MEF2B Instructs Germinal Center Development and Acts as an Oncogene in B Cell Lymphomagenesis

Abstract

The gene encoding the MEF2B transcription factor is mutated in germinal center (GC)-derived B cell lymphomas, but its role in GC development and lymphomagenesis is unknown. We demonstrate that Mef2b deletion reduces GC formation in mice and identify MEF2B transcriptional targets in GC, with roles in cell proliferation, apoptosis, GC confinement, and differentiation. The most common lymphoma-associated MEF2B mutant (MEF2B^D83V) is hypomorphic, yet escapes binding and negative regulation by components of the HUCA complex and class IIa HDACs. Mef2b^D83V expression in mice leads to GC enlargement and lymphoma development, a phenotype that becomes fully penetrant in combination with BCL2 de-regulation, an event associated with human MEF2B mutations. These results identify MEF2B as a critical GC regulator and a driver oncogene in lymphomagenesis.

Keywords: B cell; MEF2B; germinal center; lymphoma; mouse model.

Figure 1.. GC-specific deletion of Mef2b reduces GC formation and DZ/LZ ratio.

(A) Percentage of splenic GC B cells in Mef2b^+/+;Cγ1^cre/+ (WT/WT), Mef2b^fl/+;Cγ1^cre/+ (-/WT) and Mef2b^fl/fl;Cγ1^cre/+ (−/−) littermates, as measured by flow cytometry 10 days post-SRBC immunization. (B) Average GC size and numbers measured by pixels (pix) quantitation of immunofluorescence staining for the GC marker PNA on FFPE spleen sections. Analysis was performed on a subset of the mice displayed in panel A. (C) Immunohistochemistry for PNA and BCL6 on representative spleen sections. Scale bar, 2000 μm. (D) BCL6 mean fluorescence intensity (MFI) in GC B cells (B220⁺/PNA^hi/CD95^hi) measured by flow cytometry analysis and displayed as fold change relative to WT mice (Mann-Whitney test, *p≤0.05). (E) DZ (CXCR4^hi/CD86^lo)/LZ (CXCR4^lo/CD86^hi) ratios in GC B cells (B220⁺/PNA^hi/CD95^hi). (F) Gene Set Enrichment Analysis (GSEA) in Mef2b^+/+;Cγ1^cre/+ (WT/WT) and Mef2b^fl/fl;Cγ1^cre/+ (−/−) GC B cell RNA-seq data using a previously reported mouse DZ vs LZ gene signature (Victora et al., 2012). NES, Normalized Enrichment Score. (G) DZ and LZ B cell numbers normalized by spleen weight in a subset of the mice displayed in panel E. Dot plots and bar graph display average ± SD. P values in dot plots were determined by one-way ANOVA with Tukey’s multiple comparison test (^***p≤0.001, ^**p≤0.01, *p≤0.05). See also Figures S1 and S2 and Table S1.

Figure 2.. MEF2B genome-wide DNA binding in human GC B cells.

(A) MEF2B peak distribution relative to the closest transcription start site (TSS). Promoters are defined as the regions encompassing the TSS (−2/+1 kb) and classified based on their chromatin marks as active (H3K4me3⁺, H3K27me3^– and H3K27Ac⁺), weak (H3K4me3⁺, H3K27me3^– and H3K27Ac^– ), poised (H3K4me3⁺ and H3K27me3⁺), or inactive (H3K4me3^–). Intragenic and intergenic regions are classified based on enhancer-associated chromatin marks as active (H3K4me1⁺, H3K4me3^–, and H3K27Ac⁺) or poised (H3K4me1⁺, H3K4me3^– and H3K27Ac^–), while considered inactive when lacking all marks (H3K4me1^–, H3K4me3^– and H3K27Ac^–). The term “other” refers to regions that do not fulfill the above classification criteria. (B) Distribution of reads (as reads per million, RPM) for the indicated histone marks, relative to the MEF2B peaks. The displayed histone modifications are the same used in panel A to classify bound regions in promoters (H3K4me3, H3K27me3 and H3K27Ac) or intra/intergenic regions (H3K4me1, H3K4me3 and H3K27Ac). (C) Percentage of MEF2B peaks that are located in enhancers and super-enhancers, as identified by the ROSE algorithm applied to H3K27Ac data obtained from normal GC B cells. “Other” denotes all remaining regions. (D) Percentage of GC enhancers and super-enhancers that are occupied by MEF2B. See also Tables S2 and S3.

Figure 3.. MEF2B-driven transcriptional network in GC B cells.

(A) Flowchart detailing the main steps leading to the identification of MEF2B targets as bound in regulatory regions, transcriptionally regulated during the human GC transition, and affected by Mef2b deletion in mouse GC B cells. (B) Schematic representation of the main pathways affected by MEF2B as obtained by integrating manual annotations of the genes identified in panel A and the results of pathway enrichment analysis performed using the DAVID 6.7 tool and KEGG, Biocarta, Panther and Reactome databases. The genes included among the 141 up-regulated or 109 down-regulated core targets are displayed in red and blue, respectively. See also Tables S4.

Figure 4.. MEF2B^D83V displays hypomorphic features including reduced DNA binding and protein stability.

(A-B) Venn diagram illustrating the overlap between DNA regions bound by MEF2B^WT or MEF2B^D83V mutant, as identified by ChIP-sequencing analysis with anti-MEF2B or anti-FLAG and –HA (B) antibodies in SUDHL10-FLAG-HA-MEF2B^WT or -MEF2B^D83V cell lines. Motif logos summarize the MEF2 DNA motifs found in the regions bound by MEF2B^WT or MEF2B^D83V. (C) MEF2B^WT and MEF2B^D83V protein half-life as measured in SUDHL10 or SUDHL4 cell lines expressing FLAG-HA-MEF2B^WT or -MEF2B^D83V isoform A or isoform B upon treatment with cycloheximide. Data are displayed as average ± SD of at least two independent experiments. See also Figure S3 and Table S5.

