|Home | About | Journals | Submit | Contact Us | Français|
Inhibition of MAP kinase pathways by selective BRAF inhibitors, such as vemurafenib and dabrafenib, have evolved as key therapies of BRAF-mutated melanoma. However, tumor relapse and therapy resistance have remained as major problems, which may be addressed by combination with other pathway inhibitors. Here we identified the potassium channel inhibitor TRAM-34 as highly effective in combination with vemurafenib. Thus apoptosis was significantly enhanced and cell viability was decreased. The combination vemurafenib/TRAM-34 was also effective in vemurafenib-resistant cells, suggesting that acquired resistance may be overcome. Vemurafenib decreased ERK phosphorylation, suppressed antiapoptotic Mcl-1 and enhanced proapoptotic Puma and Bim. The combination resulted in enhancement of proapoptotic pathways as caspase-3 and loss of mitochondrial membrane potential. Indicating a special mechanism of vemurafenib-induced apoptosis, we found strong enhancement of intracellular ROS levels already at 1h of treatment. The critical role of ROS was demonstrated by the antioxidant vitamin E (α-tocopherol), which decreased intracellular ROS as well as apoptosis. Also caspase activation and loss of mitochondrial membrane potential were suppressed, proving ROS as an upstream effect. Thus ROS represents an initial and independent apoptosis pathway in melanoma cells that is of particular importance for vemurafenib and its combination with TRAM-34.
For decades, melanoma remained a deadly disease lacking an effective therapy in the metastatic stage. The situation changed with the identification of activating BRAF mutations in about one-half of cutaneous melanomas.1 Selective BRAF inhibitors (vemurafenib, dabrafenib) as well as inhibitors of the downstream MAP kinase MEK (trametinib, cobimetinib) have now been approved for therapy of BRAF-mutated melanomas.2, 3 Especially, combinations of BRAF and MEK inhibitors have shown substantial benefit for patients with regard to significant prolongation of overall survival.4, 5 The situation may even further improve with combinations of immune checkpoint inhibitors, such as anti-CTLA4 and anti-PD1.6 Nevertheless, tumor relapse and therapy resistance are still frequent and often follow within only a few months.7 Thus, although present clinical results are highly encouraging, further improvements and particularly new, even more effective combinations will be needed to finally overcome melanoma mortality.
Different cellular mechanisms may contribute to therapy resistance of cancer,8 of which apoptosis deficiency may be considered as the major cause. This is explained by the fact that elimination of cancer cells through proapoptotic programs represents the common end path of most anticancer strategies. For example, different chemotherapeutic drugs cause cellular or DNA damage, which induces cell-intrinsic proapoptotic pathways; and also BRAF inhibitors induce apoptosis or sensitize for proapoptotic programs.9, 10, 11
Two major apoptosis pathways have been described that are initiated by the interaction of death ligands with death receptors (extrinsic pathway) and by cellular or DNA damage (intrinsic pathway), respectively. Downstream, caspase cascades are activated consisting of initiator and effector caspases, such as caspase-3.12 In several tumor cells, including melanoma cells, efficient apoptosis induction needs proapoptotic mitochondrial activation.9, 13 Mitochondrial contribution to apoptosis is closely related to a loss of mitochondrial membrane potential (MMP) and release of proapoptotic, mitochondrial factors. This pathway is critically controlled by the family of proapoptotic and antiapoptotic Bcl-2 proteins.14
Besides the well-established pathways, there is increasing evidence for alternative and possibly supplementary pathways. Thus reactive oxygen species (ROS) may initiate independent apoptosis programs.15, 16 In melanoma cells, we have previously found enhanced ROS levels in response to an iron-containing nucleoside analog and to the phospoinositol-3 kinase inhibitor wortmannin.17, 18 Elevation of ROS was proven as critical for apoptosis induction, as apoptosis was prevented by the ROS scavenger α-tocopherol.
