Cardiovascular Characteristics of Patients with Genetic Variation in Desmoplakin (DSP)

Background: Variants in the desmoplakin (DSP) gene have been recognized in association with the pathogenesis of arrhythmogenic right ventricular cardiomyopathy (ARVC) for nearly 20 years. More recently, genetic variation in DSP has also been associated with left-dominant arrhythmogenic cardiomyopathy. Data regarding the cardiac phenotypes associated with genetic variation in DSP have been largely accumulated from phenotype-first studies of ARVC. Methods: We aimed to evaluate the clinical manifestations of cardiac disease associated with variants in DSP through a genotype-first approach employed in the University of Pennsylvania Center for Inherited Cardiovascular Disease registry. We performed a retrospective study of 19 individuals with “pathogenic” or “likely pathogenic” variants in DSP identified by clinical genetic testing. Demographics and clinical characteristics were collected. Results: Among individuals with disease-causing variants in DSP, nearly 40% had left ventricular enlargement at initial assessment. Malignant arrhythmias were prevalent in this cohort (42%) with a high proportion of individuals undergoing primary and secondary prevention implantable cardioverter defibrillator implantation (68%) and ablation of ventricular arrhythmias (16%). Probands also experienced end-stage heart failure requiring heart transplantation (11%). Conclusions: Our data suggest DSP cardiomyopathy may manifest with a high burden of heart failure and arrhythmic events, highlighting its importance in the pathogenesis of dilated and arrhythmogenic cardiomyopathies. Targeted strategies for diagnosis and risk stratification for DSP cardiomyopathy should be investigated.


Introduction
The expanding use of clinical genetic testing in patients with heart failure is facilitating new insights into the causes of dilated cardiomyopathy (DCM). DCM has been defined by the presence of: (a) fractional shortening less than 25% (>2 SD) and/or ejection fraction less than 45% (>2 SD) and (b) left ventricular end diastolic diameter greater than 117% (>2 SD of the predicted value of 112% corrected for age and body surface area), in the absence of another myocardial, valvular, or systemic etiology of cardiomyopathy [1]. Familial DCM, a subtype of DCM, is suspected when (a) two or more affected relatives with DCM meet the previously mentioned criteria or (b) a relative of a DCM patient experienced unexplained sudden death before the age of 35 years [2]. The prevalence of familial DCM is estimated to be 30-50% [1,2]. Forty percent of familial DCM has an identifiable genetic cause and over 60 genes associated with familial DCM have been reported [1].
Natural history studies have demonstrated that pathogenic variants in DCM-associated genes, such as lamin A/C (LMNA), filamin C (FLNC), cardiac sodium channel NAv1.5 (SCN5A), and RNA binding motif protein 20 (RBM20), lead to a malignant arrhythmogenic phenotype, which can be unrelated to the degree of left ventricular (LV) dysfunction [1]. Arrhythmogenic DCM, which has been found in one-third of DCM patients [3], has emerged as an overlap phenotype with arrhythmogenic cardiomyopathy (ACM) and arrhythmogenic left/left-dominant arrhythmogenic cardiomyopathy (ALVC) [4]. The lack of specific and widely accepted diagnostic criteria for ACM and ALVC has limited the recognition and categorization of these entities.
The desmosome and its components regulate cell-cell communication. Desmoplakin (DSP), a desmosomal protein essential to cardiac force transmission, functions as a linker protein and connects the desmosome to a network of cytoskeletal proteins, including intermediate filament proteins [5][6][7][8]. Disruption in these desmosomal structural components can result in myocardial structural and electrical alterations. Recently, the ACM/ALVC phenotype has been identified in association with genetic variation in desmin (DES), the gene that encodes a multipurpose intermediate filament protein [9,10].
Genetic variation in DSP has been causally implicated in arrhythmogenic right ventricular cardiomyopathy (ARVC), with more recent suspicion in the pathogenesis of ALVC. A majority of the investigation of DSP-associated disease has been performed in cohorts of individuals with ARVC [11][12][13][14][15][16][17][18]. Emerging data from ACM cohorts have suggested features of DSP-associated cardiac disease that may be distinct from DCM, ARVC, and ALVC ( Figure 1). We hypothesize that the predominantly right ventricular phenotype-first approach utilized to date to investigate DSP-associated cardiac disease has limited the identification of 2. Methods

