Downregulation of the Ca(2+)-activated K(+) channel KC a3.1 by histone deacetylase inhibition in human breast cancer cells.

The intermediate-conductance Ca(2+)-activated K(+) channel KC a3.1 is involved in the promotion of tumor growth and metastasis, and is a potential therapeutic target and biomarker for cancer. Histone deacetylase inhibitors (HDACis) have considerable potential for cancer therapy, however, the effects of HDACis on ion channel expression have not yet been investigated in detail. The results of this study showed a significant decrease in KC a3.1 transcription by HDAC inhibition in the human breast cancer cell line YMB-1, which functionally expresses KCa3.1. A treatment with the clinically available, class I, II, and IV HDAC inhibitor, vorinostat significantly downregulated KC a3.1 transcription in a concentration-dependent manner, and the plasmalemmal expression of the KC a3.1 protein and its functional activity were correspondingly decreased. Pharmacological and siRNA-based HDAC inhibition both revealed the involvement of HDAC2 and HDAC3 in KC a3.1 transcription through the same mechanism. The downregulation of KC a3.1 in YMB-1 was not due to the upregulation of the repressor element-1 silencing transcription factor, REST and the insulin-like growth factor-binding protein 5, IGFBP5. The significant decrease in KC a3.1 transcription by HDAC inhibition was also observed in the KC a3.1-expressing human prostate cancer cell line, PC-3. These results suggest that vorinostat and the selective HDACis for HDAC2 and/or HDAC3 are effective drug candidates for KC a3.1-overexpressing cancers.

4‐(acetylamino)‐N‐[2‐amino‐5‐(2‐thienyl)phenyl]‐benzamide

breast cancer

5,6‐dichloro‐1‐ethyl‐1,3‐dihydro‐2H‐benzimidazol‐2‐one

bis‐(1,3‐dibutylbarbituric acid)trimethine oxonol

nucleoside diphosphate kinase

epidermal growth factor receptor

estrogen receptor

flow cytometry

histone deacetylase

HDAC inhibitor

epidermal growth factor receptor

head and neck squamous cell carcinoma

2,2‐bis(4‐fluorophenyl)‐2‐phenyl‐acetamide

Ca2+‐activated K+ (channel)

myotubularin‐related protein

nucleoside diphosphate kinase

phosphohistidine phosphatase

phosphoinositide‐3‐kinase, class 2, β polypeptide

progesterone receptor

prolactin

repressor element 1‐silencing transcription factor

small interfering RNA

sirtuin

N‐(2‐aminophenyl)‐4‐[1‐(2‐thiophen‐3‐ylethyl)‐1H‐[1], [2], [3]triazol‐4‐yl]benzamide

(S)‐S‐7‐(adamant‐1‐ylamino)‐6‐(tert‐butoxycarbonyl)‐7‐oxoheptyl‐2‐methylpropanethioate

1‐[(2‐chlorophenyl)diphenylmethyl]‐1H‐pyrazole

2‐(4‐Iodophenyl)‐3‐(4‐nitrophenyl)‐5‐(2,4‐disulfophenyl)‐2H‐tetrazolium, monosodium salt

Introduction

The Ca2+‐activated K+ (KCa) channel contributes to the feedback mechanism required to enhance Ca2+ signaling by controlling Ca2+ entry through Ca2+ release‐activated Ca2+ channels in nonexcitable cells such as cancer and immune cells (Cahalan and Chandy 2009; Bergmeier et al. 2013; Huang and Jan 2014; Pardo and Stühmer 2014; Comes et al. 2015; Feske et al. 2015). Based on the single‐channel conductance, KCa channels have been classified into large conductance (KCa1.1), intermediate conductance (KCa3.1, also known as IKCa, SK4, and KCNN4), and small conductance (KCa2.x) channels. The blockade of KCa3.1 inhibits proliferation and invasion during metastasis and/or promotes apoptosis in several cancer types including breast cancer (Wulff and Castle 2010; Ouadid‐Ahidouch and Ahidouch 2013). Breast cancer is the most commonly occurring malignancy among women, and KCa3.1 has been identified as one of the new loci associated with susceptibility to breast cancer by large‐scale genotyping analyses (Michailidou et al. 2013). Since KCa3.1 increased with the progression of breast cancer, it may be useful as a tumor marker to detect the malignancy grade during breast cancer progression (Haren et al. 2010). The luteotropic hormone, prolactin (PRL) has been associated with an increased risk of breast cancer, and PRL‐induced KCa3.1 activation promotes the proliferation of breast cancer cells (Faouzi et al. 2010). Of the four KCa3.1 regulators identified to date (class II phosphatidylinositol 3 kinase C2β (PI3K‐C2B), nucleoside diphosphate kinase‐B (NDPK‐B), protein histidine phosphatase 1 (PHPT1), and phosphatidylinositol 3‐phosphate phosphatase myotubularin‐related protein 6 (MTMR6), NDPK‐B is involved in various cellular functions in cancer cells and is a potential therapeutic target for breast cancer (Attwood and Wieland 2015).

Histone deacetylase inhibitors (HDACis), which exhibit a broad spectrum of epigenetic activities, are emerging as anticancer drugs (Bose et al. 2014). The suberoylanilide hydroxamic acid vorinostat received FDA approval for the treatment of cutaneous T‐cell lymphoma and is a pan‐HDACi that inhibits class I, II, and IV HDAC subtypes. HDACis are a novel class of agents in the treatment of solid cancers (Slingerland et al. 2014), and several clinical studies have been conducted on vorinostat as a combination therapy (Munster et al. 2011; Ramaswamy et al. 2012). HDACis reverse DNA methylation in cancer cells, and have clinical activity in the treatment of cancers (West and Johnstone 2014). We previously reported that the transcription of the Ca2+‐activated Cl− channel TMEM16A is downregulated by vorinostat and the pharmacological and small interfering RNA (siRNA)‐based blockade of HDAC3 (Matsuba et al. 2014); however, the regulation of other ion channels by HDAC inhibition remains to be elucidated. The destabilization of DNA methylation (hypermethylation or hypomethylation) in ion channels has been correlated with tumorigenesis and a poor prognosis (Ouadid‐Ahidouch et al. 2015). Hypomethylation of the KCa3.1 promoter has recently been associated with the upregulation of KCa3.1 in lung cancer cells (Bulk et al. 2015).

We herein demonstrated that the expression of KCa3.1 was downregulated in the human breast cancer cell line YMB‐1 by treatment with the pan‐HDAC inhibitor vorinostat. Pharmacological and siRNA‐based HDAC inhibition experiments indicated that KCa3.1 transcription is regulated by HDAC2 and HDAC3 through the same mechanism. Taken together, these results suggest that vorinostat and HDAC2/3‐selective inhibitors are effective against KCa3.1‐overexpressing cancers and other KCa3.1‐overexpressing disorders such as autoimmune and inflammatory diseases.

