TEAD4 antagonizes cellular senescence by remodeling chromatin accessibility at enhancer regions

Dramatic alterations in epigenetic landscapes are known to impact genome accessibility and transcription. Extensive evidence demonstrates that senescent cells undergo significant changes in chromatin structure; however, the mechanisms underlying the crosstalk between epigenetic parameters and gene expression profiles have not been fully elucidated. In the present study, we delineate the genome-wide redistribution of accessible chromatin regions that lead to broad transcriptome effects during senescence. We report that distinct senescence-activated accessibility regions (SAAs) are always distributed in H3K27ac-occupied enhancer regions, where they are responsible for elevated flanking senescence-associated secretory phenotype (SASP) expression and aberrant cellular signaling relevant to SASP secretion. Mechanistically, a single transcription factor, TEAD4, moves away from H3K27ac-labled SAAs to allow for prominent chromatin accessibility reconstruction during senescence. The enhanced SAAs signal driven by TEAD4 suppression subsequently induces a robust increase in the expression of adjacent SASP genes and the secretion of downstream factors, which contribute to the progression of senescence. Our findings illustrate a dynamic landscape of chromatin accessibility following senescence entry, and further reveal an insightful function for TEAD4 in regulating the broad chromatin state that modulates the overall transcriptional program of SASP genes. Electronic supplementary material The online version of this article (10.1007/s00018-023-04980-9) contains supplementary material, which is available to authorized users.


Fig. S1
Fig. S1 Creation of an in vitro replicative senescence model.A Population doubling of primary human umbilical cord mesenchymal stem cells (UC-MSCs).B Percentage of SA-β-gal staining in young to senescent UC-MSCs.Error bars indicate the mean ± S.E.M. of three independently performed experiments.***p < 0.001.A Student's t-test was used for statistical analysis.C Immunofluorescence staining and statistics of Ki-67 in UC-MSCs.The nucleolus is indicated by DAPI staining.Scale bar, 10 μm.Error bars indicate the mean ± S.E.M. of three independently performed experiments.***p < 0.001.D Top: Immunoblotting of P16 expression in young to senescent UC-MSCs.βactin and ponceaus S were used as the loading controls.Bottom: CDKN2A expression in PD2, PD26, and PD42 UC-MSCs.The dashed line indicates RT-qPCR, while the solid line indicates RNA-seq data.Error bars indicate the mean ± S.E.M. of three independently performed experiments.E and F Cytokine array analysis of secreted proteins and relative quantitation of SASP factors during senescence.Error bars indicate the mean ± S.E.M of three independently performed experiments.*p < 0.05, **p < 0.01, ***p < 0.001.A Student's t-test was used for statistical analysis.

Fig. S2
Fig. S2 Exploration of genome-wide transcriptional programs upon senescence entry.A Correlation heatmap of RNA-seq data.There are two independently performed biological replicates for each passage number (rep 1 and 2).B Venn diagram indicating the number of differentially expressed genes in in young to senescent UC-MSCs.C Heatmap of RNA-seq data showing hierarchically clustered gene expression in young to senescent UC-MSCs.D and E GSEA showing the enrichment of GO (D) and RECTOME (E) pathways in PD2 versus PD42 cells.

Fig. S3
Fig. S3 Remodeling of the chromatin landscape occurs during senescence.A Principal Component Analysis (PCA) displaying the correlation and dispersion of ATAC-seq in young to senescent UC-MSCs.There were two independently performed biological replicates for each passage number (rep 1 and 2).B Representative distribution of insert size, showing clear signal modulation for mono-and di-nucleosomes.C Normalized ATAC-seq read count across the genome.Following peak calling, the mean ATAC-seq peaks present between the TSS (3 kb) and the TES (+3 kb) were calculated across the genome.TSS, transcription start site; TES, transcription end site.D Heatmap and enrichment plots showing normalized read densities for ATAC-seq peaks at PD2 and PD42.Tracks are centered at the peaks and extend ± 3 kb.E Genomic feature distributions of accessible chromatin regions that changed during senescence.UTR, untranslated region.