Subjects and Study Design
Individuals evaluated at the University of Pennsylvania Center for Inherited Cardiovascular Disease, a tertiary referral center for patients with suspected hereditary cardiomyopathies and otherwise unexplained heart muscle diseases, from 2011-2019 were eligible for inclusion. Individuals with "pathogenic" or "likely pathogenic" variants in DSP identified by Clinical Laboratory Improvement Amendments-certified clinical genetic testing and consistent with published standards by the American College of Medical Genetics and Genomics and the Association for Molecular Pathology were included [19]. Cascade screening of first-degree relatives for all individuals with cardiomyopathy was routinely recommended as per current practice guidelines [20]. Probands and first-degree family members were classified as such. Clinical genetic testing results were independently confirmed by licensed genetic counselors (J.C., L.H.-A.). Entry into the cohort was assigned by date of initial evaluation by a cardiologist in the University of Pennsylvania Health System. The Institutional Review Board of the University of Pennsylvania approved this study (protocol number 843087). Due to the retrospective nature of the study and waiver granted by the Institutional Review Board, no informed consent from the subjects was required.

Data Collection
Data obtained by review of medical records included clinical history, pedigree analysis, electrocardiography (ECG), transthoracic echocardiography (TTE), ambulatory electrocardiographic monitoring (MCOT), cardiac magnetic resonance imaging (CMR), treatment, and follow-up testing as clinically indicated. Baseline demographic data included age at diagnosis, sex, first-degree familial history of DCM or sudden cardiac death, and symptoms at initial visit. Testing was performed according to standard clinical protocols. Echocardiographic chamber quantification was performed in alignment with the 2015 American Society of Echocardiography/European Association of Cardiovascular Imaging recommendations on cardiac chamber quantification for adults [21]. Clinical outcomes were adjudicated by medical record review.

Statistical Analysis
Continuous variables were described with medians and interquartile ranges (IQR) and categorical variables as n (%) with group comparisons performed with Wilcoxon rank sum testing. Kaplan-Meier curves were produced to analyze event-free probability from date of initial evaluation to a composite event (ventricular arrhythmia ablation, heart transplantation) with right censoring at date of death or date of last follow-up and were compared using the log-rank test. A 2-sided p value < 0.05 was considered statistically significant. Analyses were performed with Stata 15.1/IC (College Station, TX, USA).

Clinical Presentation of Probands and Relatives
Between 2011 and 2019, a total of 19 patients, 11 probands and 8 relatives, from 14 different families were identified (Table 1). Data on variant type and pathogenicity are presented in Table 2. Of 14 unique variants found, 7 had not been previously cited in ClinVar, the United States National Institutes of Health public archive of human genetic variants (as of 20 March 2021) [22]. Age at diagnosis was not significantly different between relatives (median 35.5 years, IQR 31-48) and probands (42 years, IQR 30-48; p = 0.53). Five probands (26%) had a family history of sudden cardiac death, and five probands had a family history of DCM. At initial evaluation, most probands and relatives were asymptomatic or mildly symptomatic by New York Heart Association (NYHA) class. Palpitations were the most common symptom reported by both probands (27%) and relatives (63%) at initial evaluation.

