BRAF V600E Disrupts AZD6244-Induced Abrogation of Negative Feedback Pathways between Extracellular Signal-Regulated Kinase and Raf Proteins

Introduction

The Ras/Raf/mitogen-activated protein kinase (MAPK)/extracellular signal-regulated kinase (ERK) kinase (MEK)/ERK MAPK pathway plays an important role in regulating diverse cellular functions, including cell proliferation, cell cycle regulation, cell survival, and cell migration. Initiation of MAPK signaling occurs through activation of cell surface receptor tyrosine kinases that in turn induce the G protein Ras to exchange GDP for GTP. The serine/threonine kinase Raf proteins are recruited to the plasma membrane by GTP-bound Ras and are activated via a complex mechanism. Activated Raf phosphorylates and activates the dual specificity kinases MEK1 and MEK2, which successively phosphorylate and activate the proline-directed serine/threonine kinases ERK1 and ERK2, which have multiple downstream targets, including Elk-1, c-Ets1, c-Ets2, p90RSK1, MNK1, MNK2, and TOB ( 1– 3).

Hyperactivation of the MAPK signaling cascade resulting from receptor tyrosine kinase overexpression/mutation or MAPK family member mutation has a well-established role in oncogenesis and tumor growth ( 4, 5). Activating mutations in Ras and BRAF are common in many types of cancer and result in constitutive activation of MAPK signaling. K-ras mutations are found in up to 30% of all cancers and are particularly common in pancreatic cancers (90%) and colon cancers (50%; ref. 6). BRAF mutations have a more narrow distribution but are prevalent in a few specific malignancies, including melanoma (63%), papillary thyroid cancers (45%), and low-grade ovarian cancers (36%; refs. 7– 9). Of interest, BRAF V600E has also been found in up to 70% of colon cancers characterized by the presence of defective DNA mismatch repair, specifically those resulting from MLH1 promoter hypermethylation ( 10). Although many BRAF mutations have been described, a single amino acid change, resulting in a valine to glutamic acid substitution at position 600 (V600E), accounts for ∼80% of the mutations ( 11). The V600E mutation results in elevated constitutive kinase activity, which is thought to occur by disruption of the inactive conformation by the mutation ( 12).

Given its well-defined role in cancer growth, therapeutic targeting of the MAPK pathway has been an area of intense investigation, and MEK inhibitors are seen as a promising new approach for cancer treatment ( 13). The first MEK inhibitor tested clinically was PD184352 (CI-1040), but further development was discontinued due to concerns over bioavailability and pharmacodynamics ( 14). Clinical development of two additional MEK inhibitors, AZD6244 (ARRY-142886) and PD0325901, continues ( 15, 16). The biological characterization and chemical structure of this agent was recently published ( 16). AZD6244 is a potent and selective noncompetitive inhibitor of MEK1 and MEK2, with an in vitro IC₅₀ of 10 to 14 nmol/L against purified enzyme. When tested against a broad range of other serine/threonine and tyrosine kinases, little inhibition at concentrations up to 10 μmol/L is detected. This high degree of specificity is thought to occur as a result of binding to an allosteric inhibitory site adjacent to the ATP-binding site ( 17). AZD6244 has excellent preclinical activity ( 16, 18, 19). Safety and tolerability have been shown in initial reports from a phase I clinical trial in patients with advanced solid malignancies, and phase II efficacy studies are now under way. Response to MEK inhibition in vitro is variable, although the MAPK pathway is activated in a large proportion of tumors. Genotype, with respect to activating mutations in Ras and BRAF, seems to play an important role in conferring sensitivity. Tumor cell lines containing the V600E BRAF mutation are exquisitely sensitive to MEK inhibition, whereas cells containing Ras mutations display variable sensitivity, and cells with wild-type (WT) Ras and BRAF are relatively resistant ( 20). The molecular mechanisms conferring in vitro sensitivity to MEK inhibitors are not well understood. Initial studies with AZD6244 suggest that the pattern of sensitivity with respect to BRAF and KRAS mutation status is similar to other MEK inhibitors ( 16). To investigate the determinants of sensitivity to AZD6244 and enhance the clinical development of this exciting class of compounds, we undertook the present study to further characterize the in vitro effects of AZD6244 in a broad panel of cell lines.

