Diffuse large B-cell lymphomas (DLBCLs) are aggressive tumors that arise from germinal center B cells (GCBs) and post-GCBs and are noted for their heterogeneity and variable clinical outcomes. Epigenetic modifications like DNA methylation of cytosine nucleotides have emerged as important mechanisms of gene regulation and have been implicated in carcinogenesis. Our previous genome-wide studies in primary samples revealed profound alterations in the cytosine methylation patterning of DLBCLs. We also found that expression of activation-induced deaminase (AID) was significantly associated with the loss of methylation in DLBCL patients and was predominantly identified within computationally predicted AID-binding RGYW motifs. AID is a cytidine deaminase required for class switch recombination and somatic hypermutation (SHM) of immunoglobulin genes in GCBs. The enzymatic machinery that mediates these processes is error-prone and may introduce point-mutations and changes in DNA methylation, resulting in genomic and epigenomic instability. Since AID can also function as a demethylase during embryonic development, we asked whether AID has demethylase activity during transit of B cells through the GCs and if its overexpression can contribute to lymphomagenesis through disrupting DNA methylation. To address this question, we studied the epigenetic function of AID in GCBs and GC-derived lymphomas. We characterized the methylome of naïve B cells (NBs) and GCBs isolated from human tonsils and spleens of immunized mice by enhanced Reduced Representation Bisulfite Sequencing (eRRBS). We observed that the transition from NBs to GCBs was characterized by DNA hypomethylation, with 60,000 and 8,000 differentially methylated CpGs (DMCs) that were hypomethylated in GCBs compared to NBs, in human and mouse respectively. We also found that hypomethylated regions were enriched for the putative AID binding site RGYW (Wilcoxon P <.001). Furthermore, AID knockdown in lymphoma cells (RAMOS) resulted in preferential hypermethylation at AID-binding sites (Chi square P ~ 0). We then isolated DNA from splenic NBs and GCBs from Aicda-/- (AID-deficient) and Aicda+/+ (wild type) mice and performed eRRBS analysis, obtaining single nucleotide resolution for 2.5-3 million represented CpGs. We observed that most of the 8,000 hypoDMCs identified between GCBs and NBs in Aicda+/+ mice were absent in Aicda-/- mice (800 hypoDMCs between GCBs and NBs Aicda-/- cells), implying that AID is a regulator of DNA methylation in GCBs. In addition, those AID-dependent hypoDMCs were predominantly localized in introns (35%), and also in promoters (10%) and exons (10%). We then defined differentially methylated regions (DMRs) based on the following criteria: ≥ 5 DMCs and methylation difference ≥10%, with >250bp between DMRs. We identified DMRs that get hypomethylated in GCBs in the Aicda+/+ mice, but are not hypomethylated in Aicda-/- GCBs, corresponding to >200 genes that represent AID epigenetic targets. These genes include factors involved in B cell function and differentiation like PAX5, BCL2L1, IRF8 and others. Not unexpectedly, many of epigenetic targets are also known targets for SHM, but some are novel targets that only demonstrate evidence of epigenetic deregulation. We also analyzed the transcriptome of NBs and GCBs from Aicda-/- and Aicda+/+ mice by RNA-seq and detected an increase in DNMT1 expression in Aicda-/- cells compared to Aicda+/+ cells. There were no significant changes in expression of other factors involved in modification of cytosine methylation, such as DNMT3a/3b, TET1/2/3, UNG or MSH2/6. Finally, we performed bone marrow transplantation experiments using VavP-Bcl2 mice, which are known to develop GC-derived lymphomas. We transplanted VavP-Bcl2 bone marrow cells infected with AID-expressing retroviral vectors into C57BL/6 mice and monitored the progression of the resulting BCL2-driven lymphomas. Our preliminary results indicate that high AID expression is correlated with a more aggressive phenotype of the disease. We are currently analyzing the epigenetic targets of AID in both normal GCBs and tumors, in order to find genes that could be epigenetically deregulated and contribute to the formation of lymphomas. Our results demonstrate for the first time that AID functions as a demethylase in GCBs in vivo and suggest that the epigenetic role of AID could contribute to lymphomagenesis.

Disclosures

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution