Research ArticleCancer

Kinase-Impaired BRAF Mutations in Lung Cancer Confer Sensitivity to Dasatinib

See allHide authors and affiliations

Science Translational Medicine  30 May 2012:
Vol. 4, Issue 136, pp. 136ra70
DOI: 10.1126/scitranslmed.3003513


During a clinical trial of the tyrosine kinase inhibitor dasatinib for advanced non–small cell lung cancer (NSCLC), one patient responded dramatically and remains cancer-free 4 years later. A comprehensive analysis of his tumor revealed a previously undescribed, kinase-inactivating BRAF mutation (Y472CBRAF); no inactivating BRAF mutations were found in the nonresponding tumors taken from other patients. Cells transfected with Y472CBRAF exhibited CRAF, MEK (mitogen-activated or extracellular signal–regulated protein kinase kinase), and ERK (extracellular signal–regulated kinase) activation—characteristics identical to signaling changes that occur with previously known kinase-inactivating BRAF mutants. Dasatinib selectively induced senescence in NSCLC cells with inactivating BRAF mutations. Transfection of other NSCLC cells with these BRAF mutations also increased these cells’ dasatinib sensitivity, whereas transfection with an activating BRAF mutation led to their increased dasatinib resistance. The sensitivity induced by Y472CBRAF was reversed by the introduction of a BRAF mutation that impairs RAF dimerization. Dasatinib inhibited CRAF modestly, but concurrently induced RAF dimerization, resulting in ERK activation in NSCLC cells with kinase-inactivating BRAF mutations. The sensitivity of NSCLC with kinase-impaired BRAF to dasatinib suggested synthetic lethality of BRAF and an unknown dasatinib target. Inhibiting BRAF in NSCLC cells expressing wild-type BRAF likewise enhanced these cells’ dasatinib sensitivity. Thus, the patient’s BRAF mutation was likely responsible for his tumor’s marked response to dasatinib, suggesting that tumors bearing kinase-impaired BRAF mutations may be exquisitely sensitive to dasatinib. Moreover, the potential synthetic lethality of combination therapy including dasatinib and BRAF inhibitors may lead to additional therapeutic options against cancers with wild-type BRAF.


Lung cancer is the leading cause of cancer-related deaths worldwide. Researchers recently identified two molecular subpopulations of non–small cell lung cancer (NSCLC)—populations with epidermal growth factor receptor (EGFR) mutations and those with EML4-ALK translocations—for which treatment with targeted agents has produced profound clinical responses. This development has elicited a paradigm shift in treating lung cancer: The tumors’ genetic characteristics dictate therapy. Nevertheless, definition of clinically relevant genetic aberrations is lacking in about 85% of NSCLC cases.

One approach to defining these determinants and thus to discovering effective therapeutic targets is by characterizing tumors that respond significantly to targeted agents. This approach led to the discovery of an increased response to EGFR inhibitors in patients with NSCLC harboring EGFR-activating mutations in an initial characterization of nine patients (1). In a phase 2 study of the multitargeted kinase inhibitor dasatinib in 34 patients with systemic therapy–naïve stage IV NSCLC, we observed in one patient at 12 weeks of treatment an initial partial response that continued to improve long after dasatinib was discontinued for toxicity (malaise), suggesting that dasatinib induced cancer cell senescence (2). This patient [patient X (PX)] received no further therapy and at present is alive and free of active cancer 4 years after his initial diagnosis.

We hypothesized that an analysis of PX’s tumor would identify the determinants of his response to dasatinib. We thus undertook a comprehensive analysis of his tumor that included mutational analysis of 40 genes, array comparative genomic hybridization (aCGH), and immunohistochemistry and then used site-directed mutagenesis, kinase assays, cell cycle analysis, terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick end labeling (TUNEL), and β-galactosidase staining in vitro in NSCLC cell lines to characterize a mutation identified in the tumor. We discovered a previously undescribed BRAF mutation, Y472CBRAF, that results in substantial impairment of BRAF kinase activity. BRAF is a key factor in the RAS pathway: Once activated by surface receptors, RAS activates RAFs, including BRAF. RAF phosphorylates MEK (mitogen-activated or extracellular signal–regulated protein kinase kinase), which then activates extracellular signal–regulated kinase (ERK), which can promote cancer progression or induce cell senescence (3). BRAF mutations include those that cause kinase activation or impair kinase activity. Paradoxically, most BRAF mutants with reduced kinase activity still activate MEK and ERK via transactivation of CRAF (4, 5). Here, we tested whether the marked and durable clinical response of our patient was a result of dasatinib-induced senescence of Y472CBRAF carrying cancer cells.


