Enhancing Pace: Identifying and Validating the Cis-Regulatory Landscape of the Sinoatrial Node

\Since the publication of the human genome, it is clear that a far larger proportion of our genome is comprised of regulatory, non-coding information rather than coding sequences¹. One goal to understanding this regulatory information has been the identification of the cis-regulatory elements (CREs) found in the genome and the cell types in which they function². However, it has proven challenging to identify cell-type specific CREs from rare cell populations. This is an important goal because many rare cell populations play critical roles in human health. In this issue, two independent teams present their work illuminating the CRE landscape of the small, but extremely important, population of sinoatrial node (SAN) cardiomyocytes in the heart for the first time^3,4. Their combined efforts further our understanding of the complex gene regulatory networks during the development of the SAN and lay the groundwork for identifying disease causing mutations, which specifically affect SAN function, as a method for diagnostic screening and early treatment of patients.

The SAN is the pacemaker of the heart and as such, has the important distinction of regulating heart rate, a process that occurs 60–100 times a minute for an entire lifetime. Simplistically, the SAN sets the pace through leaky sodium/potassium channels known collectively as the “funny-current” channels, notably HCN1 and HCN4⁵. The SAN triggers the surrounding atrial cardiomyocytes initiating atrial depolarization and the heartbeat. Insight into the cis-regulatory landscape of the SAN is imperative to understanding the complex interactions between transcription factors, gene targets, and functional electrophysiological outcomes of the SAN relevant to cardiac health.

Initially developing at the venous pole of the heart, the SAN forms from a genetically distinct lineage of cardiac progenitors^6–9 that becomes incorporated into the upper wall of the right atrial chamber at the junction with the superior vena cava⁵. Composed of a dense network of cardiomyocytes, the SAN structure can be split into the head domain, derived from a Nkx2-5–;Tbx18+ lineage, and the tail domain from a Nkx2-5+;Tbx18– lineage⁶. The development of these structures requires the contribution of several transcription factors, including Tbx3⁶, Shox2⁷, and Isl1^8,9 with mutations in each resulting in the failure of the SAN to form appropriately.

In this issue, two groups took independent approaches to uncover the CRE landscape of the SAN (Figure. 1). The group led by Dr. Vincent Christoffels used human pluripotent embryonic stem cells (hESCs) to generate SAN-like and ventricular-like cardiomyocytes and performed a method to uncover regions of accessible chromatin, Assay for Transposase-Accessible Chromatin using sequencing (ATAC-seq),¹⁰ to generate a list of accessible CREs specific to the human SAN⁴. On the other hand, Dr. Vasanth Vedantham’s group isolated SAN pacemaker cells (PC) and right atrial cardiomyocytes from the mouse heart and conducted ATAC-seq to uncover the PC-specific chromatin accessible regions in vivo³. Using this information, both groups identified and functional validated CREs that regulate SAN-specific expression of Tbx3, Shox2, and Isl1 for SAN development and function^3,4.

Figure 1. Identification of Functional Cis-regulatory Elements in the SAN-Pacemaker Myocardium.

A. The Christoffels and Vedantham groups generated pools of pacemaker and working myocardium from human and mouse, respectively. B. ATAC-sequencing was performed to identify regions of accessible chromatin. C. Pacemaker-enriched accessible elements were identified relative to working myocardium. D. Functional validation was performed by generating enhancer-reporter transgenic mice or zebrafish to examine identified element’s ability to generate pacemaker expression patterns and CRISPR-mediated deletions to examine identified elements’ functional requirement.

The work presented by the Christoffels group in van Eif et al. identifies several SAN-specific CREs for TBX3, ISL1, and SHOX2, including the identification of a set of SAN-specific CREs in the gene desert downstream of SHOX2 required for Shox2 expression in mice and embryonic viability⁴. The Christoffels group has a rich history uncovering tissue-specific TBX3 CREs in mouse and human and has begun to paint a complete picture of the TBX3 CRE landscape in the heart¹¹. Using their previously published annotations for the topologically associating domains (TADs), which define the three dimensional structure of the genome, in the vicinity of TBX3¹¹, van Eif et al. identified multiple SAN-specific CREs at the distal end of the TBX3-containing TAD^4,11. This region is of particular interest as it contains SNPs associated with heart rate recovery following exercise in humans and is syntenic in mice¹². Homozygous deletion of a 280kb region containing these SAN sites in mice results in embryonic lethality and loss of Tbx3 expression in the SAN⁴. Of particular interest to the heart rate phenotype, heterozygous mice for the 280kb deletion are viable, but show a slowed heart rate at baseline as adults and paced animals had a prolonged sinus node recovery time. One impressive result was the identification of a single CRE containing a SNP in linkage disequilibrium (LD) for heart rate recovery. The major allele variant demonstrates SAN expression in 5/7 lines; however, when van Eif et al. mutated this site to the minor allele variant, they saw SAN expression in only 2/7 lines suggesting that this single base pair substitution was sufficient to disrupt SAN expression directed by this CRE implicating this SNP as causative⁴.

The Vedantham group in Galang et al. utilized their ATAC-seq data from the mouse PC, and found that the top SAN-specific putative CREs were associated with genes involved in SAN development and function, including Hcn4, Isl1, Shox2 and Tbx3³. Furthermore, the group showed that 5/16 of these sites had been previously associated with SAN or venous inflow expression during embryonic development³. To discover the SAN enhancers that regulate SAN Shox2, Tbx3 and Isl1 expression, the Vedantham group searched for differentially accessible regions within previously annotated TADs that contain these TFs. Interestingly, as mentioned above there are many SAN specific accessible sites for Tbx3 and Shox2, but only a single peak, which spanned 2.7kb, and was located 120kb downstream of Isl1. Using this region’s sequence, Galang et al. generated Isl1 locus SAN Enhancer (ISE) reporters in mice and demonstrated SAN-specific activity in development and maturation. Moreover, ISE reporter activity was detected in Hcn4 expressing PCs with robust “funny” current, indicating that these ISE reporter positive cells were bona fide PCs. To functionally test the importance of the ISE, Galang et al. found that ISE deletion caused reduced ISL1 protein, SAN hypoplasia, sinus arrhythmias, and slowed resting heart rate, indicating an essential role for ISE in SAN function. Furthermore, ISE is functionally conserved between zebrafish and mice as the ISE regulates gene expression in the pacemaker region of the zebrafish. Finally, as seen in the Christoffels data⁴, ISE was conserved with humans and active in the SAN derived from hESCs. Interestingly, human SNPs that are located in or near the ISE are associated with resting heart rate further suggesting a potentially important function in humans³.

In summary, these exceptional studies unveil SAN-specific CREs from mice and humans and experimentally validate several novel CREs essential for SAN development and function. These findings pave the way for future studies understanding the molecular mechanisms of PC differentiation and SAN development downstream of ISL1, SHOX2, and TBX3 and the complex epigenetic regulation each of these transcription factors impart alone and in concert.