The genetic landscape of pediatric postural orthostatic tachycardia syndrome

Abstract

Background

Postural orthostatic tachycardia syndrome (POTS) is a complex disorder with serious health consequences, while its etiology remains largely elusive.

Objective

The purpose of this study was to investigate the genetic landscape of POTS using genomic approaches in a unique pediatric cohort.

Methods

We conducted a combined genome wide genotyping and whole exome sequencing (WES) study to systemically examine the molecular mechanisms of POTS pathogenesis. The patients were genotyped as two independent cohorts: a family cohort of 100 complete families and a case–control cohort of 207 unrelated European cases and 4063 ethnicity-matched control subjects. The WES component consisted of a subset of the genotyped subjects, including 87 unrelated European cases and 2719 unrelated European control subjects.

Results

The heterogeneous phenotype of POTS made achieving genome-wide significance improbable. Instead, 5670 SNPs with nominal significance (P < 0.05) were identified in both the family and case–control cohorts, with effects in the same direction. We conducted an over-representation analysis (ORA) by considering all genes that showed nominal significance. The ORA identified gene sets linked to cell–cell junction, early estrogen response, and substance-related disorders with statistical significance. Moreover, WES revealed 55 genes with genome-wide significance through rare variant burden analysis, harboring 92 variants classified as pathogenic or likely pathogenic by ClinVar.

Conclusions

This study showcases the complex interplay between common and rare genetic variants in POTS development, marking a pioneering step forward in deciphering its complex etiologies. The insights from this research enrich our understanding of POTS, offering new avenues for precise treatment strategies and highlighting areas for further research.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10286-025-01110-2.

Introduction

Postural orthostatic tachycardia syndrome (POTS) is a dysautonomia condition characterized by persistent excessive upright tachycardia upon assuming an upright position, without concurrent orthostatic hypotension [6, 27, 86, 91]. Chronic orthostatic intolerance caused by POTS leads to severe functional impairment and psychological distress to the patients, seriously affecting patients’ lives. Patients with POTS often experience a wide array of symptoms, including severe lightheadedness, palpitations, cognitive impairment, debilitating fatigue, disruptions in sleep patterns, varying levels of pain, recurrent headaches, and gastrointestinal disturbances. POTS was only formally recognized as a distinct medical condition in 1993 [84]. POTS affects about 0.2% to 1.0% of the US population, and is more frequently seen in women [6, 27, 28]. The connection with sex remains poorly understood. Both children and adults can be affected, while a majority of patients were diagnosed prior to reaching menopause [7, 58]. Tachycardia in POTS can be due to any of the many factors affecting venous return and cardiac stroke volume, e.g., inability to maintain peripheral vascular tone, low blood volume, or increased pooling in the splanchnic circulation and extremities [92]. While anxiety is common, it is not considered as a significant causal factor. Despite extensive research, the complex pathophysiology of POTS remains only partially understood.

In light of the elusive nature of POTS etiology, our study employed an integrative OMICS approach to explore the molecular underpinnings of its pathogenesis in a unique pediatric cohort. Previous research has suggested a genetic link to POTS, such as the identification of the A457P mutation in the SLC6A2 gene (encoding norepinephrine transporter) causing POTS [85]. In our study, genome-wide genotyping of common variants and exome sequencing for rare variants were applied. Unlike traditional genome-wide association study (GWAS) phenotypes, it is unlikely that any loci will reach genome-wide significance due to the heterogeneous phenotype of POTS. Increasing the sample size may not be effective as it also complicates the heterogeneity. Instead, focusing on common variants, we aimed to identify gene sets tagged by common single nucleotide polymorphisms (SNP) that are overrepresented in POTS patients. Through the exome sequencing, we aimed at identifying rare coding genetic variants in the candidate genes uncovered that may contribute to the disorder’s heterogeneous etiology. By centering on pediatric patients with POTS, our study offers a unique perspective on the genetic landscape of this condition, potentially revealing insights into its complex mechanisms.

Methods

GWAS

Subjects

Participants were enrolled through the POTS Program at the Children’s Hospital of Philadelphia (CHOP). The diagnosis of POTS is based on the definition outlined in a previous publication by the lead clinician of this study [6]. Patients with POTS aged 18 years or younger at the time of diagnosis were eligible for this study. We invited patients and their families to join this study through letters or emails. Those with DNA samples available from both parents were included in the family cohort. Unrelated patients lacking parental DNA samples were included in the case–control cohort. None of the non-POTS control subjects had a history of diagnosis with chronic fatigue syndrome or orthostatic intolerance. All subjects were of European ancestry. We obtained informed consent from all subjects, or if subjects were under 18 years, from a parent and/or legal guardian with assent from the child if 7 years or older. The CHOP Institutional Review Board (IRB) approved this study.

