Detecting BRAF Mutations in Formalin-Fixed Melanoma: Experiences with Two State-of-the-Art Techniques.

BACKGROUND: Melanoma is characterized by a high frequency of BRAF mutations. It is unknown if the BRAF mutation status has any predictive value for therapeutic approaches such as angiogenesis inhibition. PATIENTS AND METHODS: We used 2 methods to analyze the BRAF mutation status in 52 of 62 melanoma patients. Method 1 (mutation-specific real-time PCR) specifically detects the most frequent BRAF mutations, V600E and V600K. Method 2 (denaturing gel gradient electrophoresis and direct sequencing) identifies any mutations affecting exons 11 and 15.
RESULTS: Eighteen BRAF mutations and 15 wild-type mutations were identified with both methods. One tumor had a double mutation (GAA) in codon 600. Results of 3 samples were discrepant. Additional mutations (V600M, K601E) were detected using method 2. Sixteen DNA samples were analyzable with either method 1 or method 2. There was a significant association between BRAF V600E mutation and survival.
CONCLUSION: Standardized tissue fixation protocols are needed to optimize BRAF mutation analysis in melanoma. For melanoma treatment decisions, the availability of a fast and reliable BRAF V600E screening method may be sufficient. If other BRAF mutations in exons 11 and 15 are found to be of predictive value, a combination of the 2 methods would be useful.

Introduction

BRAF gene aberrations became prominent in melanoma in 2002, when Davies et al. [1] published the results of a first genome-wide analysis of melanoma tissue. This study revealed a high frequency of BRAF mutations in melanoma patients occurring mainly at codon 600. The V600E mutation occurs in 40–60% and the V600K exchange is observed in up to 20% of BRAF mutations [2, 3, 4].

Currently, promising therapeutic agents such as vemurafenib (RG 7204; Roche, Basel, Switzerland) are available that specifically target the BRAF V600E mutation [5, 6, 7, 8, 9, 10, 11] while also exhibiting positive effects in melanoma patients with the V600K variant [7, 12].

Several studies have proven that BRAF mutations at V600 lead to reduced overall survival (OS) unless treated with a BRAF targeting drug [3, 13]. Therefore, knowing the BRAF mutation status in melanoma patients is clinically important. Various detection methods have been applied to detect BRAF V600 mutations [1, 3, 5, 13, 14, 15, 16, 17, 18, 19, 20, 21].

Here, we describe 2 different BRAF mutation detection methods that were utilized in a phase II melanoma trial [13] with regard to their advantages, reliability and limitations. The impact of a BRAF mutation on patient outcome was also analyzed.

Patients and Methods

Patient Selection

A total of 62 patients were treated with temozolomide and bevacizumab in this phase II trial; SAKK 50/07. All patients were considered for BRAF genotyping. All patients had histologically confirmed melanoma and unresectable stage IV disease. Patient demographics have been described previously [13]. The trial was approved by the local ethics review boards as well as by Swissmedic and was registered at the National Institute of Health (www.clinicaltrial.gov; identifier number: NCT00568048). All patients gave written informed consent before any trial procedure.

Material Preparation and Selection

HE-stained sections of 62 primary melanomas that were fixed in formalin and embedded in paraffin (FFPE) were reviewed by 2 pathologists (D.M.-P., H.M.). The tumor region of interest was marked on each glass slide. DNA was extracted using the DNeasy Blood & Tissue Kit (Qiagen, Basel, Switzerland) and tested for BRAF mutations in 2 independent laboratories.

BRAF Mutation Analysis from Melanoma Tumor Tissues

Two methods were used for the analysis of BRAF mutations. Method 1 was performed according to the method described by Benlloch et al. [14]. This mutation-specific real-time PCR enables the specific detection of the GTG to GAG transversion in codon 600 of BRAF, which causes an amino acid exchange from valine to glutamic acid. The sequences of the exon 15-specific primers and the MGB-TaqMan probes are shown in table . Depending on the genotype, distinct plots for the mutant and wild-type (WT) variants were generated. The threshold between WT and mutated genotypes was adapted to a heterozygous BRAF V600E mutated melanoma reference sequence. The reference cell culture was established from 1 melanoma patient. Final evaluation was based on allelic discrimination (fig. ).

