Author + information
- Received May 1, 2018
- Revision received July 25, 2018
- Accepted September 13, 2018
- Published online January 21, 2019.
- David J. Tester, BSa,∗,
- Jaeger P. Ackerman, BSa,∗,
- John R. Giudicessi, MD, PhDb,
- Nicholas C. Ackermana,
- Marina Cerrone, MDc,
- Mario Delmar, MD, PhDc and
- Michael J. Ackerman, MD, PhDa,d,e,∗ ()
- aDepartment of Molecular Pharmacology & Experimental Therapeutics, Windland Smith Rice Sudden Death Genomics Laboratory, Mayo Clinic, Rochester, Minnesota
- bDepartments of Cardiovascular Medicine and Internal Medicine (Clinician-Investigator Training Program), Mayo Clinic, Rochester, Minnesota
- cThe Leon H. Charney Division of Cardiology. New York University School of Medicine, New York, New York
- dDepartment of Cardiovascular Medicine, Division of Heart Rhythm Services, Mayo Clinic, Rochester, Minnesota
- eDepartment of Pediatric and Adolescent Medicine, Division of Pediatric Cardiology, Mayo Clinic, Rochester, Minnesota
- ↵∗Address for correspondence:
Dr. Michael J. Ackerman, Department of Molecular Pharmacology & Experimental Therapeutics, Windland Smith Rice Sudden Death Genomics Laboratory, Mayo Clinic, 200 First Street, Southwest, Guggenheim 501, Rochester, Minnesota 55905-0001.
Objectives This study determined if radical plakophilin-2 (PKP2) variants might underlie some cases of clinically diagnosed catecholaminergic polymorphic ventricular tachycardia (CPVT) and exercise-associated, autopsy-negative sudden unexplained death in the young (SUDY).
Background Pathogenic variants in PKP2 cause arrhythmogenic right ventricular cardiomyopathy (ARVC). Recently, a cardiomyocyte-specific PKP2 knockout mouse model revealed that loss of PKP2 markedly reduced expression of genes critical in intracellular calcium handling. The mice with structurally normal hearts exhibited isoproterenol-triggered polymorphic ventricular arrhythmias that mimicked CPVT.
Methods A PKP2 gene mutational analysis was performed on DNA from 18 unrelated patients (9 males; average age at diagnosis: 19.6 ± 12.8 years) clinically diagnosed with CPVT but who were RYR2-, CASQ2-, KCNJ2-, and TRDN-negative, and 19 decedents with SUDY during exercise (13 males; average age at death: 14 ± 3 years). Only radical (i.e., frame-shift, canonical splice site, or nonsense) variants with a minor allele frequency of ≤0.00005 in the genome aggregation database (gnomAD) were considered pathogenic.
Results Radical PKP2 variants were identified in 5 of 18 (27.7%) CPVT patients and 1 of 19 (5.3%) exercise-related SUDY cases compared with 96 of 138,632 (0.069%) individuals in gnomAD (p = 3.1 × 10−13). Cardiac imaging or autopsy demonstrated a structurally normal heart in all patients at the time of their CPVT diagnosis or sudden death.
Conclusions Our data suggested that the progression of the PKP2-dependent electropathy can be independent of structural perturbations and can precipitate exercise-associated sudden cardiac arrest or sudden cardiac death before the presence of overt cardiomyopathy, which clinically mimics CPVT, similar to the PKP2 knockout mouse model. Thus, CPVT and SUDY genetic test panels should now include PKP2.
- arrhythmogenic right ventricular cardiomyopathy
- catecholaminergic genetic testing
- polymorphic ventricular tachycardia
- sudden unexplained death in the young
Catecholaminergic polymorphic ventricular tachycardia (CPVT) is a potentially lethal heritable arrhythmia syndrome that classically manifests with exercise-induced syncope and/or seizures, sudden cardiac arrest (SCA), or sudden cardiac death (SCD) in the setting of a structurally normal heart (1). CPVT is associated with a normal resting electrocardiogram. CPVT is suspected clinically based on the patient’s personal and/or family history, and documentation of the emergence of significant and progressive ventricular ectopy during either treadmill or catecholamine stress testing, which can culminate in its specific but insensitive trademark arrhythmia of exercise-associated bidirectional ventricular tachycardia (2). Once thought to manifest only during childhood, more recent studies have suggested that the age of first presentation can range from infancy to 40 years (2,3). The potential lethality of CPVT is evident by its mortality rates of 30% to 50% by age 35 years and the presence of a positive family history of young (<40 years) SCD in more than one-third of CPVT individuals (4). In addition, 15% of autopsy-negative sudden unexplained death in the young (SUDY) cases may be caused by CPVT, especially when SCD occurred during exercise (5,6).