Genotyping

The genotyping was done using the Illumina Infinium Global Screening Array (Illumina, San Diego, CA) with > 700,000 SNPs genotyped. Altogether, 93.5% SNPs had a calling rate > 99%, and the average calling rate of each DNA sample was 98.2%. Genome-wide imputation was done with the TOPMed Imputation Server (https://imputation.biodatacatalyst.nhlbi.nih.gov/#!) using the TOPMed (Version R2 on GRC38) Reference Panel. Altogether, 19,537,894 autosomal single nucleotide variants (SNV) with quality R² ≥ 0.3 were included in this study.

Genotyping data analysis

In this study, the family cohort was tested by transmission disequilibrium test (TDT), which is immune to spurious associations from population stratification [25]. Kinship between family members in the family cohort was validated by identity by descent (IBD) analysis on the basis of the auto-chromosomal genotyping data. Loci with Mendelian errors > 3 were removed from association test. In addition, previous study has emphasized that replicated sequences in autosomes and sex chromosomes cause sex-related bias on the genotyping of autosomal SNPs [30]. As an additional quality filter, we tested sex effect by comparing mothers and fathers in the family cohorts. All SNPs with sex effect P < 0.05 were removed. For the case–control cohort, unrelated cases were identified of European ancestry (EA) by principal component analysis (PCA) with genome-wide SNP markers, and were confirmed of non-relationship by identical-by-descent (IBD) analysis. European control subjects were selected by matching ethnicity on the basis of the PCA analysis. Correction for population stratification was done by logistic regression using the first ten principal components as covariates [75]. The IBD analysis, TDT test, and case–control association test were done using the PLINK software v 1.9 [76]. All SNPs with Hardy–Weinberg Equilibrium (HWE) P < 0.01 in European control subjects were removed from further analysis. Gene-based association test was done by the Versatile Gene-based Association Study—2 version 2 (VEGAS2v02) software [37, 63].

WES

Subjects

On the basis of the genotyping data, we identified 87 unrelated European cases (61 female and 26 male) with non-relationship validated by the IBD analysis, and European ancestry validated by the PCA analysis. The cases were compared with the Non-Finish European (NFE) population in the Exome Aggregation Consortium (ExAC) database [44], using the Test Rare vAriants with Public Data (TRAPD) software [35]. Considering the potential inflation with the public database controls, the cases were further compared with 2719 unrelated European non-POTS control subjects sequenced by WES at the Center for Applied Genomics (CAG) of the Children’s Hospital of Philadelphia (CHOP).

Library Preparation and Sequencing

Paired-end sequencing was performed on the Illumina NovaSeq 6000 platform (Illumina, San Diego, CA), using an S4 flowcell with run parameters of 101 × 10 × 10 × 101 [Read 1 × Index 1 (i7) × Index 2 (i5) × Read 2]. Demultiplexing, alignment, and variant calling processes were performed on the Illumina DRAGEN Bio-IT Platform (version 3.3.7) using the 1000 Genomes Project Reference Human Genome Sequence (hs37d5). Alignment metrics were calculated using the Picard (version 2.18.27) CollectHsMetrics tool.

Burden analysis of variants of interest (VOI) and pathogenic (P) or likely pathogenic (LP) Variants

The genetic variants, which have minor allele frequency (MAF) greater than 0.001 in the NFE population based on the ExAC database [44], have been excluded. Functional candidate VOIs were selected by the prediction results with at least 1 of a number of genetic variant prediction software, i.e., SIFT_pred = "D" or Polyphen2_HDIV_pred = "D" or Polyphen2_HDIV_pred = "P" or Polyphen2_HVAR_pred = "D" or Polyphen2_HVAR_pred = "P" or LRT_pred = "D" or MutationTaster_pred = "A" or MutationTaster_pred = "D" or MutationAssessor_pred = "H" or MutationAssessor_pred = "M" or FATHMM_pred = "D" or PROVEAN_pred = "D" or MetaSVM_pred = "D" or MetaLR_pred = "D", on the basis of the annotation with ANNOVAR software [103]. The mutation burden in cases and controls were counted with the TRAPD software [35]. We have optimized the TRAPD algorithm with normalized genome coverage to capture causal variants with effects in the same directions [56]. Gene-wide burden test of the candidate variants in the cases was done by one-tailed Fisher exact test, compared with the ExAC NFE controls by dominant inheritance model. Multiple comparisons were corrected by Bonferroni correction. Assuming 21,306 protein-coding genes in human genome [71], the genome-wide significance of the burden test was defined as α = 0.05/21,306 = 2.347E-06. In further, deleterious variants were identified from the functional candidate variants according to the aggregated information by ClinVar annotation [48, 49], InterVar prediction [53], and the Human Gene Mutation Database (HGMD) classification [90].

Results

Subjects

All patients with POTS in this study were Caucasian, with the age of diagnosis ranging from 12 years old to 21 years old, and median diagnosis age (Q1, Q3) of 15.6 (13.2, 17.8) years [7]. The patients were evaluated as two independent cohorts, a family cohort and a case–control cohort. The family cohort included 114 POTS cases (including 28 male and 86 female patients) from 100 complete families. We included 62 unaffected siblings from these families in this study. Significant comorbidities were detected in 18 unrelated patients from the family cohort (Table 1). The case–control cohort included 207 unrelated cases (including 53 male and 154 female patients). Among the 207 cases, comorbidities were seen in 24 unrelated patients (Table 1). From the 207 cases, 194 unrelated cases (including 44 male and 150 female patients) based on genome-wide genotyping were compared with 4063 European control subjects for genetic association test. In these subjects, 14.2% of patients had a family member with POTS, with male patients being more likely to have an affected family member [8].

Table 1.

Comorbidities with POTS

Cohort	Comorbidities	Number of patients
The family cohort (case n = 114)
	Ehlers–Danlos syndrome (EDS)	11
	EDS + autoimmune alopecia	1
	Mast cell activation syndrome (MCAS)	1
	MCAS + Wolff-Parkinson-White syndrome + Leigh disease	1
	Scoliosis + Hashimoto’s thyroiditis + benign premature atrial contractions	1
	Crohn’s disease	1
	Post-concussion	1
	Benign Rolandic epilepsy	1
The case–control cohort (case n = 207)
	EDS	6
	EDS + eosinophilic esophagitis	1
	EDS + Gilbert syndrome	1
	EDS + IgA deficiency	1
	EDS + MCAS	1
	EDS + Chiari malformation + exercise-induced asthma + Asperger syndrome + gastroesophageal reflux + urticaria + and left duplicated ureter	1
	Multiple sclerosis	2
	MCAS	1
	Crohn’s disease	1
	Alport’s syndrome	1
	Asperger syndrome + seizure disorder	1
	Behҫet’s disease	1
	Post-concussion	1
	Hodgkin lymphoma	1
	Type 1 diabetes	1
	Neuromuscular disorder	1
	Congenital adrenal hyperplasia + von Willebrand’s disease + Hashimoto’s disease	1
	UTI + VUR + asthma + vitamin D insufficiency + Lyme disease	1

GWAS results

The inherently heterogeneous phenotype of POTS presented significant challenges. Achieving genome-wide significance for any particular genetic loci was improbable by increasing sample size. Instead, we conducted an over-representation analysis (ORA) by considering all genes that showed nominal significance. Our conclusions were only drawn on the basis of ORA analysis with solid statistical evidence. In this study, 5670 SNPs were identified of potential association signals with P < 0.05 in both the family cohort and the case–control cohort, with effects in the same direction. The summary statistics are available in Supplementary Table 1. As shown, none of these loci showed genome-wide significance. In the family cohort, we identified one proband with (MCAS + Wolff-Parkinson-White syndrome + Leigh disease) and another proband with (Scoliosis + Hashimoto’s thyroiditis + benign premature atrial contractions). The phenotypic heterogeneity in these two cases could potentially dilute true causal effects or lead to false positives. To address this concern, we reanalyzed the GWAS in the family cohort, excluding these two families. The results show that the majority of the associations in Supplementary Table 1 (5670 SNPs) remain statistically significant 5,024 (88.6%). Of the SNPs that lost significance, all had nominal significance, with the lowest P-value being 0.01586. Additionally, among the 646 SNPs that lost significance, 367 SNPs (6.5% of the 5670 SNPs) have other SNPs within the same gene locus that remain statistically significant. This new analysis can be considered a sensitivity analysis, as it tests the robustness of our findings by excluding the two probands with potentially confounding conditions. The results indicate that 95.1% of the SNPs either remain significant or are located within gene loci containing other significant SNPs, suggesting that the overall conclusions of our study are not materially affected by the exclusion.

Consequently, we performed gene-based association test at these loci. As presented in Supplementary Table 2, 716 genes showed association P < 0.05 in both the family cohort and the case–control cohort, a number significantly higher (P = 1.81E-128) than the 53 genes expected by chance (i.e., 21,306 × 5% × 5%), assuming there are 21,306 human coding genes. Using the WebGestalt (WEB-based Gene SeT AnaLysis Toolkit) web tool [102], over-representation analysis (ORA) of the 716 genes by the DisGeNET approach [72], the Gene Ontology (GO) Cellular Component [14], the GO molecular function [14], the MSigDB Hallmark gene sets [54], and the Human Phenotype Ontology (HPO) [82] underscored several gene sets of statistical significance (False Discovery Rate (FDR) < 0.1), with common genetic variants contributing to the susceptibility of POTS (Table 2).

Table 2.

Over-representation analysis of the 716 genes showed nominal significance in both the family cohort and the case–control cohort

a. By gene ontology cellular component
Gene set	Description	P-value	FDR	Genes
GO:0005911	Cell–cell junction	4.34E-06	0.000747	AJAP1, ANK3, APP, ATP2A2, BAIAP2L2, CD2AP, CDH13, CDH15, CDH22, CDH4, CDH8, CNN3, CNTNAP2, COL13A1, CTNNA3, DSG1, EPB41L3, F11R, FBF1, FRMD4A, GRB2, KIFC3, LYN, NCK1, NDRG1, NFASC, PAK4, PDZD2, PKP4, PPL, PRKCZ, SLC2A1, TJP2, TJP3, UBN1, VASP, WASF2
GO:0097060	Synaptic membrane	1.54E-05	0.001327	ANK3, ANKS1B, ATP2B2, ATP2B4, CADPS2, CDH8, CHRNA3, CHRNA4, CNR1, CNTN1, COL13A1, CPEB1, DENND1A, DGKI, DISC1, DLG2, DLGAP1, GABRG3, GRIK4, KCNB1, KCNC1, KCNJ3, LRRC4C, LRRTM4, NTRK3, PI4K2A, ROGDI, SEMA4F, SHC4, SHISA6, SLC1A6, SLC8A3, SYNJ2BP, SYT6, UNC13C
GO:0043025	Neuronal cell body	8.58E-05	0.004921	ADA, ADAM21, ADCY10, APP, ASIC2, BRD1, CACNA1B, CHRNA3, CHRNA4, CNN3, CNTNAP2, COBL, CRHBP, CYGB, DAB2IP, DENND1A, DGKI, FZD3, GIP, KCNB1, KCNC1, KCNN3, KNDC1, LRP8, MBP, MYO1D, NMNAT3, NPTXR, PCP2, PDE9A, PI4K2A, PRKCZ, RBFOX3, ROGDI, SLC8A3, TGFB2
GO:0031252	Cell leading edge	0.000201	0.008647	ABLIM1, AIF1L, APBB2, APP, CD2AP, CNTNAP2, COBL, CTNNA3, CUBN, EPB41L3, FERMT1, FGD2, GABRG3, IQGAP2, JMY, KCNB1, KCNC1, MACF1, MYO1D, MYO1G, PDE9A, PIEZO1, PRKCZ, SHISA6, SNTG1, SRC, SYNE2, TPM1, VASP, WASF2
GO:1,990,351	Transporter complex	0.00036	0.012401	ANO2, CACNA1B, CACNA1E, CACNA2D4, CALM1, CATSPERB, CHRNA3, CHRNA4, CNGB1, CNTNAP2, CUBN, DLG2, DPP10, DPP6, GABRG3, GRIK4, KCNB1, KCNC1, KCNJ3, KCNJ6, KCNK6, RYR2, SCN8A, SHISA6, SYNJ2BP, TTYH1
GO:0031253	Cell projection membrane	0.000942	0.027	AIF1L, CNGA1, CNGB1, CNTNAP2, CUBN, EPB41L3, EPS15, EVC, FERMT1, FGD2, GABRG3, GUCY2D, KCNB1, KCNC1, MACF1, MYO1D, PDE9A, PIEZO1, SHISA6, SNTG1, SRC, SYNE2, TPM1, TTYH1, VASP
