We used one BRAF/NRAS wild-type (MEWO) and two BRAF (V600E)-mutant (A375, A2058) melanoma cell lines. All the cell lines were purchased from ATCC and subjected to STR analysis at the Medical University of Vienna. We cultured the cells in DMEM supplemented with 10% FBS, 100 mg/ml streptomycin, 100 U/ml penicillin at 37°C, and 5% CO₂ in a humidified atmosphere.

Valproic acid (sodium salt; Sigma-Aldrich) was dissolved in distilled water at the concentration of 200 mM and stored at −20°C. From SAHA (Sigma-Aldrich), a stock solution was prepared at the concentration of 100 mM in DMSO and stored at −20°C. The mutant BRAF (V600E)-specific inhibitor, vemurafenib (PLX4032, Selleck Chemicals, Munich, Germany) was dissolved in DMSO (100 mM) and kept at −80°C. For Western blots, 1–2 × 10⁵ cells/well were seeded in 6-well plates. For Ca²⁺ signal measurements and immunofluorescence staining, 1–2 × 10⁴ cells/well were placed in an Imaging Chamber 8 Well (Ibidi). Twenty-four hours after seeding, medium was changed to fresh medium with the appropriate drug. For SAHA treatment, medium was changed daily. The final DMSO concentration did not exceed 0.01% in any of the experiments.

mRNA of several Ca²⁺ channels was isolated with TRIzol reagent (Life Technologies) from vemurafenib- and valproic acid-treated and control cells. For reverse transcription, RevertAid Reverse Transcriptase (Thermo Scientific) was used and amplification was performed with the Maxima SYBR Green master mix (Thermo Scientific) on an Applied Biosystems^® 7500 Real-Time PCR System. Primers used are described in Ref. (21).

Total protein from the cells was precipitated by the addition of 6% TCA, and equal amounts of protein were loaded on 10% acrylamide gels. Samples were analyzed by Western blot as described previously (10). The following primary antibodies were applied: mouse monoclonal anti-PMCA4b (JA3, recognizing the region between residues 1,156–1,180, which is specific to hPMCA4b (26), dilution 1:1,000), anti-pan PMCA (5F10, dilution 1:5,000) (27), rabbit polyclonal anti-PMCA1 (Affinity BioReagents, PA1-914, dilution 1:1,000), mouse monoclonal anti-SERCA2 (IID8, dilution 1:2,500, Sigma-Aldrich, S1439), mouse monoclonal anti-SERCA3 [PL/IM430, dilution 1:200 (10)], rabbit monoclonal anti-phospho-p44/42MAPK (ERK1/2) (cell Signaling, CST4370S, dilution 1:1,000), mouse monoclonal anti-ERK1/2 (MK1) (Santa Cruz, sc135900, dilution 1:500), rabbit polyclonal anti-beta-tubulin (Abcam, ab6046), rabbit polyclonal acetyl-Histone H3 (Lys 9/Lys 14) (cell Signaling, 9677). Subsequently HRP-conjugated anti-rabbit and anti-mouse secondary antibodies (Jackson ImmunoResearch, dilution 1:10,000) were used, and detection was performed with Pierce ECL Western Blotting Substrate (Thermo Scientific) and luminography. Densitometric analysis was performed by ImageJ software v1.42q.

Cells were treated with 1.0 µM SAHA or 2 mM valproic acid for 48 h. First, the cells were washed twice with PBS at 37°C; then, they were fixed with 4% paraformaldehyde for 15 min at room temperature. Immunostaining experiments were performed as described previously (9). As primary antibody, mouse monoclonal anti-PMCA4b (JA3, dilution: 1:200) antibody was used. As secondary antibody, Alexa Flour 488-conjugated anti-mouse IgG (Invitrogen) was applied. Images were taken by a Zeiss LSM500 confocal laser scanning microscope.

Cells were treated with 2.0 mM valproic acid for 48 h. In order to detect the intracellular Ca²⁺ level changes, the Fluo-4 green fluorescent Ca²⁺ indicator was used. Cells were kept in HBSS supplemented with 2 mM CaCl₂, 0.9 mM MgCl₂, and 20 mM HEPES pH7.4 during the entire experiment. First, cells were washed twice; then, they were incubated with 0.5 µM Fluo-4 AM (Molecular Probes, F14201) for 30 min at RT. After loading, cells were washed twice again. Ca²⁺ signals were induced by A23187 ionophore (2 µM). In order to specifically inhibit PMCA4 activity, cells were incubated with caloxin 1c2 (20 µM) for 10 min followed by the initiation of Ca²⁺ signaling with the addition of A23187. Time-lapse images were taken by Zeiss LSM500 and Olympus IX81 laser scanning confocal microscopes with a 60× (1.4) oil immersion objective. Z-resolution was set to 1 µm, and the images were taken every 1 s. The relative fluorescence intensities were calculated as F/F_o (where F_o was the average initial fluorescence), and the data were analyzed using the Prism4 software v4.01 (GraphPad Software).

