Aberrant somatic hypermutation in large B-cell lymphomas and its relationship to mutational signatures, including a novel haloalkane pattern.

Abstract

7076 Background: Diffuse large B cell lymphoma (DLBCL) is a genetically complex disease, often characterized by the simultaneous occurrence of multiple mutations within single driver genes. Here, while activation-induced cytidine deaminase (AID)-driven aberrant somatic hypermutation (aSHM) is recognized as a major contributor to mutation, many genes could be heavily mutated due to other mechanisms. In this study, we evaluate clinical DLBCL samples identified as hypermutated, to assess the presence of AID driven mutation and the contribution of other defined mutational signatures which may contribute to clinicopathologic heterogeneity. Methods: Clinical samples of DLBCL tested by a hybrid-capture next-generation sequencing (NGS) panel targeting up to 468 genes, were identified. Hypermutated cases were selected based on the qualifying criteria of having three or more co-occurring mutations in at least 1 gene. Based on the mutational position from transcriptional start site and inter-mutational distance, cases were initially classified as having aSHM or Kataegis. Mutational signature analysis was performed using all high-confidence somatic single-base substitutions (SBS) identified and interpreted according to COSMIC SBS signatures v3.5. A predominant signature was defined when >40% of the mutations were attributed to the signature. Wherever possible, correlation between the morphological type, molecular subtype, and mutational signatures was performed. Results: Out of the 216 samples identified as hypermutated, most (n=172, 79.6%) were classified as having aSHM, 9.3% (n=20) as Kataegis and 11.1% (n=24) were unclassifiable. The 3 most common genes targeted by aSHM were PIM1(n=66), SOCS1(n=55) and BCL2(n=35). By contrast, SGK1(n=16), MYC(n=12), GNA13(n=6) and ITPKB(n=6) were the common targets of Kataegis. The most frequent dominant molecular signatures included AID (n=82) and Clock-like aging(n=48), followed by Haloalkane (n=24), MMR deficiency (n=17) and the SBS39 (n=14) signature of unknown etiology. Based on morphology and immunophenotype, cases were primarily stratified as either Germinal Center B-cell like (GCB; n=86) or Non-GCB (n=87), with lower representation of Primary mediastinal B cell lymphomas (n=21), High grade B cell lymphomas with MYC and BCL2 translocations (n=18), and DLBCL special subtypes (n=4). The same top 3 molecular signatures were identified across all subtypes, except in GCB, where MMR deficiency constituted the 3rd most frequent signature. Conclusions: In DLBCL, mutational signatures are primarily driven by endogenous processes, namely AID and Aging. The identification of a haloalkane signature in a small proportion of cases is a novel finding, raising suspicion for the role of exogenous influences, either environmental or therapeutic in origin. Further studies are needed to better define this cohort.