GO:0098984	Neuron to neuron synapse	0.002593	0.063702	ANKS1B, ARFGEF2, ATP2B2, CHRNA3, CNN3, CPEB1, DGKI, DISC1, DLG2, DLGAP1, EPB41L3, GRIK4, LRP8, LRRC4C, LYN, NCK2, PKP4, PRKAR1B, PRKCZ, SHISA6, SRC, SYNJ2BP, SYT9, TANC2
GO:0005875	Microtubule-associated complex	0.004053	0.080183	CHURC1-FNTB, DNAH1, DNAH12, DNAH2, DNAH3, DNAH7, DNAH8, DNAH9, KIF15, KIFC1, KIFC3, LRP8, WDR78
GO:0033267	Axon part	0.004196	0.080183	ADCY10, ANK3, APBB2, APP, AUTS2, CALM1, CDH8, CNGB1, CNR1, CNTNAP2, COBL, CRHBP, DGKI, DLG2, EPB41L3, IQCJ-SCHIP1, KCNC1, MBP, MYO1D, NFASC, NPTXR, PRKCZ, PTPRN2, SCN8A, UNC13C
b. By gene ontology molecular function
Gene set	Description	P-Value	FDR	Genes
GO:0050839	Cell adhesion molecule binding	4.81E-06	0.001355	ANK3, CD2AP, CDH13, CDH15, CDH22, CDH4, CDH8, CNN3, COL5A1, CTNNA3, CXCL12, DAB2IP, ECM2, EGFR, EPS15, F11R, FRMD5, GAPVD1, LRRC4C, LYN, MACF1, NCK1, NDRG1, NRXN3, PAK4, PARVA, PFKP, PKP4, PPL, PRKCA, PTPRT, SRC, STAT1, TENM4, TJP2, TMPO, VASP, WASF2
GO:0045503	Dynein light chain binding	2.18E-05	0.003074	DNAH1, DNAH2, DNAH3, DNAH7, DNAH8, DNAH9, WDR78
GO:0003779	Actin binding	6E-05	0.00556	ABLIM1, ABLIM2, AIF1L, CNN3, COBL, CORO2B, COTL1, CTNNA3, DSTN, EGFR, EPB41L3, FERMT1, GAS7, IQGAP2, JMY, MACF1, MYO1D, MYO1F, MYO1G, MYPN, MYRIP, PACRG, PARVA, PHACTR1, SNTB2, SNTG1, SVIL, SYNE2, TPM1, TRIOBP, VASP, WASF2
GO:0045505	Dynein intermediate chain binding	7.89E-05	0.00556	BICD1, DNAH1, DNAH2, DNAH3, DNAH7, DNAH8, DNAH9
GO:0003774	Motor activity	0.000395	0.022266	DNAH1, DNAH12, DNAH2, DNAH3, DNAH7, DNAH8, DNAH9, KIF15, KIFC1, KIFC3, MYO1D, MYO1F, MYO1G, WDR78
GO:0051959	Dynein light intermediate chain binding	0.00052	0.024456	DNAH1, DNAH2, DNAH3, DNAH7, DNAH8, DNAH9
GO:0046873	Metal ion transmembrane transporter activity	0.001459	0.058774	ASIC2, ATP2A2, ATP2B2, ATP2B4, CACNA1B, CACNA1E, CACNA2D4, CNR1, GRIK4, KCNB1, KCNC1, KCNJ3, KCNJ6, KCNK6, KCNN3, RYR2, SCN8A, SLC1A6, SLC1A7, SLC23A2, SLC24A2, SLC24A3, SLC24A4, SLC28A1, SLC39A10, SLC41A2, SLC4A5, SLC8A3, TTYH1
GO:0005516	Calmodulin binding	0.00173	0.060986	ATP2B2, ATP2B4, CNN3, EGFR, IQGAP2, KCNN3, MBP, MYO1D, MYO1F, MYO1G, PLA2G6, RYR2, SLC8A3, SNTB2, SPATA17, UNC13C
GO:0046875	Ephrin receptor binding	0.002586	0.081012	ANKS1B, GRB2, LYN, NCK1, SRC
c. By hallmark
Gene set	Description	P-value	FDR	Genes
HALLMARK_ESTROGEN_RESPONSE_EARLY	Early estrogen response	3.78E-04	0.018889	ABLIM1, ADCY9, CELSR1, CXCL12, FHL2, GAB2, IGF1R, MPPED2, RAB31, SEC14L2, SLC24A3, SLC27A2, SLC2A1, SLC7A5, SVIL, TJP3, TTC39A
d. By the DisGeNET approach
Gene set	Description	P-value	FDR	Genes
C0236969	Substance-related disorders	1.02E-08	3.7E-05	ABLIM1, ADARB2, AGBL4, CADPS2, CDCP1, CDH13, CNR1, CSMD3, CTNNA3, DNAH8, FHIT, FRMD4A, MACROD2, NRXN3, PARVA, PRKCH, RAD51B, SLC2A13, SLC45A2, ZNF366

