Genetic Profile and Clinical Characteristics of Brugada Syndrome in the Chinese Population

Background: Brugada syndrome (BrS) is an inheritable arrhythmia syndrome that can lead to sudden cardiac death in patients while the heart structure is normal. However, the genetic background of more than 65% of BrS probands remains unclear. Objectives: The purpose of this study is to report the variant spectrum in a Chinese cohort with suspected BrS and to analyze their distinct clinical and electrocardiographic features. Methods: Patients with suspected BrS from Tongji Hospital between 2008 and 2021 were analyzed retrospectively. Results: A total of 79 probands were included in this study. Patients with type 1 BrS electrocardiogram (ECG) had a prolonged QRS duration compared to patients with type 2/3 BrS ECG. Of them, 59 probands underwent genetic testing. Twenty-five patients (42.37%) showed abnormal genetic testing results, and eight of them (13.56%) carried pathogenic/likely pathogenic (P/LP) mutations. Mutation carriers presented much more prominent depolarization and repolarization abnormalities than non-carriers, including a prolonged P-wave duration, QRS duration, QTc interval, decreased QRS amplitude, and deviation of the electrocardiographic axes (T-wave axis and R-wave axis). Furthermore, our study identified four novel P/LP mutations: Q3508X in TTN, A990G in KCNH2, G1220E, and D372H (in a representative pedigree) in SCN5A. Conclusions: Our study showed the variant spectrum of a suspected Chinese BrS cohort, and we identified four novel P/LP mutations in TTN, KCNH2, and SCN5A.


Introduction
Brugada syndrome (BrS) is one of the most common inherited primary arrhythmia syndromes with an extensive genetic heterogeneity. BrS is definitively diagnosed when a type 1 ST-segment elevation is observed either spontaneously or after an intravenous administration of a sodium channel blocking agent in at least one right precordial lead (V1 and V2), which are placed in the 2nd, 3rd, or 4th intercostal space [1]. The prevalence of BrS varies among continents, countries, and ethnicities, and is highest in Southeast Asia (0.37%) [2], which may be attributed to genetic polymorphisms in the SCN5A promoter region [3]. However, the real prevalence in the general population remains unclear due to the intermittent and concealed classic electrocardiogram (ECG) pattern [4].
BrS typically manifests in the third or fourth decade of life, but this syndrome may occur at any age, from 2 days old to 84 years old. The incidence in males is 8-10 times higher than that in females [4]. More than 60% of patients with a BrS ECG are asymptomatic and diagnosed incidentally by a routine evaluation, family screening, or the observation of an abnormal ECG pattern during a fever [5]. A small number of patients present various symptoms, including slight darkness, a history of syncope (30%), paroxysmal nocturnal dyspnea, ventricular tachycardia/fibrillation (VT/VF), and sudden cardiac death (SCD) (6%). An SCD is often the first manifestation of BrS, predominantly in adult males at night or during rest [6,7]. BrS is considered to be responsible for at least 4% of all SCDs and at least 20% of those occur in patients with normal hearts [4]. Patients with aborted cardiac arrest or documented spontaneous sustained VT are at the highest risk of an SCD, followed by a history of cardiac syncope. Moreover, the risk of life-threatening arrhythmias in asymptomatic patients is 0.5-1.5% per year [6]. From this, it is important to focus on identifying the genetic cause of BrS to detect asymptomatic genetic carriers at risk of an SCD.
BrS was initially thought to be a monogenic, autosomal dominant disease. However, the occurrence and prognosis of BrS are more likely affected by a combination of multiple genetic alterations and environmental factors due to the incomplete penetrance and variable expressivity [7]. Currently, more than 500 potentially disease-causing variants account for about 30-35% of BrS patients, mainly including genes regulating the sodium current (INa), the L-type calcium channels (ICa), and the transient outward potassium channels (Ito) [8].
The majority of all pathogenic mutations (>75%) reported are located in SCN5A, which is the only definitive gene for BrS, accounting for 20-30% of BrS patients [9][10][11]. About 150 additional variants proposed to be causative of BrS in other genes explain no more than 10% of cases. These genes are classified as minor genes with a limited evidence for BrS [7,11]. Thus, approximately 65-70% of BrS probands remain genetically undetermined [7].
Genetic testing is recommended for an early detection of a patient's relatives who are potentially at risk [4]. However, the data about the genetic background and clinical characteristics of Chinese BrS patients are scarce [12][13][14]. It has been reported that the prevalence rate of SCN5A mutations is around 8% (4/47) in Taiwan [14] and 14% (5/36)-34% (22/65) in Hong Kong [12,13]. Additionally, the distribution of disease-causing genes among BrS patients in the Asian population might differ from that in the Caucasian population (20-25%) [15]. However, there is a lack of large-scale genetic and clinical characteristic data in Chinese BrS patients. In the present study, we aimed to determine the prevalence and spectrum of genetic variations in BrS-susceptibility genes in a Chinese cohort with suspected BrS, and to analyze the clinical and genetic features.

