Author + information
- Received September 8, 2015
- Revision received October 25, 2015
- Accepted October 27, 2015
- Published online February 9, 2016.
- Rowida Almomani, PhDa,
- Judith M.A. Verhagen, MDb,
- Johanna C. Herkert, MDa,
- Erwin Brosens, PhDb,
- Karin Y. van Spaendonck-Zwarts, MD, PhDa,c,
- Angeliki Asimaki, PhDd,
- Paul A. van der Zwaag, MD, PhDa,
- Ingrid M.E. Frohn-Mulder, MDe,
- Aida M. Bertoli-Avella, MD, PhDb,f,
- Ludolf G. Boven, BSca,
- Marjon A. van Slegtenhorst, PhDb,
- Jasper J. van der Smagt, MDg,
- Wilfred F.J. van IJcken, PhDh,
- Bert Timmer, MD, PhDi,
- Margriet van Stuijvenberg, MD, PhDj,
- Rob M. Verdijk, MD, PhDk,
- Jeffrey E. Saffitz, MD, PhDd,
- Frederik A. du Plessis, MDe,
- Michelle Michels, MD, PhDl,
- Robert M.W. Hofstra, PhDb,
- Richard J. Sinke, PhDa,
- J. Peter van Tintelen, MD, PhDa,c,m,
- Marja W. Wessels, MD, PhDb,
- Jan D.H. Jongbloed, PhDa,∗ ( and )
- Ingrid M.B.H. van de Laar, MD, PhDb
- aUniversity of Groningen, University Medical Center Groningen, Department of Genetics, Groningen, the Netherlands
- bDepartment of Clinical Genetics, Erasmus University Medical Center, Rotterdam, the Netherlands
- cDepartment of Clinical Genetics, Academic Medical Center, University of Amsterdam, Amsterdam, the Netherlands
- dDepartment of Pathology, Harvard Medical School, Beth Israel Deaconess Medical Center, Boston, Massachusetts
- eDepartment of Pediatric Cardiology, Erasmus University Medical Center, Rotterdam, the Netherlands
- fCentogene AG, Rostock, Germany
- gDepartment of Medical Genetics, University Medical Center Utrecht, Utrecht, the Netherlands
- hCenter for Biomics, Erasmus University Medical Center, Rotterdam, the Netherlands
- iUniversity of Groningen, University Medical Center Groningen, Department of Pathology and Medical Biology, Groningen, the Netherlands
- jUniversity of Groningen, University Medical Center Groningen, Division of Neonatology, Beatrix Children’s Hospital, Groningen, the Netherlands
- kDepartment of Pathology, Erasmus University Medical Center, Rotterdam, the Netherlands
- lDepartment of Cardiology, Erasmus University Medical Center, Rotterdam, the Netherlands
- mDurrer Center for Cardiogenetic Research, Utrecht, the Netherlands
- ↵∗Reprint requests and correspondence:
Dr. Jan D.H. Jongbloed, Department of Genetics, University of Groningen, University Medical Center Groningen, P.O. Box 30.001, 9700 RB Groningen, the Netherlands.
Background Cardiomyopathies are usually inherited and predominantly affect adults, but they can also present in childhood. Although our understanding of the molecular basis of pediatric cardiomyopathy has improved, the underlying mechanism remains elusive in a substantial proportion of cases.
Objectives This study aimed to identify new genes involved in pediatric cardiomyopathy.
Methods The authors performed homozygosity mapping and whole-exome sequencing in 2 consanguineous families with idiopathic pediatric cardiomyopathy. Sixty unrelated patients with pediatric cardiomyopathy were subsequently screened for mutations in a candidate gene. First-degree relatives were submitted to cardiac screening and cascade genetic testing. Myocardial samples from 2 patients were processed for histological and immunohistochemical studies.
