Most uveal melanomas harbor mutations in Gαq and show constitutive MAPK activation. Although MEK inhibition has some efficacy against uveal melanoma, clinical responses are typically poor. The Gαq inhibitor-MEK inhibitor combination showed prolonged suppression of MAPK signaling in preclinical uveal melanoma models and led to improved therapeutic responses.
See related article by Hitchman et al., p. 1476
In this issue of Clinical Cancer Research, Hitchman and colleagues demonstrated that the combination of a Gαq inhibitor (YM-254890) and a MEK inhibitor had synergistic antitumor activity against Gαq-mutant uveal melanoma cell lines in vitro and in vivo (1).
Uveal melanoma is the most common cancer of the eye, arising from the oncogenic transformation of melanocytes of the uveal tract. Although few patients with uveal melanoma have evidence of disseminated disease at initial diagnosis, up to half will develop distant metastases—overwhelmingly localized to the liver. Once uveal melanoma becomes established in the liver, the prognosis is very poor. At this time, there are no FDA-approved therapies for advanced uveal melanoma and novel treatment strategies are urgently needed. More than 90% of uveal melanomas harbor activating mutations in the small G-proteins GNAQ or GNA11 which drive activation of multiple signaling pathways including the MAPK pathway (2). Introduction of mutant Gαq alone does not lead to oncogenic transformation of uveal melanocytes, with studies pointing to a role for cooperating, mutually exclusive “progression mutations” in the H2A ubiquitin hydrolase BAP1 or hemizygous change-of-function mutations in spliceosome factor 3b (SF3B1) or eukaryotic translation initiation factor 1A X-linked (EIF1AX).
In agreement with previous studies, the majority of the GNAQ/GNA11 mutations identified in the uveal melanoma patient cohort were at the Q209 position, with R183 mutations occurring less frequently. Mutations at these two residues are associated with impaired GTP hydrolysis and lead to Gαq adopting a constitutively active GTP-bound state (2). One additional G48L mutation in Gαq was also identified, and subsequently confirmed by analysis of The Cancer Genome Atlas and other datasets. Mutations at the G48 position of Gαq are located in the phosphate-binding loop, and have analogy to G12 position mutations in RAS. The high frequency of Gαq mutations in uveal melanoma, along with their likely role in uveal melanoma progression makes mutant Gαq an attractive target for uveal melanoma–targeted therapy. At this time, the majority of FDA-approved targeted therapies are kinase inhibitors, with considerably less progress being made in the development of small-molecule GTPase inhibitors. The recent development of mutant allele-specific inhibitors for KRAS G12 mutants has suggested that targeting mutant GTPases may also be possible. Although mutation-specific Gαq inhibitors analogous to KRAS inhibitors are not currently available, multiple natural products that can bind to and inhibit G-protein function do exist. These molecules, which tend to be derived from bacteria, also inhibit wild-type G-protein function in normal cells (e.g., pertussis toxin), making therapeutic windows very narrow and toxicity a potential issue. This study focused upon YM-254890, a compound first isolated from the culture broth of Chromobacterium sp, that inhibits ADP-induced platelet aggregation through inhibition of P2Y1 receptor–driven (Gαq-mediated) calcium release. To address the potential utility of YM in a disease-relevant setting, Hitchman and colleagues generated a series of Gαq mutants which they expressed at physiologically relevant levels in immortalized mouse melanocytes. Use of these models allowed the authors to circumvent issues associated with overexpression of mutant of Gαq in HEK-293 cells, in which the YM had little inhibitory activity. The authors demonstrated that the Gαq mutants drove transcriptional programs and signaling pathways in the Melan-a cells that were more associated with uveal melanoma than cutaneous melanoma. Having demonstrated that their model system was analogous to uveal melanoma they next demonstrated the potential of YM to inhibit MAPK signaling and noted a significant pathway inhibition in cells expressing mutant Gαq and CYSLTR2 but not those expressing mutant BRAF, NRAS, or KRAS. Inhibition of MAPK signaling in RAF- or RAS-mutant tumors is frequently associated with relief of feedback inhibition, a rebound of ERK signaling and a diminution of the drug efficacy. Similar findings were observed in uveal melanoma cells treated with YM with the initial suppression of RasGRP3, CRAF, and ERK phosphorylation being relatively short in duration, and the pathway recovering by 24 hours (Fig. 1). RNA sequencing analysis confirmed these findings and demonstrated that YM did not durably suppress MAPK transcriptional outputs. In cutaneous melanoma, the recovery of MAPK signaling can be limited through vertical pathway inhibition in which BRAF inhibitors are combined with inhibitors of MEK or MEK inhibitors are combined with ERK inhibitors. In BRAF-mutant cutaneous melanoma, the BRAF-MEK inhibitor combination is associated with sustained MAPK pathway suppression, fewer off-target effects and more durable clinical responses (∼35% survival at 5 years; ref. 3). Similarly improved responses have also been reported in BRAF-mutant colorectal carcinoma following treatment with the BRAF-MEK-EGFR inhibitor triplet combination.
The authors next asked whether MEK inhibition could increase the efficacy of YM in their engineered Gαq-driven melanocytes and uveal melanoma cell lines. As expected, the combination of YM and trametinib enhanced the suppression of phospho-ERK levels and limited transcriptional output from the MAPK pathway, leading to synergistic effects upon cell proliferation in vitro. These enhanced antitumor effects were confirmed in uveal melanoma xenograft models, with the YM-MEK inhibitor combination being more effective than either drug alone. Pharmacodynamic analyses showed that the Gαq-MEK inhibitor combination was effective at suppressing reactivation of MEK and CRAF (Fig. 1).
There is some precedent for developing MEK inhibitor–based targeted therapy combinations for uveal melanoma. In a phase II trial for advanced uveal melanoma, the MEK inhibitor selumetinib improved progression-free survival compared with either dacarbazine or temozolomide. However, a subsequent phase III double-blinded trial of selumetinib plus dacarbazine showed no improvement in PFS compared with dacarbazine alone (4). Although never directly demonstrated in clinical uveal melanoma specimens, it is likely that poor responses to MEK inhibition result from the relief of feedback inhibition and signaling adaptation through the MAPK and other pathways. Recent work from our group suggested that MEK inhibition in uveal melanoma cells led to increased receptor tyrosine kinase signaling (IGF1R and ROR1/2) and upregulation of AKT signaling (5). Interestingly, MEK inhibition in this context was also associated with the increased expression of multiple GPCRs, including the endothelin B receptor (5) and subsequent Gαq-mediated YAP signaling. In this context, escape from MEK inhibition could be limited through concurrent pan-HDAC inhibition. There is mounting preclinical evidence that multiple FDA-approved agents improve the efficacy of MEK inhibition with other studies identifying protein kinase C (PKC) inhibitors, decitabine (a DNA methyltransferase inhibitor) and chloroquine (an autophagy inhibitor) as potential combination partners. Determining which of these combinations are safe, tolerable, and efficacious in advanced uveal melanoma is a priority. The relative rarity of uveal melanoma makes large-scale randomized clinical trials a challenge, necessitating close collaboration between high-volume cancer centers and a robust system for prioritizing the most likely “winners” for future clinical evaluation. The study from Hitchman and colleagues adds to this growing knowledge base and identifies a promising new drug combination for future clinical development.
K.S.M. Smalley reports personal fees from Elsevier outside the submitted work. No disclosures were reported by the other author.
This work was supported by grants from Florida Department of Health (7BC05) and NCI Cancer Center Support Grant P30-CA076292 (both to K.S.M. Smalley).