Revisiting the definition of somatic mutational status in B-cell tumors: does 98% homology mean that a V(H)-gene is unmutated?

In B-cell tumors, accurate assessment of somatic mutational status has implications for defining tumor origin and, in some categories, has relevance for predicting the clinical outcome. In 1999, it was found that in chronic lymphocytic leukemia (CLL) cases carrying VH-genes with >98% homology to the closest germline (GL) gene identified a subset with a poorer prognosis, as compared to cases which had clearly undergone somatic mutation.1 Initially, the 98% cutoff was used to exclude potential polymorphic variant sequences. It was a short cut to avoid analyzing the corresponding GL gene in each patient, and is now used for assignment of mutational status in a range of B-cell tumors. The most integrated V-base database (http://www.mrc-cpe.cam.ac.uk/v base) describes 47 polymorphisms in 30/51 functional VH-gene segments, with nucleotide differences from GL generally of <1%, and only five less frequently used VH segments, V1-03, V1-69, V3-13, V4-28 and V4-b, with >2% variation. Aligned VH-genes readily map to the correct allele, and the 98% cutoff point seemed reasonable. However, low levels of mutations can occur in VH-genes and single-base changes (0.33–0.35% from germ line) may be important for improving the antigen recognition.2 It is possible therefore that significant changes could be occurring in cases with apparent >98% homology to GL. While it is unlikely to affect the prognostic power in CLL, it has implications for our understanding of the behavior of B-cell tumors. One problem with assessing low levels of mutation is that of the approaching Taq error rate, and, although an increased rate may suggest somatic mutation, it can still be questioned. An indicator of the reality of base changes is identification of the same change in >1 sequence derived from separate amplification reactions. However, this parameter applies mainly to mutations occurring early in tumor development, and may not be evident from those accumulating late in the life of the clone. It has revealed genuine intraclonal heterogeneity in some cases of monoclonal gammopathy of undetermined significance, distinguishing it from the more malignant counterpart, multiple myeloma.3 For B-cell tumors with a low level (<2%) of somatic mutation, this approach has not generally been applied, since there is no obvious clinical reason for investigation, and the intraclonal mutational rate is clearly below that exhibited during diversification of the antigen receptor in normal B cells. However, there is evidence for a low rate of intraclonal mutation in CLL, even in the 'unmutated' subset with >98% homology to GL sequence.4 Among these cases, only 1/5 had variants repeated in >1 clone. While lack of repeats prevents confirmation of the reality of the mutations, it might be expected if mutational events are relatively random and of low frequency. An intriguing single case of splenic marginal zone lymphoma with >99% homology also showed suggestive evidence for intraclonal heterogeneity.5 Recently, we have analyzed VH-genes in hairy cell leukemia (HCL). These tumors commonly express multiple isotypes and the majority have undergone somatic mutation in all isotype variants.6 The new finding that 2/13 HCL cases displayed germline tumor-derived VH-genes (Forconi F et al, submitted) prompted the question whether a further 3/13 cases with >98% homology were actually unmutated. Tumor VH transcripts were identified by established RT-PCR and cloning procedures described previously.6 Full sequence transcripts were analyzed and shared the same CDR3 in multiple clones. Cases 103 and 163 (Figure 1) revealed VH-genes with 98.6 and 98.2% homology to the closest germline allele, V4-31 and V1-02, respectively.6 Case 330 displayed 98.6% homology to the V2-05 gene.