Burden analysis of VOIs

By WES, 10,199 functional rare coding variants from 6566 autosomal genes were called and annotated using the ANNOVAR software [103]. The gene burden analysis of these variants is presented in Supplementary Table 3. Considering the potential inflation with public database controls by TRAPD [35], the cases were further compared with 2719 unrelated European non-POTS control subjects sequenced by WES at CAG. Significant genes were defined as genome-wide significant by comparing with both public database controls and the internal controls. As a result, 55 genes showed genome-wide significance (P < 2.347E-06, Table 3). Among the 55 genes, 7 genes (ABCA13, CELSR1, DAB2IP, DNAH1, DNAH2, DNAH3, SYNE2) had nominal significance in the gene-based GWAS study [gene enrichment: 7/716 (gene-based GWAS) versus 55/21,306 (human coding genes), OR 3.81, P = 3.52E-04], which suggests the association signals of these genes in GWAS may be explain by rare coding variants. ORA analysis of the 55 genes highlighted the roles of the genes in muscular function, emphasizing muscular dysfunction in the pathogenesis of POTS (Table 4).

Table 3.

The 55 genes showed genome-wide significance by burden analysis of rare coding variants

#GENE	CASE_COUNT_HET	CASE_COUNT_CH	CASE_COUNT_HOM	Control_COUNT_HET	Control_COUNT_CH	Control_COUNT_HOM	P_DOM
ABCA13	10	2	0	26	7	0	6.40E-08
CELSR1	5	0	0	2	0	0	5.11E-07
DAB2IP	8	0	0	8	0	0	6.50E-09
DNAH1	9	0	0	20	0	0	1.05E-07
DNAH2	7	0	0	13	0	0	1.21E-06
DNAH3	9	1	0	25	4	1	6.37E-07
SYNE2	7	0	0	14	0	0	1.77E-06
ABCA7	7	1	0	8	6	0	1.14E-07
AGRN	6	0	0	5	0	0	3.05E-07
AHNAK	8	2	0	21	0	0	1.56E-06
AP5Z1	5	0	0	3	0	0	1.33E-06
ARHGAP22	4	0	0	0	0	0	8.64E-07
ARID1B	5	0	0	3	0	0	1.33E-06
BAHCC1	9	0	0	5	0	0	3.08E-11
CACNA1A	5	1	0	3	1	0	1.33E-06
CACNA1D	6	0	0	8	0	0	1.84E-06
CFAP46	6	0	0	6	0	0	5.95E-07
CILP	4	0	0	0	0	0	8.64E-07
CMYA5	4	0	0	0	0	0	8.64E-07
COL12A1	5	0	0	2	0	0	5.11E-07
COL27A1	6	0	0	5	0	0	3.05E-07
COL7A1	8	0	0	9	0	0	1.20E-08
CSMD1	6	0	0	6	0	0	5.95E-07
DNAH10	7	0	0	5	0	0	1.51E-08
EPHB4	5	0	0	1	0	0	1.50E-07
FAT1	6	0	0	8	0	0	1.84E-06
FBXW5	6	2	0	2	1	0	1.99E-08
FIGNL1	6	0	0	7	0	0	1.08E-06
FLG	7	0	0	5	0	0	1.51E-08
HSPG2	11	1	0	25	0	0	4.29E-09
KMT2C	8	0	0	8	0	0	6.50E-09
KNTC1	5	0	0	2	0	0	5.11E-07
LAMA5	13	1	0	21	0	0	5.38E-12
LRP2	11	0	0	11	0	0	7.17E-12
MUC16	14	2	0	25	0	0	2.11E-12
MYH7B	5	0	0	2	0	0	5.11E-07
NEB	12	3	0	21	1	0	7.64E-11
NRAP	5	0	0	3	0	0	1.33E-06
NUP160	4	0	0	0	0	0	8.64E-07
OBSCN	11	1	0	14	0	0	4.20E-11
PABPC1L	4	0	0	0	0	0	8.64E-07
PKD1	7	1	0	4	0	0	6.46E-09
PKD1L2	8	1	0	19	10	2	1.56E-06
PKHD1L1	7	0	0	9	0	0	1.97E-07
PLEC	11	1	0	21	0	0	1.02E-09
PLXNA2	4	0	0	0	0	0	8.64E-07
RP1L1	6	0	0	8	1	0	1.84E-06
RYR1	8	0	0	5	0	0	7.02E-10
SACS	8	0	0	4	0	0	2.77E-10
SRRM2	5	0	0	2	0	0	5.11E-07
TG	6	0	0	5	0	0	3.05E-07
TTN	24	5	0	51	1	0	1.72E-19
USH2A	9	0	0	9	0	0	6.76E-10
XIRP2	6	0	0	7	0	0	1.08E-06
ZFHX3	6	0	0	6	0	0	5.95E-07

HET heterozygote, CH compound heterozygote, HOM homozygote, P_dom P-value of dominant model, P_rec P-value of recessive model

Table 4.

