Path Forward for RAF Therapies: Inhibition of Monomers and Dimers

Summary

Current BRAF inhibitors block signaling from monomeric BRAF^V600E, but not from oncogenic RAS, which requires RAF dimerization. In this issue of Cancer Cell, Yao and colleagues elucidate why current drugs are ineffective against RAF dimers, while Peng and coworkers describe a pan-RAF inhibitor targeting both monomeric and dimeric RAF.

Mutations in RAS family members and in BRAF are important cancer drivers in greater than 30% of human malignancies, and up-regulation of canonical RAS/RAF/MEK/ERK signaling is observed in the majority of tumors. The extensive nature of oncogenic signaling through this pathway has made the identification of RAS and RAF inhibitors a top priority of drug discovery programs for over two decades. Although agents that block RAS activity remain elusive, the development of BRAF kinase inhibitors progressed steadily, with vemurafenib being the first to gain FDA approval in 2011 for the treatment of malignant melanoma driven by BRAF^V600E, the most prevalent BRAF mutation. Vemurafenib and other first generation BRAF inhibitors exhibit good efficacy against BRAF^V600E and have been touted as another success story for targeted therapeutics; however, several early observations tempered enthusiasm.

In particular, these drugs had little activity against tumors possessing RAS mutations even though the RAF kinases are essential downstream effectors of RAS (Fedorenko et al., 2011). In cell based assays, researchers further found that while these inhibitors were effective at shutting down ERK signaling mediated by BRAF^V600E, they paradoxically upregulated ERK activity in the presence of oncogenic RAS (Gibney et al., 2013). Moreover, a subset of melanoma patients treated with these drugs developed secondary malignancies, many of which arise from cells harboring pre-existing RAS mutations. Finally, the effectiveness of current BRAF inhibitors in treating BRAF^V600E-driven melanoma is short-lived, with drug resistance invariably developing, often as a result of ERK cascade reactivation (Bucheit and Davies, 2014).

The apparent limitations to the usefulness of these drugs, however, were not without a silver lining in that they stimulated a flurry of investigation that has significantly advanced our understanding of the RAF kinases. The studies by Yao et al. (2015) and Peng et al., (2015) in the current issue of Cancer Cell are no exception, providing explanations for the limited effectiveness of current BRAF inhibitors, elucidating mechanisms for how different BRAF mutants promote tumorigenesis, and describing new RAF inhibitors with broader activity.

At the time that the first generation BRAF therapies entered the clinic, signal transmission from RAS to RAF was known to be complex, with gaps still remaining in our understanding of RAF kinase activation. Through experiments characterizing the paradoxical upregulation of ERK cascade signaling induced by these drugs in RAS mutant cells, the issue of RAF dimerization came to the forefront, with inhibitor treatment appearing to promote or stablize RAF dimer formation (Lavoie et al., 2013). Through subsequent mutant analysis and structural studies, it is now know that RAF dimerization is an obligatory step in RAS-mediated RAF activation (Freeman et al., 2013). Like most kinases, to become an active enzyme the RAF catalytic domain must assume a closed conformation with the conserved DFG motif swinging in to align the regulatory spine. Under normal signaling conditions, formation of this active conformation occurs through an allosteric transactivation mechanism that is mediated by RAF dimerization and requires RAS binding to promote dimer formation (Hu et al., 2013). This allosteric mechanism has been best characterized for Ras-induced BRAF/CRAF heterodimers where BRAF activates CRAF; however, ARAF/BRAF heterodimers as well as BRAF/BRAF and CRAF/CRAF homodimers have also been observed.