Study Population
A total of 79 Chinese probands with suspected BrS were included from Tongji Hospital Affiliated Tongji Medical College of Huazhong University of Science and Technology (Wuhan) between 2008 and 2021.
Patients were suspected to have BrS according to the following diagnostic criteria [5]: patients showed one of the three types of BrS ECG and/or presented with one of the following typical symptoms: documented VF or polymorphic VT, arrhythmic syncope or paroxysmal nocturnal dyspnea, a positive cardiac electrophysiology examination, a family history of SCD at <45 years old or type 1 BrS ECG in family members. Type 1 ("coved type"), the only diagnostic pattern for BrS, is defined as an ascending and high take-off of ≥2 mm at the end of the QRS duration in ≥1 right precordial leads (V1 to V3), followed by a coved or rectilinear down-sloping ST-segment and a negative symmetric T-wave. Type 2 ("saddle-back type"), a suggestive pattern of BrS [16], is characterized by an ST-segment elevation ≥0.5 mm (generally ≥2 mm in V2) in a ≥1 right precordial lead (V1 to V3), followed by a convex ST and a positive T wave in V2 or variable morphology in V1. Type 3 is characterized by either a saddleback or coved appearance with an ST-segment elevation <1 mm. Similarly, a type 3 ECG is only suspected of BrS. Besides these, we collected other atypical clinical manifestations of patients as supportive diagnoses, including a first-degree atrioventricular block, atrial fibrillation, positive late ventricular potentials, and fragmented QRS. Furthermore, other causes of the ST-segment elevation were excluded, such as an atypical right bundle branch block, ventricular hypertrophy, early repolarization, acute pericarditis/myocarditis, acute myocardial ischemia or infarction, arrhythmogenic right ventricular dysplasia, hypothermia, dissecting aortic aneurysm, pulmonary thromboembolism, and Duchenne muscular dystrophy and so on [17].

Sanger Sequencing
Sanger sequencing was used to confirm all the rare variants screened by the above methods. The PCR primers were designed by Primer Premiers 5.0 and listed in Table S1.

Waterfall Plot and Needle Plot
A summary waterfall plot of the variants in the genes associated with BrS was generated using the R package "maftools" (R Foundation for Statistical Computing, Vienna, Austria). A mutation needle plot of SCN5A was created with MutationMapper (https://www.cbioportal.org/mutation_mapper, accessed on 6 July 2022).

Protein 3D Structure Prediction
The structural change in the protein by the substitution of the amino acid was predicted using Missense3D (http://missense3d.bc.ic.ac.uk/missense3d/, accessed on 27 August 2022). This tool uses three-dimensional structural information from experimentally determined protein models to predict the consequences of amino acid substitutions. The variant was analyzed using the experimentally determined structure of the sodium channel protein type 5 subunit alpha (UniProt ID: Q14524, PDB code: 6LQA).

Statistical Analysis
Continuous variables were expressed as the median (Q1-Q3). Continuous variables were tested for normality using the Kolmogorov-Smirnov and Shapiro-Wilk tests. For comparing the differences in the groups, continuous variables with a normal distribution were compared by an independent samples t-test or a one-way ANOVA, with Bonferroni correction for multiple comparisons, while continuous variables with an abnormal distribution were compared by the Wilcoxon rank-sum test or Kruskal-Wallis test with Bonferroni correction for multiple comparisons. Categorical variables were expressed as the total numbers (percentages). The chi-square test or Fisher exact test was used to compare the categorical variables with Bonferroni correction if required by multiple comparisons. A statistical analysis was conducted with the SPSS version 23. GraphPad Prism 8 was used to evaluate the significance between the groups. A 2-tailed p-value < 0.05 was considered to be statistically significant.

