|Home | About | Journals | Submit | Contact Us | Français|
Autopsy-negative sudden unexplained death, including sudden infant death syndrome, can be caused by cardiac channelopathies such as Brugada syndrome (BrS). Type 1 BrS, caused by mutations in the SCN5A-encoded sodium channel, accounts for ≈20% of BrS. Recently, a novel mutation in the glycerol-3-phosphate dehydrogenase 1–like gene (GPD1-L) disrupted trafficking of SCN5A in a multigenerational family with BrS. We hypothesized that mutations in GPD1-L may be responsible for some cases of sudden unexplained death/sudden infant death syndrome.
Using denaturing high-performance liquid chromatography and direct DNA sequencing, we performed comprehensive open-reading frame/splice site mutational analysis of GPD1-L on genomic DNA extracted from necropsy tissue of 83 unrelated cases of sudden unexplained death (26 females, 57 males; average age, 14.6±10.7 years; range, 1 month to 48 years). A putative, sudden unexplained death–associated GPD1-L missense mutation, E83K, was discovered in a 3-month-old white boy. Further mutational analysis was then performed on genomic DNA derived from a population-based cohort of 221 anonymous cases of sudden infant death syndrome (84 females, 137 males; average age, 3±2 months; range, 3 days to 12 months), revealing 2 additional mutations, I124V and R273C, in a 5-week-old white girl and a 1-month-old white boy, respectively. All mutations occurred in highly conserved residues and were absent in 600 reference alleles. Compared with wild-type GPD1-L, GPD1-L mutations coexpressed with SCN5A in heterologous HEK cells produced a significantly reduced sodium current (P<0.01). Adenovirus-mediated gene transfer of the E83K–GPD1-L mutation into neonatal mouse myocytes markedly attenuated the sodium current (P<0.01). These decreases in current density are consistent with sodium channel loss-of-function diseases like BrS.
The present study is the first to report mutations in GPD1-L as a pathogenic cause for a small subset of sudden infant death syndrome via a secondary loss-of-function mechanism whereby perturbations in GPD1-L precipitate a marked decrease in the peak sodium current and a potentially lethal BrS-like proarrhythmic substrate.
An estimated 500 to 1000 people in the United States die of sudden cardiac arrest every day. Most of these deaths occur in elderly individuals as a result of ischemic heart disease. However, each year, several thousand people <40 years of age die suddenly, representing a disproportionate loss to the community and a number of lost life-years that rival ischemic heart disease–precipitated sudden deaths.1,2 For two thirds of the cases, a medicolegal examination is able to determine the cause of death, commonly attributing it to entities like hypertrophic cardiomyopathy, arrhythmogenic right ventricular cardiomyopathy, coronary artery anomalies, ruptured aortic aneurysms, or aortic valve stenosis.3–5 However, recent studies have indicated that as many as one third of sudden deaths in the young remain unexplained after a complete autopsy; such cases are labeled autopsy-negative sudden unexplained death (SUD).3,4,6,7 In addition, 60% to 80% of sudden deaths in children <1 year of age are reported as autopsy-negative sudden infant death syndrome (SIDS).8,9
Underlying molecular aberrations in cardiac ion channels implicated in heritable arrhythmia syndromes such as long-QT syndrome (LQTS), catecholaminergic polymorphic ventricular tachycardia, and Brugada syndrome (BrS) account for a significant proportion of SUD,10–14 as well as for 10% to 15% of SIDS.15,16 In addition, the Southeast Asian phenomenon of sudden unexplained nocturnal death syndrome, the most common cause of natural death in young Asians, has been shown recently to be phenotypically, genotypically, and functionally the same disorder as BrS.17 “Gain-of-function” mutations in the SCN5A-encoded sodium channel are associated with LQT3,18–20 whereas “loss-of-function” mutations in SCN5A result in type 1 BrS (BrS1). Given that SCN5A mutations account for only 20% of BrS,21 the vast majority of BrS remains pathogenetically unknown. Recently, a genome-wide linkage study involving a large pedigree with BrS phenotype revealed that the glycerol-3-phosphate dehydrogenase 1–like gene (GPD1-L) is associated with BrS and that the BrS2-associated GPD1-L mutation yields a “secondary” loss of function of the cardiac sodium channel.22
Given the recent implication of GPD1-L in the pathogenesis of BrS, a disease of known pathogenetic association with sudden death, we hypothesized that mutations in GPD1-L may confer risk for a lethal ventricular arrhythmia and thus be responsible for some cases of autopsy-negative SUD and SIDS. In this study, we sought to determine the spectrum and prevalence of GPD1-L mutations in a cohort of 83 unrelated SUD cases. The identification of a GPD1-L mutation in an SUD victim <1 year of age prompted us to explore our population-based cohort of SIDS (n=221) to formally determine the spectrum and prevalence of GPD1-L mutations in SIDS.
