Comparison of high resolution melting analysis, pyrosequencing, next generation sequencing and immunohistochemistry to conventional Sanger sequencing for the detection of p.V600E and non-p.V600E BRAF mutations

Abstract

Background: The approval of vemurafenib in the US 2011 and in Europe 2012 improved the therapy of not resectable or metastatic melanoma. Patients carrying a substitution of valine to glutamic acid at codon 600 (p.V600E) or a substitution of valine to leucine (p.V600K) in BRAF show complete or partial response. Therefore, the precise identification of the underlying somatic mutations is essential. Herein, we evaluate the sensitivity, specificity and feasibility of six different methods for the detection of BRAF mutations.

Methods: Samples harboring p.V600E mutations as well as rare mutations in BRAF exon 15 were compared to wildtype samples. DNA was extracted from formalin-fixed paraffin-embedded tissues by manual micro-dissection and automated extraction. BRAF mutational analysis was carried out by high resolution melting (HRM) analysis, pyrosequencing, allele specific PCR, next generation sequencing (NGS) and immunohistochemistry (IHC). All mutations were independently reassessed by Sanger sequencing. Due to the limited tumor tissue available different numbers of samples were analyzed with each method (82, 72, 60, 72, 49 and 82 respectively).

Results: There was no difference in sensitivity between the HRM analysis and Sanger sequencing (98%). All mutations down to 6.6% allele frequency could be detected with 100% specificity. In contrast, pyrosequencing detected 100% of the mutations down to 5% allele frequency but exhibited only 90% specificity. The allele specific PCR failed to detect 16.3% of the mutations eligible for therapy with vemurafenib. NGS could analyze 100% of the cases with 100% specificity but exhibited 97.5% sensitivity. IHC showed once cross-reactivity with p.V600R but was a good amendment to HRM.

Conclusion: Therefore, at present, a combination of HRM and IHC is recommended to increase sensitivity and specificity for routine diagnostic to fulfill the European requirements concerning vemurafenib therapy of melanoma patients.

Figure 1

Representative results for BRAF exon 15 mutation analysis. Sanger sequencing (A - C), high resolution melting (HRM) analysis (D - F), pyrosequencing (Pyro) (G - I) and immunohistochemistry (IHC) (J-L) are compared: The first column shows exemplarily p.V600E mutations, the second p.V600K mutations and the third column p.V600R mutations. In HRM, normalized and temperature shifted difference plots showing wildtype control in blue and mutant control in red. HRM can distinguish between p.V600E (red) and p.V600K (green) and p.V600R (light blue). Pyrosequencing was performed in the reverse direction with the sequence to analyze 5’- YAY TGT AGC TAG ACS AAA AYC ACC -3’. All three mutations can be detected. Immunohistochemistry shows a strong staining for p.V600E (J) but is negative for p.V600K (K) in representative melanoma sample. Pigmentation has to be clearly distinguished from a positive p.V600E staining (K). Cross reactivity was observed for p.V600R mutation (L) in a colorectal carcinoma sample. A: adenine, C: cytosine, G: guanine, T: thymine, V: valine, E: glutamic acid, K: lysine, R: arginine.

Figure 2

BRAF p.V600E mutation analysis of formalin-fixed paraffin-embedded tumor samples using Sanger sequencing. Electropherograms showing that Sanger sequencing is not only able to detect mutations with high mutant allele frequency (D - F) but also mutations with rather low mutant allele frequency (B and C) compared to a wildtype sample (A). Even 6.6% of BRAF p.V600E allele could be detected. A: adenine, C: cytosine, G: guanine, T: thymine, T: threonin, V: valine, K: lysine.

Figure 3

Melanoma sample showing different results of the BRAF p.V600E mutational analysis (case 30). Sanger sequencing as well as high resolution melting analysis show wildtype results (A and B, respectively). Pyrosequencing analysis resulted in a p.V600E mutation with a 5% mutant allele frequency having low relative fluorescent units of almost 30 (C). This result is confirmed by immunohistochemistry (D). Next generation sequencing resulted in a 2% mutant allele frequency with a coverage rate of 181 (E).

Figure 4

Effects of different extracts on the same sample. DNA from a melanoma sample (case 7) was extracted and showed wildtype results using Sanger, HRM and NGS (3% mutant allele frequency). In contrast, immunohistochemistry was positive. The second extract exhibited a p.V600E mutation with all evaluated methods. Sanger: Sanger sequencing, HRM: high resolution melting analysis, NGS: next generation sequencing, IHC: immunohistochemistry.