Author + information
- David Milan, MD∗ ()
- ↵∗Address for correspondence:
Dr. David Milan, Massachusetts General Hospital, Cardiovascular Research Center, 55 Fruit Street, GRB-109, Boston, Massachusetts 02116.
Genome-wide association studies (GWAS) have been used for more than a decade to find genetic links to complex human traits and diseases such as schizophrenia, rheumatoid arthritis, and atrial fibrillation (AF). On its face, the strategy is simple enough: compare the genetic make-up of enough people with and without a disease and you will be able to identify genetic variants linked to the disease. In this issue of the Journal, Thorolfsdottir et al. (1) performed such a study of 14,225 subjects with AF and 374,939 subjects without from the Icelandic population. Genome wide association studies of AF have been previously performed, resulting in identification of more than 30 signals that associate with AF. However, in a pattern that is commonly observed in such studies, the associated variants lie in intergenic and noncoding regions of the genome, adding to the challenge of understanding the mechanisms by which they cause AF. This well-established pattern makes the primary finding of the study by Thorolfsdottir et al. (1) more interesting because the novel signal they discovered lies in the coding region of the gene PLEC, previously unsuspected in AF, which encodes plectin, a cytoskeletal cross-linking protein thought to contribute to cardiac tissue integrity. In this case, the genetic variant results in a change in the protein composition of plectin. The implication is that the genetic signal is also the mechanistic link to the disease. The authors also identified a second sequence variant in an intergenic region between the genes METTL11B and LINC01142. This second signal lies a short distance (∼1 kb) away from a previously reported AF-associated variant and may be marking the same causal gene.
The frequency of the PLEC variant in the Icelandic population was fairly low, 3% in subjects with AF and approximately 2% in those without. Furthermore, the variant is exceedingly rare outside of Iceland, found in only 13 of 140,842 subjects in the Genome Aggregation Database. This makes replication of the Iceland findings challenging, but the investigators did identify 3 rare carriers of the PLEC variant in non-Icelandic samples, all 3 had AF. The effect size of the PLEC variant was modest, another common pattern in GWAS results, with an odds ratio of 1.55.
PLEC encodes plectin, a large cytoskeletal protein thought to be important for maintaining tissue integrity. Homozygous mutations in PLEC cause a syndrome of skin fragility and epidermolysis bullosa simplex (EBS), accompanied by muscular dystrophy, a myasthenic syndrome, or pyloric atresia. Cardiomyopathies and arrhythmias have been reported in EBS cases of PLEC mutations. However, no skin phenotypes were identified in even homozygous carriers of the Icelandic PLEC variant, suggesting that the functional consequences are subtle. Furthermore, no cardiomyopathy phenotypes were observed either, again supporting a subtle if any effect on ventricular function, although effects on atrial function would be more difficult to diagnose and therefore cannot be ruled out. The one additional phenotype that was associated with the PLEC variant was reduced circulating creatine kinase levels, which may reflect a lower skeletal and/or cardiac muscle mass.
As a gene encoding a structural protein expressed in the heart, PLEC joins MYL4 and MYH6 as AF genes that have structural/functional roles in the cardiomyocyte. These findings resonate clinically where AF as a manifestation of atrial myopathy or atrial remodeling is frequently discussed. We invariably examine the structure and function of the heart in patients with new onset atrial fibrillation to see whether the AF may be present in conjunction with structural heart disease. These genetic findings highlight the possibility that, even in the absence of overt cardiomyopathy, there may be subclinical dysfunction contributing to AF.
The investigators performed some interesting additional analyses that may shed light on the genetic signals discovered to date. First, they evaluated the 2 novel and 29 previously reported AF variants for association with electrocardiographic (ECG) traits. They found that some of the genetic variants were associated with multiple ECG parameters (PLEC and MYH6), whereas others showed no association with any ECG measurement (KCNN3 and METTL11B). These findings may provide clues to mechanisms and/or common pathways for the diverse genetic signals.
All genetic AF signals were tested for associations with secondary phenotypes. Of particular interest is the association with stroke. It seems plausible that some variants, for instance, those associated with structural/functional proteins in the atrium may not only cause AF, but the attendant stroke risk could be higher in this setting. Thus, each genetic variant was evaluated for the effect size on AF and compared to the association with stroke. In these analyses, all effects on stroke were consistent with proportionality to their effect on AF. However, this negative result does not rule out the possibility that some AF genetic risk factors may carry higher stroke risk than others, it simply means that such a finding was not supported by these data. The one AF-associated phenotype that did stand out was the risk of sick sinus syndrome (SSS) and of pacemaker implantation, which was observed in association with the MYH6 variant, originally discovered as an SSS-associated variant.
Recently the GWAS field has attracted criticism with questions about its ability to find useful disease links. GWASs are expanding to include hundreds of thousands of patients, but the argument is that increasingly large studies will find more and more genetic variants that have miniscule influences on disease. In a recent article published by Boyle et al. (2), the authors suggest that many GWAS findings have no specific biological relevance to the disease and therefore may not offer any useful insights into disease pathogenesis or serve as good therapeutic targets. Instead, they argue that many GWAS hits are peripheral to the central disease mechanisms and act through modification of core pathways through interconnected gene networks. In this “omnigenic” view, the concept is advanced that all genes expressed in disease-relevant cells can affect core disease genes and that most heritability can be explained on this basis.
When we examine AF genetics with this framework in mind, many observations come into focus. I think it would be premature to say that the current AF genetic findings are not useful or that further GWAS in AF should not be pursued. Only through ongoing studies to trace the genetic signals in AF to their fundamental mechanisms will we begin to know whether the signals are useful. Moreover, the argument can be made that GWAS itself is a valuable tool to identify members of the complex gene regulatory networks in a given disease state.
With respect to the utility of the PLEC variant in AF, the variant is found in so few people and increases risk in a modest enough fashion that genetic testing for this allele is unlikely to be clinically valuable. The subtlety of the variant and lack of clinical evidence of overt myopathy make its candidacy as a therapeutic target questionable. Thus, what have we learned from this study? These findings reinforce clinical observations that cardiac structure and function are important in AF pathogenesis. Whether as primary myopathy or as a response to hemodynamic stresses, the concept that variants in structural/functional proteins would be important in AF is intuitive and now has a greater grounding in genetic data. We have also gained further insight into the difficulties in teasing out genetic risks for stroke, apart from the risk of AF. These observations provide a better understanding of pathogenesis and perhaps more importantly the complexities of AF genetics and mechanisms.
↵∗ Editorials published in the Journal of the American College of Cardiology reflect the views of the authors and do not necessarily represent the views of JACC or the American College of Cardiology.
Dr. Milan has reported that he has no relationships relevant to the contents of this paper to disclose.
- 2017 American College of Cardiology Foundation