Frequent BRAF Gain in Low‐Grade Diffuse Gliomas with 1p/19q Loss

Abstract

Chromosomal 7q34 duplication and BRAF‐KIAA1549 fusion is a characteristic genetic alteration in pilocytic astrocytomas. 7q34 gain appears to be common in diffuse astrocytomas, but its significance is unclear. We assessed BRAF gain and BRAF mutations in 123 low‐grade diffuse gliomas, including 55 diffuse astrocytomas, 18 oligoastrocytomas and 50 oligodendrogliomas. Quantitative polymerase chain reaction (PCR) revealed BRAF gain in 17/50 (34%) oligodendrogliomas, a significantly higher frequency than in diffuse astrocytomas (7/55; 13%; P = 0.0112). BRAF gain was common in low‐grade diffuse gliomas with 1p/19q loss (39%) and those lacking any of the genetic alterations analyzed (31%), but was rare in those with TP53 mutations (2%). Logistic regression analysis showed a significant positive association between 1p/19q loss and BRAF gain (P = 0.0032) and a significant negative association between TP53 mutations and BRAF gain (P = 0.0042). Fluorescence in situ hybridization (FISH) analysis of 26 low‐grade diffuse gliomas with BRAF gain additionally revealed BRAF‐KIAA1549 fusion in one oligodendroglioma. Sequencing of cDNA in 17 low‐grade diffuse gliomas showed BRAF‐KIAA1549 fusion in another oligodendroglioma. A BRAF^V600E mutation was also detected in one oligodendroglioma, and a BRAF ^A598V in one diffuse astrocytoma. These results suggest that low‐grade diffuse gliomas with 1p/19q loss have frequent BRAF gains, and a small fraction of oligodendrogliomas may show BRAF‐KIAA1549 fusion.

INTRODUCTION

Pilocytic astrocytoma [World Health Organization (WHO) grade I], a relatively circumscribed, slowly growing, often cystic astrocytoma occurring in children and young adults (22), is genetically characterized by frequent (>60%) fusion of the BRAF and KIAA1549 genes, which are closely associated with duplication of the BRAF gene at 7q34 14, 18. BRAF^V600E mutations were also reported in a small fraction of pilocytic astrocytomas (up to 7%) 14, 29, 30.

BRAF‐KIAA1549 fusion has not been detected in any of 50 diffuse astrocytomas WHO grade II by fluorescence in situ hybridization (FISH) (18), 11 diffuse astrocytomas by reverse‐transcriptase polymerase chain reaction (RT‐PCR) (29) or 3 diffuse astrocytomas by single‐nucleotide polymorphism (SNP) array (20). However, gain of 7q34 without evidence for BRAF‐KIAA1549 fusion appears to be common in diffuse astrocytomas, although frequencies vary significantly among different studies 12, 13, 18, 25, 31. Korshunov et al (18) showed gain of 7q34 in 31 of 50 (62%) diffuse astrocytomas by FISH. In array comparative genome hybridization (CGH) analyses, Pfister et al (25) reported 7q34 duplication in 2 of 13 (15%) diffuse astrocytomas, and Jeuken et al. (13) showed gain at the BRAF locus at 7q34 in 4/9 (44%) diffuse astrocytomas. Sievert et al (31) found 7q34 duplication in three of six pediatric fibrillary astrocytomas by FISH, whereas Jacob et al (12) reported the lack of 7q34 duplication in 27 diffuse astrocytomas by SNP array and quantitative PCR.

It has been reported that gliomas with BRAF gain showed significantly increased levels of BRAF mRNA compared with tumors without gain (25). Furthermore, silencing of BRAF or pharmacological inhibition of its downstream phosphorylation targets suppressed proliferation of low‐grade glioma cells (25). These findings suggest that activation of the mitogen‐activated protein kinase (MAPK) pathway due to BRAF gain may play a role in the pathogenesis of a fraction of low‐grade diffuse gliomas.