A375 cells were seeded at 1,000 cells per well in 96-well plates (Greiner). The next day, the cell nuclei were stained with 0.1 µM Hoechst 33342 for 1 h. The cells were imaged automatically with ImageXpress Micro XL (Molecular Devices, Sunnyvale, CA, USA) high content screening system using a Nikon CFI Super Plan Fluor ELWD ADM 10× objective. The cells were kept at 37°C in a humidified atmosphere containing 5% CO₂ throughout the measurements. The fluorescence signals for Hoechst (447/60 nm) were acquired at 377/50 nm excitations. Four images per field with 2 × 2 binning were taken every 30 min for 24 h. Motility analysis was performed using the Multidimensional Motion Analysis module of MetaXpress High Content Image Acquisition and Analysis Software Version 5.3. The cell nuclei were analyzed in a size range of 9–15 µm. The maximum distance between two points on the trajectory of individual cells was set to 100 µm.

A modified Boyden chamber assay was previously described by Albini et al. (28). The assay was performed in a 48-well Boyden chamber (Neuro probe, Gaithersburg, MD, USA) on 10 μm-thick uncoated Nucleopore membranes (Whatman) with pore diameters of 8 µm. Fibronectin (100 µg/ml Millipore) was used as chemoattractant in the bottom chamber. Valproate pretreated (4 mM, 48 h) and untreated A375 melanoma cells were placed into the upper chamber at 2 × 10⁴ cells/well in DMEM containing 10% FCS. The 20 µM caloxin 1c2 was added as appropriate, and cells were allowed to migrate 6 h at 37°C. The filter was removed from the chamber, and the cells on the upper side of the filter were scraped off. The migrated cells on the lower side of the filter were fixed with methanol and stained with toluidine blue. Cell migration activity was quantified as the number of migrated cells on the lower side of the filter using a light microscope at ×200 magnification.

Viability was measured by the NucleoCounter NC-3000™ system (Chemometec). Cells were treated with 0.5 µM vemurafenib, 2 or 4 mM valproate, 1 µM SAHA alone, or in combination for 48 h in 6-well plates. After trypsinization, two dyes were added to the cells: acridine orange which stained all the cells in the solution and DAPI which was able to stain only the non-viable cells (Solution 13, Chemometec, 910-3013). 10 µl of each sample was loaded in an 8-well NC slide, and the counting of cells for the two cell populations was performed. Viability was calculated from the cell counts; viability = total cells − non-viable cells/total cells.

The number of cells in each cell cycle phase was determined based on their DNA content. Cells were treated with 0.5 µM vemurafenib, 2 or 4 mM valproate, 1 µM SAHA alone, or in combination for 48 h in 6-well plates. After trypsinization, cells were treated with lysis buffer and stained with DAPI for 5 min at 37°C. Then, the stabilization buffer was added and 10 µl of each sample was loaded onto an 8-well NC slide. Cellular fluorescence was quantified by the NucleoCounter NC-3000™ system (Chemometec).

It was previously found that inhibition of HDAC activity increased PMCA4b expression in breast and colon cancer cell lines (9, 10). More recently, we demonstrated that in BRAF-mutant melanoma cells BRAF inhibitor treatment had a similar effect (21). To investigate how HDACis affected PMCA expression in melanoma cells, we treated one BRAF wild-type (MEWO) and two BRAF-mutant (A375, A2058) cell lines either with increasing concentrations of SAHA or valproate (Figure 1A) or for different time intervals (Figure 1B). We found that both HDACis strongly increased the expression of PMCA4b in MEWO and A375 cells and to a lesser extent in A2058 cells. Interestingly, in response to HDACi treatment, the pan-PMCA antibody, 5F10, also recognized a band just above the one corresponding to PMCA4b. The upper band could be identified as PMCA1 by using the PMCA1 isoform-specific antibody NR1 (Figure 2A). Both SAHA and valproate upregulated PMCA expression already at the lowest doses tested, 0.5 µM and 1 mM, respectively; half-maximal response was reached at about 1 µM concentration for SAHA and 2 mM for valproate. According to the time curve in Figure 1B, PMCA4b reached maximum abundance after 48 h. At the longer time points, a moderate increase in PMCA4b expression was seen also in the control samples, in good accordance with previous findings, indicating that PMCA4b expression increases with cell confluency (29).

