OP0097 K4ME3-DOMINANT BIVALENT CHROMATIN REGIONS ASSEMBLE SURVIVIN WITH CHROMATIN REMODELLING COMPLEXES IN CD4 CELLS

Background

Rheumatoid arthritis (RA) is associated with aberrant epigenomic control which obstructs the proper function of CD4 T cells. Bivalent chromatin regions (BvCR), containing histone H3 tail trimethylation at lysine 4 (H3K4me3) and lysine 27 (H3K27me3), act as a major organizer of the 3D nuclear environment required for efficient T cell development. Survivin has gained importance in RA pathogenesis. Survivin is shown to interact with chromatin genome-wide and regulate gene transcription in the IFNg producing CD4 T cells [1].

Objectives

This study investigates survivin cooperation with nucleosome in BvCR and functional effect of this cooperation on CD4 cell phenotype.

Methods

BvCR were identified in CD4 cells by annotation of H3K4me3, H3K27me3 and H3K27ac binding through chromatin immunoprecipitation and sequencing (ChIPseq). Protein binding was quantified through peak score. An overlap of the H3 peaks and Survivin ChIPseq (12 individual CD4 cultures) by at least one bp was identified through R package ChIPPeakAnno. Changes in H3 peaks in CD4 cells treated with a survivin inhibitor YM155 were assessed. Tag normalization was applied to enable comparison between different ChIPseq. Tag prevalence within the BvCR defined the dominant H3 group. Changeable BvCR were defined by alteration in the dominant-H3 group after YM155 treatment. TF ChIPseq colocalization analysis was performed through ReMap2022 database. To analyse effect of BvCR changeability on transcriptome, we performed RNAseq of YM155-treated CD4 cells (n=4). Overlap of BvCR with regulatory elements (RE) was done through GeneHancer database. TF enrichment analysis was performed through GSEA (Broad Institute).

Results

Integration of H3K4me3, H3K27me3 and H3K27ac ChIP-seq identified 6199 BvCR, 66% of those contained survivin peaks. Fluorescence imaging of effector T cells validated nuclear colocalization of survivin with histone proteins. Of those 6199 BvCR, 43.8% were dominated by H3K4me3, which had also the highest survivin peak scores. Complexes of histone modifiers COMPASS, responsible for H3K4 methylation, and SWI/SNF, responsible for counteracting PRC2-guided H3K27me3, colocalized with BvCR. YM155-treatment significantly increased deposition of both H3K4me3 and H3K27me3. Changeable BvCR frequently acquired a K4me3-dominant status after YM155 treatment. This indicates that survivin contributes to regulation of methylation of lysines in H3 tail within BvCR. TF enrichment analysis of genes connected to the BvCR revealed the role of histone demethylases KDM7A (FDR=1.03e-32), PHF2 (FDR=3.47e-19), KDM5D (FDR=1.51e-15) in their transcription. Strongest overall enrichment was identified in H3K4me3-dominant BvCR colocalized with survivin. Integration of BvCR with CD4 RNA-seq revealed 2469 protein-coding genes connected through RE to H3K4me3-dominant BvCR, which sustained transcriptional activity and 346 of those were differentially expressed (DEG, nom p<0.05) in YM155-treated cells. DEG transcriptionally dependent on the TF families of histone modifiers built a strong SWI/SNF-dependent interaction network acting in biological processes of immune system regulation (FDR=1.07e-6. FOXP1, KLF3, FYB1, BCL2), DNA repair (FDR=5.57e-6. MSH6, MGMT, MCM8, UPF1), and RNA polymerase transcription (FDR=1.22e-8. TCF3, ARID5A, RARA). CD4 cell transcriptomics of randomly selected RA patients demonstrated that a) DEG controlled by the H3K4me3-dominant BvCR were upregulated in survivin-high CD4 cells with proinflammatory Th1 phenotype b) DEG had altered expression in CD4 cells of patients treated with JAK-inhibitors.

Conclusion

This study demonstrates that survivin contributes to regulation of lysine methylation in H3 tail within BvCR, which coordinates activity of the SWI/SNF complex in chromatin remodelling and transcription. We demonstrate that survivin inhibition affects this coordinated activity, and thereby impinge on CD4 activation in RA.

Reference

[1]Erlandsson et al. iScience 2022;25:105526.

Acknowledgements:

NIL.

Disclosure of Interests

None Declared.