The pathogenic substrates for CPVT involve key components of intracellular calcium-induced calcium release from the sarcoplasmic reticulum. The most common subtype of CPVT, CPVT1, stems from mutations in the RYR2-encoded cardiac ryanodine receptor/calcium release channel and accounts for approximately 60% of CPVT cases (7–9). Mutations in RYR2 confer a so-called “gain-of-function” phenotype to the calcium-release channel, which results in increased calcium leak during sympathetic stimulation, particularly in diastole. In contrast to the autosomal dominant and/or sporadic CPVT1, CPVT is rarely autosomal recessive, due to mutations in CASQ2-encoded calsequestrin or TRDN-encoded triadin (10–12). In addition, bidirectional ventricular tachycardia can be seen in patients with KCNJ2-mediated Andersen-Tawil syndrome (ATS1), which can have some overlap features with CPVT (CPVT3), although it can be distinguished typically by a high burden of ambulatory ectopy, which is generally absent in CPVT1 (13).
Arrhythmogenic right ventricular cardiomyopathy (ARVC) is an inherited heart disease characterized by a fibrous or fibro-fatty infiltration of the heart muscle and an increased risk for SCD in the young (14). ARVC-triggered SCD is often associated with exercise and can occur during the subclinical (or “concealed”) phase of the disease when overt cardiomyopathy is not yet detectable by cardiac imaging (echocardiography or cardiac magnetic resonance imaging) (15,16). ARVC is associated primarily with dysfunctional desmosomal proteins and is most commonly caused by loss-of-function mutations in the PKP2-encoded plakophilin-2 protein (PKP2) (15–18).
Recently, Delmar et al. (19) demonstrated, in a cardiomyocyte-specific PKP2 knockout mouse model, that the loss of PKP2 markedly reduced the transcriptional expression of genes critical to intracellular calcium handling, including RYR2 and TRDN, and also reduced Casq2 protein levels. While still manifesting a structurally normal heart, the PKP2-knockout mice exhibited isoproterenol-triggered polymorphic ventricular arrhythmias that mimicked CPVT. These data further supported loss-of-function PKP2 variants as a potential cause of life-threatening ventricular arrhythmias even in the absence of overt structural heart disease in humans. Therefore, we speculated that radical PKP2 variants might be present in some patients who had been diagnosed clinically with genetically elusive CPVT and in some decedents classified as exercise-associated, autopsy-negative SUDY.
Briefly, 18 (9 males; average age at diagnosis: 19.6 ± 12.8 years) unrelated patients with clinically diagnosed but RYR2-/CASQ2-/KCNJ2-/TRDN-negative (i.e., genotype-negative) CPVT and 19 (13 males; average age at death: 14 ± 3 years) medical and/or coroner cases of exercise-associated, autopsy-negative SUDY were referred to the Windland Smith Rice Sudden Death Genomics Laboratory at the Mayo Clinic in Rochester, Minnesota, for exploratory genetic testing (Table 1). All CPVT patients and the decedents’ next-of-kin signed a Mayo Clinic institutional review board approved written consent form before genetic analysis.
PKP2 mutation analysis
Comprehensive mutational analysis of PKP2’s (NM_004572) 14 coding region exons was performed on genomic DNA from these 37 cases using DNA Sanger sequencing or next-generation whole exome sequencing (WES). Consistent with the American College of Medical Genetics and Genomics and the Association of Molecular Pathology variant interpretation guidelines (20), only rare, non-synonymous radical (i.e., frame-shift, nonsense, or splice-site altering) variants with a minor allele frequency of ≤0.00005 in the genome aggregation database (gnomAD) (21) were considered to be likely pathogenic. PKP2 missense variants were excluded in this study because of their relatively high frequency in gnomAD.