Membrane ion channels, which serve fundamental cellular functions, represent additional, promising tumor targets. Thus Ca2+-dependent potassium channels contribute to cytoplasma membrane hyperpolarization facilitating Ca2+ entry, a prerequisite for cell proliferation. Overexpression of the family member KCa3.1 (IK1) was related to aberrant cell proliferation of different tumor cells, including melanoma.19, 20 TRAM-34 has been established as a selective KCa3.1 inhibitor, which avoids side effects of some other channel inhibitors, such as hepatotoxicity.21, 22 We have previously shown that TRAM-34 can act as apoptosis enhancer in melanoma cells when combined with the death ligand TRAIL (TNF-related apoptosis-inducing ligand).23 Here we demonstrate that apoptosis induction by vemurafenib in melanoma cells is strongly enhanced by TRAM-34. Unraveling the pathway identified ROS generation as a general factor in vemurafenib-induced apoptosis, which appears of particular importance for the combination with TRAM-34.
For identifying suitable combination partners, which may enhance the antitumor activities of vemurafenib, a screening with several pathway inhibitors and apoptosis agonists was performed (data not shown). Synergistic enhancement of apoptosis in the combination of vemurafenib and TRAIL, which had been described previously,11 served as positive control. Of the different effectors tested, TRAM-34, an inhibitor of the potassium channel KCa3.1, appeared of particular value. Although it did not trigger apoptosis by itself, TRAM-34 strongly enhanced and accelerated vemurafenib-induced proapoptotic effects. Vemurafenib was applied in a standard concentration of 30μM as well as in two reduced concentrations of 3 and 10μM. TRAM-34 further enhanced apoptosis at all tested concentrations. Thus apoptotic rates were stepwise increased when raising the concentration of TRAM-34 from 0 to 25 to 50μM. For example, at 48h, the numbers of apoptotic cells increased from 12% to 23% to 33% for 3μM vemurafenib, from 18% to 33% to 41% for 10μM vemurafenib and from 49% to 79% to 86% for 30μM vemurafenib (Figure 1a). These effects were clearly more than additive, particularly as there was only limited proapoptotic effect of TRAM-34 alone (Figure 1b).
In order to generalize the findings on enhanced apoptosis, two more melanoma cell lines (Mel-HO and Mel-2a) were investigated. All three cell lines carry BRAF mutations, but whereas A-375 and Mel-HO were responsive to vemurafenib single treatment, Mel-2a revealed pronounced resistance. Nevertheless, comparable enhancement of apoptosis was seen in all three melanoma cell lines for the combination of 30μM vemurafenib+50μM TRAM-34 versus vemurafenib alone. Thus, in Mel-HO, apoptosis rates increased at 48h from 27% to 62% by the combination with TRAM-34. In vemurafenib-resistant Mel-2a, apoptosis rates raised from <5% at 48h up to 40±2% (Figure 1c). This suggests that the combination effect reported here may be representative for melanoma cells, even in case of intrinsic vemurafenib resistance.
To further prove the issue of vemurafenib resistance, another model was established in A-375. Cells were cultured with successively increasing concentrations of vemurafenib up to 10μM within a period of 4 weeks. The selected cells showed continuous growth despite vemurafenib and revealed no apoptotic response to 10μM. Nevertheless, these cells were significantly responsive to the combined treatment of vemurafenib/TRAM-34. In response to 10μM vemurafenib+TRAM-34, apoptosis was induced up to 19% and 58% at 24 and 48h, respectively. Similarly, apoptosis induced by 30μM vemurafenib was enhanced (Figure 1d). Thus both intrinsic resistance (Mel-2a) and induced vemurafenib resistance (A-375-VemR) were overcome in these models.
Enhanced apoptosis went along with a reciprocal loss of cell viability, as determined by calcein staining and subsequent flow cytometry. At all vemurafenib concentrations used (3, 10, 30μM), TRAM-34 further decreased cell viability. Thus, at 48h, the numbers of viable cells in response to vemurafenib stepwise further decreased when raising TRAM-34 concentration from 0 to 25 to 50μM. Values decreased from 73% to 68% to 39% (Vem 3μM), from 69% to 42% to 17% (Vem 10μM) and from 31% to 2% to 0.5% (Vem 30μM) (Figure 2a). Also concerning cell viability, the combination effects were far more than additive (Figure 2b).
