Differential Role of Crebbp Missense and Truncating Mutations in the Malignant Transformation of Germinal Center B Cells

Somatic mutations of the CREBBP acetyltransferase are highly recurrent in germinal center (GC)-derived lymphomas, including follicular lymphoma (FL, 60% of cases) and diffuse large B cell lymphoma (DLBCL; 40% of cases in the EZB/C3 genetic subtype). Mutations include prototypical inactivating events that abrogate the C-terminal acetyltransferase (AT) domain, as well as amino-acid changes clustering within this domain, which impair the protein enzymatic activity by decreasing its affinity for acetyl-coenzyme A (Pasqualucci et al., Nature 2011). These events are acquired early during lymphomagenesis, in a putative common precursor cell (CPC) that subsequently undergoes divergent evolution to FL or tFL through the acquisition of additional genetic lesions. While extensive work has focused on understanding the lymphoma-driving role of CREBBP protein loss, mimicking the outcome of truncating mutations, the functional significance of missense mutations remains largely unexplored. This is a critical gap because, different from truncated proteins, CREBBP missense mutants are expressed, suggesting they may contribute to malignant transformation through distinct mechanisms. In support of this view, missense and truncating mutations are differentially distributed in FL and DLBCL, with the former being preferentially enriched in FL (p<0.01). The aim of this study was to investigate the in vivo role of CREBBP missense mutations during the initial phases of lymphomagenesis, and contrast it with the phenotype resulting from CREBBP loss.

To address these questions, we took advantage of a novel, conditional knock-in mouse model where the most common R1446H mutant protein (R1447 in mice, hereafter Crebbp^RH) is expressed specifically in GC B cells under the control of its endogenous promoter, following T cell-dependent immunization and Cγ1Cre-driven recombination. RT-PCR and sequencing confirmed efficient, GC-restricted activation of the knock-in allele, and Crebbp protein levels were comparable in homozygous Crebbp^RH vs WT littermates, documenting a stable protein. While modest differences were noted in the GC B cell fraction of Crebbp^KO mice, consistent with our previous work, expression of the Crebbp^RH allele led to pronounced, dose-dependent expansion of the GC (p<0.001), the structure from which FL and DLBCL arise. These GCs displayed an abnormal dark zone to light zone polarity manifesting as an increased proportion of centrocytes (p<0.001), and persisted abnormally for several weeks after immunization, when the GC reaction is normally extinguished, indicating an altered dynamics.

To investigate the mechanisms underlying this phenotype, we contrasted the transcriptomic and chromatin profiles of Crebbp^RH and Crebbp^KO GC B cells. We found shared but also unique transcriptional changes separating the two cohorts by unsupervised hierarchical clustering. Consistently, a significant number of differentially expressed genes were unique to Crebbp^RH vs Crebbp^KO GC B cells (q<0.05, FC≥1.2). In particular, Crebbp^RH GCs were positively enriched in light-zone-associated and pre-memory B cell differentiation programs (Ccr6, Gpr183, Klf2); in contrast, these signatures were attenuated upon Crebbp deletion, indicating missense mutations may favor distinct cell fates upon GC exit. While global H3K27Ac was reduced in both Crebbp^RH and Crebbp^KO cells, consistent with impaired AT activity, Cut&Run analysis revealed a significant Crebbp^RH-mediated redistribution of the H3K27Ac mark at super-enhancers, including those linked to genes involved in GC exit and B:T-cell interactions. Accordingly, expression of Crebbp^RH alleles led to a two-fold increase in memory B cell responses to the T cell-specific antigen NP-KLH (p<0.01), which was not detected in Crebbp^KO mice, as measured by immunophenotyping. Moreover, Crebbp^RH but not Crebbp^KO mice showed compositional changes in the T cell microenvironment, with expansion of T follicular helper cells (8.3%, vs 3.8% in WT mice; p<0.05). Thus, Crebbp missense mutations may skew GC B cell fate decisions toward memory B cell differentiation and favor the persistence of enlarged GCs to facilitate the malignant transformation process, in part by shaping the GC microenvironment. These data have implications for the understanding of FL/DLBCL ontogeny and for the therapeutic targeting of tumors carrying missense vs truncating mutations.