Overall, radical (i.e., nonsense, frame-shift, splice error) variants in PKP2 were identified in 5 of 18 (27.7%) patients who had been diagnosed clinically with CPVT but who were genotype-negative and in 1 of 19 (5.3%) exercise-related, autopsy-negative SUDY cases compared with 96 of 138,632 (0.069%) individuals in gnomAD p = 3.1 × 10−13 (Figure 1, Table 1, Online Table 1). The only additional PKP2 non-synonymous variant identified was a S70I-PKP2 variant identified in a 16-year-old male diagnosed with CPVT. This variant has a minor allele frequency of 0.021 in gnomAD. In addition, the variant has been seen as a homozygote in 65 of 108,411 individuals in gnomAD.
A p.R79X-PKP2 nonsense variant was identified in 2 unrelated cases (17-year-old female and 12-year-old male) diagnosed clinically with CPVT (Figure 1). Both patients experienced syncope while playing soccer. One of the 2 p.R79X-PKP2−positive CPVT patients subsequently satisfied (2 years after sentinel presentation and CPVT diagnosis) the 2010 Task Force Criteria for ARVC and required a cardiac transplantation at 16 years of age. However, the second p.R79X-PKP2−positive patient reportedly still exhibited a structurally normal heart.
A p.E85fs-PKP2 frame-shift variant was identified in a 28-year-old female who experienced SCA while jogging (Figure 1). A p.A418fs-PKP2 variant was identified in a 19-year-old male who experienced syncope during exercise and had a family history of autopsy-negative SUDY. A p.H689fs-PKP2 variant was identified in a 15-year-old female with exercise-induced ventricular fibrillation (Figure 1). A p.N634fs*22-PKP2 variant was identified in a 16-year-old white male who died suddenly while participating in noncontact football drills (Figure 1). His autopsy examination failed to reveal any primary congenital anomaly, primary pathological disease or abnormality, or any definitive traumatic injury to account for death. There was no evidence of fibrous or fibro-fatty infiltration of the heart muscle. His heart valves, chamber sizes, and myocardial wall thicknesses were within normal limits. Microscopic analysis of the heart was normal. The toxicology report was negative. The medical examiner concluded that the cause of death was “primary cardiac electrical dysfunction.”
Using a combination of Sanger DNA sequencing and/or WES and a PKP2 gene-specific analysis, we identified an ultra-rare, radical (i.e., frame-shift or nonsense) PKP2 pathogenic variant in approximately 28% (5 of 18 patients) of our cohort of unrelated patients who had been diagnosed clinically with CPVT by at least 1 pediatric and/or adult heart rhythm specialist, but who remained genetically elusive following standard CPVT genetic testing. We also identified 5% (1 of 19 cases) of our exercise-related, autopsy-negative SUDY cases.
These observations were consistent with previous reports that indicated that during the course of first-degree relative evaluations that occurred following a reportedly autopsy-negative SUDY, clinical evidence for ARVC emerged (22). In addition, PKP2 truncation variants were reported previously in autopsy-negative SUDY cases (22). In 2012, Skinner et al. reported on a 16-year-old male who died suddenly during exercise. Despite a negative autopsy, the decedent’s father and sibling both had cardiac magnetic resonance imaging that was suggestive for ARVC (22). In 2016, Bagnall et al. (23) identified PKP2 frame-shift variants in 2 of their 113 exertional and non-exertional SCDs in the young cases (1.8%). One of these was a 17-year-old male autopsy-negative SUDY case that hosted the same E85fs-PKP2 variant that was identified in 1 of our patients, a 28-year-old woman diagnosed with CPVT after a sentinel event of SCA while jogging. In 2017, Lahrouchi et al. (24) identified radical PKP2 variants in 1 of 30 autopsy-negative SUDY cases (3.3%) with either exercise- or extreme emotion-related sudden death, a R609fs-PKP2 variant in a 22-year-old man who died during exercise.
