mendelian ratios.
Analysis of primary cultures established from dissociation of multiple MMTV-PyMT tg/+ breast tumours confirmed SK4 expression in tumour-derived cells, although at relatively lower levels compared to the respective tumours in situ (Fig. 1A). As expected, corresponding samples from MMTV-PyMT tg/+ SK4 KO animals were negative for SK4 mRNA (Fig. 1A). Immunostaining of MMTV-PyMT tg/+ breast tumour tissue using SK4-specific antibodies further confirmed SK4 protein abundance in transformed mouse mammary epithelium (Fig. 1B). Antibody specificity for the SK4 epitopes was verified by parallel analysis of breast tumours derived from MMTV-PyMT tg/+ SK4 KO animals (Fig. 1B). To determine the functionality of SK4 channels in primary breast tumour cells, plasma membrane K + currents were recorded with the cellattached patch clamp technique (Fig. 1C,D). Currents were elicited in cells in the presence and absence of the SK4 inhibitor TRAM-34 (10 lM) by voltage square pulses ranging from À100 to +100 mV (Fig. 1C). TRAM-34-sensitive inward currents were observed in MMTV-PyMT tg/+ SK4 WT-derived breast tumour cells (Fig. 1D, upper tracings and Fig. 1E). Importantly, TRAM-34, albeit at very high concentration, did not affect inward currents of the SK4deficient primary breast tumour cells (Fig. 1D, lower tracings and Fig. 1F). In accordance with an SK4generated current fraction in MMTV-PyMT tg/+ tumour cells, inward currents and calculated inward conductance were significantly lower in cells lacking endogenous SK4 channels (Fig. 1G, red symbols/column) than in SK4-proficient cells (Fig. 1G, black symbols/column). The SK4-dependent current fraction (difference between red and black symbols in Fig. 1G) exhibited strong inward rectification as reported for canonical SK4 channels which are usually present at the plasma membrane of, for example, endothelial cells and smooth muscle cells (Grgic et al., 2005;Toyama et al., 2008).

Function of SK4 channels in growth and cell cycle control
We next examined whether SK4 activity plays a role in the growth factor-induced proliferation of PyMT breast cancer cells. Cell growth was assessed quantitatively by a mini-grid-based set-up ( Fig. 2A,B). Cell numbers were determined after 72 h of serum starvation (t 0 ), and following readdition of serum every 24 h thereafter for four consecutive days by taking micrographs of previously defined areas of the cell layer. Cell counts upon pharmacological inhibition of SK4 using 10 lM TRAM-34 showed a tendency to lower numbers at all time points analysed (Fig. 2B). Importantly, this difference reached statistical significance at 72 h for PyMT transgenic cells derived from five different SK4 WT tumours treated with 0.1 to 10 lM TRAM-34 ( Fig. 2B and Fig. S2A). Accordingly, cell cycle analyses using propidium iodide to quantify the amount of DNA suggest a G 1 /S arrest in PyMT transgenic breast tumour cells after SK4 inhibition (Fig. 2C,D), whereas apoptotic parameters such as cytochrome c release and caspase 3 activation were not affected by TRAM-34 (Fig. S1E). In line with the pharmacological blockade of SK4 channels in MMTV-PyMT tg/+ SK4 WT cells, lack of endogenous SK4 channels attenuated the growth of serum-restimulated breast tumour cells as well as the antigrowth effects of TRAM-34 (Fig. 2E,F and Fig. S2B). This essential role of SK4 as regulator of growth was further substantiated by a significantly lower number of Ki-67-positive nuclei in SK4 KO cells compared to WT cells (Fig. 2G and Fig. S3). Finally, a reduced response of the G 1 cell cycle phase markers c-fos and c-jun (Fig. 2H,I) to acute serum restimulation in PyMT transgenic SK4 KO cells confirmed the role of SK4 for the control of cell growth, that is replication and division.

