Gastrointestinal stromal tumours (GISTs), the most common primary gastrointestinal (GI) mesenchymal tumours, are KIT positive and KIT-signalling driven neoplasms. The stomach represents the most common site involved (60–70%), followed by the small bowel (20–30%).1 GISTs harbour activating KIT mutations in a majority of cases (≈80%).2 Less commonly (≈10%), mutations in the homologue oncogene platelet-derived growth factor receptor α (PDGFRA) are detected in GISTs that often display gastric location and epithelioid morphology.2 3 Activating mutations of KIT or PDGFRA represent early initiating molecular events and are mutually exclusive. On the other hand, a minority of the cases (10–15%), particularly GISTs in paediatric patients,4 those affected by the Carney triad,2 4 neurofibromatosis 1 (NF1)-associated GISTs,5 and a small subset of sporadic adult GISTs,2 are wild-type for KIT and PDGFRA. The pathogenetic mechanisms underlying these wild-type GISTs are poorly understood.

BRAF is a member of the RAS–RAF–MEK–ERK pathway that is involved in cell cycle regulation and oncogenic modulation of cellular responses to growth signals by activating the mitogen-activated protein kinase (MAPK) pathway.6 7 BRAF mutations are commonly detected in diverse benign and malignant tumours, including benign congenital and acquired melanocytic nevi (81%),7^–9 malignant melanoma (50–70%),7 10^–12 papillary thyroid carcinoma (35–60%),7 13 colorectal adenoma/carcinoma (5–20%)7 14 and others.7 Most BRAF mutations cluster in a hot-spot at nucleotide 1799 in exon 15 leading to substitution of valine by glutamic acid at codon 600 (V600E).7 It has been demonstrated that BRAF and KIT mutations represent mutually exclusive alternative oncogenic events in malignant melanoma.10 15 This led Agaram et al to investigate a series of KIT/PDGFRA wild-type GISTs for BRAF mutations in a recent study.16 They detected V600E BRAF mutations in 4% of wild-type GISTs, and in one patient with acquired imatinib resistance who lacked secondary KIT/PDGFRA mutations or other apparent cause of resistance. However, no KIT/PDGFRA-mutated imatinib-naive GISTs were analysed for BRAF mutations. Accordingly, it remains unclear whether KIT/PDGFRA mutations are mutually exclusive with BRAF mutations. Furthermore, there is evidence that BRAF mutations may act in a synergistic manner with other secondary concomitant genetic events, including KRAS mutations and other genetic alterations, to enhance tumour progression and aggressiveness in melanoma, colorectal cancer and thyroid carcinoma.11 12 17^–19 Taking the above observations into consideration, we analysed a series of KIT/PDGFRA mutants and sporadic adult wild-type GISTs for BRAF and KRAS mutations to further explore the frequency of BRAF mutations in GISTs, and to assess whether coexisting KRAS mutation might be responsible for the reportedly aggressive clinical course of BRAF-mutated GISTs.

METHODS

Clinicopathological features and KIT/PDGFRA mutation status of the cohort

The study included 69 tumours from 49 patients. Seventeen patients had 2–3 distinct tumours. Some of the latter group of patients who had multiple primary sporadic gastric tumours have been included in a previous study.20 In all, 41 tumours had a known tyrosine kinase mutation (35 had KIT exon 11 mutations, 4 had KIT exon 9 mutations, and 2 had PDGFRA exon 18 mutations) and 28 tumours were wild-type for KIT exons 9, 11, 13, 17 and PDGFRA exons 12, 14 and 18. KIT/PDGFRA mutant tumours were of gastric (n = 29), small bowel (n = 9) and metastatic or unsure location (n = 3). KIT/PDGFRA wild-type tumours originated in the stomach (n = 11), small bowel (n = 10), colon (n = 1), oesophagus (n = 1), and uncertain or metastatic sites (n = 5). According to the National Institutes of Health (NIH) consensus criteria,21 KIT/PDGFRA-mutated tumours fitted the very low risk (n = 13), low risk (n = 10), intermediate risk (n = 1) and high risk (n = 9) categories. On the other hand, wild-type tumours with a known risk stratification were at very low risk (n = 12), low risk (n = 3), intermediate risk (n = 2) and high risk (n = 6) for aggressive behaviour. None of the patients had evidence of NF1, Carney triad or paediatric GISTs, and none had imatinib therapy prior to surgery.