A recent study described a cardiomyocyte-specific, tamoxifen-activated PKP2 conditional knockout (PKP2-cKO) murine model that allowed control of the onset of PKP2 loss of expression (19). The data showed that PKP2 deficiency in adult ventricular myocytes was sufficient to cause ARVC. However, shortly after loss of PKP2 expression, mice mimicked the concealed stage of ARVC and showed a high propensity for ventricular arrhythmias, which were most apparent during a catecholaminergic challenge. Polymorphic ventricular arrhythmias were seen frequently in the PKP2-cKO mice. Interestingly, lethal arrhythmias were only observed in the early stage after loss of PKP2, when the mice did not yet show evidence of overt structural disease. One-third of the PKP2-cKO mice in the concealed stage died during the isoproterenol (ISO) challenge. The echocardiographic images of these PKP2-cKO mice did not show structural heart disease. Importantly, the ISO-induced polymorphic ventricular ectopy, couplets, triplets, and runs of polymorphic nonsustained ventricular tachycardia, together with the ISO-induced ventricular fibrillation documented lethal arrhythmias in PKP2-cKO mice with structurally normal hearts, mimicked the profile of CPVT. At this early stage of PKP2 loss, the mice fully satisfied the diagnostic criterion for CPVT. Moreover, the ISO-induced arrhythmias were prevented by flecainide treatment.
Because previous studies indicated that flecainide can limit the outflow of calcium through RyR2 channels, these results supported the notion that ventricular arrhythmias in the PKP2-cKO mice that occurred during the concealed stage of PKP2-mediated ARVC might result from dysregulation of intracellular calcium cycling by increased RyR2-dependent calcium release, which is similar to RYR2-mediated CPVT. Considering the phenotypic overlap between ARVC and CPVT, it was noteworthy that approximately 9% of patients with an overt ARVC phenotype might have RYR2 mutations as their predominant pathogenic substrate (25).
Although flecainide is widely viewed as the first addition to beta-blockers when breakthrough cardiac events occur in patients with RYR2-mediated CPVT (Class IIa recommendation by Priori et al. [26,27]), there is currently insufficient evidence to comment on the safety or efficacy of flecainide in PKP2-mediated CPVT and/or ARVC. Anecdotally, a 9-month trial of nadolol plus flecainide in 1 of the patients included in this study resulted in a marked reduction in the burden of ventricular ectopy and nonsustained ventricular tachycardia. However, flecainide was discontinued subsequently when a diagnosis of p.R79X-PKP2-ARVC was rendered out of concern that accentuation of an underlying PKP2-mediated reduction in sodium current (INa) by flecainide could be pro-arrhythmic.
Nevertheless, the potentially beneficial use of flecainide does appears to fit with: 1) the prevention of ISO-induced arrhythmias in PKP2 null mice by flecainide; and 2) the suggestion that flecainide plus sotalol and/or metoprolol might be an effective antiarrhythmic strategy for the control of refractory ventricular arrhythmias in patients with ARVC, including 2 patients with unspecific PKP2 variants (28). However, the long-term safety of flecainide in patients with PKP2-mediated CPVT and/or ARVC, particularly with regard to the potential for synergistic and potentially pro-arrhythmic effects on the cardiac sodium channel that may worsen with progression to ARVC (desmosome destabilization) or with age, remains undefined. As such, there is insufficient evidence to recommend the use of flecainide in individuals with PKP2-mediated CPVT and/or electrical disease at this time.
PKP2-mediated CPVT or PKP2-mediated electrical phase (concealed) ARVC
Despite the PKP2-cKO mouse model that completely mimics CPVT, do these human patients with truncating mutations in PKP2-encoded PKP2 merely have classic PKP2-mediated ARVC, albeit in the concealed phase of the disease’s natural history? It is well known that ARVC may present with exercise-associated potentially lethal arrhythmias during the phase of disease progression when structural abnormalities are undetectable. Such exercise-triggered events, which occur in the young patient in the setting of a structurally normal heart, may compel the heart rhythm specialist to diagnose the patient with CPVT. This may happen especially if the patient’s stress test shows normal sinus rhythm at rest with exercise-associated ventricular ectopy, when, in fact, the patient may have concealed ARVC.
The 12-year-old male patient with a p.R79X-PKP2 variant, only 2 years after his sentinel presentation and initial diagnosis of CPVT, subsequently satisfied the 2010 Task Force Criteria for ARVC. However, the remaining PKP2-positive patients have not yet converted to an overt cardiomyopathy, and, apart from their genetic test result, they still do not satisfy 2010 Task Force Criteria for ARVC. In addition, their electrocardiographic evaluation lacks T-wave inversions, epsilon waves, and late potentials, making it difficult and/or impossible to suspect PKP2-mediated electrical phase ARVC. In contrast, they were all diagnosed by at least 1 pediatric and/or adult heart rhythm specialist as having CPVT secondary to exercise-associated cardiac events in the setting of a structurally normal heart (and all but the aforementioned individual still do), a normal resting 12-lead electrocardiogram, and a treadmill stress test that showed exercise-associated ventricular ectopy.