Combination of HDACis with targeted therapies for the treatment of melanoma is a possibility to overcome resistance (13). Since both HDAC and BRAF inhibition were able to increase PMCA4b expression in BRAF-mutant melanoma cells, we tested their combined effects. We treated two BRAF-mutant and one BRAF wild-type melanoma cell lines with the mutant BRAF inhibitor vemurafenib (0.5 µM) and with the HDACis SAHA (1 µM) and valproate (2 mM) alone, or in combination for 48 h (Figure 2). We found that in the BRAF-mutant A375 cell line vemurafenib had a stronger effect on PMCA4b abundance than the HDACis, and this was not further increased by combination with valproate or SAHA, while in A2058 cells HDACi treatment and BRAF inhibition had a similar effect and combination treatment with valproate had a slightly stronger effect than either drug alone (Figure 2B). In the BRAF wild-type MEWO cells, vemurafenib had no effect and combination treatments were similar to that of the HDACis alone, as expected. In good accordance with our previous results, PMCA1 abundance was not influenced by the BRAF inhibitor, vemurafenib, treatment; however, it was increased substantially by HDAC inhibition. Importantly, the staining with the antibody pair pERK/ERK clearly illustrates that HDACi treatment did not reduce ERK activation in any of the cell lines, while vemurafenib strongly decreased ERK phosphorylation in the BRAF-mutant cells (Figure 2C). In the BRAF wild-type MEWO cells, pERK level was slightly increased by the BRAF inhibitor treatment, which is in accordance with previous findings (30). This suggests that HDACi treatment increases PMCA abundance in an ERK-independent manner. To verify that HDACi treatment increased histone protein acetylation, we compared the acetylation level of histone 3 protein among control and HDACi-treated cells (Figure S3 in Supplementary Material). We found that both SAHA and valproate treatment increased H3 acetylation levels in all the cell lines in line with previous findings (31). In contrast, vemurafenib treatment did not enhance H3 acetylation in these cells.

We also tested the effect of HDAC inhibition alone or in combination with vemurafenib on PMCA4b expression at the mRNA level (Figure S1 in Supplementary Material). Cells were treated with valproate (2 mM) alone or in combination with vemurafenib (0.5 µM) for 48 h. In MEWO cells, PMCA4b expression was enhanced already after 8 h and further increased up to 48 h but then started to decline. In the A375 cells, the mRNA level of PMCA4b strongly increased after 24 h treatment and also started to decline after 48 h. In A2058 cells also PMCA4b mRNA level increased after 24 h but this increase was less substantial, except in the combination treatment, which induced strong increase.

The sarco/endoplasmic reticulum Ca²⁺ ATPases (SERCAs) are also important regulators of intracellular Ca²⁺ homeostasis, and the SERCA3 isoform was shown to be modified in cancer. Therefore, we analyzed the protein levels of SERCA2 and SERCA3, as well (Figure 2). We found that before treatment only SERCA2 was present in the melanoma cell lines and its protein level moderately decreased in response to HDACis in the BRAF-mutant cells (Figure 2D). SERCA3 expression, which is a known differentiation marker (32, 33), was induced by HDACi treatment in A2058 cells but was not detected in the other two cell lines.

To assess the effect of HDAC inhibition on Ca²⁺ channels—the “on” mechanisms of the Ca²⁺ signaling toolkit—we analyzed their mRNA expression before and after valproate treatment alone, or in combination with vemurafenib (Figure S2 in Supplementary Material). We found that in MEWO and A375 cells none of the treatments affected the expression of the inositol 1,4,5-trisphosphate receptor type 1–3 (IP3R1, IP3R2, IP3R3), the ORAI1, the stromal interaction molecule 1 and 2 (STIM1, STIM2), and the TRPM1. In A2058 cells, the mRNA level of IP3R3 was substantially increased by valproate, whereas TRPM1 expression was induced only by the combination treatment. Previously, we found that vemurafenib induced TRPM1 expression in this cell line suggesting that the increase was caused by the BRAF inhibitor in the combination treatment.

We treated one BRAF wild-type (MEWO) and one BRAF-mutant (A375) cell line with the HDACis, SAHA (1 µM), or valproate (2 mM), for 48 h and then used confocal microscopy to analyze the intracellular localization of PMCA4b (Figure 3). We found that HDAC inhibition strongly increased the abundance of PMCA4b in the plasma membrane of both cell lines.

