Top Advances in Functional Genomics and Translational Biology for 2012

Introduction

The potential therapeutic and diagnostic insights that modern “omics” techniques can provide are well recognized. The amount of data that can be generated with these technologies is increasing, as is the number of manuscripts being produced. For example, in 2012, there were >2700 articles archived in PubMed that used the term genome-wide association study. In 2008, there were just 750 such papers published. The most recognizable advance of 2012, however, was not the growing volume of data but rather the increased ability to implement the data obtained from genomics, transcriptomics, proteomics, and metabolomics into clinically applicable discoveries.

Such applicability is at the core of the American Heart Association Functional Genomics and Translational Biology (FGTB) Council (http://www.my.americanheart.org/fgtbcouncil). The FGTB council provides a multidisciplinary forum that brings together clinicians and researchers with expertise in translational science focusing on genetics, genomics, transcriptomics, metabolomics, and proteomics to target the worldwide burden of heart disease and stroke. With input from the Early Career Committee of the Council on FGTB, in choosing the research highlights of 2012, we considered many articles and selected 10 outstanding sets of articles that we felt exemplified the greatest advances of the past year and that we present below in no particular order.

The catalog of genome-wide associations with human diseases includes many variants that are distant from known genes. May et al¹ hypothesized that these associations could be explained by alterations of distal enhancer activity that lead to changes in gene expression. Using chromatin immunoprecipitation followed by massively parallel sequencing (ChIP-seq), they identified thousands of cardiac enhancer elements in human fetal and adult heart tissues. They showed that these enhancers are enriched near genes that are known to be expressed in the heart and can drive reproducible expression of a reporter in mouse cardiac tissues. Identification of these elements will serve as a valuable road map for prioritizing genomic variants of uncertain significance in future postgenomic studies.

Myocardial infarction (MI) remains a leading cause of death worldwide, and observational studies have found that, after an initial MI, the risk of a repeated MI within 1 year is increased. Dutta et al² examined this phenomenon by studying the effect of MI on established atherosclerosis in a mouse model. After Apoe^–/– mice with atherosclerosis underwent MI by coronary ligation, there was a noted increase in plaque size and an induction of vulnerable plaque morphology with an increased inflammatory cell content and protease activity, the latter of which was found with hybrid fluorescence molecular tomography–x-ray computed tomography. The acceleration of disease was found to last for several weeks and was characterized in particular by an increase in monocyte recruitment secondary to the sympathetic nervous system stimulating the release of progenitor cells and hematopoietic stem cells from the bone marrow. The translational relevance of these findings lies in the insight they provide into the previously uncharacterized secondary proatherogenic effects of MI and may result in new therapeutic anti-inflammatory or anti-adrenergic options in the post-MI setting.

In line with the inflammatory role of atherosclerosis, Ferguson et al³ studied lipoprotein-associated phospholipase A₂ (Lp-PLA₂) in both inflammation and atherosclerosis. Lp-PLA₂ has been identified as a potential therapeutic target for coronary artery disease, but whether Lp-PLA₂ is causal in coronary artery disease has not been established, and differences in Lp-PLA₂ expression in humans and rodent models of coronary artery disease had not been addressed. In this work, Lp-PLA₂ was found not to increase in the setting of inflammation alone in humans (through analysis after administering endotoxin). This contrasts with rodent findings in which Lp-PLA₂ is increased in endotoxemia and inflammation. Thus, Lp-PLA₂ does not seem to contribute to the acute-phase response of humans. In addition, inflammatory macrophages and foam cells were identified as the primary leukocyte source of Lp-PLA₂, not circulating monocytes. Single nucleotide polymorphisms of the gene that codes Lp-PLA₂, PLA2G7, were genotyped, and their relevance to coronary artery calcification was studied. Interestingly, although common variants in PLA2G7 were found to be associated with coronary artery calcification, these variants were not related to circulating Lp-PLA₂ activity or mass. Thus, the atherogenic role of PLA2G7 and Lp-PLA₂ seems to be independent of the inflammatory role and not related to the levels of circulating Lp-PLA₂ mass or activity. Although the role played by Lp-PLA₂ in atherosclerosis is established, this study highlights the challenges in translating this knowledge into an Lp-PLA₂–based biomarker or therapeutic intervention against coronary artery disease.

Chen et al⁴ presented a tour de force pilot study of a single individual, using a process they describe as integrative personalized omics profiling. By combining whole-genome sequencing with serial profiles of the peripheral blood transcriptome, metabolome, and proteome, they were able to correlate dynamic changes in gene expression with clinical events, including infections and nascent diabetes mellitus, over an observation period of 2 years. These analyses presage an era of personalized risk management based on a comprehensive view of a patient’s genetic and molecular data.

Previous work had shown that a combination of 3 genes encoding transcription factors—Gata4, Mef2c, and Tbx5—could reprogram various types of fibroblasts into cells that had cardiomyocyte-like properties in vitro and that could be successfully transplanted into mouse heart. Qian et al⁵ and Song et al⁶ hypothesized that delivery of transcription factors directly into mouse heart after MI could reprogram proliferating cardiac fibroblasts into functional cardiomyocytes in situ and thereby improve cardiac function. This is of particular interest because cardiac fibroblasts represent a large proportion of the cells in the heart and thus would serve as a ready pool for the generation of new muscle. The 2 studies showed that reprogramming of cardiac fibroblasts into functional cardiomyocytes can be successfully performed in a living mammal in the setting of MI; indeed, there was improved left ventricular contractile function and smaller infarct zones 2 to 3 months later, and there was no ECG evidence of more arrhythmias resulting from the reprogramming. As a translational application, vectors encoding the reprogramming factors could be delivered into myocardial tissue via a catheter-based approach in a patient who has experienced an acute coronary syndrome.