From September 1998 to January 2006, 83 medical examiner/coroner cases of SUD (26 females, 57 males; average age, 14.6±10.7 years; range, 1 month to 48 years) from 57 medical examiner offices were referred to the Mayo Clinic Windland Smith Rice Sudden Death Genomics Laboratory for a cardiac channel molecular autopsy. Seven of these decedents were <1 year of age and would be classified technically as SIDS. By definition, to be accepted as a case of SUD, the death had to be sudden, unexpected, and unexplained after the conclusion of a comprehensive medicole-gal autopsy. Decedents with a premortem diagnosis of a cardiac channelopathy either in himself or herself or in a family member were excluded from this study.
This study was approved by the Mayo Foundation Institutional Review Board. Although informed consent is waived for investigations involving decedents, written informed consent from the decedent’s parents or appropriate next of kin was obtained.
Between January 1996 and December 2000, 221 cases from a population-based cohort of unexplained infant deaths (84 females, 137 males; average age, 3±2 months; range, 3 days to 12 months) were submitted for postmortem genetic testing. SIDS was defined as for SUD, with the additional requirement of occurring before 1 year of age. Infants whose death was due to asphyxia or specific disease were excluded.
The present study was approved by the Mayo Foundation Institutional Review Board as an anonymous study. As such, only limited medical information was available, including sex, ethnicity, and age at death. Time of day, medication use, and position at death were not available. By definition, the infant’s medical history and family history were negative.
DNA was extracted from autopsy blood with the Puregene DNA Isolation Kit (Gentra, Minneapolis, Minn) or from frozen necropsy tissue with the Qiagen DNeasy Tissue Kit (Qiagen, Inc, Valencia, Calif). The entire coding region (8 translated exons) of GPD1-L was examined with polymerase chain reaction, denaturing high-performance liquid chromatography, and direct DNA sequencing as previously described.23 Primer sequences, polymerase chain reaction conditions, and denaturing high-performance liquid chromatography conditions are available on request.
To be considered a possible SUD/SIDS-predisposing mutation, the genetic variant had to (1) be a nonsynonymous variant (synonymous single-nucleotide polymorphisms were excluded from consideration), (2) involve a highly conserved residue, (3) be absent among 600 reference alleles from 200 healthy white and 100 healthy black control subjects, and (4) have resulted in a functionally altered, proarrhythmic cellular phenotype. Control genomic DNA was obtained from the Human Genetic Cell Repository sponsored by the National Institute of General Medical Sciences and the Coriell Institute for Medical Research (Camden, NJ). None of these control alleles were obtained from infants.
The full-length coding sequence of GPD1-L (GenBank accession No. BC028726) was subcloned into pIRES2EGFP (Clontech Laboratories, Palo Alto, Calif) to generate pGPD1-L–IRES2EGFP. For transduction into primary cardiac myocytes, the bicistronic adenovirus shuttle vector of GPD1-L and green fluorescent protein (GFP) was generated. GPD1-L cDNA was subcloned into an entry vector of pENTR1A-IRES2EGFP, which was created by incorporation of the multicloning sites, an internal ribosome entry site, and GFP coding region into pENTR1A (Invitrogen, Carlsbad, Calif). A recombination reaction between pAd/CMV/V5-DEST (Invitrogen) generated pAdGPD1-L–IRES2EGFP. The E83K–, I124V–, and R273C–GPD1-L mutants were incorporated into the wild-type (WT) vector by site-directed mutagenesis. All clones were sequenced to confirm integrity.
WT–GPD1-L and GPD1-L mutant constructs in pIRES2EGFP were transfected into HEK cells, along with the SCN5A-expressing vector (GenBank accession No. AB158469) using FuGENE6 reagent according to the manufacturer’s procedure. Neonatal ventricular myocytes were isolated from the hearts of 1- to 2-day-old mice through collagenase digestion (Worthington Biochemical Corp, Freehold, NJ) in calcium- and magnesium-free Hanks’ balanced salt solution (pH 7.4). The isolated cells were washed 3 times with PBS and plated in medium 199 (Gibco-BRL, Gaithersburg, Md) supplemented with (in mmol/L) 5 creatine, 2 L-carnitine, and 5 taurine and 0.2% BSA at a field density of 10 000 cells/cm2 on 35-mm culture dishes precoated with laminin (Sigma, St Louis, Mo). After 1 hour, the media was changed to remove the nonadherent cells and then infected with adenovirus shuttle vectors of the WT–GPD1-L and E83K–GPD1-L mutant.
