Introduction
Most melanomas have mutually-exclusive activating mutations in the mitogen-activated protein kinase (MAPK) pathway involving
NRAS or
BRAF genes in melanomas of skin primary,
c-Kit in acral and mucosal melanomas, and
GNAQ and
GNA11 in uveal melanomas [
1‐
5]. These mutations render melanoma cells independent of the normal receptor tyrosine kinase (RTK)-mediated pathway regulation, and constitutively drive melanoma cells to oncogenic proliferation and survival [
6]. The most common of these mutations is the
BRAF
V600E
mutation, present in approximately 50% of melanomas of skin origin.
BRAF
V600E
mutant cutaneous melanomas are dependent on MAPK signaling for cell-cycle progression and proliferation, and have high sensitivity to type I BRAF inhibitors and to MEK inhibitors [
7‐
10]. Very high response rates and improved survival have been noted with the administration of the type I BRAF inhibitor vemurafenib (formerly PLX4032/RG7204) to patients with BRAF
V600E mutant cutaneous metastatic melanoma [
11‐
13]. Tumor responses were dependent on the presence of the
BRAF
V600E
oncogene and efficient inhibition of the MAPK pathway as detected by decreased phosphorylation of ERK [
8]. Inhibition of the immediately downstream MEK1/2 kinases in BRAF
V600E mutant cutaneous melanoma was shown to lead to marked inhibition of cell proliferation in cell lines [
7]. The attractiveness of inhibiting at the level of MEK is supported by the very high kinase specificity of allosteric MEK inhibitors and the fact that MEK1/2 kinases are critically positioned as a funnel in the MAPK pathway downstream of the three RAS isoforms and the three RAF isoforms. Therefore, the inhibition of MEK1/2 with specific MEK inhibitors may result in blocking MAPK signaling from multiple upstream oncogenes. Preclinical studies suggest that some
NRAS-mutant cutaneous melanomas may also exhibit sensitivity to RAF or MEK inhibition [
14], whereas
KRAS mutations have conferred only marginal sensitivity [
15]. Gene expression profiling studies mapping the gene signatures downstream of a constitutively activated MAPK pathway suggested that cutaneous melanoma cell lines with NRAS mutations are less dependent in signaling through this pathway compared to
BRAF
V600E
mutant cutaneous melanoma cell lines [
10,
16], explaining in part the differential sensitivity of NRAS and BRAF mutant cells to MEK inhibitors [
7].
BRAF and
NRAS mutations are absent in melanomas arising in the uveal layer of the eye, but mutually exclusive somatic mutations in the heterotrimeric G protein alpha-subunit, GNAQ, or in GNA11, are present in the great majority of uveal melanomas [
4,
5]. It had long been noted that uveal melanomas have constitutive MAPK signaling [
17,
18], and it is now understood that it is because of the presence of GNAQ or GNA11 mutations. These mutations occur in codons 183 or 209 in the Ras-like domain and result in constitutive activation, turning the GNA proteins into dominant-acting oncogenes signaling through the MAPK pathway [
4]. GNAQ knockdown, as well as treatment with the U0126 MEK inhibitor, resulted in inhibition of MAPK signaling and loss of viability [
4]. Therefore, MEK inhibition may be a way to treat metastatic melanoma of uveal origin, a disease that has been highly refractory to most therapies tested to date.
TAK733 represents a novel and distinct inhibitor of MEK that is capable of allosteric inhibition of the RAF substrates MEK-1 and MEK-2 [
19]. This compound has been characterized extensively and shown to possess desirable drug-like attributes [
20]. In the current studies we have analyzed the sensitivity and resistance of human cutaneous and uveal melanoma cell lines to this novel MEK inhibitor, with analysis of the oncogenic driver mutations and downstream signaling alterations and functional effects.
Discussion
Initial data testing MEK inhibitors in melanoma cell lines suggested a high level and selective sensitivity in
BRAF
V600E
mutant melanoma cell lines, with low sensitivity in melanoma cell lines with other driver oncogenes [
7]. Further testing with expanded panels of cell lines has confirmed a trend towards higher sensitivity in
BRAF
V600E
mutant melanoma, but has also provided evidence that some melanoma cell lines with
NRAS activating mutations are sensitive to MEK inhibitors [
10,
14]. The higher sensitivity of
BRAF mutant cell lines compared to
NRAS mutant cell lines is generally represented in our series, but some
BRAF mutants have high resistance to the MEK inhibitor while some
NRAS mutants are sensitive. It is certainly possible that our
BRAF
V600E
mutant cutaneous melanoma panel is skewed for cell lines with natural resistance to inhibition of the MAPK pathway, since we have previously reported a similar greater than expected frequency of cutaneous cell lines resistant to the type I BRAF inhibitor vemurafenib [
9,
22]. The molecular basis for this relative high frequency of natural resistance of
BRAF
V600E
mutant cutaneous melanoma cell lines in our series is currently not well understood. Initial exploration of secondary oncogenic events in the PI3K/AKT pathway (such as PTEN deletions) did not clearly differentiate naturally sensitive and resistant
BRAF
V600E
mutant cutaneous melanomas to the BRAF inhibitor vemurafenib, but downstream signaling studies did suggest that the PI3K/AKT pathway may be involved [
9,
22]. In the current studies we noted the same phenomenon, a lack of correlation between natural sensitivity and resistance to TAK733 based solely on oncogenic analysis of the cell lines using SNP arrays or targeted oncogene sequencing for mutations frequently present in cancer. However, there was a suggestion from Western blot analyses of signaling pathways that differential effects of MEK inhibitor altering signaling through the PI3K/AKT pathway may be related to resistance. This observation may provide means to explore combinations of MEK inhibitors with PI3K or AKT inhibitors that may be useful in
NRAS or
BRAF mutant melanomas, which could be due to hyperactive receptor tyrosine kinase signaling leading to resistance [
22‐
24].
