Mefloquine induces ER stress and apoptosis in BRAFi-resistant A375-BRAF^V600E/NRAS^Q61K malignant melanoma cells targeting intracranial tumors in a bioluminescent murine model

Abstract

Molecularly targeted therapeutics have revolutionized the treatment of BRAF^V600E-driven malignant melanoma, but the rapid development of resistance to BRAF kinase inhibitors (BRAFi) presents a significant obstacle. The use of clinical antimalarials for the investigational treatment of malignant melanoma has shown only moderate promise, attributed mostly to inhibition of lysosomal-autophagic adaptations of cancer cells, but identification of specific antimalarials displaying single-agent antimelanoma activity has remained elusive. Here, we have screened a focused library of clinically used artemisinin-combination therapeutic (ACT) antimalarials for the apoptotic elimination of cultured malignant melanoma cell lines, also examining feasibility of overcoming BRAFi-resistance comparing isogenic melanoma cells that differ only by NRAS mutational status (BRAFi-sensitive A375-BRAF^V600E/NRAS^Q61 versus BRAFi-resistant A375-BRAF^V600E/NRAS^Q61K). Among ACT antimalarials tested, mefloquine (MQ) was the only apoptogenic agent causing melanoma cell death at low micromolar concentrations. Comparative gene expression-array analysis (A375-BRAF^V600E/NRAS^Q61 versus A375-BRAF^V600E/NRAS^Q61K) revealed that MQ is a dual inducer of endoplasmic reticulum (ER) and redox stress responses that precede MQ-induced loss of viability. ER-tracker^™ DPX fluorescence imaging and electron microscopy indicated ER swelling, accompanied by rapid induction of ER stress signaling (phospho-eIF2α, XBP-1s, ATF4). Fluo-4 AM-fluorescence indicated the occurrence of cytosolic calcium overload observable within seconds of MQ exposure. In a bioluminescent murine model employing intracranial injection of A375-Luc2 (BRAF^V600E/NRAS^Q61K) cells, an oral MQ regimen efficiently antagonized brain tumor growth. Taken together, these data suggest that the clinical antimalarial MQ may be a valid candidate for drug repurposing aiming at chemotherapeutic elimination of malignant melanoma cells, even if metastasized to the brain and BRAFi-resistant.

1. Introduction

Metastatic melanoma is a malignant tumor originating from neural crest-derived melanocytes causing the majority of skin cancer-related deaths.¹ Molecularly targeted therapeutics have revolutionized the treatment of BRAF^V600E-driven malignant melanoma, but BRAF inhibitor (BRAFi) resistance compromises long-term survivorship of patients.^2–4

Importantly, in treatment-naïve melanoma tumors, NRAS mutations rarely occur together with either BRAF or PTEN mutations, a phenomenon referred to as oncogene exclusion.^5–8 In contrast, during the course of BRAF^V600E-targeted BRAFi-treatment, emergence of resistance often involves NRAS mutations [predominantly in codon 61 (Q61R, Q61K) observable in approximately 20% of patients], a unique co-occurrence of BRAF and NRAS mutations facilitated and selected for by BRAFi-based therapy.^{3, 9}

In addition to BRAFi-resistance compromising clinical outcomes in melanoma patients, an urgent need exists for the identification of molecular therapeutics that can target these drug-resistant cells at metastatic sites including the brain.^10–12 Indeed, brain metastases are a frequent complication in patients with metastatic melanoma. Up to 45% of patients develop clinically documented brain tumors during their lifetime, and it has been reported that brain lesions are observable in more than 70 % of melanoma patients at autopsy.^{10, 12–15}

The use of clinical antimalarials for the investigational treatment of malignant melanoma has shown moderate promise, attributed mostly to inhibition of lysosomal-autophagic adaptations of cancer cells.^16–18 Artemisinin-combination-therapeutics (ACT) are now standard of care medications used as antimalarials worldwide.¹⁹ Beyond clinically used artemisinin-endoperoxides (including dihydroartemisinin acting through induction of oxidative stress), numerous ACT agents of the amino-quinoline class (such as amodiaquine, primaquine, chloroquine, piperaquine, mefloquine, among others) have shown autophagy-directed inhibitory activity that might be harnessed for experimental melanoma therapy, particularly in the context of BRAFi-resistance that might render these cells more sensitive to autophagy inhibition.^{16–18, 20–24} However, identification of specific antimalarials displaying single agent activity targeting BRAFi-resistant melanoma has remained challenging.

Anti-melanoma activity of the amino-quinoline antimalarial chloroquine has been investigated in preclinical models focusing on therapeutic induction of apoptosis and autophagic blockade but clinical efficacy benefitting melanoma patients remains unsatisfactory.^{22, 24, 25} Interestingly, the combinatorial use of temozolomide and chloroquine targeting BRAFi-resistant post-treatment melanoma cells (from patients who failed treatment with either vemurafenib or dabrafenib) has recently shown promise in murine models attributed to induction of inflammasome-dependent pyroptotic cell death facilitated by chloroquine-dependent inhibition of autophagy.²⁴

Here, we have screened a focused library of clinical ACT antimalarials [including chloroquine, amodiaquine, piperaquine, lumefantrine, and mefloquine (MQ)] for the apoptotic elimination of cultured malignant melanoma cell lines focusing on BRAFi-resistance. To this end, we have employed a stringent genetic model of BRAFi-resistance using isogenic melanoma cells that differ only by NRAS mutational status (BRAFi-sensitive A375-BRAF^V600E/NRAS^Q61 vs. BRAFi-resistant A375-BRAF^V600E/NRAS^Q61K) as explored and validated by us recently.^{3, 26} Here, we demonstrate feasibility of using the clinical antimalarial mefloquine for the therapeutic elimination of BRAFi-resistant malignant melanoma cells accompanied by early induction of ER stress and apoptosis targeting brain metastases in a bioluminescent murine model of the human intracranial disease stage.

2. MATERIALS AND METHODS

2.1. Chemicals

All chemicals including clinical antimalarials were purchased from Sigma Chemical Co (St. Louis, MO, USA). Mefloquine was the (±) -racemic mixture (as used in clinical antimalarial formulations).

2.2. Cell culture

Human malignant melanoma cells were purchased as follows: (BRAF^V600E/NRAS wildtype) LOX-IMVI (SCC201, Millipore Sigma), G361 (CRL-1424, ATCC), SK-MEL-28 (HTB-72, ATCC).^{3, 27–32} Moreover, vemurafenib-sensitive BRAF^V600E-mutated A375-Luc2 (CRL-1619-LUC2) and vemurafenib-resistant A375-Luc2 (isogenic variant containing both BRAF^V600E and NRAS^Q61K mutations; CRL-1619IG-2-LUC2) were purchased from ATCC.^{3, 27–32} For generation of CRL-1619IG-2-LUC2, the parental BRAF^V600E/NRAS^Q61 human malignant melanoma cells were CRISPR/Cas9 gene edited, generating an isogenic variant containing both BRAF^V600E and mutated NRAS^Q61K, followed by transduction using a lentiviral vector encoding firefly luciferase (LUC2) under control of the EF-1alpha promoter; differential sensitivity to vemurafenib treatment as a function of NRAS mutational status was confirmed as published before.³ All cells were maintained as published recently.^{33, 34}

2.3. Flow Cytometric Analysis of Cell Viability

After 24 h treatment with clinical antimalarials [15 μM; amodiaquine, chloroquine, lumefantrine, piperaquine and mefloquine (MQ)], viability of melanoma cells was assessed by annexinV-FITC/propidium iodide (PI) dual staining with flow cytometric analysis using an apoptosis detection kit according to manufacturer’s specifications (APOAF, Sigma Aldrich).^{33, 35}

2.4. Caspase-3 activation assay

MQ-induced (15 μM; 24 h) caspase-3 activation was examined by flow cytometric analysis using a cleaved/activated caspase-3 (asp 175) antibody (Alexa Fluor 488 conjugate, Cell Signaling) following our published procedure.^{33, 35}

2.5. Immunoblot analysis

Cellular protein extraction, 4–15% gradient SDS-PAGE gel electrophoresis (Bio-Rad laboratories), transfer to PVDF membrane, and immunoblot development were performed as published recently.^{3, 33} The following rabbit anti-human antibodies were used (obtained from Cell Signaling): cleaved PARP-1 (5625), p-eIF2α (3398), total eIF2α (5324), SQSTM1 (8025), LAMP1 (9091), LC3-I/II (12741), cytochrome c (11940), ATF-4 (11815), and XBP-1s (40435). Equal protein loading was examined by β-actin detection using a mouse monoclonal antibody (Sigma Aldrich); secondary antibodies: HRP-conjugated goat anti-rabbit or goat anti-mouse (Jackson ImmunoResearch Laboratories). Densitometric image analysis was performed using Image Studio^™ Lite quantification software (LI-COR Biosciences). For cytochrome c immunodetection, cell fractionation (mitochondrial versus cytosolic) was performed using the Mitochondria/Cytosol fractionation kit (ab65320, Abcam). Purity of fractions was confirmed according to kit instructions.

2.6. Assessment of mitochondrial transmembrane potential (Δψm)

The potentiometric dye 5,5’,6,6’-tetrachloro-1,1’,3,3’-tetraethylbenzimidazolyl-carbocyanine iodide (JC-1; Sigma, T4069) was used to examine the mitochondrial transmembrane potential [Δψm; JC-1 monomeric green fluorescence (depolarized mitochondria, detector FL-1) vs. JC-1 aggregate red fluorescence (polarized mitochondria, detector FL-2) fluorescence] following our previously published procedure.³⁵

2.7. Human Stress & Toxicity PathwayFinder RT² Profiler^™ gene expression array analysis

After MQ treatment (15 μM; 6 h), total mRNA from cultured A375-BRAF^V600E/NRAS^Q61 and A375-BRAF^V600E/NRAS^Q61K (200,000 in 35 mm dish format) was isolated using the RNeasy Mini kit (Qiagen) following our published standard procedures. Reverse transcription was then performed using the RT² First Strand kit (Qiagen) from 500 ng total RNA. For gene expression array analysis, the human Stress & Toxicity PathwayFinder RT² Profiler^™ technology (Qiagen), assessing expression of 84 stress response-related genes, was used as published before.^{35, 36} Quantitative PCR was run using the following conditions: 95 °C (10 min), followed by 40 cycles at 95 °C (15 s) alternating with 60 °C (1 min) (Applied Biosystems). Gene-specific products were normalized to a group of 5 housekeeping genes (ACTB, B2M, GAPDH, HPRT1, RPLP0) and quantified using the comparative ΔΔCt method (ABI Prism 7500 sequence detection system user guide).

