Nội dung

Rezania et al. BMC Cancer (2016) 16:628 DOI 10.1186/s12885-016-2664-8 RESEARCH ARTICLE Open Access Overexpression of KCNJ3 gene splice variants affects vital parameters of the malignant breast cancer cell line MCF-7 in an opposing manner S. Rezania1,8, S. Kammerer1,8, C. Li1,8, B. Steinecker-Frohnwieser1,8,9, A. Gorischek1,8, T. T. J. DeVaney1,8, S. Verheyen1,8,9, C. A. Passegger2, N. Ghaffari Tabrizi-Wizsy2, H. Hackl3, D. Platzer1, A. H. Zarnani4, E. Malle5, S. W. Jahn6, T. Bauernhofer7,8 and W. Schreibmayer1,8* Abstract Background: Overexpression the KCNJ3, a gene that encodes subunit 1 of G-protein activated inwardly rectifying K+ channel (GIRK1) in the primary tumor has been found to be associated with reduced survival times and increased lymph node metastasis in breast cancer patients. Methods: In order to survey possible tumorigenic properties of GIRK1 overexpression, a range of malignant mammary epithelial cells, based on the MCF-7 cell line that permanently overexpress different splice variants of the KCNJ3 gene (GIRK1a, GIRK1c, GIRK1d and as a control, eYFP) were produced. Subsequently, selected cardinal neoplasia associated cellular parameters were assessed and compared. Results: Adhesion to fibronectin coated surface as well as cell proliferation remained unaffected. Other vital parameters intimately linked to malignancy, i.e. wound healing, chemoinvasion, cellular velocities / motilities and angiogenesis were massively affected by GIRK1 overexpression. Overexpression of different GIRK1 splice variants exerted differential actions. While GIRK1a and GIRK1c overexpression reinforced the affected parameters towards malignancy, overexpression of GIRK1d resulted in the opposite. Single channel recording using the patch clamp technique revealed functional GIRK channels in the plasma membrane of MCF-7 cells albeit at very low frequency. Discussion: We conclude that GIRK1d acts as a dominant negative constituent of functional GIRK complexes present in the plasma membrane of MCF-7 cells, while overexpression of GIRK1a and GIRK1c augmented their activity. The core component responsible for the cancerogenic action of GIRK1 is apparently presented by a segment comprising aminoacids 235–402, that is present exclusively in GIRK1a and GIRK1c, but not GIRK1d (positions according to GIRK1a primary structure). Conclusions: The current study provides insight into the cellular and molecular consequences of KCNJ3 overexpression in breast cancer cells and the mechanism upon clinical outcome in patients suffering from breast cancer. Keywords: KCNJ3, Breast cancer, GIRK1, MCF-7, Splice variant * Correspondence: W.Schreibmayer@gmx.at 1 Institute of Biophysics, Molecular Physiology Group, Medical University of Graz, Harrachgasse 21/4, Graz, Austria 8 Research Unit on Ion Channels and Cancer Biology, Medical University of Graz, Graz, Austria Full list of author information is available at the end of the article © 2016 The Author(s). Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Rezania et al. BMC Cancer (2016) 16:628 Background Four gene loci within the human genome encode for subunits of G-protein activated K+ channels (GIRK1-4). GIRK subunits form homo- or hetero-tetrameric ion channel complexes in the plasma membrane, act as classical Gprotein effectors thereby mediating the regulation of cellular activity and/or excitability via hormones and neurotransmitters [1]. So far, known physiological roles of GIRKs comprise the vegetative regulation of the heartbeat, pain perception, learning and memory, anxiety behaviour and reward mechanisms [1]. Also in electrically nonexcitable tissues physiological functions, like pancreatic insulin secretion [2, 3], blood platelet aggregation [4, 5] and regulation of lipid metabolism in fat cells [6] have been reported. Two of the gene loci encoding GIRK subunits in humans have been proven to be related to tumorigenesis and tumor growth: somatic mutations in the KCNJ9 gene (encoding the GIRK4 subunit) have been identified to induce endocrine renal adenomas that cause primary aldosteronism and severe hypertension [7]. Overexpression of mRNA encoding the GIRK1 subunit, the product of the KCNJ3 gene, may contribute significantly to the malignant properties of breast cancers: using expression profiling, Stringer et al. [8] observed that RNA derived from KCNJ3 was aberrantly and highly