Patient’s tumor did not carry common mutations

In our phase 2 study of dasatinib in 34 patients with systemic therapy–naïve stage IV NSCLC, the sole responder was a male former smoker (PX) who had a profound, durable response (2). Over the 12 weeks of dasatinib-based therapy, PX had a partial response as assessed by both tumor size and metabolic activity, and his metastatic tumor (in paraspinal muscle) continued to shrink after therapy was stopped. At the end of therapy, the diameter of the metastasis was 2.8 cm, with a standardized uptake value (SUV) of 17. At 17 weeks, accurately measuring the metastasis on a computed tomography (CT) scan was difficult, but the SUV was 11. At 21 weeks, the SUV was 4.5. At 32 weeks, the mass was undetectable on CT and PET (positron emission tomography) scans. PX remains free of active cancer 4 years after the initial diagnosis and has not received any other cancer therapy. PX still has a 2-cm lung nodule that has no detectable metabolic activity on PET and that has been stable on CT scans for 4 years (fig. S1A). The median progression-free survival for the entire study was 1.4 months, and the median overall survival was 15.6 months (fig. S1B). We performed additional studies of PX’s tumor tissue to identify the underlying mechanism of dasatinib sensitivity.

PX’s tumor did not harbor any EGFR or KRAS mutations by intron-based polymerase chain reaction (PCR) of K-Ras exons 1 and 2 (codons 12, 13, and 61) and EGFR exons 18 to 21 as published (2). We did not detect any ALK gene rearrangements by fluorescence in situ hybridization, c-Src mutations by intron-based PCR of exons 7 to 10, nor any discoidin domain receptor 2 (DDR2) mutations with conventional Sanger sequencing of all exons as described (6). Immunohistochemistry revealed that the tumor expressed total and phosphorylated EphA2, c-Kit, and PDGFRα (platelet-derived growth factor receptor α) (table S1). PX did not harbor any germ-line BRAF or KRAS mutations by intron-based PCR of BRAF exons 11 and 15 and K-Ras exons 1 and 2 (codons 12, 13, and 61) of DNA isolated from his peripheral blood lymphocytes.

To identify novel mutations or changes in gene copy number in PX’s tumor, we used the MassARRAY system (Sequenom) and performed aCGH. We identified no mutations among the 40 genes tested (table S2). Using aCGH, we identified several regions of increased and decreased copy numbers (fig. S2 and table S3). We also observed increased copy numbers of the known direct dasatinib targets HCK, DDR1, EPHA3, and ARG (ABL2). We found no copy number changes for LYN, FGR, FYN, SRC, DDR2, EPHB1, EPHB2, EPHB3, EPHA1, EPHA2, EPHA4, TNK2, PTK6, GAK, KIT, PDGFR, KRAS, EGFR, or BRAF.

Y472CBRAF is a kinase inactivating mutation

Because the Sequenom MassARRAY technology can only identify candidate mutations for which assays are specifically designed and given the known role of BRAF in oncogene-induced senescence, we sequenced exons 11 and 15 of BRAF. These two exons have many known mutations not included in our panel. We identified the mutation Y472CBRAF, which has not been described previously and occurs in a highly conserved region of exon 11 (figs. S3 and S4). We also analyzed BRAF in 19 patients from our original clinical trial for whom DNA adequate for analysis was available and found no other inactivating mutations (table S4).

To determine the functional significance of Y472CBRAF, we used site-directed mutagenesis to create Y472CBRAF, G466VBRAF (kinase-impaired), and V600EBRAF (constitutively active) in a Flag-tagged wtBRAF construct. We transfected the constructs into COS7 cells, isolated the Flag-tagged proteins, and tested for kinase activity. As expected, V600EBRAF had increased kinase activity and G466VBRAF had reduced kinase activity. Y472CBRAF showed severely reduced kinase activity that was less than 10% that of wtBRAF (Fig. 1A).

Fig. 1

Y472CBRAF is biochemically similar to G466VBRAF but different from V600EBRAF. (A) Kinase activity in Flag-tagged BRAF proteins with the indicated mutations. After immunoprecipitation (IP) from COS7 cells and incubation with kinase-dead MEK (substrate) and ATP, MEK phosphorylation was measured using Western blotting and quantitated with Image J. (B to D) Effect of the expression of mutated BRAF proteins on MEK and ERK activity in intact cells. Western blotting was performed in (B) COS7, (C) H661, and (D) H226 cells transfected with BRAF constructs. (E) Activity of CRAF in H661 cells transfected with mutated BRAF, as measured by incubating immunoprecipitated CRAF with kinase-dead MEK (substrate) and ATP, and detecting phosphorylated MEK1/2 (pMEK1/2). (F) CRAF kinase activity and BRAF-CRAF binding, as measured in COS7 cells transfected with Flag-tagged BRAF with the indicated mutations by kinase activity [in vitro kinase assay (IVKA)] or IP of CRAF followed by blotting with anti-Flag and anti-BRAF antibodies (binding). WT, wild type.

Lung cancer cells expressing Y472CBRAF express activated MEK, ERK, and CRAF

Cancer cells that express kinase-impaired mutated BRAF genes still activate MEK and ERK via transactivation of CRAF (4, 5). To determine whether Y472CBRAF functions similarly, we transfected COS7, H226, and H661 cells (all with wtBRAF) with the BRAF construct panel and measured their ERK and MEK activity (Fig. 1, B to D). As we expected, cells transfected with V600EBRAF showed markedly more phosphorylation of ERK and MEK than those transfected with wtBRAF. Both Y472CBRAF and G466VBRAF activated MEK and ERK to levels at or above those observed after transfection with wtBRAF. Cells transfected with wtBRAF had levels of total BRAF similar to those in cells transfected with mutant BRAF (Fig. 1, C and D).

