Abstract
Most B cell non-Hodgkin’s lymphomas (B-NHL) derive from germinal center (GC) B cells and their pathogenesis is associated with the accumulation of distinct genetic lesions, including chromosomal translocations and a more recently identified mechanism of genomic instability, termed aberrant somatic hypermutation. These alterations are thought to be due to mistakes occurring during two GC-associated immunoglobulin (Ig) genes remodeling processes: class switch recombination (CSR) and somatic hypermutation (SHM). However, this model has never been formally proven. To conclusively investigate the role of CSR and SHM in the pathogenesis of B-NHL, we examined whether lymphoma development in mice requires the function of activation induced cytidine deaminase (AID), a DNA editing enzyme expressed specifically in GC and activated B cells and essential for both processes. Three transgenic mouse models were generated by crossing lymphoma-prone mice (λMYC, λMYC/IμHABCL6 and IμHABCL6) with mice (AID−/−) that are unable to undergo both SHM and CSR. The λMYC mice develop a diffusely infiltrating monoclonal proliferation of pre-GC origin, with unmutated IgV genes and lack of BCL6 expression, and therefore presumably independent from AID-associated DNA remodeling events. Conversely, lymphomas in λMYC/IμHABCL6 and IμHABCL6 mice recapitulate GC/post GC-derived malignancies, in that the former display somatically mutated IgV genes and upregulation of post-GC markers (CD138) in most of the cases, while the latter develop a splenic lymphoproliferative syndrome that culminates, past 12 months of age, in clonal B cell lymphomas with DLBCL morphology and somatically mutated IgV genes (~70% of the animals) (Cattoretti et al., Cancer Cell 7:445–455, 2005). Mice were monitored for tumor incidence and survival, and a combination of histologic, immunophenotypic and gene expression profiling analysis was used for tumor characterization. As expected, no significant differences in event-free survival and lymphoma type were observed between AID-proficient and AID-deficient λMYC mice, in agreement with their pre-GC derivation. Conversely, a phenotypic shift of the tumor was observed in λMYC/IμHABCL6 mice when bred into an AID−/− background, with >80% of the cases (N=21/26) reverting to a pre-GC phenotype (loss of GC/post GC markers) undistinguishable from that of the λMYC and λMYC/AID−/− mice. Gene expression profile analysis on representative cases (N=10 λMYC/IμHABCL6 and 5 each for λMYC, λMYC/AIDKO, λMYC/IμHABCL6/AIDKO) confirmed significant phenotypic similarities between pre-GC derived λMYC lymphomas and the λMYC/IμHABCL6/AID −/− lymphomas, which co-segregated in a separate cluster from λMYC/IμHABCL6 tumors. Analogously, a significant reduction in DLBCL frequency was observed in the IμHABCL6/AIDKO cohort as compared to IμHABCL6 mice (N= 4/19, 21% vs 8/14, 57%; p=0.03).
Taken together, these results indicate that GC-derived lymphomas cannot develop in the absence of AID, thereby providing direct support to the notion that AID-mediated mistakes in antigen receptor gene modification events (CSR and SHM) represent major contributors to B-NHL pathogenesis.
Somatic Hypermutation Shapes the Antibody Repertoire of Memory B Cells in Humans