Burden of Cardiomyopathic Genetic Variation in Lethal Pediatric Myocarditis

Abstract

Background -

Acute myocarditis (AM) is a well-known cause of sudden death and heart failure, often caused by prevalent viruses. We previously showed that some pediatric AM correlates with putatively damaging variants in genes related to cardiomyocyte structure and function. We sought to evaluate whether deleterious cardiomyopathic variants were enriched among fatal pediatric AM cases in New York City compared to ancestry-matched controls.

Methods -

Twenty-four children (aged 3 weeks to 20 years) with death due to AM were identified through autopsy records; histologies were reviewed to confirm that all cases met Dallas criteria for AM and targeted panel sequencing of 57 cardiomyopathic genes was performed. Controls without cardiovascular disease were identified from a pediatric database and matched by genetic ancestry to cases using principal components from exome sequencing. Rates of putative deleterious genetic variation (DV) were compared between cases and controls. Where available, AM tissues underwent viral analysis by PCR.

Results -

DV were identified in 4 of 24 AM cases (16.7%), compared to 2 of 96 age and ancestry-matched controls (2.1%, P=0.014). Viral etiologies were proven for 6 of 8 AM cases (75%), including the one DV+ case where tissue was available for testing. DV+ cases were more likely to be female, have no evidence of chronic inflammation, and associate with sudden cardiac death than DV− cases.

Conclusions -

Deleterious variants in genes related to cardiomyocyte integrity are more common in children with fatal AM than controls, likely conferring susceptibility. Additionally, genetically-mediated AM may progress more rapidly and be more severe.

Introduction

Acute myocarditis (AM) is an inflammatory myocardial disease typically due to viral infection, causing up to 16% of dilated cardiomyopathies (DCM) and 12% of sudden cardiac death (SCD) in adults.^{1, 2} Its impact is sizeable as affected children have only a 60% chance of transplant-free survival at 10 years.³ Viruses are the most common causes of AM, which is typically due to coxsackievirus, adenoviruses, herpesviruses or parvovirus B19.^{4, 5} Although approximately 1.5 million people are affected with AM per year,⁶ the incidence is relatively low in proportion to the pervasiveness of these pathogens. For example, while over half the population is seropositive for coxsackievirus B3 (CVB3)⁷, 76% of infected individuals have no clinical manifestations⁸ and fewer still develop AM. Presentations of AM also vary from mild systolic dysfunction to decompensated heart failure or SCD. Some patients recover cardiac function fully while others progress to transplant or death.

Explanations for the seemingly random susceptibility, variable expressivity and wide-ranging clinical courses and outcomes in AM have not been found. Human genetic factors may underlie individual predilection towards versus protection from myocardial injury following exposure to cardiotropic viruses. Indeed, we have previously identified a link between deleterious variation in cardiomyopathy (CMP)-associated genes and incidence of AM. Using whole exome sequencing (WES), we found that a significant proportion (16.7%) of children with AM harbored rare biallelic non-synonymous or splice-site variations in six CMP-associated genes (BAG3, DSP, PKP2, RYR2, SCN5A, and TNNI3)⁹. Viral etiology was proven in three of these cases. Interestingly, pathogenic variation in three of the identified genes (DSP, PKP2, and RYR2) are associated with arrhythmogenic cardiomyopathy (ACM); pathogenic variation in the other three have been observed in cases of DCM. This intriguing result aligns with previous observations about AM and CMP.

