Gastrointestinal stromal tumors (GISTs) are mesenchymal tumors characterized by mutations of KIT or PDGFRA. The objectives of this study were to evaluate BRAF mutations in GISTs and then to correlate BRAF mutational status in the tumor with clinical parameters, with B-raf expression, and with activation of some cellular pathways. BRAF mutation was screened in 321 GISTs with 70 wild-type GISTs. BRAF V600E was detected in 9 (13%) of 70 wild-type GISTs. No mutations were detected in GISTs bearing KIT or PDGFRA mutations. BRAF V600E detection in the tumor does not induce a higher expression of the B-raf protein or the preferential activation of the p42/44 mitogen-activated protein kinase (MAPK) signaling pathway compared with GISTs without the BRAF mutation. In comparison with the GIST group with KIT or PDGFRA mutation or the wild-type GIST group without BRAF mutation, the wild-type GIST group with a BRAF mutation is not different in terms of B-raf expression or the p44/42 MAPK- or AKT-activated signaling pathway.
Gastrointestinal stromal tumors
Gastrointestinal stromal tumors (GISTs) are mesen-chymal tumors arising in the intestinal tract, mainly in the stomach and the small intestine. They are characterized by mutations in the KIT (80%) or the PDGFRA gene (8%).1,2 Nevertheless, approximately 10% to 15% of GISTs are free of KIT or PDGFRA mutations.3 The downstream signaling cascade activated by KIT includes the Ras-Raf–mitogen-activated protein kinase (MAPK) pathway, the phosphati-dylinositol-3-kinase pathway, and the Jak-STAT kinase pathway.4,5 According to the type of KIT mutation, one of these pathways is activated preferentially.5
The BRAF gene encodes for a serine/threonine-protein kinase.6 Its function is to control proliferation and differentiation through the Ras-Raf-MAPK pathway. It is mutated in a wide range of cancers, with a high rate in malignant melanoma, thyroid carcinoma, and colorectal cancer with microsatel-lite instability.7 Most mutations lie within the kinase domain with a single nucleotide substitution at position 1799 in exon 15, leading to the V600E amino-acid substitution (98%). The remaining 2% are located in exon 11. Germline mutations of BRAF have been described in the cardio-facio-cutaneous syndrome within exons 6, 11, 12, and 14.8,9 Recently, the BRAF mutation was detected in 3 GISTs without KIT or PDGFRA mutation and in 1 tumor after imatinib treatment failure.10
Imatinib mesylate is the first drug to target directly the mutated protein responsible for the cancer.11 It inhibits KIT, PDGFRA, and BCR-ABL tyrosine kinase in GISTs and chronic myelogenous leukemia. It is also very effective in GISTs bearing the exon 11 KIT mutation and in GISTs bearing the exon 9 mutation, but at higher dose.12 GISTs with KIT mutations in exon 13 or 17, with p.Asp842Val mutations in PDGFRA or without mutations in KIT or PDGFRA are more frequently imatinib-resistant. The challenge is not only to find new drugs able to act on GISTs resistant to the available drugs but also to identify other genomic alterations responsible for this cancer. In this study, we analyzed a series of 321 GISTs for BRAF mutations. BRAF mutation detection in GISTs would be interesting because therapeutic targets against BRAF V600E are under investigation and seem promising.13,14
Materials and Methods
GISTs were collected from the Department of Pathology, Institut Bergonié (Bordeaux, France). Tumors were collected in agreement with the national ethical committee. The diagnosis of GIST was based on tumor location, histologic aspect, c-KIT and protein kinase C θ immunoreactivity, and molecular analysis. Most of the GISTs reported in this study were immunohistochemically positive for c-KIT, except 13 cases in which PDGFRA mutations were identified and 2 cases (with no KIT nor PDGFRA mutations) positive for protein kinase C θ immunostaining.