Macroscopic sodium currents were measured 48 hours after transfection with the standard whole-cell patch clamp method at room temperature (22°C to 24°C). The extracellular solution contained the following: 140 mmol/L NaCl, 4 mmol/L KCl, 1.8 mmol/L CaCl, 0.75 mmol/L MgCl, and 5 mmol/L Hepes and was adjusted to pH 7.4 with NaOH. The pipette (intracellular) solution contained the following: 120 mmol/L CsF, 20 mmol/L CsCl, 2 mmol/L EGTA, and 5 mmol/L Hepes and was adjusted to pH 7.4 with CsOH. Electrodes used were manufactured from borosilicate glass using a puller (P-87, Sutter Instrument Co, Novato, Calif) and were heat polished with a microforge (MF-83, Narishige, Tokyo, Japan). The electrode resistances ranged from 1 to 2 mol/LΩ. Cells were mounted on an inverted microscope (Nikon Instruments Inc, Melville, NY) in a Faraday cage and were perfused continuously with the bath solution. Voltage clamp data were generated with an Axopatch 200B amplifier with series-resistant compensation and pClamp software (9.2 Axon Instruments, Foster City, Calif). Membrane current data were digitalized at 100 kHz, low-pass filtered at 5 kHz, and then normalized to membrane capacitance.
Activation was measured by clamp steps of −120 to 70 mV in 10-mV increments from a holding potential of −140 mV and fitted to the Boltzmann function GNa = [1 +exp(V1/2−V)/k]−1, where V1/2 and k are the midpoint and slope factor, respectively, and GNa=INa (norm)/(V−Vrev), where Vrev is the reversal potential and V is the membrane potential. Steady-state inactivation was measured using a two-step protocol with a holding potential of −140 mV. The first step was a 1-second conditioning pulse from −150 to 0 mV in 10-mV increments. The second step to 0 mV recorded the available current after the conditioning pulse. The line represents a fit to the Boltzmann function: INa=INa-max[1+exp(Vc−V1/2)/k]−1, where Vc is the membrane potential. The numbers and fit parameters are given in the Table. An indication of adequate voltage control was the graded response of the peak current-voltage relationship.
All data points are shown as the mean value, and the bars represent the SEM. Determinations of statistical significance were performed with Student t test for comparisons of 2 means or with ANOVA for comparisons of multiple means. A value of P<0.05 was considered statistically significant. With exact binomial confidence intervals used for an allele frequency, absence of the variants of interest in at least 600 reference alleles indicates with a 95% confidence interval that the true allelic frequency is <1%, which satisfies the genetic distinction between annotation as a mutation rather than a polymorphism.
The authors had full access to and take full responsibility for the integrity of the data. All authors have read and agree to the manuscript as written.
Overall, mutational analysis of GPD1-L revealed a putative pathogenic mutation in 1 of 83 cases (≈1%) of SUD. This GPD1-L mutation–positive decedent was among the 7 cases who died before 1 year of age (ie, SIDS cases). Therefore, we examined our population-based cohort of SIDS to further investigate the spectrum and prevalence of GPD1-L mutations in SIDS. Here, 2 of 221 SIDS cases (≈1%) hosted novel missense mutations in GPD1-L.
Figure 1 details the molecular characterization of the 3 missense mutations. An abnormal denaturing high-performance liquid chromatography elution profile (Figure 1A) and subsequent DNA sequencing (Figure 1B) led to the identification of a nucleotide substitution (307 G→A) producing a lysine (K) for glutamic acid (E) substitution (E83K) in a 3-month–old white boy. A nucleotide substitution (nucleotide 430 A→G) causing an I124V (isoleucine, I, to valine, V) missense mutation was identified in a 5-week–old white girl. Similarly, a nucleotide substitution (nucleotide 877 C→T) yielded an R273C (arginine, R, to cysteine, C) missense mutation in a 1-month–old white boy. All of these missense mutations were absent in 600 reference alleles and involved highly conserved residues across a variety of species (Figure 1C). Figure 1D depicts the linear topology of GPD1-L and location of the mutations.
Although the family history of the E83K victim included a history of poorly documented cardiac arrhythmias, no diagnosis of any arrhythmia syndrome, including BrS, had been assigned to any family members before the victim’s sudden death. Although we assume on the basis of the family history that this mutation may be heritable rather than sporadic, we were unable to acquire DNA from the infant’s parents because they declined participation. Because of the blinded, anonymous nature of the SIDS population-based cohort, it was not possible to obtain further information on the I124V-positive and R273C-positive decedents and their families.