Molecular analysis for KIT, PDGFRA, BRAF and KRAS mutations

Extraction of tumour DNA from formalin-fixed paraffin-embedded material, PCR amplification, and direct sequencing of KIT exons 9, 11, 13, 17 and PDGFRA exons 12, 14, 18 were performed using the methods previously described.22 For BRAF and KRAS mutation analysis, the sequences of the primers used for the PCR are listed below. BRAF exon 15 forward: 5′-CTTCATGAAGACCTCACAGTAAAAATAGG-3′ BRAF exon 15 reverse: 5′-TAGCCTCAATTCTTACCATCCACAAA-3′ KRAS exon 2 forward: 5′-TTTTTATTATAAGGCCTGCTGAAAATG-3′ KRAS exon 2 forward nested: 5′-TATTATAAGGCCTGCTGAAAATGACTG-3′ KRAS exon 2 reverse: 5′-AATGGTCCTGCACCAGTAATATGCATAT-3′ KRAS exon 2 reverse nested: 5′-GTCCACAAAATGATTCTGAATTAGCTGTA-3′.

BRAF and KRAS amplification required one step and two steps of amplification, respectively. For the first and the nested PCR, a 50 cycle amplification step was performed using an AmpliTaq Gold (Applied Biosystems, Foster City, California, USA) under the following conditions: 20 s at 95°C for denaturation, 10 s at 55°C for annealing, and 40 s at 72°C for elongation. Sequencing PCR and sequence analysis were performed as described previously.22 To investigate for a possible coexistence of different mutations in the same tumour, all tumours were examined for mutations in each of the above four genes, even if a mutation was detected in one gene.

Verification PCR

Each sample in which a mutation could be detected by PCR was reconsidered for amplification using the respective exon primers exclusively. Four independent amplifications were done, and all confirmed or excluded the mutations that were initially suspected.

RESULTS

BRAF and KRAS mutation status

Two of the 28 KIT/PDGFRA wild-type tumours (7%) revealed a heterozygous BRAF V600E mutation in exon 15 of the gene. On the other hand, none of the 41 tumours with known KIT or PDGFRA mutations had a BRAF mutation. None of the 69 tumours had a mutation in exon 2 of the KRAS gene. Normal mucosa was analysed for BRAF mutation in both mutated cases and demonstrated a wild-type sequence, thus excluding a germline BRAF mutation.

Pathological features of the BRAF mutant GISTs

The two patients were males with a mean age of 76 years (table 1). Patient no. 1 had two tumours (3 and 4 mm in diameter) originating in the gastric body; both tumours were incidental autopsy findings (fig. 1A). One tumour harboured a BRAF mutation and the other was wild-type for KIT/PDGFRA/BRAF/KRAS. Patient no. 2 had a 4 mm jejunal GIST found during surgery for gastric carcinoma. All tumours were mitotically inactive spindle-cell GISTs that were indistinguishable histologically from their KIT-mutated counterparts detected at comparable early stage. The gastric tumour from patient no. 1 showed variable stromal hyalinisation (fig. 1B). The small bowel tumour from patient no. 2 was moderately cellular with numerous skeinoid fibres (fig. 1C). All tumours strongly expressed CD117 (fig. 1D). CD34 was expressed by the gastric tumours, but not by the small bowel tumour. The latter expressed focally smooth muscle actin. P16 immunostaining showed weak, but variable, nuclear staining in scattered tumour cells in both cases (<20%, data not shown). Both tumours qualified as very low risk for aggressive behaviour according to the NIH consensus criteria,21 corresponding to prognostic group 1 in the Miettinen’s classification.1 The molecular findings from BRAF-mutated gastric GIST from patient no. 1 are illustrated in fig 2.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Gross, histological and immunohistochemical features of BRAF-mutated GIST. (A) Two minute incidental gastric GISTs forming plaque-like subserosal nodules, one displaying V600E BRAF mutation (right), and the other KIT/PDGFRA/BRAF/KRAS wild-type (left). (B) The BRAF-mutated tumour showed sclerosing spindle morphology. (C) Moderately cellular spindled jejunal GIST with BRAF mutation from patient no. 2 (note skeinoid fibres). (D) Diffuse immunoreactivity for KIT/CD117 in the BRAF-mutated tumour in patient no. 1 (note scattered Cajal cells within non-staining muscle tissue on the left).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Complementary reverse sequence of the heterozygous V600E mutation detected in the tumour illustrated in fig. 1A, B (patient no. 1). An additional red peak (see below the red asterisk) was observed in the BRAF-mutated sequence. The amino acid sequence corresponds to the reverse complementary sequence of the chromatogram. The depicted wild-type (WT) reverse sequence TTT CAC TGT corresponds to the forward ACA GTG AAA sequence encoding for the amino acid sequence T599 E600 K601.