We found previously that the vemurafenib-induced increase in PMCA4b abundance resulted in a faster Ca²⁺ clearance from the cells after stimulation. In order to assess whether HDACis had a similar effect, we measured intracellular Ca²⁺ signaling in A375 cells after valproate treatment with confocal imaging (Figure 4A). The Ca²⁺ signal was evoked by the Ca²⁺ ionophore A23187 that initiates Ca²⁺ uptake independently of the Ca²⁺ entry machinery. The Ca²⁺-sensitive fluorescent Fluo-4 indicator was used to monitor the changes in intracellular Ca²⁺ concentrations. For each measurement, we acquired the data with identical acquisition parameters and settings. We found that the Fluo-4 fluorescence signal declined to the basal level much faster after the peek in the valproate treated cells than in the control cells. Furthermore, this effect could be reverted to a significant extent by the addition of the PMCA4-specific inhibitor, caloxin 1c2, which inhibited the decay phase of the Ca²⁺ signal (Figures 4B,C). These results show that PMCA4b is primarily responsible for the increased Ca²⁺ clearance after HDACi treatment in these cells.

Histone deacetylase inhibition was previously shown to induce apoptosis and to decrease proliferation in a variety of cancer cells. Furthermore, the combination of BRAF and HDACis increased cell death and even overcame resistance in BRAF-mutant melanoma cells (13). We treated the three melanoma cell lines (MEWO, A2058, and A375) with HDAC and BRAF inhibitors alone and in combination for 48 h and tested their viability (Figure 5). We found that none of the treatments affected significantly the viability of the BRAF wild-type MEWO cells, and BRAF-mutant A2058 cells (Figures 5A1,C1). In the A375 cells, vemurafenib alone or in combination with SAHA decreased the cell viability slightly, but significantly (Figure 5B1).

We also analyzed the distribution of the cells among the cell cycle phases after the various treatments (Figure 5). In these measurements, we distinguished four cell cycle phases; the resting G0/G1 phase, the synthesis S phase, the G2/M phase with cells having two sets of paired chromosomes, and the sub-G1 phase with dying cells. We found that the effect of the treatments was cell line dependent. In A375 cells, as expected, proliferation strongly decreased in the presence of vemurafenib (G2/M phase cells) (Figure 5C3), while HDACis alone had no effect. Vemurafenib induced cell death (sub-G1) in approximately 10% of the cells (Figure 5B3). This effect was further increased by combination with valproate but not with SAHA. HDACis alone only slightly increased cell death compared to control in this cell line. In A2058 cells, SAHA decreased significantly the number of cells in the G2/M phase, valproate alone did so to a lesser extent (Figure 5C3), but in combination with vemurafenib the effect was stronger. Treatment with SAHA alone and also in combination with vemurafenib increased the ratio of dying cells in this cell model (Figure 5C2). MEWO cells responded exclusively to SAHA treatment with decreased proliferation and none of the treatments induced cell death in these cells (Figures 5A2,A3). Based on these data, we conclude that HDACis alone decreased only modestly the viability and proliferation of melanoma cells, while combination treatments had a more substantial effect, although in a cell-type specific manner.

We found previously that increased expression of PMCA4b in A375 cells resulted in decreased migratory activity. We assessed the effect of HDAC inhibition on both the random and directional migration of A375 cells (Figure 6). In all the experiments, the cells were treated with valproate (4 mM) for 48 h before the migration analysis. Random migration was recorded by using automated fluorescence microscopy (ImageXpress Micro XL) for 24 h. Representative trajectories of single cells after 48 h of valproate treatment clearly show that valproate strongly inhibited the migration of A375 cells (Figure 6A1). The velocity of valproate-treated cells decreased significantly compared to the untreated control cells (Figure 6A2). This effect was similar to that described previously for vemurafenib-treated A375 cells (21).

Directed migration of A375 cells was assessed using modified Boyden chamber assay (Figure 6B), in which fibronectin was used as a chemoattractant. Cells were allowed to migrate through the pores of the Boyden Chamber membrane for 6 h toward the fibronectin coat. We observed a sharp decrease in the directional movement of A375 cells pretreated with valproate, compared to that of the untreated cells (Figure 6B1). A half-maximal response of cells could be achieved at approximately 2 mM valproate treatment (Figure 6B2), which was comparable to that of the PMCA4b protein expression response to valproate under similar conditions (Figure 1A, middle panels). To evaluate the role of PMCA in this effect, we added the PMCA4-specific inhibitor, caloxin 1c2, to the cells immediately before the assay. Inhibition of PMCA4 resulted in a modest but significant increase in the directional migration activity in both the control and valproate-treated cells (Figure 6C) suggesting that PMCA4 activity is involved in the control of A375 cell motility.