Materials and Methods

Cell culture and cell viability assay

The breast cancer cell lines MDA‐MB‐453, YMB‐1, MCF‐7, Hs578T‐Luc, and BT‐549 and the prostate cancer cell lines PC‐3 and LNCaP (clone FGC) were supplied by the RIKEN BioResource Center (RIKEN BRC) (Tsukuba, Japan) and Health Science Research Resources Bank (HSRRB) (Osaka, Japan). They were maintained at 37°C, in 5% CO2 with RPMI 1640, Dulbecco's modified Eagle's (DMEM), or Leibovitz's L‐15 medium (Wako, Osaka, Japan) containing 10% fetal bovine serum (Sigma, St. Louis, MO) and a penicillin (100 units/mL)‐streptomycin (0.1 mg/mL) mixture (Wako) (Matsuba et al. 2014). A cell viability assay using WST‐1 was performed as described in our previous study (Matsuba et al. 2014). Briefly, using a density of 4 × 105 cells/mL, cells were cultured in duplicate on 96‐well plates for 48 h (Fig. 1D) or 72 h (Fig. 1C, E). Absorbance was measured 2 h after the addition of WST‐1 reagent into each well using the microplate reader MULTSCAN FC (Thermo Fisher Scientific, Yokohama, Japan) at a test wavelength of 450 nm and reference wavelength of 620 nm. A pair of control and treated samples was prepared from different passage cells, and then the same protocol was repeated on another day. Cell viability of the vehicle (0.1% dimethyl sulfoxide)‐treated cells was arbitrarily expressed as 1.0.

Figure 1

Expression levels of K a3.1 transcripts in human breast tumors and breast cancer cell lines and effects of K a3.1 blockade on cell proliferation in YMB‐1 cells. (A and B) Real‐time PCR assay for K a3.1 in normal and tumor breast tissues (A) and human breast cancer cell lines (YMB‐1, MCF‐7, Hs578T, BT549 and MDA‐MB‐453). (C) Effects of the selective K a3.1 blocker, TRAM‐34 (1 and 10 μmol/L) on cell viability in YMB‐1 cells. (D) Effects of the K a3.1 activator, 10 μmol/L DCEBIO on cell viability in YMB‐1 cells. (E) Effect of 1 μmol/L TRAM‐34 on cell viability in YMB‐1 cells pretreated with PRL (200 nmol/L). In (A and B) the expression levels of K a3.1 transcripts are expressed as K a3.1/β‐actin (ACTB) ratio. In (C, D, and E) the cell viability in vehicle control (without treatment with PRL and TRAM‐34) was expressed as 1.0. Results are expressed as means ± SEM. The numbers used for the experiments are shown in parentheses. The significance of differences was evaluated using the Student's t‐test. **P < 0.01 versus normal tissues and vehicle control. # P < 0.05 versus PRL.

Transfecting cells with siRNA

Lipofectamine RNAiMAX reagent was used for all siRNA transfection procedures (Matsuba et al. 2014). Commercially available siRNA oligonucleotides against KCa3.1 (sc‐91691), HDAC2 (sc‐29345), HDAC3 (sc‐35538), and control siRNA (type A) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). The expression levels of the target transcripts were assessed 48 h after transfection of siRNAs using a real‐time PCR assay, and cell viability was measured by the WST‐1 assay.

RNA extraction, reverse transcription, and real‐time PCR

Total RNAs from normal human and tumor tissues were purchased from TaKaRa BIO (Osaka, Japan), BD Biosciences (San Jose, CA), and BioChain (Hayward, CA). Total RNAs from the primary tumor and corresponding metastatic tumor of the same donor (a 65‐year‐old female) were purchased from BioChain, who confirmed the expression of tumor‐specific genes and tumor metastatic genes, respectively, in these tumors. Total RNA extraction from cell lines and reverse transcription were performed as previously reported (Matsuba et al. 2014). cDNA products were amplified with gene‐specific PCR primers, designated using Primer Express™ software (Ver 3.0.1; Life Technologies, Carlsbad, CA). Quantitative, real‐time PCR was performed using SYBR Green chemistry on an ABI 7500 sequence detector system (ThermoFisher Scientific). The following gene‐specific PCR primers of human origin were used for real‐time PCR: KCa3.1 (GenBank accession number: NM_002250, 1475‐1595), amplicon = 121 bp; NDPK‐B (NM_002512, 288‐408), 121 bp; repressor element 1‐silencing transcription factor (REST) (NM_005612, 1415‐1545), 131 bp; insulin‐like growth factor 5 (IGFBP5) (NM_000599, 1088‐1208), 121 bp; histone deacetylase (HDAC) 2 (NM_001527, 298‐405), 108 bp; HDAC3 (NM_003883, 699‐819), 121 bp; β‐actin (ACTB) (NM_001101, 411‐511), 101 bp. Unknown quantities relative to the standard curve for a particular set of primers were calculated as previously reported (Matsuba et al. 2014), yielding the transcriptional quantitation of gene products relative to the endogenous standard, ACTB.

Overexpression of IGFBP5 in YMB‐1 cells

The cDNA products from YMB‐1 cells were amplified with KOD FX DNA polymerase (TOYOBO, Osaka, Japan) and PCR primers specific for human IGFBP5: IGFBP5 (NM_000599, 717‐1629), 913 bp. The amplification profile was follows: a 2 min predenaturation step at 94°C, followed by 40 cycles of a 10 sec denaturation step at 98°C and a 1 min primer extension step at 68°C for 40 cycles. Obtained amplified products of full‐length IGFBP5 were ligated into pcDNA3.1(+)/Neor (Invitrogen) (pcDNA‐IGFBP5). The nucleotide sequences of plasmid constructs were determined by Eurofins Genomics (Tokyo, Japan). YMB‐1 cells were transiently transfected with pcDNA‐IGFBP5 using Lipofectamine 2000 (Invitrogen). Forty‐eight hours after transfection, the expression levels of IGFBP5 and KCa3.1 transcripts were determined by real‐time PCR assay.

Protein expression analyses by Western blotting and immunocytochemistry

Proteins lysates were prepared from YMB‐1 cells for Western blot analysis, as previously reported (Matsuba et al. 2014). Protein expression levels were determined 48 h after the compound treatment. Equal amounts of protein (20 μg/lane) were subjected to SDS‐PAGE (10%). Blots were incubated with anti‐KCa3.1 (Alomone Labs, Jerusalem, Israel) and anti‐ACTB (Medical & Biological Laboratories (MBL), Nagoya, Japan) antibodies, then incubated with anti‐rabbit and anti‐mouse horseradish peroxidase‐conjugated IgG (Merck Millipore, Darmstadt, Germany), respectively. An enhanced chemiluminescence detection system (GE Healthcare Japan, Tokyo, Japan) was used to detect the bound antibody. The resulting images were analyzed by a VersaDoc5000MP device (Bio‐Rad Laboratories, Hercules, CA). In the immunocytochemical examination, living YMB‐1 cells were harvested using a sterile cell scraper, and stained by an ATTO 488‐conjugated anti‐KCa3.1 (extracellular) antibody for live cell imaging (Alomone Labs). Stained cells were subjected to an analysis on a FACSCalibur flow cytometer using CellQuest software (BD Biosciences).