For the study, 524 GISTs from Institut Bergonié were analyzed for KIT and PDGFRA mutations from 2004 to 2008. KIT and PDGFRA mutations were detected in 86.6% of the cases, approximately 75% and 25%, respectively, whereas GISTs with no KIT or PDGFRA mutations were detected in 13.4% of the cases (70 cases). Among GISTs bearing a KIT or PDGFRA mutation and analyzed for BRAF mutation (251 tumors), 12 tumors were metastatic; 5 represented relapse, 4 tumors after treatment with imatinib mesylate; 2 were primary tumors after treatment with imatinib mesylate; and others were primary untreated tumors. For GISTs without KIT or PDGFRA mutations (70 tumors), 3 tumors involved relapse, 2 were metastatic, 2 were primary tumors after treatment, and others were primary untreated tumors. Protein S-100 staining was analyzed for wild-type (WT) GISTs and was negative or focal in all cases.
Morphologic, immunohistologic, and clinical data for GISTs without KIT or PDGFRA mutations are shown in Table 1.
For fixed tissues, fifteen to thirty 10-μm sections were cut from the paraffin block, depending on tumor cell density. Sections were scraped off the block, deparaffinized twice in toluene, rinsed twice with absolute ethanol, and washed with TNE (10 mmol/L tris(hydroxymethyl)aminomethane [Tris], 1 mmol/L EDTA, and 100 mmol/L sodium chloride). Tissues were resuspended with 500 μL TNE with added Proteinase K (final concentration, 10 μg in 100 μL of TNE, Promega, Madison, WI) and incubated overnight at 55°C. Finally, DNA was purified through columns (Wizard DNA Clean-up System, Promega), rinsed through minicolumns (Ultrafree-MC30, Millipore SA, Bedford, MA) with TE (Tris, pH 8, 10 mmol/L; EDTA, pH 8, 1 mmol/L), and suspended in 100 μL TE.
For frozen tissues, DNA was extracted from 50 mg of tumor according to standard procedure using phenol-chloroform and DNA precipitation with ethanol. The DNA concentration was measured by UV absorbance at 260 nm.
Polymerase Chain Reaction
Polymerase chain reaction (PCR) amplifications were performed on exons 11 and 15 of the BRAF gene using the primer sets BRAF 11F (5′-TTT CTG TTT GGC TTG ACT TGA CT-3′) and BRAF 11R (5′-GTC ACA ATG TCA CCA CAT TAC ATA CT-3′) for BRAF exon 11, BRAF 15F (5′-TCA TAA TGC TTG CTC TGA TAG G-3′) and BRAF 15R (5′-AGT AAC TCA GCA GCA TCT CAG G-3′) for BRAF exon 15. For the studies, 50 ng of DNA was amplified with the following thermal cycling profile: 95°C for 5 minutes; then 45 cycles of 95°C for 30 seconds, 60°C for 1 minute, and 72°C for 1 minute; with a final extension at 72°C for 10 minutes. A temperature cycle was added for PCR products run on denaturing high-pressure liquid chromatography (DHPLC) for heteroduplex formation: 98°C for 10 minutes, 60°C for 30 minutes, and cooling at 4°C.
Denaturing High-Pressure Liquid Chromatography
The WAVE DNA Fragment Analysis System (Transgenomic, Omaha, NE) was used to detect the hetero-duplex PCR products. The elution temperature was calculated using WAVEMAKER software (Transgenomic) and was 56.5°C for BRAF exon 15 and 55.5°C for BRAF exon 11.
Purification of the PCR product was performed using the GFX PCR DNA and gel band purification kit (GE Healthcare Lifesciences, Buckinghamshire, England). Automated cycle sequencing for both strands was performed using the Big Dye Terminator v1.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA). Sequencing reactions were carried out on the ABI Prism 310 Genetic Analyzer (Applied Biosystems). Sequences were compared with the normal sequence by using SeqScape v2.5 software (Applied Biosystems).
PCR amplifications were performed on exon 15 of the BRAF gene using primer sets of BRAF ASF (5′-GGT GAT TTT GGT CTA GCT ACA TA-3′) and BRAF ASR (5′-GGC CAA AAA TTT AAT CAG TGG A-3′). For the studies, 50 ng of DNA was amplified with the following thermal cycling profile: 95°C for 5 minutes; then 35 cycles of 95°C for 30 seconds, 60°C for 1 minute, and 72°C for 1 minute; with a final extension at 72°C for 10 minutes. The reliability of the results was determined for each experiment using a positive control sample (DNA extracted from a papillary carcinoma of the thyroid positive for the BRAF V600E mutation). Moreover, each sample was previously amplified for BRAF exon 15 as described, to ensure that it was possible to obtain a PCR product with a similar size.