In the present study, to provide further information on the frequencies of BRAF alterations in low‐grade diffuse gliomas and to correlate these with other common genetic alterations, we assessed BRAF gain, BRAF‐KIAA1549 fusion, and BRAF mutations in low‐grade diffuse gliomas with different histology (diffuse astrocytomas, oligoastrocytomas and oligodendrogliomas) and genetic features (IDH1/2 mutations, TP53 mutations and 1p/19q loss).

MATERIALS AND METHODS

Tumor samples

A total of 123 low‐grade diffuse gliomas of WHO grade II (109 tumors in patients older than 20 years, and 14 cases in those younger than 20 years) were obtained from the Department of Neuropathology, University Hospital Zurich, Switzerland; the Department of Neuropathology, University Hospital Frankfurt, Germany; the Departments of Neuropathology and Neurosurgery, University Hospital Essen, Germany; the Department of Pathology, Gunma University, Japan; the Institute of Neuropathology and Department of Neurosurgery, University Hospital Munster, Germany; the Institute of Neuroscience, Bordeaux, France; and the Department of Neurosurgery, University Hospital Bern, Switzerland.

Histologically, these tumors were classified as diffuse astrocytoma (55 cases), oligoastrocytoma (18 cases) and oligodendroglioma (50 cases). Genetic alterations in these tumors have been published previously (16). Thirty‐four cases had IDH1/2 plus TP53 mutations, 27 cases had IDH1/2 mutation plus 1p/19q loss, 12 cases showed IDH1/2 mutation only, 7 cases had TP53 mutations only and 14 cases had 1p/19q loss only. Twenty‐nine cases lacked any of these changes (IDH1/2 mutations, TP53 mutations and 1p/19q loss).

DNA extraction

DNA was extracted from typical tumor areas that were manually scraped off from formalin‐fixed, paraffin‐embedded (FFPE) tissue sections as previously described (16). DNA concentration was determined by spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA). Absorption was measured at 230, 260 and 280 nm and DNA quality was evaluated by A₂₆₀/A₂₃₀ and A₂₆₀/A₂₈₀ ratios.

BRAF gain

BRAF gain was assessed by quantitative PCR, using three reference sequences at different chromosomal locations (CF at 7q31.2, β‐globin at 11q15.5 and GAPDH at 12p13.31) 24, 26, 33, 34. Primer sequences were as follows: 5′‐TTC ATG AAG ACC TCA CAG TAA AAA‐3′ (sense) and 5′‐CCA CAA AAT GGA TCC AGA CA‐3′ (antisense) for BRAF (PCR product, 107 bp), 5′‐GGC ACC ATT AAA GAA AAT ATC ATC TT‐3′ (sense) and 5′‐GTT GGC ATG CTT TGA TGA CGC TTC‐3′ (antisense) for the CF (PCR product, 79 bp), 5′‐GTG CAT CTG ACT CCT GAG GAG A‐3′ (sense) and 5′‐CCT TGA TAC CAA CCT GCC CAG‐3′ (antisense) for the β‐globin (PCR product, 102 bp), and 5′‐TCA AGA AGG TGG TGA AGC AG‐3′ (sense) and 5′‐TGT CGC TGT TGA AGT CAG AG‐3′ (antisense) for the GAPDH (PCR product, 96 bp). Quantitative PCR was carried out in a total volume of 20 µL with 10 µL of iQ^TM SYBR green (Bio‐Rad, Hercules, CA, USA), 6.4 µL of primers (1.25 µmol/L of each primer) and approximately 20 ng of DNA with initial denaturation at 95°C for 12 minutes followed by 40 cycles of denaturation at 95°C for 20 s, annealing at 55°C for 20 s and extension at 72°C for 45 s. PCR was performed in triplicate on a 96‐well optical plate with an iCycler iQ5 Detection System (Bio‐Rad). The copy‐number calculation was carried out using the comparative Ct (threshold cycle) method, as described previously 3, 24. Results using quantitative PCR with three different references were concordant in >92% of cases. Tumors were considered to have BRAF gain when PCR reactions using two or three references showed significant copy‐number gain of BRAF.

