Dissecting chromatin architecture for novel cardiovascular disease targets

summary:

Unraveling the principles of cardiac chromatin organization for next generation therapies in heart disease

Conserved gene expression changes in response to myocardial stress suggest that epigenetic regulatory mechanisms integrate disparate heart failure signaling pathways arising from various genetic and environmental triggers (Figure 1). The protein families that write, erase and read chromatin marks have been shown to precisely regulate gene expression during cardiac disease, yet these actions have, until recently, been studied only in the context of the linear genome. In vivo, however, chromatin folds into the confines of the nucleus through an architecture that regulates gene expression by facilitating physiologically relevant interactions between enhancers and gene promoters, while insulating genes to prevent aberrant enhancer interactions. The principles for the folding of chromatin have begun to emerge from non-cardiovascular cell types and include the formation of chromatin loops and topologically associating domains (TADs), which are regions of high local interaction. These folding principles are thought to segment active, euchromatic regions apart from inactive, heterochromatic regions, enabling the formation of specialized transcriptional factories, such as those involved in mRNA splicing.

Figure 1.

Epigenomes as integrators of genetic and environmental cues during development and disease.

CTCF is a ubiquitously expressed zinc finger DNA binding protein that is involved in gene regulation through its actions forming semi-stable loops within, and potentially between, chromosomes. Key questions remain unanswered with regard to the role of CTCF in cell type specific genome packaging (and in cellular phenotype; the germline knockout is embryonic lethal) as well as with regard to how cardiac chromatin architecture relates to cardiovascular homeostasis. Examining chromatin architecture in the heart should answer fundamental questions in epigenomics and cardiovascular pathophysiology, such as: Which genes exist in the same physical neighborhood in the cardiac nucleus? And conversely, which are insulated from each other? Which genes share the same distal enhancers? How are distal enhancers engaged during the myocardial stress response?

In light of this, our labs have undertaken studies of cardiac chromatin organization in the context of cardiovascular disease ^{1, 2}. The approach to this question involved using HiC technology to examine chromatin organization, in a pressure overload model (transverse aortic constriction) and a genetic model (cardiac specific Ctcf knockout) in the adult mouse. Adult cardiomyocytes are terminally differentiated and very rarely proliferate, meaning that the effects of Ctcf perturbation on chromatin and gene expression can be examined without accounting for cell division. In both studies, transcriptomic changes in TAC were characteristically associated with stress response gene expression and changes in histone acetylation enrichment both at promoters and distal enhancer regions—however, the overall genome-wide 3D chromatin architecture, whether measured by TADs or active/inactive compartmentalization, remained relatively stable. Importantly, these findings are unlike the more pronounced changes of active/inactive compartments, and to a lesser degree TAD changes, evident during cardiac development ³. Analysis of the gene-enhancer interactome in the TAC model also unraveled the spectrum of regulatory interactions necessary for baseline cardiac gene expression and for the myocardial stress response gene program. Indeed these studies ^1–3 now provide the cardiovascular community with a rich epigenomic resource, which when coupled with publicly available genome-wide maps of cardiac transcription factor occupancy from ChIP-seq, and transcriptional data from RNA-seq, will accelerate our understanding of the epigenetic control of cardiac development and disease.

Although CTCF expression is not significantly altered in TAC, the depletion of CTCF leads to dire cardiac pathology,^{1, 2} in agreement with its necessary role in cardiovascular development⁴. The effect of CTCF depletion on chromatin architecture has been a topic of debate in the field of chromatin biology, with different studies reporting varying degrees of chromatin architecture instability following CTCF depletion. A seminal study, using the degron-system allowing for controlled CTCF protein degradation, resolved this debate by showing that CTCF depletion needs to be nearly complete to elicit the most substantial architecture defects ⁵. Indeed, these investigators showed that since many insulator boundaries are enriched for CTCF, some residual CTCF protein (even as little as 15%), is sufficient to preserve some structures. This observation was reflected in the recent studies in cardiomyocytes, wherein greater CTCF depletion leads to greater disruption of chromatin features, including loops and TADs. An open question from these studies regards the relationship between chromatin disruption as a primary insult and the activation of compensatory stress response pathways as a secondary insult—both of which can lead to cardiomyocyte dysfunction at the cell level and heart failure at the organ level. Reduced CTCF gene expression or unique DNA mutations inferred to produce aberrant CTCF binding have been implicated in some studies of chromosomal syndromes, including intellectual disability and congenital heart disease; chromatin rearrangements in these affected individuals have not been investigated. What is clear from studies to date is that disrupting chromatin structure by knocking out CTCF in myocytes causes heart failure in mice, supporting the hypothesis that stable chromatin architecture is necessary for cellular homeostasis.

How can we leverage these chromatin architecture maps to better understand disease and discovery therapeutic targets? A promising area unlocked by 3C and Hi-C technology is in the analysis of genome wide association studies (GWAS) datasets: many disease-associated single nucleotide polymorphisms (SNPs) reside in non-coding regions, often far from expressed genes. Chromatin architecture maps enable interrogation of how SNPs in non-coding loci may affect gene expression, perhaps by influencing enhancer-promoter interactions or rearrangement of chromatin boundaries, leading to aberrant gene interactions and expression. If the actions of non-coding SNPs can be mechanistically understood using chromatin architecture maps (i.e. identifying the genes whose expression is altered by physical interaction with the mutated region), novel drug targets could emerge. Another promising approach is the use of chromatin architecture and/or accessibility and a functional readout of chromatin modifying drugs, towards the engineering of a new class of drugs that modulate sub-epigenomes underpinning transcriptome remodeling events across the spectrum of heart failure phenotypes.