HEK cells were transiently cotransfected with SCN5A and GPD1-L. Representative current traces for SCN5A coex-pressed with WT–, E83K–, I124V–, or R273C–GPD1-L are shown in Figure 2A, with typical INa for a step depolarization from −140 mV to various potentials. Coexpression of WT–GPD1-L caused robust peak INa density. However, consistent with a BrS1-like loss-of-function phenotype, INa density was reduced significantly with the GPD1-L mutants (Figure 2 and the Table). Peak INa-voltage relationships (Figure 2B) show an 8-mV positive shift in the midpoint of the activation relationship for R273C but no differences for the other mutations (the Table). Inactivation kinetics were not affected by the mutations (the Table). Unlike LQT3 gain-of-function SCN5A mutations, increased late INa was absent.
When WT–GPD1-L and E83K–GPD1-L were transfected into a stable SCN5A-expressing HEK293 cell line, similar decreases in INa density were obtained (−1137±186 versus −357±50 pA/pF, WT–GPD1-L [n=4] versus E83K–GPD1-L [n=7], respectively; P<0.001). To assess the effects in a more native myocardial cell environment, neonatal cardiomyocytes were transduced for 48 hours with adenovirus containing the WT– or the E83K–GPD1-L-IRES2EGFP recombinants. Only beating GFP-positive heart cells were selected. Representative current traces (Figure 3A and 3B) and summary data from 8 cells from 2 infections (Figure 3C) also show dramatically reduced INa density in native cardio-myocytes as a result of mutant E83K–GPD1-L.
Sodium channel–interacting proteins (ChIPs) have been implicated recently in the pathogenesis of arrhythmias and sudden death, including SIDS. In particular, mutations in 2 SCN5A-associated ChIPs, namely caveolin-3 (CAV3)24 and the sodium channel β-4 subunit (SCN4B),25 represent novel LQTS-susceptibility genes, and CAV3 mutations have been reported in black SIDS cases.26 In addition, a mutation in SCN5A that disrupts sodium channel association with the cytoskeletal protein ankyrin-G results in a BrS-like loss of function.27 Here, we detail 3 novel SIDS-associated mutations (E83K, I124V, and R273C) in a novel sodium ChIP, GPD1-L. Whether these mutations are familial or sporadic is indeterminate. In the case of E83K, a family history of unspecified cardiac arrhythmias remains unexplored because of nonparticipation. The anonymous nature of the SIDS population-based cohort prevented further investigation.
In 2002, Weiss et al,28 using linkage analysis in a large multigenerational family, discovered a novel BrS locus (BrS2) on chromosome 3, distinct from the BrS1/LQT3 locus. Subsequent candidate gene analysis identified a novel missense mutation, A280V, in GPD1-L that was shown to cosegregate with the BrS phenotype.22,28 Although little is known about GPD1-L, it is expressed in cardiac tissue and colocalizes with the SCN5A-encoded cardiac sodium channel at the plasma membrane.22
Overexpression studies of the identified BrS2-associated GPD1-L mutant with the SCN5A-encoded cardiac sodium channel confirmed a BrS phenotype of reduction in peak sodium current and normal inactivation kinetics, suggesting a loss of function of the sodium channel. Surface membrane expression of the sodium channel in the presence of mutant GPD1-L also was shown to be reduced, suggesting that this particular GPD1-L mutation disrupts trafficking and localization of the sodium channel to the membrane.22
Here, we provide molecular and functional evidence detailing GPD1-L as a novel susceptibility gene for SIDS. After elucidation of the E83K–, I124V–, and R273C–GPD1-L mutations in highly conserved residues that were absent in 600 reference alleles, a marked (≈60%) reduction in peak INa was demonstrated not only in heterologous expression systems for all 3 mutants but also in native cardiomyocytes (E83K–GPD1-L). The demonstration of surface membrane colocalization with SCN5A raises the possibility of a direct interaction with the sodium channel. Alternatively, mutant GPD1-L may exert its effects via a number of possible indirect mechanisms. For instance, GPD1-L may interact with any of a number of known sodium ChIPs such as the cytoskeletal protein ankyrin-G, scaffolding proteins such as the syntrophins, the caveolae-associated caveolin-3, or sodium channel degradation machinery proteins such as Nedd4–2.24,29 In addition, GPD1-L may affect protein kinase A–dependent phosphorylation of the channel, which blocks retention signals in the endoplasmic reticulum.30,31 Third, oxidation state has been shown to play a role in ion channel function and the arrhythmogenesis of atrial fibrillation.32,33 Thus, modulation via NAD+/NADH levels within the cell may contribute to what is already an increasingly complex picture of cardiac sodium channel regulation.