BRAF mutation

The mutational hotspot codons of BRAF were amplified by PCR. Primer sequences were as follows: 5′‐TGC TTG CTC TGA TAG GAA AAT G‐3′ (sense) and 5′‐CCA CAA AAT GGA TCC AGA CA‐3′ (antisense) (PCR product, 173 bp) (30). PCR amplification products were subjected to the direct sequencing on ABI 3100 PRISM DNA sequencer (Applied Biosystems, Foster City, CA, USA) with the Big Dye Terminator cycle sequencing kit (ABI PRISM, Applied Biosystems).

BRAF‐KIAA1549 fusion by FISH

Tumors showing BRAF gain according to quantitative PCR were further screened for the BRAF‐KIAA1549 fusion gene by FISH using previously published probes and methods with minor modifications (18). Two‐color interphase FISH analysis was performed on 5‐micron thick paraffin tissue sections pairing two home brew locus‐specific probes: FITC‐labeled locus‐specific probe RP11‐355D18 (CHORI BACPAC Resources Center, Oakland, CA, USA) corresponding to fluorescein isothiocyanate (FITC) labeled KIAA1549 (green) and rhodamine‐labeled locus‐specific probe 726N20 corresponding to BRAF (red). Pretreatment of slides, hybridization, posthybridization processing and signal detection were performed as reported elsewhere (25). Samples showing sufficient FISH efficiency (>90% nuclei with signals) were evaluated, and signals were scored in at least 100 nonoverlapping, intact nuclei. Non‐neoplastic brain biopsy specimens were used as controls. Chromosomal gains at 7q34 region were defined as >5% of nuclei containing three or more signals for both locus‐specific probes. The BRAF‐KIAA1549 fusion gene was scored in cases showing 7q34 gain in combination with overlap of at least one red signal and one green signal, resulting in a yellow signal. Because these two probes are normally in close proximity, signals were designated as fused only when the red and green signals were completely or nearly completely overlapping. Based on the median number of fusion signals encountered in control specimens plus three standard deviations, we scored cases as positive for BRAF‐KIAA1549 fusion when >25% of cells had both yellow fusion signals and associated copy‐number gains (at least three green and/or red signals). Additionally, in order to distinguish polysomy 7 from a more specific gain of the BRAF region, a second FISH analysis was performed pairing the BRAF probe with a commercial SpectrumGreen labeled centromere enumerating probe (CEP7; Abbott Laboratories, Abbott Park, IL, USA). Copy‐number gains associated with an overall BRAF to CEP7 ratio >1.15 were considered BRAF specific gains, while the remaining cases were classified as polysomy 7.

BRAF‐KIAA1549 fusion by sequencing

RNA was extracted from paraffin sections of 17 cases for which sufficient materials were available (eight diffuse astrocytomas, nine oligodendrogliomas). For preparation of RNA extraction from FFPE samples, the RNA RNeasy FFPE kit (QIAGEN GmbH, Hilden, Germany) was used according to the manufacturers' recommendations. cDNA was constructed with Superscript® II RT (Invitrogen, Carlsbad, CA, USA). Primer sequences were as follows: 5′‐GCG ATG GCA CCT ACA GGA‐3′ (sense) for KIAA1549 exon 15, 5′‐CAG TGG GGG TCC TTC TAC AG‐3′ (sense) for KIAA1549 exon 16, 5′‐TGC CAG AGG GAT CTA CTC G‐3′ (sense) for KIAA1549 exon 18 and 5′‐CCT TCG TAC GGG GAG GAC‐3′ (sense) for KIAA1549 exon 19, and 5′‐CCA CGA AAT CCT TGG TCT CT‐3′ (antisense) for BRAF exon 9, 5′‐GGG GGT AGC AGA CAA ACC T‐3′ (antisense) for BRAF exon 10 and 5′‐TCA CTC GAG TCC CGT CTA CC‐3′ (antisense) for BRAF exon 11. The sizes of the PCR products were 88 bp for KIAA1549 exon 15–BRAF exon 9, 80 bp for KIAA1549 exon 16–BRAF exon 9, 80 bp for KIAA1549 exon 16–BRAF exon 11, 97 bp for KIAA1549 exon 18–BRAF exon 10 and 70 bp for KIAA1549 exon 19–BRAF exon 9.