2.8. Individual RT-qPCR analysis

Total cellular mRNA was isolated using the Qiagen RNeasy Mini Kit (Qiagen) according to the manufacturer’s protocol. Human primer probes [DDIT3 (Hs_00358796_g1), HSPA6 (Hs_00275682_s1), HSPA1A (Hs_00359163_s1), EGR1 (Hs_00152928_m1), HMOX1 (Hs_00157965_m1), SOD2 (Hs_00167309_m1), RSP18 (housekeeping gene; Hs_01375212_g1)], were obtained from Thermo Fisher Scientific. After cDNA synthesis, quantitative PCR reactions were performed as follows: 10 min (95 °C) followed by 15 s (95 °C), 1 min (60 °C), 40 cycles, using the ABI7500 Real-Time PCR System (Applied Biosystems). Amplification plots were generated, and Ct values were recorded as published before.³⁴

2.9. Cellular ER and Ca²⁺ Fluo-4 AM imaging

For ER organelle imaging, photostable ER-Tracker^™ Blue-White DPX dye staining (ThermoFisher Scientific) was used. Briefly, cells were seeded in 6-well plate format (100,000/well), drugged accordingly (15 μM MQ; 6 h) and stained with 1 mM ER-tracker probe for 1 h in the dark (37 °C; 5% CO₂). DAPI filter settings were used to image cells (EVOS fluorescent microscope; ThermoFisher Scientific). For Calcium imaging, cells were stained with Fluo-4 AM dye (3 μM in HBSS; ThermoFisher Scientific) for 1 h in the dark at room temperature.³⁷ The Fluo-4 loaded cells were washed with HBSS and then exposed to MQ (15 μM) before fluorescent image acquisition (using a FITC filter).

2.10. Transmission electron microscopy

Cells were fixed in situ with 2.5% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.4), post-fixed in 1% osmium tetroxide in cacodylate buffer, washed, scraped and pelleted as described recently.²¹ Cells were then stained in 2% aqueous uranyl acetate, dehydrated through a graded series (50, 70, 90, and 100%) of ethanol and infiltrated with Spurr’s resin (Sigma, EM0300), then allowed to polymerize overnight at 60 °C. Sections (50 nm) were cut, mounted onto uncoated 150-mesh copper grids, and stained with 2% lead citrate. Sections were examined in a CM12 transmission electron microscope (FEI) operated at 80 kV with digital image collection.

2.11. Detection of intracellular oxidative stress

Levels of MQ-induced intracellular oxidative stress were analyzed by flow cytometry using 2′,7′ - dichlorodihydrofluorescein diacetate (DCFH-DA) as a sensitive nonfluorescent precursor dye according to a published standard procedure.^{34, 35} Briefly, cells were treated with MQ (15 μM, ≤ 3 h), followed by DCFH-DA loading (5 μg/ml), incubated for 1 h in the dark (37 °C, 5% CO₂), then harvested and analyzed immediately by flow cytometry.

2.12. MQ chemotherapeutic intervention in a bioluminescent murine model of human intracranial malignant melanoma

A published standard procedure generating and assessing melanoma brain metastases in live mice was employed.³⁸ To this end, anesthetized SCID mice (male, 10 weeks old, 10 in total) received A375-Luc2 (BRAF^V600E/NRAS^Q61K; CRL-1619IG-2-LUC2) cells (100,000 cells/animal) through stereotaxic methods.^{39, 40} The cells were implanted unilaterally into the striatum [coordinates: anterior-posterior (AP): +0.74; medial-lateral (ML): +1.2; dorsal-ventral (DV): −3.0] via intracranial injections (50,000 cells/ μl of sterile PBS; 2 μl volumes delivered at 0.5 μl/ min). Control animals received 2 μl PBS alone. One day after injection, bioluminescent imaging (Lago CDC bioluminescent imaging system, Spectral Instruments Imaging) was performed followed by pair matching. Starting on d 7 after cell injection until d 17, mice received MQ [50 mg/kg/d in H₂O (10% DMSO; 200 μl via gavage); n=5] or solvent control (n=5). On d 18, mice were imaged for final tumor growth assessment and tissue was harvested for IHC analysis. This study was performed in accordance with the recommendations of the National Institutes of Health (University of Arizona Institutional Animal Care and Use Committee; mouse protocol number: IACUC 17–298)

2.13. Immunohistochemistry

Brain tissue was analyzed for detection of proliferating cell nuclear antigen (PCNA) (2586, Cell Signaling Technology) and MELTF (NBP1-85777, Novus Biologicals) as published.^20,22 Nuclear staining was performed using Hematoxylin QS counterstain (H3404; Vector Laboratories, Burlingame, CA). Images were captured using an Olympus BX50 and Spot (Model 2.3.0) camera.

2.14. Statistical analysis

Numerical data were analyzed as published recently.^{3, 34} Unless stated differently, data sets were analyzed employing analysis of variance (ANOVA) with Tukey’s post-hoc test using the Prism 8.4.3 software (Prism Software Corp., Irvine, CA, USA); in respective bar graphs (analyzing more than two groups), means without a common letter differ (p < 0.05). For bar graphs comparing two groups only, statistical significance was calculated employing the Student’s two-tailed t-test, utilizing Excel (Microsoft^™). Experiments involved six individual replicates per data point, except for gene expression array analysis (using three independent biological replicates). Nonparametric data analysis of murine experimentation was performed using the Mann–Whitney test. The level of statistical significance was marked as follows: p * < 0.05; p ** < 0.01; p *** < 0.001.