Fig. S4
Fig. S4 ATAC-seq data are highly similar to public Encyclopedia of DNA Elements (ENCODE) data for enhancers decorated by histone modification H3K27ac.A and B Integrative Genomics Viewer (IGV) snapshot displaying the H3K4me3 and H3K27ac peaks of MSCs from ENCODE, in addition to our ATAC-seq and RNA-seq at the IL6 and IL8 loci in young to senescent UC-MSCs.Vertical gray boxes indicate enhancer and promoter ATAC-seq peaks.Chromatin states were obtained from ENCODE: yellow, weak enhancer; orange, strong enhancer; red, active promoter; green, transcribed region; and gray, heterochromatin.

Fig. S5
Fig. S5 Enrichment analysis of differentially accessible regions during senescence.A Insertion tracks of senescence-activated accessibility regions (SAAs) and senescenceinactivated accessibility regions (SIAs) at loci on chromosomes 7 and 9. B Distribution of the distances between SAA or SIA ATAC-seq peaks and the TSS.C Annotations of SAAs and SIAs showing genomic features of the differentially accessible regions.D

Fig. S6
Fig. S6 Comparison between DNase-seq and ATAC-seq data from ENCODE and other types of mesenchymal stem cells and our chromatin accessibility data.A and B Integrative Genomics Viewer (IGV) snapshot displaying the DNase-seq and ATACseq peaks of AMSCs and hMSCs from ENCODE and the ATAC-seq peaks in young to senescent UC-MSCs at loci on chromosomes 2 and 6.

Fig. S8
Fig. S8 TEAD4 binding at SAAs accelerates the transcription of SASP genes.A Topranked enriched motifs from MEME motif analysis of SAAs and SIAs.The circle size represents the percentage of regions with the motif and the color represents the p value (shrink to 10×).B ATAC-seq pattern in the bZIP motif at PD2 and PD42.Normalization was performed to ensure that each motif possessed the same mean number of insertions 200-500 bp away.C Expression of other TEAD family members, TEAD1, TEAD2, and TEAD3 in young to senescent UC-MSCs.The dashed line indicates RT-qPCR, while the solid line indicates RNA-seq data.Error bars indicate the mean ± S.E.M. of three independently performed experiments.D Heatmaps and enrichment plots showing normalized read densities for TEAD4 CUT&Tag peaks performed at PD2 and PD42.Tracks are centered at the peaks and extend ± 3 kb.E Venn diagram showing overlap of TEAD4 CUT&Tag peaks between PD42 and PD2, in addition to SAA signals defined in ATAC-seq peaks.

Fig. S9
Fig. S9 Suppression of TEAD4 leads to the upregulation of SAA-adjacent SASP genes.A Relative density of immunoblotting showing the expression of P16 following knockdown or overexpression of TEAD4.β-actin served as controls.Error bars indicate the mean ± S.E.M. of three independently performed experiments.**p < 0.01, ***p < 0.001.B Immunofluorescence statistics of the expression of Ki-67 following knockdown or overexpression of TEAD4.Error bars indicate the mean ± S.E.M. of three independently performed experiments.**p < 0.01, ***p < 0.001.C Percentage of SA-β-gal staining following TEAD4 suppression or overexpression.Error bars indicate the mean ± S.E.M. of three independently performed experiments.**p < 0.01, ***p < 0.001.A Student's t-test was used for statistical analysis.D RNA-Seq scatter diagram showing the gene expression profiles of control versus siTEAD4 cells.Representative SASP genes are labeled.E RT-qPCR data showing the relative expression of SASP genes following downregulation of TEAD4 expression.Error bars indicate the mean ± S.E.M. of three independently performed experiments.*p < 0.05, **p < 0.01, ***p < 0.001.One-way ANOVA followed by Dunnett's multiple

Fig. S11
Fig. S11 YAP/TAZ are not involved in the regulation of TEAD4-mediated SAAs flanking SASP genes.A Lysates from PD2 were subjected to immunoprecipitation (IP) with an anti-YAP antibody and subsequent immunoblotting with the indicated antibodies.B Lysates of HEK293T cells co-overexpressing pcDNA3.1-YAP-Flagand pSIN-TEAD4-GFP were subjected to IP and subsequent immunoblotting with anti-Flag