Because pathogenic PKP2 variants might result in a CPVT-like phenotype during an early pre-cardiomyopathic phase of the disorder and during the natural progression of the disease toward a cardiomyopathic condition, segregation of both a CPVT and ARVC phenotype within a single pedigree might be seen. However, there were no examples in the literature or in our study of CPVT and ARVC phenotypes tracing through the same pedigree. Our limited familial cascade genetic testing and clinical evaluation for 3 of our index cases with pathogenic PKP2 variants did not illustrate this. For the 12-year-old male patient with the R79X-PKP2 nonsense variant, neither his variant positive brother nor mother has any phenotypic expression of CPVT or ARVC, including normal electrocardiogram, stress test, and echocardiogram. For the 28-year-old female patient with the E85fs-PKP2 frameshift variant, there was no phenotypic expression of ARVC in her family. Several of her first-degree relatives were genetically tested for this disease-causative variant. The patient’s older brother, younger sister, and her mother are E85fs-PKP2 variant positive; however, they do not express any apparent cardiomyopathic or arrhythmic phenotype. For the 19-year-old male patient with the A418Pfs*2-PKP2 variant, his variant positive 22-year-old brother who died suddenly had a negative autopsy. The patient’s 56-year-old variant positive father has no ARVC phenotypic expression. There is a paternal history of sudden death in the young. The patient’s paternal uncle died at age 50 years while running. In addition, there were fifth- and sixth-degree relatives from the paternal family line who died at age 31 years while playing basketball and another relative who died in their 50s. However, it is currently unknown if these relatives possessed the A418Pfs*2-PKP2 variant or if there was any cardiomyopathic expression of ARVC.
Importantly, in contrast to the characteristically progressive ventricular ectopy seen in patients with CPVT1 (i.e., RYR2-mediated CPVT), the stress-test associated ectopy was less predictable. As such, we suspected that all of the patients in this present study who were diagnosed with CPVT, yet hosted a PKP2 truncation variant, were likely to have concealed ARVC. However, we suspect that these patients will continue to be diagnosed with CPVT by many heart rhythm specialists who do not have the genetic test result. The difference in clinical diagnosis is not trivial, because patients with syncope due to ARVC may warrant an implantable cardioverter-defibrillator, whereas patients with syncope and/or seizures stemming from CPVT may be treated with beta-blockers, flecainide, and/or left cardiac sympathetic denervation as the first line of defense. Hence, a wrong diagnosis might lead to the wrong treatment. Therefore, we suggest that PKP2 should be added to the CPVT genetic test panel.
As seen in the public compendia of exomic databases like gnomAD, radical variants in PKP2 that result in premature truncation of PKP2 and anticipated haploinsufficiency have a very low, but non-zero, background rate of approximately 0.7 per 1,000 individuals. Although this is far lower than the 2% to 3% background rate of radical truncating variants in TTN-encoded titin (29), the phenotype of the patient must still be considered carefully before reflexively viewing any such PKP2 variant as a probable cause. Although PKP2 mutations may represent genetic disease susceptibility and/or modifiers with phenotypes ranging from none to severe, as well as being dependent on age, environment, and other disease states, when obtained as merely an incidental finding, a radical PKP2 variant is most likely just part of the background genetic noise (especially if buttressed by a normal stress test and normal imaging studies). Before such a cardiologic evaluation is conducted, a purely incidental finding of a radical PKP2 variant in an asymptomatic person with no family history would be associated with only an estimated 1% to 2% chance of having PKP2-mediated CPVT and/or ARVC. However, in the setting of an individual clinically diagnosed with CPVT previously and with a negative CPVT genetic test, the identification of a radical PKP2 variant would have an estimated signal-to-noise ratio of nearly 400:1. In contrast, when the phenotype is that of a decedent with exercise-associated SCD but a normal conventional autopsy, the signal-to-noise ratio decreases to an estimated 75:1. In such a decedent-centered context, there would be about a 1% to 2% chance that the PKP2 finding is irrelevant.
Limitations of this study included the small sample size and the lack of genetic evaluation of PKP2 copy number variations. Not including a PKP2 copy number variations analysis might have resulted in an under-estimate of the prevalence of pathogenic loss-of-function PKP2 variants in our genotype-negative CPVT and exercise-related, autopsy-negative sudden death in the young cohorts.
