A novel PRRX1 loss-of-function variation contributing to familial atrial fibrillation and congenital patent ductus arteriosus

Abstract Atrial fibrillation (AF) represents the most common type of sustained cardiac arrhythmia in humans and confers a significantly increased risk for thromboembolic stroke, congestive heart failure and premature death. Aggregating evidence emphasizes the predominant genetic defects underpinning AF and an increasing number of deleterious variations in more than 50 genes have been involved in the pathogenesis of AF. Nevertheless, the genetic basis underlying AF remains incompletely understood. In the current research, by whole-exome sequencing and Sanger sequencing analysis in a family with autosomal-dominant AF and congenital patent ductus arteriosus (PDA), a novel heterozygous variation in the PRRX1 gene encoding a homeobox transcription factor critical for cardiovascular development, NM_022716.4:c.373G>T;p.(Glu125*), was identified to be in co-segregation with AF and PDA in the whole family. The truncating variation was not detected in 306 unrelated healthy individuals employed as controls. Quantitative biological measurements with a reporter gene analysis system revealed that the Glu125*-mutant PRRX1 protein failed to transactivate its downstream target genes SHOX2 and ISL1, two genes that have been causally linked to AF. Conclusively, the present study firstly links PRRX1 loss-of-function variation to AF and PDA, suggesting that AF and PDA share a common abnormal developmental basis in a proportion of cases.


Introduction
Atrial fibrillation (AF), characteristic of rapid and disorganized electrical activation and inefficient contraction of the atria, is the most common form of clinical dysrhythmia that affects approximately 1% of the general population globally (Tian et al., 2020;Zhang et al., 2021).Its global prevalence is low in individuals aged <40 years, but increases abruptly beyond the age of 65 years, reaching over 10% in subjects ≥80 years old (Huang et al., 2020).The lifetime for development of AF is ~25% in subjects ≥40 years of age and ~37% in those ≥55 years of age (Lloyd-Jones et al., 2014;Weng et al., 2018).Given that about one third of the total AF population is silent or subclinical asymptomatic, the global prevalence of AF is certainly underestimated (Dilaveris and Kennedy, 2017).AF confers a significantly increased risk for ischemic or hemorrhagic stroke, dementia, venous thromboembolism, acute myocardial infarction, congestive heart failure and premature death with substantial socioeconomic costs (Kornej et al., 2020).Nevertheless, existing therapeutic regimens for AF are considerably limited in effectiveness and seldom curative, which reflects a poor understanding of the molecular mechanisms underpinning this complex supraventricular arrhythmia (Huang et al., 2020).
Epidemiological investigations have revealed that environmental risk factors predispose to the occurrence and perpetuation of AF, such as advancing age, obesity, obstructive sleep apnea, diabetes mellitus, arterial hypertension, valvular heart diseases, coronary artery disease, dilated cardiomyopathy, hyperthyroidism, heart failure, smoking, alcohol consumption, psychological stress and extreme sports (January et al., 2014;Kornej et al., 2020).However, in ~30% of patients, no well-established cardiovascular pathologies or precipitating factors for AF can be identified, which suggests possible genetic basis underlying AF (Ragab et al., 2020).During the past two decades, multiple epidemiological investigations have demonstrated familial aggregation of individuals with AF and the heritability of AF has been estimated to be as high as 62%, highlighting a strong heritable component responsible for AF (Roselli et al., 2020).By genotyping with a few hundred polymorphic microsatellite markers scattered throughout the genome and genetic linkage analysis of AF families, Brugada and his partners located the first locus for AF at human chromosome 10q22-q24 (Brugada et al., 1997).Subsequently, similar genetic studies linked more genetic loci to AF, including human chromosome 5p13, 5p15, 6q14-16, 10p11-q21 and 20q12-13 (Ellinor et al., 2003;Oberti et al., 2004;Volders et al., 2007;Darbar et al., 2008).By genetical analysis of a large Chinese family inflicted with AF, Chen and his coworkers mapped a new locus for AF to chromosome 11p15.5 and in this chromosomal region discovered the first AF-causative gene, S140G-mutant KCNQ1, which encodes an α subunit of voltage-gated potassium channel (Chen et al., 2003).Functional analysis of the S140G-mutant KCNQ1 unveiled a gain-of-function impact on the currents of KCNQ1/ KCNE1 and KCNQ1/KCNE2 channels, which significantly shorten the action potential duration of atrial myocytes thereby increasing the vulnerability to AF (Chen et al., 2003).Up to now, in addition to the association of ~140 genetic loci with increased predisposition to AF revealed by genome-wide association studies (Kim et al., 2021), rare variations in over 50 distinct genes have been discovered to contribute to AF, amidst which the majority encode cardiac potassium ion channels, sodium channels, gap junction channels, calcium channels, signaling molecules, structural proteins and transcription factors (Choi et al., 2020;Ghazizadeh et al., 2020;Hansen et al., 2020;Huang et al., 2020;Jiang et al., 2020;Ragab et al., 2020;Roselli et al., 2020;van Ouwerkerk et al., 2020;Wu et al., 2020;Yang et al., 2020;Chalazan et al., 2021;Lazarte et al., 2021;Li et al., 2021a, b;Ziki et al., 2021).Interestingly, multiple variations in or near the PRRX1 gene, has recently been associated with an enhanced susceptibility to AF in humans (Tucker et al., 2017;Guo et al., 2021;Wu et al., 2021).However, due to pronounced genetical heterogeneity, the genetic determinants underlying AF remain largely elusive.This study was sought to identify a novel genetic variation predisposing to AF.

