Summary: ERK inhibitors have enormous therapeutic potential against tumors that are BRAF mutant, BRAF–MEK inhibitor resistant, or RAS mutant. In this issue of Cancer Discovery, Sullivan and colleagues report on the first-in-human dose-escalation study of the ERK inhibitor ulixertinib, which they show to be well tolerated with clinical activity against a wide range of tumor types. Cancer Discov; 8(2); 140–2. ©2018 AACR.
See related article by Sullivan et al., p. 184.
The MAPK pathway is the major signal transduction cascade that regulates cell growth; its activity is frequently upregulated in cancer. Under normal physiology, activation of the MAPK pathway occurs following the binding of growth factors to their cognate receptors, and the recruitment of the small GTPase RAS, which then recruits the serine threonine kinase RAF (1). This, in turn, activates MEK, which phosphorylates ERK, the major effector kinase of the pathway. ERK then phosphorylates and activates multiple cytoplasmic substrates and transcription factors (Fig. 1; ref. 1). The MAPK cascade is highly amplified at each signaling module, so that a relatively small input at the level of RAS leads to maximal stimulation of ERK. When activated in cancer, signaling through the MAPK pathway contributes to tumor progression through multiple mechanisms, including the maintenance of activated cyclin D1/CDK4 complexes (which increases cell growth), the suppression of proapoptotic molecules such as BIM (which increases cell survival), and through direct regulation of the cytoskeleton (which contributes to invasion and metastasis). The aberrant activation of the MAPK pathway observed in cancer occurs through multiple mechanisms, including amplification and overexpression of receptor tyrosine kinases (RTK), amplification of BRAF, mutations in RAS and BRAF, and the inactivation/mutations in tumor suppressors such as NF1.
BRAF mutations occur in 7% of all cancers, and 50% of all cutaneous melanomas. To date, over 200 mutant BRAF alleles have been identified, and these have been subcategorized into three functional groups (2). The first two classes comprise the activating BRAF mutations that signal in a RAS-independent manner, as either active monomers (class I mutations, such as V600 mutations) or dimers (class II mutations, such as K601E/N/T, L597Q/V, and G469A/V/R; Fig. 1; ref. 2). The third class of BRAF mutations are kinase-dead and instead depend upon RAS for their signaling (class III, such as G466V/E, G469E, N581S, D594A/G/H; ref. 2). In cutaneous melanoma, the majority of BRAF mutations are V600E mutations, with a further 10% to 35% being non-V600 BRAF mutations. In other cancers, up to 50% of the BRAF mutations are non-V600. Among the three classes of BRAF mutations, only the class I are sensitive to BRAF inhibitors, with therapeutic strategies currently lacking for cancers with class II and class III BRAF mutations (2). Much of the MAPK-centric drug development to date has centered upon BRAF and MEK inhibitors, with the BRAF–MEK inhibitor combination now FDA-approved for multiple BRAFV600-dependent diseases, including melanoma, non–small cell lung cancer (NSCLC), and Erdheim–Chester disease. Although attempts were also made to develop MEK inhibitors for NRAS-mutant melanoma, this was halted on account of poor efficacy.
Multiple clinical and preclinical studies have shown that BRAF inhibitors abrogate the growth and survival of melanoma cells with BRAFV600E mutations. Although effective in the short term, therapeutic escape occurs, and this is often secondary to the relief of feedback inhibition in the MAPK pathway leading to increased RTK, RAS, and CRAF signaling and recovery of ERK signaling (Fig. 1). These effects can be overcome through vertical targeting of the MAPK pathway using the BRAF–MEK inhibitor combination, a strategy associated with an improved duration of antitumor response. In cases of acquired BRAF and BRAF–MEK inhibitor resistance, reactivation of the MAPK pathway is the most common mechanism of therapeutic escape, occurring in up to 70% of cases (3). It is now well established that melanoma cells have multiple mechanisms to reactivate the MAPK pathway and that even small changes in the MAPK signaling inputs can overcome the effects of BRAF inhibition. With this in mind, new strategies have been sought to suppress the recovery of MAPK signaling, with the goal of improving overall survival.
