LOH analysis should not be used as a tool to assess whether UVs of BRCA1/2 are pathogenic or not

Abstract

I have read with interest the article by Dworkin et al. [1] entitled ‘‘Methylation not a frequent ‘second hit’ in tumors with germline BRCA mutation’’ published in your journal. This article states that loss of the wild type allele is common in tumors in inherited germline mutation carriers. In fact, these authors used this criterion to assess whether an unclassified variant was pathological or neutral, as previously reported by other authors (Osorio et al. [2]; Chenevix-Trench et al. [3]; Osorio et al. [4]). They thus analyzed loss of heterozygosity (LOH) in 35 tumors from carriers of a deleterious mutation in BRCA1 and 19 carriers of a deleterious mutation in BRCA2, and found loss of the wild type allele in 82% of BRCA1 carriers and in 61% of BRCA2 carriers. Our laboratory has a wide experience in molecular characterization of BRCA1 and BRCA2 in patients with hereditary breast and ovarian cancer (Beristain et al. [5]; Beristain et al. [6]), and this is why we wanted to analyze LOH in different types of both pathological and neutral variants in order to use this feature to determine, as the authors state, the pathogenicity or neutrality of the unclassified variants. DNA was extracted from paraffin-embedded breast tumor tissue and adjacent non-tumor tissue taken from 15 patients. DNA fragments including the 17 different abnormalities found in these 15 patients were amplified (Table 1). Age of samples and the hospital providing them were assessed, in case these variables interfered with the results, and none of these two variables seemed to have a direct relationship with the results obtained. In non-tumor tissue, LOH was not found in any case. Table 1 only shows the results found in tumor tissue. Thus, this table shows the variants tested, both in BRCA1 and BRCA2, and the results of the LOH analysis. The nomenclature based on the protein level proposed by the Human Genome Variation Society (HGVS) was used to designate the variants [7], except for those that must be named according to cDNA (synonymous, intronic, and splicing variants). The effect of each variant on its implication in breast/ovarian cancer was taken from the Breast Cancer Information Core (BIC) database [8]. The term ‘‘likely’’ precedes the cases for which there is ample literature showing an effect different from that proposed by BIC. Variants found for the first time in our study population are termed ‘‘NEW’’. The results showed no homogeneous LOH pattern. Interestingly, LOH was found in most neutral variants, which should not show LOH, although the allele containing the variant, rather than the wild type allele, was lost in most cases. Pathological variants, which should show LOH in the wild type allele, also did not follow the expected pattern. Thus, since a different result was not obtained for the neutral variants as compared to the pathological variants, E. Beristain M. I. Tejada (&) Molecular Genetics Laboratory, Department of Biochemistry, Hospital de Cruces, Pza. de Cruces s/n, 48903 Barakaldo, Bizkaia, Spain e-mail: MARIAISABEL.TEJADAMINGUEZ@osakidetza.net