Recruitment and clinical evaluation of study participants
For this investigation, a three-generation family affected with AF and congenital patent ductus arteriosus (PDA) was identified, from which 18 available family members were enlisted.A total of 306 unrelated healthy volunteers, who had neither AF nor congenital heart defect (CHD), the most common type of birth defects (Oliveira-Brancati et al., 2020;Virani et al., 2021), were enrolled as control subjects.All study participants experienced a comprehensive clinical assessment, including review of medical histories, physical examination, electrocardiography and echocardiography as well as routine laboratory tests.The healthy control individuals were exactly matched with the cases for gender, ethnicity and age.Clinical diagnosis and classification of AF or CHD were made as previously described (January et al., 2014;Benjamin et al., 2019;Wang et al., 2020;Zhao et al., 2021).This case-control research was carried out in conformity with the ethical tenets outlined in the Declaration of Helsinki and was approved by the Medical Ethics Committee of Shanghai Chest Hospital (with an approval number of KS1101).Prior to collection of peripheral venous blood samples, informed consent was provided by the study participants or their parents.

Whole-exome sequencing and bioinformatical analysis
Genomic DNA was extracted from the venous blood leucocytes of every test person utilizing a genomic DNA extraction kit (Promega, USA).Whole-exome sequencing (WES) and bioinformatical analysis in five affected family members (I-1, II-1, II-6, III-2 and III-4, Figure 1A) and four unaffected family members (I-2, II-2, II-5 and III-1, Figure 1A) were performed as described previously (Di et al., 2020;Qiao et al., 2020;Linhares et al., 2021;Wang et al., 2021;Xian et al., 2021).In brief, each exome library was constructed using 5 μg of genomic DNA from a study subject, enriched by ligation-mediated polymerase chain reaction (PCR) and captured with the SureSelect Human All Exon V6 Kit (Agilent Technologies, USA).Each exome library was sequenced on the Illumina HiSeq 2000 Genome Analyzer (Illumina, USA) by utilizing the HiSeq Sequencing Kit (Illumina, USA) as per the manufacturer's instructions.Raw image files were processed using the software Pipeline (Illumina, USA) to call bases and the sequences of each subject were generated as a set of reads.By using the Burrows-Wheeler Aligner (BWA) software (Li andDurbin, 2009, 2010), sequencing reads were aligned to the sequences of referential human genome (GRCh37/hg19).Variation calling was performed with the SAMtools (Sequence Alignment/Map Tools, version 0.1.18)software (Li et al., 2009;Li, 2011) and the Genome Analysis Toolkit (GATK, version 4.0.10.1)software (McKenna et al., 2010).The genetic variants that passed the pedigree analysis with any reasonable inheritance pattern of AF and PDA were annotated with the ANNOVAR (annotation of variance, version 20170221) software (Wang et al., 2010).Deleterious variations with a minor allele frequency of <0.001 (in such databases as the Genome Aggregation Database and the Single Nucleotide Polymorphism database) annotated by ANNOVAR were selected as candidate disease-causing variants subject to confirmation by Sanger sequencing analysis in the whole family.The entire coding region and splicing donors/acceptors of the gene harboring a confirmed candidate causative variant were PCR-sequenced in all the available family members and 306 unrelated healthy persons.For an identified rare damaging variation, the Single Nucleotide Polymorphism database (https://www.ncbi.nlm.nih.gov/), the 1000 Genomes Project database (https://www.internationalgenome.org), the Human Gene Mutation Database (http://www.hgmd.cf.ac.uk/ac/index.php), and the UK Biobank database (https://www.ukbiobank.ac.uk/) were consulted to check whether it was novel.