Results We identified 5 patients from 3 unrelated families with pediatric cardiomyopathy caused by homozygous truncating mutations in ALPK3, a gene encoding a nuclear kinase that plays an essential role in early differentiation of cardiomyocytes. All patients with biallelic mutations presented with severe hypertrophic and/or dilated cardiomyopathy in utero, at birth, or in early childhood. Three patients died from heart failure within the first week of life. Moreover, 2 of 10 (20%) heterozygous family members showed hypertrophic cardiomyopathy with an atypical distribution of hypertrophy. Deficiency of alpha-kinase 3 has previously been associated with features of both hypertrophic and dilated cardiomyopathy in mice. Consistent with studies in knockout mice, we provide microscopic evidence for intercalated disc remodeling.
Conclusions Biallelic truncating mutations in the newly identified gene ALPK3 give rise to severe, early-onset cardiomyopathy in humans. Our findings highlight the importance of transcription factor pathways in the molecular mechanisms underlying human cardiomyopathies.
Cardiomyopathies represent a clinically and genetically heterogeneous group of disorders affecting the ventricular myocardium. Among children <18 years of age, overall incidence of cardiomyopathy is 1.13 cases per 100,000 annually in the United States (1). Pediatric cardiomyopathy is associated with significant morbidity and mortality: up to 40% of affected children die or undergo cardiac transplantation within 5 years of diagnosis (2,3). Cardiomyopathy can be classified into 5 clinical phenotypes based upon morphological and functional characteristics: hypertrophic cardiomyopathy (HCM); dilated cardiomyopathy (DCM); restrictive cardiomyopathy; arrhythmogenic right ventricular (RV) cardiomyopathy; and unclassified cardiomyopathy, including left ventricular (LV) noncompaction (4). Extremely diverse, particularly in the pediatric population, the etiology of cardiomyopathy encompasses both nongenetic and genetic causes, including myocarditis, neuromuscular diseases, inborn errors of metabolism, malformation syndromes, and familial forms caused by mutations in genes encoding the specialized, often structural, components of cardiomyocytes. Because of the routine incorporation of genetic testing in the diagnostic evaluation of patients with cardiomyopathy, a causal diagnosis can be identified in more than 70% of children (5). Interestingly, the same genetic causes that result in cardiomyopathy in adults are prevalent in the pediatric population (e.g., sarcomeric or cytoskeletal gene mutations) (6).
Despite recent advances in understanding the genetic etiologies of pediatric cardiomyopathy, a substantial number of cases remain unsolved, suggesting that other genes await discovery. Establishing an underlying genetic cause for cardiomyopathy allows pre-symptomatic identification of family members at risk and facilitates reproductive decision making. Genetic and genomic studies continue to provide new insights into the pathophysiological processes contributing to cardiomyopathy and will ultimately facilitate development of patient-specific prevention and treatment strategies.
To identify new genes contributing to pediatric cardiomyopathy, we used a combined approach of homozygosity mapping and whole-exome sequencing.
We studied 4 individuals with pediatric cardiomyopathy from 2 consanguineous families of Dutch and Moroccan descent, respectively (Figures 1A and 1B). A third family of Turkish descent was identified by subsequent cohort screening (Figure 1C). Patients underwent extensive clinical investigations including high-resolution prenatal ultrasound, physical examination, 12-lead electrocardiography, transthoracic echocardiography, and post-mortem examination. Mutation screening of up to 48 known cardiomyopathy-related genes per individual was negative. A complete overview of the genetic and metabolic tests performed before this study is provided in the Online Appendix. All asymptomatic siblings and parents underwent echocardiographic screening. HCM was defined by the presence of increased LV wall thickness (>2 SD above the mean for body surface area in children or ≥13 mm in adult relatives) in the absence of hemodynamic stresses sufficient to account for the degree of hypertrophy. DCM was defined by the presence of LV dilation (LV end-diastolic dimension >2 SD above the mean for body surface area) and systolic dysfunction (fractional shortening or LV ejection fraction >2 SD below the mean for age) in the absence of abnormal loading conditions sufficient to cause global systolic impairment (4). A cohort of 60 unrelated patients with idiopathic pediatric cardiomyopathy from diverse ethnic backgrounds was available for mutational screening of candidate genes. These patients had previously been screened for mutations in 8 to 55 known cardiomyopathy-related genes. The medical ethical committees of the University Medical Center Groningen and the Erasmus University Medical Center approved this study. Written informed consent was obtained from all participants or their legal guardians.