RT‐PCR was performed with 40 cycles of denaturation for 50 s at 94°C, annealing for 45 s at 58°C and extension for 50 s at 72°C. PCR products were visualized by 8% acrylamide‐gel electrophoresis, staining with gel red. RT‐PCR products were subjected to direct sequencing on an ABI PRISM®3100 DNA sequencer (Applied Biosystems) with the BigDye® Terminator Cycle Sequencing kit (ABI PRISM, Applied Biosystems).

Statistical analyses

The χ² test or the Fisher's exact test was conducted to analyze the significance of the association of age, histology or genetic features with BRAF gain. Logistic regression analysis was carried out to assess associations between different genetic alterations. Statistical analysis was performed with StatView® for Windows 5.01 software (SAS Institute Inc., Cary, NC, USA).

RESULTS

BRAF gain

Quantitative PCR revealed BRAF gain in a total of 28 of 123 (23%) low‐grade diffuse gliomas (Figure 1A). BRAF gain was significantly more frequent in oligodendrogliomas than in diffuse astrocytomas (34% vs. 13%; P = 0.0112; Table 1). BRAF gain was common in low‐grade diffuse gliomas with 1p/19q loss (16/41; 39%) and in those lacking any of the common genetic alterations (9/29; 31%). In contrast, only 1 of 41 (2%) low‐grade diffuse gliomas with TP53 mutation showed BRAF gain (Table 1). Logistic regression analysis showed a significant positive association between 1p/19q loss and BRAF gain [OD = 3.733 (1.553–8.973); P = 0.0032], and a significant negative association between TP53 mutations and BRAF gain [OD = 0.051 (0.007–0.391); P = 0.0042].

Figure 1.

A. Quantitative PCR showing BRAF gain in low‐grade gliomas. Note that normal CF (normal brain) and tumor CF show similar CTs, while BRAF in a low‐grade diffuse glioma (tumor) shows a significantly smaller CT compared with BRAF in normal brain; this indicates a BRAF gain in tumor DNA (left). Relative CT values for BRAF and CF at different concentrations of normal control DNA are shown. The slopes of the curves are similar, suggesting equal efficiencies of the two PCR reactions at CT (right). B. FISH analysis of BRAF gain and BRAF‐KIAA1549 fusion. Most cases demonstrated gain of BRAF and KIAA‐1549 without any evidence of fusion (left). In one oligodendroglioma with 1p/19q co‐deletion, FISH demonstrated fusion of BRAF (red) and KIAA1549 (green) signals, resulting in a yellow signal for the fusion gene (arrow) in addition to copy‐number gains (middle). FISH in another case demonstrates selective gain of the red BRAF in comparison with green CEP7 signals (right). C. RT‐PCR showing KIAA1549 exon 16–BRAF exon 9 fusion in an oligodendroglioma (left). MS, molecular size marker; T1, fusion‐negative tumor; T2, fusion‐positive oligodendroglioma. Sequencing confirming a KIAA1549 exon 16–BRAF exon 9 fusion (right). D. BRAF ^V600E mutation in an oligodendroglioma (left), and BRAF ^A598V mutation in a diffuse astrocytoma (right). CT = cycle threshold; PCR = polymerase chain reaction; RT‐PCR = reverse‐transcriptase PCR.

Table 1.

BRAF gain in low‐grade gliomas.

	No. of cases with BRAF gain
Histology
Diffuse astrocytoma (n = 55)	7 (13%)*
Oligoastrocytoma (n = 18)	4 (22%)
Oligodendroglioma (n = 50)	17 (34%)*
Genetic alterations
TP53 mutation ± IDH1/2 mutation (n = 41)	1 (2%)†‡
1p/19q loss ± IDH1/2 mutation (n = 41)	16 (39%)†
IDH1/2 mutation only (n = 12)	2 (17%)
No alteration§ (n = 29)	9 (31%)‡

*P = 0.0112; †P = 0.0001; ‡P = 0.0011.