Cellular transfection with expression plasmids and dual-luciferase assay
Hela cells were cultivated in Dulbecco's modified Eagle's medium (Invitrogen, USA) containing 10% fetal calf serum (Thermo Fisher Scientific, USA) together with 1% penicillin/streptomycin (Thermo Fisher Scientific, USA) in an incubator with an air of 5% CO₂ at 37 °C.Hela cells were grown in a 12-well plate at an initial density of 1×10 5 cells per well 24 h before transient transfection.Cells were transfected with various amounts of expression plasmids as described previously (Guo et al., 2021).The plasmid pGL4.75(Promega, USA) expressing renilla luciferase was co-transfected as an internal control to normalize transfection efficiency.The activities of firefly and renilla luciferases were measured on a luminometer (Promega, USA) employing a dual-luciferase reporter assay kit (Promega, USA).The activity of a promoter was expressed as fold activation of firefly luciferase relative to renilla luciferase.For each plasmid, three independent transfections were performed and the resultant data for promoter activity were presented as mean ± standard deviation (SD) of three independent transfection experiments.

Statistical analysis
Unpaired Student's t-test was applied to the comparison of promoter activities between two groups.A two-tailed p<0.05 was considered to indicate statistical difference.

Clinical characteristics of the pedigree with AF and PDA
As shown in in Figure 1A, a three-generation pedigree with high incidence of AF and PDA was recruited, which comprised 18 living family members, of whom 6 members, including 4 male members and 2 female members with a mean age of 42 years ranging from 19 to 71 years, were diagnosed with AF and PDA in terms of the electrocardiographic and echocardiographic findings.Within this family, AF and PDA were inherited in an autosomal-dominant mode with complete penetrance.Of note, in this pedigree AF began with paroxysmal episodes, but in two family members (members II-1 and II-6) AF became persistent and in one family member (member I-1) AF became permanent over time.No family members had well-established environmental risk factors prone to AF, such as arterial hypertension, valvular heart disease, coronary heart disease, pulmonary heart disease, hyperthyroidism, diabetes mellitus nor obstructive sleep apnea.The proband (member II-6), a forty-three-year-old female member with nineteen years of AF history, was hospitalized due to recurrent syncope and received a successful radiofrequency ablation of AF.The proband's other affected relatives had a history of taking anti-arrhythmic drugs but none of them underwent interventional treatment for AF at the time of enrollment.Additionally, catheter-based closure of PDA was performed in all the affected family members before 6 years of ages except for family member I-1, who underwent closure of PDA at the age of 22.The unaffected family members (six male members and six female members with an average age of 40 years varying from 15 to 68 years) had neither a history of AF nor a history of CHD, with their electrocardiograms and echocardiograms being normal.The clinical features of the pedigree members suffering from AF and PDA are provided in Table 1.

