Sequence-Based Evidence for Antigen Selection in Mantle Cell Lymphoma: Remarkable Immunoglobulin Gene Repertoire Biases, Stereotyped Antigen-Binding Sites and Recurrent Hypermutations in Certain Subsets.

Abstract 1933

Poster Board I-956

According to the 2008 WHO Classification, mantle cell lymphoma (MCL) most likely derives from neoplastic transformation of a peripheral B cell of the inner mantle zone, mostly of naïve pre-germinal center type. Analysis of the immunoglobulin (IG) repertoire in MCL has revealed a bias in IG heavy variable (IGHV) gene usage and a predominance of unmutated rearrangements, with a variable minor proportion of cases carrying mutated IGHV genes (according to the 98% identity cutoff). However, limited knowledge exists regarding antigen-binding site sequence restrictions or specific somatic hypermutation (SHM) patterns in MCL. We examined the productive IGHV-D-J rearrangements of 651 MCL cases (including 398 cases from the European MCL Network), the largest series to date. We confirm and significantly extend previous observations that the IGHV gene repertoire is remarkably biased, with only three genes accounting for 40.3% of cases (IGHV3-21, 16.7%; IGHV4-34, 15.3%; IGHV1-8: 8.3%). Using bioinformatics approaches previously applied to CLL, biased associations of certain IGHV, IGHD and IGHJ genes with restricted (stereotyped) heavy complementarity-determining regions (HCDR3s) were identified in 68/651 cases (10.4%). Overall, 24 subsets of cases with stereotyped HCDR3s were recognized. Eight of 24 subsets included 3-7 cases each (“confirmed”), while the remaining 16 subsets included two cases each and were considered “provisional”. Stereotyped HCDR3s were found predominantly amongst rearrangements utilizing the IGHV3-21 and IGHV4-34 genes and the IGHJ6 gene. Notably, the MCL HCDR3 stereotypes identified here were distinct from those previously reported in CLL. Based on SHM analysis, the sequences were divided into three groups: (i) truly unmutated (100% germline identity, GI): 189/651 sequences (29.2%); (ii) minimally/borderline mutated (98-99.9% GI): 306/651 sequences (47%); and (iii) mutated (<98% GI): 155/651 sequences (23.8%), of which 93 had a GI<97%. In keeping with previous reports, the IGHV gene repertoire of the three identity groups differed considerably. For instance, the IGHV3-21 gene was used by 7% of rearrangements with <98% identity vs. 22.8% of rearrangements with 100% identity. In contrast, the IGHV3-23 gene was over-represented among mutated rearrangements (26.4%). Shared (“stereotyped”) amino acid (AA) changes across the entire IGHV gene sequence were identified for certain groups of sequences, especially those utilizing the IGHV1-8, IGHV3-21 and IGHV3-23 genes. A comparison to published series from other entities, in particular CLL, revealed that several stereotyped mutations identified in MCL were “disease-biased”. Importantly, stereotyped AA changes were also observed in the borderline or minimally mutated groups, indicating that even a low level of mutations may be functionally relevant. In conclusion, MCL is characterized by a highly distinctive IG gene repertoire with restricted HCDR3s and very precisely targeted and, probably, functionally driven SHM, strongly implying a role for antigen-driven selection of the clonogenic progenitors. Based on the evidence presented here, an antigen-driven origin of MCL could be envisaged, at least for subsets of cases.