Materials and Methods

Materials and cells. AZD6244 was provided by AstraZeneca. U0126 and PD98059 were purchased from Calbiochem. All non–small cell lung cancer (NSCLC) cell lines, except MV522, were obtained from the American Type Culture Collection (ATCC). MV522 cells were provided by Dr. Julian Molina (Mayo Clinic, Rochester, MN). HT29, Colo205, HCT116, MiaPaCa-2, MDA-MB-231, CAKI-1, and BxPC3 were obtained from ATCC. SK-MEL 3 and SK-MEL 28 cells were provided by Dr. Svetomir Markovic (Mayo Clinic). All cells were cultured in RPMI 1640 containing 10% fetal bovine serum, penicillin, and streptomycin and incubated in a humidified incubator with 5% CO₂ at 37°C. Genotypes for all cell lines, except MV522, have been previously determined ( 11, 21). The genotype of MV522 relative to K-ras and BRAF was determined by direct sequencing of genomic PCR products encompassing exons 2 and 3 of K-ras and exons 11 and 15 of BRAF using previously described primers and techniques ( 22).

Cell growth assays. For 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium salt (MTS) cell proliferation assays, cells were plated in 96-well culture dishes at a density of 500 to 1,000 per well. The following day, AZD6244 or DMSO was added to the cells at concentrations ranging between 1 nmol/L and 10 μmol/L depending on the experiment. Cell proliferation was determined 4 to 5 d later using CellTiter 96 AQueous One Solution (Promega). For clonogenic assays, cells were plated at 500 to 1,000 in 35-mm culture dishes. AZD6244 or DMSO was added the next day, and cells were grown for 7 to 10 d to allow colonies to form. The number of viable colonies was counted after staining with Coomassie blue.

Cell cycle analysis. Cells were plated at ∼50% confluence in 100-mm culture dishes. AZD6244 (1 μmol/L) or DMSO was added 24 h later and cells were collected after a 24-h incubation. Cells were rinsed in PBS, fixed in 95% ethanol, pelleted by centrifugation, and resuspended in RNase A at 1 mg/mL in 0.1% sodium citrate. Following incubation at 37°C for 15 min, propidium iodide was added to 100 μg/mL in 0.1% sodium citrate. DNA content was determined by flow cytometry and the proportion of cells in each cell cycle phase was determined using ModFit software (Verity Software House).

Immunoblot analysis. Following drug treatment, cell lysates were prepared and subjected to SDS-PAGE immunoblot analysis using standard techniques. The following primary antibodies were used: rabbit polyclonal antibodies to MEK1/2, p27, phospho-ERK (Thr²⁰²/Tyr²⁰⁴), and phospho-MEK1/2 (Ser^217/221) from Cell Signaling Technology; rabbit polyclonal antibodies to ERK, actin, and BRAF and murine monoclonal antibodies to Raf-1, tubulin, and cyclin D1 from Santa Cruz Biotechnology; murine monoclonal antibody to p21^WAF1 from NeoMarkers; and rabbit polyclonal antibody to sprouty 2 (Spry2) from Abcam. To ensure equivalent loading and transfer, blots were probed with anti-actin where indicated. For quantitation, blots were exposed using a Syngene bioimaging system.

Short interfering RNA knockdown. Short interfering RNA (siRNA) against BRAF, Raf-1, ERK1, and ERK2 was purchased from Ambion. Cells were transfected with siRNAs at a final concentration of 20 nmol/L using siPORT NeoFX Transfection Agent (Ambion) following the manufacturer's protocol. Cells were retransfected 24 h later and cell lysates were harvested 72 h after initial transfection.

Immunoprecipitation kinase assay. Kinase assays were performed using a nonradioactive assay kit for BRAF activity from Promega, and the manufacturer's protocol was followed with the following exceptions. The antibody for immunoprecipitation was a rabbit polyclonal anti-BRAF from Santa Cruz Biotechnology. For analysis, levels of phosphorylated glutathione S-transferase-MEK were determined in the fraction not bound to Sepharose beads.

Exogenous BRAF expression. Cells were plated at ∼80% confluence in a six-well culture dish. After 24 h, the cells were transfected with hemagglutinin-tagged plasmids expressing either WT or V600E BRAF (Biomyx Technology) using Lipofectamine 2000 (Invitrogen) following the manufacturer's protocol.