Electrocardiography and Arrhythmias
Eight of eleven probands (73%) and six of eight relatives (75%) had abnormal 12-lead ECGs on initial electrocardiographic screening. Over half of the probands (55%) and 25% of relatives had T wave inversions in leads II, III, and aVF. T wave inversions in the lateral precordial leads (V4, V5, V6) were more frequent among probands compared with T wave inversions in leads V1, V2, and V3 (p < 0.01).
Of the seven individuals who met 2010 ARVC Task Force criteria for "definite" ARVC [23], three had no phenotypic evidence of RV disease and four had evidence of biventricular cardiomyopathy (Table 3).

Cardiovascular Imaging
Left ventricular ejection fraction (LVEF) was higher for relatives (46%,  at first contact compared to probands (30%, , though this difference was not statistically significant (p = 0.59). Four probands (36%) and three relatives (38%) had left ventricular enlargement by TTE at the time of first contact. Left ventricular end diastolic diameter (LVEDD) was larger for probands (57 mm, IQR 50-58) at first contact compared to relatives (48.5 mm, IQR 45-57), though this difference was not statistically significant (p = 0.13). LVEF and LVEDD did not significantly change for either probands or relatives over a median follow-up of 36.3 months (IQR 9.9-72.5). At the time of most recent follow-up, most probands and relatives had normal RV size (82% and 63%, respectively) and normal RV function (82% and 75%, respectively) on TTE.
Thirteen patients underwent CMR either at our or an outside institution. Twelve patients (92%) had LV late gadolinium enhancement (LGE), while only one proband and one relative (unrelated) had RV LGE (Table 4). Among probands, LV LGE occurred most commonly in the subepicardial and mid-myocardial layers ( Figure 2).

Histopathology
Myocardial histopathology was available from 5 of 19 patients: right ventricle (Patient 1), left ventricle (Patients 15 and 19), and native explanted heart (Patients 7 and 9). All samples demonstrated myocyte hypertrophy and mild to moderate endocardial and interstitial fibrosis. No sample demonstrated granulomas, giant cells, significant inflammatory infiltrates, amyloid, or iron deposition. Periodic acid-Schiff staining with and without diastase on Patient 15′s sample revealed presence of intracellular glycogen; however, this was not demonstrated on other LV samples. Histopathology findings are summarized in Table 5.

Outcomes
Over a median follow-up of 36.3 months (IQR 9.9-72.5), one relative required extracorporeal membrane oxygenation for cardiogenic shock and eventually died. Two probands underwent heart transplantation for NYHA IV/Stage D heart failure and refractory ventricular arrhythmias (Table 6). Kaplan-Meier curves showing a comparison of the incidence of the composite outcome of PVC ablation, VT ablation, and heart transplantation for probands and relatives are shown in Figure 3.

