To the editor: We read with great interest the recently published article by Mesteri et al1 regarding immunohistochemical detection of BRAF mutations in serrated colonic polyps. Some types of serrated polyps, particularly sessile serrated polyps (sessile serrated adenomas), frequently harbor BRAF c.1799T>A (p.V600E) mutations and presumably represent precursor lesions to sporadic colonic carcinomas with a high degree of microsatellite instability (MSI-H). Detection of BRAF mutations helps distinguish sporadic MSI-H tumors from those associated with Lynch syndrome since Lynch-related adenocarcinomas virtually never harbor BRAF mutations. BRAF status combined with testing for microsatellite instability also provides important prognostic information. Recent evidence indicates that MSI-H tumors with wild-type BRAF have a better prognosis than those associated with BRAF mutations. Patients with microsatellite stable (MSS) tumors and mutations in BRAF have lower 5-year survival than those with MSI-H tumors and those with BRAF wild-type MSS tumors.2 (For a complete discussion of the molecular carcinogenesis in colorectal carcinoma, the reader is referred to a recent review by Colussi et al.3) Evaluation of BRAF is currently limited to molecular techniques, such as PCR and sequencing assays, which often require specialized laboratories and are not widely available to practicing pathologists. Thus, simple and inexpensive methods to evaluate BRAF mutational status are of substantial clinical interest. The potential clinical utility of detecting BRAF mutations in sessile serrated polyps is less clear. Although up to 80% of sessile serrated polyps and microvesicular hyperplastic polyps contain mutated BRAF, most of these do not progress to adenocarcinoma; thus, BRAF mutations are neither diagnostic of sessile serrated polyps, nor do they reliably identify polyps that are at risk for malignant progression.4, 5

In this study, Mesteri et al used the BRAF p.V600E mutation-specific antibody, VE1, to subclassify serrated polyps according to BRAF mutational status. This antibody was originally developed in the laboratory of one of the co-authors, Dr Andreas von Deimling, and is now commercially available. The authors correlated their immunohistochemical findings with the results of sequencing assays in a subset of these polyps. They reported positive VE1 staining in 100% of sessile serrated polyps and the majority of traditional serrated adenomas and microvesicular hyperplastic polyps, as well as perfect correlation between immunohistochemical and sequencing assays. The authors concluded that immunohistochemistry is a feasible means of detecting BRAF mutations in serrated polyps in routine practice. The immunostaining results depicted by Mesteri et al are convincing and strongly support their claim. However, we would like to compare their experience with the VE1 antibody with that of our group and point out major obstacles to the routine use of this antibody that were not addressed by these authors.

We recently evaluated BRAF mutational status by sequencing and immunohistochemistry in sessile serrated polyps, using the commercially available VE1 antibody from Spring Bioscience (Pleasanton, CA, USA). We first tested seven sessile serrated polyps with BRAF c.1799T>A (p.V600E) mutations, including two with cytologic dysplasia, as well as one sessile serrated polyp that was BRAF wild-type by sequencing. Cytoplasmic staining was observed in only one of seven mutation-positive cases when using the VE1 antibody at 1:50 dilution, as recommended by the manufacturer, and the staining reaction was focal and weak. All eight cases also showed moderately intense nuclear staining in both lesional and non-lesional colonic epithelial cells, thereby complicating interpretation. We then carried out additional experiments using various antigen retrieval methods and concentrations in order to optimize the antibody, but these measures failed to improve stain sensitivity. Overall, we found that VE1 immunostaining was concordant with molecular analyses in only 3 (7%) of 43 BRAF-mutated serrated polyps.6 Thus, in our hands, the results reported by Mesteri et al are not reproducible when the commercially available VE1 antibody is used.

Previous studies evaluating this antibody in invasive colonic adenocarcinomas have also produced mixed results. Affolter et al7 reported 100% concordance between immunohistochemical VE1 stains and pyrosequencing of BRAF V600E in colorectal adenocarcinomas, but noted that staining was either weak or heterogeneous in 43% of BRAF-mutated tumors. Sinicrope et al8 also reported complete concordance between VE1 staining and BRAF mutational status, although weak staining of 10–30% of tumor cells was considered a ‘positive’ result in 22% of BRAF-mutated cancers. Other authors have observed false-positive and false-negative results with the VE1 antibody. Adackapara et al9 reported weak staining in 26% of BRAF wild-type colorectal carcinomas, whereas 29% of BRAF-mutated tumors were negative and 35% showed weak staining. Finally, Kuan et al10 observed VE1 staining in 54 of 57 (95%) invasive adenocarcinomas with BRAF mutations, but also found that three of four tumors with weak cytoplasmic staining proved to have wild-type BRAF by molecular analysis.

The critical difference between our immunohistochemical staining methods and those of Mesteri et al reflects the nature of the antibody utilized. Mesteri et al used undiluted VE1 hybridoma supernatant for their experiments, which is available only to Dr Andreas von Deimling and collaborators. In contrast, the commercially available VE1 antibody (Spring Bioscience) that we and others have used was produced by protein A/G purification of the VE1 hybridoma antibody and reconstituted in Tris-HCl pH 7.5 with carrier proteins. The product specification sheet does not indicate whether the antibody used for purification was obtained through hybridoma supernatants or in ascites form, and the antibody concentration of this mouse IgG2a antibody is also not specified.

We suspect that the difference between the original research antibody and the commercial antibody from Spring Biosciences is the fundamental reason why our results are so different from those of Mesteri et al. We conclude that routine use of the currently available VE1 antibody for detection of BRAF p.V600E in serrated colonic polyps and invasive colonic adenocarcinomas is premature for several reasons. First, most serrated polyps that harbor BRAF mutations do not show VE1 immunohistochemical expression. Second, others have reported that occasional invasive colonic adenocarcinomas with mutated BRAF are negative for VE1, whereas weak or focal VE1 staining occurs in both BRAF wild-type and BRAF-mutated tumors. Last, but not least, the VE1 antibody is a costly immunohistochemical stain ($2000 per 0.5 ml), which mitigates its potential advantage over molecular analyses. We are aware that Ventana Medical Systems (Tucson, AZ, USA) now markets a VE1 antibody as well. It may be worthwhile to evaluate the performance of this new antibody in BRAF-mutated carcinomas, including colonic adenocarcinomas, and in serrated colonic polyps.