Measurements of KCa3.1 K+ currents by whole‐cell patch‐clamp recordings

Whole‐cell patch‐clamp experiments were performed as reported previously (Matsuba et al. 2014). A pipette solution was used, with the following composition (in mmol/L): 110 K‐aspartate, 30 KCl, 1 MgCl2, 10 HEPES, 10 EGTA, and 2 Na2ATP (pH 7.2 by KOH). Ca2+ concentrations (pCa 6.5) were determined using the WINMAXC program (Stanford University, Stanford, CA). Membrane currents were digitized using an analog‐digital converter (PCI6229, National Instruments, Tokyo, Japan), and data acquisition and analysis were performed using WinWCP4.65, developed by Dr. John Dempster (University of Strathclyde, UK). The liquid junction potential between the pipette and bath solutions (−10 mV) was corrected. Cells were held at −80 mV and currents were evoked by pipette solutions containing Ca2+ buffered at 300 nmol/L. Outward K+ currents were determined by stepping the cell from −80 to +40 mV for 100 msec, and then to −40 mV to measure the tail current (see Fig. S3A). A standard HEPES‐buffered bathing solution was used, with the following composition (in mmol/L): 137 NaCl, 5.9 KCl, 2.2 CaCl2, 1.2 MgCl2, 14 glucose, and 10 HEPES (pH 7.4 by NaOH). All experiments were performed at 25 ± 1°C.

Measurement of the intracellular Ca2+ concentrations by the fluorescent Ca2+ indicator dye

Cells were seeded onto 35 mm glass bottom dishes and cultured at 37°C in 5% CO2 humidified incubator. Intracellular Ca2+ concentration was measured using the fluorescent Ca2+ indicator dye, Fura 2‐AM. Cells were incubated with 10 μmol/L Fura 2‐AM in normal HEPES solution for 30 min at room temperature. Cells loaded with Fura 2‐AM were alternatively illuminated at wavelengths of 340 and 380 nm, and fluorescence images were recorded on the ORCA‐Flash2.8 digital camera (Hamamatsu Photonics, Hamamatsu, Japan). Data collection and analyses were performed using an HCImage system (Hamamatsu Photonics). Images were measured every 5 sec. The fluorescent intensity of Fura 2 was expressed as measured 340/380 nm fluorescence ratios [ratio (340/380)] after background subtraction. After treatment with a sarco/endoplasmic reticulum Ca2+‐ATPase (SERCA) inhibitor, thapsigargin (1 μmol/L) under 0 mmol/L Ca2+ in a bathing medium, application of 2.2 mmol/L Ca2+ caused store‐operated Ca2+ entry (SOCE) (see Fig. S3B). Ten minutes after 1st application of 2.2 mmol/L Ca2+, second one was performed in the presence of 1 mmol/L TRAM‐34. The value of area under the curve (AUC) of the ratio (340/380) was obtained by the integration on the computer. The AUC value in the second application was normalized by that of the first one (relative AUC of [Ca2+]i).

Chemicals

The sources of pharmacological agents were as follows: TRAM‐34 (Santa Cruz Biotechnology), ICA‐17043 (Cheminstock Ltd., Shanghai, China), DiBAC4(3) (Invitrogen), DCEBIO (TOCRIS Bioscience, Ellisville, MO), paclitaxel (Sigma‐Aldrich, St. Louis, MO, USA), PRL (Wako), WST‐1 (Dojindo, Kumamoto, Japan), and Fura2‐AM (Dojindo). HDAC inhibitors (vorinostat, AATB, T247, NCT‐14b and NCO‐04) were supplied by Professor Suzuki (KPUM). Others were obtained from Sigma‐Aldrich or Wako Pure Chemical Industries (Tokyo, Japan).

Statistical analysis

The significance of differences among two and multiple groups was evaluated using the Student's t‐test and Tukey's test after the F‐test or ANOVA, respectively. Significance at P < 0.05 and P < 0.01 is indicated in the figures. Data are presented as the means ± SEM.

Results

High level expression of the KCa3.1 K+ channel in breast tumor tissues and the breast cancer cell line, YMB‐1

Figure 1A shows that KCa3.1 transcripts were more strongly expressed in human breast tumor tissues than in normal breast tissues. Of the five human breast cancer cell lines examined, the high level expression of KCa3.1 transcripts was detected in YMB‐1 cells (Fig. 1B). YMB‐1 and MDA‐MB‐453 cells were human epidermal growth factor receptor 2 (HER2)‐positive; however, no correlation was found between KCa3.1 and HER2. We next examined the effects of the selective KCa3.1 blocker, TRAM‐34 and KCa2.x and KCa3.1 activator, DCEBIO on the viability of YMB‐1 cells. Cell viability significantly decreased in YMB‐1 cells treated for 72 h with 10 μmol/L TRAM‐34 (Fig. 1C), and significantly increased in those treated with 1 μmol/L DCEBIO (Fig. 1D). Faouzi et al. (2010) have shown that KCa3.1 activity is significantly enhanced by a treatment with PRL and PRL‐induced proliferation is inhibited by the pharmacological blockade of KCa3.1. In this study, PRL receptor (PRLR) transcripts were strongly expressed in YMB‐1 cells (Fig. S1A). As shown in Figure 1E, the enhancement induced in cell viability by the treatment with 200 nmol/L PRL for 72 h was significantly decreased by that with TRAM‐34 (10 μmol/L). Of 51 K+ channel subtypes (28 voltage‐gated Kv subtypes, 15 two‐pore domain K2P subtypes, 7 Ca2+‐activated KCa subtypes, and inward‐rectifier Kir2.1), YMB‐1 cells predominantly expressed KCa3.1 transcripts (Fig. S2A–F), and no significant decrease in cell viability were detected in YMB‐1 cells treated for 48 h with 10 mmol/L tetraethylammonium (TEA) (a Kv inhibitor) and 1 mmol/L Ba2+ (a K2P inhibitor) (Fig. S2G).

Downregulation of KCa3.1 gene and protein expression in YMB‐1 cells by the treatment with vorinostat

As shown in our previous study (Matsuba et al. 2014), vorinostat (1 and 10 μmol/L) significantly inhibits the viability of YMB‐1 cells in a concentration‐dependent manner. Real‐time PCR examinations revealed that the treatment of YMB‐1 cells with vorinostat for 24 h significantly downregulated the expression levels of the KCa3.1 transcripts in a concentration‐dependent manner (n = 4, P < 0.05 and 0.01 vs. vehicle control) (Fig. 2A). Similarly, the expression levels of the KCa3.1 proteins in YMB‐1 cells were suppressed by the treatment with vorinostat for 48 h in a concentration‐dependent manner (Fig. 2B). Similar results were obtained from three independent experiments. The optical density of the KCa3.1 protein band signal relative to that of the ACTB one was calculated by ImageJ software (Ver. 1.42, NIH, Bethesda, MD), and protein expression levels in the vehicle control were then expressed as 1.00. Relative optical densities in vorinostat‐treated groups (1 and 10 μmol/L) were 0.80 ± 0.04 (n = 3, P < 0.05) and 0.59 ± 0.05 (n = 3, P < 0.01). We also measured the protein expression levels of KCa3.1 in the plasma membrane with an ATTO 488‐conjugated anti‐KCa3.1 (extracellular) antibody (Alomone Labs), which has the ability to distinguish the extracellular region of KCa3.1 by a FCM analysis. A decrease in the percentage of KCa3.1‐positive cells was observed by the treatment with 1 μmol/L vorinostat for 48 h (Fig. 2C and D). In MDA‐MB‐453 cells weakly expressing KCa3.1, the percentage of ATTO 488‐positive cells was less than 3% (not shown). The chemotherapy drugs paclitaxel (100 nmol/L) and bafilomycin‐A (10 nmol/L, a vacuolar‐ATPase inhibitor) markedly suppressed cell proliferation in these cell lines, whereas no suppressive effects were observed on KCa3.1 transcription (Fig. S1B). Additionally, the inhibition of SIRT1 and 2 (class III HDAC subtypes) by NCO‐04 (50 μmol/L) (Suzuki et al., 2012) did not significantly downregulate the expression of KCa3.1 in YMB‐1 cells (Fig. S1C and D).

Figure 2

Effects of vorinostat on K a3.1 gene and protein expression levels in YMB‐1 cells. (A) Effects of vorinostat (1 and 10 μmol/L) on the expression levels of K a3.1 transcripts. Expression levels are expressed as K a3.1/ACTB ratio. (B–D) Effects of vorinostat on the expression levels of K a3.1 proteins in YMB‐1 cells in Western blotting (B) and flow cytometric analyses (C and D). In (B) proteins from the plasma membrane fraction of YMB‐1 cells were probed by immunoblotting with anti‐K a3.1 (upper panel) and anti‐ACTB (lower panel) antibodies on duplicate filters. Arrowheads indicate the migrating positions of K a3.1 and ACTB. Molecular mass standards are shown in kDa on the right side of the panels. In (C) living YMB‐1 cells were stained with an ATTO 488‐fused anti‐K a3.1 (extracellular) antibody, and ATTO 488‐positive cells were measured using flow cytometry. Data were expressed as the percentage of K a3.1‐positive cells (D). Results are expressed as means ± SEM. The numbers used for the experiments are shown in parentheses. The significance of differences among two and multiple groups was evaluated using the Student's t‐test and Tukey's test. *,**P < 0.05, 0.01 versus vehicle control.

Decrease in KCa3.1 activities in YMB‐1 cells by the treatment with vorinostat

Outward K+ currents were recorded by the application of depolarizing steps (from −80 to +40 mV for 100 msec) using the whole‐cell configuration. Tail currents were then measured at −40 mV after depolarization to +40 mV for 100 msec. KCa3.1 K+ currents were estimated as 1 μmol/L TRAM‐34‐sentitive components. Both peak and tail KCa3.1 K+ currents sensitive to TRAM‐34 were significantly smaller in 10 μmol/L vorinostat‐treated than in vehicle‐treated YMB‐1 cells (Fig. 3). Both peak (Fig. 3A) and tail (Fig. 3B) current amplitudes were over 80% decreased in vorinostat‐treated YMB‐1 cells. Next, we examined the effects of TRAM‐34 on the SOCE in vehicle‐ and 10 μmol/L vorinostat‐treated YMB‐1 cells. In vehicle‐treated YMB‐1 cells, significant decrease in relative AUC of [Ca2+]i due to membrane depolarization elicited by KCa3.1 inhibition was observed (Fig. 3C), whereas it was disappeared in vorinostat‐treated YMB‐1 cells (Fig. 3D). TRAM‐34‐induced decrease in the relative peak amplitude of transient Ca2+ rise (peak 2/peak 1, see Fig. S3A) was also disappeared by the treatment with 10 μmol/L vorinostat (not shown).

Figure 3

Inhibition of K a3.1 activity in YMB‐1 cells by the treatment with vorinostat. Pharmacological experiments using 1 μmol/L TRAM‐34 were performed using whole‐cell patch‐clamp recording and Ca2+ imaging systems. (A and B) TRAM‐34 (1 μmol/L)‐sensitive outward peak (A) and tail (B) currents at +40 mV in vehicle‐ and 10 μmol/L vorinostat‐treated YMB‐1 cells for 30–48 h. (C and D) TRAM‐34 (1 μmol/L)‐induced changes in store‐operated Ca2+ entry in vehicle‐ (control) (C) and 10 μmol/L vorinostat‐treated (D) YMB‐1 cells for 30–48 h. Relative AUC of [Ca2+]i was calculated as described in “Materials and Methods.” Results are expressed as means ± SEM. Numbers used for the experiments are shown in parentheses. The significance of differences was evaluated using the Student's t‐test. *, **P < 0.05, 0.01 versus vehicle control.

Identification of HDAC subtypes involved in the downregulation of KCa3.1 in YMB‐1 cells by pharmacological and siRNA‐based blockade of HDACs