In addition to BRAF V600E, other less frequent BRAF mutations in exons 11 and 15 have been described for melanoma [1]. Method 2 was used to find such mutations. This method is based on a denaturing gel gradient electrophoresis (DGGE) of amplified PCR products of exons 11 and 15, and direct DNA sequencing as previously described [22]. PCR products showing reduced or increased mobility shifts compared to WT BRAF were subjected to direct sequencing (fig. ).

Statistical Analysis

Endpoints considered for the statistical analysis were disease stabilization (either complete response, partial response or stable disease) rate at 12 weeks, best overall response, and 3 time-to-event type endpoints: duration of response stabilization, progression free survival (PFS) and OS. PFS was defined as the time from trial registration until either disease progression or death, with patients censored at the time of starting second-line therapy or the last time they were known to be alive without progression.

The BRAF mutation status identified by the 2 methods was used as a descriptive secondary endpoint: 3 types of BRAF status groups were considered based on either BRAF method 1, BRAF method 2 or combined BRAF methods 1 and 2. All endpoints were evaluated by BRAF status by means of the log-rank test for survival-type endpoints and χ2 or Spearman's rank correlation coefficient allowing for the ordered outcome of tumor response. No adjustment for multiple comparisons was performed. The data was analyzed in SAS (Statistical Analysis Systems, version 9.2).

Results

Detection of BRAF Mutation Status by Methods 1 and 2

DNA could be extracted from 52 of 62 (84%) FFPE tumor samples. Due to insufficient or no tumor cells, 10 cases could not be analyzed. The BRAF mutation status was analyzable in 36 DNA samples with both methods. Eighteen BRAF V600E mutations (50%) and 15 WT mutations (42%) were identified with both methods. The results of 3 samples (8%) were discrepant between the methods. Additional mutations (V600M, K601E) were obtained by using method 2. Notably, 1 of the V600E melanoma results had a GTG to GAA double point mutation. Eight DNA samples were analyzable with method 1, and 8 with method 2. No mutations were seen in BRAF exon 11. The results of the 2 BRAF mutation analyses are shown in table . Examples of mutations are shown in fig. .

BRAF Mutation Status and Tumor Response

No statistically significant association was seen between BRAF mutation status obtained by method 1, 2 or both and overall best tumor response with the use of angiogenesis inhibition in combination with an alkylating chemotherapy.

BRAF Mutation Status and Time-to-Event Endpoints

Among patients with a tumor response and a known BRAF mutation status, there was no statistically significant difference in response duration between BRAF status groupings. No statistically significant difference in median PFS was observed between BRAF status using BRAF method 1, 2 or both.

The median OS was 9.6 months (95% CI: 8.0, 11.9). When stratified by method 1, BRAF mutation status showed a statistically significant difference (p = 0.0137). Using method 2, BRAF mutation status was borderline statistically significant (p = 0.0546) and when combining methods 1 and 2, BRAF mutation statuses also resulted in statistical significance (p = 0.0102). The data are shown in table .

Discussion

Advantages and Limitations of the Used Methods

Method 1 allows the specific detection of the BRAF hotspot mutation in melanoma at codon 600. In addition to GAG (V600E) and AAG (V600K) mutations, the DNA sequencing results obtained with method 2 indicate that the TaqMan probe is obviously also able to bind a third variant, namely GAA which also codes glutamic acid. This double mutation at codon 600 is the third most frequent BRAF sequence alteration and occurs in about 2% of malignant melanomas [4].

The significant advantage of method 1 lies in the TaqMan approach, which allows the rapid identification of BRAF V600E without any additional analysis such as the time-consuming and costly DGGE and DNA sequencing. In contrast, method 2 allows the detection of any possible mutations occurring within exon 15, including other V600 variants. If any other BRAF mutations besides that at codon 600 also become therapeutically relevant for melanoma patients, the TaqMan approach would not be sufficient. However, the occurrence of other mutations in melanoma, such as K601E or V600M, is quite low and the long-term relevance has not yet been documented.

Frequency of V600 Mutations

The frequency of the BRAF V600E mutation of 50% is comparable to that described in the literature. Recent results show that up to 20% of all BRAF mutations lead to a V600K exchange, which is also targetable [23, 24]. We were unable to find this alteration in our melanoma samples. The lack of finding V600K mutations in our patient cohort may be explained due to the relatively low number of analyzable DNA samples. However, our data are in line with the results of other studies in which much lower frequencies of the V600K mutation were found [23, 24]. Based on these data, it is also possible that the frequency of the V600K mutation is overestimated.

