HIGHER-ORDER IMMUNOGLOBULIN SEQUENCE RELATIONS FOR MAJOR SUBSETS OF CHRONIC LYMPHOCYTIC LEUKEMIA: UNIQUENESS VERSUS EQUIVALENCE

Background
Studies of the B cell receptor immunoglobulin (BcR IG) in chronic lymphocytic leukemia (CLL) revealed that ~30% of cases can be assigned to stereotyped subsets. In our previous publication of 7,596 CLL BcR IG gene sequences, we focused on 19 major subsets, collectively accounting for 12% of the cohort. This process was based on the implementation of stringent criteria focusing on high sequence relatedness. However, we have been aware of BcR IG sequences similar to those in or even between major subsets, yet set apart due to not fulfilling the adopted criteria.

Aims
We systematically searched for previously obscured sequence relatives of major BcR IG stereotypes with the aim of obtaining a wider, better connected, and thus more insightful view into IG repertoire restriction in CLL.

Methods
Sequence relatives of major subsets were sought for through applying a series of criteria in reference to each respective subset: (i) usage of the same IGH genes, (ii) VH CDR3 length difference ranging from -2 to +2 amino acids (aa); and, (iii) expression of the same VH CDR3 sequence motif. A total of 20331 productive IGHV-IGHD-IGHJ gene rearrangements from a multi-institutional series of CLL patients were evaluated, with assignment to major subsets performed with purpose-built bioinformatics methods. 

Results
Overall, 2450/20331 (12.1%) CLL sequences formed the 19 major subsets, each containing 44 to 548 sequences. The relative frequency of each major subset was highly similar between the present and the previous analysis, justifying their consideration as major. The quest for sequence relatives to major subsets was performed individually for each subset. For major subsets with mutated IGHV genes (M-CLL), generally, very few non-subset sequences showed adequate VH CDR3 similarity, and even these utilized different IGHV genes, thus not satisfying our search criteria. Amongst subsets with unmutated IGHV genes (U-CLL), a striking case of relatedness was observed between major subsets #1 (501 cases) and #99 (70 cases), characterized by the usage of the same IGH genes and the presence of a shared QWL motif at the same positions within VH CDR3. Their single important difference concerned VH CDR3 length (subset #1: 13 aa, subset #99: 14 aa). For the remaining major U-CLL subsets, certain subsets remained “unique”, contrasting others for which we identified closely related sequences we now deem as “satellite” subsets. In particular, in the case of subset #8 (IGHV4-39) we identified two “satellite” subsets that consisted of 35 and 10 cases and displayed relative sizes of 37.6% and 10.8% (compared to subset #8), respectively. Even more pronounced was the case of subset #31 (IGHV3-48), with its “satellite” comprising of 51 sequences and a relative size of 86.4% compared to subset #31, thus almost equal to its established relative. 

Summary
Major M-CLL stereotyped subsets remained isolated, with unique sequence characteristics. In contrast, a clear trend was evident for many major U-CLL subsets in having “satellites”, potentially functionally equivalent, and often large in size. These findings could be attributed to the intrinsic differences between M-CLL versus U-CLL subsets in terms of IG gene repertoire and VH CDR3 features. Considering that major stereotyped subsets may indeed represent distinct disease subgroups, our highlighting of higher-order sequence relations has implications for clinicobiological research aimed at dissecting the heterogeneity of CLL towards identifying distinct profiles that would assist in clinical decision making.

Keyword(s): CDR3, Chronic lymphocytic leukemia, Immunoglobulin gene

www.eha.org