Differential sensitivity of melanoma cell lines with BRAFV600E mutation to the specific Raf inhibitor PLX4032

Blocking oncogenic signaling induced by the BRAFV600E mutation is a promising approach for melanoma treatment. We tested the anti-tumor effects of a specific inhibitor of Raf protein kinases, PLX4032/RG7204, in melanoma cell lines. PLX4032 decreased signaling through the MAPK pathway only in cell lines with the BRAFV600E mutation. Seven out of 10 BRAFV600E mutant cell lines displayed sensitivity based on cell viability assays and three were resistant at concentrations up to 10 μM. Among the sensitive cell lines, four were highly sensitive with IC50 values below 1 μM, and three were moderately sensitive with IC50 values between 1 and 10 μM. There was evidence of MAPK pathway inhibition and cell cycle arrest in both sensitive and resistant cell lines. Genomic analysis by sequencing, genotyping of close to 400 oncogeninc mutations by mass spectrometry, and SNP arrays demonstrated no major differences in BRAF locus amplification or in other oncogenic events between sensitive and resistant cell lines. However, metabolic tracer uptake studies demonstrated that sensitive cell lines had a more profound inhibition of FDG uptake upon exposure to PLX4032 than resistant cell lines. In conclusion, BRAFV600E mutant melanoma cell lines displayed a range of sensitivities to PLX4032 and metabolic imaging using PET probes can be used to assess sensitivity.


Background
Improved knowledge of the oncogenic events in melanoma indicates that a majority of mutations activate the mitogen-activated protein kinase (MAPK) pathway [1,2]. The most frequent mutation in the MAPK pathway is in the BRAF gene, present in 60-70% of malignant melanomas [3]. NRAS mutations occur in approximately 15% of melanomas [1,4,5] and are mutually exclusive with BRAF mutations [6,7]. The majority of mutations in BRAF are accounted for by a single nucleotide transversion from thymidine to adenosine leading to a substitution of valine by glutamic acid at position 600 (termed BRAF V600E ) [3,4,8], which leads to a 500-fold increase in activity compared to the wild type protein kinase [8].
PLX4032 (also known as RG7204) was developed as a specific inhibitor of Raf. It is an analogue of the pre-clinically tested PLX4720 [9]. PLX4720 inhibits the mutated B-Raf kinase at 13 nM, while the wild type kinase requires tenfold higher concentration (160 nM) [9], thus predicting high specificity for BRAF V600E mutant cell lines. The basis of this specificity for the mutated kinase is thought to be the preferential inhibition of the active conformation of B-Raf. In addition, its access to a Raf-selective pocket accounts for the selectivity against most other non-Raf kinases, which require concentrations 100 to 1000 times higher for kinase inhibition. The only exception is the breast tumor kinase (BRK), which is inhibited at 130 nM, a one-log difference compared to the V600E mutated B-Raf kinase [9].
In the current studies we analyzed a panel of human melanoma cell lines with defined oncogenic alterations for sensitivity to PLX4032. In addition, with a view to development of a biomarker to indicate response to tar-geted therapy, we investigated a non-invasive method of imaging resistance versus sensitivity in vivo. We describe that PLX4032 works differentially in melanoma cell lines with BRAF V600E mutations and that the positron emission tomography (PET) tracer 2-fluoro-2-deoxy-D-glucose (FDG) can be used in non-invasive PET imaging to distinguish between sensitive and resistant cell lines.

Reagents and cell lines
PLX4032 (also known as RG7204 or RO5185426) was obtained under a materials transfer agreement (MTA) with Plexxikon (Berkeley, CA) and dissolved in DMSO (Fisher Scientific, Morristown, NJ) to a stock concentration of 10 mM. SKMEL28 was obtained from American Type Culture Collection (ATCC, Rockville, MD), and the remaining human melanoma cell lines (M series) were established from patient's biopsies under UCLA IRB approval #02-08-067. Cells were cultured in RPMI 1640 with L-glutamine (Mediatech Inc., Manassas, VA) containing 10% (unless noted, all percentages represent volume to volume) fetal bovine serum (FBS, Omega Scientific, Tarzana, CA) and 1% penicillin, streptomycin, and amphotericin (Omega Scientific). All cell lines were mycoplasma free when periodically tested using a Mycoalert assay (Lonza, Rockland, ME).

Oncomap 3 core mass-spectrometric genotyping
Samples were run through OncoMap 3 which interrogates 396 somatic mutations across 33 genes. Whole genome amplified DNA at 5 ng/μl was used as input for multiplex PCR as described previously [11]. Single-basepair primer extension (iPLEX) was performed in a 2 μl reaction volume using iPLEX Gold single base extension enzyme (Sequenom, San Diego, CA). Products were resined and transferred to SpectroCHIPs for analysis by MALDI-TOF mass spectrometry [11]. All mutations were confirmed by direct sequencing of the relevant gene fragment.

