Targeting of signaling pathways has come of age in cancer research, with sustained effort in understanding signaling identifying a number of therapeutic targets that are beginning to improve clinical outcomes for patients with cancer. Leading the way has been one of the most important pathways in cancer—the RAS/RAF/MEK/ERK pathway. Signaling is typically complex, with many pathways showing multiple points of divergence to intersect with other molecules and signaling pathways. One of the best-understood and most simple pathways is the RAS/RAF/MEK pathway, summarized in Figure 1. In this pathway multiple, signals are funneled into MEK and ERK kinases, allowing a nodal point for therapeutic targeting.
A wide range of cell-surface molecules activate RAS (KRAS, NRAS, and HRAS), a family of GTPases that act as a molecular switch, turning on downstream RAF protein kinases (BRAF, CRAF, and ARAF). The dominant substrates of RAF kinases are the MAPK/ERK kinases, MEK1, and MEK2. Intriguingly, MEK kinases appear to have only one main substrate extracellular signal–regulated kinase (ERK). This is unusual for protein kinases that are often relatively promiscuous but has led to the use of phosphorylation of ERK to identify small molecules capable of inhibiting signaling of the RAS/RAF/MEK/ERK pathway.
Downstream of ERK, the pathway becomes more complex due to ERK’s many substrates.1 Activation of ERK generates extensive changes in gene expression mediated by transcription factors that control cell cycle progression, differentiation, protein synthesis, metabolism, cell survival, cell migration, and invasion and senescence.1 Indeed, activation of ERK leads to cells acquiring many of the hallmarks of cancer. Moreover, ERK directly regulates by phosphorylation pro-apoptotic molecules such as BAD.2 As such, targeting the pathway is an extremely attractive option to overcome the malignant phenotype, and it is not surprising that upstream components of the pathway, such as the RAS GTPases and RAF kinases, are potent oncogenes.
RAS Mutations in Cancer
RAS proteins containing activating mutations that prevent cycling between GTP and GDP and lead to constitutive activation of RAF kinases in cells are the most common oncogenes in human cancer.3 It is estimated that approximately 30% of human cancers contain such mutations, resulting in a staggering 3 million new cancers diagnosed worldwide each year with RAS mutations. Importantly, mutated RAS proteins are also the most potent of human oncogenes, leading to profound changes in cells that in cooperation with other events can transform normal cells into fully malignant cancer cells.4 It is no wonder that targeting RAS has attracted the interest of the pharmaceutical industry and remains a holy grail of the therapeutic targeting of oncogenes.
RAS mutations have been described in both hematologic and solid-tumor malignancies. Different cancers based on cell type of origin show a propensity to mutate different RAS isoforms. For example, NRAS mutations dominate in hematologic malignancies and melanoma, whereas KRAS is the dominantly mutated isoform in colorectal and lung cancers.5 The reason for this is not clear but may relate to levels of expression of the isoforms in different cell types or differences in the capacity of each isoform to activate downstream signaling in a cell type–dependent manner. So despite the fact that RAS mutations have been described in the vast majority of types of cancer, the cellular context remains important with respect to which isoform is mutated.
Cancers with the most frequent RAS mutations are pancreatic cancer (90%), colorectal cancer (40%), non-small cell lung cancer (30%), bladder cancer (30%), peritoneal cancer (30%), cholangiocarcinoma (25%), and melanoma (15%). In contrast, lymphomas, acute lymphoblastic leukemia, hepatocellular carcinoma, osteosarcoma, and prostate cancer less commonly contain RAS mutations.5
One critical area of study involved with improved understanding of signaling in RAS-mutated cancers and more efficient exploitation of this knowledge therapeutically is the dependence of RAS signaling on RAF, MEK, and ERK kinases. Understanding this is essential because RAS can activate other signaling pathways, and attempts to target RAF, MEK, or ERK downstream of RAS would be futile if these kinases were not critical components of RAS signaling. Elegant studies from Dr. Barbacid and colleagues5 on using sophisticated gene-targeting techniques to generate RAS-less cells have been very informative. Strikingly, the proliferative defects in RAS-less cells can be completely restored with activated MEK, telling us that MEK, and by inference ERK, are indeed critical components of RAS signaling. Similar studies have also implicated cyclin-dependent kinases, critical regulators of cell cycle progression, downstream of ERK in RAS signaling6; therefore, inhibition of MEK has shown activity in RAS-mutated cancers7,8 and preclinical studies implicate inhibition of CDK4 as one approach to target cancers with RAS mutations.9 Early clinical data is promising, with the MEK inhibitors trametinib, cobimetinib, pimasertib, and MEK-162 all showing some activity in RAS-mutated cancers.10,11 Although escalation of the dose of MEK inhibitors is limited by cutaneous and gastrointestinal toxicity,10 these clinical studies support the importance of MEK as a key downstream mediator of the cellular effects of mutant RAS in cancer cells.
BRAF Mutations in Cancer
The pioneering work of Drs. Stratton, Futreal, and colleagues12 in identifying mutations in BRAF in human cancer has opened up profound new therapeutic opportunities for the management of cancer. Overall, 6% of human cancers contain activating mutations in BRAF13 that result in more than 500,000 new cases of BRAF-mutated cancer diagnosed worldwide each year.5 Similar to RAS mutations, BRAF mutations are profoundly oncogenic in cooperation with other genetic events and are capable of fully transforming normal cells. In contrast to RAS, a GTPase that has not been successfully targeted directly with small molecules, the BRAF protein kinase has been successfully directly targeted with a number of BRAF inhibitors that have striking clinical results including vemurafenib,14 dabrafenib,15 and LGX818.
Intriguingly, despite expression of the CRAF isoform in most cells, mutations in CRAF are extremely rare.5 One possible explanation for this observation might relate to structural differences in BRAF and CRAF with respect to the ability of these kinases to function as monomers or dimers, with mutated BRAF able to strongly activate downstream signaling as a monomer, a property not shared by CRAF. However, CRAF is an important component of the signaling pathway and is particularly important in mediating signaling through RAS and upstream cell-surface molecules.
BRAF is activated most commonly by single amino acid substitutions but can also be involved in translocations with other proteins resulting in active fusion proteins. Fusion proteins have been described in pilocytic astrocytomas16 and spitzoid melanocytic neoplasms.17 The various point mutations lead to variable biochemical activation of BRAF with amino acid substitutions at and around valine 600, potently activating the kinase activity as assessed biochemically in vitro and, most importantly, generating a kinase resistance to feedback inhibition that typically occurs when the pathway is activated upstream in cells with wild-type BRAF (feedback inhibition is described in greater detail further in this editorial).18 The most common mutation in BRAF by far is the substitution of valine 600 by glutamic acid (V600E), which accounts for more than 85% of the BRAF mutations in melanoma, more than 50% of the mutations in non-small cell lung cancer, and more than 95% of mutations in colorectal cancer, cholangiocarcinoma, and hairy-cell leukemia.5 Lower-activity mutations have been described in a variety of malignancies including melanoma, lung cancer, and prostate cancer, but whether these mutated proteins activate signaling of the MEK/ERK pathway sufficient to induce dependence on MEK and ERK remains untested.19
Just as RAS-mutated cells are dependent on MEK, it is clear that cells harboring the common- and high-activity BRAF mutation, V600E, are dependent on MEK and, by inference, ERK signaling for cell survival and proliferation. This is clearly validated in the landmark METRIC clinical trial that randomly assigned patients with melanoma who harbored BRAF V600E mutations to the MEK inhibitor trametinib or to dacarbazine.20 Not only did patients in the trametinib arm show tumor regressions, proving the dependence of BRAF V600E on MEK for cell survival, but there was an overall survival advantage for these patients as well. There is no doubt that melanomas with V600E mutations are dependent on MEK. However, response rates for trametinib (22%) are clearly lower than those observed for BRAF inhibitors vemurafenib, dabrafenib, or LGX818 (approximately 40%-60%).20,21 The reasons for this most likely relate to the inability to escalate the doses of MEK inhibitors to sufficiently inhibit MEK because of adverse events, whereas BRAF inhibitors can be escalated to a dose that inhibits BRAF, as well as downstream MEK and ERK signaling in the tumor, more potently.
Targeting the RAS/RAF/MEK pathway
Both BRAF (vemurafenib and dabrafenib) and MEK inhibitors (trametinib) are now approved for the treatment of BRAF-mutated melanoma, clinically validating the ability of inhibiting the RAS/RAF/MEK pathway to achieve meaningful benefit for patients. Yet a number of key questions and challenges remain. First and foremost is whether the pathway can be targeted to achieve clinically meaningful outcomes in patients with RAS mutations. The response rates to the first-generation of MEK inhibitors in RAS-mutated cancer is modest7,8 (< 10%) and may be insufficient to achieve clinically meaningful results. The first RAF inhibitor sorafenib, a type 2 inhibitor, does not have significant activity in RAS-mutated cancer. Like the newer type I RAF inhibitors, sorafenib may in fact promote signaling in RAS-mutated cells22; therefore, it is clearly not sufficient to inhibit the pathway in RAS-mutant cells. ERK inhibitors are in clinical development, but whether ERK inhibitors can be escalated to a dose to sufficiently inhibit ERK activity without significant toxicity in patients with RAS-mutated cancers is likely to be a challenge, given the relationship between MEK and ERK in signaling. As discussed further, combination therapy may be the best approach to adequately inhibit the pathway in cells and tumors with RAS mutations.
