Triple Therapy to Outwit the BRAF Oncogene

Since the discovery of mutant BRAF as a major driver oncogene in 2002, waves of progress have led to improved therapy, yet resistance still generally wins out in the end. Single-agent BRAF inhibitors (BRAFi) revealed both the vulnerabilities of certain BRAF-mutant tumors and also the tenacious resurgence of rewired tumors. Addition of MEK inhibitors (MEKi) improved the durability of tumor regressions, yet again eventual tumor progression prevailed (1). In parallel, detailed investigation of tumor-resistance mechanisms also exposed a new vulnerability: feedback regulation of the mitogen-activated protein kinase (MAPK) pathway.

In this issue of Cancer Discovery, Adamopoulos and colleagues thus exploit structural chemistry and the complexity of BRAF^V600E signaling to devise a novel approach to therapy of BRAF-driven tumors (2). The innovation here is to add a dimer-selective BRAFi (dBRAFi) to a monomer-selective BRAFi (mBRAFi) plus MEKi optimized to disrupt the BRAF–MEK complex (triple therapy). This triple therapy is attractive because it exploits several arms of the RAF signaling circuit to both overcome adaptive resistance in the tumor and ameliorate unintended side effects in normal cells. If it plays out correctly, this approach can achieve a relatively high therapeutic index, although three different compounds with intrinsic liabilities are used.

The mBRAFi/MEKi combo suppresses ERK, but concomitantly leads to an increase in RAS activation due to relief of negative feedback; activated RAS enhances dimerization of BRAF, which is not inhibited by mBRAFi. The adaptive activation of RAS thus introduces a liability in the therapeutic benefit of the mBRAFi/MEKi combo and supports the important benefit of adding dBRAFi. The structural implications of these combinations are discussed below, and the precise targeting of distinct RAF conformations is critical and largely understood (Fig. 1A). RAF molecules in cells exist in pools with distinct conformations and the structural heterogeneity limits the effectiveness of a single-conformation inhibitor. Combining two conformationally complementary RAF inhibitors makes both structural and pharmacologic sense.

Before wading into the challenging waters of cocktail therapies, it is useful to consider the current clinical landscape for BRAF-mutant melanomas. FDA-approved therapies started with vemurafenib single-agent therapy and were substantially enhanced by double-agent therapy such as dabrafenib–trametinib. Indeed, despite the much-lauded promise of immunotherapy, mBRAFi/MEKi dual therapy has been remarkably effective (3). Subsequent clinical studies revealed the promising result that vemurafenib/cobimetinib/atezolizumab triple therapy shows improved survival over vemurafenib/cobimetinib dual therapy (4), yet another triple therapy (dabrafenib/trametinib/spartalizumab) did not demonstrate benefit versus dual therapy (https://bit.ly/2YsIgCq). Nonetheless, there is significant unmet need in the BRAF-mutant tumor space.

Does the MAPK pathway hold yet untapped therapeutic opportunities beyond the current mBRAFi/MEKi backbone treatment? To answer this question, researchers have focused on raising the fitness barrier for BRAF-mutant tumors by exploring lateral blockade of MAPK signaling through pan-RAF inhibition (5) and deeper vertical pathway blockade with ERK inhibitors (6) or MEKi that affect BRAF–MEK interactions (7). These approaches thus far have had limited success, because indiscriminately blocking the MAPK pathway, though effective in forestalling resistance, also causes stress on normal tissues and dose-limiting toxicities.

Adamopoulos and colleagues turn their attention to a group of type II kinase inhibitors (including sorafenib and regorafenib) that have previously been viewed as nonviable BRAF-targeting agents because of their minimal activities against BRAF-mutant tumor cells. These compounds show selective inhibition of BRAF dimers as opposed to monomers, the preferred target of the type I inhibitors vemurafenib, dabrafenib, and encorafenib (Fig. 1A). This property primarily derives from the diaryl-urea or -amide tails that they share. Extensive molecular dynamics simulations suggest that these tail moieties keep the αC-helix “in” by exposing and accessing the allosteric back pocket in BRAF dimers. In BRAF monomers, the orientation of the αC-helix is more dynamic, so access to the allosteric back pocket is impeded—this structural mechanism was used to explain the dimer selectivity as defined in this paper. Note αC-helix stabilization also contributes to the activity against many other kinases that also have the allosteric back pocket, hence the less selective nature of some dBRAFi (sorafenib and regorafenib in particular) versus other kinases.