Modulation of intracellular Ca 2+ levels by SK4 activity
Intracellular Ca 2+ signals are associated with different phases of cell cycle progression (Roderick and Cook, 2008;Whitfield et al., 1995). Elevated [Ca 2+ ] i and consecutive SK4 activation result in hyperpolarization, which, in turn, may favour the entry of Ca 2+ by increasing its electrochemical driving force across the plasma membrane or by keeping voltage-dependent Ca 2+ entry pathway in an active and/or highly conductive state. To assess whether SK4 controls the Ca 2+ dynamics of PyMT transgenic breast tumour cells, growth factor-induced Ca 2+ responses were determined in the presence and absence of TRAM-34. Changes in [Ca 2+ ] i were monitored after serum starvation in FURA-2-AM-loaded cells by changing the Ca 2+ -free loading buffer to a buffer containing Ca 2+ (1.8 mM) in the presence and absence of FBS. Ca 2+ superfusion alone resulted in a transient [Ca 2+ ] i rise showing an incomplete decay over 10 min (Fig. 3A). FBS augmented the Ca 2+ buffer-induced [Ca 2+ ] i amplitude of PyMT transgenic breast tumour cells (Fig. 3A,B), which is in line with findings from Lepple-Wienhues et al. (1996). Conversely, TRAM-34 suppressed the FBS-induced fraction of the [Ca 2+ ] i response to levels observed under FBS-free conditions, suggesting that changes in SK4 activity are relevant for growth factor-dependent control of [Ca 2+ ] i homeostasis (Fig. 3A,B). Accordingly, we observed a significantly lower FBS-induced [Ca 2+ ] i amplitude in SK4deficient PyMT transgenic breast tumour cells (Fig. 3C,D). In the next series of experiments, spontaneous changes of [Ca 2+ ] i were investigated at different time points in nonarrested PyMT breast cancer cells.

Impact of SK4 channels on survival time in breast cancer mouse models
Our in vitro data on the role of SK4 in cell growth, cell cycle control and Ca 2+ signalling collectively point to an antitumour effect of SK4 inhibition in vivo. Performing Kaplan-Meier analyses, we demonstrate that neither tumour-free survival nor overall survival of breast cancer-developing MMTV-PyMT tg/+ mice was significantly affected by the global lack of SK4 (Fig. 4A,B). Median survival rates of the SK4 KO mice tend to result in a better outcome, but this tendency did not reach statistical significance. As SK4 channels are known to be important determinants in the activation of T and B cells (Begenisich et al., 2004;Kang et al., 2014;Xu et al., 2014), decreased tumour formation in SK4 KO mice may be masked by the inability of the SK4-deficient immune system to recognize and/or control tumorigenesis in breast cancer. To prove this hypothesis, we took a different approach and established an orthotopical transplantation model. In vitro propagated breast cancer cells derived from tumour-bearing PyMT transgenic SK4 WT or SK4 KO mice were allowed to grow into the same origin site as the primary tumours by their implantation into the mammary fat pad of SK4-proficient FVB/N wild-type mice. Beginning 14 days postinjection, FVB/N wild-type hosts that had received SK4 WT cells developed palpable tumours, whereas animals implanted with SK4 KO cells exhibited a significant longer tumour-free survival time (Fig. 4C).
The MMTV-PyMT model usually develops multiple tumours in asynchronous manners that are spreading to different organs. As the impact of stromal and/or immune cell SK4 channels on tumour development might differ between highly malignantly and slowly progressing tumours, we analysed whether SK4 activity affects the survival rates in an alternative mouse model of spontaneously occurring tumours, that is the MMTV-cNeu tg/+ breast cancer model (Guy et al., 1992b). This model develops monofocal breast tumours with a later onset and a slower tumour progression than the MMTV-PyMT model. Notably, tumour-free survival and overall survival were significantly prolonged in MMTV-cNeu tg/+ mice lacking SK4 channels (Fig. 4D,E) as compared to the SK4 WT mice. Moreover, the time span between first tumour palpation and reaching of the abort criterion as a measure of tumour growth was significantly prolonged in MMTV-cNeu tg/+ SK4 KO mice (Fig. 4F). The important role of SK4 in this model was further confirmed ex vivo in MMTV-cNeu tg/+ -derived breast cancer cells by demonstrating the TRAM-34 sensitivity of their growth rate (Fig. S4A,B).