In order to distinguish between direct cytotoxic effects (cell lysis) and apoptosis induction, release of lactate dehydrogenase (LDH) into the cell culture supernatant was quantified by an enzymatic assay at 24 and 48h. In contrast to induced apoptosis, LDH activity in cell culture supernatant remained at a low level, thus indicating that cells were not lysed up to 24h. Only at 48h of combination treatment, LDH levels were slightly enhanced by 2.3-fold as compared to control cells, which may be explained by secondary cell lysis following apoptosis in cell culture (Figure 2c). Complete cell rounding and detachment, characteristic of apoptotic cells, were seen already at 24h of combination treatment (Figure 2d), and monitoring attachment of cells in real time (xCELLigence system) confirmed that both single treatments resulted in only intermediate effects, while the combination of TRAM-34 and vemurafenib resulted in complete cell detachment (Figure 2e).
Thus apoptosis appeared as the dominant effect of the combination vemurafenib/TRAM-34 demanding an unravelling of the pathways involved. As expected, vemurafenib almost completely abolished ERK phosphorylation. Total ERK was also downregulated at 48h of combination treatment, which may be explained by strong apoptosis induction at this time. In contrast, TRAM-34 remained without effect on ERK. For 30μM vemurafenib applied, only little rebound ERK activation was seen at 48h, indicating a largely sustained effect (Figure 3a).
BRAF inhibition in A-375 was correlated to downregulation of the antiapoptotic Bcl-2-related protein Mcl-1 as well as to upregulation of the proapoptotic BH3 domain-only proteins Puma and Bim (24kDa, isoform EL), as also shown previously.10 The finding that Puma was downregulated again after 48h is explained by strongly induced apoptosis, which may lead to degradation of many proteins. Of the other apoptosis regulators investigated by Western blotting, antiapoptotic Bcl-2 and the proapoptotic transcription factor p53 were not regulated by TRAM-34 or vemurafenib. But antiapoptotic XIAP as well as proapoptotic Bax were downregulated at 48h of combination treatment in course of massive apoptosis induction (Figure 3a). Thus Bcl-2 proteins appear of particular note for vemurafenib activity, but there was no indication that altered expression of Bcl-2 proteins was also critical for the combination effects seen here with TRAM-34.
Significant activation of the major effector caspase-3 upon combination treatment (24 and 48h) was seen by its characteristic cleavage products of 17 and 15kDa, suggesting a contribution of caspase-mediated pathways. Nevertheless, caspase activation appeared as not complete, seen by still high levels of procaspase-3 (Figure 3a). The role of caspases was further addressed by using the pancaspase inhibitor QVD-OPh. Pretreatment of A-375 with QVD-OPh (10μM) for 1h resulted in significant but not complete reduction of apoptosis by vemurafenib alone (9%→3%) as well as by the combination (47%→13%). Similarly, cell viability was partly restored (16%→33% Figure 3b).
Proapoptotic and antiapoptotic Bcl-2 proteins control mitochondrial apoptosis pathways, for which loss of the MMP represents a critical step. Whereas in A-375, TRAM-34 and vemurafenib as single treatments remained without significant effect on MMP at 4 and 24h, the combination resulted in a complete loss of MMP at 24h (>90%), indicating full activation of mitochondrial apoptosis pathways at this time (Figure 3c). Similarly, 60% cells with low MMP were obtained in Mel-2a at 24h. Only here, cells also showed response at 4h (Figure 3d).
The decisive roles of Bcl-2 proteins and mitochondrial pathways were further confirmed by the results of Bcl-2 overexpression. Thus Vem/TRAM-induced apoptosis was completely prevented in a stably Bcl-2-overexpressing A-375 cell line (A375-Bcl-2), as compared with a mock-transfected cell clone (A375-pIRES; Figure 3e). Both cell clones had been previously established and characterized.13
Despite the obvious significance of lost MMP at 24h, it should be considered at least in A-375 as a delayed effect, as less evident at 4h (Figure 3c). Thus we looked for more initial effects. By this, we identified the generation of ROS as an immediate early event, already visible at 1h of treatment. Thus ROS levels were consistently upregulated by vemurafenib single treatment raising from 27% of cells with high ROS at 1h to 49% cells at 24h. This effect further increased with the combination treatment resulting in 48% cells with high ROS at 1h and 72% cells at 4h (Figure 4a). Thus the whole cell population shifted toward higher ROS levels upon treatment (Figure 4b). The reproducible finding that ROS levels upon combination treatment were again downregulated at 24h underlined the transient appearance of ROS, which was thus not a consequence of induced apoptosis. In parallel, we obtained ROS generation in Mel-2a by vemurafenib at 2h, which was, with 19±2%, however, less than in vemurafenib-sensitive A-375. Nevertheless, upon combination treatment, ROS levels were raised up to 42±4% (Figure 4c).