DISCUSSION

This study represents the second report of BRAF mutations in GISTs, and the first one to investigate KIT/PDGFRA-mutated GISTs for BRAF and KRAS mutations. In a recent study, Agaram et al detected the novel V600E BRAF mutation in 3 of 61 KIT/PDGFRA wild-type GISTs (4%) and in 1 of 25 patients with secondary imatinib resistance without detectable secondary mutations in KIT/PDGFRA (4%).16 All three patients in that study were middle-aged women with high-risk spindled small bowel GISTs, with occasional sclerosis and calcification in one case (table 1). The patient with secondary resistance had a PDGFRA-mutated primary epithelioid gastric GIST. The resistant clone harboured the primary PDGFRA mutation and the secondary V600E BRAF mutation. No mutations in the NRAS gene were detected in that study.16

In the current study, we demonstrated that the V600E BRAF mutation may occur in early-stage incidental GISTs at very low risk for malignant behaviour, thus suggesting gate-keeper alternative molecular events in a subset of KIT/PDGFRA wild-type GISTs. The two BRAF mutant tumours in our study were ⩽5 mm in greatest diameter, mitotically inactive, KIT-positive and KIT/PDGFRA wild-type with spindle morphology. Their histology was indistinguishable from their KIT-mutated incidental counterparts reported in more detail previously.23 Analysis of normal tissue in both cases revealed a wild-type BRAF sequence, thus ruling out a germline mutation. None of our 41 KIT/PDGFRA-mutated GISTs revealed a BRAF mutation indicating that these mutations are mutually exclusive, similar to the observation for KIT and BRAF mutations in malignant melanoma.10 15

The oncogenic transforming properties of the BRAF V600E mutation have been demonstrated in diverse cancer types including melanocytes, colorectal cancer and thyroid cancer.11 12 17^–19 24 In particular, it has been shown that mutant BRAF (V600E) cooperates functionally with Rac1b,25 Akt311 and other genetic pathways to sustain cell viability and proliferation in tumours and cancer cell lines. However, the prevalence of V600E BRAF mutations in congenital and acquired melanocytic naevi8 9 suggests that the mutation per se is not sufficient for tumour progression, and that additional genetic alterations and activation of other cooperating oncogenic pathways are needed for a tumour to gain a malignant potential.6 8 11 19 Our observations were in line with the previous report that considered loss of nuclear expression of the tumour suppressor oncoprotein p16 as a possible cofactor in the progression of BRAF-mutated GISTs16 and melanocytic lesions.24 Unfortunately, no frozen tissue was available to evaluate the oncogenic activity of the mutant BRAF in our cases. However, future studies might shed light on the mechanisms by which constitutively activated BRAF might influence KIT activation in this subset of GISTs.

Our observation, that BRAF mutations may occur in minute clinically benign GISTs of small bowel and stomach location without a significant predilection for certain age groups or gender, contrasts with the observations of Agaram et al. It is likely that this represents selection bias in both studies and that BRAF mutations probably occur in a wider range from benign incidental to clinically malignant GISTs in the spectrum of the disease. It is remarkable from the previous study and the current study that 4 of 5 GISTs with a primary BRAF mutation were of small bowel origin, suggesting a strong predilection for tumours arising in the small bowel. Whether BRAF-mutated GISTs would be more prone to genetic instability, and hence would follow a malignant course in later stages of the disease, remains unclear. However, there is sound evidence that BRAF mutations may facilitate the acquisition of secondary genetic events through induction of a higher genetic instability, thereby resulting in a more aggressive clinical course.17 18 Lack of NRAS mutations in the previous study16 and of KRAS mutations in our study suggests that mutations in the RAS family of oncogenes probably play no or little role in GIST pathogenesis, but this needs further validation.

In summary, we demonstrated that V600E BRAF mutations represent alternative early molecular events in a subset of KIT/PDGFRA wild-type GISTs and that they are mutually exclusive to KIT/PDGFRA mutations. The clinicopathological profiles of BRAF-mutated GISTs remain to be further delineated in larger future studies.