Insight Into HCDR3 Restrictions in CLL by Analysis of Incomplete IGHD-IGHJ Rearrangements: Further Evidence that Somatic Selection Shapes the Expressed CLL Immunoglobulin Repertoire.

Abstract 2346

Poster Board II-323

A strong sequence-based evidence supporting a role for antigen in the development of CLL is the existence of subsets of patients with stereotyped heavy complementarity-determining region 3 (HCDR3) sequences. Stereotyped HCDR3s are often defined by the selective association of certain IGHD genes in specific reading frames (RF) with certain IGHJ genes, especially IGHJ6. To gain insight into the mechanisms shaping the IG repertoire and also determine the developmental stage when restrictions in HCDR3 are imposed, we investigated the molecular features of incomplete IGHD-IGHJ rearrangements (IDJR) in a series of 830 patients with CLL. IDJRs were detected in 272/830 cases (32.7%). No associations were identified between the occurrence of IDJRs and IGHV gene usage or mutational status in the complete IGHV-D-J rearrangement from the coding IGH allele. A trend for higher IDJR frequency was evident, however, in certain subsets with stereotyped HCDR3s, in particular subset #1 (IGHV1-5-7/IGHD6-19/IGHJ4; 13/33 cases, 40%), #7 (IGHV1-69/IGHD3-3/IGHJ6; 10/21 cases, 48%) and #8 (IGHV4-39/IGHD6-13/IGHJ5; 5/12 cases, 41%). Sequence analysis of the IDJRs revealed: (i) increased frequency of IGHD2 subgroup genes (115/238 cases, 48%), especially IGHD2-2; (ii) equal distribution of the three RFs of the IGHD genes; (iii) increased recombination frequency between 5`genes of the IGHD cluster and 3` genes of the IGHJ cluster, suggestive of secondary rearrangements on the same allele. Overall, 205/238 (86%) IDJRs were considered as potentially functional (PF), since they did not carry a stop codon at the IGHD-J junction. Of note, 26/28 (93%) IDJRs detected in cases from subsets #1, 7 and 8 could be assigned to the PF category. In the group of CLL cases carrying PF IDJRs, comparison of the IGHD gene repertoire in IDJRs vs. complete, expressed IGHV-D-J rearrangements (CE-VDJRs) revealed: (i) statistically significant (p<0.001) selection of the IGHD3-3 and IGHD6-19 genes in RF2 and RF3, respectively, among CE-VDJRs (especially those assigned to subsets #7 and #1, respectively); (ii) preferential usage of RFs encoding for hydrophilic peptides among CE-VDJRs. At a subsequent stage, we compared the repertoire of the IDJRs from the CLL cohort to that of 174 IDJRs obtained from patients with pre-B acute lymphoblastic leukemia (ALL). Except for higher frequency of (i) the IGHD7-27 and IGHJ6 genes and (ii) IGHD-IGHD gene fusions in pre-B ALL, the overall configuration of IDJRs did not differ significantly in CLL vs. pre-B ALL. In conclusion, these results document that the early stages of IG gene rearrangements in pre-B ALL and CLL do not show intrinsic, disease-specific differences. The detailed molecular characterization and comparison of the IGHD and IGHJ gene repertoires in IDJRs vs. CE-VDJRs in CLL provides further support for the notion that CLL development is not stochastic but directed by selection operating at the IG protein level.