Overcoming MITF-conferred drug resistance through dual AURKA/MAPK targeting in human melanoma cells

Abstract

MITF (microphthalmia-associated transcription factor) is a frequently amplified lineage-specific oncogene in human melanoma, whose role in intrinsic drug resistance has not been systematically investigated. Utilizing chemical inhibitors for major signaling pathways/cellular processes, we witness MITF as an elicitor of intrinsic drug resistance. To search kinase(s) targets able to bypass MITF-conferred drug resistance, we employed a multi-kinase inhibitor-directed chemical proteomics-based differential affinity screen in human melanocytes carrying ectopic MITF overexpression. A subsequent methodical interrogation informed mitotic Ser/Thr kinase Aurora Kinase A (AURKA) as a crucial regulator of melanoma cell proliferation and migration, independent of the underlying molecular alterations, including TP53 functional status and MITF levels. Crucially, assessing the efficacy of investigational AURKA inhibitor MLN8237, we pre-emptively witness the procurement of a molecular program consistent with acquired drug resistance. This involved induction of multiple MAPK (mitogen-activated protein kinase) signaling pathway components and their downstream proliferation effectors (Cyclin D1 and c-JUN) and apoptotic regulators (MITF and Bcl-2). A concomitant AURKA/BRAF and AURKA/MEK targeting overcame MAPK signaling activation-associated resistance signature in BRAF- and NRAS-mutated melanomas, respectively, and elicited heightened anti-proliferative activity and apoptotic cell death. These findings reveal a previously unreported MAPK signaling-mediated mechanism of immediate resistance to AURKA inhibitors. These findings could bear significant implications for the application and the success of anti-AURKA approaches that have already entered phase-II clinical trials for human melanoma.

Figure 1

MITF confers intrinsic drug resistance in melanocytes. (a) Immortalized human melanocytes transformed with constitutively active BRAF(V600E) (HMEL-B) or a combination of BRAF(V600E) and MITF(HMEL-B/M) were analyzed for the expression of the indicated proteins by immunoblotting. (b) HMEL-B and HMEL-B/M cells were treated with the indicated inhibitors (Supplementary Table S1) for 24 h, followed by determination of the resistivity index (see Methods) of HMEL-B/M cells in comparison to the HMEL-B cells (n=3)

Figure 2

Integrative differential drug affinity-based proteomics. (a) Flowchart depicting the multi-step strategy (DEFINIT) for the identification of the kinase(s) targets circumventing MITF-mediated intrinsic drug resistance. (b) HMEL-B/M and HMEL-B cells were treated as indicated for 4 weeks in a soft agar colony formation assay followed by colony count (n=3). (c) List of top 10 kinases exhibiting higher relative affinity for midostaurin in comparison to sunitinib based on iTRAQ score. (d) Venn diagram depicting the exclusive and shared kinases from iTRAQ and (1D) LC–MS studies. (e) HMEL-B/M cells were treated as indicated for 4 weeks in a soft agar colony formation assay followed by colony count (n=3). (f) HMEL-B and HMEL-B/M cells were treated with DMSO control or AURKA inhibitor (MLN8237, 100 nM) and GSK3A inhibitor (SB415286, 5 μM) for 24 or 48 h followed by assessment of relative viability (n=3). (g) Relative gene expression levels of AURKA in the indicated stages of melanoma. All error bars indicate ±S.D.; ns, non-significant; *P⩽0.05, **P⩽0.01, ***P⩽0.001

Figure 3

Role of AURKA in melanoma biology. (a) Indicated melanoma cell lines were treated with DMSO control or MLN8237 (1 μM) for 48 h followed by measurement of relative viability (n=3). (b) (Left) Indicated wild-type (wt) TP53 melanoma cell lines were treated with DMSO control or MLN8237 (1 μM) for 16 h followed by cell cycle analysis. (Right) Quantification of the cell cycle distribution. (c) Immunoblotting-based analysis of TP53 and p21^Cip1 in the indicated wt TP53 melanoma cell lines upon treatment with DMSO control or MLN8237 (1 μM) for 48 h. (d) (Left) Cell cycle analysis of mut-TP53 Sk-Mel2 and Sk-Mel28 cells upon treatment with DMSO control or MLN8237 (1 μM) for 16 h. (Right) Quantification of the cell cycle distribution. (e) Indicated mut-TP53 melanoma cell lines were treated with DMSO control or MLN8237 (1 μM) for 48 h followed by assessment of TP53 and p21^Cip1 levels by immunoblotting. (f) Indicated melanoma cell lines were treated with DMSO control or MLN8237 (1 μM) for 48 h followed by Annexin V/propidium iodide (PI) staining. Percentages on the bottom correspond to the early apoptotic (Annexin V positive)+late apoptotic (Annexin V+PI positive) cells. (g) A375 and UACC-62 cells were treated as indicated for 48 h followed by immunoblotting-based assessment of TP53 and p21^Cip1 levels. MLN8237 (1 μM) was used. (h) A375 and UACC-62 cells were treated as indicated for 48 h followed by Annexin V/PI staining. MLN8237 (1 μM) was used. (i) A375 and UACC-62 cells were treated as indicated for 48 h followed by Annexin V/PI staining. Pifithrin-α and MLN8237 were used at 10 μM and 1 μM, respectively. (j) Soft-agar colony formation assay with the indicated melanoma cell lines treated with DMSO control or MLN8237 (40 nM) for 7 days (n=3). Quantification was performed as explained in the methods section. Error bars indicate ±S.D.; **P⩽0.01, ***P⩽0.001