Discussion
Herein, we examined the role of endogenous SK4 channels for breast tumour cell growth, cell cycle progression, Ca 2+ signalling and survival of breast tumourbearing PyMT and cNeu transgenic mice. In accordance with previous reports (Haren et al., 2010;Huang and Jan, 2014), we found high levels of SK4 expression in primary tumours of both animal models and in tumour The mean maximal FL ratio of the 340/380-nm wavelength was determined relatively to the respective baseline FL ratio. Statistical analysis by one-way ANOVA followed by Bonferroni correction (**P < 0.01, ***P < 0.001 and # P < 0.05 or ###/ § § § P < 0.001 for the comparison to the SK4 WT group stimulated with Ca 2+ and FBS ). cells derived thereof. Electrophysiological evidence for TRAM-34-sensitive, inwardly rectifying currents in SK4 WT but not SK4 KO PyMT oncogene-driven breast tumour cells (Fig. 1) supports the notion that the SK4 channel is functional in breast tumour cells. Based on these initial experiments, tumour formation and development were evaluated in the absence of SK4 in vivo. Ablation of SK4, however, had a minor effect on mean survival rates but did not significantly prolong tumour-free survival nor overall survival in the MMTV-PyMT tg/+ model (Fig. 4), possibly due to an impaired immune and/or stromal cell response, which results in a tumour progression-promoting microenvironment (Jinushi et al., 2011;Loeffler et al., 2006;Wiseman and Werb, 2002) that counteracts the loss of the oncogenic SK4 functions in the tumour cells. The difference in tumour formation time between SK4-deficient and SK4-proficient MMTV-PyMT tumour cells transplanted orthotopically into WT mice strongly supports this hypothesis. Indeed, recent evidence from a melanoma-bearing mouse model suggests that the potassium efflux via SK4 and Kv1.3 is essential for tumour-specific T cells in order to impair malignant growth (Eil et al., 2016). The oncogenic role of SK4 was further confirmed by the MMTV-cNeu tg/+ breast tumour model where SK4 ablation prolonged the survival rates. Unfortunately, the role of SK4 in different cells and compartments of the tumour microenvironment, that is in noncancerous cells such as tumour-associated immune cells, endothelial cells and/or cancerassociated fibroblasts (CAFs) (Koshy et al., 2013;Xu et al., 2014), could not be tested by the present study due to the lack of a conditional SK4 KO mouse line allowing for the generation of cell-specific SK4 mutants. Our observation that both growth factor-evoked and spontaneous Ca 2+ transients in PyMT breast tumour cells do require SK4 activity is consistent with the TRAM-34 sensitivity of the cell cycle and abundance of c-fos and c-jun (Lai et al., 2011;Ouadid-Ahidouch et al., 2004;Wang et al., 2007b). Previous work identified hERG1, Kv1.1, BK and several other K + channels including SK4 to cause membrane hyperpolarization, which modifies the passive Ca 2+ entry in nonexcitable cancer cells (Ouadid-Ahidouch and Ahidouch, 2013). Although the molecular identity for the SK4-controlled Ca 2+ entry routes, that is the plasma Tumour-free survival of FVB/N WT mice that were orthotopically transplanted with MMTV-PyMT tg/+ SK4 WT (mean survival 20 AE 2 days) or SK4 KO (mean survival 45 AE 7 days) breast tumour cells, respectively (n = 5). (D-E) Tumour-free and overall survival of MMTV-cNeu tg/+ mice (FVB/N background) with a genetic ablation of SK4 (322 AE 21 days and 391 AE 19 days, respectively) compared to SK4 WT littermates (247 AE 9 days and 291 AE 9 days, respectively) (n = 8/34). (Statistical analysis for A-E was performed by the log-rank test followed by Pvalue calculation from v 2 with one degree of freedom (**P < 0.01, ***P < 0.001). (F) Tumour progression of MMTV-cNeu tg/+ SK4 WT (46 AE 4 days) and SK4 KO (68 AE 11 days) mice. Tumour progression was defined as the time between first palpation of a tumour until abort criterion (15-mm-diameter tumour size) was reached. Statistical analysis was performed by an unpaired two-tailed Student's t-test (*P < 0.05). SK4, calcium-activated potassium channel with intermediate conductance; WT, wild-type; KO, knockout; MMTV, mouse mammary tumour virus; PyMT, polyoma virus middle T-antigen; LR, log-rank value. membrane channels and store-operated mechanisms in the PyMT and Her2/cNeu breast tumour cells, is ill-defined, SK4 was previously colocalized or functionally linked to Ca 2+ release-activated Ca 2+ (CRAC)like channels or TRP (transient receptor potential) family such as TRPV6, TRMP7 and TRMP3 in T lymphocytes, prostate cancer cells (Kuras et al., 2012;Lallet-Daher et al., 2009) and MMTV-PyMT tg/+ breast tumour cells (data not shown).