To prove the significance of ROS production, we performed experiments with the ROS scavenger α-tocopherol (VitE). In fact, pretreating A-375 cells for 1h with 1mM VitE reduced ROS levels upon vemurafenib and combination treatment by almost twofold (Figure 5a). This resulted in a concomitant twofold reduction of apoptosis (Figure 5b) and largely recovered cell viability (Figure 5c). Thus ROS produced by vemurafenib appeared as largely responsible for enhanced apoptosis upon combination treatment. Further demonstrating its signaling effect, caspase-3 activation by vemurafenib/TRAM was completely abolished by vitamin E, just as it was abolished by the pancaspase inhibitor QVD-OPh (Figure 6a). Also MMP was largely restored by vitamin E, locating also loss of MMP as downstream of ROS in the signaling cascade (Figure 6b).
The development of selective inhibitors for BRAF and MEK as well as immune checkpoint inhibitors such as anti-CTLA4 and anti-PD1 have revolutionized melanoma therapy in the past years leading to significant improvements for melanoma patients.2, 3, 6, 7 To further overcome therapy resistance and tumor relapse, combination therapies are of principle value as actually being demonstrated by combinations of BRAF and MEK inhibitors.4, 5 Also, combinations of targeted therapy and immune checkpoint inhibitors are considered.24, 25 Besides, many other combinations have been tested in melanoma experimental models. Enhanced efficiency was described for combinations of vemurafenib with different strategies, for example, with cyclin-dependent kinase 4 and 6 inhibitor,26 with inhibitors of epigenetic regulators27 or radiotherapy.28 Often toxicity is a limiting factor for in vivo use. Given a high number of possible targets in cancer cells, the repertoire of, in principle, suitable combination partners will grow tremendously in near future, giving much hope for further improvement of cancer therapies.
Inhibitors of membrane channels expressed in cancer cells represent additional, promising candidates that are still awaiting their full exploration. Potassium channels have critical roles in a wide variety of physiological processes. The group of Ca2+-dependent K+ channels contributes to cytoplasmic membrane hyperpolarization thus facilitating Ca2+ entry, a prerequisite for cell proliferation.19 The family member KCa3.1 is inhibited by clotrimazole, which itself is, however, not suitable for systemic application as hepatotoxicity results from non-specific effects on cytochrome P450. In contrast, the clotrimazole analogs TRAM-34 and ICA-17043 are more selective for KCa3.1 and lack P450-inhibitory activity.21, 29 Specificity of TRAM-34 was also proven in KCa3.1-negative HEK-293 cells, which did not respond, whereas responsiveness was recovered in KCa3.1-transfected cells.23
Knockout mice for KCa3.1 are viable and ICA-17043 was well tolerated in a clinical trial, showing almost no adverse effects.29, 30 Thus KCa3.1 may not be required for steady-state physiological balance but may be upregulated in disease situations. It may have critical roles in aberrant cell proliferation and cytokine synthesis in cancer and autoimmune disease.29 Blockade of KCa3.1 frequently results in G0/G1 cell cycle arrest and inhibitory effects on cytokine synthesis, which may be caused by suppressed calcium inflow.31, 32 KCa3.1 mRNA is overexpressed in different cancer cell lines, for example, of glioblastoma, pancreatic carcinoma, breast carcinoma and prostate carcinoma.20, 29 In a series of melanoma cell lines, mRNA and protein expression was shown, suggesting KCa3.1 expression as characteristic for melanoma.23 Also in nude mice, KCa3.1 inhibition by clotrimazole inhibited melanoma growth.31
Antitumor effects of KCa3.1 inhibition were correlated to suppression of cell proliferation and migration in breast cancer, endometrial carcinoma and hepatocellular carcinoma cells.33, 34, 35 In combinations, it increased the sensitivity of glioblastoma cells for radiotherapy36 and of glioma cells for temozolomide.37 In most instances, KCa3.1 blockers were cytostatic rather than cytotoxic or proapoptotic,33, 38 which may limit their use as monotherapy but may encourage for combination with proapoptotic agents. This was also seen in melanoma cells, namely, KCa3.1 inhibition by TRAM-34 decreased cell proliferation without directly affecting apoptosis, but it strongly sensitized melanoma cell lines for TRAIL-induced apoptosis.23
Here we prove TRAM-34 as a suitable combination partner for vemurafenib. Upon combination, apoptosis was strongly enhanced and even vemurafenib-resistant cells could be targeted. Cell viability and overall cell proliferation were completely abolished, whereas direct cytotoxc effects (cell lysis) remained at a low level. Concerning the proapoptotic mechanisms, activation of caspase-3, the mitochondrial pathway and elevation of ROS levels were shown. Enhanced apoptosis and loss of cell viability were evident not only for the standard concentration of 30μM vemurafenib but also for reduced concentrations, not indicating a threshold for mutual enhancement. Also, the combination effect appeared as representative for melanoma cells as seen in 3/3 BRAF-mutated melanoma cell lines.