3. RESULTS

3.1. MQ induces cell death in BRAFi-resistant A375 malignant melanoma cells

First, in search of ACT antimalarials displaying single agent apoptogenicity targeting malignant melanoma cells, we examined a focused library of clinical amino-quinoline ACT-antimalarials (15 μM, 24 h) for the ability to induce A375-BRAF^V600E/NRAS^Q61 melanoma cell death as assessed by annexin V/PI-flow cytometric analysis (Figure 1A). Among the chemical entities tested, only MQ [± mefloquine (R/S, S/R-racemic mixture)] caused pronounced cell death detectable upon 24 h continuous exposure to low micromolar concentrations. Next, the detailed dose-response relationship of MQ-induced cell death was established in a panel of cultured human melanoma cells [BRAF^V600E/NRAS wildtype: A375, LOX-IMVI, G361, SK-MEL-28; BRAF^V600E/mutated NRAS: A375-BRAF^V600E/NRAS^Q61K; Figure 1B).^{3, 27–32} All melanoma cell lines contained in our test panel displayed sensitivity to MQ-induced cell death observable in the low micromolar concentration range [Figure 1B; MQ (0–30 μM]. Remarkably, MQ displayed cytotoxic efficacy irrespective of BRAFi-resistance status (A375-BRAF^V600E/NRAS^Q61 versus A375-BRAF^V600E/NRAS^Q61K), and a moderate (up to two-fold) increase in MQ-sensitivity of A375-BRAF^V600E/NRAS^Q61K cells (as compared to their isogenic controls) was observed (Figure 1B,C).

Figure 1. MQ impairs viability of cultured human malignant melanoma cell lines including BRAFi-sensitive and NRAS-based BRAFi-resistant A375 isogenic variants.

(A) Annexin V/PI-flow cytometric analysis of A375-BRAF^V600E/NRAS^Q61 melanoma cell death induced by exposure to clinical ACT-antimalarials (15 μM, 24 h). The numbers indicate viable (AV⁻, PI^-) in percent of total gated cells (mean ± SD). (B) Dose-response relationship of MQ-induced cell death in A375-BRAF^V600E/NRAS^Q61, A375-BRAF^V600E/NRAS^Q61K, LOX-IMVI, G361, SK-MEL-28 (MQ ≤ 30 μM, 24 h) as assessed in (A). (C) Time course analysis of A375-BRAF^V600E/NRAS^Q61 versus A375-BRAF^V600E/NRAS^Q61K cell death induction [MQ, 15 μM, ≤ 24 h; (left panels: flow cytometric analysis; right bar graph: quantitative analysis)]. For all bar graph depictions, quantitative data analysis employed ANOVA with Tukey’s post hoc test; means without a common letter differ from each other (p<0.05).

Given the remarkable single agent activity of MQ impairing melanoma cell viability (observable even in BRAFi-resistant NRAS isogenic variant), we focused our further studies on comparative examination of MQ-induced cytotoxicity employing our A375 isogenic variants. First, in order to inform the subsequent gene expression array analysis (Figure 2), a detailed time course of cell death induction by MQ was established comparing A375-BRAF^V600E/NRAS^Q61 versus A375-BRAF^V600E/NRAS^Q61K cells (Figure 1C; ≤ 24 h continuous exposure). This analysis indicated that cell viability was fully maintained at 6 h exposure time, followed by loss of viability detectable at 12 h and after. As observed above (Figure 1B), loss of viability induced by MQ exposure (24 h) was almost two-fold more pronounced in A375-BRAF^V600E/NRAS^Q61K cells as compared to the isogenic NRAS wildtype line (Figure 1C).

Figure 2. Comparative array analysis profiles MQ-induced stress response gene expression in isogenic A375-BRAF^V600E/NRAS^Q61 versus A375-BRAF^V600E/NRAS^Q61K cells.

Comparative stress response gene expression analysis in MQ-treated A375-BRAF^V600E/NRAS^Q61K and A375-BRAF^V600E/NRAS^Q61 cells. Volcano blots (panels A and B) depict differential gene expression as detected by the Human Stress and Toxicity Profiler^™ RT² PCR array technology (MQ: 15 μM, 6 h; relative to untreated; cut-off criteria: fold change ≥ 2; p < 0.05). (C) Venn diagram comparing statistically significant MQ-induced expression changes. (Top panels are scatter blot depictions.) (D) Quantitative summary of MQ-induced expression changes comparing isogenic cell lines (cut-off criteria: fold change ≥ 2; p value < 0.05). (E) MQ-induced proteotoxic and oxidative stress response at the single gene expression level confirmed by independent RT-qPCR. Bar graphs depict fold change versus untreated control as a function of genotype. (F) Expression of top three upregulated genes (EGR1, HMOX1, DDIT3; from panel D) examined at the protein level using immunoblot analysis [MQ (15 μM); t ≤ 6 h]; bar graph depiction summarizes densitometric analysis. For bar graph depictions, quantitative data analysis employed ANOVA with Tukey’s post hoc test; means without a common letter differ from each other (p<0.05). For bar graphs comparing two groups only, statistical significance was calculated employing the Student’s two-tailed t-test (**p < 0.01; ***p < 0.001).