Discovery of a new causative variation in PRRX1
WES was conducted in five affected family members (I-1, II-1, II-6, III-2 and III-4, Figure 1A) and four unaffected family members (I-2, II-2, II-5 and III-1, Figure 1A), generating a mean of 22.9 Gb of sequence for each family member, with ~97% mapping to the referential human genome (GRCh37/hg19) as well as ~73% mapping to the target DNA sequences.An average of 18,632 exonic variations (range 17,105-19,148) per family member passed filtering by inheritance model, of which 12 heterozygous nonsense and missense variations passed filtering by ANNOVAR, shared by the five affected family members and predicted to be pathogenic variants, with minor allele frequencies of <0.001 (as summarized in Table 2).Sanger sequencing analysis in the family revealed that only the variant chr1:170,688,998G>T (GRCh37/hg19: NC_000001.10),equivalent to chr1:170,719,857G>T (GRCh38/hg38: NC_000001.11)or NM_022716.4:c.373G>T;p.(Glu125*) in the PRRX1 gene, was in co-segregation with AF and PDA in the whole family, with complete penetrance.The electropherogram traces exhibiting the heterozygous PRRX1 variation as well as its homozygous wild-type sequence (used as a control) are exhibited in Figure 1B.The schemas displaying the homeobox domains of wild-type and mutant PRRX1 proteins are shown in Figure 1C.The truncating variation was neither observed in 306 unrelated healthy individuals, nor found in the Single Nucleotide Polymorphism database, the Single Nucleotide Polymorphism database, the Human Gene Mutation Database, the 1000 Genomes Project database, and the UK Biobank database, indicating it was a novel variation.