Individuals with dystrophin deficiency in Duchenne muscular dystrophy (DMD) have an increased likelihood of AM,¹⁰ and the occurrence of AM accelerates progression to DCM. AM may also be a factor in the pathogenesis of ACM; cardiotropic viruses are more frequently identified in biopsies of ACM patients,¹¹ and AM is associated with cardiac deterioration.¹² Interestingly, computational models predict the most likely protein targets for CVB3 protease 2A to be 1) desmocollin-2 (DSC-2)—a desmosomal protein with mutations linked to ACM—and 2) dystrophin.¹³ In fact, protease 2A preferentially cleaves dystrophin in murine and human tissues,¹⁴ and dystrophin deficiency predisposes to enterovirus-induced CMP.^{15, 16} Our previous genetic sequencing results coupled with these associations establishes a link between “cardiomyopathic” genes and AM. However, in the prior WES study, AM diagnoses were based on a range of criteria including symptomatology, relevant historical features, cardiac MRI findings, biomarker elevation and, for a limited percentage of cases, histologic evidence. We sought to validate these findings by investigating whether the rate of deleterious genetic variation within CMP-associated genes was increased in fatal, autopsy-confirmed pediatric AM cases.

Methods

Full methods are available as supplemental data. Human subject procedures were in accordance with institutional guidelines. Control subjects were identified through previous enrollment at Columbia University’s Institute for Genomic Medicine under different protocols approved by the Columbia Institutional Review Board. The data that support the findings of this study are available from the corresponding author upon reasonable request.

Results

Case and control demographics

Decedents with AM (n=24) ranged in age from 3 weeks to 20 years, were predominantly male and primarily of self-reported Black or Hispanic race/ethnicity (Table 1). Clinical history preceding death was variable; 58% of cases had clinical symptoms characteristic of AM prodrome prior to death (Supplemental Table I). All cases had histologic evidence of myocyte injury, lymphocyte infiltration and/or necrosis, meeting the Dallas criteria for myocarditis¹⁷ (Supplemental Table II). Genetic ancestries of control subjects were well-matched to cases according to PCA mapping (Supplemental Figure I). The number of males (55) and females (41) in the final control cohort was similar to that of the case cohort (P=0.82).

Table 1.

Patient demographics

Age		Sex		Self-reported ancestry		Prodrome
N (%)		N (%)		N (%)		N (%)
0 – 12 mo	4 (16.7)	M	15 (62.5)	Black	12 (50)	Yes	14 (58.3)
1 – 5	4 (16.7)	F	9 (37.5)	Hispanic	4 (16.7)	No	9 (37.5)
6 – 10	5 (20.8)			White	3 (12.5)	Unkn	1 (4.2)
11 – 15	2 (8.3)			E. Asian	1 (4.2)
16 – 20	9 (37.5)			S. Asian	1 (4.2)
				Unkn	3 (12.5)

Unkn- unknown

Genomic characteristics of cases and controls

Sequencing depth of down-sampled targeted gene panel (TGP) case data (mean 87.6X, s.d. 24.3) was similar to control WES (mean 92.5X, s.d. 49.7) for the 57 genes of interest (Table 2). Because of the inherent differences in sequencing technology between TGP and WES, we queried for batch effect by comparing overall rates of non-synonymous and synonymous variation between cases and controls. Average rates of non-synonymous and synonymous variation were similar between cases and controls (15.7 vs. 15.1 and 17.1 vs. 17.4, respectively), indicating that data sets were suitable for comparison.

Table 2.

List of structural cardiomyocyte genes analyzed in cases and controls.

ABCC9	LAMA4	SGCD
ACTC1	LAMP2	TAZ
ACTN2	LDB3	TCAP
ANKRD1	LMNA	TGFB3
BAG3	MYBPC3	TMEM43
CALR3	MYH6	TMPO
CASQ2	MYH7	TNNC1
CAV3	MYL2	TNNI3
CRYAB	MYL3	TNNT2
CSRP3	MYLK2	TPM1
CTF1	MYOZ2	TRDN
DES	MYPN	TTN
DSC2	NEBL	VCL
DSG2	NEXN
DSP	PKP2
DTNA	PLN
EMD	PRDM16
FHL2	PRKAG2
GATAD1	PTPN11
GLA	RBM20
JPH2	RYR2
JUP	SCN5A

Putatively damaging variants are identified in fatal AM cases

A single heterozygous deleterious variant (DV) was identified in 4 out of 24 AM cases (16.7%) in the following genes: CTF1, SCN5A, TTN (Table 3). Demographic, clinical and autopsy characteristics of DV+ case subjects are listed in Table 4. Each of the four DVs was found in only one case. One DV was found in each of two ancestry-matched controls (2.1%, Supplemental Table III). DVs were significantly enriched in AM cases compared to genetically matched controls (P=0.014, Table 5).

