E6201 is a novel MEK1/2 inhibitor which inhibits selected cancer-specific kinases that is currently in clinical trials for solid tumours and, as a result of the data presented herein, is undergoing Phase I expansion in BRAF-mutant malignancies (NCT00794781, ClinicalTrials.gov). In the current study, we established a diverse cell line panel to not only represent the known genetic heterogeneity in melanoma, but also to enrich for rare mutations or genotypes in which to test the effectiveness of E6201 in vitro and in vivo. From this genetically diverse panel, we demonstrate for the first time that sensitivity to MEK1/2 inhibition in vitro correlated with wildtype PTEN suggesting parallel signalling of the PI3K/Akt/mTOR pathway may play a role in the resistance of melanoma cell lines to E6201 and MEK1/2 inhibitors in general. To this end we demonstrate that concurrent targeting of the Ras/Raf/MAPK and the PI3K/Akt/mTOR pathways was more effective than targeting either of the pathways alone in all six cell lines studied with the greatest synergy observed in E6201 resistant cell lines. These results underscore the power of heterogeneous cell line panels, such as the NCI60, to identify potential biomarkers of sensitivity and resistance in a clinical setting.
There is a general consensus that genomic analysis of tumours through The Cancer Genome Atlas (TCGA) and the International Cancer Genome Consortium (ICGC) will identify the core pathways activated in each tumour. Previous work in pancreatic cancer indicates that only 12 pathways need to be activated. This has been interpreted as molecular targeting of only a few pathways may be needed to effectively treat cancer. Emerging N-Ras/BRAF/ERK data would suggest that some therapies will only work on pathways activated at a certain “node”. For example, melanoma cells demonstrate marked differences in response to MEK1/2 inhibition, with BRAF and RAS mutational status thought to predict sensitivity and resistance, respectively. Melanomas harbouring mutant BRAF and wildtype RAS are intimately dependent on ERK signalling for their growth and survival and selective RAF inhibition in these lines efficiently blocks ERK activation and growth. Conversely, RAF inhibitors paradoxically enhance ERK activation and proliferation in BRAF-wildtype, RAS-mutant melanoma (and normal) cells through a mechanism that involves the interaction of these drugs with RAF dimers[7, 30, 31]. In this setting, concurrent treatment with a MEK inhibitor may prevent this paradoxical activation[21, 32].
The exquisite sensitivity of BRAF mutant cell lines to E6201 is consistent with that reported for other MEK inhibitors, including CI-1040 and AZD6244 (ARRY-142886). Similar to these MEK inhibitors, RAS mutant cell lines do not display the same sensitivity to E6201 as BRAF mutant cell lines[11, 13]. It is possible that the resistance of RAS mutant tumour lines in this study and others is the result of compensatory signalling by a parallel or non-canonical pathway, such as PI3K/Akt/mTOR. Indeed, the importance of intact PI3K signalling has recently been established for Ras-driven lung tumourigenesis in vivo. Interestingly, those cell lines with wildtype BRAF and RAS were not all resistant to E6201 in contrast to previously published data, suggesting that these cell lines may carry activation of the MAPK pathway through additional mechanisms, such as receptor tyrosine kinase or MEK1 activation. Perhaps only the combination of genome-wide expression profiling, exome mutation data and phospho-protein status will allow us to unravel these complex pathway interactions and their relative roles in drug sensitivity.
Strangely, despite correlating BRAF mutational status to anti-tumour activity with E6201, phosphorylated ERK1/2 levels did not correlate with the magnitude of cell growth inhibition. Similarly, the cytostatic response of melanoma cell lines to other MEK inhibitors has been shown previously not to correlate with pERK levels before or after treatment. Taken together these results support the notion that the upstream mechanism of ERK activation is important in predicting sensitivity to MEK inhibition. These findings also suggest that the cytostasis induced by MEK inhibition may be mediated by modulation of parallel signalling pathways potentially via ERK-mediated autoregulatory processes. To this end, Gopal and co-workers demonstrated reduced efficacy of MEK inhibition in melanoma cell lines as a result of PI3K pathway activation via a MEK-IGF-1R-mediated feedback loop.