To determine whether Y472CBRAF transactivates CRAF, we assayed for CRAF kinase activity in H661 cells transfected with our panel of BRAF constructs. Cells expressing G466VBRAF or Y472CBRAF had higher CRAF activity than did those cells expressing wtBRAF or V600EBRAF (Fig. 1E). Likewise, cells expressing G466VBRAF or Y472CBRAF had higher BRAF-CRAF binding than did cells expressing wtBRAF or V600EBRAF (Fig. 1F). These results confirmed that Y472CBRAF functions in the RAS/RAF/MEK/ERK pathway similarly to previously characterized BRAF proteins with inactivating mutations.

Dasatinib leads to senescence and apoptosis in lung cancer cells expressing kinase-impaired BRAF

To determine whether Y472CBRAF was related to PX’s response to dasatinib, we tested a panel of NSCLC cell lines for their sensitivity to dasatinib. After extensive searching in multiple databases, we identified only two NSCLC cell lines with known inactivating BRAF mutations: H1666 and Cal12T. These two lines, along with one cell line expressing wtBRAF (H322), were sensitive to dasatinib (Table 1). Researchers previously characterized H322 cells as having a c-Src amplification that drives their sensitivity to dasatinib (7). All other lines tested were resistant to dasatinib, including those with activating BRAF mutations (Table 1).

Table 1

Lung cancer cell sensitivity to dasatinib. IC50, half-maximal inhibitory concentration; wt, wild type; ND, not done.

View this table:

We did not observe significant apoptosis when we treated H1666 or Cal12T cells with dasatinib (fig. S5), but the cells did undergo cell cycle arrest with an increased proportion of cells in G1 and decreased proliferation with reduced bromodeoxyuridine (BrdU) incorporation. Further, they stained for β-galactosidase and heterochromatin protein 1–γ (HP1-γ), indicating senescence (Fig. 2, A to D, and fig. S6). Consistent with this finding, dasatinib reduced phosphorylated retinoblastoma (Rb) in H1666 and Cal12T (Fig. 2E).

Fig. 2

NSCLC cells with an inactivating BRAF mutation undergo senescence when exposed to dasatinib. (A to D) Biological effects of dasatinib in NSCLC cells with BRAF mutations. NSCLC cells were incubated with 150 nM dasatinib or a control vehicle for 72 hours followed by β-galactosidase (A) staining and (B) quantification, (C) cell cycle analysis with propidium iodide staining and FACS analysis, or (D) BrdU incorporation. (E) Effect of dasatinib on downstream signaling proteins, as measured by Western blotting of NSCLC cells incubated with 150 nM dasatinib for 24 hours. (F) Reversibility of dasatinib-induced senescence. Cal12T cells were incubated with 150 nM dasatinib or fresh medium (vehicle, but no drug) for the indicated times (left panel), and senescence was measured by β-galactosidase staining and light microscopy. *P < 0.05; **P < 0.001.

To determine whether dasatinib-induced senescence was reversible, we incubated Cal12T and H1666 cells with dasatinib for 6 to 96 hours followed by drug removal and measurement of β-galactosidase expression (Fig. 2F and fig. S7). Induction of significant tumor cell senescence required 72 hours of exposure to dasatinib and was largely irreversible by 96 hours.

To further study the effects of inactivating BRAF mutations on dasatinib sensitivity, we transfected H661 and H226 cells with the panel of BRAF constructs. Expression of the inactivating mutations increased the cells’ sensitivity to dasatinib, whereas expression of V600EBRAF induced further resistance (Fig. 3, A and B, and fig. S8). In contrast, H1666 and Cal12T cells, which have endogenous BRAF inactivation (G466VBRAF), and H661 cells, which we transfected with kinase-impaired BRAF, underwent apoptosis and not senescence when exposed to dasatinib (Fig. 3C).

Fig. 3

Expression of kinase-inactive and kinase-active BRAF leads to sensitivity and resistance to dasatinib, respectively, in NSCLC cells. (A, B, and D) Cell viability was assayed after dasatinib treatment in NSCLC cells transfected with BRAF constructs. After transfection with the noted BRAF constructs, (A) H661, (B) H226, and (D) H1666 cells were incubated with increasing doses of dasatinib for 72 hours, and cell viability was estimated with the MTT assay. The half-maximal inhibitory concentration (IC50) of dasatinib was calculated for H661. (C) Apoptosis in H661 cells expressing BRAF with the indicated mutations, as measured by the TUNEL assay.

To confirm the importance of the inactivating BRAF mutations in mediating dasatinib sensitivity in H1666 cells, we transfected these cells with kinase-active BRAF (that is, V600EBRAF or wtBRAF) and treated them with dasatinib (Fig. 3D). Overexpression of kinase-active BRAF increased resistance to dasatinib, confirming the role of kinase-inactive BRAF in mediating dasatinib sensitivity. None of the constructs in the BRAF construct panel had a significant effect on cell number in untreated cells (fig. S9).