overrepresented in primary invasive breast carcinomas when compared to the corresponding healthy breast tissue. GIRK1 mRNA overexpression correlated both with occurrence and number of lymph node metastases. Later on, Brevet et al. [9] observed a positive correlation between the immunohistochemical staining of GIRK1 in breast tumor specimen and lymph node metastasis and an inverse correlation with overall survival of the patients. A retrospective study, based on data from 905 invasive breast cancers derived from The Cancer Genome Atlas (TCGA) confirmed the findings delineated above at an appreciably larger scale. This corroborates the correlation between KCNJ3 expression and breast cancer progression [10]. Malignant breast cancer cell lines express mRNAs encoding GIRK1 (but also GIRK2 and GIRK4) subunits [11] and several splice variants of the KCNJ3 gene transcript [12]. In addition, the occurrence of GIRK1 and GIRK4 protein has been demonstrated in several breast cancer cell lines, including MCF-7 [12, 13]. Increasing evidence for KCNJ3 expression in cancerous, compared to normal breast tissue and for its correlation with disease progression has accumulated. Comparatively little is known on a possible causal relationship between KCNJ3 expression, tumorigenesis and cancer progression. GIRK1 protein may drive benign mammary epithelial cells (MECs) towards hallmarks of malignancy. In order to investigate a presumable role of GIRK1 in oncogenesis and metastasis of MECs, we overexpressed full length human GIRK1a as well as two splice variants, GIRK1c and GIRK1d (known to be abundant in breast cancer cells Page 2 of 15 [12]), in the MCF-7 breast cancer cell line. This cell line was chosen, as GIRK1 mRNA levels are high, but expression of the corresponding protein(s) is low [12, 13] with the prospect to further strengthen potential malignant predicates due to pronounced overexpression. Analysis and comparison of selected vital parameters were performed in order to pinpoint characteristic features of MCF-7 that were possibly influenced by KCNJ3 overexpression. By identification of peculiar properties that may be affected, we anticipated insight into the mechanism(s) how GIRK1 accomplishes its malignant task. Methods Solutions (concentrations in mmole/L): Zeroing Bathing Solution (ZBS) K+/Asp-(120), KCl (20), MgCl2 (4), NaCl (10), EGTA−/K+ (10), HEPES− (10), buffered with K+ to pH:7.4. Pipette Filling Solution (PFS): KCl (153), MgCl2 (4), CaCl2 (1), GdCl3 (0.2), HEPES− (10) buffered with K+ to pH: 7.4. Neutral buffered formalin (NBF): 10 % formalin, PO−4 (75) buffered with Na+ to pH:7.0. Cell culture MCF-7 cell line was obtained from ATCC (American Type Culture Collection) and maintained in minimal essential medium (MEM; Gibco, Life Technologies, Grand Island, NY, USA; Ordering No: 31095_029) supplemented with 10 % fetal bovine serum (Sigma Aldrich, St. Louis, USA, cat.No.: F2442), 1 mmole/L sodium pyruvate (Sigma Aldrich; St. Louis, USA, cat.No.: S8636) and penicillin/streptomycin (100 U.mL−1/100 ng.mL−1; Sigma Aldrich; St. Louis, USA, cat.No.: P0781) in 5 % CO2 atmosphere at 37 °C. Constructs N-terminal (N-T) fusions of GIRK1a, GIRK1d and GIRK4 with enhanced yellow fluorescence protein (eYFP) and enhanced cyan fluorescence protein (eCFP) were expressed in MCF-7 cells using the pEYFP-C1 and pECFP-C1 based constructs described in detail in [12]. C-terminal (C-T) fusions of GIRK1a and GIRK1c with eYFP were produced by cloning the corresponding coding DNA sequence (CDS) into the plasmid pEYFP-N1 (Clontech Laboratories, Inc., Mountain View, CA, USA) using XhoI and EcoRI restriction sites. For fluorescence labelling of subcellular compartments plasmids encoding glycosylphosphatidyinositol/eCFP (GPIeCFP; for lipid rafts within plasma membrane [14]) and signal recognition particle receptor ß-subunit/ eCFP (SrßeCFP; for endoplasmic reticulum (ER) [15]) were used. A vector for mammalian overexpression of fluorescence labelled G-protein β/γ subunits was generated by cloning Gγ2 CDS (Genbank Acc.No.: M37183) into the multiple cloning site (MCS) B of the pIRES vector (Clontech Laboratories, Inc., Mountain View, CA, USA) via XbaI Rezania et al. BMC Cancer (2016) 16:628 and SalI restriction sites. Subsequently, the CDS of a fusion protein of eYFP with Gβ1 (Genbank Acc.No.: M313236; Nterminal with respect to Gβ1) was inserted into MSC B via NheI and EcoRI restriction sites. Integrity of the construct was verified by sequencing. Biological activity of fluorescence labelled G-protein β/γ subunits was verified by coexpression of the corresponding synthetic mRNAs in Xenopus laevis oocytes and subsequent electrophysiological testing for their ability to activate coexpressed GIRK ion channels composed of the GIRK1a and GIRK4 subunits (data not shown). Transfection MCF-7 cells were transfected with the different constructs using TransFast reagent (Promega, Madison, USA, Cat. No.: E2341) and studied approx. 