Genetic Characteristics Analysis
As specific SCN5A mutations are linked to cardiac conduction disorders and electrocardiographic phenotypes, we analyzed the clinical characteristics of patients with or without SCN5A variants. We found that patients with SCN5A variants had a longer Pwave duration (112 [92-120] ms and 98 [87-102] ms, respectively; p = 0.038) and a higher RV1 + SV5 (0.76 [0.55-0.90] mV and 0.49 [0.31-0.72] mV, respectively; p = 0.037) than patients without SCN5A variants (Table S5). Moreover, SCN5A genetic-positive patients tended to present a longer QTc interval and larger T axis deviation than negative patients, though the results were not significant (Table S5).
Next, we divided these 59 patients into three groups based on the mutation types: (1) P/LP: patients with P/LP mutations (n = 8, 13.56%); (2) VUS: patients with uncertain significant variants (n = 17, 28.81%); and (3) Table 3). We further analyzed the clinical and electrocardiographic characteristics of the three subgroups in patients with a type 1 BrS ECG (Table S6)

Genetic Characteristics Analysis
As specific SCN5A mutations are linked to cardiac conduction disorders and electrocardiographic phenotypes, we analyzed the clinical characteristics of patients with or without SCN5A variants. We found that patients with SCN5A variants had a longer Pwave duration (112 [92-120] ms and 98 [87-102] ms, respectively; p = 0.038) and a higher RV1 + SV5 (0.76 [0.55-0.90] mV and 0.49 [0.31-0.72] mV, respectively; p = 0.037) than patients without SCN5A variants (Table S5). Moreover, SCN5A genetic-positive patients tended to present a longer QTc interval and larger T axis deviation than negative patients, though the results were not significant (Table S5).
Next, we divided these 59 patients into three groups based on the mutation types: (1) P/LP: patients with P/LP mutations (n = 8, 13.56%); (2) VUS: patients with uncertain significant variants (n = 17, 28.81%); and (3) Table 3). We further analyzed the clinical and electrocardiographic characteristics of the three subgroups in patients with a type 1 BrS ECG (Table S6) Figure S1E) compared with non-carriers. However, in the type 2/3 BrS ECG group, we did not find any significant difference (Table S6).

Clinical and Genetic Features of Four Probands Carrying Novel P/LP Mutations
In our cohort, we screened eight novel variants, and four of them were classified as P/LP mutations: one was a titin (TTN) mutation, p.Q3508X (c.10522C > T); one was a KCNH2 mutation, p.A990G (c.2969C > G); and two were SCN5A mutations, p.G1220E (c.3659G > A) and p.D372H (c.1114G > C) (Table S4). All four mutations were absent from the controls in ExAc and gnomAD and were predicted deleterious with more than eight bioinformatic tools (Table S4). The details of these mutation carriers were as follows: Case 1: The first mutation was p.Q3508X, a nonsense mutation in TTN (c.10522C > T) ( Figure 3E, Table S4). The carrier was a 66-year-old male, diagnosed with a right brachial plexus injury without any other symptoms. His 12-lead ECG exhibited a coved-type ST elevation in the V1 lead and a saddleback-type ST elevation in the V2 lead ( Figure 3A). The novel c.10522C > T variant in the TTN gene was classified as pathogenic according to the ACMG.  Case 2: The proband, a 46-year-old man, was admitted to the hospital for a recurrent fever, which induced a type 2 BrS ECG in the V2 lead ( Figure 3B). He was asymptomatic and did not receive a drug provocation test, but his nephew died suddenly at night when he was 16. We identified a heterozygous p.A990G (c.2969C > G) in the KCNH2 by genetic screening ( Figure 3F, Table S4). However, more detailed clinical and genetic information about his relatives was not available. The novel c.2969C > G variant in the KCNH2 gene was classified as being likely pathogenic according to the ACMG.
Case 3: The third mutation, a heterozygous p.G1220E (c.3659G > A) of SCN5A, was identified in an asymptomatic 40-year-old man ( Figure 3G, Table S4). His 12-lead ECG Case 2: The proband, a 46-year-old man, was admitted to the hospital for a recurrent fever, which induced a type 2 BrS ECG in the V2 lead ( Figure 3B). He was asymptomatic and did not receive a drug provocation test, but his nephew died suddenly at night when he was 16. We identified a heterozygous p.A990G (c.2969C > G) in the KCNH2 by genetic screening ( Figure 3F, Table S4). However, more detailed clinical and genetic information about his relatives was not available. The novel c.2969C > G variant in the KCNH2 gene was classified as being likely pathogenic according to the ACMG. Case 3: The third mutation, a heterozygous p.G1220E (c.3659G > A) of SCN5A, was identified in an asymptomatic 40-year-old man ( Figure 3G, Table S4). His 12-lead ECG showed a coved-type ST elevation in the V1 lead and a saddleback-type ST elevation in the V2 lead ( Figure 3C). The novel c.3656G > A variant in the SCN5A gene was classified as being likely pathogenic according to the ACMG.
Case 4: A heterozygous p.D372H (c.1114G > C) of the SCN5A was identified in a 48-year-old man ( Figure 3H, Table S4) who was admitted to the hospital for sudden syncope. This patient had sleep apnea syndrome and a paroxysmal II degree I atrioventricular block. His 12-lead ECG revealed a significant coved-type ST elevation ( Figure 3D). His father died at 81 years old without any cardiovascular disease. His mother had a history of syncope and died suddenly at the age of 50 due to unexplained heart discomfort ( Figure 3I). We also performed genetic testing on his sister, son, and daughter. The patient's sister and daughter had no gene variants or symptoms ( Figure 3I). Only his son carried the same variant, suffering a first-degree atrioventricular block, and a slight ST elevation ( Figure 3J). The son showed a longer P-wave duration (116 ms), QRS duration (152 ms), T-wave duration (200 ms), and PR interval (212 ms). Furthermore, we used Missense3D to detect the structural change due to the substitution of the amino acid. The Missense3D tool predicted that this substitution replaced a buried negative-charged residue (ASP, RSA 1.8%) with a positive-charged residue (HIS) ( Figure 3K). Thus, the novel c.1114G > C variant in the SCN5A gene was classified as being likely pathogenic according to the ACMG.

