Distinct VDJ-H Gene Usage between Mutated and Unmutated Patients in B-CLL: Implications for Pathogenesis and Correlation with Clinical Outcome.

In B cell chronic lymphocytic leukemia (B-CLL) somatic hypermutation in the immunoglobulin variable heavy chain (IgVH) identifies two distinct prognostic variants with mutated (M) patients having better survival. Ig recombination in this disease demonstrates biased V, D and J gene usage. To determine features of VDJ recombination associated with changes in biologic behavior we sequenced and analyzed the IgH of 356 patients with B-CLL. Of 375 sequences obtained (19 patients express two VH genes), 202 (54%) were unmutated (UM) (<2% from germline) with 84 (41%) in 100% germline configuration. UM CLL overused VH1 gene family (e.g. VH1-69, VH1-2) compared to VH3 and VH4 (e.g. VH3-7, VH3-23, VH3-48, VH4-34) in M CLL (p<0.0001). Focusing on V gene segment location, D distal VH genes were overused in CLL compared to normal B cells (p=0.05). However, this was accounted for mostly by the marked over usage of a single V gene, VH1-69, used in 66 of 202 (33%) UM sequences, but 3 of 173 (2%) M cases. V gene segments usage in M CLL was similar to normal B cells (p=0.04). CLL cases overused J distal D genes (p<0.0001), mostly accounted for by overuse of DH2-2 (15%) and DH3-3 (27%) in UM cases. JH4 (37%) and JH6 (33%) were most commonly used (p<0.0001), with JH6 used more in UM cases (p<0.0001). JH5 was overused in CLL (18%) compared to healthy B cells (7%). Analysis revealed more mutations in the framework regions, but the ratio of silent to replacement mutations was elevated in the complementarity-determining regions (CDR) (3.45 vs. 1.39), consistent with antigen driven selection. Of note, this was also true in those cases with <2% mutations (5.20 vs. 1.57), normally classified within the UM group. In keeping with the notion that UM cases might also be antigen driven, we identified patients using identical CDR3 regions including 3 patients using VH3-11, DH3-10, and JH1, and 2 cases using VH5-51, DH3-3 and JH4. Among 21 cases using VH1-69, D3-3, and JH6, 4 were identical and 17 others had marked CDR3 homology. CDR3 length was significantly longer in UM than M cases (p=0.0001) with more negatively charged amino acids and lower calculated isoelectric point (median pI 3,8 vs. 4,8; p=0.005), which might reflect selection for specific structural motifs that facilitate antigen binding. We observed no statistically significant differences in predicted binding affinity of MHC Class I binding nonameric peptides in M compared to UM CLL patients (p=0.41), suggesting improved outcome in M CLL patients is not due to an increased T cell response. In keeping with the data of others, the median time from diagnosis to treatment of UM patients (22 months) was significantly shorter than of M patients (70 months) (p<0.001). Patients using D distal V genes had worse outcome (p<0.001), although this was accounted for entirely by the poor outcome of the patients using VH1-69. Time to treatment was significantly shorter for patients using the VH1 gene family (p<0.001), even when VH1-69 patients were excluded from analysis (p=0.05). We conclude that CLL is characterized by VDJ use patterns that differ distinctly between M and UM CLL patients, but with the exception of VH1-69, and potentially other VH1 family members, the location of the V gene did not influence outcome. Therefore, mechanisms other than gene localization must be responsible for the poorer outcome of CLL patients utilizing VH1-69.