The HDAC1, HDAC2, HDAC3, and HDAC6 subtypes are predominantly expressed in YMB‐1 cells and human tumor breast tissues (Müller et al. 2013; Matsuba et al. 2014; Seo et al. 2014). In order to identify the HDAC subtypes involved in the downregulation of KCa3.1 in YMB‐1 cells, the following selective HDAC inhibitors were used: 30 nmol/L AATB, 300 nmol/L AATB, 1 μmol/L T247, and 1 μmol/L NCT‐14b for the selective inhibition of HDAC1, HDAC1/2, HDAC3, and HDAC6, respectively (Itoh et al. 2007; Methot et al. 2008; Suzuki et al. 2013). As shown in Figure 4A, the transcription of KCa3.1 was significantly decreased by the treatment with 300 nmol/L AATB or 1 μmol/L T247 (for 48 h). The treatment with 300 nmol/L AATB plus 1 μmol/L T247 did not exhibit additive or synergistic inhibitory effects on KCa3.1 transcription (Fig. 4A). The treatments with 30 nmol/L AATB and 1 μmol/L NCT‐14b did not elicit any significant changes in KCa3.1 transcription. Similarly, the cell surface protein expression of KCa3.1 was significantly suppressed by the treatment with 300 nmol/L AATB or 1 μmol/L T247 (Fig. 4B). In consistent with these results, decrease in relative AUC of [Ca2+]i elicited by KCa3.1 inhibition was prevented in AATB (300 nmol/L)‐ and T247 (1 μmol/L)‐treated YMB‐1 cells (Fig. S3C and D). We further examined the siRNA‐based inhibition of HDAC subtypes. We first determined transfection conditions with more than 90% cell viability efficacy using Lipofectamine RNAiMAX. In YMB‐1 cells, transfection efficacy was 50–60% under this optimum condition. HDAC2 or HDAC3 were 50% inhibited by HDAC2 or HDAC3 siRNAs (Fig. 5A and B). Similar to the results shown in Figure 4, KCa3.1 transcripts were downregulated by HDAC2 or HDAC3 siRNA (Fig. 5C). The transcription of KCa3.1 was suppressed to a similar level by the transfection of KCa3.1 siRNA (Fig. 5C). These results suggest that the inhibition of HDAC2 or HDAC3 may cause the downregulation of KCa3.1 in YMB‐1 cells through the same mechanism. The expression levels of KCa3.1 transcripts were higher in the androgen‐independent prostate cancer cell line PC‐3 than in the androgen‐dependent prostate cancer cell line LNCaP: 0.047 ± 0.002 and less than 0.0002 in arbitrary units, respectively. The downregulation of KCa3.1 by the treatment with vorinostat (Fig. 6A) and HDAC2 or HDAC3 inhibition (Fig. 6B) was also detected in PC‐3 cells.

Figure 4

Effects of selective HDAC inhibitors on K a3.1 gene and protein expression levels in YMB‐1 cells. (A) Effects of AATB (30 nmol/L, selective for HDAC1; 300 nmol/L, selective for HDAC1 and 2), T247 (1 μmol/L, selective for HDAC3), NCT‐14b (1 μmol/L, selective for HDAC6) and AATB (300 nmol/L) plus T247 (1 μmol/L) on the expression levels of K a3.1 transcripts in YMB‐1 cells. Expression levels were expressed as K a3.1/ACTB ratio. (B) Effects of AATB (300 nmol/L), T247 (1 μmol/L), and AATB (300 nmol/L) plus T247 (1 μmol/L) on the cell surface expression of K a3.1 proteins by flow cytometric analysis. Living YMB‐1 cells were stained with an ATTO 488‐fused anti‐K a3.1 (extracellular) antibody and ATTO 488‐positive cells were measured by flow cytometry. Data were expressed as the percentage of K a3.1‐positive cells. Results are expressed as means ± SEM. The numbers used for the experiments are shown in parentheses. The significance of differences was evaluated using the Tukey's test. **P < 0.01 versus vehicle control.

Figure 5

Effects of HDAC2 and HADC3 siRNAs on the expression levels of K a3.1 transcripts in YMB‐1 cells. (A and B) Effects of HDAC2 and/or HDAC3 siRNAs on the expression levels of HDAC2 (A) and HDAC3 (B) transcripts. (C) Effects of HDAC2 and/or HDAC3 siRNAs on the expression levels of K a3.1 transcripts. Control siRNA‐A (Santa Cruz Biotechnology) was used as control siRNA (si‐ctrl). Expression levels were expressed as K a3.1/ACTB ratio. Results are expressed as means ± SEM. The numbers used for the experiments are shown in parentheses. The significance of differences was evaluated using the Tukey's test. **P < 0.01 versus control siRNA (si‐ctrl).

Figure 6

Effects of HDACis on the expression levels of K a3.1 transcripts in the prostate cancer cell line, PC‐3. (A) Effects of 1 and 10 μmol/L vorinostat on the expression levels of K a3.1 transcripts. (B) Effects of AATB (300 nmol/L) and T247 (1 μmol/L) on the expression levels of K a3.1 transcripts in PC‐3 cells. Expression levels were expressed as K a3.1/ACTB ratio. The results are expressed as means ± SEM. The numbers used for the experiments are shown in parentheses. The significance of differences was evaluated using the Tukey's test. *,**P < 0.05, 0.01 versus vehicle control.

Effects of HDACis on the expression of KCa3.1 regulators in YMB‐1 cells

Nucleotide diphosphate kinase (NDPK)‐B is strongly expressed in many cancer types and, thus, is a potent therapeutic target for anticancer drugs (Attwood and Wieland 2015). In T cells, NDPK‐B is a well‐known positive regulator of KCa3.1 (Di et al. 2010b; Ohya et al. 2014). As shown in Figure 7A, NDPK‐B transcripts were strongly expressed in human breast tumor tissues and YMB‐1 cells. The pharmacological and siRNA‐based blockade of HDAC2 and HDAC3 did not significantly affect the expression levels of NDPK‐B transcripts (Fig. 7B–D). Furthermore, the other positive and negative regulators of KCa3.1, and the transcriptional expression levels of PI3K‐C2B, PHPT1, and MTMR6 (Srivastava et al. 2006, 2008, 2009) were not affected by the treatment with 10 μmol/L vorinostat (Fig. S4).

Figure 7

Effects of HDAC inhibitors on the expression levels of NDPK‐B transcripts in YMB‐1 cells. (A) Expression levels of NDPK‐B transcripts in normal and tumor breast tissues. (B–D) Effects of vorinostat (1 and 10 μmol/L) (B), selective HDAC inhibitors (C), and HDAC siRNAs (D) on the expression levels of NDPK‐B transcripts in YMB‐1 cells. Expression levels were expressed as NDPK‐B/ACTB ratio. Results are expressed as means ± SEM. The numbers used for the experiments are shown in parentheses. The significance of differences among two and multiple groups was evaluated using the Student's t‐test and Tukey's test. **P < 0.01 versus normal tissues.

Effects of HDAC inhibition on the expression of REST and IGFBP5 in YMB‐1 cells

Repressor element 1‐silencing transcription factor, REST (also known as neuron‐restrictive silencer factor, NRSF) is known as a transcriptional repressor of KCa3.1 (Cheong et al. 2005; Ohya et al. 2011a), and the transcription or protein degradation of REST is regulated by the treatment with HDACis (Taylor et al. 2012). Therefore, we determined whether the downregulation of KCa3.1 by HDAC inhibition is due to an increase in REST expression in YMB‐1 cells. As shown in Figure 8A and B, no significant changes in the transcriptional and protein expression levels of REST were found in vorinostat (1 and 10 μmol/L)‐, AATB (300 nmol/L), and T247 (1 μmol/L)‐treated YMB‐1 cells. When REST transcription was approximately 50% inhibited by the transfection of REST siRNA in YMB‐1 cells (not shown), no significant changes in the expression levels of KCa3.1 transcripts were found (Fig. 8C).