Genomic deoxyribonucleic acid (DNA) was extracted from peripheral blood samples (A-VIII:1, VIII:2, IX:1, and IX:2; B-III:1, IV:3, and IV:4), buccal swabs (A-IX:3 and IX:4), amniotic fluid (B-IV:1), or fibroblasts (B-IV:2) (Figure 1). Genome-wide genotyping was performed using the Human610-Quad BeadChip (family A) or HumanCytoSNP-12 v2.1 BeadChip (family B) array with raw data normalized and converted into genotypes using GenomeStudio data analysis software (Illumina, San Diego, California). Genotype files were uploaded into the web-based tool HomozygosityMapper to detect homozygous stretches (>5 Mb) in patients that were not present in their unaffected relatives (7).
Genomic DNA was sheared by sonication, then the size distribution of the fragmented DNA was analyzed, targets were captured, and paired-end sequencing (2 × 101 bp) performed. The sequence reads were mapped and aligned against the human reference genome GRCh37/hg19 (Online Appendix). Variant calling was performed, and variants were filtered on quality (read depth ≥6) and location (within an exon or first/last 20 bp of introns). Variants with a minor allele frequency ≥1% in at least 200 chromosomes in dbSNP, the Genome of the Netherlands (8), the 1000 Genomes Project (9), or an in-house reference set, or a minor allele frequency ≥3% in the National Heart, Lung, and Blood Institute’s Go Exome Sequencing Project were excluded from further analysis. Assuming an autosomal recessive model of inheritance, we selected variants found in the heterozygous state in the father and in the homozygous state in the patient. Only variants predicted to change the protein sequence (nonsynonymous single nucleotide variants and short insertions and deletions) or the recognition of consensus ribonucleic acid (RNA) splice sites were considered. Finally, we filtered for variants present in the homozygous regions.
Bidirectional Sanger sequencing of the entire coding region and exon–intron boundaries of the alpha-kinase 3 gene (ALPK3) was performed using polymerase chain reaction (PCR) primers (Online Table 1). PCR amplification was performed; PCR fragments were purified and sequenced, and then data were analyzed (Online Appendix). For annotation of DNA and protein changes, the mutation nomenclature recommendations from the Human Genome Variation Society were followed (10). Nucleotide numbering reflects complementary DNA (cDNA) numbering with +1 corresponding to the adenine (A) of the ATG translation initiation codon in the reference sequence NM_020778.4.
To assess the effect of the sequence variants identified on protein structure and function, we used prediction programs integrated in the Alamut Visual v2.6.1 software (Interactive Biosoftware, Rouen, France) (11–13). Furthermore, possible effects on splicing were predicted using various integrated algorithm tools.
Reverse transcriptase PCR analysis
To investigate the effect of the splice site variant c.4736-1G>A at the RNA level, we performed reverse-transcription PCR analysis on RNA isolated from fibroblasts of the affected individual (A-IX:2) and 2 age-matched controls. Cells were cultured in Dulbecco’s Modified Eagle Medium supplemented with 10% fetal bovine serum, 1% penicillin/streptomycin, 1% glucose, and 1% glutamax, in atmospheric oxygen and 5% CO2 at 37° C. Total RNA was extracted, and reverse-transcription PCR was performed using gene-specific primers designed to amplify the exon expected to be affected by the mutation and flanking sequences (Online Figure 1). The resulting PCR products were examined by 2% agarose gel electrophoresis and subsequently subjected to Sanger sequencing.