Another key question is whether the pathway can be inhibited when other upstream components of signaling are activated, most prominently cell-surface receptors including receptor tyrosine kinases and G-protein–coupled receptor (GPCR) complexes. Here the same challenges as with RAS emerge as limiting factors. These challenges are made greater by the ability of these receptors to activate other signaling pathways, most prominently the PI3K/AKT/mTOR pathway23 but also protein kinase C-signaling in the case of GPCRs.24 In these cases, dual or multiple pathway inhibition may be required. This is being pursued clinically, although to date it remains uncertain if, once again, toxicity in normal tissue will allow sufficient dose escalation to adequately inhibit the dual or multiple pathways. Early clinical results suggest this might be the case with, for example, the combination of MEK and PI3K inhibitors due to significant toxicity without impressive clinical efficacy.25
Paradoxical Activation of the RAS/RAF/MEK Pathway by BRAF Inhibitors
If targeting the RAS/RAF/MEK pathway is difficult because of toxicity in normal tissues as doses are escalated, then why can type 1 BRAF inhibitors, such as vemurafenib and dabrafenib, be delivered at doses that provide such profound efficacy? The answer was a surprise and relates to the ability of these compounds to potently inhibit activity of mutated BRAF proteins while actually activating RAF-dependent signaling in normal cells.26,27 Therefore, these compounds display an ideal property for a drug that inhibits signaling: potent inhibition of signaling in cancer with no significant inhibition of signaling in normal cells.
Type I inhibitors bind to the kinase in its active conformation (type 2 inhibitors bind to the inactive conformation) and result in the formation of active heterodimers with unbound wild-type BRAF or unbound CRAF.26,27 These heterodimeric complexes between an inhibited molecule and an active molecule will result in enhanced activation of the active molecule when upstream signaling is activated by mutation in RAS or by activation of cell-surface receptors (Fig. 2). As normal cells use ligand-dependent activation of receptors to regulate cell proliferation of normal tissues (e.g., EGFR in the skin), type I BRAF inhibitors can activate MEK and ERK signaling in normal cells, resulting not only in the absence of hypoproliferative toxicities but actual promotion of hyperproliferative toxicities.14,28
This phenomenon of “paradoxical activation” reaches the point of clinical importance when considering the ability of type I BRAF inhibitors to promote proliferation of RAS-mutated– cutaneous cancer and possible other RAS-mutated cancers.14,28,29
Feedback Control of the Output of the RAS/RAF/MEK Pathway—Implications for Treatment
A increasingly studied concept in the understanding of signaling in normal and malignant cells is the ability of pathways to regulate their own output through feedback mechanisms. This phenomenon may be critical to allow cells to achieve the fine balance between dysregulated signaling and uncontrolled cell proliferation (a hallmark of cancer) as well as the capacity to switch pathways on or off when needed for physiologic purposes.
Seminal work from Dr. Rosen and colleagues18 has identified a profound difference in signaling through RAF -kinases, in particular BRAF, when it is activated by mutations compared to activation by more physiologic upstream processes. Mutated BRAF phosphorylates MEK and activates ERK, with notably different downstream molecular consequences when compared with activation of BRAF or other RAF kinases by upstream processes such as engagement of receptor tyrosine kinases by their ligands.18,30 These differences can be quantitatively assessed by examination of gene expression at the mRNA level. Why is there such a fundamental difference? It appears to be related to feedback control by a series of downstream molecules that inactivate upstream components of the pathway. In particular ERK itself and the ERK-regulated phosphatases DUSPs and SPRYs can affect upstream kinases to regulate their activity (Fig. 3). However, these molecules have limited capacity to do this when BRAF is activated by mutation.18 This creates a signaling cascade profoundly different between normal and malignant cells and has contributed the exciting outcome of selectivity of BRAF inhibitors for mutant versus wild-type kinases allowing inhibition of signaling in melanoma, colorectal cancer, thyroid cancer, lung cancer, and other malignancies harboring activation mutations in BRAF at position V600 with relatively limited effects on signaling in normal cells. Other mechanisms may also be involved including inhibition of ERK, which increases expression of ligands for receptor tyrosine kinases and provides a further feedback mechanism that may reduce some of the consequences of ERK inhibition.18
What do we know about feedback control of the cells with RAS mutations? Mutations of RAS lock it in the GTP-bound form and prevent the cycling of GDP and GTP that normally occurs with wild-type RAS. This constitutively activates the RAF/MEK/ERK pathway and drives proliferation, survival, and other changes in cells that contribute to the malignant phenotype. RAS also activates other pathways, such as the PI3K/AKT and RAC1 pathways.31 However, the effects on the RAF/MEK/ERK pathway are important for transformation. Interestingly, the model presented above (Fig. 3) would suggest that activation of the RAF/MEK/ERK pathway by mutant RAS would result in feedback inhibition of RAF kinases, and indeed this appears to occur with significantly lower output of the pathway in RAS-mutated compared with BRAF-mutated cells.8 However this lower activation of ERK appears to be sufficient to mediate the consequences of enhanced signaling induced by mutant RAS because inhibition of MEK and ERK can reverse the malignant phenotypes of some RAS-mutated cells lines.