Dasatinib indirectly inhibits CRAF

The mechanism by which dasatinib induces senescence in NSCLC cells with kinase-deficient BRAF is unknown. We hypothesized that dasatinib induces senescence by affecting CRAF function (8). Although we found that dasatinib did not directly affect CRAF or BRAF kinase activity at relevant concentrations (Fig. 4A), which is consistent with published studies (7, 9), dasatinib did lead to decreased CRAF activity in intact cells that express kinase-inactive BRAF (Fig. 4B). To establish the importance of CRAF in NSCLC cells with kinase-impaired BRAF, we knocked down CRAF expression using small interfering RNA (siRNA) in NSCLC cells with kinase-active or kinase-inactive BRAF. CRAF knockdown in H1666 (G466VBRAF) and Cal12T (G466VBRAF) cells, but not H322 (wtBRAF) or H661 (wtBRAF) cells, affected their viability as estimated with an MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] assay (Fig. 4C), although the effects on Cal12T cells were modest. Given the limited number of cell lines expressing endogenous mutated BRAF, we used a less physiologic approach and transfected H661 cells with kinase-active or kinase-inactive BRAF. Only cells expressing kinase-deficient BRAF showed reduced cell number after treatment with CRAF siRNA (fig. S10).

Fig. 4

Dasatinib indirectly inhibits CRAF activity in cells with an inactivating BRAF mutation but not in cells with wtBRAF or V600EBRAF. (A) Kinase activity of isolated BRAF and CRAF incubated with dasatinib. Purified recombinant BRAF and CRAF were incubated with 100 nM dasatinib in the presence of inactive MEK1 and ATP. Kinase activity was measured by phosphorylation of MEK1. (B) Effect of dasatinib on the kinase activity of CRAF in intact cells expressing various BRAF constructs. COS7 cells were transfected with Flag-tagged CRAF along with HA-tagged BRAF constructs with the noted mutations 24 hours before dasatinib treatment. Transfected cells were incubated with 150 nM dasatinib or vehicle for 24 hours followed by IP with an anti-Flag antibody. The resulting lysate was subjected to a Western blot (WB) with anti-Flag and anti-HA antibodies (lower two rows), and CRAF kinase activity was measured by incubating Flag-immunoprecipitated CRAF with kinase-dead MEK1 (substrate) and ATP (IVKA, top row). MEK activation was quantitated and normalized with total Flag. (C) Effect of CRAF knockdown (KD) on NSCLC cell number in cells with endogenous BRAF mutations. Untransfected H1666, Cal12T, H322, and H661 cells were incubated with CRAF siRNA, and their viability was estimated with an MTT assay. In all cases, Western blotting was used to confirm CRAF knockdown. At 120 hours, Cal12T cells lost dasatinib sensitivity and CRAF expression recovered.

Dasatinib induces RAF dimerization

Consistent with the recent findings of Packer et al., we found that dasatinib induced RAF dimerization in cells with a KRAS mutation (A549) (10). In addition, dasatinib induced RAF dimerization in NSCLC cells with kinase-impaired BRAF (Fig. 5A). Although dasatinib caused a modest inhibition of RAF kinase activity, when the data were corrected for the increase in total RAF dimers, there was no net inhibition of RAF in Cal12T and H1666 cells. Similarly, ERK was activated and p21 expression was induced after dasatinib treatment in cells with kinase-impaired BRAF, consistent with the oncogene-induced senescence that is observed with KRAS or BRAF activation (Fig. 2E) (11). ERK and MEK were inhibited in H661 and A549 cells, which is consistent with their lack of senescence after dasatinib exposure.

Fig. 5

Dasatinib induces RAF dimerization in NSCLC cells, which is necessary for sensitivity in cells expressing kinase-impaired BRAF. (A) Dasatinib induced RAF dimerization. NSCLC cells were incubated with 150 nM dasatinib for 72 hours, and then CRAF was immunoprecipitated. The resulting lysate was split with part used for blotting with antibodies to BRAF or CRAF (to measure dimerization) and part used to measure CRAF kinase activity using MEK as a substrate. (B and C) Introduction of a mutation that interferes with RAF dimerization abrogates the effects of kinase-deficient BRAF on dasatinib sensitivity. H661 cells transfected with the BRAF constructs as indicated were subjected to Western blotting (B) and an MTT assay (C).

To further investigate the role of BRAF/CRAF heterodimerization in dasatinib-induced senescence, we transfected a BRAF mutant that prevents dimerization (R509HBRAF) into NSCLC cells with endogenous wtBRAF either as a single mutation or in cis with Y472CBRAF (Fig. 5B) (12). As before, the expression of Y472CBRAF led to the activation of ERK and increased sensitivity to dasatinib. The addition of R509HBRAF to Y472CBRAF inhibited Y472CBRAF’s ability to activate ERK. NSCLC cells expressing the double mutations were not more sensitive to dasatinib than those expressing wtBRAF (Fig. 5C).

