Introduction: Changes in the transcriptome of the heart orchestrate cardiac remodeling, hence counteracting these processes is necessary to prevent heart failure progression. Identifying maladaptive transcriptional regulators is therefore of high clinical relevance. The transcription factor SOX9 guides mesodermal differentiation during embryogenesis, but recent data point towards an additional role in maladaptive cardiac remodeling. While inactivation of cardiomyocyte Sox9 was shown to delay cardiac hypertrophy and fibrosis, the functional role and the molecular mechanisms are less clear.

Methods and Results: SOX9 protein level were increased in cardiomyocytes after TGF-β stimulation or co-culture with activated fibroblasts. To investigate the functional role of cardiomyocyte SOX9, previously described mice with loxP sites flanking exon 2-3 of Sox9 or mice with loxP sites around transgenic insertion of Sox9 at the genomic Hprt locus were interbred with mice expressing cre recombinase controlled by the Myh6 promoter to obtain cardiomyocyte specific deletion (Sox9-CM-KO) or overexpression (Sox9-CM-OE) of Sox9. While Sox9-CM-OE mice had significantly aggravated heart failure (EF 24.5% vs 38.4%, p<0.001) and enhanced mortality along with histological signs of increased cardiac remodeling after 9 weeks of transverse aortic constriction (TAC) compared to controls, Sox9-CM-KO mice were protected from TAC induced left ventricular dysfunction (EF 42.5 % vs 32.3%, p<0.05). Surprisingly, we found no changes in cardiac fibrosis or hypertrophy in Sox9-CM-KO mice after TAC compared to controls. We investigated the transcriptional role of SOX9 in cardiomyocytes using a multi-OMICS approach. RNA-sequencing after adenoviral overexpression of Sox9 in vitro revealed downregulation of key genes for cardiomyocyte contraction and excitation (e.g. Myh6, Adcy5, Gja1, Hcn4), while upregulated genes were associated with cell adhesion (e.g. Ctnnd1, Itgb1) and stress fiber assembly (e.g. Pxn, Rhoc). ChIP-sequencing identified 10.901 distinct genes with enrichment of SOX9 and comparison with differentially expressed genes from RNA-sequencing showed significant overlaps, indicating direct control of gene expression by SOX9. We additionally investigated the SOX9 interactome in cardiomyocytes using SOX9 immunoprecipitation (IP) followed by mass spectrometry. Interestingly, we found strong enrichment of proteins from chromatin remodeling complexes BAF (e.g. BRG1, BAF60, BAF250) and NuRD (e.g. CHD4, HDAC2) in SOX9-IP samples compared to controls. Importantly, we confirmed direct interaction of SOX9 with BRG1 and HDAC2 after co-expression in HEK cells. Furthermore, downregulation of Brg1 or Chd4 with siRNA in vitro partially restored expression level of Sox9 target genes in cardiomyocytes.

Conclusion: Fibrotic signaling enhances cardiomyocyte SOX9 and upregulation or downregulation of Sox9 in cardiomyocytes aggravates or ameliorates TAC induced heart failure, respectively. Increased SOX9 causes downregulation of key genes for cardiomyocyte function, while raising expression of cell-adhesion and stress fiber assembly genes. Moreover, SOX9 interacts with chromatin remodeling complexes and inhibition of BAF- or NuRD complex core units partially restores gene expression. Therefore, therapeutic inactivation of SOX9 in the heart could not only serve to prevent fibrotic remodeling, but also to protect the distinct cardiomyocyte transcriptome from fibrosis induced disruption.