^§ Absence of IDH1/2 mutations, TP53 mutations, 1p/19q loss.

Patients with low‐grade diffuse glioma with BRAF gain tended to be younger, but the age difference was significant only among pediatric patients (<20 years; Table 2). The median survival of patients was not significantly different between cases with and without BRAF gain (data not shown).

Table 2.

Age and BRAF gain in low‐grade gliomas.

	No. of cases (%)	Mean age ± SD (years)	P‐value
All patients (n = 116)
With BRAF gain	28 (24)	34.3 ± 18.2	0.0802
Without BRAF gain	88 (76)	40.4 ± 15.2
Pediatric patients (<20 years; n = 14)
With BRAF gain	5 (35)	4.8 ± 1.5	0.0038 *
Without BRAF gain	9 (65)	13.3 ± 5.1
Adult patients (≥20 years; n = 102)
With BRAF gain	23 (22)	40.7 ± 12.8	0.3570
Without BRAF gain	79 (77)	43.5 ± 12.7

^* Statistically significant.

BRAF‐KIAA1549 fusion

FISH analyses were carried out in 26 of 28 cases in which BRAF gain was detected by quantitative PCR. In all cases, BRAF gain was confirmed by FISH (Figure 1B left). Furthermore, FISH detected a BRAF–KIAA1549 fusion in one oligodendroglioma (male, aged 40 years; frontal right location; with 1p/19q loss and BRAF gain but no IDH1/2 mutation) with 65% of cells showing BRAF‐KIAA1549 fusion signals in addition to copy‐number gains (Figure 1B middle). Additionally, CEP7/BRAF FISH studies demonstrated that at least six of the positive cases were associated with specific gains of the BRAF region, rather than polysomy 7 (Figure 1B right). Sequencing of cDNA in 17 low‐grade diffuse gliomas (11 cases with BRAF gain and 6 cases without gain) showed BRAF‐KIAA1549 gene fusion in one oligodendroglioma (female aged 22 years; thalamus; with BRAF gain but no TP53 mutations, no 1p/19q loss and no IDH1/2 mutations) (Figure 1C).

BRAF ^V600E mutation

Sequencing analyses revealed a BRAF ^V600E mutation in one oligodendroglioma that lacked BRAF gain. This tumor was located in the right occipital lobe and had 1p/19q loss, but lacked IDH1/2 mutation (Figure 1D). A rare BRAF ^A598V mutation was also detected in a diffuse astrocytoma that lacked BRAF gain (temporal location; with TP53 mutation plus IDH1 mutation) (Figure 1D).

DISCUSSION

The MAPK/extracellular signal‐regulated kinase (ERK) pathway regulates a wide range of biological activities, including cell differentiation, proliferation, senescence and survival 6, 7, 15, 27. This pathway consists of a small GTP protein of the RAS family that is activated in response to extracellular signaling to recruit a member of the RAF kinase family to the cell membrane (6). Mutations in the BRAF or RAS genes have been found as activating mutations in approximately 30% of all human cancers (6).

BRAF^V600E, the most common mutation in this gene, is frequent in hairy cell leukemias (100%) (32), melanomas (60%–80%) 4, 9, 23, 28 and papillary thyroid cancers (35%–70%) 8, 17. BRAF mutations are additionally associated with BRAF gene amplification in melanomas 10, 21. Lin et al (21) showed copy‐number gain at 7q34 in 65% of melanomas and activating BRAF^V600E mutations in 56% of cases, and a co‐occurrence of these two events was observed in 46% of cases, suggesting that the mutated BRAF gene may be amplified in melanomas. CGH analysis by Bastian et al (1) also showed frequent gain of the BRAF gene in 16/32 (50%) melanomas.