Drug-induced RAF dimerization alone did not induce senescence in Cal12T or H1666 cells. Nilotinib induced more robust BRAF/CRAF dimerization than did dasatinib in Cal12T cells but did not induce senescence (fig. S11). Together, these experiments demonstrate that dasatinib-induced RAF dimerization is essential but not sufficient for dasatinib sensitivity.

A preliminary investigation of ERK-independent pathways downstream of RAF demonstrated no mutation-specific changes in BAD or JNK (c-Jun N-terminal kinase) (13). Activated Aurora A and PLK1 (Polo-like kinase 1) were not detected by Western blotting (fig. S12) (14). Src family kinase (SFK) inhibition with AZD0530 or knockdown with siRNA was not adequate to induce significant senescence or cytotoxicity (figs. S11 and S13).

A BRAF inhibitor plus dasatinib leads to augmented cytotoxicity in cancer cells resistant to dasatinib

The sensitivity of cancer cells with inactivating BRAF mutations to dasatinib suggests BRAF’s synthetic lethality with a dasatinib target. We treated cell lines with wtBRAF and marked resistance to dasatinib (15, 16) with dasatinib plus the pan-RAF inhibitor sorafenib or dasatinib plus the BRAF inhibitor PLX4032 (vemurafenib) and then measured the treatment’s effect on these lines’ viability in vitro. In all cases, dasatinib enhanced the effect of the RAF inhibitors at clinically relevant doses (Fig. 6 and fig. S14). Although specific CRAF and pan-RAF inhibitors may be as effective as dasatinib in treating cancers with inactivating BRAF mutations, no credible direct CRAF inhibitors are currently in clinical development, and pan-RAF inhibitors have poor in vivo activity and are not specific (17, 18).

Fig. 6

Dasatinib enhances the cytotoxicity of PLX4032 in cancer cells resistant to dasatinib. Dasatinib-resistant NSCLC cells were incubated with single-agent dasatinib (50 to 600 nM) or single-agent PLX4032 (1000 to 12,000 nM) or with a combination of both (dasatinib/PLX4032 = 1:20) for 72 hours. Their viability was estimated with an MTT assay.


Metastatic NSCLC is a common and fatal disease with a 4-year survival rate of only 2%, but personalized therapies that target specific genetic aberrations in NSCLC tumors are remarkably successful. We identified a patient with stage IV NSCLC with long-term disease control by single-agent dasatinib therapy alone. His tumor harbored a kinase-inactivating Y472CBRAF mutation that was likely responsible for its unusual sensitivity to dasatinib. Although effective treatments are available for melanoma patients with activating BRAF mutations, no such therapies are available for patients with cancers that harbor inactivating mutations.

We did not observe any inactivating BRAF mutations in any other patients in the clinical trial, all of whom were nonresponders. As with previously characterized kinase-inactivating BRAF mutations, Y472CBRAF expression led to CRAF, MEK, and ERK activation. Also, NSCLC cell lines with endogenous inactivating BRAF mutation underwent senescence when exposed to dasatinib; the cell lines’ dasatinib sensitivity was reversed with the overexpression of active BRAF. Whereas NSCLC cells transfected with an activating BRAF mutation were more resistant to dasatinib than were controls, transfection of NSCLC cells with kinase-impaired BRAF led to their increased dasatinib sensitivity. These data and PX’s pattern of clinical response are consistent with the conclusion that his tumor underwent senescence when exposed to dasatinib owing to its inactivating BRAF mutation. The mechanism of dasatinib-induced senescence and apoptosis in NSCLC cells expressing kinase-impaired BRAF is unknown but may relate to increased RAF dimerization leading to ERK activation, consistent with the paradigm of oncogene-induced senescence that occurs after moderate BRAF or KRAS activation (11). However, the modest effects of dasatinib on ERK suggest that ERK-independent pathways are involved as well.

BRAF mutations occur in only about 4% of NSCLC cases but are more common in other tumors, such as melanoma (50%) and papillary thyroid cancer (40%) (19, 20). Most of the BRAF mutations cause activation of the kinase (20). In NSCLC, 57% of mutations are V600EBRAF and the remaining 43% are a mixture of kinase-inactivating, activating, and uncharacterized mutations (19). CRAF mutations are rare in all cancers (21).

Interactions between CRAF and BRAF are complex and incompletely understood despite several recent elegant studies in melanoma (4, 5, 22, 23). The pathway is not linear and is further complicated by the multiplicity of signaling molecule components; BRAF and CRAF are similar to each other but do not function identically and are not interchangeable, and feedback pathways inactivate upstream components of the pathway (that is, a negative feedback amplifier) (24). Additionally, MEK is not the only RAF substrate (25).

The complexity of BRAF-CRAF interactions is epitomized by the RAF inhibitor paradox. Specifically, inhibition of BRAF activity in cells that express V600EBRAF results in expected inhibition of ERK activity and subsequent apoptosis; melanomas with V600EBRAF respond clinically to BRAF inhibition (26). Paradoxically, BRAF inhibition in cells with active RAS leads to ERK activation via activation of CRAF (4). These observations are supported by previous research demonstrating that CRAF and BRAF heterodimers are more active than is either the CRAF or the BRAF homodimer even when the BRAF protomer has an inactivating mutation (23).