Over-representation analysis of the 55 genes burdened with VOIs

a. By the DisGeNET approach
Gene set	Description	P-value	FDR	Genes
C1864711	Muscle biopsy shows dystrophic changes	2.41E-05	0.045614	PLEC, RYR1, SYNE2, TTN
C0026850	Muscular dystrophy	3.92E-05	0.045614	PLEC, RYR1, SYNE2, TTN
C0221629	Proximal muscle weakness	5.55E-05	0.045614	NEB, RYR1, SYNE2, TTN
C1838869	Proximal neurogenic muscle weakness	5.55E-05	0.045614	NEB, RYR1, SYNE2, TTN
C0746674	Generalized muscle weakness	0.000109	0.045614	NEB, PLEC, RYR1, TTN
C0151576	Elevated creatine kinase	0.000113	0.045614	HSPG2, PLEC, RYR1, SYNE2, TTN
C0241005	Creatine phosphokinase serum increased	0.000113	0.045614	HSPG2, PLEC, RYR1, SYNE2, TTN
C0376175	Bell palsy	0.000117	0.045614	COL12A1, NEB, RYR1, TTN
C1858719	Facial muscle weakness of muscles innervated by CN VII	0.000117	0.045614	COL12A1, NEB, RYR1, TTN
C0427055	Facial paresis	0.000125	0.045614	COL12A1, NEB, RYR1, TTN
b. By gene ontology cellular component
Gene set	Description	P-value	FDR	Genes
GO:0043292	Contractile fiber	3.00E-10	5.16E-08	AHNAK, CACNA1D, CMYA5, MYH7B, NEB, NRAP, OBSCN, PLEC, RYR1, SYNE2, TTN, XIRP2
GO:0042383	Sarcolemma	3.78E-04	0.018834	AHNAK, CACNA1D, OBSCN, PLEC, RYR1
GO:0031012	Extracellular matrix	4.02E-04	0.018834	AGRN, CILP, COL12A1, COL27A1, COL7A1, FLG, HSPG2, LAMA5, USH2A
GO:0016528	Sarcoplasm	4.38E-04	0.018834	CMYA5, PLEC, RYR1, SYNE2
GO:0030055	Cell–substrate junction	5.55E-04	0.019084	AHNAK, ARHGAP22, FAT1, HSPG2, NRAP, PLEC, SYNE2, XIRP2
c. By gene ontology molecular function
Gene set	Description	P-value	FDR	Genes
GO:0005201	Extracellular matrix structural constituent	9.23E-07	0.000197	AGRN, CILP, COL12A1, COL27A1, COL7A1, HSPG2, LAMA5
GO:0045503	Dynein light chain binding	1.40E-06	0.000197	DNAH1, DNAH10, DNAH2, DNAH3
GO:0051959	Dynein light intermediate chain binding	2.60E-06	0.000211	DNAH1, DNAH10, DNAH2, DNAH3
GO:0045505	dynein intermediate chain binding	3.00E-06	0.000211	DNAH1, DNAH10, DNAH2, DNAH3
GO:0042805	Actinin binding	8.80E-06	0.000496	CACNA1D, NRAP, TTN, XIRP2
GO:0008307	Structural constituent of muscle	1.43E-05	0.000674	NEB, OBSCN, PLEC, TTN
GO:0030506	Ankyrin binding	4.03E-05	0.001624	CACNA1D, OBSCN, PLEC
GO:0003774	Motor activity	9E-05	0.003173	DNAH1, DNAH10, DNAH2, DNAH3, MYH7B
GO:0003779	Actin binding	0.000481	0.01508	MYH7B, NEB, NRAP, PLEC, SYNE2, TTN, XIRP2
GO:0016887	ATPase activity	0.000627	0.017687	ABCA13, ABCA7, DNAH1, DNAH10, DNAH2, DNAH3, FIGNL1
d. By hallmark
Gene set	Description	P-value	FDR	Genes
None
e. By human phenotype ontology
Gene set	Description	P-value	FDR	Genes
HP:0003306	Spinal rigidity	4.94E-07	0.002311	AGRN, COL12A1, HSPG2, NEB, SYNE2, TTN
HP:0003458	EMG: myopathic abnormalities	3E-05	0.034381	AGRN, COL12A1, NEB, RYR1, SYNE2, TTN
HP:0003457	EMG abnormality	3.84E-05	0.034381	AGRN, CACNA1D, COL12A1, HSPG2, NEB, RYR1, SYNE2, TTN
HP:0003701	Proximal muscle weakness	0.000042	0.034381	AGRN, COL12A1, NEB, PLEC, RYR1, SYNE2, TTN
HP:0100285	EMG: impaired neuromuscular transmission	4.27E-05	0.034381	AGRN, CACNA1D, RYR1, TTN
HP:0003324	Generalized muscle weakness	4.41E-05	0.034381	AGRN, COL12A1, NEB, PLEC, RYR1, TTN
HP:0003236	Elevated serum creatine phosphokinase	6.24E-05	0.03722	AP5Z1, COL12A1, HSPG2, NEB, PLEC, RYR1, SYNE2, TTN
HP:0003690	Limb muscle weakness	6.48E-05	0.03722	AGRN, AP5Z1, NEB, RYR1, SACS, SYNE2, TTN
HP:0040081	Abnormal levels of creatine kinase in blood	7.72E-05	0.03722	AP5Z1, COL12A1, HSPG2, NEB, PLEC, RYR1, SYNE2, TTN
HP:0011021	Abnormality of circulating enzyme level	7.96E-05	0.03722	AP5Z1, COL12A1, HSPG2, NEB, PLEC, RYR1, SYNE2, TTN