Future Prospects for Achieving Cancer’s Holy Grail: Targeting RAS
The most common mechanism of acquired resistance to BRAF inhibitors in BRAF-mutant melanoma is the acquisition of new genomic events that reactivate MEK and ERK signaling though the emergence of activating mutations in NRAS, amplification of BRAF, increased expression of a truncated form of BRAF, and in some instances new mutations in MEK. This has led to the idea that resistance may be overcome, or indeed prevented, by more effective targeting of the pathway. The leading approach to achieve better inhibition of the pathway has been the combination of type 1 BRAF inhibitors with MEK inhibitors such as dabrafefenib plus tranetinib, vemurafenib plus cobimetinib, and LGX818 plus MEK162. Both the combinations of dabrafefenib plus tranetinib and vemurafenib plus cobimetinib show activity in patients acquiring resistance to single-agent BRAF inhibition, and dabrafefenib plus tranetinib has been shown to be superior to single-agent dabrafenib in delaying the emergence of resistance with improvement in median progression-free survival from 5.8 to 9.4 months32 that has led to recent U.S. Food and Drug Administration approval of this combination.
It is worth reflecting on why this combination is successful. The ability of inhibiting the same pathway at two points may not only lead to greater inhibition of the pathway though addictive effects but may also affect feedback mechanisms in a favorable way, allowing more sustained inhibition of output of the pathway. Strikingly, the toxicity of the combination of dabrafenib and trametinib appears to be associated with less toxicity than single-agent dabrafenib,32 suggesting that the downstream inhibition of MEK activated by paradoxical activation allows for normalization of signaling in normal cells and less toxicity.
The most common mutation associated with reactivation of the pathway in patients who develop acquired resistance to BRAF inhibitors is acquisition of mutations in NRAS that activate ERK by signaling through CRAF.33-35 The addition of a MEK inhibitor in this scenario, therefore, can reduce this reactivation of MEK and ERK through direct inhibition of MEK and can also potentially affect feedback pathways to reduce signaling. As such, the combination of RAF and MEK inhibitors may have activity in RAS-mutated cancers.
Given all of this information, it has been postulated that MEK inhibitors may have activity in RAS-mutated cancers. However, results have been disappointing, with low response rates. Recently, data have emerged suggesting that feedback activation of MEK may limit activity of MEK inhibitors in RAS-mutated cancers.36 The allosteric MEK inhibitor GDC-0623 inhibits feedback phosphorylation of MEK and, therefore, appears to be more active in RAS-mutated cells. If this observation is broadly applicable in RAS-mutated cancers, it could lead to the development of the first effective targeted approach for the treatment of RAS-driven malignancy.
The other approach to target mutant RAS aims to inhibit multiple pathways activated by RAS, such as the combination of MEK inhibitors and PI3K inhibitors.37,38 However, this approach is challenging because toxicity from this combined-pathway approach is significant and contrasts to the combination of BRAF and MEK inhibitors, in which toxicity of the combination was reduced compared with BRAF inhibitors alone. The paradoxical activation phenomenon as seen with BRAF inhibitors—with direct targeting of the consequences of paradoxical activation by immediate downstream inhibition of MEK—contrasts to the diverse roles of both MEK and PI3K pathways in normal cells, with limited cross talk between the PI3K and MEK pathways. Nonetheless, RAS-mutated cells might be more sensitive to the combined inhibition of these pathways when compared with normal cells, so the combination approach continues to be worthy of further attention.
The rapid advancement in strategies to target signaling in cancer cells led by small molecules targeting the RAF/MEK/ERK pathway is fundamentally changing the approach to treating cancer. With major advances in targeting cancers with BRAF mutations now achieved, it remains pertinent to not only more effectively target this pathway to reduce toxicity and improve efficacy but also to turn our gaze to the role of this pathway in RAS-mutated cancers.