Dissecting Mechanistic Differences between EZH2 Gain-of-Function Events in Non-Hodgkin B Cell Lymphomas

Non-Hodgkin Lymphoma (NHL) is one of the most common cancers in the United States, with approximately 80,000 new cases and 20,000 deaths occurring each year. The two most common types of NHL, Diffuse Large B Cell Lymphoma (DLBCL) and Follicular Lymphoma (FL), arise from transformed mature B lymphocytes. The Germinal Center (GC) B cell subtype, a transient population that undergoes transcriptional changes to produce high-affinity antibodies in response to an antigen, requires tight epigenetic regulation to prevent transformation into DLBCL and FL; indeed, DLBCL and FL are characterized by mutations in epigenetic modifier genes. Up to 30% of GC-DLBCL and FL cases exhibit gain-of-function alterations in the histone methyltransferase EZH2, including a heterozygous hotspot mutation (EZH2^Y641F/+) or overexpression of wild-type protein (EZH2^OE). A major gap in knowledge is whether EZH2^Y641F/+ and EZH2^OE mediate lymphoma by shared or distinct mechanisms. EZH2^Y641F/+ and EZH2^OE have not been directly compared in any cancer context, and the molecular and phenotypic consequences of EZH2^OE in B cells are unknown. Therefore, our work directly compares EZH2^Y641F/+ and EZH2^OE in mouse B cells to understand how both events contribute to lymphoma development and to illuminate new strategies for targeting DLBCL and FL. We hypothesize that EZH2^Y641F/+ and EZH2^OE are non-equivalent gain-of-function events that drive lymphoma by distinct mechanisms.

We have established two separate and tissue-specific mouse models of EZH2^Y641F/+ and EZH2^OE that serve as the foundation of this work. We use conditional alleles for both Ezh2^Y641F/+ and Ezh2^OE that are activated upon Cre-mediated recombination. We crossed each conditional allele with CD19-Cre, which induces the alleles in the B lineage at the pro-B cell stage. To study normal B cell populations before lymphoma onset, we evaluated cells in the bone marrow, spleen, and peripheral blood of 3-month-old adult mice using flow cytometry. We observed minimal differences in the bone marrow and blood populations at this age. In the spleen, both CD19-Cre;Ezh2^Y641F/+ and CD19-Cre;Ezh2^OE cause an increase in follicular B cells compared to CD19-Cre controls. CD19-Cre;Ezh2^Y641F/+ also causes an increase in marginal zone B cells. In summary, each EZH2 gain-of-function event increases the B cell populations in the spleen by early adulthood in mice.

Because EZH2^Y641F/+ and EZH2^OE occur in GC B cell-derived lymphomas, we then investigated how these events impact GC cells specifically. We crossed the Ezh2^Y641F/+ and Ezh2^OE alleles to the GC-specific Cγ1-Cre. Using the two Cre cohorts together, we can model Ezh2^Y641F/+ and Ezh2^OE as both early and late events in the B lineage. To stimulate GC cells, we immunized adult mice across each CD19-Cre and Cγ1-Cre genotype with sheep red blood cells and analyzed their GC populations by flow cytometry. Intriguingly, early and late activation of Ezh2^Y641F/+ cause opposite impacts on GC expansion. CD19-Cre;Ezh2^Y641F/+ mice display a striking decrease in the GC population compared to CD19-Cre control, while Cγ1-Cre;Ezh2^Y641F/+ increases this population compared to Cγ1-Cre control. In contrast, neither early nor late activation of Ezh2^OE changes the overall GC population compared to Cre-only control. These results suggest that the timing of EZH2^Y641F/+ in the B lineage influences its activity and provide evidence that EZH2^Y641F/+ and EZH2^OE are functionally distinct events in the development of B cell lymphoma.

In summary, we modeled EZH2^Y641F/+ and EZH2^OE as early and late events in the B lineage and found that they both increase splenic B cell populations. In response to antigen stimulation, EZH2^Y641F/+ impacts overall GC expansion depending on the timing of EZH2^Y641F/+ in the B lineage. The results so far support the hypothesis that EZH2^Y641F/+ and EZH2^OE promote B cell lymphoma by distinct mechanisms. Ongoing progress includes investigating how these alterations impact gene expression and chromatin methylation patterns, using RNA-sequencing and ChIP-sequencing, respectively. By completing this work, we will elucidate the oncogenic mechanisms that distinguish EZH2^Y641F/+ and EZH2^OE in B cells through multiple angles and identify downstream targets that may lead to new therapeutic approaches in patients with these alterations.

No relevant conflicts of interest to declare.