24 h after transfection. For stable transfection, pEYFP-C1 and pEYFP-N1 based constructs were linearized with AseI and pIRES construct with SalI, respectively, prior to transfection. Selection was started by adding 3 mg/mL G418 (Gibco, life technologies, Grand Island, USA, Ordering No.: 11811031) to the medium 24 h after the transfection. Single cell sorting was done two weeks after G418 addition. Individual clones were chosen by visual inspection using confocal Laser Scan Microscopy (cLSM). See Additional file 1: Table S1 for list of clonal cell lines that were used for the present study. Confocal laser scan microscopy Fluorescence images of transfected MCF-7 cells were obtained in-vivo using Leica inverted microscope with 63x H2O immersion objective (NA: 1.20) with attached laserscan module (DMIRE2 and TCS SL2; Leica Microsystems, Heidelberg, Germany) as described previously [12]. qPCR RNA isolation and cDNA synthesis were performed as described previously [12]. qPCR has been described in [16]. Primer sequences were as follows: GIRK1a_f: 5′-G TGGAAACAACTGGGATGAC-3′; GIRK1a_r: 5′-GTT GCATGGAACTGGGAGTA-3′; GIRK1c_f: 5′- CAAGC TGCTCAAATCTCGGC-3′; GIRK1c_r: 5′-AGTTGATC TGCCCCTGTACT-3′; GIRK1d_f: 5′-CAAGCTGCTCA AAGGATGAC-3′; GIRK1d_r: 5′-GTTGCATGGAACT GGGAGTA-3′; GAPDH_f: 5′-ATGGGGAAGGTGAAG GTCG-3′; GAPDH_r: 5′-GGGGTCATTGATGGCAAC AATA-3′. Immunohistochemistry (IHC) Cells were fixed in NBF, embedded in agarose gel (7 %) and then processed for paraffin embedding. Target retrieval solution (pH: 9.0; Dako, Glostrup, Denmark; Product No: S236884) heated for 40 min at 150 W in a microwave was used for antigen retrieval. Slides were then washed in Page 3 of 15 washing buffer (Dako, Glostrup, Denmark; Product No: S3006) and incubated with a monoclonal antibody against GIRK1 (Abcam, Cambridge, UK; cat.No: 119246; 1:50; clone 3E11). For visualization, the EnVision + dual link reagent (rabbit/mouse horse radish peroxidase, Glostrup, Dako, Denmark; Product No: K406311) was used according to manufacturer’s protocols. Immunohistochemical staining was developed by incubation of sections with diaminobenzidine (DAB; Glostrup, Dako; Product No: K406511) as a chromogenic substrate. Slides were then washed in Dako wash buffer, counterstained with Meyer’s hematoxylin (from the pharmacy of the Medical University of Graz), rinsed in tap water, dehydrated and mounted with Entellan® (Merck, Darmstadt, Germany). Sections incubated without primary antibody served as negative controls. Analysis of vital parameters of cell lines In order to avoid possible deviations of vital parameters that might be due to the cloning procedure itself instead of differential overexpression of GIRK1 variants, assessment of these parameters was always conducted on more than one cell line overexpressing identical constructs (Additional file 1: Table S1). As no difference in vital parameters between cell lines expressing identical GIRK1 variants was observed, these data were pooled and analyzed collectively. In order to monitor eventual effects of stable eYFP overexpression alone or of the manipulation of cellular genome on the vital parameters tested, all vital assays were performed using both MCF7WT and MCF-7eYFP as controls. Adhesion assay MCF-7 cells were washed with PBS and plated into each well of Corning® BioCoat™ Fibronectin 96 Well Clear Flat Bottom (Corning, NY. USA, Cat No: 354409). Nonadherent cells were removed 150 min later by washing with PBS. Adherent cells were fixed with 2 % formaldehyde, air dried and stained with 0.1 % crystal violet (Sigma Aldrich, St.Louis, USA; Cat No: C0775) in PBS. Bound dye was solubilized with 10 % acetic acid and absorbance was measured at 550 nm using a plate reader (Labsystem Mutiskam MS). Cell-free wells served as blanks. Proliferation assay Cells