BRG1-SWI/SNF-dependent regulation of the Wt1 transcriptional landscape mediates epicardial activity during heart development and disease

Epicardium-derived cells (EPDCs) contribute cardiovascular cell types during development and in adulthood respond to Thymosin β4 (Tβ4) and myocardial infarction (MI) by reactivating a fetal gene programme to promote neovascularization and cardiomyogenesis. The mechanism for epicardial gene (re-)activation remains elusive. Here we reveal that BRG1, the essential ATPase subunit of the SWI/SNF chromatin–remodelling complex, is required for expression of Wilms’ tumour 1 (Wt1), fetal EPDC activation and subsequent differentiation into coronary smooth muscle, and restores Wt1 activity upon MI. BRG1 physically interacts with Tβ4 and is recruited by CCAAT/enhancer-binding protein β (C/EBPβ) to discrete regulatory elements in the Wt1 locus. BRG1-Tβ4 co-operative binding promotes optimal transcription of Wt1 as the master regulator of embryonic EPDCs. Moreover, chromatin immunoprecipitation-sequencing reveals BRG1 binding at further key loci suggesting SWI/SNF activity across the fetal epicardial gene programme. These findings reveal essential functions for chromatin–remodelling in the activation of EPDCs during cardiovascular development and repair.

results are presented as fold enrichment over input, whereas re-ChIP results are present in fold enrichment over the level of ChIP with negative control IgG antibody. All error bars are data ± s.d. Significant differences (P values) were calculated using two-tailed Student's t-test (*P ≤ 0.05; **P ≤ 0.01).

Supplementary Figure 4. Wt1 ECRs drive gene expression in the developing embryo.
(a-c) -42bp ECR-driven LacZ expression (n=2 embryos) recapitulates Wt1 pattern of expression in the developing embryo at E11.5, comprising domains in the spinal cord and brain (arrows in a), gut mesothelium (arrow in b) and epicardium (arrowheads in c). (d-f) +4.0kb ECR-driven LacZ expression (n=5 embryos) is restricted to the developing epicardium (arrowheads in e and f). (g-l) +5.8kb ECR-driven LacZ expression (n=6 embryos) is mostly restricted to the developing epicardium (5 out 6 embryos; arrowheads in h and i), but can also be detected in the urogenital region (1 out 6 embryos; arrow in l), where Wt1 plays an essential role during development. Asterisk in k indicates lack of β-gal activity in the epicardium. 2-6 transgenic mice with independent ECR integration were examined to assess the reproducibility of any given reporter activity pattern. Scale bars: a,d,g,j 1mm, b,e,h,k 500µm and c,f,i,l 160µm.
in the developing embryo.
(a-i) Whole-embryo in situ hybridization for Wt1 mRNA in the developing embryo at E11.5, showing expression in the limb buds (arrows), head and spinal cord domains (arrows) and epicardial layer covering the ventricles and atria (arrowheads). (j-l) -42bp ECR-driven LacZ expression recapitulates Wt1 pattern of expression in the developing embryo at E11.5, comprising domains in the spinal cord and brain (arrows) and epicardium covering ventricles and outflow tract (arrowheads). Similarities in the spatial pattern of expression of the Wt1 mRNA and -42bp ECR-driven LacZ expression support the role of this ECR as the Wt1 core promoter. Scale bars: a,d,g,j 1mm, b,c,e,f,h,k 500µm, i 160µm and j 100µm.  ) in a-d,h-j were calculated using two-tailed Student's t-test, whilst a Chi-square test (χ 2 ) was performed in g to evaluate statistically significant differences between the observed and expected (25%; grey line in g) numbers of Gata5 cre ;Brg1 F/F mutant embryos, assuming Mendelian inherence of a single copy number of Cre (*P ≤ 0.05; **P ≤ 0.01). co, control; ctrl, control; ep, epicardium; lv, left ventricle; mut, mutant; my, myocardium; rv, right ventricle. All scale bars 100µm, except k-n 1mm.  note that these reporter assays are heterologous system, utilising episomal vectors, as opposed to studying naked plasmid DNA. Most luciferase reporter assays focus on the latter, and the expectation is to observe significant fold-changes in expression with addition of the factors being tested. Here the effect of chromatin-remodelling on Wt1 ECRs is probed and the system is at the limit of detection given the need for synthesis/packaging of chromatin-like DNA (via the episomal vectors) and appropriate expression levels of the reporters (firefly luciferase in pREP4 and TK-Renilla luciferase in pREP7), remodellers (Brg1 or Brm) and cofactors (Tβ4). Despite these technical limitations, relative increases with addition of Brg1/Brm and Tβ4 combined were observed, as compared to Brg1/Brm alone across the three ECRs, providing confidence-in-the-data as it stands. All error bars are data ± s.d.

Supplementary
Significant differences (P values) were calculated with one-way ANOVA followed up by the Tukey multiple comparison test (*P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001; ****P ≤ 0.0001). r.l.u., relative luciferase unit, after normalization to Renilla. showing comparatively more peaks in the Tβ4-primed, -injured adult heart samples, compared to the non-primed (PBS), -injured adult heart samples. Please note that zoomed-in views are shown for the Wt1 (b) and Myh7 (e) loci in order to highlight previously described BRG1-bound regulatory elements located within 5kb of the transcriptional start site (Myh7 1 ) or intron 1 (Wt1; this study).