Divergence from the germ-line sequence in unmutated chronic lymphocytic leukemia is due to somatic mutation rather than polymorphisms.

Division of chronic lymphocytic leukemia (CLL) into 2 subsets, with and without mutations of the IgVH genes, has important prognostic significance. The unmutated subset, with an inferior prognosis, was originally defined by 98% or more homology to the germ-line sequence based on the observation that polymorphisms may lead to 2% difference in the DNA sequence of alleles within the immunoglobulin loci.1 The degree of homology to the germ-line sequence that optimally delineates these 2 clinical subsets of CLL remains controversial. Both of the original studies2,3 demonstrating the prognostic significance of IgVH gene mutational status found that a 98% cutoff provided the optimum discrimination, but recently a 97% or 95% cutoff was found to provide better discrimination.4,5 In addition, cases using the VH3-21 and possibly the VH3-23 gene segments are associated with a poor prognosis regardless of their mutational status.6,7 In this study we have sequenced the corresponding germ-line IgVH gene from the neutrophils of a number of patients with CLL whose rearranged IgVH gene showed between 96.6% and 99.6% homology to the germ-line database sequence, in order to determine the relative incidence of polymorphisms and point mutations in cases with a low mutational load. Germ-line IgVH sequences were obtained from all 13 cases analyzed. Clinical stage, IgVH gene usage, and percent homology to the germ-line sequence are shown in Table 1. Analysis of the granulocytes sequence showed that 70 of the 72 base changes observed in the lymphocytes were absent in the corresponding germ-line sequence. However, polymorphisms did account for the single base pair change in the VH1-02 gene in case 1 and for 1 of the 7 changes in the VH3-48 gene in case 10 (Table 1). In this study the frequency of IgVH gene polymorphisms is 0.0005%, suggesting that a cutoff value of 99.99% homology to germ line would be needed to separate truly unmutated IgVH genes from those that had undergone some degree of somatic hypermutation. Our results confirm that the leukemic B cells in most cases of CLL are exposed to a mutational mechanism. However, cases whose IgVH genes show complete homology to the germ-line sequence are indistinguishable both clinically and in vitro from cases with 99% or 98% homology. This suggests that factors that determine mutational load rather than mutational status per se account for the differences in biology and clinical behavior in CLL. Until these factors are identified, it would seem prudent to choose a cutoff for IgVH gene homology based on the clinical outcome of patients with known IgVH gene status rather than an arbitrary figure that implies that leukemic cells either have or have not undergone somatic hypermutation.