were plated in six well plates and incubated in a cell culture incubator until reaching 80 % confluency. Proliferation of the cells were assessed based on the incorporation of the thymidine analog, 5-bromo-2′-deoxyuridine (BrdU) into the newly synthesized DNA of replicating cells (during S phase). Labeling of DNA was done by adding 10 μL of BrdU solution directly to each mL of cell culture media. (APC BrdU flow kit; BD Pharmingen, San Diego, CA, USA; Cat No: 552598). The treated MCF-7 cells were then Rezania et al. BMC Cancer (2016) 16:628 incubated for 3 h in cell culture incubator. The cells were then stained with Anti-BrdU APC. 7-AAD (7-Aminoactinomycin D) (DNA binding dye) used in order to define cell cycle (G1/G0, S, G2/M). Invasion assay Corning® BioCoat™ Matrigel® Invasion Chambers (Corning, NY. USA, Cat No: 354480) were rehydrated in MEM for 2 h at 37 °C. 1.25x105 cells/mL in 2 mL MEM (without fetal bovine serum (FBS)) were seeded into the upper compartment and 2 mL MEM with 5 % FBS as chemoattractant were added to the lower compartment of Matrigel. After 24 h incubation in cell culture incubator, non-invading cells were removed from the upper surface of the membrane by scrubbing with cotton tipped swabs. To stain the invading cells, membranes beneath the insert were cut, fixed with ice-cold methanol and stained with 0.1 % crystal violet. Invading cells were counted under microscope. Wound healing assay Cells were plated into 24 well plates (1x105 cells/well) and incubated for 24–72 h to achieve a confluent monolayer. The cell monolayer was scratched in a straight line with a pipet tip (VWR, USA; Cat No #53508-910). Debris was removed by washing the cells once with PBS followed by adding MEM growth medium. The plate was put into the cell observer (Zeiss Axiovert 200 M). Images were taken over the course of 72 h at 1 h interval. ImageJ software was used for analysis of the resulting time lapse videos [17]. Motility assay Cellular velocities and motility coefficients were assessed by cell observer (Axiovert200M, Zeiss, Germany) over a time period of 72 h, as described previously [16]. Ex ovo chorioallantoic membrane (CAM) assay Fertilized white leghorn chicken eggs from local hatchery (Schropper GmbH, Gloggnitz, Austria) were incubated at 37.6 °C and 70 % humidity (J. Hemel Brutgeräte GmbH & Co KG, Am Buschbach, Germany). The egg shell was cracked on day 3 of chicken embryo development, the embryo was decanted to a sterile dish and further incubated as indicated before. On day 10 of cultivation cell onplants with volumes of 20 μl were applied on vascular branches of the CAM within sterile silicon rings of 5 mm diameter (1x106 cells in 10 μl PBS mixed 1:2 with Matrigel® (Corning, NY, USA, cat.No: 356237), allowing subsequent tumor growth for 3 days [18]. The intensity of the angiogenic response was analyzed under a stereomicroscope according as described previously [19]. Page 4 of 15 Electrophysiology Single channel recording from MCF-7 cells in the cell attached configuration was performed exactly as described previously for PIEZO1 mechanosensitive ion channel protein, but without application of mechanical stress to the membrane [16]. Statistics Statistical evaluation was performed using SPSS software in the SigmaStat environment (SigmaPlot 13.0; Systat Software GmbH, Erkrath, Germany) or using “R” software (version 3.2.1; https://www.r-project.org/). Depending on variance and distribution of the dataset, the appropriate tests were performed, as specified in the legends to the figures. Results Characterization of the generated MCF-7 based cell lines Effective overexpression of GIRK1 protein in the stably transfected MCF-7 cell lines was verified by several independent methods. GIRK1 is an integral membrane protein with two transmembrane α-helices per subunit; therefore, the subcellular localization of fluorescent chimaeras in membranes is expected to indicate successful overexpression of the entire protein. Indeed, expression of C-terminal labelled GIRK1a was different compared to soluble eYFP alone, GIRK1 being located mostly in the endoplasmic reticulum (ER) and, to some extent, also in the plasma membrane, but not in the nucleus. Pure eYFP distributed evenly within the cytosol and the nucleus, with the exception of vacuoles (Fig. 1). Localization identical to C-terminal labelled GIRK1a was also observed for