These compounds provide Adamopoulos and colleagues with the tools needed to test a novel rational drug combination that explicitly addresses RAF dimer–mediated mBRAFi/MEKi escape. Triple combination of dBRAFi with mBRAFi/MEKi results in potent suppression of both MAPK signaling and tumor growth in several models that resist mBRAFi/MEKi treatment. Although the alternative double combination dBRAFi/MEKi is similarly efficacious to the triple combination, weight loss indicative of toxicity is observed in mice receiving the double combination but not in those receiving the triple combination. Cotargeting two conformational states of RAF thus appears to have differential effects on tumor and normal cells.

Why would the triple combination be tolerated by noncancer cells? Perhaps “monomer” and “dimer” are simplified representations of a much more complex array of states RAF can adopt, supported and dynamically regulated by a cast of auxiliary molecules. Alone, both mBRAFi and dBRAFi have minimal effect on normal cells. The preponderance of RAF signaling in BRAF wild-type cells occurs through RAS-induced RAF dimers, yet nondividing cells rarely have active RAS. When the RAS/MAPK pathway is activated upstream, the presence of mBRAFi may reduce the inhibitory effect of dBRAFi/MEKi in BRAF wild-type cells to maintain normal MAPK signaling. This paradoxical activation phenomenon has been observed among all clinically approved monomer-selective BRAFi (8). Similarly, although single-agent MEKi and likely mBRAFi/MEKi actually mute the activity of cytotoxic T cells, the addition of dBRAFi can restore and augment T cell–mediated attack on the tumor cells (9).

Can this innovative approach be addressed in patients without requiring new therapeutic compound development? In principle, yes. Yet the compounds that have been used in the current studies may not be optimal for this triple-therapy approach. Regorafenib, a relatively nonspecific RAF and receptor tyrosine kinase inhibitor, works in the mouse models, but likely adds many liabilities in the human trials. A few more selective compounds are currently in clinical development, such as naporafenib (LHX254) and belvarafenib (GDC-5573), yet these were primarily designed as pan-RAF inhibitors, binding to both monomeric and dimeric RAF. Therefore, the current discovery that calls for triple therapy could likely benefit from a novel third dBRAFi. This presents a new opportunity for drug-discovery efforts.

Ultimately, the success of a combination regimen depends on the quality and compatibility of the components. In the current report, the recommended triple therapy includes dabrafenib, selected based on a synergy screen of top mBRAFi drugs with the pan-RAF inhibitor (TAK632), and trametinib, the most potent MEKi in suppressing ERK phosphorylation and disrupting RAF–MEK interactions (Fig. 1B). Although the selection of dabrafenib/trametinib makes strong clinical sense, the question lingers of whether the choices would have been different if mBRAFi and MEKi were optimized in the context of the triple combination. Perhaps the inclusion of a MEKi that promotes the RAF–MEK complex (Fig. 1B) could be more effective in this context because this type of MEKi has been shown to limit adaptive resistance to MEK inhibition (7).

With the discovery of ever more selective kinase inhibitors—particularly conformation-specific inhibitors—the opportunity for safer combination regimens becomes feasible. BRAF oncogene–dependent tumors illustrate the insidious nature of advanced cancers: potent inhibition of the driver pathway can be readily circumvented by adaptive or acquired cancer-resistance mechanisms. Thus, future cancer treatment regimens will require cocktails that preempt resistance pathways. The innovative approach described above attacks the driver pathway through three distinct tactics. In principle, this could be accomplished with a pan-RAF inhibitor equally potent against both BRAF monomers and dimers, but the therapeutic index is likely to be lower than combining two specific inhibitors. Adding immune-checkpoint inhibitors has also recently become feasible. Finally, strategies that add inhibitors of orthogonal pathways have so far lacked therapeutic index, but once the liabilities of such combinations are understood and circumvented, durable cancer treatments, perhaps even cures, will be within reach. The triple therapy highlighted in the present article provides a promising leap forward, and we look ahead to clinical validation of this strategy.

No disclosures were reported.