In this issue of Cancer Discovery, Sullivan and colleagues report on the phase I clinical trial of the ERK inhibitor ulixertinib and provide the first clinical evidence that ERK inhibitors are active against cancers harboring diverse BRAF mutations, as well as melanomas with BRAF–MEK inhibitor resistance (4). The clinical trial consisted of an accelerated 3 + 3 rapid dose-escalation cohort and 6 expansion cohorts. A total of 135 patients were recruited, with 108 of these being enrolled into the expansion cohort. At the recommended phase II dose of ulixertinib (600 mg), ERK was completely inhibited in whole blood (4). Among those who responded in the dose-escalation phase, the best response was a partial response in 3 of 25 patients, all of who harbored a BRAF mutation. Of note, one of these responding patients had previously failed vemurafenib and dabrafenib and continued to respond to ulixertinib for >24 months. A further 6 patients had stable disease for >6 months as their best response, including one patient with a bronchoalveolar NSCLC. In the dose-expansion cohort, 11 of 81 (14%) patients experienced a partial response. Of those who responded, 3 of 18 had NRAS-mutant melanoma, 3 of 12 had BRAF-mutant NSCLC (including BRAF V600E and L597Q mutations), and 4 of 21 had other BRAF-mutant cancers, including (L485W-mutant gallbladder cancer, V600E-mutant glioblastoma multiforme, G469A-mutant head and neck cancer, G469A-mutant small-bowel cancer; ref. 4). The clinical activity of ulixertinib was consistent with prior preclinical data in which good antiproliferative activity was primarily seen in cell lines with MAPK pathway mutations (defined here as RAS- or RAF-mutant). The spectrum of genotypes showing sensitivity to ERK inhibition was broader than that of the BRAF inhibitors, with ulixertinib being effective in cancers with both non-V600 BRAF mutations (particularly class II BRAF mutations) and in BRAF-mutant cancers other than melanoma. In many instances, this was the first report that these genotypes were clinically actionable. Whether or not this represents a broader range of actionable genotypes than has been reported for MEK inhibitors (which also have activity against L597S BRAF-mutant melanoma) remains to be determined (5).
Ulixertinib was also active in the cohort of patients with melanoma who had failed BRAF–MEK inhibitor therapy, with 3 of 19 showing a partial response (4). These observations were supported by xenograft and cell-culture studies in which ulixertinib showed efficacy against BRAF-mutant A375 melanoma cells engineered to express mediators of BRAF inhibitor resistance, such as the MEKQ56P mutation, as well as cell lines with RAS mutations (6). It therefore seems likely that ERK inhibitors may be of use in patients who have failed BRAF and BRAF–MEK inhibitor therapy, particularly when the resistant tumors remain addicted to the MAPK pathway. The potential effects of ulixertinib against BRAF and BRAF–MEK inhibitor resistance mediated through alternate resistance mechanisms, such as activation of the PTEN–PI3K–AKT pathway, were not investigated.
It is possible that the true magnitude of ulixertinib's effects against BRAF and BRAF–MEK inhibitor resistance was not seen in the current study. It is known that patients who have failed BRAF inhibitor therapy can be successfully rechallenged with the BRAF–MEK inhibitor combination and that up to 32% can experience a second durable response, provided that there is a sufficient washout period (7). Mechanistic studies have shown that certain types of BRAF inhibitor resistance, such as that mediated through increased RTK signaling, are dependent upon continuous drug selection pressure and that removal of drug restores sensitivity (8). Most of the BRAF inhibitor–resistant patients reported here by Sullivan and colleagues had relatively brief washout periods (∼6 weeks), and this may have been too short for sensitivity to be reestablished.
The evolving experience of targeted cancer therapy development suggests that resistance is inevitable when kinase inhibitors are used in a monotherapy setting, and it is likely that ulixertinib will ultimately be used as part of a drug combination. There is already preclinical evidence that the combination of ulixertinib and the BRAF inhibitor dabrafenib improves responses in BRAF-mutant melanoma xenograft models, with curative responses observed in multiple mice (6). Similarly, the combination of an ERK inhibitor with a MEK inhibitor leads to enhanced antitumor responses in cell culture models of NRAS-mutant melanoma (9). As reactivation of the MAPK pathway is the most common resistance mechanism to BRAF and BRAF–MEK inhibitor therapy, there is clearly a rationale for using ulixertinib with the BRAF–MEK inhibitor combination. As increased toxicity is a concern when drugs are combined, an argument can be made for using these three drugs on an intermittent dosing schedule, in which the inhibitors are administered in sequence, with drug holidays. This approach has already been explored in preclinical models and appeared successful at delaying resistance in melanomas for which resistance was mediated by the expansion of preexisting cells with BRAF amplification (10). Whether this would work for all melanomas, or just a subset that utilize this particular resistance mechanism, remains to be determined. The continued development of ulixertinib as an ERK inhibitor, with a tolerable safety profile, provides another avenue to treat MAPK-driven cancers that were not previously deemed actionable. This, along with the expanded possibility of new MAPK pathway–targeted drug combinations and dose schedules, also offers hope for patients whose tumors are driven by BRAFV600, non-V600 BRAF, and NRAS mutations.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
This work was supported by R21 CA198550, R21 CA216756, and P50CA168536-05 from the NIH and a Career Development Award from the Melanoma Research Foundation.