Nterminal labelled one, indicating that the position of the fluorophore did not affect the proteins’ synthesis or subcellular localization (Additional file 1: Figure S1). Similar subcellular localization was observed for the shorter splice variants GIRK1c and GIRK1d (Fig. 1). In addition to fluorescence microscopy, qPCR corroborated overexpression of mRNAs encoding all GIRK1 variants in the established cell lines (Fig. 2a; Additional file 1: Figure S2). Since the epitope recognized by the antibody used is present exclusively in full length GIRK1a, IHC may not be considered a suitable tool to confirm expression of the GIRK1c/1d as these splice variants lack the corresponding part of the Cterminal portion, respectively. Under our experimental conditions native MCF-7WT cells did not exert detectable GIRK1a expression in IHC, underscoring the scale of protein overexpression in MCF-7GIRK1a cells by orders of magnitude; these results are in line with qPCR results (Fig. 2a). Finally, IHC clearly demonstrated overexpression of the protein as well as its subcellular localization (Fig. 2b). Thereupon we conclude that the cell lines generated represent valid models to study the biological effect(s) of GIRK1 variant overexpression in human breast cancer. Rezania et al. BMC Cancer (2016) 16:628 a Page 5 of 15 GIRK1a marker overlay transmission GIRK Srß overlay transmission eYFP Srß overlay transmission GpI Srß b 1c 1d control Fig. 1 cLSM in-vivo reveals overexpressed GIRK1 protein to localize in a large part to the ER. Horizontal sequences of images show identical cells. The sequence of channels (from left to right is: eYFP (green)/ eCFP (red)/ overlay / transmission. Scalebar : 15 μm in all images. a Upper sequence: subcellular localization of GIRK1a, labelled with eYFP at the C-terminus (MCF-7GIRK1a; transient transfection with GpI-eCFP was used as marker for lipid rafts in the plasma membrane). Lower sequence: subcellular localization of GIRK1a, labelled with eYFP at the C-terminus (MCF-7GIRK1a; transient transfection with Srß-eCFP was used as ER marker). b Upper sequence: subcellular localization of hGIRK1c, labelled with eYFP at the C-terminus (MCF-7GIRK1c; transient transfection with Srß-eCFP was used as ER marker). Middle sequence: subcellular localization of hGIRK1d, labelled with eYFP at the N-terminus (MCF-7GIRK1d; transient transfection with Srß-eCFP was used as ER marker). Lower sequence: subcellular localization of eYFP alone (MCF-7eYFP; transient transfection with Srß-eCFP was used as ER marker) Effects of GIRK1 overexpression on adhesion and proliferation of MCF-7 cells Modification of selected cellular vital parameters in culture by candidate proteins indicates whether a protein under consideration may act as an oncoprotein causing and promoting particular features of malignancy [20, 21]. Cell adhesion (Fig. 3) remained unaffected upon overexpression of GIRK1a, GIRK1c and GIRK1d variants. When monitoring cell cycle and proliferation via gated cell sorting, it became evident that the parameters tested where, for the major part, unaffected (Fig. 4). The exception was MCF-7GIRK1d, that exerted an increased period between cell division and start of DNA replication (G0/G1) when compared to all other groups. This increase was moderate and statistically significant only by comparison with MCF7GIRK1a. The difference is in line with our observation (SR; AG) that MCF-7GIRK1d cells require longer time intervals to grow, when compared to the other cell lines under identical conditions. We conclude that upon GIRK1 overexpression, proliferative signaling remained practically unchanged in MCF-7 under our experimental conditions. GIRK1 overexpression interferes with wound healing and invasion Both wound healing and tumor development, two processes that may seem unrelated at the first glance, are based on similar molecular mechanisms and signaling pathways Rezania et al. BMC Cancer (2016) 16:628 Page 6 of 15 p < 0.05 p < 0.001 1000 (7) normalized mRNA (12) (6) 100 10 n.s. (20) (18) 1 WT eYFP b MCF-7WT MCF-7GIRK1a hG1a hG1c hG1d Fig. 2 Overexpression of GIRK1 mRNA and protein in the stably transfected MCF-7 cells assessed by IHC. a Quantification of mRNA expression encoding different GIRK1 variants via qPCR in stably transfected cells. Expression was normalized to the expression level in the MCF-7WT cell line. WT: MCF-7WT; eYFP: data derived from two cell lines with stably integrated pEYFPN1 vector alone (MCF-7eYFP); hG1a: data derived from two cell lines overexpressing N-terminal and two cell lines overexpressing C-terminal fusions of eYFP with the GIRK1a protein (MCF-7GIRK1a); hG1c: data derived from two cell lines overexpressing C-terminal fusions of eYFP with the GIRK1c protein (MCF7GIRK1c); hG1d: data derived from two cell lines overexpressing N-terminal fusions of eYFP with the GIRK1d protein (MCF-7GIRK1d). Mean values ± SEM were plotted (number of experiments is given in parenthesis above each bar). Statistical significances between groups are indicated. hG1a, as well as hG1d differs from both WT, as well as from eYFP, statistically significant at the p < 0.001 level. hG1c differs from both WT, as well as from eYFP, statistically significant at the p < 0.05 level. Kruskal-Wallis one way analysis on ranks was used for analysis of statistical significance. b Detection of GIRK1a protein in stably transfected cells via immunohistochemistry. Brown: Immunoreactivity. Blue: hematoxylin stain. Scale bar corresponds to 100 μm throughout. Upper image: MCF-7WT; Lower image: MCF-7hGIRK1a [22]. Wound healing, however, is a rather transient process, while tumors pursue to evolve and spread. Hence, wound healing assay was performed to identify additional characteristic vital parameters, potentially affected by GIRK1 expression. As shown in Fig. 5a the rate of wound closure was markedly enhanced by overexpression of the full length GIRK1a protein when compared to control (see also Additional files 2, 3 and 4). While overexpression of GIRK1c produced a similar, albeit statistically not significant increase, overexpression of GIRK1d did not result in (12) (6) 0.5 (3) (6) (6) OD550nm p < 0.001 a 0.0 WT eYFP hG1a hG1c hG1d Fig. 3 Surface adhesion of MCF-7 cells is unaffected by GIRK1 overexpression. Quantification of cells adhering to fibronectin coated substrate via OD550nm. WT: MCF-7WT; eYFP: MCF-7eYFP; hG1a: MCF-7GIRK1a; hG1c: MCF-7GIRK1c; hG1d: MCF-7GIRK1d. Mean values ± SEM were plotted (number of experiments is given in parenthesis above each bar). The mean values do not differ statistically significantly Rezania et al. BMC Cancer (2016) 16:628 a Page 7 of 15 MCF-7GIRK1d MCF-7eYFP 62,5% 30,5% 7,0% 66,1% 27,3% 6,6% MCF-7GIRK1a MCF-7GIRK1c 58,8% 32,7% 8,5% 56,8% 33,4% 9,8% Cell Cycle % b : G0/G1 :S : G2/M p < 0.05 50 0 WT eYFP hG1a hG1c hG1d Fig. 4 Survey of cell cycle and proliferation upon GIRK1 overexpression in MCF-7 cells. a Original results from the assessment of cell cyle using gated cell sorting according to fluorescence intensities for PerCP-A (x-axis) and APC-A (y-axis) for different experimental groups. % of cells for the given experiment is stated in respective colors besides the plot. b Statistics for the percentage of time spent in the different phases of the cell cycle Mean values ± SEM were plotted. N was (in parenthesis behind each experimental group): was: MCF-7WT (8) / MCF-7eYFP (12) / MCF-7GIRK1a (16) / MCF-7GIRK1d (12) / MCF-7GIRK1d (6). G1/G0 fraction of MCF-7GIRK1d differs statistically significant at the p < 0.05 level from the one of MCF-7GIRK1a. One way ANOVA was used for analysis of statistical significance an increase of wound closure rate that was even slightly reduced when compared to control (Fig. 5b). Next, the Matrigel invasion assay regarded to be indicative for activation of invasion and metastasis was performed. This assay unveiled that GIRK1 overexpression affected invasion towards a chemoattractant in a bimodal manner, depending on the respective splice variant: overexpression of GIRK1d greatly reduced the number of cells with invasive phenotype, while overexpression of GIRK1a and GIRK1c slightly promoted invasion, although not statistically significant (Fig. 6; see Additional file 1: Figure S3 for representative micrographs of all the groups tested). Taken together, both assays uncover remarkable differences between the larger, higher molecular mass, splice variants GIRK1a and GIRK1c, which significantly promoted wound healing and invasive phenotype when compared to GIRK1d. Rezania et al. BMC Cancer (2016) 16:628 Page 8 of 15 Start 48h 72h b p < 0.001 p < 0.05 (19) p < 0.01 (13) GIRK1 overexpression affects angiogenesis 0.8 %.h-1 (7) (7) (17) 0.7 WT eYFP hG1a hG1c [23, 24]. When cellular velocities were