3.2. Comparative array analysis: MQ-induced stress response gene expression in isogenic A375-BRAF^V600E/NRAS^Q61 versus A375-BRAF^V600E/NRAS^Q61K melanoma cells

In order to explore the early stress response elicited by short term MQ treatment [15 μM, 6 h; a time point at which viability was not yet impaired (Fig. 1C)], we profiled parental A375-BRAF^V600E/NRAS^Q61 and isogenic BRAFi-resistant A375-BRAF^V600E/NRAS^Q61K cells employing comparative gene expression analysis (Figure 2). Using the Human Stress and Toxicity Profiler^™ RT² PCR array technology, interrogating expression of 84 stress-related genes contained on the array, it was observed that MQ-induced expression changes in isogenic A375 cells affected 20 genes by at least two-fold over untreated control cells (Figure 2A,B). Venn diagram depiction comparing MQ-induced expression changes identified nine genes displaying MQ-induced expression changes shared between the two genotypes including (Figure 2C,D):

1. proteotoxic stress response genes [i.e. DDIT3 (DNA-damage-inducible transcript 3/ CHOP), HSPA6 (Heat shock 70kDa protein 6), HSPA5 (Heat shock 70kDa protein 5/ GRP78), HSPA1A (Heat shock 70kDa protein 1A);

2. redox stress-related genes [i.e. EGR1 (early growth response protein 1), HMOX1 (heme oxygenase-1), SOD2 (Superoxide dismutase 2, mitochondrial), MT2A (Metallothionein 2A)].

Remarkably, MQ-induced upregulation of gene expression was generally more pronounced in BRAFi-resistant A375-BRAF^V600E/NRAS^Q61K cells as compared to their isogenic controls (Figure 2D). For example, upregulation of DDIT3 by more than fifty-fold was contrasted by only ten-fold upregulation in the isogenic parental cells. Likewise, upregulation of EGR1 (93-fold) was contrasted by only six-fold upregulation in controls, a trend observed also with other genes including MT2A and HSPA6. Moreover, the MQ-induced proteotoxic and oxidative stress response was also confirmed at the mRNA and protein levels (using independent RT-qPCR and immunoblot analyses), indicating pronounced upregulation of select genes as already observed by array analysis (DDIT3, EGR1, HMOX1, etc.) in both A375 isogenic cell lines (Figure 2E,F).

3.3. MQ elicits ER stress signaling and early calcium dysregulation in isogenic BRAFi-resistant A375-BRAF^V600E/NRAS^Q61K and BRAFi-sensitive A375-BRAF^V600E/NRAS^Q61 melanoma cells

Next, informed by our gene expression data (Figure 2), the early occurrence of MQ-induced ER stress was examined by immunoblot analysis focusing on key effector signaling pathways (phospho-eIF2α, XBP-1s, ATF4; Figure 3A).^{41, 42} MQ-initiation of ER-related signaling was observed in both cell lines, with A375-BRAF^V600E/NRAS^Q61K cells displaying a more pronounced and rapid MQ-induced upregulation of phospho-eIF2α (detectable within 1 h exposure to MQ); in contrast, parental A375-BRAF^V600E/NRAS^Q61 cells displayed a more pronounced and rapid upregulation of ATF4 (detectable within 1 h exposure to MQ); rapid XBP-1s upregulation was also observed in both isogenic lines detectable within 3–6 h exposure times. Consistent with the early occurrence of ER stress in response to MQ treatment, ER-swelling and enlargement were detectable by quantitative ER-tracker^™ DPX fluorescence imaging and electron microscopy (Figure 3B,C).

Figure 3. MQ induces ER stress signaling and early calcium dysregulation in isogenic BRAFi-resistant A375-BRAF^V600E/NRAS^Q61K and BRAFi-sensitive A375-BRAF^V600E/NRAS^Q61 human malignant melanoma cells.

(A) MQ-induced ER stress signaling (phospho-eIF2α, XBP-1s, ATF4) as revealed by immunoblot analysis [MQ (15 μM); t ≤ 6 h]; bar graph depiction summarizes densitometric analysis. (B) ER organelle visualization by ER-tracker^™ DPX fluorescence imaging [MQ (15 μM; t ≤ 6 h; 20x magnification)]; bar graph depiction summarizes quantitative analysis. (C) Electron microscopy indicating ER organelle swelling and enlargement observed in untreated versus MQ-exposed (15 μM; 6 h) A375-BRAF^V600E/NRAS^Q61K cells [11,500-fold magnification; scale bar: 500 nm; N; nucleus; M: mitochondria; V: vesicle; ER: endoplasmic reticulum (marked by arrowheads)]. (D) Fluo-4 AM-fluorescence visualization of cytosolic free calcium observable within seconds of exposure to MQ (15 μM; 120 s); phase contrast and DAPI nuclear staining; representative images]. (E) Flow cytometric analysis (histograms) of MQ induced calcium signal (Fluo-4 AM-fluorescence intensity); upper left: representative histograms [0 – 600 s MQ (15 μM) exposure time]; upper right: bar graph depiction with statistical analysis; bottom: time resolved (60 – 360 s) Fluo-4 AM-fluorescence intensity as a function of MQ exposure time. (F) Flow cytometric detection of 2’,7’-dihydrodichlorofluorescein-diacetate (DCFH-DA) oxidation [producing dichlorofluorescein (DCF) fluorescence] as an indicator of cellular oxidative stress in response to MQ exposure (15 μM; ≤ 180 min exposure time) examined in both genotypes; bar graph depiction with statistical analysis. (G) Effect of calcium chelation [BAPTA-AM (20 μM)] or antioxidant intervention [NAC (10 mM)] on MQ-induced cellular oxidative stress (DCF-assay as performed in panel F) examined in both genotypes. For all bar graph depictions, quantitative data analysis employed ANOVA with Tukey’s post hoc test; means without a common letter differ from each other (p<0.05). For bar graphs comparing two groups only, statistical significance was calculated employing the Student’s two-tailed t-test (**p < 0.01).