Most genome-wide association studies explain a relatively small fraction of the risk for cardiovascular diseases because the effect sizes of single variants tend to be small and the phenotypes are complex and heterogeneous. Inouye et al⁷ proposed to overcome some of these limitations by leveraging multivariate analysis and systems-based approaches to increase the statistical power of associations. In this pilot study, they identified highly correlated groups of serum metabolites in 2 Finnish cohorts and used multivariate methods to identify 7 novel loci that are associated with blood lipid levels. All but one of these loci would have been missed if they had used single-marker, single-phenotype association methods. They showed that the 2 most highly associated genes, SERPINA1 and AQP9, drive the expression of metabolic profiles that promote atherosclerosis and are enriched in mouse and human atherosclerotic plaques. Their work paves the way for similar approaches using accumulated “omics” data for deep phenotyping of other complex disorders.

Epidemiological studies have identified a diurnal variation in sudden cardiac death and ventricular arrhythmia. In the work by Jeyaraj et al,⁸ the mechanism through which this endogenous diurnal variation occurs was examined, with important, novel, and clinically relevant insights. Although the mice in this study were kept in darkness for 36 hours while telemetry was continued, corrected QT interval was noted to exhibit a diurnal variation. The authors serendipitously observed that Kruppel-like factor 15 (Klf15) also demonstrated diurnal variation. Klf15 was then shown to transcriptionally control rhythmic expression of Kv channel interacting protein 2 (KChIP2), which is the regulatory β-subunit for the repolarizing transient outward potassium current (I_to). There was a decrease in I_{to fast} density in Klf15^–/– mice, as well as a prolongation of action potential duration. After identifying the diurnal variation in corrected QT and the Klf15-mediated pathophysiologic mechanism, Jeyaraj et al showed that both Klf15^–/– and Klf15-overpressing mice have increased ventricular arrhythmia, and in the case of Klf15-overpressing mice, an increased incidence of sudden death was observed. This increased risk of sudden death occurred without a change in ventricular function or fibrosis and likely represented abnormalities in repolarization. The significance of this work lies in the outlining of a plausible mechanism for the diurnal variation in sudden cardiac death, and by identifying the role played by I_to, this study offers potential therapeutic targets.

Previous genetic studies of patients with mendelian forms of hypertension had identified causal genes involving renal electrolyte transport, highlighting new mechanisms for the regulation of blood pressure in humans. Pseudohypoaldosteronism type II, a rare genetic disorder causing hypertension, hyperkalemia, and metabolic acidosis, had been found in some cases to be caused by mutations in WNK1 or WNK4, but most cases had remained unexplained. Boyden et al⁹ used exome sequencing, linkage analyses, and gene resequencing in a large number of kindreds of individuals with pseudohypoaldosteronism type II to attempt to uncover novel casual genes. Of 52 pseudohypoaldosteronism type II kindreds, 48 yielded single or double allelic mutations in 1 of 4 genes—2 known (WNK1 and WNK4) and 2 novel (KLHL3 and CUL3)—thereby accounting for the vast majority of cases of the mendelian disorder. The investigators demonstrated that judicious use of a combination of techniques could unravel the genetic basis of a disease when 1 technique alone would have been insufficient. Furthermore, in their identification of 2 novel, interacting genes (KLHL3 and CUL3) that can underlie pseudohypoaldosteronism type II, they uncovered a new molecular pathway regulating renal electrolyte transport and blood pressure in humans.

Cardiac development depends on thousands of genes during differentiation. However, the manner in which these genes are epigenetically regulated and transcribed had not been comprehensively analyzed. Wamstad et al¹⁰ defined transcriptional and epigenetic changes during cardiac differentiation. Their study is noteworthy because of the manner in which the authors studied and interpreted the large amount of data generated from mouse embryonic stem cells, the genome-wide localization of histone modifications, and the high-throughput sequencing used. Using these techniques, the authors found that the fetal expression of genes is coordinated in a stage-specific manner by unique, functionally related chromatin patterns. New regulators of cardiac development were identified, including an enrichment of the MEIS1BHOXA9 motif; this motif predicts the binding of Meis factors, with genome-wide association studies having previously identified Meis1 as being associated with arrhythmia. Thus, this study provides a means to understand the mechanisms underlying cardiomyocyte differentiation, and the techniques used can potentially be used in conjunction with genome-wide association studies to define mechanisms of disease.

Protein phosphorylation can change protein structure and function by magnifying environmental impulses. Zhang et al¹¹ reported the phosphorylation of amino acids of cardiac troponin I in hearts from patients with ischemic and dilated cardiomyopathy hearts. This study is noteworthy as much for the successful techniques used as for the results obtained. The authors used targeted proteomics by multiple-reaction monitoring-mass spectrometry. Protein phosphorylation has historically been analyzed through Western blots with known phosphorylation sites assessed with site-specific antibodies. Mass spectrometry allows simultaneous measurements at multiple phosphorylation sites. In addition, the multiple reaction monitoring used facilitates quantification of a potentially phosphorylated residue at the specific site. This provides a high degree of specificity and allows quantification of the unphosphorylated and phosphorylated residues. This study assessed the levels of phosphorylation at 14 sites of cardiac troponin I, 6 of which had not previously been identified. Protein kinase C sites displayed increased phosphorylation, whereas protein kinase A sites were less phosphorylated. The translational relevance of this study is further enhanced by the authors’ observation that cardiac resynchronization therapy reversed the phosphorylation levels of S165, T180, and S198 in a canine model of heart failure. Few studies have assessed the role played by long-term phosphorylation in chronic disease, and the comprehensive manner in which the authors performed this work makes it the high-water mark in performing proteomic research to identify biomarkers and therapeutic targets.