directly quantified it became evident that average cellular velocities were greatly augmented upon overexpression of GIRK1a and GIRK1c, when compared to control (MCF-7eYFP), MCF7WT and MCF-7GIRK1d (Fig. 7). Average velocities of MCF7GIRK1d cells were indistinguishable from MCF-7WT or control cells. Similar results were obtained for cellular migration, as depicted by cellular motility coefficients (MCs) that were also considerably increased by GIRK1a and GIRK1c overexpression (see Additional files 5 and 6). Again, cellular MCs were indistinguishable when MCF7GIRK1d, MCF-7WT and control cells were compared (Fig. 8). In order to rule out the remote possibility that differences in average cellular velocities and MCs observed were produced by differences between individual MCF-7 cell clones, MCF-7WT cells were transiently transfected with GIRK1a and the corresponding control plasmid. Time-lapse microscopy revealed that GIRK1a transfected cells exhibited massively elevated average velocities and MCs, when compared to control cells (transfected with eYFP alone) or non-transfected ones (Additional file 1: Figure S4). hG1d Fig. 5 GIRK1 overexpression affects wound healing rate in a differential manner, depending on the variant tested. a Representative view of monolayers at different time intervals after scratching. Scale bars correspond to 200 μm. Left: control; right: GIRK1a overexpressors. b Wound healing rates (in % healing per h). WT: MCF-7WT; eYFP: MCF-7eYFP; hG1a: MCF-7GIRK1a; hG1c: MCF-7GIRK1c; hG1d: MCF-7GIRK1d. Mean values ± SEM were plotted (number of experiments is given in parenthesis above each bar). Statistical significances between groups are indicated. hG1a differs from eYFP statistically significant at the p < 0.05 level. hG1a differs from hG1d statistically significant at the p < 0.001 level. hG1c differs from hG1d statistically significant at the p < 0.01 level. Kruskal-Wallis one way analysis on ranks was used for analysis of statistical significance Cellular motilities and velocities are affected by GIRK1 overexpression Both wound healing and Matrigel invasion assays gather properties of cells related to cell migration in vitro, proliferation, cell/extracellular-matrix and cell/cell interactions Induction of angiogenesis represents another peculiar feature of cancer cells: while absent in normal tissue, except for embryogenesis, wound healing and endometrial cycling, the “angiogenic switch” is permanently turned on in tumor cells in order to provide tumor associated neovasculature [21]. To study the potential of the different MCF-7 based cell lines to induce angiogenesis, the CAM assay was performed (Fig. 9). No apparent difference in macroscopic vascularization scores (MVSs) was found for GIRK1 overexpressors when compared individually to either control (MCF-7eYFP) or MCF-7WT. MVS was, however, substantially decreased in MCF7GIRK1d cells when they are compared to MCF-7GIRK1a that showed the highest MVSs amongst all the cell lines studied. This finding suggests that the “angiogenic switch” becomes down-regulated by GIRK1d overexpression. Alternatively, GIRK1a overexpression has no effect or may even increase angiogenesis, when compared to MCF-7WT cells that have moderate angiogenic activity per se (see Fig. 9a and as an example, reference [25]). Are there functional GIRK ion channels in MCF-7 cells? In order to investigate whether GIRK1 is able to form functional K+ ion channels in the plasma membrane of MCF-7 cells, single channel recording was performed. At present limited information is available on G-protein coupled receptors (GPCRs) that may activate GIRKs in the MCF-7 cell line. Therefore we produced a MCF-7 based cell line, permanently overexpressing free GProtein β/γ subunits (Gβ/γ). Free Gβ/γ is known to Rezania et al. BMC Cancer (2016) 16:628 Page 9 of 15 Fig. 6 Effects of GIRK1 variant overexpression on chemoinvasion through Matrigel. a Comprehensive view of fixated cells, stained with crystal violet, that managed to invade the Matrigel layer (10x magnification). The Scalebar corresponds to 100 μm for both images. Upper: MCF-7GIRK1a, lower: MCF-7GIRK1d. b Quantification of the number of invasive cells after 24 h. WT: MCF-7WT; eYFP: MCF-7eYFP; hG1a: MCF7GIRK1a; hG1c: MCF-7GIRK1c; hG1d: MCF-7GIRK1d. Mean values ± SEM were plotted (number of experiments is given in parenthesis above each