Paraffin-embedded or frozen cardiac tissue was available from 2 affected individuals (A-IX:2 and B-IV:2). Muscle biopsy specimens from the lateral portion of the quadriceps femoris muscle were available from another affected individual (B-IV:3) and her healthy sister (B-IV:4). Tissues from age-matched donors were used as controls. All samples were histologically examined after hematoxylin and eosin and desmin staining using standard techniques.
Immunohistochemical analysis was performed on myocardial samples from individual A-IX:2 and 2 age-matched controls with no clinical or pathological evidence of heart disease, as described previously (14). In brief, glass-mounted sections of formalin-fixed, paraffin-embedded myocardium (5 μm) were deparaffinized and rehydrated before antibody retrieval. Primary antibodies included mouse monoclonal anti-plakoglobin (1:1,000 dilution), mouse monoclonal anti-plakophilin2 (2a+2b, undiluted), mouse monoclonal anti-desmoplakin (1:10 dilution), mouse monoclonal anti-pan cadherin (1:400 dilution), rabbit polyclonal anti-43 (1:400 dilution), and rabbit polyclonal anti-desmin (1:100 dilution). The slides were then incubated with indocarbocyanine-conjugated goat anti-mouse or goat anti-rabbit secondary antibodies (1:400 dilution). Immunostained preparations were analyzed by laser scanning confocal microscopy.
Results are summarized in Table 1. In family A, the second child of healthy Dutch parents was born at 35 weeks following an uneventful pregnancy (IX:2) (Figure 1A). At birth, he presented with respiratory insufficiency and cyanosis. Echocardiography showed severe LV dilation (LV end-diastolic dimension 23.5 mm; z-score unreliable because of massive edema) with markedly reduced contractility of both ventricles, and mitral and tricuspid regurgitation. He died due to progressive heart failure at 5 days. The parents agreed on autopsy (discussed in the Histology section later). Genealogical evaluation revealed that the parents were sixth-degree cousins. Family history was negative for sudden death or cardiomyopathies. Echocardiography revealed no abnormalities in his parents (ages 34 and 31 years), 2 brothers (ages 9 and 6 years), and sister (age 7 years).
In family B, the first pregnancy of these healthy consanguineous Moroccan parents (IV:1) (Figure 1B) included prenatal ultrasound examination at 33 weeks of gestation, revealing generalized hydrops fetalis and cardiomegaly with reduced contractility. The pregnancy ended in intrauterine fetal death at 35 weeks. External examination of the fetus did not show any gross abnormalities except massive skin edema. Autopsy was declined by the parents. Their second pregnancy resulted in a spontaneous abortion. During the third pregnancy, ultrasound examinations at 20 and 26 weeks of gestation showed an enlarged heart with severely reduced contractility, a thickened myocardium with spongy appearance, and severe tricuspid regurgitation (Figure 2A). At 30 weeks of gestation, severe fetal hydrops had developed. The patient was born at 36 weeks (IV:2) (Figure 1B) and presented with severe respiratory distress. Echocardiography revealed severe hypertrophy and dilation of both ventricles. She died 2 h post-partum. Her parents agreed to a detailed post-mortem examination of the heart (discussed in the Histology section later). In the fourth dizygotic twin pregnancy, extensive prenatal ultrasounds revealed no abnormalities. The twin girls were born at 37 weeks. Echocardiography at day 4 showed severe concentric HCM in 1 of the twins (IV:3) (Figure 1B), which remained stable over subsequent years. Cardiac examination in patient IV:3 at age 10 years showed increased thickness of the interventricular septum (IVS) (IVS thickness 13 mm, z-score +6) and LV posterior wall (LVPW) (LVPW thickness 15 mm, z-score +9.5) (Figures 2B and 2C), biventricular diastolic dysfunction, and repolarization abnormalities. There were no ventricular arrhythmias. Family history was negative for sudden cardiac death or cardiomyopathies. Echocardiography revealed no abnormalities in either parent (ages 39 and 22 years) or the twin sister (age 2 months).