Approximately 28% of our genotype-negative CPVT cohort and 5% of our exercise-related. autopsy-negative sudden death in the young cases hosted a likely pathogenic PKP2 truncation variant. Based on our observations as well as others (22–24), we estimated that PKP2 pathogenic variants might be responsible for 2% to 5% of autopsy-negative SUDY. Our data suggested that the progression of the PKP2-mediated electropathy could be independent of structural perturbations and could precipitate exercise-induced SCA before the presence of an overt cardiomyopathy, similar to the PKP2-knockout mouse model that mimics CPVT. Consequently, both the CPVT genetic test panel and the molecular autopsy panel for autopsy-negative SUDY cases should now include PKP2.
COMPETENCY IN MEDICAL KNOWLEDGE: Progression of the PKP2-mediated electropathy can be independent of structural perturbations and can precipitate exercise-induced SCA and/or SCD before the presence of an overt cardiomyopathy, similar to the PKP2-knockout mouse model that mimics CPVT. Both the CPVT genetic test panel and the molecular autopsy panel for autopsy-negative SUDY cases should now include PKP2.
TRANSLATIONAL OUTLOOK: Further studies are needed to distinguish clinically PKP2-mediated concealed ARVC from the possibility of PKP2-mediated CPVT or to determine that before manifestation of an overt cardiomyopathy, these 2 designations reflect the same underlying disease phenotype for patients with PKP2 truncations.
The authors thank the Genome Aggregation Database (gnomAD) and the groups that provided exome and genome variant data to this resource. A full list of contributing groups can be found at http://gnomad.broadinstitute.org/about.
↵∗ Mr. Tester and Mr. J.P. Ackerman contributed equally to this work and are joint first authors.
Dr. M.J. Ackerman was supported by the Mayo Clinic Windland Smith Rice Comprehensive Sudden Cardiac Death Program. Dr. Delmar was supported by grants RO1HL134328 and RO1HL136179 from the National Institutes of Health; has been a consultant for Audentes Therapeutics, Boston Scientific, Gilead Sciences, Invitae, Medtronic, MyoKardia, and St. Jude Medical; and he and the Mayo Clinic have an equity/royalty relationship (without remuneration so far) with AliveCor, Blue Ox Health Corporation, and StemoniX. Dr. Cerrone was supported by grant AHA14SDG18580014 from the American Heart Association. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
All authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the JACC: Clinical Electrophysiology author instructions page.
- Abbreviations and Acronyms
- arrhythmogenic right ventricular cardiomyopathy
- catecholaminergic polymorphic ventricular tachycardia
- genome aggregation database
- sudden cardiac arrest
- sudden cardiac death
- sudden unexplained death in the young
- whole exome sequencing
- Received May 1, 2018.
- Revision received July 25, 2018.
- Accepted September 13, 2018.
- 2019 American College of Cardiology Foundation
- Leenhardt A.,
- Lucet V.,
- Denjoy I.,
- et al.
- Priori S.G.,
- Napolitano C.,
- Memmi M.,
- et al.
- Tester D.J.,
- Ackerman M.J.
- Laitinen P.J.,
- Brown K.M.,
- Piippo K.,
- et al.
- Priori S.G.,
- Napolitano C.,
- Tiso N.,
- et al.
- Ackerman M.,
- Priori S.,
- Willems S.,
- et al.
- Lahat H.,
- Eldar M.,
- Levy-Nissenbaum E.,
- et al.
- Philips B.,
- Cheng A.
- Groeneweg J.A.,
- Bhonsale A.,
- James C.A.,
- et al.
- Delmar M.,
- McKenna W.J.
- Cerrone M.,
- Montnach J.,
- Lin X.,
- et al.
- ↵Exome Aggregation Consortium (ExAC). Cambridge, MA. Available at: http://exac.broadinstitute.org. Accessed July 1, 2018.
- Lahrouchi N.,
- Raju H.,
- Lodder E.M.,
- et al.
- Priori S.G.,
- Wilde A.A.,
- Horie M.,
- et al.
- Priori S.G.,
- Blomström-Lundqvist C.,
- Mazzanti A.,
- et al.
- Ermakov S.,
- Gerstenfeld E.P.,
- Svetlichnaya Y.,
- Scheinman M.M.