B-raf Immunohistochemical Analysis
First, 4-μm-thick paraffin sections were cut and mounted on glass slides (SuperFrost+, Menzel-Gläser, Braunschweig, Germany). Preparations were dried for 1 hour at 58°C and then overnight at 37°C. Sections were deparaffinized with toluene and rehydrated with ethanol. Preparations were preheated, and a heat-based antigen-retrieval method was used before incubation. Endogenous peroxidase was blocked using a 3% hydrogen peroxidase solution for 5 minutes. All slides were immunostained for B-raf (monoclonal F-7, Santa Cruz Biotechnology, Santa Cruz, CA) at a dilution of 1:50. Sections were incubated for 1 hour at room temperature with the primary antibodies followed by staining with a streptavidin-biotin peroxidase kit (Chem-Mate detection kit/DAB, DakoCytomation, Glostrup, Denmark). Sections were then revealed in a diaminobenzidine solution for 15 minutes and stained with hematoxylin for 1 minute.
For Western blotting, 50 mg of tumor was prepared by grinding in the following: a RIPA buffer (Sigma-Aldrich, St Louis, MO) with phosphatase inhibitors (sodium fluoride, 30 mmol/L; β-glycerophosphate, 40 mmol/L; sodium pyrophosphate, 20 mmol/L; and sodium orthovanadate, 1 mmol/L) and with some protease inhibitors (PIC, Sigma). Equivalent amounts of protein (50 μg) from clarified lysates were resolved with sodium dodecyl sulfate–polyacrylamide gel electrophoresis, transferred to polyvinylidene fluoride membranes, and immunoblotted sequentially with the following antibodies: anti-panAKT (clone 11E7, dilution 1/1,000; Cell Signaling Technology, Danvers, MA), anti–p44/42 MAPK (clone 137F5, dilution 1/1,000; Cell Signaling Technology), anti-AKT (clone 11E7, dilution 1/1,000; Cell Signaling Technology) and anti-phospho-AKT (clone 193H12, dilution 1/1,000; Cell Signaling Technology), anti–phospho-p44/42 MAPK (clone 20G11, dilution 1/1,000; Cell Signaling Technology), and anti-β-actin (dilution 1/10,000; Cell Signaling Technology). After washes, membranes were probed with antirabbit immuno-globulin–horseradish peroxidase conjugate and incubated with Amersham ECL Plus substrate (GE Healthcare Lifesciences).
BRAF Mutation Detection
Among the 454 GISTs bearing KIT or PDGFRA, 251 were analyzed for BRAF mutation by DHPLC, sequencing, and/or V600E-allele specific PCR, and no V600E mutation was detected. Three polymorphisms were detected in 2 tumors in intron 14 as c.IVS14-32G>A and c.IVS14-39G>A in one tumor and IVS14-37G>A in the other; owing to localization in the intron, the polymorphisms do not likely induce abnormal splicing. A similar polymorphism has been previously reported.15
We screened 70 GISTs without KIT or PDGFRA mutation for the BRAF mutation by DHPLC, direct sequencing, and V600E allele-specific PCR. Nine cases were positive for V600E mutation, and no other mutations were detected in exon 15 of BRAF. The V600E mutation was only detected by allele-specific PCR for 2 cases. All cases with BRAF mutations were primary tumors. The KIT and PDGFRA mutation detection analysis in GISTs is based on prescreening by DHPLC. Nevertheless, the absence of mutation in the WT GISTs has been confirmed by direct sequencing of KIT (exons 9, 11, 13, and 17) and PDGFRA (exons 12, 14, and 18).
The BRAF exon 11 mutation was screened by direct sequencing in 85 GISTs with a KIT or PDGFRA mutation and in 40 WT GISTs. No mutation was detected, but 2 polymorphisms (p.Arg444Arg) were detected in the WT GIST group.