BRAF has only MEK as a substrate for activation [
6], and as discussed cutaneous cell lines with the
BRAF
V600E
mutation frequently have high sensitivity to MEK inhibitors
in vitro[
7]. However, patients with
BRAF
V600E
mutant cutaneous metastatic melanoma enrolled in clinical trials testing MEK inhibitors [
25,
26] have lower response rates than the use of the type I BRAF inhibitors vemurafenib or dabrafenib (GSK2118436) in the same population [
11,
13,
27]. The reason for this discrepancy between
in vitro and
in vivo results with MEK inhibitors is not clearly understood at this time, but it may be related to a lower therapeutic window of MEK inhibitors in the clinic compared to type I BRAF inhibitors. This could be explained by the paradoxical activation of the MAPK pathway in
BRAF wild type cutaneous cells, where type I BRAF inhibitors increase (or do not change) MAPK signaling in normal cells, while they efficiently block the MAPK pathway downstream of oncogenic BRAF
V600. On the contrary, MEK inhibitors can equally block the MAPK pathway downstream of both oncogenic and wild type BRAF. This lack of differentiation most likely causes the dose limiting toxicities (DLT) at exposures
in vivo that do not adequately block the MAPK pathway in
BRAF
V600
mutant melanoma. Despite this, MEK inhibitors are likely to have a role in the treatment of cancers with constitutive MAPK signaling from oncogenic mutations upstream of MEK. In particular the combination of MEK and RAF inhibitors may be beneficial by inducing higher MAPK inhibition in mutant cells and therefore lowering the cancer escape mechanisms and also decreasing toxicities from paradoxical MAPK activation [
28], such as the development of cutaneous squamous cell carcinomas [
29].
The majority of uveal melanomas bear a mutually exclusive activating mutation in either
GNAQ or
GNA11, resulting in overlapping functions in melanoma cells with the constitutive upregulation of the MAPK pathway [
5]. In preclinical models it was shown that at least the
GNAQ mutation resulted in sensitivity to downstream blocking of the MAPK pathway with a MEK inhibitor [
4]. Our data demonstrating the sensitivity of uveal melanoma cell lines to TAK733 provides further evidence that it may be a clinical strategy to use MEK inhibitors to treat metastatic uveal melanomas. However, the same issues of a lack of correlation between the
in vitro and clinical results when blocking oncogenic MAPK signaling using MEK inhibitors may apply to uveal melanomas.
The differential uptake of 3H-radiolabeled compounds that are trapped intracellularly upon metabolic processing allows testing their potential future use as PET probes in the clinical development of a new agent. It is anticipated that these radiolabeled metabolic probes can provide non-invasive pharmacodynamic information with the use of clinical PET scanners. In our studies, the highly sensitive cell lines had a decrease in the uptake of radiolabeled thymidine and deoxy-glucose that seemingly correlated with the cell viability and cell cycle results. However, there were variable changes in the highly resistant cell lines that did not directly correlate with the cell viability assay results (ex. M263 with marked decrease in the uptake of both tracers despite its resistance to TAK733). The metabolic tracer uptake studies were performed at a slightly earlier time point than the proliferation/viability assays to capture earlier events, and may be the reason of the discrepancy in results. These results raise the point that earlier PET scans with these tracers to detect early pharmacodynamic changes may not fully predict the later restaging imaging CT scan results.
In conclusion, inhibition of oncogenic MAPK signaling through MEK1 and MEK2 by TAK733 results in antitumor activity in vitro against a large subset of melanoma cell lines. We confirmed the previously reported cytotoxic effect of a MEK inhibitor against cell lines with BRAF
V600E
mutations, but in addition the cytotoxic activity was evident in a high proportion of melanoma cell lines with NRAS, GNAQ or GNA11 driver mutations. The antiproliferative and cell metabolism effects of this MEK inhibitor against melanoma cell lines can be detected with metabolic probes that could be tested with caution in the clinical development of this agent using PET imaging.