Table 3.

Deleterious variants (DV) identified in cases

Case	Age	Gene	cDNA change	Function	ClinVar	gnomAD AF	hg38 coordinates
1	5	SCN5A	c.G1231A	nonsynonymous SNV	P	.	38606058
2	9	CTF1	c.G548A	stopgain	.	3.38E-05	30902484
3	9	TTN	c.38122+2T>C	splicing	VUS	0.0001	178654910
4	12	TTN	c.10600delA	frameshift deletion	no Badge entries	8.14E-06	178756362

SNV- single nucleotide variation; P- pathogenic; VUS- variant of uncertain significance; . - not listed in database; NM numbers for cDNA variants are as follows: SCN5A, NM_000335; CTF1, NM_001142544; TTN (frameshift deletion), NM_133437; PKP2, NM_001005242

Table 4.

Demographic, clinical history and post-mortem findings of subjects with deleterious variation

Case	Age (yrs)	Sex	Race	Clinical vignette	Gross findings	Microscopic findings	Viral etiology
1	5	F	Hispanic	Found unresponsive in bed.	Acute angle RCA; enlarged tonsillar and adenoid tissue; obesity	Multifocal dense mononuclear inflammatory infiltrate (lymphocytes and plasma cells) with myocyte necrosis.	parvo (H)
2	9	F	E. Asian	Observed to be shaking; ventricular tachycardia.	Acute angle RCA	Multifocal dense mononuclear inflammatory infiltrate (predominantly lymphocytic) with myocyte necrosis.	N/A
3	9	F	White	Respiratory distress. Pulseless apneic when EMS arrived. Streptococcal infection with fever 10 days prior treated with Keflex. Vomiting and diarrhea for 1 month prior.	Normal	Multifocal dense mononuclear inflammatory infiltrate (lymphocytes and plasma cells with occasional eosinophils) with myocyte necrosis. Involvements of the pericardial and endocardial surfaces and AV node region.	N/A
4	12	F	Hispanic	Wide-complex VT to PEA arrest in ED. Had presented with abdominal pain, vomiting (diagnosed with gastritis), progressed to weakness of lower extremities, cough, running nose.	Slight myxomatous degeneration of mitral valve	Diffuse dense inflammatory infiltrate (lymphocytes and plasma cells) with microscopic foci of neutrophils and myocyte necrosis throughout myocardium. Inflammatory infiltrate most marked under epicardial surface.	N/A

H=heart; parvo= parvovirus B19; RCA= right coronary artery; VT= ventricular tachycardia; N/A = tissue not available for viral analysis; PEA= pulseless electrical activity; ED= emergency department

Table 5.

Enrichment of DV in AM cases compared to ancestry-matched or healthy pediatric controls

Cohort	N	No. with DV (%)	p-value
AM cases	24	4 (16.7)	-
Matched controls	96	4 (4.17)	0.014

Viral etiologies are responsible for some fatal AM cases

Analysis for viral RNA was performed on banked tissue for a subset of AM cases (N=8, ages 3 weeks through 5 years). A viral etiology was confirmed in six cases (75%; Supplemental Table IV). One of these was DV+ (Table 4) and due to infection with parvovirus, which, unlike in adults, does not remain latent in the myocardium of children and is deemed causal. The other three DV+ cases did not have tissue available for viral analysis and had not undergone viral testing.

