Refining Targeted Therapy Opportunities for BRAF-Mutant Melanoma

First-line therapies for metastatic melanoma include BRAF/MAPK–directed small-molecule inhibitors for melanomas harboring oncogenic BRAF mutations, and immune checkpoint blockade for both BRAF-mutant and BRAF wild-type melanoma (1). Oncogenic BRAF mutations are present in roughly half of all cutaneous melanomas, with 90% of all BRAF mutations occurring in Val⁶⁰⁰ within the catalytic domain, which enhances BRAF catalytic activity, thereby promoting MAPK pathway activation and unrestricted cell growth (1). BRAF inhibitors and MEK inhibitors are active as single agents in BRAF-mutant melanoma, but combination BRAF–MEK inhibition yields higher response rates and improved progression-free survival and overall survival. Despite initial response rates of 70% to 80% with combination BRAF–MEK inhibition, resistance develops in most patients within 1 year, and long-term benefit is seen in a limited number of patients (22% progression-free survival and 44% overall survival at 3 years; ref. 2).

Several resistance mechanisms to BRAF/MAPK–directed therapy have been reported, including NRAS mutations, bypass pathway activation of MAPK through alternate receptor tyrosine kinases (e.g., EGFR), activation of COT (MAP3K8), MAP2K1 (MEK1) and MAP2K1 (MEK2) mutations, loss of NF1, target amplification, target alteration (e.g., alternative splicing), and secretion of growth factors acting in autocrine or paracrine fashion to promote outgrowth of resistant cells (3). Given these diverse mechanisms of resistance, identification of a single molecularly defined and druggable target with activity in BRAF-mutant melanoma with innate and/or acquired resistance to BRAF/MAPK pathway inhibition, as well as BRAF wild-type melanoma, would represent a major advance with significant mechanistic and therapeutic implications.

In their article in this issue of Cancer Discovery, Eskiocak and colleagues describe a novel molecularly defined subclass of melanoma characterized by sensitivity to TBK1/IKKϵ inhibition (4). TBK1 and IKKϵ are homologous Ser/Thr kinases with roles in innate immune signaling, cell proliferation and growth, xenophagy/autophagy, and cancer pathogenesis (5). Oncogenic KRAS co-opts TBK1 signaling to promote tumorigenesis, and further studies have demonstrated that TBK1 inhibition impairs cytokine-fueled growth of KRAS-mutant lung adenocarcinoma, especially in combination with MEK inhibition (5). Dysregulated TBK1 signaling is also intertwined with autophagy and may contribute to the dual role of autophagy in KRAS-mutant pancreatic carcinogenesis (5).

In a search for copy number–driven survival genes, Eskiocak and colleagues initially identified a SOX10-addicted subset of melanomas and demonstrated that SOX10 addiction defines sensitivity to BRAF/MAPK pathway inhibition in melanomas harboring oncogenic BRAF mutations. Conversely, SOX10-independent cell lines exhibit resistance to BRAF/MAPK inhibition regardless of BRAF mutational status, consistent with a previous report in which shRNA-mediated SOX10 knockdown induced resistance to BRAF/MAPK–directed therapy through upregulation of RTK signaling, including TGFβ and EGFR signaling (6). Further evaluation of the SOX10-addicted phenotype yielded a 5-gene biomarker that was able to differentiate sensitivity and resistance to BRAF/MEK–directed therapy. To identify novel therapies for melanomas exhibiting innate resistance to BRAF/MEK therapy, the authors used the predictive feature set coupled with matched drug sensitivity data for 130 small-molecule inhibitors. The compound BX795, an inhibitor of TBK1/IKKϵ and PDK1, was shown to preferentially inhibit cell growth of the targeted therapy–resistant class, and the importance of TBK1/IKKϵ was confirmed using a second inhibitor (compound II) that lacked PDK1-inhibitory activity, as well as two additional multitargeted TBK1/IKKϵ inhibitors (MRT67307 and momelotinib).

Cell lines with innate or acquired resistance to MAPK pathway inhibition exhibited marked sensitivity to TBK1/IKKϵ inhibition, suggesting a selective vulnerability in melanoma cell lines and xenografts with innate or acquired resistance to BRAF/MEK inhibition. The authors also leveraged The Cancer Genome Atlas (TCGA) data and determined “MITF^lo” and “immune” TCGA melanoma subtypes that were enriched for the predicted TBK1/IKKϵ–sensitive subtype. Gene set enrichment analysis (GSEA) revealed correlation with innate immune signaling and Toll-like receptor signaling in the TBK1/IKKϵ–sensitive melanoma subtype. Interestingly, the “MITF^lo” state has previously been associated with cell state change associated with upregulation of the receptor tyrosine kinase AXL and resistance to targeted therapies. Thus, the TBK1/IKKϵ–sensitive subtype may be related to the MITF^lo/AXL^hi transcriptional cell state associated with NF-κB upregulation and resistance to MAPK pathway inhibition (7). This cell state may also be related to the dedifferentiated cell state associated with decreased MITF expression and enhanced c-JUN expression capable of promoting feed-forward responses to TNFα with associated inflammatory cytokine secretion, and recruitment of immunosuppressive myeloid cells (8).