In pilocytic astrocytomas, the BRAF^V600E mutation is present in only a small fraction (up to 7%) 2, 14, 29, 30, but BRAF fusion genes (>60%) are the most common genetic alterations leading to abnormal activation of the MAPK/ERK pathway 14, 18. BRAF fusion in pilocytic astrocytomas is considered to occur as a result of tandem BRAF duplication at chromosome 7q34 14, 18. Several reports have suggested that gain at 7q34 (60%–80%) 13, 18, 25, 31 and gain/amplification of the BRAF gene (50%–80%) 11, 14, 18, 25 are frequent genetic alterations in pilocytic astrocytomas.

Gain or amplification of the BRAF gene has also been reported in other gliomas, including diffuse astrocytomas (15%–62%) 13, 18, 25, oligoastrocytomas (14%) (13), anaplastic oligoastrocytomas (56%) (13), anaplastic oligodendrogliomas (18%) (13) and glioblastomas (76%) (13). In the present study, we present evidence suggesting that BRAF gain is common in oligodendrogliomas (34%) and in low‐grade diffuse gliomas with 1p/19q loss (39%). In contrast, BRAF gain is infrequent in diffuse astrocytoma (13%) and very rare in low‐grade diffuse gliomas with TP53 mutations (2%). We found a significant positive association between 1p/19q loss and BRAF gain (P = 0.0032), and a significant negative association between TP53 mutations and BRAF gain (P = 0.0042). The finding of infrequent BRAF gain in diffuse astrocytomas or in low‐grade diffuse gliomas with TP53 mutations in the present study was consistent with the results of a study by Jacob et al (12), in which quantitative PCR revealed absence of 7q34 duplication in the 27 diffuse astrocytomas analyzed (12). In contrast, several previous studies using FISH with centromere probes showed frequent 7q polysomy in diffuse astrocytomas (62%–76%) 5, 18, 19, while other studies using array CGH showed gain at the BRAF locus in 15%–44% of diffuse astrocytomas 13, 25. Discrepancies of frequencies of BRAF gain in different studies may be at least in part due to variation in the specificity and sensitivity of the different methods used.

BRAF gain may be due to polysomy 7 or specific gain at the BRAF region. As we used BRAF‐specific primers for quantitative PCR, the results in the present study indicate the specific gain of the BRAF gene. Our FISH analysis using the CEP7/BRAF probe demonstrated that at least 6 of 26 cases with BRAF gain were associated with specific gain of the BRAF region, rather than polysomy 7. Irrespective of the mechanisms involved, BRAF gain itself appears to have significant biological implications. In thyroid tumors, BRAF gain detected by FISH analysis was associated with higher levels of BRAF protein as detected by Western blot; BRAF gain and RAS mutations were mutually exclusive (6).

The present study also shows that, although rare, BRAF fusion genes and BRAF mutations may be present in low‐grade diffuse gliomas. FISH analysis showed the BRAF‐KIAA1549 gene fusion to be present in one oligodendroglioma. In addition, RT‐PCR analysis revealed KIAA1549 exon 16–BRAF exon 9 fusion in another oligodendroglioma with BRAF gain. We also found BRAF ^V600E mutation in an oligodendroglioma, and a rare BRAF ^A598V mutation in a diffuse astrocytoma. This is consistent with the results of a previous study on 162 low‐grade diffuse gliomas, which showed that only one oligodendroglioma had a BRAF^V600E mutation (30).

It has been shown that in pilocytic astrocytomas, younger patients more frequently show 7q34 duplication or BRAF rearrangement 11, 25, 31. In the present study, we assessed the relationship between age and BRAF gain in low‐grade diffuse gliomas. In adults, there was no significant correlation between age and BRAF gain, whereas among pediatric cases (aged <20 years), BRAF gain was associated with the youngest patients (mean 4.8 vs. 13.3 years; P = 0.0038) (Table 2).

In summary, this study suggests that BRAF gain is common in oligodendrogliomas and in low‐grade diffuse gliomas with 1p/19q loss, suggesting that activation of the MAPK signaling pathway may be involved in their pathogenesis. As with pilocytic astrocytomas, this finding raises potential therapeutic implications.