These complex interactions of RAS, BRAF, and CRAF explain the existence and function of inactivating BRAF mutations in cancer cells, which superficially seem to be contrary to natural selection. Activation of CRAF not only allows cells that express kinase-impaired BRAF to survive but also may further promote the cancer phenotype by activating non–ERK-dependent pathways or promoting aneuploidy (25, 27). CRAF activation may also be responsible for BRAF inhibitor–induced squamous cell carcinomas (26).

Although much of the research on BRAF has focused on melanoma, the BRAF pathway undoubtedly is active and important in NSCLC. RAS is commonly activated in NSCLC via upstream growth factor receptors or activating mutations. Active KRAS signals predominantly through CRAF, not BRAF (28). The role of CRAF in NSCLC has yet to be elucidated, although CRAF is overexpressed in NSCLC cells, and its forced overexpression leads to lung adenomas in a transgenic mouse model (29).

Oncogene-induced senescence occurs after activation of oncogenes such as RAS and RAF. Classically, it is mediated by activation of the p16INK4A/Rb and/or p14ARF/p53 tumor suppressor pathway. Senescence can also be induced and mediated by other pathways, such as those involving c-Src, STAT3 (signal transducer and activator of transcription 3), c-Myc, FOXO4, Chk2, and JNK. Here, dasatinib induced senescence in H1666 and Cal12T cells but induced apoptosis in H661 cells into which we had transfected kinase-impaired BRAF mutations. Possible reasons for this discrepancy are that the absolute level of endogenous wtBRAF, the level of mutant BRAF, or the wild-type/mutant BRAF ratio may influence the outcome of dasatinib-based treatment. The level of mutant BRAF expression was higher in the transfected cells, which also harbored a full complement of endogenous wtBRAF, than in nontransfected cells.

A limitation of our study was a paucity of cell lines and patients with endogenous inactivating BRAF mutations. Our conclusions would be greatly enhanced by demonstrating BRAF mutations in patients with other cancers that responded clinically to dasatinib. Another limitation of our study is that we did not fully elucidate the molecular mechanisms underlying dasatinib’s biological effects in NSCLC cells with mutant BRAF.

Occasionally, spontaneous tumor regression occurs in melanoma and renal cell carcinoma cases and is thought to be immune-mediated. Also, the activating V600EBRAF mutation in melanoma may induce an immune response. Although we cannot exclude such a possibility in the case of PX, spontaneous tumor regression is rare in NSCLC cases. Instead, the dasatinib sensitivity of NSCLC cell lines with kinase-impaired BRAF mutations is consistent with the patient having experienced a direct antitumor effect of dasatinib rather than an immune-mediated mechanism.

Other potential mechanisms that we could not fully explore are the roles of other dasatinib targets in mediating dasatinib sensitivity. Although no protein expression has been linked to dasatinib sensitivity, Sos et al. demonstrated that copy number gains influence sensitivity of NSCLC cells to kinase inhibitors (7). H322 cells, which have c-Src amplification, are sensitive to dasatinib. PX’s tumor had copy number gains in several dasatinib targets. Because live tumor tissue specimens are not available from PX, determining whether his tumor was dependent on HCK, DDR1, EPHA3, or ARG for survival is impossible (30).

In conclusion, we demonstrated that the kinase-inactive mutation Y472CBRAF functions similarly to other known inactivating BRAF mutations and was likely responsible for PX’s response to dasatinib. Moreover, the potential for synthetic lethality of combination therapy including dasatinib and BRAF inhibitors may lead to additional therapeutic options. Given that dasatinib and BRAF inhibitors are in clinical use, our work has potential for direct clinical applications.

Materials and Methods

Please see the Supplementary Materials for immunofluorescence microscopy, copy number variation analysis, mutational analysis, cell culture, genomic DNA preparation, immunohistochemistry, and fluorescence in situ hybridization.


Antibodies used in this study were phospho-SFK (Y416), phospho-ERK1/2 (T202/Y204), p21cip, Bcl2, total ERK, phospho-MEK1/2 (S217/221), phospho-EphA2 (Y594), and phospho–c-Kit (Y719) (Cell Signaling Technology); total CRAF (BD Biosciences); total BRAF, phospho-PDGFRα (Y754), total PDGFRα, c-Kit, Flag M2, and agarose-conjugated CRAF (Santa Cruz Biotechnology); Flag and β-actin (Sigma Chemical Co.); p53 (Dako); HP1-γ (Millipore); and anti-mouse Alexa Fluor 594 (Molecular Probes). Dasatinib, PLX4032, and sorafenib were purchased from Selleck Chemicals and prepared as 10 mM stock solutions in dimethyl sulfoxide (DMSO).

