Analysis of Stereotyped IGHV Distribution In a Series of 1133 Chronic Lymphocytic Leukemia Patients: The Experience of a Multicenter Italian Study Group

Chronic lymphocytic leukemia (CLL) is characterized by an extremely variable clinical course. Mutational status of the immunoglobulin heavy-chain variable (IGHV) region defines two disease subsets with different prognosis. A fraction of CLL cases carries highly homologous B-cell receptors (BCR), i.e. characterized by non-random combinations of immunoglobulin heavy-chain variable (IGHV) genes and heavy-chain complementarity determining region-3 (HCDR3). We performed sequence analysis to characterize IGHV regions in a panel of 1133 CLL patients investigated by a multicenter Italian study group. A total of 1148 rearrangements were identified; the analysis of stereotyped subsets was performed based on previously reported criteria (Messmer et al, J Exp Med 2004; Stamatopuolos et al, Blood 2007). Specifically, we compared all our sequences with those found in three different publicly available data sets (Stamatopoulos et al, Blood 2007; Murray et al, Blood 2008 and Rossi et al, 2009 Clin Cancer Res). In addition, a pairwise alignment within all sequences was performed in order to discover novel potential subsets (HCDR3 identity > 60%). Based on the 2% cut-off used to discriminate between Mutated (M) and Unmutated (UM) cases, 777 sequences (67.59%) were classified as M, while 371 sequences (32.3%) as UM. The most represented IGHV genes within mutated cases were IGHV4-34 (104/118) and IGHV3-23 (85/96), whereas IGHV1-69 (97/112) was the most frequently used in the UM group. Interestingly, the IGHV3-21 gene, reported to be frequently expressed in CLL patients from Northern Europe, was present in only a small fraction of cases (24; 2.07%), confirming a previous finding reported by Ghia et al (Blood 2005) in a smaller panel. In our series, stereotyped HCDR3 sequences were found in 407/1148 (35.45%) patients, 177 of whom were M and 230 were UM cases. Overall, we observed that stereotyped sequences were significantly associated with UM IGHV status (Fisher’s exact test, P<0.0001). Among the 407 stereotyped HCDR3 sequences, 345 belong to the clusters reported by Murray et al and 14 to those described by Rossi et al., 2009 Clin Cancer Res. The most frequent stereotyped subsets identified in our panel were #1 (35 cases), #7 (28 cases), #4 (24 cases), #3 and #9 (16 cases), #28 (13 cases), and #2 (12 cases), together with subsets #5, #8, #10, #12, #13, #16 and #22 (all ranging from 6 to 9 cases). Finally, we were able to identify by auto-matching analysis 48 sequences potentially specific for 23 novel putative stereotype subsets. In our series we identified 407/1148 (35.45%) stereotyped HCDR3 sequences. The percentage was higher than that reported by Stamatopoulos et al and Murray et al. This discrepancy may partially be due to the different approach used in our analysis, namely the matching to a general data set including all published stereotyped subsets instead of the auto-matching performed by those Authors. We demonstrated a significant association between IGHV status and stereotyped sequences and confirmed the finding that #1 is the most frequent subset identified so far. Finally, we were able to identify a series of 23 novel putative subsets that will require further confirmation.