Targeting Histone Acetyltransferase Gene Inactivation in Diffuse Large B Cell Lymphoma

Diffuse Large B-cell Lymphoma (DLBCL) is the most common form of non-Hodgkin lymphoma, accounting for ~30% of de-novo diagnoses and also arising as a frequent clinical evolution of indolent lymphomas. Although curable in a substantial fraction of cases, one third of patients do not achieve durable remissions, highlighting the need for novel, targeted therapies. Over the past decade, we and others have identified the CREBBP acetyltransferase and, less frequently, its paralogue EP300 as highly recurrent targets of inactivating somatic mutations/deletions in DLBCL and follicular lymphoma (FL) (30% and 60% of patients, respectively), indicating a prominent role in the pathogenesis of these tumors (Pasqualucci et al., Nature 2011). In most cases, mutations are heterozygous and the residual wildtype (WT) allele is expressed, suggesting a haploinsufficient tumor suppressor role. Indeed, germinal center (GC)-specific loss of Crebbp perturbs the expression of genes that are relevant to the normal biology of this structure, i.e. the lymphoma cell of origin, and cooperates with BCL2 deregulation to increase the incidence of tumors recapitulating the features of the human disease (Zhang et al., Cancer Discovery 2017). Intringuingly, while CREBBP binds to virtually all GC-specific superenhancers, no detrimental effects were observed upon its deletion in mice, suggesting the existence of compensatory mechanisms. Consistent with this hypothesis, inactivation of CREBBP and EP300 rarely coexist in human DLBCL and FL, suggesting that cells require a certain amount of acetyltransferase activity.

To investigate whether EP300 compensates for CREBBP loss in the GC, we analyzed the GC responses in compound mouse models engineered to specifically delete these two genes (alone and in combination) upon SRBC immunization and induction of a Cγ1-driven Cre-recombinase. While Crebbp^KOmice showed a mild increase in GC formation, as reported, loss of Ep300 led to ~40% reduction in the percentage of GC cells (mean: 1.8% vs 3.1% in WT littermates; p<0.05), documenting that these two enzymes play non-entirely overlapping roles in this population. Importantly, GC formation was completely abrogated in Crebbp^KOEp300^KO mice and dramatically impaired in Crebbp^HETEp300^KO mice, as compared to both WT and single EP300^KO mice. These data suggest that GC B cells require a minimum amount of acetyltransferase activity, and reveal a potential therapeutically exploitable dependency of Crebbp-mutated GC B cells on Ep300.

In order to probe if a similar dependency exists in neoplastic GC B cells, we used an inducible CRISPR/Cas9 system to delete EP300 (or a control non-genic region) in 4 DLBCL cell lines representative of the various CREBBP genotypes found in DLBCL, and monitored cell proliferation and survival in competition assays over 12 days. Compared to CREBBP^WT, CREBBP heterozygous and homozygous mutant cells were significantly counter-selected from the total population following doxycycline induced EP300 deletion (~30% at day 7). Moreover, no EP300-edited clones were recovered from the CREBBP mutant lines in single cell plating assays, compared to CREBBP WT (p<0.01). Thus, DLBCL cells remain addicted to the residual EP300 aceyltransferase activity, supporting the existence of a therapeutic window for EP300 inhibitors.

To explore this concept further, we generated isogenic DLBCL clones carrying WT or defective CREBBP alleles (n=4 each), and performed drug-sensitivity assays with 2 novel small molecule inhibitors that specifically target the CREBBP/EP300 HAT or BRD domain. While, at higher doses, both inhibitors interfered with cell growth in all clones, CREBBP^KO cells were significantly more sensitive than their isogenic WT pairsat low nanomolar ranges (IC50: 60nM vs 300nM). Importantly, we were able to design an in vitro protocol that was toxic to CREBBP^KO cells but tolerated by CREBBP^WT cells, providing a proof of concept for therapeutically targeting these molecules.

In conclusion, we show that CREBBP and EP300 have differential roles in normal GC B cell development and that CREBBP mutated cells are addicted to the residual EP300 activity. This dependency is maintained in DLBCL cells, providing the basis for the potential application of acetyl transferase inhibition into the clinical settings.