Retroviral-mediated short hairpin RNA expression. Retroviral plasmids containing control or BRAF-specific short hairpin RNAs (shRNA) were provided by Dr. David Tuveson (Cambridge Research Institute, Cambridge, United Kingdom). BRAF-specific constructs used in this study include COM-4, which targets both WT and V600E BRAF, and MU-A, which targets only V600E BRAF ( 23). Retroviruses were prepared by transient transfection of Phoenix helper virus–free amphotropic producer cells (ATCC), and cell lines were infected with retroviral supernatants using previously described techniques ( 24).

Allele-specific PCR. Total cellular RNA was prepared from cells using the RNeasy Mini kit (Qiagen) following the manufacturer's protocol. cDNA was prepared with 5 μg RNA and oligo(dT) primers using the SuperScript III First-Strand Synthesis System following the manufacturer's protocol (Invitrogen). Primers used in PCRs included 5′-GTGATTTTGGTCTAGCTACAGT-3′ and 5′-GTGATTTTGGTCTAGCTACAGA-3′, which amplify the WT allele or BRAF V600E mutation, respectively ( 25), and the reverse primer 5′-GTCCCTGTTGTTGATGTTTGAATA-3′, which produces a 212-bp PCR product. Amplification was carried out using FastStart Taq DNA polymerase (Roche) following the manufacturer's guidelines. Annealing temperature was 58°C for 29 or 31 cycles in WT or mutant samples, respectively. Control reactions using glyceraldehyde-3-phosphate dehydrogenase (GAPDH) primers were done as previously described for 23 cycles ( 26). PCR samples were separated using agarose gel electrophoresis and the products were visualized using ethidium bromide staining. For quantitation, gel images were captured and analyzed using a Syngene bioimaging system.

Results

In vitro activity of AZD6244. Sensitivity to the MEK inhibitor CI-1040 seems to depend on the presence of the BRAF activating mutation V600E ( 20); however, sensitivity among cell lines containing activating Ras mutations is variable and cells containing WT MAPK pathway proteins are resistant. Sensitivity to AZD6244 seems to follow the same general characteristics based on recently published data ( 16). To more fully characterize the in vitro activity of AZD6244, we used a panel of NSCLC cell lines as well as multiple additional cell lines that harbor a variety of Ras and BRAF mutations ( Table 1 ). For each cell line, the IC₅₀ was determined in a MTS assay using a 4- to 6-day continuous exposure to AZD6244. All cell lines containing BRAF V600E mutation were exquisitely sensitive to AZD6244 (IC₅₀, <100 nmol/L), whereas cells containing non-V600E BRAF mutations or activating N-ras or K-ras mutations with WT BRAF were variably sensitive to AZD6244. Cells containing WT BRAF and Ras proteins were resistant to AZD6244.

To examine the determinants of sensitivity to AZD6244, we chose to examine two NSCLC cell lines that harbor mutations resulting in constitutively active ERK, MV522 (BRAF V600E), and A549 (K-ras G12S) in detail. Both cell lines show dose-dependent growth inhibition with AZD6244 using either a colony-forming or MTS proliferation assay ( Table 1; Fig. 1A ). The IC₅₀s in A549 versus MV522 cells differed by ∼20-fold using the clonogenic assay (1.3 versus 0.06 μmol/L, respectively) or by ∼80-fold using the MTS assay (6.3 versus 0.08 μmol/L, respectively). Consistent with a previous report on MEK inhibitors in NSCLC cell lines ( 27), AZD6244 induced a G₀-G₁ cell cycle arrest in the sensitive cell line MV522 and to a lesser extent in A549 cells ( Fig. 1B). In neither of these cell lines was AZD6244 able to induce apoptosis, as the percentage of sub-G₀-G₁ cells in these experiments was <1% in all samples. In colony-forming assays, significantly fewer colonies were seen in the presence of AZD6244, but many isolated cells and small groups of cells were observed, suggesting that our results were due to growth inhibition rather than induction of apoptosis.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Effects of AZD6244 in NSCLC cell lines. A, MV522 and A549 cells were treated with AZD6244 at varying concentrations and response was determined using a clonogenic colony-forming assay. After 7 d, colonies were counted and the data are presented as the percentage of colonies formed relative to vehicle-treated cells (% Control). B, MV522 and A549 cells were treated with the indicated concentrations of AZD6244 for 24 h. The percentage of cells in each phase of the cell cycle was determined by quantitation of DNA content by propidium iodide staining and flow cytometry. C, MV522 and A549 cells were treated with the indicated concentrations of AZD6244 for 2 h and cell extracts were analyzed by immunoblotting for ERK and phospho-ERK (P-ERK). D, MV522 and A549 cells were treated for the indicated times with 500 nmol/L AZD6244. Cell extracts were analyzed by immunoblotting for the indicated proteins.