Cytotoxicity assay

Cytotoxicity in NSCLC cells was assessed with an MTT assay as described (31). Under each experimental condition, at least four independent wells were treated. Median effects of drugs on viability were calculated using the Chou-Talalay equation (32) with the CalcuSyn software program (Biosoft).

Senescence-associated β-galactosidase staining

NSCLC cells were processed with a senescence-associated β-galactosidase staining kit (Cell Signaling Technology) according to the manufacturer’s instructions and visualized under an Olympus 1X71 phase microscope (Olympus America). In brief, upon completion of dasatinib-based treatment, cells were washed with phosphate-buffered saline to remove residual medium and fixed. A β-galactosidase staining solution containing X-galactosidase was then added to the fixed cells and incubated at 37°C overnight in a dry incubator without CO2. Fields with at least 100 cells were counted in triplicate.

BRAF and CRAF kinase assays

The kinase activity of immunoprecipitated endogenous BRAF and CRAF protein from NSCLC cells, purified recombinant proteins, or immunoprecipitated Flag-tagged BRAF or Flag-tagged CRAF protein expressed in COS7 cells was measured with an in vitro kinase assay (IVKA) kit for RAF (Millipore). Purified recombinant BRAF and CRAF (Sigma Chemical Co.) and immunoprecipitated proteins [technique described below with anti-Flag, anti-hemagglutinin (HA), anti-BRAF, or anti-CRAF antibodies] were incubated with 100 to 250 μM adenosine triphosphate (ATP)/Mg2+ along with 1 μg of inactive recombinant glutathione S-transferase (GST)–MEK1 at 30°C for 30 min and subsequently boiled with 1× sample buffer to stop the reaction. MEK1 activation was quantified by measuring the phospho-specific MEK1 band (ImageJ software program; National Institutes of Health) after Western blotting (described below).

Cell cycle, proliferation, and apoptosis assays

For cell cycle analysis, cells were harvested, fixed, and stained with propidium iodide, and their DNA content was analyzed with a cytofluorimeter and fluorescence-activated cell sorter (FACS) (FACScan; Becton-Dickinson) and the ModFit software program (Verity Software House) (31). To measure apoptosis, we subjected the fixed cells to TUNEL staining according to the manufacturer’s procedure (APO-BRDU kit; Phoenix Flow Systems) and quantitated them with the FACS (31). BrdU incorporation was measured according to the manufacturer’s instruction (BrdU Flow Kits, BD Biosciences). Briefly, subconfluent cell cultures were treated with DMSO or 150 nM dasatinib for 72 hours and then labeled with 10 μM BrdU for 4 hours. Cells were trypsinized, fixed, and stained with fluorescein isothiocyanate–conjugated anti-BrdU antibody and 7-aminoactinomycin D (7-AAD). Samples were analyzed by two-dimensional flow cytometry to detect both fluorescein and 7-AAD.

Transfection of NSCLC cells with an siRNA

Cells were harvested, washed, and suspended (106 cells/100 μl) in Nucleofector Solution V (Amaxa), and siRNA (200 pmol/100 μl) was added to the cells. Cells were then electroporated with the Nucleofector program U24 (Amaxa), diluted with a prewarmed 500-μl RPMI medium supplemented with 10% serum, and plated onto 60-mm plates, and the medium was changed 16 hours later. The c-Src and CRAF siRNAs were predesigned as sets of four independent sequences (siGENOME SMARTpool; Dharmacon). Controls were cells transfected with a nontargeting (scrambled) siRNA and mock-transfected cells (that is, no siRNA).

Site-directed mutagenesis

Flag-tagged wtBRAF and V600EBRAF plasmids were provided by W. Kolch (Systems Biology Ireland and The Conway Institute, University College Dublin). The wtBRAF construct was used as a template to create the Y472CBRAF, G466VBRAF, R509HBRAF, and Y472C/R509HBRAF mutations with the QuikChange XL Site-Directed Mutagenesis kit (Stratagene) according to the manufacturer’s instructions. The sense mutagenic primers used were 5′-GATCATTTGGAACAGTCTGCAAGGGAAAGTGGCATGGT-3′ for Y472CBRAF, 5′-GTGGGACAAAGAATTGGATCTGTATCATTTGGAACAGTC-3′ for G466VBRAF, and 5′-GTAGGAGTACTCAGGAAAACACACCATGTGAATATCCTACTCT-3′ for R509HBRAF and Y472C/R509HBRAF. To confirm successful introduction of the mutations, we Sanger-sequenced six different plasmids for each mutant.

Immunoprecipitation and Western blot analysis

For both Western blot and immunoprecipitation (IP) analysis, cells were lysed on ice, and the lysates were centrifuged at 20,000g for 5 min at 4°C as described previously (31). For IP, equal amounts of the protein cell lysate supernatant (500 mg) were precleared with protein A and G Sepharose beads (Invitrogen). The precleared lysate was incubated with an IP antibody (anti-Flag or anti-CRAF) overnight. The immunoprecipitates were washed four times with an immunocomplex wash buffer [50 mM tris-HCl (pH 7.5), 100 mM NaCl, 1% Triton X-100, 1 mM egtazic acid, 1 mM EDTA, 1% glycerol, leupeptin (20 μg/ml), aprotinin (10 μg/ml), 1 mM phenylmethylsulfonyl fluoride, and 1 mM sodium vanadate] and boiled with 1× sample buffer for 5 min. For both the IP and the Western blot analysis, equal protein aliquots were resolved with SDS–polyacrylamide gel electrophoresis, transferred to nitrocellulose membranes, immunoblotted with a primary antibody, and detected with a horseradish peroxidase–conjugated secondary antibody (Bio-Rad Laboratories) and ECL reagent (Amersham Biosciences).