A functionally altered “SCN5A-centric” view of SIDS pathogenesis is beginning to emerge for the subset of SIDS cases stemming from heritable arrhythmia syndromes. Perturbed SCN5A channels resulting from primary mutations or rare genetic variants have now been implicated as a pathogenic mechanism for as much as 6.5% of white SIDS cases,15,16 a surprisingly high number considering that SCN5A mutations account for only 5% to 10% of LQTS.15 Additionally, the common sodium channel polymorphism, S1103Y, previously associated with increased risk of arrhythmia and sudden cardiac death in blacks,34,35 has been demonstrated to confer increased risk of SIDS in black infants.36 Moreover, functional perturbations in the cardiac sodium channel via mutations in the sodium ChIP caveolin-3 have been identified in 6% of black SIDS cases.26
Although these SIDS-associated mutations have been reported most commonly as gain-of-function LQT3-like mutations, evidence increasingly suggests that loss-of-function mutations in SCN5A also can result in sudden death–predisposing clinical phenotypes, including BrS and sick sinus syndrome. Here, we report the discovery of novel mutations in the sodium ChIP GPD1-L in 3 SIDS victims. The 3 mutations in GPD1-L precipitate a marked loss of function with respect to the peak sodium current akin to that described as the mechanism for primary BrS-causing mutations in the cardiac sodium channel. Given the functional data for these mutations and that reported by London et al,22 we hypothesize that these mutations conferred the principal substrate for a lethal ventricular arrhythmia during infancy. Furthermore, in addition to BrS, loss-of-function mutations in SCN5A can result in sick sinus syndrome and bradycardia, providing an alternative arrhythmogenic mechanism.37–39
Presently, it appears that ≈5% to 10% of SIDS is precipitated by disruption of the sodium channel macromolecular complex via mutations in either its pore-forming α subunit (SCN5A) or in 2 of its ChIPs, caveolin-3 and now GPD1-L. Genes encoding other ChIPs warrant further scrutiny as novel candidate genes for unexplained sudden death during infancy.
Sudden infant death syndrome (SIDS) claims >2000 seemingly healthy infants each year in the United States and is considered a multifactorial syndrome, with both environmental and genetic factors converging on the vulnerable infant during the first year of life. Among the sources underlying this vulnerability, cardiac channelopathies, particularly congenital long-QT syndrome, are believed to contribute to ≈5% to 10% of SIDS. Perturbations in the cardiac sodium channel represent the most common channelopathic cause of SIDS. Although “gain-of-function” mutations in the SCN5A-encoded pore-forming α subunit cause type 3 long-QT syndrome, loss-of-function mutations cause another potentially lethal channelopathy known as Brugada syndrome. Beyond this α subunit, the sodium channel is increasingly appreciated as a macromolecular complex. For this reason, sodium channel–interacting proteins represent viable candidates in the pathogenesis of SIDS. Here, we detail the discovery of the new Brugada syndrome–susceptibility gene, GPD1-L, in a population-based cohort of SIDS. Three novel missense mutations in GPD1-L were discovered and are shown to cause a marked reduction in the cardiac sodium current not only in a standard in vitro heterologous expression system but also in native cardiocytes. The molecular/cellular phenotype was consistent with a loss-of-function sodium channel disease like Brugada syndrome, providing probable cause and manner (ventricular fibrillation secondary to a Brugada syndrome–like mechanism) for ≈1% of this SIDS cohort.
We gratefully acknowledge the medical examiners, coroners, and forensic pathologists from across the United States for their referrals of SUD and SIDS cases. We also thank Jing Wang (University of Wisconsin) for technical assistance in the myocyte experiments.
Sources of Funding
Dr Ackerman’s research program is supported by the Mayo Clinic Windland Smith Rice Comprehensive Sudden Cardiac Death Program, the CJ Foundation for SIDS, the Dr Scholl Foundation, the Hannah Wernke Memorial Foundation, the American Heart Association (Established Investigator Award), and the National Institutes of Health (HD42569). The project described was supported by National Institute of Child Health and Development Grant No. R01HD042569 (Dr Ackerman) and National Heart, Lung, and Blood Institute Grant No. R01HL71092 (Dr Makielski).
The content of this work is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Child Health and Development, the National Heart, Lung, and Blood Institute, or the National Institutes of Health.