Perhaps not surprising, the signaling activity of BRAF^V600E bypasses this Ras-mediated dimerization step, and structure modeling studies would suggest that the valine to aspartic acid substitution itself allows this mutant to adopt the active kinase conformation in the absence of the allosteric mechanism, thus functioning as an activated monomer (Hu et al., 2013). Although V600E is the most prevalent BRAF mutation, many other BRAF mutations have been detected in human cancers, the majority of which have increased kinase activity. In this issue of Cancer Cell, Yao et al. (2015) investigate the mechanisms by which various activated BRAF mutants promote tumorigenesis. Utilizing elegant cell based systems and mutational analysis, they find that a common property of these activated proteins is their ability to signal in a RAS-independent manner, thus evading normal mechanisms of pathway attenuation (Figure 1). Under physiological conditions, signaling through the RAS/RAF/MEK/ERK is regulated by ERK-mediated feedback inhibition (Lito et al., 2012). Through direct phosphorylation events and by increasing the expression of pathway inhibitors, ERK acts at multiple points to limit RAS-GTP levels, which in turn modulates the amplitude and duration of pathway signaling. Like BRAF^V600E, Yao and coworkers found that all oncogenic substitutions in V600 provided RAS-independence by allowing BRAF to function as an activated monomer. In contrast, other activated BRAF mutants formed constitutive homodimers that did not require RAS activity for dimerization. Moreover, they found that the mechanism by which the mutants gained their RAS independence determined their sensitivity to current BRAF drugs. More specifically, activated monomers were inhibited by these drugs but homodimeric mutants were not. The BRAF drugs currently in the clinic are classified as Type I inhibitors that bind to the active “DGF-in” kinase conformation, and further investigation by Yao and colleagues revealed that binding of these drugs to one protomer in the dimer significantly reduced the affinity for binding to the second protomer, demonstrating negative cooperativity (Figure 1). The authors go on to identify a RAF inhibitor BGB659 with equivalent efficacy against both the monomeric and homodimeric BRAF mutants; however, the IC₅₀ for monomeric BRAF was found to be higher than current BRAF drugs and BGB659 was ineffective against WT RAF dimers in cells expressing RAS mutations.

Figure 1.

Next Generation RAF Inhibitors. (Left panel) Oncogenic BRAF mutants evade normal ERK-mediate feedback inhibition and gain RAS-independence by functioning as either activated monomers or constitutive homodimers. Current BRAF drugs block signaling from activated monomers but not dimeric RAF, due to negative cooperativity in protomer binding. (Right panel) Next generation pan-RAF inhibitors target RAF monomers and dimers.

Also in this issue, Peng et al. (2015) characterize a new pan-RAF inhibitor LY3009120 that is capable of inhibiting monomeric BRAF^V600E as well as WT and mutant RAF dimers. LY3009120 binds all RAF family members with similar affinities and inhibits their kinase activity with IC₅₀s in the low nM range. Although LY3009120 strongly induced RAF dimerization in the presence of oncogenic RAS, minimal activation of MEK and ERK was observed, indicating that the activity of the induced dimers was effectively blocked. In cell culture assays and xenograph studies, LY3009120 inhibited the growth of tumor cells expressing BRAF or RAS mutations. Peng and coworkers propose that the high affinity for CRAF and the way in which LY3009120 binds the RAF kinase domain likely explains its effectiveness against both BRAF and RAS mutant cells. Structural analysis revealed that LY3009120 could occupy both protomers in a RAF dimer and that it was selective for the inactive, “DFG-out” kinase conformation, designating it a Type II inhibitor. The accessibility of the ATP binding pocket allowed by this conformation may explain how this and other Type II inhibitors, including BGB659 described above and two recently reported RAF inhibitors (Girotti et al., 2015), can bind both protomers with similar affinity.

It is clear that the path forward for more effective RAF therapies will require drugs that can inhibit both monmeric and dimeric RAF(Figure 1). Moreover, inhibitors that target all RAF members may expand their use to tumors with upstream pathway activation. Determining the efficacy of these drugs in patients as well as the potential mechanisms of drug resistance will be critical. For the first generation BRAF inhibitors, drug resistance has often involved RAF dimerization. However, now with compounds that can inhibit RAF dimers, what mechanisms will tumor cells use to evade death? Will mutations that disable inhibitor binding, such as gatekeeper mutations, now be observed? Only time will tell.