Heart disease, a leading cause of morbidity and early death in babies, teenagers, and adults, occurs in association with other conditions or as a primary disorder of the myocardium. By applying genomic technologies, we have demonstrated that gene mutations cause many primary heart disorders, including congenital malformations, cardiomyopathies, and heart failure. Allelic variation in heart disease genes also contributes to common cardiovascular phenotypes. By combining genetic insights with mechanistic analyses of patient tissues and model systems, we have defined molecules and pathways that are critical for heart development and lifelong physiological functions and novel therapeutic opportunities in heart disease.
Congenital Heart Disease
Each year more than 1.3 million babies are born with a heart malformation, accounting for a third of all major developmental anomalies. In the vast majority, the cause of congenital heart disease (CHD) is unknown. To consider gene etiologies for CHD, we initially studied families with inherited heart malformations and identified highly penetrant mutations in cardiac transcription factors, including NKX2-5, TBX5, and GATA4. Most inherited CHD mutations alter gene (and protein) dosage, typically reducing normal levels by half. Notably, identical mutations in the same gene often cause very different malformations.
Improved genomic technologies allowed expansion of these studies to include severe cardiac malformations that were usually lethal before modern clinical interventions. As severe CHD often occurs sporadically, we suspected that de novo deleterious mutations caused some malformations. To test this model, we studied children with tetralogy of Fallot (TOF; a condition that combines a malpositioned aorta overriding both ventricles, a ventricular septal defect, pulmonary stenosis, and right ventricular hypertrophy) and their unaffected parents by using genomic SNP (single-nucleotide polymorphism) arrays. We identified rare de novo copy number variants (CNVs) in 10 percent of TOF patients, including a recurrent CNV at chromosome 1q21.1 that occurred in 1 percent of unrelated TOF cases. CNVs at the chromosome 1 locus also cause neurocognitive and psychiatric phenotypes, an observation that implicates a shared genetic etiology for these disorders, as occurs in DiGeorge syndrome (CHD and neurocognitive deficits), from CNVs at chromosome 22q11.2.
Most recently, we studied, with colleagues in the National Heart, Lung, and Blood Institute Pediatric Cardiac Genomics Consortium, exome sequences in children with a broad spectrum of severe CHD. We compared rare de novo nonsense, frameshift, and splice variants found in CHD children of healthy parents with those found in control child-parent trios. Among genes that are highly expressed in the developing heart, we observed sevenfold deleterious de novo mutations in CHD cases as in controls. Approximately 25 percent of mutated genes encode molecules involved in activating (H3K4 methylation), inactivating (H3K27 methylation), or reading chromatin marks, a finding that implies that CHD mutations compromise the regulated activation of promoters and enhancers in primordial heart cells. As chromatin modification genes are also mutated in autism and neurocognitive disorders, these studies, like the CNV analyses, suggest shared developmental pathways in the heart and central nervous system. Other damaging de novo mutations altered BCL9 (encoded in the chromosome 1q21.1 CNV interval and involved in canonical WNT signaling) and SMAD2 (a downstream target of NODAL signaling), indicating critical roles for other developmental pathways in heart formation.
Collectively, genetic studies indicate that inappropriate dosage of transcriptional regulatory molecules, rather than defective structural proteins, cause CHD. This model also suggests how different heart malformations might arise from the same mutation. Given the multiple levels of molecular regulation (epigenetic markers, transcription factor complexes, promoter/enhancer elements in target proteins) of cardiac transcription throughout development, genetic variation in many molecules may complement or accentuate the clinical consequences of each CHD mutation.
Idiopathic dilated cardiomyopathy (DCM) weakens and enlarges the heart. DCM often results in heart failure, a diagnosis associated with 50 percent morality within five years and the most common cause for cardiac transplantation. Studies have identified more than 40 DCM genes that encode components of the sarcomere, the cytoskeleton, or the nuclear lamina, but pathogenic mutations in these are identified in fewer than 30 percent of patients. Titin is an abundant 33,000–amino acid sarcomere protein that spans half (~0.5 μm) of this contractile unit. The gene encoding titin (TTN) had not been comprehensively studied in cardiomyopathy patients because of its monumental size (>100,000–base pair coding sequence).
We harnessed next-generation sequencing to define TTN variants in more than 600 hundred patients with DCM or hypertrophic cardiomyopathy (HCM) and control subjects. We found that rare nonsense, frameshift, or splicing TTN mutations were strongly enriched (P = 9x10-14) in DCM (27 percent) but not in HCM (1.3 percent) or control (2.8 percent) subjects. TTN mutations in DCM patients were nonrandomly distributed but were overrepresented in the carboxyl A band. These findings imply that foreshortened titin molecules, not haploinsufficiency, cause DCM, possibly because the carboxyl M band, which senses and modulates sarcomere force, is absent. We are developing models to test this hypothesis.
Newer sequencing methodologies have also expanded genetic analyses of cardiac hypertrophy. HCM, a dominant disorder characterized by unexplained ventricular hypertrophy, which increases risk for arrhythmias, stroke, heart failure, and sudden death, is caused by pathogenic mutations in eight sarcomere protein genes. We showed that mice with HCM mutations have primary increases in sarcomere force and secondarily in Tgf-β activation that result in myocardial fibrosis and heart failure. Antibodies or pharmacological agents can abrogate Tgf-β effects. Recently we developed a more direct strategy to attenuate disease: allelic-specific RNA silencing. By suppressing only 25 percent of mutant transcript levels in mice, we prevented HCM for 5 months.
Ventricular hypertrophy also occurs in 3 percent of the general population, often along with other cardiovascular disorders, and increases risk for adverse clinical events. To consider whether genetic variation in sarcomere proteins contributed to these common clinical phenotypes, we sequenced sarcomere protein genes in 3,600 Framingham Heart Study (FHS) and Jackson (JHS) Heart Study participants and assessed longitudinal information on cardiac morphology, function, and outcomes. The allelic spectrum of sarcomere protein genes among FHS and JHS participants was broad: more than 10 percent of participants had a rare nonsynonymous sarcomere gene variant, and 0.6 percent had variants predicted to be pathogenic—twice the population estimates for HCM. Notably, most study participants with "pathogenic" variants did not have overt HCM, but they did have increased adverse cardiovascular events (hazard ratio: 2:3). While our findings show that familial-based estimates of the penetrance of HCM mutations poorly predict the risk of HCM in the general population, they also indicate that sarcomere protein gene variants are not without consequence. With better definition of specific sarcomere variants that impair cardiac performance, the incorporation of genotype may improve risk stratification for many patients with heart disease.
Grants from the National Institutes of Health provide partial support for these projects.
As of October 21, 2013