Characteristics of DV+ vs. DV− fatal AM cases

Although the present study was not powered to detect statistically significant changes across smaller subgroups, there was a higher rate of sudden death in the DV+ vs. DV− AM cases (50% vs. 35% respectively; Supplemental Table I). Although the average age of DV+ subjects (8.8 years) was similar to that of DV− subjects (10.4 years), females with AM were more like than males to harbor a DV (4/9 vs. 0/15 respectively). Finally, histologic evidence of chronic myocarditis (associated with fibrosis) was similarly frequent between DV+ and DV− AM cases (0/4 vs. 6/20 respectively).

Discussion

Occurrence and outcome of AM is seemingly capricious, taking a fatal turn for some. Until recently, the role of human genetic factors in the pathophysiology of this disease was merely speculative. However, ample mechanistic studies in rodents and cell culture models supported the potential importance of cytoskeletal proteins in AM pathogenesis.^{14–16, 18–21} Our prior work demonstrated, for the first time, enrichment of deleterious variation in genes related to cardiomyocyte structure and function among children with AM. This confirmation of genetic susceptibility to AM in humans dovetailed findings from cell and rodent studies linking viral vulnerability to derangements in critical cytoskeletal proteins. However, one limitation of the prior retrospective human study was that the 42 included AM cases were diagnosed by various measures depending on the clinical information and diagnostic testing performed on a case-by-case basis. Where cardiac MRI images were available, diagnoses were based on modified Lake Louise criteria, which may be less specific in children.^{22, 23} Importantly, only a minority of cases were diagnosed using the gold standard histologic Dallas criteria, either because of a paucity of myocardial tissue available for analysis or not having met criteria. Diagnosing AM is generally challenging due to the limited sensitivity of the various criteria. False positive diagnoses are possible absent histologic confirmation since presentations and imaging features of AM overlap with other conditions.²⁴ In the present study, we refined our case selection by focusing only on fatal AM cases unequivocally diagnosed at autopsy based on cardiac histologic findings according to Dallas criteria. Unlike in the prior study where the case outcomes were mixed (some recovered complete myocardial function and others died), all cases in the present cohort shared death as a common outcome.

Another limitation of the prior study was that the enrichment analysis was performed against a publicly available database (1000 Genomes) or in-house control database of 2,324 exomes from heathy individuals and patients with severe infectious disease of childhood other than AM; neither of these control cohorts were specifically matched to the cases by either genetic ancestry or age. Since recent studies have highlighted the importance of contextualizing genetic variants to individual ancestry backgrounds in order to avoid misclassification,^{25, 26} we further refined our experimental design by performing enrichment analysis against controls matched to cases by genetic ancestry as well as age. Indeed, we found a significantly increased rate of DVs in the AM cohort compared to these well-matched controls.

Overall, the rates of DV within AM cases were similar in the present cohort (16.7%) and the original WES study (16.7%). However, all of the DV enrichment was biallelic in the original WES cohort, a finding ultimately not surprising in the context of known inbreeding in the studied population and confirmed autosomal recessive inheritance patterns. Our present investigation studying a more typically outbred population and utilizing a more strictly defined and fatal AM group identified only heterozygous DVs. Heterozygous mutations may therefore confer an increased risk for worsened AM severity. This hypothesis will need to be tested by performing additional genomic studies in larger cohorts and comparing genetic findings in cases stratified by outcome severity.

Although no specific mutation was recurrent, one DV-containing gene (SCN5A) were common between the present fatal AM cohort and the prior mixed WES cohort. Putatively damaging variants in SCN5A are linked to both ACM and DCM.^27–29 The finding of DVs within this gene across two distinct AM cohorts suggests it may represent a particularly relevant genetic locus for AM susceptibility. Additional genes in which we now identified AM-associated DVs include CTF1 (associated with DCM³⁰) and TTN (associated with ACM and DCM³¹); neither of these genes were interrogated amongst WES data in the original cohort. Altogether, we observe a consistent pattern of DV in ACM-associated genes, adding to the canon of literature demonstrating overlap between AM and inflammatory CMP.^{12, 32–35} Further supporting our overall findings connecting CMP gene mutations with AM is a recent retrospective report of eight children with presumed myocarditis by Brown et al., in which six were found to have a likely pathogenic or pathogenic variant in a non-syndromic CMP-related gene, including SCN5A and TTN³⁶. Follow-up mechanistic studies in cell culture and animal models are needed to investigate the mechanisms by which aberrations in CMP-related proteins may promote AM.