Next, since calcium dysregulation causing cytosolic overload is an established consequence of ER stress, we employed Fluo-4 AM fluorescence-based calcium visualization (Figure 3D,E).³⁷ Indeed, cytosolic free calcium was observable within seconds of exposure to MQ as detected by fluorescence microscopy, an effect observable in both isogenic A375 melanoma cell variants irrespective of NRAS-genotype, quantified by flow cytometric analysis. Interestingly, calcium homeostasis was restored within 300 s of observation time as revealed by time-resolved (60 – 360 s) continuous monitoring of Fluo-4 AM-fluorescence intensity as a function of MQ exposure time (Figure 3E, bottom panel).³⁷

Next, the occurrence of redox dysregulation as a result of MQ treatment was examined employing flow cytometric detection of 2’,7’-dihydrodichlorofluorescein-diacetate (DCFH-DA) oxidation, an established indicator of cellular oxidative stress known to occur in the context of ER disruption (Figure 3F).^{41, 43} DCF fluorescence (resulting from indicator dye oxidation) was detectable starting at 60 min of MQ exposure, an early occurrence that places redox dysregulation downstream of earlier calcium dysregulation (observable within seconds of MQ exposure). As observed with calcium release, increase in DCF fluorescence signal was more pronounced in the NRAS-mutated genotype. Furthermore, treatment with the membrane-permeable calcium chelator BAPTA-AM (quenching the consequences of cytosolic overload) prevented MQ-induced increase in DCF fluorescence indicating that early calcium dysregulation is a causative factor upstream of ROS formation, an effect observable in both isogenic variants. (Figure 3G). As a positive control, antioxidant intervention using N-acetyl-L-cysteine (NAC) suppressed ROS formation that occurs as a consequence of MQ exposure.

Taken together these data document that MQ treatment causes early ER stress signaling associated with calcium dysregulation and subsequent induction of oxidative stress, observable in A375-BRAF^V600E/NRAS^Q61 melanoma cells with analogous yet more pronounced effects in the isogenic A375-BRAF^V600E/NRAS^Q61K variant.

3.4. MQ induces apoptotic cell death (with mitochondrial cytochrome C release, caspase activation, and PARP-1 cleavage) targeting BRAFi-resistant A375-BRAF^V600E/NRAS^Q61K melanoma cells, and oral administration blocks tumor growth in a bioluminescent murine model of intracranial BRAFi-resistant malignant melanoma.

Next, in preparation of testing the possibility of therapeutic elimination of BRAFi-resistant melanoma cells (located intracranially in SCID mice) employing systemic administration of MQ, we further explored molecular mechanisms of MQ-induced A375-BRAF^V600E/NRAS^Q61K melanoma cell death in vitro (Figure 4A–E). First, MQ-induced proteolytic activation of caspase-3 activation was confirmed by flow cytometric analysis (Figure 4A). Moreover, cells exposed to the combined action of MQ (15 μM, 24 h) and the pan-caspase inhibitor zVAD-fmk displayed significantly increased survival indicative of an early involvement of caspase-dependent pathways (Figure 4B). Importantly, immunoblot analysis at early time points of MQ exposure confirmed PARP-1 cleavage, an established target of caspase 3 involved in apoptotic execution (Figure 4C). PARP-1 cleavage was observable as early as within 1 h exposure time followed by further increase over 6 h; at the same time, markers of autophagic-lysosomal dysregulation [SQSTM1 (p62), LAMP1] remained largely unchanged during this early observational time frame (with the exception of LC3-I/II displaying mild inhibition indicative of interference with autophagic function), an observation consistent with the rapid occurrence of caspase-dependent pathways preceding disruption of autophagy (Figure 4C).²¹ Furthermore, mitochondrial release of cytochrome C (accompanied by the corresponding increase in cytosolic cytochrome C level) was detected, an observation consistent with involvement of the mitochondrial pathway of apoptosis (Figure 4D), confirmed further by a concomitant MQ-induced loss of mitochondrial transmembrane potential as detected by JC-1 flow cytometric analysis (Figure 4E).

Figure 4. MQ induces apoptosis in cultured A375-BRAF^V600E/NRAS^Q61K melanoma cells, and oral administration blocks tumor growth in a bioluminescent murine model of intracranial BRAFi-resistant malignant melanoma.