Figure 4

MAPK signaling-elicited resistance program. (a) Indicated melanoma cell lines were treated with DMSO control or MLN8237 (1 μM) for 48 h followed by immunoblotting for MITF. (b) UACC-62 cells were treated with DMSO control or MLN8237 (1 μM) for 18 h followed by qRT-PCR based transcript analysis for the indicated molecules (n=2). The representative experiment is shown. (c) Indicated melanoma cell lines were treated for 48 h with DMSO control or MLN8237 (1 μM) followed by the assessment of Cyclin D1 and Bcl-2 levels by immunoblotting. (d) HT144 and A375 cells were transfected with si-MITF. After 24 h MLN8237 (1 μM) was added for another 24 h. After a total of 48 h relative viability was assessed (n=3). (e) A375 and UACC-257 cells were treated as indicated followed by immunoblotting for MITF, Cyclin D1, p21^Cip1 and Bcl-2. MLN8237 used at 1 μM concentration. (f) Indicated melanoma cell lines were treated with DMSO control or MLN8237 (1 μM) for 48 h followed by immunoblotting for the indicated molecules. (g) Indicated melanoma cell lines were treated with DMSO control or MLN8237 (1 μM) for 48 h followed by immunoblotting for BRN2 and p(Ser-133)CREB1. (h) (Left) Box diagram summarizing the expression changes in MAPK signaling and the downstream components in the indicated cell lines upon AURKA inhibition. (Right) Schematic representation of AURKA inhibitor-mediated MAPK signaling-elicited potential resistance program. All error bars indicate ±S.D.; *P⩽0.05, **P⩽0.01, ***P⩽0.001

Figure 5

Combinatorial AURKA and MAPK targeting in melanomas. (a) A375, UACC-62 and LOX-IMVI cells were treated as indicated for 48 h followed by immunoblotting-based analysis of the indicated proteins. MLN8237 (1 μM) and PLX-4032 (1 μM) were used. (b) HT144, A375, UACC-257 and UACC-62 cells were treated as indicated for 48 h followed by the assessment of relative cell viability (n=3). MLN8237 (10 nM) and PLX-4053 (10 nM) were used. (c) (Left) HT144 and UACC-62 cells were treated as indicated for 48 h followed by Annexin V/PI staining. MLN8237 (0.5 μM) and PLX-4053 (1 μM) were utilized. (Right) Bar diagrams showing the corresponding quantification. The representative experiment is shown. (d) (Left) NRAS(Q61R)-mutated Sk-Mel2 cells were treated as indicated for 48 h followed by western blotting for the indicated proteins. MLN8237 (1 μM) and U0126 (1 μM) were used. (Right) Sk-Mel2 cells were treated as indicated for 48 h followed by the measurement of relative viability (n=3). MLN8237 (100 nM) and U0126 (100 nM) were used. (e) (Upper panel) Unlike all the tested small molecule inhibitors (Figure 1b), AURKA inhibition overcomes MITF-mediated intrinsic drug resistance. However, after some time lapse, AURKA inhibition triggers MAPK signaling activation through enhanced expression/activity of several MAPK signaling components (Figure 4). MAPK signaling induction in-turn upregulates downstream proliferative/drug resistance molecular signature, including MITF, c-Jun, Cyclin D1 and Bcl-2 induction. (Lower panel) A concomitant treatment with MAPK inhibitors (BRAF inhibitor for mut-BRAF and MEK inhibitor for mut-NRAS) successfully relieves the induction of MAPK signaling components and the associated downstream expression signature; overrides the acquired resistance build-up; and potentiates the anti-proliferative efficacy of AURKA inhibition. All error bars indicate ±S.D.; ns, non-significant; *P⩽0.05, **P⩽0.001