Discussion
The main findings in this study were as follows: (1) we reported the spectrum of genetic variations in 29 BrS-susceptibility genes in a suspected BrS cohort from China; (2) we identified one novel mutation in TTN, one novel mutation in KCNH2, and two novel mutations in SCN5A; and (3) we found that the electrocardiographic axes should also be considered when predicting the risk of arrhythmias in BrS patients.
BrS can be found all over the world, but is more prevalent in Asia, including a higher prevalence of a type 1 BrS ECG (0-0.36%) and a type 2/3 BrS ECG (0.12-2.23%) [22]. The risk prediction for an SCD is a key issue in the management of patients with BrS. It is well known that males [23,24], a spontaneous type 1 BrS ECG [25], the presence of symptoms (such as arrhythmic syncope and documented VT/VF) [26], SCN5A mutations [27], and various electrocardiographic markers [28] are shown to be significant predictors of an SCD. To date, two principal hypotheses have been proposed: the repolarization hypothesis and the depolarization hypothesis [29], which may together lead to BrS under the influence of other factors such as sex, age, genetic background, and a fever.
In this study, we enrolled a Chinese cohort suspected of having BrS and carried out multiple subgroup analyses to compare their clinical and electrocardiographic characteristics. Depolarization abnormalities were prominent in our study, including a prolonged QRS duration, P-wave duration, and a decreased QRS amplitude. The repolarization abnormalities were mainly exhibited as a QT or QTc interval prolongation. The details are described as follows: First, we compared the common ECG parameters in patients with a spontaneous type 1 BrS ECG and patients with a type 2/3 BrS ECG. Consistent with the previous report [25], compared with the type 2/3 BrS ECG group, patients with a type 1 BrS ECG exhibited a prolonged QRS duration and QTc interval. A prolonged QRS is attributed to sodium current dysfunction in the conduction system and specifically in the His-Purkinje system, which is linked to a poor prognosis, as has been confirmed by several studies [30,31]. A QTc prolongation, reflecting a delayed cellular repolarization, has been associated with an increased risk of VT/VF and an SCD in BrS [32,33]. The results indicated that the association between a spontaneous type 1 BrS ECG and a delayed activation [34] could lead to a more severe phenotype, as we showed. In addition, we found that males had a longer QRS duration than females, whereas females exhibited longer QT and QTc dura-tions. Similarly [33], the symptomatic group displayed longer QT and QTc intervals and suffered more arrhythmic events with a higher rate of ICD implantation. The above results indicated that males, patients with a spontaneous type1 BrS ECG, and symptoms likely had a greater risk of an SCD.
Next, we explored the differences in ECG features between groups of 59 patients divided by their mutation types. Consistent with previous studies [35,36], we observed a greater P-wave duration and QTc interval in the SCN5A variant carriers than in the noncarriers. It has been confirmed that sodium channels dysfunction or downregulatory can impair atrial and ventricular conduction [37]. In addition, we found a higher RV1 + SV5 in patients with SCN5A variants. RV1 + SV5 indicates right ventricular hypertrophy if the value is greater than 1.05 mV. While the values were within the normal ranges in our patients, carriers tended to show a delayed right ventricular conduction, as reported in [38]. Moreover, we observed that patients with SCN5A variants had a larger T axis deviation to the right. The T axis, an index of primary repolarization abnormality, could be affected by an action potential duration shortening or prolongation in any ventricular region. Thus, the T axis has been reported as a strong predictor of fatal and nonfatal cardiac events [13,39], even when the value is greater than 45 • [40]. In our study, the larger T axis deviation to the right in patients with SCN5A variants might also suggest a higher risk of arrhythmic events.
Without any significant difference among three subgroups of all patients, we further explored the electrocardiographic characteristics in patients with a spontaneous type 1 ECG and found marked depolarization abnormalities of the right ventricle in patients with pathogenic mutations, including an increased QRS duration, decreased S-wave amplitude in lead V1, and decreased RV5 + SV1. Moreover, patients with P/LP mutations showed a prominent deviation to the left of the R axis. This is an index for cardiac conduction, which had been reported with a left shift in males [41].
Here, we described the genetic profile in our suspected BrS cohort. The top three genes were SCN5A (20%), SCN10A (8%), and DSP (7%). The SCN5A gene encodes the α-subunit of the cardiac voltage-gated sodium channel (Nav1.5) protein, which consists of four homologous domains (DI-DIV) that are connected by intracellular linkers. Each domain contains six transmembrane-spanning segments (S1-S6). We identified five pathogenic mutations in SCN5A, accounting for 8% of BrS patients, and four of them were localized to the transmembrane and pore-forming domains, similar to a previous report [10,14]. SCN5A is responsible for initiating the cardiac action potential. Pathogenic mutations result in a sodium channel dysfunction, which slows the impulse conduction throughout the myocardium [42]. The burden of a rare variation in SCN10A varied across the regions, from 2.5% in Japanese probands [43] to 16.7% in American probands [44]. Different from the previous study [45], we found that fewer patients carried variants in CACNA1C (2%), whereas more carried variants in DSP (7%) and HCN4 (3%). In addition, we found two nonsense mutations in TTN (3%) and classified them as pathogenic mutations. TTN is associated with several diseases, including inherited arrhythmias. Currently, only one frameshift mutation in TTN has been reported, which is likely to be pathogenic for BrS [46].
In summary, we reported four novel P/LP mutations in our cohort. However, it is important to determine the clinical significance by functional studies. All the results showed that Chines BrS patients had a different spectrum of genetic variations.