Discussion
Current knowledge regarding clinical phenotypes and outcomes of desmoplakin cardiomyopathies is largely based on published reports of small populations [5], including (1) large cohorts of individuals with ARVC [11][12][13]15,17,18,24], (2) other cardio-myopathy cohorts with fewer than 10 probands [25][26][27][28], and (3) single family/single DSP variant cohorts [29,30]. Our study is one of the larger single-center descriptions of the clinical characteristics of DSP cardiomyopathy and aligns with recent work describing the distinctive presentation and course of this disease and the importance of molecular diagnosis for appropriate detection and risk stratification [8,31].
In this cohort of 19 individuals with pathogenic or likely pathogenic non-missense DSP variants, we found a high prevalence of left ventricular pathology by TTE and CMR. Probands and relatives exhibited LV systolic dysfunction relatively early, in the third to fourth decades of life, and similar to the average age at diagnoses of other genetic ACM, such as LMNA-associated DCM. Right ventricular dilation and dysfunction by TTE and/or CMR were rare in this cohort, supporting emerging knowledge that there may be a subset of individuals with DSP cardiomyopathy who do not manifest with the classical ARVC phenotype. The possibility of a distinct pathogenesis of DSP-associated disease, compared to other desmosomal cardiomyopathies, has been recently suggested [32,33]. Only 7 of 19 individuals met the 2010 Task Force criteria for a "definite" diagnosis of ARVC, despite each individual having one major criterion fulfilled at baseline with the presence of a pathogenic/likely pathogenic DSP variant. Our findings align with a recent multicenter study from Smith et al. that similarly highlighted the poor sensitivity of the 2010 Task Force criteria in identifying individuals with DSP cardiomyopathy [8]. The high prevalence of LV LGE on CMR in our cohort indicates that LV scarring and fibrosis, particularly in the subepicardial and mid-myocardial layers, might warrant consideration as a characteristic injury pattern in DSP cardiomyopathy. The etiology of this scarring and fibrosis remains unclear; however, autoimmunity and myocarditis have been proposed [34,35]. Further efforts to link biomarkers, histopathology, and imaging longitudinally in individuals with DSP cardiomyopathy will be critical to our understanding of the etiology and sequelae of the desmosomal cardiomyopathies.
We describe a high prevalence of atrial and ventricular arrhythmias and sudden cardiac arrest in this cohort. In addition, we report substantial utilization of therapies related to the treatment of malignant ventricular arrhythmias and heart failure including, secondary prevention ICD implantation, appropriate ICD shocks, ventricular arrhythmia ablation, and heart transplantation. Our observations support recent work from Wang et al., derived from a single center prospective registry of ARVC [31], which demonstrated that individuals with disease-causing variants in DSP were at high risk for sustained ventricular arrhythmia and heart failure. Importantly, the three individuals in our cohort who underwent secondary prevention ICD implantation had LVEFs above the guideline recommended primary prevention ICD implantation threshold of <35%. Although a small sample, this raises the question of using LVEF thresholds established for DCM in DSP cardiomyopathy, similar to the exception recommended for LMNA cardiomyopathy [4]. Our findings also add further evidence that left-sided and biventricular involvement, and not just right ventricular disease, should be suspected in DSP cardiomyopathies.
Limitations of our study include its single center retrospective nature with a small sample size, resulting in limited direct genotype-phenotype associations, possible referral bias, and possible lead time bias with differences between probands and relatives. Complete clinical data, including dermatologic phenotyping, was not available for all individuals. We acknowledge that the TNNT2 variant found in Patient ID 7 has been reported in association with DCM and may have influenced this individual's phenotype. In frame deletions in DMD are associated with the X-linked recessive Becker muscular dystrophy, and this female patient (Patient ID 13) did not exhibit signs or symptoms of skeletal myopathy. Digenic heterozygosity in desmosomal genes has been reported as a potential contributor to the development of an ARVC phenotype [36][37][38][39]; however, the characteristics of individuals with digenic heterozygosity in DSP and either TNNT2 or DMD has not been previously reported.

Conclusions
Variants in DSP are estimated to occur in 3% of European individuals with DCM; although they have been historically implicated in ARVC, we demonstrate that disease-causing nonmissense DSP variants manifest with a high burden of LV injury, LV dysfunction, heart failure, and arrhythmic events. We highlight the distinct features of DSP cardiomyopathy and association of DSP in the pathogenesis of DCM and ALVC, supporting the need to incorporate a genotype-first approach and molecular diagnostics into risk assessment and clinical care.

Data Availability Statement:
The data presented in this study are available on reasonable request from the corresponding author and are not publicly available due to the potential to compromise the privacy of research participants.   Kaplan-Meier estimates of event-free survival for the composite outcome of ventricular arrhythmia ablation or heart transplantation with right censoring at date of death or date of last follow-up. Tick marks indicate censored individuals. Over the observed time of follow-up, no relative underwent ventricular arrhythmia ablation or heart transplantation. Reza   Cardiovascular imaging characteristics of 19 probands and relatives with pathogenic or likely pathogenic variants in the desmoplakin gene (DSP).  Table 6.

Characteristic
Clinical outcomes of 19 probands and relatives with pathogenic or likely pathogenic variants in the desmoplakin gene (DSP).