Clinical tissue specimens

Patients’ tissue specimens were obtained in an M. D. Anderson Institutional Review Board–approved clinical trial in which all participants gave permission for testing of residual tumor tissue. For PX, residual tissue from a metastatic axillary lymph node resected 2 months before treatment was used. The metastasis was formalin-fixed and paraffin-embedded according to a standard protocol. One 4-μm tissue section was cut and stained with hematoxylin and eosin. Histopathological tumor examination and visual estimation of the tumor area and tumor cell percentage were conducted by a pathologist. The tissue section had a lymph node (2.5 cm by 1.8 cm) extensively infiltrated with malignant epithelial cells representing ~60% of the whole tissue section. Five 10-μm unstained tissue sections were cut and placed in sterile tubes for genomic DNA (gDNA) extraction and isolation.

Supplementary Materials


Fig. S1. Radiographic images of PX’s primary lung tumor and paraspinal metastasis and overall survival.

Fig. S2. Gene copy number variation assessed across all chromosomes in PX’s metastatic lymph node.

Fig. S3. Sequencing chromatograms demonstrate BRAF mutations.

Fig. S4. Alignment of the human BRAF amino acid sequence.

Fig. S5. NSCLC cells with kinase-inactive BRAF mutations did not undergo apoptosis.

Fig. S6. Cells with inactive BRAF undergo senescence in the presence of dasatinib.

Fig. S7. H1666 cells undergo irreversible senescence in the presence of dasatinib at 72 hours.

Fig. S8. H661 cells transfected with inactivating BRAF mutations show increased sensitivity to dasatinib.

Fig. S9. Transfection with mutant or wild-type BRAF does not affect cell number.

Fig. S10. NSCLC cells with an inactivating BRAF mutation are sensitive to CRAF knockdown.

Fig. S11. Kinase inhibitor–induced RAF dimerization does not result in drug sensitivity or senescence.

Fig. S12. Dasatinib does not have any BRAF mutation–specific changes in BAD or JNK phosphorylation.

Fig. S13. Inhibition of c-Src does not affect cell viability in NSCLC harboring a kinase-inactive BRAF mutation.

Fig. S14. Dasatinib enhances the cytotoxicity of sorafenib in cancer cells resistant to dasatinib.

Table S1. Immunohistochemistry scores for PX’s tumor.

Table S2. Genes analyzed using mass spectroscopy single-nucleotide polymorphism analysis.

Table S3. Copy number variation for genes associated with dasatinib targeting and/or NSCLC.

Table S4. Mutational statuses of patients with NSCLC treated with dasatinib.

References and Notes

  1. Acknowledgments: The Flag- and HA-tagged wtBRAF, V600EBRAF, and wtCRAF plasmids were provided by W. Kolch (Systems Biology Ireland and The Conway Institute, University College Dublin). We thank L. M. Solis (M. D. Anderson) for histopathological examination of the tissue specimen, C. L. McDowell (M. D. Anderson) for assistance with gDNA preparation, U. Giri (M. D. Anderson) and M. Peyton (University of Texas Southwestern) for assistance with cell line acquisition, and R. Thomas and M. Sos (Max Planck Institute, University of Cologne) for performing the DDR2 mutational analysis. We thank D. R. Norwood and D. Hackett (M. D. Anderson) for editorial assistance and L. Strauss and W. Geese (Bristol-Myers Squibb) for critical reading of the manuscript. Funding: This work was supported by the donations of patients with cancer and their families. Our previous clinical trial was supported by the National Cancer Institute’s Cancer Therapy Evaluation Program (N01 CM-62202), the University of Texas Lung SPORE (P50 CA097007), and Bristol-Myers Squibb. FACS, microscopy, and sequencing were supported in part by the NIH through M. D. Anderson’s Cancer Center Support Grant CA016672. Author contributions: F.M.J., S.P., and B.S. designed and executed the experiments and wrote the paper. T.M. executed the experiments and edited the manuscript. X.T., H.S.E., H.G., and I.W. analyzed the patients’ tissues and interpreted those data. D.J.S. was involved in the original clinical trial design and edited the manuscript. Competing interests: D.J.S. has received consulting fees from the following: Align2Action Inc., Bridger Capital LLC, Easton Associates, Trinity Partners LLC, Warburg Pincus LLC, and AstraZeneca Taiwan. The other authors declare that they have no competing interests. A patent has been filed on this work: “The novel use of dasatinib for patients with tumors that harbor kinase-inactive BRAF mutations.”
View Abstract

Navigate This Article