Study Limitations
Several limitations of the study should be noted. First, this was a retrospective study, and we did not perform drug challenges for patients with a type 2/3 BrS ECG. This might hamper their clinical diagnosis and might mask some significant differences between the groups. Second, most patients underwent target sequencing rather than WES. As we all know, the genetic background of more than 60% of BrS patients remains unclear, thus it is important to screen novel disease-causing genes for BrS. Third, we have to acknowledge that the study scale is relatively small, and multiple subgroup analyses may lead to an over-interpretation of the results. Finally, we did not conduct functional studies to determine the pathogenicity of the novel variants.

Conclusions
In this study, we compared the clinical and electrophysiologic characteristics of suspected Chinese BrS patients grouped by the types of BrS ECG, sex, symptoms, and genetic features. Furthermore, we identified four novel pathogenic mutations in the cohort and showed a representative pedigree. Our results suggest that BrS patients have significant depolarization and repolarization abnormalities, which may increase the risk of arrhythmic events and SCD.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/jcdd9110369/s1, Figure S1: Comparision of electrocardiographic characteristics between P/LP and Negative groups in patients with type 1 BrS ECG; Table S1: PCR primers for Sanger sequencing for all rare variants; Table S2: Clinical and ECG characteristics of the included subjects classified by sex; Table S3: Clinical and ECG characteristics of the included subjects classified by symptoms; Table S4: Rare variants of the included subjects; Table S5: Clinical and ECG characteristics of the included subjects classified by SCN5A variants; Table S6: Clinical and ECG characteristics of different ECG groups classified by mutation pathogenicity.  Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data of this study are available on request from the corresponding authors.