For family C, the index patient (C-V:2) was the second child of consanguineous Turkish parents. He was diagnosed with severe HCM at 4 years of age with subsequent slow disease progression. Out-of-hospital cardiac arrest occurred at age 7 years due to ventricular arrhythmias, after which an implantable cardiac-defibrillator (ICD) was placed; he has since experienced several appropriate ICD shocks. Cardiac examination at age 11 years showed severe concentric LV hypertrophy (IVS 20 mm, z-score +11.2; LVPW 19 mm, z-score +13.1) without LV outflow tract obstruction (Figures 2D and 2E), moderate hypertrophy of the right ventricle (RV), and repolarization abnormalities. His father (C-IV:2) was diagnosed with HCM at age 27 years. Cardiac magnetic resonance imaging at age 37 years showed concentric LV hypertrophy, most pronounced in the septum (IVS 32 mm) with relative sparing of the basal segments, and RV hypertrophy (Figure 2F). Echocardiographic screening in the mother (age 42 years) and older brother (age 19 years) revealed no abnormalities. A paternal uncle (C-IV:1) had been diagnosed with HCM at 29 years. Cardiac examination at 45 years of age showed asymmetric septal hypertrophy (IVS 17 mm; LVPW 11 mm) and RV hypertrophy (7 to 8 mm; normal cutoff: 5 mm).
In family A, we identified 2 chromosomal regions that were homozygous in the affected boy (IX:2) and heterozygous in his parents (VIII:1 and VIII:2) and unaffected brother (IX:1): 15q25.2q25.3 (84,058,592 to 89,164,874) and 21q21.3q22.11 (29,440,969 to 35,838,907). In family B, we identified 3 homozygous chromosomal regions shared by all affected individuals (IV:1, IV:2, and IV:3), whereas the unaffected father (III:1) and sister (IV:4) were heterozygous: 5q14.3q15 (90,263,863 to 95,782,290), 15q25.1q25.3 (79,333,228 to 87,165,178), and 16q23.2q24.3 (81,602,681 to 90,148,796) (NCBI Build GRCh37/hg19). A small region (3.1 Mb) of homozygosity on chromosome region 15q25 overlapped between the affected individuals from both families.
Exome variant filtering
Exome sequencing was performed to target all exons and exon–intron junction sequences of known genes in the human genome for potential disease-causing variants. Using the filtering pipeline discussed earlier, affected individuals from families A and B were found to harbor a homozygous variant in the ALPK3 gene (Online Table 2), located in the overlapping region of homozygosity on chromosome 15q25. Patient A-IX:2 was found to carry a homozygous intronic variant in the consensus acceptor splice site sequence of exon 10 of the ALPK3 gene (c.4736-1G>A) (Figure 3), and this result was validated using Sanger sequencing. Both parents and 2 unaffected siblings (A-IX:1 and IX:3) were heterozygous for the mutation. One unaffected sibling (A-IX:4) carried the normal genotype. Patient B-IV:3 harbored a homozygous variant in exon 6 of the ALPK3 gene introducing a premature stop codon (c.3781C>T, p.Arg1261*) (Figure 3). Sanger sequencing proved that all affected children were homozygous for this mutation. The father and unaffected sibling (B-IV:4) were heterozygous for the mutation. The mother was unavailable for testing, but is presumed to be a heterozygous carrier of the mutation as well.
ALPK3 mutational screening
Subsequent Sanger sequencing of all coding regions and flanking intronic sequences of the ALPK3 gene in a cohort of 60 unrelated patients with childhood-onset HCM or DCM uncovered the homozygous nucleotide substitution c.5294G>A in exon 12 introducing a premature stop codon (p.Trp1765*) in a third patient (family C) (Figure 1C, Table 1). His affected father and paternal uncle were heterozygous carriers; no additional mutation in the known cardiomyopathy genes was found (Online Appendix). In the remaining cohort, no additional mutations in ALPK3 were identified. Further studies will be performed in these patients to discover novel disease genes.