(A) MQ-induced (15 μM, 24 h) caspase-3 activation as examined in cultured A375-BRAF^V600E/NRAS^Q61K cells by flow cytometric detection. (B) Cell death in A375-BRAF^V600E/NRAS^Q61K melanoma cells exposed to the combined action of MQ (15 μM, 24 h) and the pan-caspase inhibitor zVAD-fmk (40 μM; mean ± SD]. (C) PARP-1 cleavage and modulation of autophagic-lysosomal mediators (SQSTM1, LAMP1, LC3-I/II) in response to MQ-exposure (15 μM, ≤ 6 h) in A375-BRAF^V600E/NRAS^Q61K cells as assessed by immunoblot analysis (left panel); bar graph depiction summarizes densitometric analysis (right panel). (D) MQ-induced (15 μM; 24 h) cytochrome C release from cellular fractions (mitochondrial versus cytosolic). (E) Loss of mitochondrial transmembrane potential (Δψm) was assessed by flow cytometric analysis of JC-1 stained cells (15 μM; 6–24 h). Numbers indicate percentage of cells inside the circle displaying intact Δψm [mean ± SD; (p < 0.05)]. (F) Chemotherapeutic MQ-regimen in vivo: A375-Luc2 [BRAF^V600E/NRAS^Q61K] melanoma cells were intracranially injected (n=5 per group) followed by bioluminescent image analysis of brain tumors on d1 (pair matching) and d18 (harvest). Starting on d7 after cell injection until d18, mice received MQ [50 mg/kg/d in H₂O (10% DMSO; 200 μL via gavage)] or solvent control. (G) Mouse body weights as a function of treatment group and time (n=5 per group; d7–18). (H) Bioluminescent imaging (d18); bar graph (right panel) depicts quantitative image analysis of bioluminescent intracranial signal. Nonparametric data analysis (Figure 4H) was performed using the Mann–Whitney test (**p < 0.01). For all other bar graph depictions, quantitative data analysis employed ANOVA with Tukey’s post hoc test; means without a common letter differ from each other (p<0.05); for bar graphs comparing two groups only, statistical significance was calculated employing the Student’s two-tailed t-test (*p < 0.05; ***p < 0.001).

Next, based on prior experimental and clinical evidence documenting brain availability of systemic MQ, we tested feasibility of targeting BRAFi-resistant A375-BRAF^V600E/NRAS^Q61K melanoma in a bioluminescent murine model (Figures 4F–H; 5).^{44, 45} To this end, A375-Luc2 [BRAF^V600E/NRAS^Q61K] melanoma cells were intracranially injected followed by bioluminescent image analysis of brain tumors in SCID mice receiving an oral MQ-regimen as compared to administration of vehicle control only (Figure 4F–H). Indeed, quantitative image analysis at the end of the experiment revealed pronounced tumor growth suppression as substantiated by an almost seven-fold reduction in bioluminescent signal in the MQ-treated group (Figure 4H), a chemotherapeutic effect that occurred without negative impact on average mouse body weight (Figure 4G).

Subsequently, IHC analysis of transverse cuts of brain tissue was performed at the end of the experiment (d 18) in order to confirm suppression of melanoma tumor growth [involving H&E (Figure 5A) and PCNA/ MELTF staining (Figure 5B,C)]. The presence (‘- MQ’ treatment group) or absence (‘+ MQ’ treatment group) of proliferating melanoma cells was substantiated assessing PCNA expression in cells staining positive for the melanocyte-specific antigen melanotransferrin [MELTF (p97); Figures 5B,C].

Figure 5. Immunohistochemical analysis of murine brain tissue confirms MQ-induced attenuation of intracranial BRAFi-resistant malignant melanoma.

At the end of the experiment (d18), brain tissue was processed for IHC [H&E (panels A, C); PCNA, MELTF (panels B, C). (A) Representative transverse cuts of brain specimens (H&E); top panels: complete specimens (‘- MQ’ versus ‘+ MQ’); bottom panels: enlarged areas (10x magnification; arrows point towards site of cell injection; scale bar: 200 μm). (B) Immunohistochemical staining (PCNA; transverse cuts of brain specimens) from control and MQ-exposed mice. (C) Transverse cut of tumor bearing area [untreated (‘- MQ’) mouse; one representative specimen]; top panel: PCNA (tumor marked by arrow; scale bar: 200 μm); bottom panels (left to right): H&E, PCNA, MELTF.

4. Discussion

Repurposing of clinical antimalarials has shown promise for cancer-directed molecular intervention, and inhibitory modulation of autophagic-lysosomal function has been substantiated as a major molecular mechanism of action underlying cancer cell-directed activity of antimalarials.²⁴ Among various aminoquinoline-antimalarials, cancer cell-directed cytotoxicity of MQ is well documented, involving numerous tentative molecular mechanisms, including autophagic-lysosomal blockade, inhibition of P-glycoprotein-dependent drug efflux, NFκB pathway attenuation through IKK inactivation, and induction of redox dysregulation.^46–49

Here we demonstrate for the first time that MQ induces cell death in cultured malignant melanoma cells displaying apoptogenic activity observable even in BRAFi-resistant BRAF^V600E/NRAS^Q61K A375 melanoma cells (Figures 1 and 4), an effect associated with rapid induction of proteotoxic and oxidative stress response gene expression (Figure 2). Moreover, ER stress signaling and rapid onset of calcium dysregulation were detected in response to MQ treatment (Figure 3).