Figure 8

Effects of vorinostat on the expression levels of REST and IGFBP5 transcripts in YMB‐1 cells, and effects of REST siRNA and IGFBP5 overexpression on the expression levels of K a3.1 transcripts in YMB‐1 cells. (A) Effects of vorinostat (1 and 10 μmol/L) (treatment for 24 h) on the expression levels of REST transcripts in YMB‐1 cells. (B) Effects of vorinostat (1 μmol/L), AATB (300 nmol/L), and T247 (1 μmol/L) (treatment for 48 h) on the expression levels of REST proteins in YMB‐1 cells. (C) Effects of REST siRNA on the expression levels of K a3.1 transcripts in YMB‐1 cells. (D) Comparison of the expression levels of IGFBP5 transcripts between human “primary” and “metastatic” breast tumors. (E) Effects of vorinostat (1 and 10 μmol/L) (treatment for 24 h) on the expression levels of IGFBP5 transcripts in YMB‐1 cells. (F) Effects of overexpression of IGFBP5 on the expression levels of K a3.1 transcripts in YMB‐1 cells: mock, transfected with pcDNA3.1(+) alone; IGFBP5, transfected with pcDNA‐IGFBP5. Expression levels were expressed as the ratio to ACTB. Results are expressed as means ± SEM. The numbers used for the experiments are shown in parentheses. The significance of differences among two and multiple groups was evaluated using the Student's t‐test and Tukey's test. *, **P < 0.05, 0.01 versus control siRNA (si‐ctrl) (C), primary breast tumor (D), and vehicle control (E).

It has been reported that insulin‐like growth factor‐binding protein 5 (IGFBP5) plays an important role in breast cancer progression and metastasis (Wang et al. 2008; Ghoussaini et al. 2014). Recently, Akkiprik et al. (2015) have shown KCa3.1 gene expression is negatively correlated with IGFBP5 one in breast cancer tissues. Using total RNAs from the primary breast tumor and corresponding metastatic breast tumor of the same donor, we found that the expression levels of IGFPB5 transcripts were significantly higher in metastatic tissue than in primary tumor one (Fig. 8D). As shown in Figure 8E, the treatment with 1 and 10 μmol/L vorinostat for 24 h caused significant increase in IGFBP5 transcripts. Therefore, we determined whether the downregulation of KCa3.1 by HDAC inhibition is related to the expression levels of IGFBP5 in YMB‐1 cells. As shown in Figure 8F, no significant changes in the expression levels of KCa3.1 transcripts were found in IGFBP5‐overexpressing YMB‐1 cells (Fig. 8C) when IGFBP5 transcription was over 10‐fold increased by the transfection of pcDNA‐IGFBP5 (not shown). These results indicated that KCa3.1 transcription was not regulated by REST and IGFBP5 in YMB‐1 cells.

Expression of HDAC2, HDAC3, and KCa3.1 in human primary and metastatic breast tumor tissues

In various cancer types including breast cancer, the high level expression of KCa3.1 is associated with a high metastatic risk (Rabjerg et al. 2015). Since KCa3.1 levels increase with the progression of breast cancer, KCa3.1 may be useful for detecting malignancy grade during breast cancer progression (Haren et al. 2010). Using total RNAs from the primary breast tumor and corresponding metastatic breast tumor of the same donor, we found that the expression levels of KCa3.1 transcripts were lower in metastatic tissues than in primary tumor tissues (Fig. 9A). The expression levels of HDAC2 (Fig. 9B) and HDAC3 (Fig. 9C) transcripts were also significantly lower in metastatic tissues. These results suggest that the upregulation of KCa3.1 through HDAC2 and HDAC3 enhancements may be responsible for the pathogenesis of aggressive rather than metastatic breast cancer.

Figure 9

Comparison of the expression levels of K a3.1, HDAC2, and HDAC3 transcripts between human primary and metastatic breast tumors. (A–D) Real‐time PCR assay for K a3.1 (A), HDAC2 (B), and HDAC3 (C) in human primary and metastatic breast tumors. Expression levels were expressed as the ratio to ACTB. Results are expressed as means ± SEM. The numbers used for the experiments are shown in parentheses. The significance of differences was evaluated using the Student's t‐test. **P < 0.01 versus primary breast tumors.

Discussion

The intermediate‐conductance Ca2+‐activated K+ channel, KCa3.1 has oncogenic potential and is a therapeutic target for KCa3.1‐positive cancers such as renal (Rabjerg et al. 2015), lung (Bulk et al. 2015), hepatic, prostate cancers (Ohya et al. 2009), and glioblastoma (D'Alessandro et al. 2013). In breast cancer, KCa3.1 is associated with tumor progression and metastasis (Ouadid‐Ahidouch et al. 2004; Faouzi et al. 2010; Haren et al. 2010). HDACis have potential as agents for the treatment of solid tumors including breast cancer (Slingerland et al. 2014), however, their effects on ion channel expression have not yet been elucidated in detailed. The main results of this study are as follows. (1) The clinically available pan‐HDAC inhibitor, vorinostat significantly downregulated KCa3.1 transcription in the human breast cancer cell line YMB‐1 (see Fig. 2). (2) The class I HDACs, HDAC2, and HDAC3 were responsible for the overexpression of KCa3.1 in YMB‐1 cells (see Figs. 4 and 5). (3) HDAC2 and HDAC3 are also involved in the epigenetic regulation of KCa3.1 in the KCa3.1‐expressing human prostate cancer cell line, PC‐3 (see Fig. 6). We previously identified the dominant‐negative splicing isoform of KCa3.1, KCa3.1B from mammalian lymphoid cells (Ohya et al. 2011b). KCa3.1B was expressed less abundantly in vehicle‐ and HDACis‐treated YMB‐1 cells (not shown).

HDAC2 and HDAC3 are strongly expressed in aggressive breast tumor subtypes (Müller et al. 2013). Their expression levels are also higher in HER2‐positive breast tumor subtypes: 44% and 52% for HDAC2 and HDAC3, respectively. Consistent with this finding, HER2‐positive YMB‐1 and MDA‐MB‐453 cells strongly expressed HDAC1, 2, and 3 (Fig. S5A and B) (Matsuba et al. 2014). However, the expression levels of KCa3.1 transcripts markedly differed between these cell types (Fig. 1B). No correlation was detected between the expression levels of HER2 and KCa3.1 in human primary breast cancer tissues (Fig. S5C). Furthermore, HDAC1, HDAC2, and HDAC3 are highly expressed in human prostate cancer (Weichert et al. 2008). The expression levels of HDAC2 were previously reported to be high in patients with Gleason score 7 (Weichert et al. 2008). HDAC3 is also expected to become a potent therapeutic target for prostate cancer due to the consistently high rate of HDAC3 positivity (Weichert et al. 2008). We previously demonstrated that HDAC2 and HDAC3 were both predominatly expressed in KCa3.1‐expressing PC‐3 cells (Matsuba et al. 2014), while the expression level of KCa3.1 was found to be prominent in biopsy samples from patients with midgrade malignancy (Gleason score 5–6) (Ohya et al. 2009). These findings suggest that selective HDAC2 and 3 inhibitors have potential as therapeutic agents for KCa3.1‐positive cancer.