Splice variant analysis
The c.4736-1G>A variant in intron 9 (family A) was predicted to impair the natural acceptor splice site recognition using in silico tools. Sanger sequencing of cDNA from cultured fibroblasts of patient A-IX:2 confirmed skipping of exon 10, leading to a deletion of 281 nucleotides (Online Figure 1). At the protein level, a frameshift starting at codon Val1579 and ending in a stop codon 29 position downstream (p.Val1579Glyfs*30) is predicted.
Post-mortem macroscopic examination of the heart of individual A-IX:2 revealed severe cardiomegaly (total weight 34.0 g; normal 13.7 g) and biventricular dilation. Microscopic examination of the myocardium showed subendocardial fibroelastosis without fatty replacement in both ventricles, fragmented elastin fibers, and myxoid degeneration of the stroma (Figure 4A). Post-mortem macroscopic examination of the heart of individual B-IV:2 revealed severe cardiomegaly (total weight 34.7 g; normal 15.5 g) and biventricular dilation with a large adherent thrombus to the left ventricle. Microscopic examination showed focal hypertrophy of the cardiomyocytes and extensive fibroelastosis in the subendocardial region without fatty infiltration (Figures 4B and 4C). No apparent myocyte disarray was observed both on evaluation with hematoxylin and eosin staining, and immunohistochemical staining for desmin (Figure 4D). Intercalated discs were difficult to recognize in both the patients and age-matched controls. Muscle biopsies from individual B-IV:3 (affected) and B-IV:4 (unaffected carrier) did not show any changes at routine histological and enzyme histochemical evaluation, including acid phosphatase, nicotinamide adenine dinucleotide + hydrogen-tetrazolium reductase, succinate dehydrogenase, and cytochrome C oxidase.
To determine the effects of ALPK3 mutations at the level of the intercalated discs, we performed immunohistochemical analysis of junctional proteins in cardiac tissue from 1 patient (A-IX:2) and 2 age-matched controls. Immunoreactive signal levels for the desmosomal proteins plakoglobin and desmoplakin were absent at intercalated discs in the patient sample. Signal levels for the other desmosomal protein plakophilin-2, the adhesion molecule N-cadherin, and the gap-junction protein connexin-43 were normal compared with controls (Figure 5). Results of desmin staining were inconclusive (data not shown), as intercalated discs were difficult to recognize in both the patient and control samples.
We identified a novel gene, ALPK3, involved in human cardiomyopathy (Central Illustration). Biallelic truncating ALPK3 mutations were identified in 5 patients with diverse ethnic backgrounds who presented with severe, early-onset cardiomyopathy. Four patients were diagnosed during fetal life or within hours of birth. The fifth patient was asymptomatic until the age of 4 years. Three of 5 patients died due to progressive heart failure between 35 weeks of gestation and 5 days after birth. The patients who died exhibited features of DCM or a combination of DCM and HCM. Two patients, who were still alive at age 11 years, showed severe concentric HCM. Although classic HCM and DCM primarily affect the left ventricle, 3 patients also displayed significant RV involvement. Interestingly, 2 heterozygous family members (C-IV:1 and IV:2) were diagnosed with an atypical form of HCM at a young adult age. Cardiac examination of the other proven and obligate heterozygotes (n = 8; age range 2 months to 42 years) revealed no evidence of cardiomyopathy. Although these numbers are limited, one may surmise that ALPK3 mutation carriers have an increased risk of developing cardiomyopathy; therefore, periodic screening should be considered.