Discussion
In the present investigation, a heterozygous PRRX1 variation, NM_022716.4:c.373G>T;p.(Glu125*),was found to be in co-segregation with AF and PDA in a family.The truncating variation was neither detected in 612 referential chromosomes nor retrieved in the Single Nucleotide Polymorphism database, the Single Nucleotide Polymorphism database, the Human Gene Mutation Database, the 1000 Genomes Project database or the UK Biobank database.Functional research unveiled that Glu125*-mutant PRRX1 lost transcriptional activation on the promoters of SHOX2 and ISL1, two genes where variations have been discovered to result in AF and CHD (Blaschke et al., 2007;Hoffmann et al., 2016Hoffmann et al., , 2019;;Li et al., 2018;Ma et al., 2019;Wang et al., 2019;Wu et al., 2020).These findings indicate that genetically defective PRRX1 contributes to AF and PDA in this family.
In humans, PRRX1 is localized to chromosome 1q24.2and encodes paired related homeobox 1, as a member of the paired homeobox-containing family of transcription factors (Grueneberg et al., 1992).The PRRX1 protein is highly expressed in the cardiovascular system throughout embryonic development, predominantly in the mesenchymal tissues, including the heart, endothoracic great arteries and pulmonary veins (Leussink et al., 1995;Bergwerff et al., 1998;Libório et al., 2011), a common source of ectopic beats triggering AF in the majority of patients (Haïssaguerre et al., 1998).It has been validated that PRRX1 regulates the epithelial to mesenchymal transition, a hallmark of human cardiovascular morphogenesis (Ocaña et al., 2012).Notably, in PRRX1-knockout mice, cardiovascular developmental malformations occurred, encompassing awkward curvature and abnormal positioning of the aortic arch, an aberrant retroesophageal right subclavian artery as well as a misdirected and elongated ductus arteriosus, highlighting the crucial role of PRRX1 in the proper development of vessels and perivascular matrices (Bergwerff et al., 2000).Moreover, a recent study has demonstrated that PRRX1 physically binds to the promoters of SHOX2 and ISL1 and transcriptionally activates the expression of SHOX2 and ISL1 (Guo et al., 2021), two key downstream target genes responsible for the normal development of the heart, especially for its pacing and conducting system (Blaschke et al., 2007;Liang et al., 2015;Vedantham et al., 2015;Galang et al., 2020) and variations in both SHOX2 and ISL1 have been causally linked to AF and CHD (Blaschke et al., 2007;Hoffmann et al., 2016Hoffmann et al., , 2019;;Li et al., 2018;Ma et al., 2019;Wang et al., 2019;Wu et al., 2020).In the present research, a new PRRX1 loss-offunction variation was discovered to lead to AF and PDA.Collectively, these observational results support that PRRX1 haploinsufficiency is involved in the molecular pathogenesis of AF and CHD in some cases.
Recently, multiple genome-wide association studies and a meta-analysis consistently revealed that a common single nucleotide polymorphism (rs3903239) about 63 kb upstream of the PRRX1 gene, a top genetic variation at the locus of AF on chromosome 1q24, was associated with significantly increased risk of AF in both Europeans and Asians (Tucker et al., 2017;Wu et al., 2021).Functional analyses unveiled that this variant diminished the transcriptional activity of the promoter of PRRX1, resulting in reduced expression of PRRX1 in human left atrial tissue (Tucker et al., 2017).Moreover, loss of PRRX1 was shown to shorten the action potential duration as well as effective refractory period in human atrial cardiomyocytes and zebrafish embryonic myocardium, forming a substrate vulnerable to AF (Tucker et al., 2017), which was further substantiated in a mouse model with deletion of the noncoding AF-associated genomic region (Bosada et al., 2021).Additionally, two loss-of-function variations in PRRX1 have been uncovered to cause familial AF (Guo et al., 2021).In this research, a new PRRX1 loss-of-function variation was identified to give rise to AF and PDA, therefore expanding the phenotypic spectrum linked to PRRX1 and supporting PRRX1 as a causative gene for AF and CHD.Notably, heterozygous loss-of-function variations in PRRX1 have already been described in patients with agnathia-otocephaly complex, a rare condition characterized by mandibular hypoplasia or agnathia, ear anomalies (melotia/synotia) and microstomia with aglossia (Dubucs et al., 2021).It is interesting that the same kind of variation has been associated with a quite different phenotype (AF) in the present study, which may be explained in part by the distinct genetic backgrounds.

Conclusions
This study firstly associates PRRX1 loss-of-function variation with AF and PDA in humans, which suggests that AF and PDA may share a common basis of anomalous cardiovascular development in a subset of cases, implying potential implications for early precise prophylaxis and improved prognostic risk stratification of patients affected with AF and PDA.

Figure 1
Figure 1 -A new PRRX1 variation predisposing to familial atrial fibrillation and congenital heart defect.(A) Pedigree structure of the family inflicted with atrial fibrillation and congenital heart disease."+", carriers of the PRRX1 variation; "-", non-carriers.(B) Sequence electropherogram traces showing the heterozygous PRRX1 variation (mutant) as well as its homozygous wild-type control base (wild type).A rectangle delimits a codon comprising three nucleotides.(C) Schemas exhibiting the structural domains of the PRRX1 proteins.NH2, amino-terminus; COOH, carboxyl-terminus.

Figure 3 -
Figure 3 -No transcriptional activation on the promoter of ISL1 by Glu125*-mutant PRRX1.Biological measurement of the transactivation of the ISL1 promoter-driven luciferase in cultivated Hela cells expressing various expression plasmids by wild-type PRRX1 (PRRX1) or Glu125*-mutant PRRX1 (Glu125*), alone or in combination, revealed that Glu125* lost the ability to transcriptionally activate the promoter of the ISL1 gene.Here # and ## mean p<0.0001 and p<0.0005, respectively, in comparison with wild-type counterpart (100 ng).

Table 2 -
Nonsynonymous variations in the candidate genes for familial atrial fibrillation and congenital patent ductus arteriosus identified by wholeexome sequencing analysis.