Sensitivity to AZD6244 correlates with MEK inhibition and the lack of phospho-MEK accumulation. To examine the differential sensitivity of MV522 and A549 cells to AZD6244-induced growth arrest, immunoblot analysis was used to determine if AZD6244 had similar target effects ( Fig. 1C). The IC₅₀ for MEK inhibition in MV522 was approximately 10 to 20 nmol/L and 20 to 50 nmol/L in A549 cells, a difference of ∼2-fold, which is unlikely to explain the 20- to 80-fold difference in IC₅₀ determined in growth assays and is probably within the inherent variability of this assay. Next, we examined target effects as well as downstream effectors in the MAPK pathway that have previously been implicated in the response of NSCLC cell lines to MEK inhibition ( Fig. 1D; ref. 27). AZD6244 effectively inhibited MEK, as detected by the absence of phospho-ERK, after only 2 h of treatment in both cell types. In MV522 cells, phospho-ERK inhibition was accompanied by increased p21^WAF1 and p27^Kip1 protein expression and decreased cyclin D1 levels, consistent with the observed G₀-G₁ arrest. Limited changes in p27^Kip1 and cyclin D1 levels were also seen in A549 cells. Interestingly, levels of phospho-MEK remained constant in MV522 cells throughout the time course but were significantly increased in A549 cells by 24 h. This result suggests that AZD6244 abrogates negative feedback between ERK and upstream targets in A549 cells, thus resulting in increased phospho-MEK. The negative feedback loop does not seem to be active in MV522 cells, as there is no corresponding change in phospho-MEK levels. To determine if phospho-MEK accumulation following treatment with AZD6244 is a class effect relative to MEK inhibitors, A549 cells were treated with AZD6244 or the alternative commercially available laboratory-grade MEK inhibitors U0126 or PD98059 ( Fig. 2A ). In each case, MEK inhibitor treatment substantially increased phospho-MEK levels.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

MEK inhibitors disrupt negative feedback pathways. A, A549 cells were treated with vehicle, 500 nmol/L AZD6244 (AZD), 25 μmol/LU0126, or 30 μmol/L PD98059 (PD) for the indicated times. Cell extracts were analyzed by immunoblotting for the indicated proteins. B, SW900 or HeLa cells were transfected with control (Ctrl), ERK1, ERK2, or a combination of ERK1 and ERK2 siRNAs. After 72 h, protein extracts were analyzed by immunoblotting for the indicated proteins. C, baseline levels of phospho-MEK and phospho-ERK in the absence of AZD6244 were determined for each cell line in Table 1. Data are presented as the expression relative to phospho-MEK (P-MEK) and phospho-ERK (P-ERK) levels in MV522 cells. Columns, mean expression in cell lines containing BRAF V600E and cell lines containing WT proteins or non-V600E mutations in BRAF or Ras; bars, SE. Student's t test revealed a significant difference between the expression of phospho-MEK in the two groups (*, P < 0.0001). D, phospho-MEK levels were determined for each cell line from immunoblots and expressed as the relative phospho-MEK level in AZD6244 versus vehicle-treated cells. Each data point represents an individual cell line and cell lines are separated as those containing BRAF V600E and those containing WT proteins or non-V600E mutations in BRAF or Ras. Student's t test revealed a significant difference between the means of each group (P < 0.05).

Accumulation of phospho-MEK after treatment with a MEK inhibitor may be due to abrogation of negative feedback between ERK and upstream targets; alternatively, the effect could result from pharmacologic effects on the MEK protein itself that inhibit phosphatase activity. To show the phospho-MEK accumulation results from abrogation of negative feedback, ERK function was directly inhibited using siRNAs to down-regulate ERK1 and ERK2 protein expression. Because knockdown of ERK was inefficient in A549 cells, two additional cell lines, SW900 and HeLa, were used ( Fig. 2B). Significant accumulation of phospho-MEK in these cell lines was seen after knockdown of both ERK proteins, whereas knockdown of either ERK1 or ERK2 alone resulted in intermediate phospho-MEK accumulation. These results confirm that phospho-MEK accumulation following treatment with a MEK inhibitor is due to abrogation of ERK-dependent feedback inhibition.