The ALPK3 gene is the human ortholog of the murine Myocytic induction/differentiation originator (also known as Midori). In 2001, the Alpk3 gene was identified by applying a differential display technique in the P19CL6 cell line, which can be differentiated into beating cardiomyocytes upon incubation with dimethyl sulfoxide. Overexpression of Alpk3 promoted their differentiation into cardiomyocytes, whereas differentiation was inhibited by blockade of Alpk3 expression (15). The gene has been mapped to chromosome 15q25.2 and contains 14 exons (Figure 3). The ALPK3 protein contains 2 immunoglobulin (Ig)-like domains and an alpha-type protein kinase domain. All 3 private homozygous mutations identified in this study are predicted to create premature stop codons, leading to nonsense-mediated decay or truncated proteins with partial or complete removal of the kinase domain (Figure 3). We therefore hypothesize that these mutations result in loss of function of ALPK3. Alpha-kinase 3 belongs to a family of atypical protein kinases that recognize phosphorylation sites in the context of alpha-helices. Alpha-kinases are known to regulate a wide range of cellular processes, including cell migration, adhesion and proliferation, protein translation, vesicular transport, and magnesium homeostasis (16,17).
In mouse embryos, expression of Alpk3 was restricted to fetal and adult hearts and adult skeletal muscle. Interestingly, Alpk3 knockout mice display cardiomyopathy with striking similarities to the human phenotype described here (18). Although concentric cardiac hypertrophy of both the left and right ventricle was the predominant feature in Alpk3-deficient mice, other changes typically associated with DCM were also observed such as increased end-diastolic and end-systolic LV volume and a reduction in cardiac outflow, stroke volume, and ejection fraction (18).
Detailed histological and ultrastructural analysis of hearts from Alpk3−/− mice showed markedly reduced numbers of indistinct and fragmented intercalated discs and diffuse mild myofibrillar disarray, resulting in looser arrangement of adjacent myofibrils (18). Interstitial fibrosis, 1 of the pathological hallmarks of cardiomyopathies, was absent. In contrast to the observations in Alpk3 knockout mice, we observed extensive fibroelastosis without apparent myocyte disarray in our human cardiac samples (Figure 4).
Intercalated discs play an essential role in the mechanical and electrochemical coupling between adjacent cardiomyocytes. There are 3 main junctional complexes within the disc: fascia adherens, desmosomes, and gap junctions. At an early age, intercalated discs are difficult to visualize under light microscopy, as exemplified in this study. Unfortunately, no patient material was available for ultrastructural examination using electron microscopy. However, immunohistochemical analysis in myocardial tissue from 1 patient revealed the absence of desmosomal proteins plakoglobin and desmoplakin signals at the intercalated discs, whereas signals for the other junctional proteins were normal (Figure 5). Reduced immunoreactive signals for plakoglobin at cell–cell junctions have been observed in the majority of patients with arrhythmogenic RV cardiomyopathy. The specificity of this finding, however, remains under debate (14,19). Redistribution of plakoglobin from intercellular junctions to the intracellular space is believed to suppress canonical Wnt/beta-catenin signaling, leading to enhanced fibrogenesis and myocyte apoptosis (20). Reduced signal levels for desmoplakin at intercalated discs, on the other hand, have mainly been demonstrated in patients with end-stage DCM due to mutations in the gene encoding desmin (21). Although our findings support a role for ALPK3 in the formation and maintenance of intercalated discs, further studies in a larger cohort of patients are needed for validation.
The molecular mechanisms involved in the earliest stages of cardiac development are still largely unknown. ALPK3 is expressed from the very early stage of cardiogenesis, before key transcription factors, such as the NK-2 class homeobox transcription factor NKX2-5, the GATA-binding protein GATA4, and the MADS-box transcription enhancer factor MEF2C (15). Some cardiac transcription factors (e.g., HEY and HAND proteins) contain a helix-loop-helix domain, characterized by 2 alpha helices separated by a flexible loop structure, and might therefore be targets of ALPK3. By regulating the expression of these transcription factors, ALPK3 might control initial induction of differentiation and maturation of cardiomyocytes. Mutations in these putative downstream targets of ALPK3 have previously been associated with various congenital heart malformations and arrhythmias in humans (22). Structural heart abnormalities were not observed in the patients described in this study. In recent years, however, some of these genes have also been implicated in familial forms of cardiomyopathy, including NKX2-5, GATA4, GATA6, and the T-box transcription factors TBX5 and TBX20 (23–27). Interestingly, mice lacking Xin actin-binding repeat-containing protein 1 (Xirp1), a downstream target of Nkx2-5 and Mef2c, display intercalated disc abnormalities and cardiac hypertrophy with conduction defects (28). Further research into the transcription factor pathways involved in heart development and the putative role of ALPK3 in the related processes will provide more insight into the molecular pathways involved in pediatric cardiomyopathy and may have implications for the development of new therapeutic strategies.