HDACis upregulate tumor suppressor genes via histone hyperacetylation. It has been reported that the expression levels of REST and IGFBP5 is negatively correlated with that of KCa3.1 in proliferative cells including breast cancer calls (Cheong et al. 2005; Ohya et al. 2011a; Akkiprik et al. 2015). The transcription or protein degradation of REST is regulated by the treatment with HDACis in medulloblastoma (Taylor et al. 2012), however, this study showed no significant changes in the expression level of REST by the treatment with HDACis in YMB‐1 cells (Fig. 8A and B). In addition, no significant changes in the expression level of KCa3.1 transcripts were detected by the siRNA‐based blockade of REST expression in YMB‐1 cells (Fig. 8C). These suggest that the downregulation of KCa3.1 by HDAC inhibition is not due to the upregulation of REST in YMB‐1 cells. On the other hand, the expression level of IGFBP5 transcripts was significantly increased by vorinostat (Fig. 8D). In gastrointestinal stromal tumors, the expression level of the Ca2+‐activated Cl− channel ANO1/TMEM16A was negatively correlated with that of IGFBP5 (Simon et al. 2013). This study showed no significant downregulation of KCa3.1 by the overexpression of IGFBP5 in YMB‐1 cells (Fig. 8F). Of seven IGFBP subtypes (IGFBP1‐7), IGFBP2 is predominatly expressed in YMB‐1 cells, and no significant changes in KCa3.1 expression was caused by the overexpression of IGFBP2 (not shown). The destabilization of DNA methylation (hypermethylation or hypomethylation) in ion channels has been correlated with tumorigenesis and a poor prognosis (Ouadid‐Ahidouch et al. 2015). Hypomethylation of the KCa3.1 promoter has recently been associated with the upregulation of KCa3.1 in aggressive nonsmall‐cell lung cancer (Bulk et al. 2015), suggesting that the upregulation of KCa3.1 may mediate the hypomethylation of its promoter gene in KCa3.1‐expressing cancer cells. Further studies are needed in order to elucidate the mechanism underlying the epigenetic regulation of the KCa3.1 gene.

Similar to previous findings by Haren et al. (2010), this study showed that KCa3.1 was associated with the enhancement of breast cancer cell proliferation (Fig. 1). Large‐conductance KCa1.1 and small‐conductance KCa2.3 have been implicated in metastasis by facilitating breast cancer cell migration (Khaitan et al. 2009; Chantôme et al. 2013). Metalloproteinases 2 and 9 (MMP‐2 and MMP‐9) have been associated with breast cancer invasion (Jaafar et al. 2014; Vucemilo et al. 2014), and the reduction in MMP‐2 may play a role in the suppressive effects of the KCa3.1 blocker on breast cancer migration and invasion (Zhang et al. 2015). Similar findings have been reported in glioma cells (D'Alessandro et al. 2013; Aroui et al. [Link]). However, the expression levels of MMP‐2 and MMP‐9 were very low in YMB‐1, and no significant changes were found in their expression by HDAC inhibition or pharmacological KCa3.1 blockade (not shown).

HDACis were primarily developed as antitumor agents, but are currently being explored for the treatment of neurodegenerative, immunological, metabolic, inflammatory, atopic, and cardiovascular disorders. Felice et al. (2015) recently showed that HDAC2 and HDAC3 isoforms were involved in chronic intestinal inflammation, and HDAC inhibition decreased disease responses in inflammatory bowel disease (IBD). HDACis including vorinostat ameliorate intestinal inflammation by inhibiting inflammatory cytokine production in the CD4+ T cells of IBD model mice (Glauben et al. 2006). Recent studies including our previous study demonstrated that the upregulation of KCa3.1 in CD4+ T cells was involved in the pathogenesis of IBD, and the pharmacological blockade of KCa3.1 elicited a significant decrease in IBD disease severity including intestinal inflammation (Di et al. 2010a; Ohya et al. 2014). Gillespie et al. (2012) also found that a selective HDAC3 inhibitor reduced the production of the proinflammatory cytokine, IL‐6 in mononuclear cells from rheumatoid arthritis patients. The pharmacological and genetic inhibition of KCa3.1 was previously shown to suppress IL‐6 levels in rheumatoid arthritis patients and allergic rhinitis model mice (Friebel et al. 2015; Lin et al. 2015). Furthermore, we reported the inhibition of IL‐6 transcription by the pharmacological blockade of KCa3.1 in colonic inflammation in IBD model mice (Ohya et al. 2014). These suggest selective HDAC inhibition of HDAC2 and/or HDAC3 as a new therapeutic strategy to improve chronic inflammatory diseases.

In conclusion, this study demonstrated that the selective inhibition of HDAC2 and/or HDAC3 is a novel therapeutic strategy for drug development focused on KCa3.1‐expressing cancer cells such as aggressive breast and prostate cancers. Moreover, HDAC2 and/or HDAC3 may become therapeutic targets for the treatment of chronic inflammatory and atopic disorders.

Author Contributions

Participated in research design: S. Ohya, T. Suzuki, K. Muraki.

Conducted experiments: S. Ohya, S. Kanatsuka, N. Hatano, H. Kito, A. Matsui, M. Fujimoto, S. Matsuba, S. Niwa, P. Zhan.

Performed data analysis: S. Ohya, S. Kanatsuka, N. Hatano, H. Kito, A. Matsui, M. Fujimoto, S. Matsuba, S. Niwa, P. Zhan, T. Suzuki, K. Muraki.

Wrote or contributed to the writing of the manuscript: S. Ohya, S. Kanatsuka, N. Hatano.

Disclosures

None declared.

Figure S1. Expression levels of PRL receptor (PRLR) transcripts in human breast cancer cell lines, and effects of chemotherapy agents (paclitaxel and bafilomycin‐A) and the SIRT1/2 inhibitor, NCO‐04. (A) Expression levels of PRLR transcripts in YMB‐1, MCF‐8, Hs578T, BT549, and MDA‐MB‐453 cells were determined. (B) Effects of paclitaxel (100 nmol/L) and bafilomycin‐A (10 nmol/L) on the expression levels of KCa3.1 transcripts in YMB‐1 cells. (C) Expression levels of class III HDACs, sirtuin subtype (SIRT1‐7) transcripts in YMB‐1, (D) Effects of NCO‐04 on the expression levels of KCa3.1 transcripts in YMB‐1 cells. Expression levels were expressed as a ratio to ACTB. Results are expressed as means ± SEM (n = 3–4). The following gene‐specific PCR primers of human origin were used: PRLR (NM_000949, 1904‐2025), 122 bp; SIRT 1 (NM_012238, 2004‐2124), 121 bp; SIRT2 (NM_012237, 977‐1097), 121 bp; SIRT3 (NM_012239, 704‐824), 121 bp; SIRT4 (NM_012240, 888‐1008), 121 bp; SIRT5 (NM_012241, 1071‐1191), 121 bp; SIRT6 (NM_016539, 188‐322), 135 bp; SIRT7 (NM_016538, 968‐1088), 121 bp.