BRAF V600E mutation disrupts ERK-directed negative feedback. Our initial studies showed that accumulation of phospho-MEK in response to AZD6244 occurred in a cell line containing an activating K-ras mutation but did not occur in a cell line containing the BRAF V600E mutation. At least one study has suggested that the BRAF V600E mutation disrupts feedback by disrupting interactions between the MAPK inhibitory proteins Spry2 or Spry4 and BRAF ( 28). To determine if AZD6244-induced phospho-MEK accumulation was correlated to the precise mutation contained in each cell line, MEK, phospho-MEK, ERK, and phospho-ERK protein expression levels were determined using immunoblot analysis in the presence and absence of AZD6244 for all cell lines described in Table 1. If a negative feedback loop was inactive in cells containing BRAF V600E, then these cells would be expected to have higher baseline levels of phospho-MEK and there should be little change in phospho-MEK after treatment with AZD6244. Indeed, baseline levels of phospho-MEK are significantly higher in cell lines containing BRAF V600E relative to cell lines containing non-V600E mutations or WT proteins, whereas baseline levels of phospho-ERK are similar ( Fig. 2C). In response to AZD6244, phospho-MEK levels are essentially unchanged in V600E-containing cell lines but increase in all other cell lines tested as a result of abrogation of feedback inhibition ( Fig. 2D). To determine if the effect on feedback in cell lines containing non-V600E or WT proteins was only a characteristic of cancer cell lines, we also examined the effect of AZD6244 on phospho-MEK expression in mouse embryonic fibroblasts (MEF). MEFs were resistant to growth inhibition due to AZD6244 (IC₅₀, >10 μmol/L) in vitro and accumulated phospho-MEK similar to cancer cells containing WT BRAF (data not shown). Thus, BRAF V600E seems to disrupt a negative feedback loop in the MAPK pathway.

AZD6244 abrogates negative feedback between ERK and Raf. Raf proteins are a likely target of AZD6244-induced feedback inhibition because direct feedback inhibition between and ERK and Raf-1 or BRAF has been described ( 29, 30). We used two approaches to test this hypothesis. First, siRNAs were used to knock down Raf-1 and BRAF expression in HeLa cells treated with vehicle or AZD6244 ( Fig. 3A ). In cells treated with AZD6244, phospho-MEK accumulation was greatest in cells transfected with a control siRNA. Knockdown of Raf-1 or BRAF reduced phospho-MEK accumulation, and combined knockdown resulted in the greatest effect. These results show that ERK feedback inhibition targets both Raf-1 and BRAF. Second, BRAF kinase activity was determined using an immunoprecipitation kinase assay from MV522 and A549 cell extracts treated with vehicle or AZD6244 ( Fig. 3B). As expected, levels of phospho-MEK increased substantially in whole-cell extracts from A549 cells treated with AZD6244 but remained relatively unchanged in MV522 cells. BRAF kinase activity was much greater in MV522 cells compared with A549 cells, but the activity was unchanged by AZD6244. However, A549 BRAF activity increased substantially after AZD6244 treatment. We also attempted to assay for changes in Raf-1 activity with AZD6244 treatment in A549 cells but were unable to detect a change in Raf-1 kinase activity after AZD6244 treatment (data not shown). Nevertheless, our results show that AZD6244 induces increased BRAF kinase activity in a non-V600E cell line.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

AZD6244-induced phospho-MEK accumulation requires both BRAF and Raf-1. A, HeLa cells were transfected with control, Raf-1, BRAF, or a combination of Raf-1 and BRAF (R/B) siRNAs. After 48 h, vehicle (DMSO) or 500 nmol/L AZD6244 was added to the cells. After 24 h of incubation, protein extracts were analyzed by immunoblotting for the indicated proteins. B, MV522 or A549 cells were treated with vehicle or 500 nmol/L AZD6244 for 24 h. Whole-cell lysates were analyzed by immunoblotting for MEK, phospho-MEK, and actin. BRAF activity was determined in cell extracts using an immunoprecipitation (IP) kinase assay. Background activity was determined in a control (Ctrl) sample that did not contain cell lysate. Immunoprecipitation and BRAF activity were shown by immunoblotting for the indicated proteins. GST, glutathione S-transferase. C, A549 and MV522 cells were treated with the indicated concentrations for 24 h. Cell extracts were analyzed by immunoblotting for the indicated proteins. Tubulin is included as a loading control.