Although our finding that ALPK3 is the causal gene in 3 unrelated families of different ethnic backgrounds with pediatric cardiomyopathy underscores its contribution to this severe disease, studies in other and larger cohorts to substantiate our findings are warranted. Further protein studies and electron microscopy examination are needed to fully understand the effects of ALPK3 mutations and to enhance our understanding of underlying mechanisms leading to the clinical phenotype. For example, immunolabeling with ALPK3 antibodies would give insight into its expression in cardiac tissue. Moreover, identifying interacting partner proteins of ALPK3 will lead to identifying downstream-acting proteins involved in the pathogenesis and remodeling of intercalated discs. Unfortunately, these studies are hampered by the limited availability of cardiac tissues of the patients.
We have shown that biallelic truncating mutations in ALPK3 led to severe, early-onset pediatric cardiomyopathy in humans. Carriers may develop cardiomyopathy as adults. ALPK3-related cardiomyopathy has morphological features of both HCM and DCM, and is characterized by biventricular involvement and atypical distribution of hypertrophy. Our findings emphasize the essential role of cardiac transcription factor pathways in normal myocardial development. Further studies are necessary to understand how dysregulation of the transcriptional regulatory network results in cardiomyopathy (Central Illustration).
COMPETENCY IN MEDICAL KNOWLEDGE: Biallelic mutations in ALPK3, a transcriptional regulator involved in early cardiomyocyte differentiation, can cause cardiomyopathy in children.
TRANSLATIONAL OUTLOOK: Further research is needed to describe the molecular and cellular mechanisms responsible for the development of cardiomyopathy in patients with ALPK3 mutations and explore potential therapeutic strategies that target these processes.
The authors thank the families for participating in this study. The authors would like also to thank Kate McIntyre for editing this manuscript, and Tom de Vries Lentsch and René Frowijn for photographic artwork.
This work was supported by funding from the Dutch Heart Foundation (2007B132, 2010B164, and 2014T007). The authors have reported that they have no relationships relevant to the contents of this paper to disclose. Drs. Almomani and Verhagen contributed equally to this work. Drs. Jongbloed and van de Laar are joint senior authors.
- Abbreviations and Acronyms
- dilated cardiomyopathy
- deoxyribonucleic acid
- hypertrophic cardiomyopathy
- implantable cardiac-defibrillator
- interventricular septum
- left ventricular
- left ventricular posterior wall
- polymerase chain reaction
- ribonucleic acid
- right ventricular
- Received September 8, 2015.
- Revision received October 25, 2015.
- Accepted October 27, 2015.
- American College of Cardiology Foundation
- Daubeney P.E.,
- Nugent A.W.,
- Chondros P.,
- et al.
- Elliott P.,
- Andersson B.,
- Arbustini E.,
- et al.
- Seelow D.,
- Schuelke M.
- Ng P.C.,
- Henikoff S.
- Hosoda T.,
- Monzen K.,
- Hiroi Y.,
- et al.
- Van Sligtenhorst I.,
- Ding Z.M.,
- Shi Z.Z.,
- Read R.W.,
- Hansen G.,
- Vogel P.
- Gustafson-Wagner E.A.,
- Sinn H.W.,
- Chen Y.L.,
- et al.