Remarkably, feasibility of targeting intracranial melanoma upon systemic administration of MQ was demonstrated in a bioluminescent murine model, an important preclinical observation given the notoriously unsatisfactory clinical outcomes currently achieved using standard of care interventions in large patient populations afflicted by brain metastases (Figures 4 and 5).^10–12 Importantly, in the context of oral MQ targeting brain melanoma in a relevant murine model, it remains to be seen if MQ treatment may also impact molecular pathways involved in invasion and metastasis (such as regulation of EMT-related gene expression) blocking the occurrence of intracranial tumors, a topic of current research activities pursued in our laboratory.³⁴

Interestingly, molecular mechanisms of MQ-associated calcium dysregulation and ER-directed toxicity have been explored before in the context of MQ-induced neurotoxic symptoms observed in malaria patients, representing a well-documented adverse effect of this antimalarial intervention.^{44, 45, 50–52} This adverse drug action has been attributed to specific molecular interactions including: (i) noncompetitive acetylcholine esterase inhibition, (ii) interaction with voltage-dependent channels, and (iii) ER-associated Ca-ATPase (SERCA) antagonism (mimicking activity of the SERCA-antagonist thapsigargin), all of which might be causing cytosolic overload and dysregulation of calcium homeostasis.^{44, 52–54} Indeed, ER stress and calcium dysregulation have been identified as critical factors in signal transduction and determination of cancer cell fate, connecting ER, mitochondrial, and lysosomal stress signaling with therapeutic efficacy, a topic with emerging relevance to melanomagenesis and therapy.^55–60 However, specific mechanism and molecular targets underlying MQ-induced calcium dysregulation achievable in malignant melanoma cells remain to be identified. It is tempting to speculate that a shared vulnerability to MQ-induced calcium dysregulation might characterize both neurons and melanoma/ melanocytic cells given their common neural crest origin. Also, it should be mentioned that based on the stereochemical diversity of mefloquine [employed here as the clinically-used (±) racemic mixture] either one of the 4 possible stereoisomers might display differential ligand activity modulating a tentative calcium regulatory target, a topic that has been addressed to some extent before comparing mammalian SERCA-directed inhibitory activity of (+)- and (−)- MQ isomers.⁵⁴

Therapeutic induction of ER stress has shown promise in overcoming BRAFi resistance, and, consistent with our observations (Figure 2D), upregulation of ER stress-related mediators including CHOP (encoded by DDIT3) and GRP78 (encoded by HSPA5) has been observed before in MQ-exposed MDA-MB-231 breast carcinoma cells, but a causative involvement of ER stress in MQ-induced cytotoxicity was not substantiated in that study.⁶¹ In our experiments targeting BRAF^V600E/NRAS^Q61K BRAFi-resistant A375 melanoma cells using MQ, ER stress response gene expression (Figure 2), disruption of calcium homeostasis (Figure 3D,E,G), induction of oxidative stress (Figure 3F,G), and mitochondrial dysregulation were observable at early time points (Figure 4D,E), with calcium dysregulation preceding all other changes (upstream of ROS production; Figure 3F,G). However, specific molecular interdependencies and targets relevant to MQ-induced apoptotic elimination of BRAFi-resistant malignant melanoma cells remain to be explored in adequate detail allowing stringent mechanistic conclusions that might inform future clinical decision making. It is also worth mentioning that recent drug discovery efforts have generated MQ-derivatives with improved experimental cancer therapeutic activity, emphasizing the need for more detailed identification of molecular mechanisms specific to MQ that would explain its unique antimelanoma profile distinct from other aminoquinoline antimalarials as described by us here for the first time.⁶² In addition, for further preclinical and clinical development aiming at repurposing MQ for melanoma intervention, it will be essential to conduct more detailed dose optimization studies identifying clinically relevant MQ regimens and to explore potential MQ synergism with standard of care melanoma therapeutics (including BRAFi and immune checkpoint inhibitors) assessed in treatment-naïve and treatment-resistant tumors.^{1, 24}

Taken together, these novel preclinical data suggest that the clinical antimalarial MQ may be a valid candidate for drug repurposing aiming at chemotherapeutic elimination of malignant melanoma cells, even if BRAFi-resistance and metastasis to the brain have occurred. However, relevance of the cellular BRAFi-resistance model (A375-BRAF^V600E/NRAS^Q61K), employed by us for the first time for drug screening and in vivo testing, might be limited since the specific NRAS mutation examined here is observable only in approximately 20% of patients displaying BRAFi-resistance necessitating a more comprehensive coverage of patient-relevant resistance models to be pursued in the near future.^{3, 9}

Abbreviations:

ACT

artemisinin-combination therapeutic

ANOVA

analysis of variance

BAPTA-AM

1,2-Bis(2-aminophenoxy) ethane-N,N,N’,N’-tetraacetic acid tetrakis(acetoxymethyl ester)

BRAFi

BRAF kinase inhibitors

DAPI

4′,6-diamidino-2-phenylindole

DCF

dichlorofluorescein

DCFH-DA

2′,7′ - dichlorodihydrofluorescein diacetate

ER

endoplasmic reticulum

JC-1

5,5’,6,6’-tetrachloro-1,1’,3,3’-tetraethylbenzimidazolyl-carbocyanine iodide

MQ

mefloquine

NAC

N-acetyl-L-cysteine

PI

propidium iodide

ROS

reactive oxygen species

DATA AVAILABILITY STATEMENT

The data sets used or analyzed in the context of the current study are available from the corresponding author by reasonable request.