In addition to altered connection to immune signaling, the TBK1/IKKϵ–sensitive subtype exhibited lower levels of PGC1α, a member of the peroxisome proliferator γ family of transcriptional coactivators, involved in regulating mitochondrial biogenesis. As might be expected, this subtype exhibited reduced mitochondrial content and function, whereas the BRAF/MEK inhibitor–sensitive subtype was enriched for genes involved in oxidative phosphorylation (OXPHOS). These findings are particularly intriguing given previous reports linking oncogenic BRAF signaling to altered mitochondrial content/function and regulation of OXPHOS (9). Therefore, metabolic reprogramming may be a feature of BRAF-mutant melanoma that is altered as part of the cell state change associated with shifting toward the AXL/NF-κB^hi cell state associated with TBK1/IKKϵ sensitivity.

The authors also provide insights into the downstream effectors of TBK1/IKKϵ signaling, including regulation of cell survival pathways (e.g., AKT), as well as suggesting that sensitivity to TBK1/IKKϵ inhibition is associated with an altered epigenetic state defined by chromatin reorganization. The authors initially identified higher levels of 1-methylnicotinamide (1-MNA) in the TBK1/IKKϵ sensitive subset in an impressive metabolic profiling effort, and link this metabolite to altered histone methylation and suggest a subsequent impact on chromatin structure. Although altered chromatin structure was not shown directly, the authors demonstrated correlation with NNMT expression and H3K27 trimethylation in TBK1/IKKϵ–sensitive cell types. Treatment with an inhibitor of the H3K27 methyltransferase EZH2 also sensitized cells to TBK1/IKKϵ inhibition. Global mass spectrometric analysis of TBK1/IKKϵ–sensitive and resistant cells ± treatment with the TBK1/IKKϵ inhibitor “compound II” identified several putative effectors, including several proteins involved in epigenetic regulation and associated with mesenchymal phenotype, suggesting a connection between innate immune signaling, altered epigenetic state, and sensitivity to TBK1/IKKϵ inhibition.

The findings presented have several potential implications. First, the association of BRAF resistance to a cell state defined by enhanced immune signaling suggests possible implications for both BRAF and MAPK therapy on the efficacy of immunotherapy. Resistance to MAPK pathway inhibition has previously been shown to drive immune evolution with loss of CD8 effector T cells (10), and transcriptional changes associated with innate resistance to PD-1 blockade include higher expression of AXL and EMT-associated transcription factors (11), in addition to immunosuppressive cytokines and proangiogenic factors. Thus, resistance to MAPK pathway inhibition may promote cell-state changes that render cells resistant to both targeted therapy and immune therapy. Conversely, therapies that sensitize melanoma cells to MAPK pathway inhibition may similarly enhance response to PD-1 blockade. This would be particularly exciting for uveal melanoma, in which BRAF mutations are exceedingly rare and which is largely resistant to MAPK pathway inhibition and immune checkpoint blockade. The authors confirmed TBK1/IKKϵ sensitivity in a subset of uveal melanomas using TCGA data and confirmed sensitivity to TBK1/IKKϵ inhibition using several uveal melanoma cell lines.

Although this study demonstrates a number of exciting findings, several unanswered questions remain. Perhaps most fundamentally, the mechanism whereby cell-state changes associated with upregulation of AXL promote TBK1/IKKϵ sensitivity is still unknown. Although Eskiocak and colleagues provide correlative data regarding altered histone methylation in this subtype, the specific molecular mechanisms underlying TBK1/IKKϵ sensitivity in this subgroup will require further investigation to define the upstream pathways beyond correlation with NNMT and EZH2 protein levels. In addition, the precise identity of proximal upstream regulators and downstream effectors of TBK1/IKKϵ will require further refinement. The authors demonstrate coimmunoprecipitation of TBK1 with YAP and LATS1 using overexpressed TBK1; however, the physiologic nature, subcellular localization, and regulation of these protein–protein interactions remain unclear. Moreover, the inhibitors utilized by these authors are tool compounds, and further validation will require the use of more potent/selective clinical stage drugs. The relevance of TBK1/IKKϵ in other molecularly defined subtypes of melanoma also requires further investigation. For example, 15% to 30% of melanomas harbor mutations in NRAS, and a study from the Aplin laboratory has already suggested a potential role for combined TBK1/MEK inhibition in NRAS-mutant melanoma (12). A phase II study of binimetinib (MEK162) demonstrated modest clinical activity in patients with NRAS-mutant melanoma (13), raising the question of whether dual TBK1/IKKϵ and MAPK inhibition may be superior in NRAS-mutant and/or BRAF-mutant melanoma with innate or acquired resistance to BRAF/MAPK pathway inhibition. Finally, given the importance of the immune system in influencing response to both small molecules and immune checkpoint inhibitors, further investigation into the role of TBK1/IKKϵ in a more complex tumor microenvironment will be required to determine whether targeting TBK1/IKKϵ is an effective strategy in preclinical models with an intact immune system.

D.A. Barbie is a consultant/advisory board member for N-of-One. No potential conflicts of interest were disclosed by the other author.