Whereas caspase cascades and proapoptotic mitochondrial activation represent standard pathways in apoptosis control, the regulation of apoptosis by ROS is less understood. The significance of ROS for apoptosis induction has been proven by us previously,17, 18 and increased ROS levels in response to vemurafenib were also reported.39, 40 Here we prove ROS as a major factor in vemurafenib-induced apoptosis, in particular, in its combination with TRAM-34. This was shown by the ROS scavenger α-tocopherol, which strongly diminished apoptosis induction and partly restored cell viability. ROS was also upstream of other signaling steps, as it appeared already within 1h, and α-tocopherol blocked both caspase activation and loss of MMP. As a possible origin of ROS, the endoplasmatic reticulum (ER) may be considered, as ER stress has already been linked to vemurafenib-induced apoptosis.10
In conclusion, these data shed more light on a frequently used but not completely understood melanoma therapy. ROS represents an initial and independent apoptosis pathway that is of particular importance for vemurafenib and its combination with TRAM-34 in melanoma cells. Better understanding may be helpful for further improving targeted therapy in melanoma.
Three BRAF mutated human melanoma cell lines were used in the present study: A-375,41 Mel-HO42 and Mel-2a.43 A-375 subclones had been established previously by stable transfection of a pIRES-Bcl-2 plasmid for Bcl-2 overexpression (A375-Bcl-2) or by transfection of an empty pIRES plasmid (A375-pIRES, mock control).13 Cells were cultured at 37°C, 5% CO2 with DMEM (4.5g/l glucose; GIBCO, Invitrogen, Karlsruhe, Germany) supplemented with 10% FCS and antibiotics (Biochrom, Berlin, Germany). For analyses, 5 × 104 and 2 × 105 cells were seeded in 24- and 6-well plates, respectively.
The selective BRAF (V600E) inhibitor vemurafenib/PLX4032 (Selleck Chemicals, Houston, TX, USA) was used at a standard concentration of 30μM. This concentration theoretically corresponds to doses applied in patients (960mg/12h; MW=490g/mol). In several experiments, also reduced concentrations (3 and 10μM) were included. TRAM-34, a selective inhibitor of the Ca2+-dependent potassium channel KCa3.1 (IK1), has been kindly provided by Dr. Heike Wulff, University of California, Department of Pharmacology, Davis, CA, USA.21 According to a previous study in melanoma cells,23 TRAM-34 was used at a standard concentration of 50μM. For caspase inhibition, the pan-caspase inhibitor QVD-OPh (Sigma-Aldrich, Taufkirchen, Germany) was used at 10μM, 1h before other treatments started.
Quantification of apoptosis was performed by cell cycle analyses.44 Trypsinized cells were lysed in hypotonic buffer, and isolated nuclei were stained for 1h with 40mg/ml propidium iodide (Sigma-Aldrich). Cell fractions in G1, G2 and S-phase as well as sub-G1 cells were quantified by flow cytometry at FL3A with a FACS Calibur (BD Bioscience, Bedford, MA, USA). Owing to the washing out of small DNA fragments, nuclei with less DNA than in G1 (sub-G1) correspond to apoptotic cells with fragmented DNA.
Direct cytotoxicity (cell lysis) was determined by quantifying released LDH activity in cell supernatants by an LDH activity enzymatic assay (Roche Diagnostics, Penzberg, Germany), which was quantified in an ELISA reader later.
Cell viability was determined by staining cells with calcein-AM (PromoCell, Heidelberg, Germany), which is converted in viable cells to green-fluorescent calcein by intracellular esterases. Cells, grown and treated in 24-well plates, were trypsinized and stained with 2.5μg/ml calcein-AM at 37°C for 1h. Labeled cells were washed with PBS and measured by flow cytometry (FL2H).
For monitoring cell growth and attachment in real time, the xCELLigence system (OMNI Life Science, Bremen, Germany) was applied. Melanoma cells were seeded in special 96-well E-plates (50000 cells per well), which have microelectrodes integrated in the bottom of the wells. Continuous measurement of electric resistance corresponds to attached cell numbers.
MMP (Δψm) was determined by staining cells with the fluorescent dye TMRM+ (Sigma-Aldrich). Cells, grown and treated in 24-well plates, were harvested by trypsinization and stained for 20min at 37°C with 1μM TMRM+. After washing two times with PBS, cells were measured by flow cytometry (FL2H).
For determination of intracellular ROS, attached cells in 24-well plates were preincubated for 1h with the fluorescent dye H2DCFDA (D-399, Thermo Fisher Scientific, Hennigsdorf, Germany, 10μM) before starting treatment with effectors. After treatment, cells were trypsinized, washed with PBS and analyzed by flow cytometry (FL1H). As positive control, cells were treated with H2O2 (1mM, 1h). For ROS scavenging, cells were pretreated for 1h with 1mM α-tocopherol (vitamin E, Fluka, Steinheim, Germany).
For Western blotting, total protein extracts were obtained by cell lysis in 150mM NaCl, 1mM EDTA, 2mM PMSF, 1mM leupeptin, 1mM pepstatin, 0.5% SDS, 0.5% NP-40 and 10mM Tris-HCl, pH 7.5. Western blotting on nitrocellulose membranes was performed as described previously.45 Primary antibodies were: Cleaved caspase-3 (9664, rabbit, 1:1000, Cell Signaling, Danvers, MA, USA), caspase-3 proform (9662, rabbit, 1:1000, Cell Signaling), Mcl-1 (sc-12756, mouse, 1:200, Santa Cruz Biotech, Dallas, TX, USA), Bcl-2 (sc-492, rabbit; 1:200, Santa Cruz Biotech), Bax (sc-20067, mouse, 1:200, Santa Cruz Biotech), GAPDH (sc-32233, mouse, 1:1000, Santa Cruz Biotech), Puma (ab33906, rabbit, 1:1000, Abcam, Cambridge, UK), XIAP (no. 2042, rabbit, 1:1000, Cell Signaling), p53 (sc-126, rabbit, 1:500, Santa Cruz Biotech), ERK (no. 4695, rabbit, 1:1,000, Cell Signaling), pERK (no. 9101, rabbit, 1:1,000, Cell Signaling), and Bim (no. 559685, rabbit, 1:200, BD Biosciences, Heidelberg, Gemany). Secondary antibodies were: peroxidase-labeled goat anti-rabbit and goat anti-mouse (Dako, Hamburg, Germany; 1:5000).
All assays were carried out in triplicate determinations and at least two independent experiments were performed. Also, Western blotting data were verified by at least two independent series of cellular extracts. Statistical significance was proven by Student's t-test (normal distribution), and P-values of <0.05 were considered as statistically significant.
The work was supported by the research fund of the Charité-Universitätsmedizin Berlin as well as in parts by the German Cancer Aid (Deutsche Krebshilfe, Melanomverbund; 10-8008, TP7). We thank Heike Wulff, PhD, University of California, Department of Pharmacology, Davis, CA, USA and Roland Schönherr, PhD, Department of Biophysics, Jena University Hospital, Jena, Germany for supplying us with TRAM-34.
Edited by A Finazzi-Agro'
FK has got an active Consulting Contract with Roche Pharma. The other authors declare no conflict of interest.