Figure 5.. MEF2B^D83V mutant escapes binding to and repression by the HUCA complex and HDACs class IIa.

(A) Immunoblot analysis of MEF2B interactors upon co- immunoprecipitation (IP) with anti-FLAG antibody in SUDHL10-FLAG-HA-MEF2B^WT or - MEF2B^D83V cells. Input is 10% of total lysate used for IP. EV, Empty Vector. (B-C) Immunoblot analyses of normal human GC B cells (B) and U2932 (MEF2B^WT/WT), SUDHL4 (MEF2B^WT/D83V), SUDHL10 (MEF2B^WT/WT), DB (MEF2B^WT/D83V) DLBCL cell lines (C) before (input; 10% of total sample) and after IP with anti-MEF2B antibody or IgG. Numbers under immunoblots represent IP enrichment relative to the input, as measured by densitometry. (D) Luciferase activity driven by synthetic wild-type (WT) or mutated (Mutant) MEF2 binding sites (3xMEF2-luc), as measured in HEK293T cells upon transfection with EV or plasmids expressing MEF2B^WT, MEF2B^D83V or MEF2B^ΔMADSMEF (ΔΔ) mutants alone or with CABIN1^WT, CABIN1^L2172A, HDAC5 or HDAC7. Data are shown as average ± SD of two independent experiments relative to EV control. Significant differences were assessed by unpaired two- tailed t-test (MEF2B^WT- or MEF2B^D83V-driven inductions vs EV control p≤0.001; CABIN1, HDAC5 or HDAC7-mediated repressions of MEF2B^WT-driven induction p≤0.01). See also Figure S4.

Figure 6.. The Mef2b^D83V mutant allele acts dominantly in GC formation.

(A) Percentage of splenic GC B cells (PNA^hi/CD95^hi) in Mef2b^+/+;Cγ1^cre/+ (WT/WT), Mef2b^fl/+;Cγ1^cre/+ (-/WT), Mef2b^+/stopD83V;Cγ1^cre/+ (WT/D83V) and Mef2b^fl/stopD83V;Cγ1^cre/+ (-/D83V) mice, as measured by flow cytometry 10 days post-SRBC immunization. (B) Average GC size and numbers as measured by pixels (pix) quantitation of immunofluorescence staining for the GC marker PNA on FFPE spleen sections in a subset of the mice displayed in panel A. (C) Immunohistochemistry for PNA on representative spleen sections. Scale bar, 2000 μm. (D) DZ (CXCR4^hi/CD86^lo)/LZ (CXCR4^lo/CD86^hi) ratios in GC B cells (B220⁺/PNA^hi/CD95^hi). (E) Gene Set Enrichment Analysis (GSEA) in Mef2b^+/+;Cγ1^cre/+ (WT/WT) and Mef2b^+/stopD83V;Cγ1^cre/+ (WT/D83V) GC B cell RNA-seq data using a previously reported mouse DZ vs LZ gene signature (Victora et al., 2012). NES, Normalized Enrichment Score. (F) DZ and LZ B cell numbers normalized by spleen weight in a subset of the mice displayed in panel D (Mann-Whitney test, *p≤0.05). (G) GSEA of the MEF2B core targets (as identified in Figure 3A) in the GC B cell transcriptional profiles of Mef2b^+/+;Cγ1^cre/+ (WT/WT), Mef2b^+/stopD83V;Cγ1^cre/+ (WT/D83V) and Mef2b^fl/stopD83V;Cγ1^cre/+ (-/D83V) mice. (H) Schematic representation of the results of the pathway enrichment analysis performed on RNA-seq profiles of GC B cells from Mef2b^+/+;Cγ1^cre/+ (WT/WT) compared to Mef2b^+/stopD83V;Cγ1^cre/+ (WT/D83V) and Mef2b^fl/stopD83V;Cγ1^cre/+ (-/D83V) 4 mice each, using GSEA and MSigDB and SignatureDB databases.Dot plots display average ± SD. P values in dot plots were determined by one-way ANOVA with Tukey’s multiple comparison test (^***p≤0.001, ^**p≤0.01, *p≤0.05), unless otherwise specified. See also Figure S5 and Tables S6-S8.

Figure 7.. MEF2B^D83V expression in mice leads to the development of B cell lymphomas.

(A) Percentage of splenic GC B cells (PNA^hi/CD95^hi) in Mef2b^+/+;CD21^cre/+ (WT/WT), Mef2b^+/stopD83V;CD21^cre/+ (WT/D83V),Mef2b^+/+;CD21^cre/+;Bcl2^tg (WT/WT;Bcl2), Mef2b^+/stopD83V;CD21^cre/+;Bcl2^tg (WT/D83V; Bcl2) mice, as measured by flow cytometric analysis in 3-month-old mice 10 days after a single SRBC immunization. Data are displayed as dot plots, including average ± SD. ^***p≤0.001, *p≤0.05 (one-way ANOVA with Tukey’s multiple comparison test). (B) Frequency of development of B cell lymphomas in the animals with the indicated genotypes. Colored bars indicate different diagnosis, namely Follicular Lymphomas (FL), Diffuse Large B Cell Lymphomas (DLBCL) and Plasmablastic lymphoma (PBL). ^**p≤0.01, *p≤0.05 (Fisher’s Exact test). (C) Event-free survival curve of mice reported in panel (B) (Gehan-Breslow-Wilcoxon test). (D) Hematoxylin and eosin (H&E) staining and immunohistochemical analysis of representative tumors from mice diagnosed with FL, DLBCL or PBL.See also Figure S6.