bar). Statistical significant differences between groups are indicated. hG1d differs from eYFP statistically significant at the p < 0.05 level. hG1a and hG1c differ from hG1d statistically significant at the p < 0.001 level. Kruskal-Wallis one way analysis on ranks was used for analysis of statistical significance MCF-7GIRK1a activate GIRK channels even in the absence of GPCR activation (MCF-7Gβγ; Additional file 1: Figure S5) [1]. Gβ/ γ induced K+ channel activity could be observed only in one single patch out of 53 experiments, making further experimentation unpromising. Transient overexpression of GIRK1 and GIRK4 subunits increased the frequency of observations of this K+ channel activity in MCF-7Gβγ to some extent, betokening that it had been produced by GIRK protein expression (Additional file 1: Figure S5). Of note a K+ channel with comparable properties has not yet been observed in MCF-7WT cells (Table 1). This observation suggests that endogenous GIRKs do not display appreciable activity in the absence of GPCR activation. In summary we have demonstrated that functional GIRK ion channels can be formed in MCF-7 cells, nevertheless their abundancy is, in the absence of ancillary overexpression, apparently very low. MCF-7GIRK1d b p < 0.05 p < 0.001 p < 0.001 (6) number of cells 500 (12) (3) (6) (6) 0 eYFP WT hG1a hG1c hG1d Discussion Our results clearly corroborate that overexpression of GIRK1 protein exerts profound effects on wound healing, chemoinvasion and cellular motility in the MCF-7 breast cancer cell line suggesting a role to promote invasion and metastasis. Induction of angiogenesis was also affected. Most noteworthy is the fact that all vital parameters affected by GIRK1 overexpression are manipulated in opposite direction, depending on the GIRK1 variant tested. Overexpression of either GIRK1a or GIRK1c reinforces vital properties of MCF-7 cells towards the malignant phenotype, while GIRK1d overexpression seemingly counteracts upon the opposite direction. Hence, differential features of GIRK1 variant proteins could be responsible for this antithetical behavior and comparison of their established functional properties may provide insight. While homo- and heterotetrameric K+ channels containing the full length GIRK1a subunit have lengthily been studied [1], little is known on the function(s) and essentially nothing on the possible (patho)physiological role of the smaller GIRK1c and GIRK1d variants. In the few studies undertaken so far by several groups using different expression systems, homotetramers composed of GIRK1c or GIRK1d Rezania et al. BMC Cancer (2016) 16:628 Page 10 of 15 v (µ m .m in -1 ) a 2 cell#01 0 2 cell#02 20 40 60 20 40 60 t (h) 0 2 cell#03 0 v (µ m .m in -1 ) GIRK1a 2 cell#01 0 2 cell#02 t (h) 0 2 cell#03 0 -1 b v (µm.min ) 0,0 (150) (395) p < 0.001 (75) p < 0.001 (70) hG1c p < 0.001 hG1a hG1d p < 0.001 eYFP (75) n.s. WT 0,2 Fig. 7 Effect of GIRK1 variant overexpression on single cell velocities. a Cellular velocities for three representative cells during the entire observation interval of 72 h. Asterisks denote cell divisions. Upper: MCF-7eYFP; lower: MCF-7GIRK1a. b Graphical representation of mean cellular velocities for different experimental groups. WT: MCF-7WT; eYFP: MCF-7eYFP; hG1a: MCF-7GIRK1a; hG1c: MCF-7GIRK1c; hG1d: MCF-7GIRK1d. The median value is represented by the black line within the box, box margins represent 75 and 25 % percentiles, whiskers indicate 90 and 10 % percentiles, values outside the 10–90 % interval are plotted individually as crosses. The grey line represents the mean value. The number of individual cells is given in parenthesis besides each box. Statistical significant differences between groups are indicated. hG1a, as well as hG1c, differ from both WT, as well as from eYFP, statistically significant at the p < 0.001 level. hG1d differs from both hG1a, as well as from hG1c, statistically significant at the p < 0.001 level. Kruskal-Wallis one way analysis on ranks was used for analysis of statistical significance subunits proved themselves to be inactive as ion channels (despite of expression at the protein level), and, in addition, entirely silenced both homo- and heterotetrameric GIRK complexes by acting as dominant negative constituents (see [12] for functional testing of splice variants and a subsumption of existing literature). Thereupon we suggest that