Material and methods
Reagents and cell lines
TAK-733 was obtained under a materials transfer agreement (MTA) from Millennium Pharmaceuticals, Inc. (Cambridge, MA) and dissolved in dimethyl sulfoxide (DMSO, Fisher Scientific, Morristown, NJ) to a stock concentration of 10 mM. The cutaneous melanoma cell lines of the M series were established from biopsies of metastatic melanoma of cutaneous origin as previously described [
9] under the UCLA IRB approval #02-08-067 following the Declaration of Helsinki. SKMEL28, Wn1366 and SBCL2 were obtained from the American Type Culture Collection (ATCC, Rockville, MD). The uveal melanoma cell lines of the Mel20 series were established from fine needle aspirates of primary uveal melanoma lesions or from a metastatic uveal melanoma lesion (Mel20-09-196), obtained under the UCLA IRB approval #04-12-084. In the case of uveal melanoma cell lines, cells were cultured in DMEM with L-glutamine and 4.5 g/liter glucose (Mediatech Inc., Manassas, VA) containing 10% (unless noted, all percentages represent volume to volume) fetal bovine serum (FBS, Invitrogen, Carlsbad, CA) and 1% penicillin, streptomycin and amphotericin (Omega Scientific), with the addition of 5 μg/ml of bovine insulin (Sigma-Aldrich, St. Louis, MO). All cell lines were mycoplasma free when periodically tested using a Mycoalert assay (Lonza, Rockland, ME).
Oncogenic analysis of cell lines
Cell lines were analyzed for known oncogenic activating mutations and deletions using multiplex PCR as well as by MALDI-TOF mass spectrometry (Sequenom, San Diego, CA) [
30]. Point mutations were confirmed by PCR and direct sequencing as previously described [
9]. In addition, most cell lines were analyzed by SNP arrays with DNA extracted from the cell lines hybridized onto Illumina Beadchip Human Exon 510 S-Duo (Illumina Inc., San Diego, CA).
Cell proliferation and viability assays
Melanoma cell lines were treated with TAK-733 or parallel DMSO vehicle control at the given concentrations for 72 hours. Cell viability was measured using a tetrazolium compound [3-(4,5-dimethylthiazol-2-yl)-5- (3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2 H-tetrazolium (MTS)-based colorimetric cell proliferation assay (Promega, Madison, WI) as previously described [
9].
Cell cycle analysis
Cells were treated with different concentrations of TAK-733 (50 and 500 nM) or parallel vehicle control for 48 hours, fixed by Cytofix/Cytoperm solution and washed by Perm/Wash buffer according to fixation and pereabilization method recommended by BD bioscience, and then stained in sterile PBS containing 1.0% albumin bovine serum, 0.1% Nonidet P-40 (Sigma-Aldrich) and 3 μM DAPI (4',6-diamidino-2-phenylindole). Flow cytometry was analyzed using FlowJo (Tree Star Inc, Asland, OR).
Western blotting
Western blotting was performed as previously described [
31]. Primary antibodies included pAkt (Ser473), pAkt (Thr308), Akt, pS6K (Thr389), S6K, pS6 (Ser235/236), S6, pMEK (Ser217/221), MEK, pERK1/2 (Thr204/205), and ERK (all from Cell Signaling Technology, Danvers, MA), and α-actin (Sigma-Aldrich). Immunoreactivity was revealed using the ECL kit (Amersham Biosciences Co, Piscataway, NJ).
3 x 104 cells/well were plated on 0.001% poly-L-lysine (Sigma-Aldrich) pre-incubated filter bottom 96-well plates (multiscreen HTS GV 0.22 μm opaque, Millipore, Billerica, MA) and rested for 24 hours. 0.1 and 1 μM of TAK733 or parallel DMSO vehicle control were added in triplicate for 20 hours. Cells were incubated for 1 hour with 2.0 μCi with metabolic tracers chosen as analogues of PET tracers: 3H-DDG (American Radiolabeled Chemicals Inc., St. Louis, MO) in glucose-free RPMI 1640 (Invitrogen), or methyl-3H-thymidine (thymidine, Moravek Biochemicals Inc., Brea, CA) in RPMI 1640. Extracellular metabolic tracer was washed off using a multiscreen HTS vacuum manifold system (Millipore). 100 μL scintillation fluid (Perkin Elmer, Waltham, MA) was added to each well and tritium count was measured on a 1450 microbeta trilux microplate (Perkin Elmer).
Competing interests
Erika von Euw, Mohammad Atefi, Narsis Attar, Connie Chu, Sybil Zachariah, Barry L. Burgess, Stephen Mok, Charles Ng, Deborah J.L. Wong, Bartosz Chmielowski, Richard C. Koya, Tara A. McCannel: Have no competing interests. David I. Lichter, Elena Izmailova: Are employees of Millennium Pharmaceuticals, Inc., the manufacturer of TAK733. Antoni Ribas: Is a compensated ad hoc advisor to Millennium Pharmaceuticals, Inc., the manufacturer of TAK733.
Authors’ contributions
EE, MA, NA, CC, SZ, BLB, SM, CN, DJLW, DIL, RCK: Performed experiments. EE, MA, TAM, EI, AR: Designed the studies. BLB, BC, EI, AR: Provided key reagents. EE, MA, AR: Wrote the manuscript. EE, MA, NA, CC, SZ, BLB, SM, CN, DJLW, BC, DIL, RCK, TAM, EI, AR: All authors approved the final manuscript.