Figure S2. Expression levels of various types of K+ channel transcripts in YMB‐1 cells. (A) voltage‐gated Kv1 subtypes. (B) voltage‐gated Kv2, Kv3, and Kv4 subtypes. (C) Kv10, 11, and 12 subtypes. (D) Kv7 subtypes. (E) K2P subtypes. (F) KCa subtypes and Kir2.1. Expression levels were expressed as a ratio to ACTB. Results are expressed as means ± SEM (n = 4). The following gene‐specific PCR primers of human origin were used for real‐time PCR: Kv1.1 (GenBank accession number: NM_000217, 967‐1072), 106 bp; Kv1.2 (NM_004974, 94‐194), 101 bp; Kv1.3 (XM_084080, 1411‐1517), 107 bp; Kv1.4 (NM_002233, 1145‐1245), 101 bp; Kv1.5 (XM_006988, 1968‐2092), 125 bp; Kv1.6 (XM_018513, 1401‐1504), 104 bp; Kv1.7 (AJ310479, 1456‐1598), 143 bp; Kv2.1 (NM_004975, 950‐1050), 101 bp; Kv2.2 (NM_004770, 221‐321), 101 bp; Kv3.1 (NM_004976, 647‐747), 101 bp; Kv3.2 (AF268897, 1418‐1519), 102 bp; Kv3.3 (AF055989, 2649‐2771), 123 bp; Kv3.4 (NM_004978, 839‐946), 108 bp; Kv4.1 (NM_004979, 1494‐1598), 105 bp; Kv4.2 (NM_012281, 2275‐2384), 110 bp; Kv4.3 (NM_004980, 1707‐1807), 101 bp; Kv7.1 (NM_000218, 1556‐1656), =101 bp; Kv7.2 (NM_004518, 536‐664), 129 bp; Kv7.3 (NM_004519, 1958‐2058), 101 bp; Kv7.4 (NM_004700, 1741‐1841), 101 bp; Kv7.5 (NM_019842, 341‐443), 103 bp; Kv10.1 (NM_002238, 1640‐1761), 122 bp; Kv10.2 (NM_139318, 2109‐2209), 101 bp; Kv11.1 (NM_000238, 2718‐2823), 106 bp; Kv11.2 (NM_030779, 32‐172), 141 bp; Kv11.3 (NM_033272, 3558‐3658), 101 bp; Kv12.1 (NM_144633, 688‐798), 111 bp, Kv12.2 (XM_035483, 1137‐1237), 101 bp; Kv12.3 (NM_012285, 1340‐1444), 105 bp; K2P1.1 (NM_002245, 948‐1048), 101 bp; K2P2.1 (NM_014217, 849‐949), 101 bp; K2P3.1 (NM_002246, 622‐744), 123 bp; K2P4.1 (NM_016611, 413‐538), 126 bp; K2P5.1 (EU978936, 577‐697), 121 bp; K2P6.1 (NM_004823, 1136‐1252), 117 bp; K2P7.1 (NM_033347, 827‐966), 140 bp; K2P9.1 (NM_016601, 1025‐1146), 122 bp; K2P10.1 (NM_021161, 702‐802), 101 bp; K2P12.1 (NM_022055, 382‐497), 116 bp; K2P13.1 (NM_022054, 1587‐1722), 136 bp; K2P15.1 (NM_022358, 1094‐1206), 113 bp; K2P16.1 (NM_001135105, 805‐925), 121 bp; K2P17.1 (NM_031460, 444‐563), 120 bp; K2P18.1 (NM_181840, 849‐969), 121 bp; KCa1.1 (NM_001014797, 679‐798), 120 bp; KCa2.1 (NM_002248, 649‐764), 116 bp; KCa2.2 (NM_021614, 1492‐1612), 121 bp: KCa2.3 (NM_002249, 2042‐2146), 105 bp; KCa4.1 (NM_020822, 1719‐1845), 127 bp; KCa4.2 (NM_198503, 1826‐1946), 121 bp; Kir2.1 (NM_000891, 765‐865), 101 bp.

Figure S3. Measurement of 1 μmol/L TRAM‐34‐sensitive currents and store‐operated Ca2+ entry (SOCE) in YMB‐1 cells and effects of 1 μmol/L TRAM‐34 on relative AUC of [Ca2+]i in HDACis‐treated YMB‐1 cells. (A and B) schematic diagrams of the analytical methods to calculate 1 μmol/L TRAM‐34‐sensitive currents (A) and relative AUC of [Ca2+]i (B) (see “Materials and Methods”). 0 Ca2+: 0 mmol/L Ca2+, 2.2 Ca2+: 2.2 mmol/L Ca2+, TG: 1 μmol/L thapsigargin, AUC: area under the curve. (C and D) Effects of TRAM‐34 (1 μmol/L) on relative AUC of [Ca2+]i SOCE in 300 nmol/L AATB‐ and 1 μmol/L T247‐treated YMB‐1 cells for 30–48 h. Results are expressed as means ± SEM. Numbers used for the experiments are shown in parentheses.

Figure S4. Transcriptional expression levels of the regulatory molecules of KCa3.1 in human breast normal and tumor tissues. Real‐time PCR assay for PI3K‐C2B (A), PHPT1 (B), and MTMR6 (C). The expression levels were expressed as the ratio to ACTB. The results are expressed as the means ± SEM (n = 3 for each). The following gene‐specific PCR primers of human origin were used: PI3K‐C2B (NM_002646, 4052‐4172), 121 bp; PHPT1 (NM_001135861, 783‐903), 121 bp; MTMR6 (NM_004685, 1476‐1607), 132 bp.

Figure S5. Expression levels of (A) HER2 transcripts in human breast cancer cell lines (YMB‐1, MCF‐7, Hs578T, BT549 and MDA‐MB‐453) (n = 3), (B) HDAC1, HDAC2, HDAC3, and HDAC6 transcripts in MDA‐MB‐453 cells (n = 4), and (C) KCa3.1 and HER2 transcripts in human primary breast tumor tissues (n = 5). Expression levels were expressed as a ratio to ACTB. Results are expressed as means ± SEM. The following gene‐specific PCR primers of human origin were used: HER2 (NM_004448, 1440‐1559), 120 bp; HDAC1 (NM_004964, 708‐824), 117 bp; HDAC6 (NM_006044, 3517‐3637), 121 bp.

Click here for additional data file.