The ages of DV+ subjects were consistent with the overall age distribution of the full cohort. Other demographic features appear skewed. Namely, although there were nearly double the number of males as females in the full cohort, all four DV+ cases were female. Curiously, rodent models have previously found that females are less susceptible to AM (at least to coxsackievirus B3) due to the protective effect of sex hormones.³⁷ Whether this translates to human adults and children remains unknown, although in human studies, females produce more interferon in response to immune triggers than men.³⁸ Since the number of DV+ cases here is small, we cannot conclude that genetic AM susceptibility is more prevalent in females. However, future investigations in cell culture or rodent knock-down models harboring DV such as those identified in this study may shed further light on the effect of sex in modulating this disease. Larger human genomic studies will also be necessary to determine if this trend is reproducible.

While the full cohort was comprised of more than half with self-reported Black ancestry, DV+ cases were only observed among those with Hispanic, White and Asian self-reported ancestry. Since the numbers of subjects in each category are small, larger replication studies in heterogeneous samples are needed to determine whether DVs for AM cluster within certain populations.

Viral testing could only be performed in 8 of the 24 cases, due to limited tissue availability. Indeed, only one of the DV+ cases was included for viral analysis; myocardial tissue from this decedent was positive for parvovirus. Among the 20 DV− cases, viral etiologies were proven for 5 out of 7 included for testing. The overall rate of detection of viral genomes in this study (75%) is higher than what has previously been published in AM cohorts with mixed clinical outcomes (26-38%)^{4, 39}. To our knowledge, this is the first study to report the prevalence of viral genomes in fatal pediatric cases of AM.

Two alternative hypotheses have emerged for why cardiomyopathy gene mutations are prevalent in AM patients: 1) a vulnerable genetic substrate predisposes patients to myocyte injury in the face of viral infection or 2) individuals harboring such mutations have a silent cardiomyopathy that is “unmasked” by viral-induced decompensation. Small numbers within each sub-group of our study and the lack of uniform viral testing in all cases preclude drawing conclusions about the relationship between genotype and viral status. However, since fibrosis was not a feature in any of the DV+ cases—in contrast to the finding of fibrosis in 30% of DV− cases—this would support the notion that these individuals did not have pre-existing cardiomyopathy. The interaction between genetic risk and viral exposure is a topic ripe for further study.

Finally, we observed a trend toward higher incidence of sudden death without preceding clinical AM prodrome in the DV+ cases. We hypothesize that mutations in cardiomyocyte genes leading to cytoskeletal derangements affect viral replication and delay viral clearance, leading to more acute and rapid cardiac compromise. Further human and in vitro studies are needed to test this hypothesis.

In summary, cases of fatal pediatric AM are associated with significantly higher rates of putatively damaging cardiomyopathy gene variants than seen in controls. Although additional studies are needed before clinical genetic testing of AM patients can be broadly recommended, this approach may be appropriate in select circumstances. For patients who recover from AM, the finding of a pathogenic variant would justify ongoing surveillance for emergence of cardiomyopathy and/or arrhythmia. Female AM patients in particular with TTN mutations may warrant further counseling about the risks of peripartum cardiomyopathy⁴⁰. Furthermore, cascade screening would be warranted for relatives of these individuals, who may be at risk for developing cardiac disease associated with such mutations.

Nonstandard Abbreviations and Acronyms

AM

acute myocarditis

ACM

arrhythmogenic cardiomyopathy

CMP

cardiomyopathy

DCM

dilated cardiomyopathy

EBV

Epstein-Barr virus

NGS

next-generation sequencing

PCA

principal component analysis

SCD

sudden cardiac death

TGP

targeted gene panel

WES

whole exome sequencing