Negative feedback in the MAPK pathway is partially mediated by ERK-directed positive regulation of the inhibitory protein Spry2. To determine if Spry2 expression is reduced with AZD6244 treatment, MV522 and A549 cells were treated with AZD6244 and the change in Spry2 expression was determined by immunoblotting ( Fig. 3C). In both cell lines, Spry2 expression is decreased following AZD6244 treatment.

Effect of exogenous BRAF V600E expression on feedback and sensitivity to AZD6244. Cell lines containing the BRAF V600E mutation are exquisitely sensitive to AZD6244 and do not show feedback inhibition. To determine if exogenous expression of BRAF V600E disrupts normal feedback inhibition and confers sensitivity to AZD6244, cell lines containing WT proteins or an activating Ras mutation were transiently transfected with either WT BRAF or BRAF V600E. In both H1299 (N-ras Q61K, Fig. 4A ) and U2OS (WT, Fig. 4C) cells, transfection with WT BRAF did not substantially change baseline phospho-MEK or AZD6244-induced phospho-MEK accumulation, but transfection with BRAF V600E substantially increased baseline phospho-MEK and there was no change in the phospho-MEK level after AZD6244 treatment. These results are consistent with the findings we have observed in cell lines naturally containing BRAF V600E. In neither H1299 nor U2OS did exogenous BRAF V600E expression alter the sensitivity to AZD6244 in proliferation assays ( Fig. 4B and D). Thus, BRAF V600E disrupts negative feedback pathways but is not able to confer sensitivity to MEK inhibition in resistant cell lines.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Exogenous BRAF V600E expression disrupts negative feedback. H1299 (A and B) or U2OS (C and D) cells were transfected with control (Ctrl), WT BRAF, or V600E expression plasmids. A and C, after 24 h, 0.5 μmol/L AZD6244 or vehicle was added. After an additional 24-h incubation, cell extracts were analyzed by immunoblotting for the indicated proteins. HA, hemagglutinin. B and D, after 24 h, cells were plated on a 96-well dish and AZD6244 was added to the indicated concentrations. After 4 d, cell proliferation was determined using a MTS assay.

Mutation-specific suppression of V600E expression. Our data as well as the work of other groups show that cell lines containing BRAF V600E are exquisitely sensitive to MEK inhibitors ( 16, 20). Studies have also shown that suppression of BRAF V600E expression results in growth suppression both in vitro and in vivo ( 31, 32). However, it has not been shown whether specific suppression of the BRAF V600E allele in cell lines that are heterozygous for the BRAF mutation would render those cells resistant to MEK inhibition. Several shRNAs have been described that specifically suppress expression of the BRAF V600E protein but have no effect on WT BRAF expression ( 32). We used these constructs to determine if BRAF V600E is required for the cytostatic effects of AZD6244 in cell lines containing the mutant protein. MV522 and HT29 were infected with retroviruses expressing shRNAs that target total BRAF expression (COM-4) or just the mutant protein (MU-A), and the effects on growth and responsiveness to AZD6244 were determined. COM-4 expression reduced the expression of both the mutant and WT RNA efficiently in MV522 and HT29 cells, whereas MU-A only affected the mutant RNA ( Fig. 5A ). Effects on the RNA were confirmed at the protein level. Currently, a mutation-specific antibody is not available, but expression of COM-4 significantly reduced total BRAF protein levels and MU-A had an intermediate effect as expected because both of theses cell lines are heterozygous ( Fig. 5B). Downstream effects for COM-4 and MU-A were similar in both cell lines. BRAF V600E knockdown significantly reduces both phospho-MEK and phospho-ERK expression, and knockdown of the WT protein does not enhance the effect. Treating the cells with AZD6244 further reduced phospho-ERK but did not alter phospho-MEK expression.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