Cell proliferation and viability assays
Melanoma cell lines were treated in triplicates with PLX4032 and parallel vehicle control in the given concentrations for 120 hours. Viable cells was measured using a tetrazolium compound [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS)-based colorimetric cell proliferation assay (Promega, Madison, WI). Cell line doubling time were determined from cell numbers measured in duplicates every 24 hours for a period of 9 to 12 days using a Vi-CELL series cell viability analyzer (Beckman Coulter). The doubling time in 24 hours was calculated by the formula 1/{[((logC2)-(logC1))×3.32]/T}, where C1 = the initial cell number, C2 = the final cell number, and T = 24 hours. The average of day 3, 4, 5 was used as the optimal doubling time for the given experimental condition.

Cell cycle analysis
Cells were treated with 1 μM PLX4032 and parallel vehicle control for 20 to 120 hours, fixed in 70% ethanol (Pharmco-Aaper, Shelbyville, KY), and then resuspended in sterile PBS containing 0.5% albumin bovine serum, 180 μL/ml propidium iodide staining solution (BD Biosciences) and 50 μg/mL ribonuclease A from bovine pancreas (Sigma-Aldrich). Flow cytometry was performed on FACSCalibur or FACScan and data was analyzed using FlowJo.

Apoptosis analysis
Melanoma cell lines were treated with increasing concentrations of PLX4032, DMSO vehicle control, or 1 μM of staurosporine as a positive control, for 120 hours. Cells were trypsinized and transferred to FACS tubes and stained with Annexin V-FITC and propidium iodide fol-lowing the manufacturer's instructions (BD Biosciences) and analyzed by flow cytometry using FACSCalibur as described [13].

In vivo microCT and microPET studies
Mice with established subcutaneous human melanoma xenografts were treated for 3 days with 100 mg/kg PLX4032 in corn oil or vehicle control twice daily by oral gavage. The last treatment was given one hour prior to intraperitoneal injection of 200 μCi [ 18 F]-FDG, which was allowed to distribute in the tissues for 1 hour before microPET scanning as previously described [15].

Statistical analysis
Continuous variables were compared using a paired Student's t-test with two-tailed P values.

PLX4032 specifically blocks the MAPK pathway in melanoma cell lines with the BRAF V600E mutation
We tested the ability of PLX4032 to differentially block MAPK pathway signaling in a panel of human melanoma cell lines (Table 1) by quantitating the inhibition of phosphorylated Erk (pErk), a downstream target of B-Raf activity, using intracellular phosphospecific flow cytometry ( Figure 1A). As expected, cell lines with BRAF V600E mutation had a fast (detectable at 1 hour) and sustained (persistent at 20 hours, Figure 1B) inhibition of pErk, although one of the cell lines (M263) had lower inhibition of pErk than the rest. There was no pErk inhibition in two cell lines with NRAS Q61L mutation (M202 and M207) and a cell line wild type for both oncogenes (M257). In fact, there was a markedly increased pErk signal in one NRAS Q61L mutated cell line (M207), an observation consistent with data from others that has been attributed to loss of negative regulatory pathways [16,17] and enhanced signaling through C-Raf [18,19]. Therefore, PLX4032 inhibits MAPK pathway signaling specifically in cell lines that harbor the BRAF V600E mutation.

Differential sensitivity to PLX4032 in BRAF V600E mutated melanoma cell lines
Melanoma cell lines with different NRAS/BRAF mutational status were treated in vitro with a range of concentrations of PLX4032 for 5 days. The three cell lines without BRAF V600E mutation were resistant to PLX4032. Seven BRAF V600E mutant cell lines were sensitive to PLX4032, including four highly sensitive cell lines with half maximal inhibitory concentration (IC 50 ) values below 1 μM. Surprisingly, in three cell lines with BRAF V600E mutation we could not determine an IC 50 with increasing concentrations of PLX4032 up to 10 μM, suggesting that these cell lines are resistant to this agent in a 5-day exposure in vitro ( Figure 1C). Similar results have been obtained in 3-day viability assays and when PLX4032 is added daily to the cultures or just at the beginning of the experiment (data not shown).

PLX4032 has similar inhibitory effects on cell cycle in sensitive and resistant BRAF V600E mutant cell lines
To study effects of PLX4032 on cell cycle progression downstream of B-Raf signaling we used propidium iodide flow cytometric staining. As expected, PLX4032 had no effect on cell cycle progression in melanoma cell lines without a BRAF V600E mutation (Figure 2A). In contrast, PLX4032 exposure for one (data not shown) or 20 hours ( Figure 2B and 2C) led to a similar and profound G1 arrest in all BRAF V600E mutant cell lines regardless of their in vitro sensitivity to PLX4032.

PLX4032 leads to apoptotic death in sensitive BRAF V600E but not in resistant BRAF V600E mutated melanoma cell lines
We then analyzed the ability of PLX4032 to differentially induce apoptotic effects against melanoma cell lines with the BRAF V600E mutation. Using a BRAF V600E mutant melanoma cell line with a good response to PLX4032 (M249) and another one that was poorly responsive to PLX4032 (M233) based on cell viability assays, we analyzed apoptotic induction using flow cytometry based on the incorporation of propidium iodide and Annexin V. After PLX4032-treatment, the increase in Annexin V positive cells, with or without being double positive for propidium iodide, was greater in the PLX4032-responsive M249 cells compared to the poorly responding M233 cells (Figure 2D and 2E). Similar results were obtained with M238 and M263 (data not shown). Taken together with the data on cell cycle inhibition, these data suggest that PLX4032 has cytostatic effects in BRAF V600E mutant cell lines with a poor response, while it has cytostatic and cytotoxic effects in cell lines with a good response to PLX4032 in cell viability assays.

Functional and genomic characterization of BRAF V600E mutated cell lines with different sensitivity to PLX4032
We tested if the differences in sensitivity to PLX4032 were due to markedly different doubling times. Resistant BRAF V600E mutated cell lines tended to have a slower doubling time compared to the sensitive BRAF V600E mutated cell lines (P = 0.24, Table 1). The lack of significance was due to outliers in a small group, most notably the highly sensitive cell line M262 having a doubling time close to 50 hours. Interestingly, all cell lines homozygous for the BRAF V600E mutation were moderately to highly sensitive to PLX4032, and cell lines resistant to PLX4032 were all heterozygous for BRAF V600E (P = 0.16). However, there were also two highly sensitive heterozygous cell lines with IC 50 values below 1 μM of PLX4032, and the sensitivity of homozygous cell lines spreads through one-log differences in PLX4032 concentrations (Table 1). We then used high throughput analysis of over 500 gene mutations using mass-spectrometry based genotyping [11] and high-density SNP arrays to explore other genomic alterations. Two different platforms (Illumina and Affymetrix) gave highly concordant results (data not shown) demonstrating that out of the 10 cell lines with BRAF V600E mutation, four have amplification of the BRAF locus (Table 1). There was no clear relationship between these amplification events and the BRAF V600E zygosity or the sensitivity to PLX4032. There were very few secondary mutations in this group of cell lines, with one cell line having a mutation in EGFR, and one cell line with a mutation in AKT (Table 1). In addition, the M257 cell line, which is wild type for both NRAS and BRAF and is highly resistant to PLX4032, was found to have 3 copies of wild type BRAF and a point mutation in CDKN2A. The distribution of amplification events in MITF and EGFR were also spread among the cell lines. Of note, there was no clear trend regarding the activation of the PI3K/Akt pathway based on activating mutations, or amplifications of AKT1/2 segregating the resistant and sensitive cell lines. Supervised hierarchical clustering comparing SNP array data from PLX4032-resistant and -sensitive BRAF V600E mutant cell lines did not point to specific genomic areas with concor-dant alterations differentiating the two groups of cell lines.

Modulation of MAPK and PI3k/Akt signaling pathways in sensitive and resistant cell lines
To further explore how cell lines with BRAF V600E mutation respond differently to PLX4032 we chose two extreme examples of cell lines with similar growth kinetics to perform an extended analysis of signaling pathways (Figure 3). M229 is one of the two most sensitive cell lines, while M233 proved to be very resistant despite having a short in vitro doubling time (  [20]. There was a low level of induction of pAMPK. These studies demonstrate that PLX4032 has complex effects on MAPK and PI3k/Akt signaling pathways that may be dependent on secondary oncogenic events beyond B-Raf.