GISTs with BRAF mutations were mainly localized in the small intestine (5/9 [56%]) but also in the stomach (2/9 [22%]) Table 2. There was statistically no difference in tumor location between WT GISTs with or without BRAF mutations (Fisher exact test). Other histologic and clinical parameters (age, mitoses/50 high-power fields, and tumor size) between the 2 groups were statistically similar (Student t test) Table 3. Three patients with BRAF mutations had a high risk of malignancy (cases 6, 33, and 64), 3 had intermediate risk (cases 11, 42, and 47), and 2 had low risk (cases 28 and 55). For the last patient (case 27), the size of the tumor and the number of mitoses for 50 high-power fields were not determined.
B-raf expression was studied in a total of 37 GISTs. Eight tumors were WT GISTs with BRAF V600E mutations, 9 cases were WT GISTs without BRAF mutations, and 20 cases were GISTs with KIT or PDGFRA mutations. The localization of the staining was cytoplasmic Image 1. The immunohistochemical results for B-raf detection in GISTs are summarized in Table 4. B-raf expression was observed in all cases examined. The intensity of the staining varied from moderate to high, whatever the KIT, PDGFRA, or BRAF mutational status. In the majority of cases, more than 80% of the tumor cells expressed B-raf. No difference for B-raf expression detected by immunohistochemical analysis was observed between the different groups of tumors according to KIT or PDGFRA and BRAF mutational status.
Cellular Signaling Pathway Activation
Activation of the p44/42 MAPK and AKT pathways was investigated in GISTs bearing V600E BRAF mutations in comparison with other GISTs Image 2. Only 1 frozen tumor with a BRAF mutation was available. Frozen tumor was also available for 2 GISTs without KIT or PDGFRA mutation, 2 GISTs with PDGFRA mutations, and 3 with KIT mutations. The p44/42 MAPK signaling pathway was activated in all cases, and the intensity of the signal for MAPK and phos-phorylated MAPK was similar in all cases, except for phos-phorylated MAPK in a case of WT GIST without a BRAF mutation. The AKT signaling pathway was similar in all cases except 2 GISTs with PDGFRA mutations in which the signal for phosphorylated AKT was higher than in the other cases. Nevertheless, the AKT signaling pathway in the case with a BRAF V600E mutation was similar to the WT GIST and GISTs with KIT mutations.
BRAF mutations have been found in a wide range of tumors, but at different levels.7 They are mainly found in melanoma (60%), thyroid papillary carcinoma (40%), and colorectal tumors with microsatellite instability (10%). BRAF is a serine-threonine kinase that activates the MAPK cascade. Mutations of the BRAF gene are mainly localized at exon 15 (nucleotide 1799), replacing a valine at position 600 with an aspartic acid. This modification mimics the phosphorylation of the kinase activation domain leading to permanent activation of the kinase. Recently, KIT mutations were detected in melanoma at a low percentage (2%).16 Some of these mutations (eg, p.L576P) have also been detected in GISTs. KIT and BRAF mutations seem to be exclusive, although Curtin et al17 described a tumor with the BRAF V600E and KIT K642E mutations. It cannot be ruled out that the KIT K642E mutation is germline because it has been detected in a French family.18
Phospho-p44/42 MAPK and phospho-Akt expression in gastrointestinal stromal tumor (GIST) according to BRAF, KIT, and PDGFRA mutational status. No difference in phospho-p44/42 MAPK or phospho-AKT expression was observed according to BRAF mutational status. The p44/42 MAPK pathway is not strongly activated in response to a BRAF V600E mutation presence in the GIST. Findings for specific cases (from Table 4) were as follows: case 6, wild type, BRAF V600E; cases 23 and 39, wild type; cases 72 and 80, PDGFRA; and cases 89, 91, and 92, KIT.