Our data point to an oncogenic role of SK4 in mice that may have implications for human breast cancer. In humans, overexpression is considered the main mechanism of oncochannels to contribute to the malignancy of cancer (Huber, 2013;Pardo and Stuhmer, 2014), but activating mutations in the SK4-encoding KCNN4 gene may also contribute to the expression of abnormally regulated channels. Notably, KCNN4 promoter hypomethylation is associated with high levels of SK4 in aggressively growing non-small-cell lung carcinoma cell lines and advanced stages of lung cancer in patients (Bulk et al., 2015). With regard to breast cancer, similar epigenetic changes may account for progression; however, such mechanisms have not been elucidated as of yet. Rather, at the germline level, several intronic KCNN4 single nucleotide polymorphisms (SNPs) have been shown to alter women's breast cancer risk in general and particularly of ER-positive breast cancer. The observed effect size is small, but is of genome-wide significance (P < 10 À6 ) (Lo et al., 2016;Michailidou et al., 2015). How the genetic variants affect SK4 expression in normal breast tissue and breast tumour is currently unknown.
Together, our findings from animal models suggest that the role of SK4 channels should be further explored within the context of breast cancer progression and as a new strategy for future breast cancer management. It seems to be necessary to define the effect of such an approach on SK4-dependent pathways in nontumour cells and in the context of the mutated oncogenes. This is important, as SK4 and SK4-dependent pathways might hold the potential to establish novel breast cancer drug targets.

Supporting information
Additional Supporting Information may be found online in the supporting information tab for this article: Fig. S1. (A) Relative SK4 expression in malignant breast tumour tissue compared to healthy mammary gland tissue of female non-transgenic FVB/N wildtype mice. Statistical analysis was performed using an unpaired student's t-test (**P < 0.01). (B) SK4 expression in MMTV-cNeu tg/+ breast tumour cells. mRNA isolated from SK4 WT and SK4 KO thymus was used to control the specificity of the primer pairs used. (C-D) Relative expression levels of SK1-3 and SK4 mRNA in MMTV-PyMT tg/+ WT breast tumour samples. (E) Apoptosis induction by TRAM-34 was monitored by cytochrome c release of mitochondria into the cytosol. Cells were treated for 4, 24 and 48 h with TRAM-34 (10 lM) and staurosporine 1 lM as positive control. Mitochondrial and cytosolic fraction were separated and analysed by SDS-PAGE and Western blot. As expected, staurosporine induced cytochrome c release and cleavage of caspase-3, whereas no corresponding signals were observed in the TRAM-34 treated cells. HSP60 was used as mitochondrial marker and a-Tubulin as a marker for the cytosolic fraction. Sufficient separation of mitochondrial and cytosolic fractions in this setup was proven by Western blot analysis of mitochondrial samples (data not shown).    (A) were counted with IMAGEJ Software version 1.46 and cell numbers were normalized to t 0 for each time point and treatment (n = 10). Statistical analysis was performed by one-way ANOVA followed by Bonferroni correction (**P < 0.01).