Tumor Heterogeneity and Protocol Standardization

An important drawback is the fact that the tumor tissue is often infiltrated and surrounded by non-tumorous cells. Depending on the level of contamination, a BRAF mutation may be overlooked, thus producing a false-negative (WT) result. We aimed to circumvent this problem by punching exclusively into areas of the tumor tissue that contained more than 80% tumor cells.

The necessity of standardized protocols for fixation, dehydration and paraffin embedding of tissues is a critical issue, particularly if the specimens to be analyzed derive from different pathology institutes. A reliable BRAF mutation analysis was problematic or even impossible for about one-third of the tissue samples, suggesting more or less differing protocols used at the different pathology centers involved in this phase II trial. Standardized and optimal operating procedures are of utmost importance for reliably performing BRAF mutation analysis in all tissue samples to guarantee the best benefit for melanoma patients.

Outcome Data

In our trial, statistical differences were between BRAF WT and mutant patients regarding OS. Whether the survival difference between BRAF WT and mutant patients is only due to the fact that BRAF mutational status has prognostic value or if BRAF mutational status may also be predictive for a combination treatment with bevacizumab and temozolomide cannot be concluded definitively from our data set, but should be prospectively studied. If the 2 most frequent BRAF mutations, V600E and V600K, are the alterations with predictive value in melanoma patients, techniques identical or similar to method 1 described in this paper may be the detection system of choice.

Disclosure Statement

Study design, data collection, analysis and interpretation of the data, final approval of the manuscript and the decision to submit the paper for publication were the responsibility of the authors and not the sponsors. R. von Moos: Consultancy/advisory role at Roche, Bristol-Myers-Squibb and Essex AG, and speaking honoraria at Roche. R. Dummer: Had consultancy/advisory role and research funding at Astra-Zeneca and Novartis. S. Gillessen: Advisory role at Novartis. H. Moch: Currently conducting research sponsored by Roche and Novartis. All other authors: no conflict of interest to declare.

Table 1

Primer sequences and MGB-TaqMan probes for BRAF exon 15 applied in method 1

Primer	Sequence
BRAF exon 15 forward	5′-CTACTGTTTTCCTTTACTTACTACACCTCAGA-3′
BRAF exon 15 reverse	5′-ATCCAGACAACTGTTCAAACTGATG-3′
BRAF WT probe	5′-VIC-CTAGCTACAGtGAAATC-MGB-3′
BRAF V600E probe	5′-FAM-TAGCTACAGaGAAATC-MGB-3′

BRAF WT (GtG) and mutated codon 600 (GaG) are indicated in bold lowercase letters.

Table 2

Patient BRAF characteristics: a comparison between methods 1 and 2

BRAF analysis results		Frequency, % (n = 52)
method 1	method 2
V600E	V600E*	18 (34.62)
WT	WT	15 (28.85)
WT	NA	5 (9.52)
NA	WT	6 (11.54)
NA	V600E	1 (1.92)
NA	V600M	1 (1.92)
V600E	WT	1 (1.92)
WT	V600E	1 (1.92)
WT	K601E	1 (1.92)
V600E	NA	3 (5.77)

NA = Not analyzable (no/too little DNA, no PCR product).

One tumor with GAA mutation. Discrepant results are in bold.

Table 3

BRAF mutation status and patient outcome

Analysis	Method	BRAF status [median event-free survival time, months (95% CI)]		Log-rank p value
		BRAF mutation	BRAF WT
PFS	1	4.0 (1.6, 5.4)	5.3 (2.6, 8.1)	0.0986
	2	4.0 (1.2, 5.5)	5.4 (2.6, 8.1)	0.0571
	1 + 2	4.0 (1.7, 6.6)	5.5 (1.3, 10.9)	0.0647
OS	1	9.2 (6.5, 11.9)	12.0 (7.4, 16.4)	0.0137
	2	9.6 (6.6, 11.9)	12.6 (8.4, 15.6)	0.0546
	1 + 2	9.2 (6.6,11.8)	12.6 (7.2, 16.4)	0.0102