Recently, Agaram et al10 identified BRAF mutations in 4% of GISTs without KIT or PDGFRA mutations and in 1 case of relapse after imatinib treatment. We analyzed BRAF mutations in a series of 321 GISTs (70 tumors with no KIT or PDGFRA mutation and 251 tumors with a KIT or PDGFRA mutation). The BRAF mutation was detected in 9 (13%) of 70 WT GISTs (69 adult and 1 pediatric GIST), and all were the V600E mutation, but was not detected in GISTs with a KIT or PDGFRA mutation. This percentage is higher than previously described.10 Excluding GISTs in children in the study by Agaram et al,10 the frequency of BRAF mutants in adults with GISTs was 7% (3/46 cases). In our series, by excluding the pediatric GIST case (case 62), the frequency of BRAF mutation is 13% (9/69 cases), twice that described by Agaram et al.10
Moreover, for 2 cases in which the percentage of tumor cells was more than 80%, the BRAF V600E mutation was probably present in a low percentage of tumor cells because it was detected only by allele-specific PCR, and the sensitivity has been evaluated to be 1% in our hands. The sensitivity of the technique was evaluated by successive dilutions of the colorectal carcinoma cell line HT29 (bearing the p.Val600Glu mutation) in the colorectal carcinoma HCT15 cell line that does not bear this mutation (data not shown). This observation suggests that some subsets of cells gain this specific genetic alteration in the tumor with consequences that remain to be determined. In the thyroid, the BRAF mutation has been correlated with aggressive pathologic data such as lymph node metastasis, extrathyroidal invasion,19 worse outcome,20 and recurrence. In melanoma, the BRAF mutation has been associated with increased tumor cell proliferation and shorter survival.21,22 A larger cohort of WT GISTs with clinical follow-up should be studied for BRAF mutations to evaluate the prognostic significance of this mutation.
It seems that KIT and BRAF mutations are exclusive as in melanoma tumors because of 251 GISTs with a KIT or PDGFRA mutation, no BRAF mutations for exon 15 or 11 were detected.
The significance of BRAF mutations in GISTs negative for KIT or PDGFRA mutation is unknown, although there should be a biologic impact in the tumor because this mutation has been selected during tumor development. BRAF is an important protein that interacts with ras to transduce a signal through the MAPK pathway,6 a signaling pathway that is activated in GISTs. In evaluating the activation of the MAPK and AKT pathways, no difference was observed between WT GISTs with the V600E mutation (although only 1 WT GIST with a V600E mutation was available as frozen material) and other WT GISTs or GISTs with a KIT or PDGFRA mutation. Although a BRAF mutation is detected, it does not seem to preferentially activate the MAPK pathways. The same observation has been made in melanoma23 and in thyroid; 170 melanomas24 and 42 papillary carcinomas25 were screened for BRAF mutations and compared with MAPK activation pathways by immunohistochemical analysis. No obvious relation between BRAF mutational status and MAPK activation has been found. We have also investigated B-raf immunostaining in WT GISTs with or without BRAF mutations and in GISTs with a KIT or PDGFRA mutation (Table 4). Whatever the mutational status of the tumor, no significant difference in the percentage of tumor cells stained or the intensity of the staining was observed. The same observation was made in thyroid carcinoma.26
As reported by Agaram et al,10 we observed a predominant small intestinal location of GISTs with BRAF V600E (56%) followed by location in the stomach (22%). Nevertheless, the repartition of tumor localization is similar to that reported for GISTs.27 No clinicopathologic data such as tumor size, mitotic count, or age seem to be specific to GISTs bearing V600E BRAF mutations. Unfortunately, we were not able to learn the clinical follow-up for these patients.
We report BRAF mutation status in GISTs without a KIT or PDGFRA mutation. The level of mutation affects 13% of GISTs without a KIT or PDGFRA mutation. This observation may be very important because there is as yet no treatment for this set of GISTs. Although this observation as been reported previously, we observed a higher V600E detection rate in this subset of tumors. Although the preferential location to the small intestine for this group of tumors is not statistically relevant, it confirms what was observed by Agaram et al.10 A drug targeting BRAF is being tested in phase 1 and 2 trials in melanoma28 and renal carcinoma29 and might also be effective in BRAF-mutated GISTs. Further studies are needed to evaluate more precisely the incidence of BRAF mutations in the WT GIST subset and to investigate the function of the mutation in GISTs.
Supported by CONTICANET (Connective Tissue Cancers Network to Integrate European Experience in Adults and Children), Lyon, France.
Antitumor efficacy of the novel REF inhibitor GDC-0879 is predicted by BRAFV600E mutational status and sustained extracellular signal-regulated kinase/mitogen-activated protein kinase pathway suppression. Cancer Res. 2009;69:3042–3051.