Consistent with the role of the MAPK pathway in G1/S transition, E6201 exerted cytostatic effects, resulting in G1 arrest in vitro and tumour growth inhibition in vivo. E6201 also induced cell death in the majority of E6201-sensitive cell lines. It would be interesting to perform a functional genomics screen in those cell lines that only showed growth arrest but not cell death to identify the genes or pathways that could be targeted alongside MEK to induce synthetic lethality. There are previous reports of MEK inhibitors leading to cell death in a subset of sensitive melanoma cell lines. For example, CI-1040 treatment resulted in cell death in 1 out of 4 melanoma cell lines evaluated, and cell death in melanoma cell lines has also been reported with its daughter compound, PD0325901. The MEK inhibitor UO126 has also been reported to lead to caspase-independent cell death in melanoma cell lines. Thus, the cell death we see upon E6201 treatment reflects the potential for MEK inhibition to result in cell death in a specific subset of melanoma cell lines. The cytocidal activity of E6201, however, may also reflect the “multi-target” nature of E6201, such that the cell death observed is due to inhibition of other cancer-specific kinases, such as Src. Indeed, while treatment of melanoma cell lines with the Src inhibitor dasatinib has been shown to inhibit proliferation and invasion[37, 38], in some melanoma cell lines it did induce apoptosis. Although clinical responses have been seen in a subset of patients in Phase I and II trials of Dasatinib, biomarkers that predict sensitivity have not yet been identified[40, 41]. To validate our findings with E6201 in monolayer culture, we created mouse xenograft models. We hypothesized that E6201 would induce tumour regression in xenografts of sensitive melanoma cell lines, as most of the sensitive melanoma lines in our panel demonstrated cell death (Annexin positivity) in response to E6201 in vitro. To this end, we evaluated the in vivo activity of E6201 in two melanoma cell lines that exhibited a cytocidal response (MM540, MM604) and two melanoma cell lines that exhibited a cytostatic response (SKMEL13, BL) to E6201 in vitro. E6201 dose-dependently inhibited tumour progression in all four of these melanoma xenografts. Furthermore, transient regression was also observed in those cell lines that demonstrated a cytocidal response to E6201 in vitro. This is in accordance with previous work showing transient, partial tumour regression in BRAF mutant xenografted tumours with MEK1/2 inhibition[13, 15]. Furthermore, higher doses of inhibitor were required to limit tumour progression in BRAF wildtype and also NRAS mutant melanoma xenografts.
The cell line panel in this study was selected to include a subset of melanoma cell lines with PTEN mutations so that we could evaluate whether PTEN mutational status was associated with resistance to E6201. PTEN is a tumour suppressor protein and an important negative regulator of PI3K signalling as it inhibits Akt phosphorylation and activation indirectly by hydrolysing the secondary messenger phosphatidylinositol 3,4,5-trisphosphate (PIP3). Indeed, using this cell line panel, we found that insensitivity to E6201 was not only associated with mutant PTEN but also high phospho-Akt levels. This finding is consistent with the pro-survival function of Akt signalling and has been observed previously in lung cancer as well as melanoma. Interestingly, two of our resistant cell lines demonstrated no basal PI3K/Akt activation, suggesting an alternative pathway to resistance. It is possible, however, that these resistant cell lines simply activated PI3K/Akt in response to MAPK inhibition, as observed by Gopal et al. in melanoma cell lines. Conversely, E6201 induced cell cycle arrest and cell death in some cell lines with constitutively active Akt, suggesting that although high pAkt does correlate with E6201 insensitivity, cell lines with high pAkt (as well as mutant PTEN) can still undergo a cytocidal response to E6201. Nonetheless, our findings highlight the possible clinical utility of mutational and oncogenic pathway screening to stratify patients to particular treatments.
PI3K inhibitors have previously been shown to be effective in melanoma cell lines not only in combination with MAPK inhibitors[9, 12, 45, 46], but also in monotherapy. In a mouse model of cutaneous melanoma, Bedogni and colleagues demonstrated that combined targeting of MAPK and PI3K significantly decreased tumour development and incidence more so than either agent given alone. Our findings confirm and expand on this previous work. We show that inhibition of the PI3K pathway in E6201-resistant cell lines with high levels of phosphorylated Akt can sensitize these cell lines to E6201. Indeed, synergy between the PI3K inhibitor, LY294002, and E6201 was evident in all 6 cell lines tested, irrespective of PTEN mutation status, pAkt levels, or E6201 sensitivity. Interestingly, the greatest enhancement of E6201 activity by LY294002 occurred in those cell lines that were resistant to E6201 alone. On this note, multiple pharmaceutical companies are testing the effectiveness of combined MEK inhibition and PI3K or AKT inhibition in solid tumours including melanoma. There is also a Phase II trial testing the efficacy of the AZD6244 MEK inhibitor and MK-2206 AKT inhibitor in patients with relapsed BRAF V600E melanoma (NCT01510444, ClinicalTrials.gov).
Recent experience with vemurafenib has demonstrated that personalized cancer therapy can have a significant impact on patient response in this emerging era of molecularly targeted therapy. It is yet to be determined, however, whether MEK inhibitors can also impart meaningful clinical benefits to melanoma patients. To this end, recent preliminary results from a phase I clinical trial of the MEK1/2 inhibitor GSK1120212 in selected solid malignancies with a high frequency of BRAF mutation (melanoma, pancreatic, non-small cell lung and colorectal cancers) were impressive with just under three quarters of BRAF mutant melanoma patients demonstrating either a partial response or stable disease with therapy. Furthermore, several phase I trials (NCT01271803, NCT01072175 and NCT01231594, ClinicalTrials.gov) are currently assessing dual BRAF and MEK inhibition to target this oncogenic pathway at multiple levels.