Non-invasive imaging of PLX4032 anti-tumor activity
We analyzed the uptake profile of three different metabolic tracers that can be used in PET scans: two nucleoside analogs (thymidine and FAC [21]) and FDG, a glucose analog widely used as a PET tracer. As expected, BRAF wild type cell lines had no significant change in uptake of thymidine or FAC upon PLX4032-exposure. Conversely, all BRAF V600E mutated cell lines, irrespective of their sensitivity to PLX4032, had markedly decreased uptake of these two nucleoside analogues (Figure 4a and 4b). The greatest difference between PLX4032-sensitive and -resistant BRAF V600E mutants was in FDG uptake. The percentage decrease in FDG uptake was roughly double in PLX4032-sensitive BRAF V600E mutants compared to PLX4032-resistant cell lines (P = 0.009, Figure  4c). Finally, we tested if [ 18 F]-FDG uptake could be used as a pharmacodynamic marker of B-Raf V600E inhibition by PLX4032 in vivo. Mice with established subcutaneous M249 melanoma xenografts, a cell line highly sensitive to PLX4032 in vitro, were treated for 3 days with PLX4032 twice daily by oral gavage, and then analyzed by com-

Discussion
The BRAF V600E mutation is one of the most common kinase domain mutations in human cancer with a particularly high incidence in malignant melanoma [3,7]. The Raf-inhibitors PLX4720 and PLX4032 have the preclinical characteristics of functioning as specific inhibitors of the BRAF V600E mutated kinase with a favorable profile compared to wild type kinases [9,22]. Understanding the patterns of sensitivity and resistance in melanomas with different oncogenic events is of high importance for clinical translation. Our studies confirmed the high specificity of PLX4032 for a subset of BRAF V600E mutant cell lines [22]. Surprisingly, we noted differences in the sensitivity to PLX4032, with some BRAF V600E mutants demonstrating resistance to the cytotoxic effects of PLX4032. In most cases, these cells had a tendency towards slower growth kinetics and being heterozygous for BRAF V600E . This differential response to PLX4032 in BRAF V600E mutant melanoma cell lines may be explained by several mechanisms. It may be that there is preferential MAPK pathway-addiction in sensitive cell lines, and cells with lower sensitivity are less dependent on the BRAF V600E oncogenic signaling, relying on the co-activation of other signaling pathways including the PI3K/Akt pathway. We explored this possibility with SNP arrays and high throughput oncogene sequencing with a particular interest in looking at this pathway. The genomic analysis revealed that alterations in PI3K/Akt, including deletions of PTEN, amplifications of AKT and activating mutations in AKT were distributed throughout the cell line list with no clear pattern of correlation with sensitivity or resistance to PLX4032. However, in two cell lines phosphospecific Western blot staining suggested that the resistant cell line had increased Akt signaling upon PLX4032 exposure. Another possibility is that PLX4032-resistant BRAF V600E mutants have alternative signaling at the level of Raf, as has been described for cell lines with acquired resistance to a different Raf-inhibitor, AZ628, which show increased signaling through C-Raf [23]. The increase in pErk in an NRAS Q61L mutant cell line could be explained by abrogation of negative feedback loops mediated mainly by dual specificity phosphatases (MKPs/ DUSPs), as reported with Mek inhibitors [17,24], and the recent description of increased C-Raf signaling when heterodimerizing with inhibited B-Raf in BRAF wild type cells [18,19]. Therefore, the modulation of feed-back loops and alteration of Raf dimerization upon treatment with Raf inhibitors may also have a role in the differential sensitivity to PLX4032 in BRAF V600E mutant cell lines. Finally, differences in expression of pro-and anti-apoptotic molecules like Bim and Bad [25] may allow some BRAF V600E mutant cell lines to undergo growth arrest but not die from apoptosis upon exposure to PLX4032. Studies are ongoing to further explore these possibilities. We explored the use of PET imaging as a mean to noninvasively detect PLX4032-sensitivity. In vitro we found that any of the three PET tracers FDG, FLT and FAC could be used to distinguish between melanomas with a NRAS or a BRAF V600E mutation based on the differential effects of PLX4032 on cell cycle and metabolism. FDG could furthermore be used to distinguish between BRAF V600E mutant melanomas with high or low sensitivity to PLX4032. The PI3K/Akt pathway has been widely regarded as having a role in the regulation of glucose metabolism through mTOR, but recently the LKB1-AMPK pathway has been found to be regulated by oncogenic BRAF V600E signaling [20], which together may explain the marked and rapid effects of PLX4032 on inhibiting FDG uptake. We explored this possibility in two cell lines. Our data suggests a minor increase in pAMPK upon PLX4032 exposure, which may be in line with the proposed hypothesis [20].

Conclusions
These studies in melanoma cell lines may allow to better interpret the results of a recently reported phase I clinical trial with PLX4032 [26], with an objective response in excess of 70% of patients with BRAF V600E positive metastatic melanoma. The characterization of PLX4032-sensitive and -resistant BRAF V600E mutant melanoma cell lines may provide information about the molecular mechanisms that dictate sensitivity and resistance to PLX4032. In addition, molecular imaging with [ 18 F]FDG PET scans may help in providing an early readout of complete or incomplete pharmacodynamic effects of PLX4032 and therefore predict lesions that may or may not respond to therapy.