Endogenous BRAF V600E suppression does not reverse AZD6244 sensitivity. A, MV522 and HT29 cells were infected with retroviral supernatants containing control (CT), COM-4 (C), or MU-A (M) shRNAs. Seventy-two hours after infection, RNA was isolated and analyzed for expression of WT and V600E BRAF by allele-specific PCR. H661 RNA was used as a control as it does not contain BRAF V600E. GAPDH was used as a control to show equivalent input RNA. Samples containing only 10% of the input RNA for MV522 (MV) and HT29 (HT) are shown to show that amplification occurred in a quantitative manner. Numbers below each gel represent the relative expression in each sample compared with cells infected with control shRNA. B, after 72 h, cells infected with the indicated retroviral shRNAs were treated for 24 h with vehicle or 0.5 μmol/L AZD6244. Cell extracts were analyzed for the indicated proteins by immunoblotting. C and D, cells infected with the indicated shRNA retroviruses were plated on 96-well dishes. C, relative cell proliferation was determined using a MTS assay on subsequent days. Data are presented as the absorbance determined after a 60-min incubation in MTS reagent in three independent experiments. D, responsiveness to AZD6244 was determined by incubating the cells in various concentrations of AZD6244 for 4 d and determining relative proliferation using a MTS assay. Points, mean compared with vehicle-treated cells; bars, SE. For each cell line, the differences observed between control and COM-4 or MU-A were significant based on a paired t test (P < 0.05).

To determine the effect of BRAF V600E suppression on growth, cells were infected with COM-4, MU-A, or control retroviruses and the effects on cell growth were determined using a MTS assay. BRAF V600E suppression clearly inhibits growth in both MV522 and HT29 cells by approximately 50% to 75% at 4 days ( Fig. 5C). Although growth is inhibited, the cells remain sensitive to AZD6244 ( Fig. 5D). Thus, cell lines containing BRAF V600E clearly depend on activation of phospho-ERK via the activating mutation that supports a model in which BRAF V600E confers sensitivity to AZD6244 by marking cells that are “addicted” to MAPK signaling to support in vitro growth. In fact, the sensitivity to AZD6244 is enhanced by BRAF V600E suppression, which supports this model. The effect of suppressing the WT BRAF allele on sensitivity to AZD6244 is not clear because we did not test a construct that specifically suppresses on the WT allele.

Discussion

AZD6244 is a potent and selective MEK inhibitor. Prior studies have suggested that cell lines harboring activating Ras and BRAF mutations were more likely to be sensitive to AZD6244 or other MEK inhibitors ( 16, 20, 33, 34). In particular, cells containing the BRAF V600E mutation are exquisitely sensitive to MEK inhibitors, whereas the response in cell lines containing Ras mutations is variable. In this study, we have confirmed these observations in a panel of NSCLC cell lines as well as a panel of cell lines representing a broad array of malignancies. Although all activating Ras and BRAF mutations are expected to result in constitutive activation of the MAPK pathway, the mechanism responsible for the patterns of sensitivity observed for the various mutations is not understood. Ras proteins have pleiotropic effects on multiple signaling pathways, whereas the target effects of Raf proteins are more limited. Raf-1 seems to have specific targets in addition to MEK, such as MST2 and ASK1 ( 35, 36), but no MEK-independent targets of BRAF have been described. Therefore, the variable resistance to MEK inhibitors in cell lines containing Ras mutations may be due to the activation of proliferative MEK-independent pathways that are not activated in cell lines containing BRAF mutations. However, cell lines containing BRAF mutations distinct from V600E, such as H1755 (G468A) and H1666 (G465V), do not have the same degree of sensitivity as V600E-containing cell lines, suggesting that other mechanisms may also be involved.

Specific suppression of BRAF V600E expression by RNA interference abrogates growth of cell lines containing the mutation in vitro and in vivo ( 23, 31, 37). Our results in NSCLC (MV522) and colon (HT29) are similar. Interestingly, in cell lines heterozygous for the V600E allele, specific knockdown of the mutant protein almost completely blocks phospho-MEK and phospho-ERK expression despite the presence of WT BRAF protein ( Fig. 5). Other groups have shown similar findings ( 32, 37). Activation of WT BRAF typically occurs through upstream activation of Ras by cell surface receptors ( 6, 38), pathways that are induced by the presence of high concentrations of FCS. Yet, despite growth in medium containing high serum concentrations, we observe near-complete abrogation of MAPK signaling by suppression of the BRAF V600E allele, which suggests that the mutant allele or other pathways may suppress upstream signaling pathways in these cell lines. V600E-containing cell lines seem to rely solely on activation of MAPK by the activating mutation in BRAF. Whether other BRAF mutations have the same effect is not known. BRAF mutations have been separated into two classes: those that produce high intrinsic kinase activity toward MEK, such as V600E and G468A, and those that have intermediate/low intrinsic kinase activity that activate MAPK through heterodimerization with Raf-1 ( 12, 39). The NSCLC cell line H1755 contains the high-activity BRAF mutation G468A and is resistant to AZD6244 (IC₅₀, 8.1 μmol/L) relative to cell lines containing the V600E mutation. Therefore, solely the presence of a high-activity mutant does not confer sensitivity to AZD6244, suggesting that V600E may have additional effects. Thus, our results show that cell lines containing V600E are dependent on the mutant allele for activation of MAPK and are dependent on MAPK for in vitro growth, which explains their exquisite sensitivity to AZD6244.

Feedback regulation is a critical component of intracellular signal transduction networks and may be involved in determining the specificity of cellular responses to upstream receptor activation. Within the MAPK signaling network, both positive and negative regulatory feedback loops have been described that affect differential cellular responses ( 40). Here, we have observed the presence of a negative feedback loop between ERK and upstream components of the MAPK pathway, including Raf proteins, in a broad range of cancer cell lines. Additionally, the presence of a feedback loop was confirmed using specific down-regulation of ERK1 and ERK2 expression by RNA interference, which also resulted in phospho-MEK accumulation. Phospho-MEK accumulation after treatment with a MEK inhibitor has been described by several groups ( 41, 42) and was presumed to be due to disruption of negative feedback pathways, but the phenomenon has not been systematically examined. Both direct and indirect feedback pathways between ERK and Raf have been described. ERK activation results in the hyperphosphorylation of multiple proline-directed target sites of Raf-1 that reduce Raf-1 kinase activity and disrupt Ras/Raf interactions ( 30). Similar results have also been observed for BRAF, whereby ERK-directed phosphorylation of a SPKTP motif in the COOH terminus of BRAF reduces activity ( 29).

Indirect pathways may also mediate the negative feedback between ERK and Raf. Spry1 and Spry2 are growth factor regulators, whose expression is positively regulated by ERK, which seem to be involved in the negative feedback regulation of both Ras and Raf proteins ( 28, 43– 45). Interestingly, sprouty proteins are unable to bind to the V600E isoform of BRAF ( 44), and this may underlie the differences we have observed in AZD6244-induced phospho-MEK changes between cell lines having BRAF V600E and WT/non-V600E mutations. In cell lines heterozygous for BRAF V600E, we expected to see intact negative feedback after V600E-specific suppression, but in both HT29 and MV522, phospho-MEK levels were significantly decreased after BRAF V600E suppression and there was no change with AZD6244 treatment. Although the reasons for this finding are not clear, it likely occurs because there is no upstream activation of MAPK other than that provided by BRAF V600E. Although we observed a small amount of residual phospho-ERK in cells after BRAF V600E suppression, this likely occurs due to incomplete knockdown. Aside from the exon 11 BRAF mutations G468A and G465V, the effect of other BRAF or Raf-1 mutations on AZD6244-induced phospho-MEK accumulation is not clear. Non-V600E BRAF exon 15 mutations may have a similar effect because prior studies have shown that they can also disrupt binding to Spry2 and Spry4 ( 44).

In summary, we show that the MEK inhibitor AZD6244 displays preclinical activity in a broad range of human cancer cell lines, particularly in lines containing the BRAF V600E activating mutation. AZD6244 disrupts negative feedback regulation between ERK and upstream targets resulting in accumulation of phospho-MEK. The BRAF V600E mutation disrupts normal feedback inhibition and results in elevated baseline phospho-MEK expression and interrupts the AZD6244-induced phospho-MEK accumulation. Although exogenous expression of BRAF V600E disrupts feedback signaling, it does not sensitize cells to AZD6244-induced growth arrest. In addition, specific down-regulation of endogenous BRAF V600E does not reverse the impaired feedback pathway nor does it confer resistance to AZD6244. In fact, suppression of endogenous BRAF sensitizes the cells to AZD6244. Thus, BRAF V600E is not sufficient to confer MEK inhibitor sensitivity but does mark cells that are dependent on MAPK for in vitro growth. Finally, feedback regulation is an important component of multiple cellular signaling cascades and our results highlight the possibility that sequential targeting of multiple proteins within a cascade may be an important clinical goal to ensure sufficient target inhibition and to block upstream activation of additional pathways that may be up-regulated by abrogation of negative feedback.

Disclosure of Potential Conflicts of Interest

A. A. Adjei: Honoraria, Array Biopharma. The other authors disclosed no potential conflicts of interest.