Pathogenic/Likely Pathogenic (P/LP) variants

Our study identified 107 deleterious P/LP variants, of which 99 deleterious variants were supported by a minimum of two databases, including ClinVar (clinvar_20231230), InterVar, or HGMD_Pro_2023.3. When we concentrated on variants classified as P/LP variants by ClinVar classification, 92 P/LP variants were highlighted, including 3 variants that were previously reported of dominant genetic effects (Supplementary Table 4). Among the 87 WES patients, 53 (60.9%) have at least one P/LP variant (Supplementary Fig. 1). The 92 P/LP variants are from 87 genes, and 5 genes, i.e., GAA, GALT, GYS2, PAH, and USH2A, each have two P/LP variants from two different individuals. The gene with P/LP variant, OTOG, has also been identified of nominal significance in the gene-based GWAS study. ORA analysis of the 87 genes highlighted a number of gene sets with statistical significance (FDR < 0.1), suggesting the possibility of several underexplored gene pathways and networks in the pathogenesis of POTS (Table 5).

Table 5.

Over-representation analysis of the 87 genes with P/LP variants

a. By the DisGeNET approach
Gene set	Description	P-value	FDR	Genes
C4020899	Autosomal recessive predisposition	< 2.2e-16	< 2.2e-16	ABCA4, ABCC6, ABCC8, ADSL, ALDOB, APRT, ASL, ASS1, ATM, BLM, BTD, C6, CAPN3, CBLIF, CFTR, COQ4, CTSA, DARS2, DBT, DPM1, FANCI, FASTKD2, FMO3, GAA, GALT, GBA, GBE1, GGCX, GYS2, ITGA2B, LAMB3, LIPT1, MC2R, MUTYH, PAH, PCCB, PDE6B, PEPD, PEX5, PLOD1, PMS2, POLG, POLR3A, POMT1, PRF1, PROM1, RAD50, RPE65, RYR1, SLC12A3, SLC17A5, SLC3A1, TG, TNFRSF13B, TRMU, TSFM, TYR, USH2A
C0019209	Hepatomegaly	4.70E-12	8.56E-09	ABCC8, ALDOB, ASL, ASS1, BTD, DPM1, FASTKD2, GAA, GALT, GBA, PCCB, PEPD, PEX5, POLG, PRF1, SLC17A5, TNFRSF13B, TRMU, TSFM
C0014544	Epilepsy	1.14E-11	1.38E-08	ABCC8, ADSL, ALDOB, ASL, ASS1, ATM, BTD, CTSA, DBT, DPM1, FASTKD2, GBA, GYS2, MC2R, PAH, PCCB, PEX5, PMS2, POLG, POMT1, PRF1, RPE65, SLC12A3, SLC17A5, SLC3A1, TSFM
C0007758	Cerebellar ataxia	2.25E-11	2.05E-08	ABCC8, ASL, ASS1, ATM, BTD, DARS2, DBT, DPM1, FASTKD2, GBA, HEXB, PEX5, POLG, POLR3A, PRF1, RAD50, SLC17A5, TSFM
C0036572	Seizures	5.23E-11	3.81E-08	ABCC8, ADSL, ALDOB, ASL, ASS1, ATM, BTD, CTSA, DBT, DPM1, FASTKD2, GBA, GYS2, MC2R, PAH, PCCB, PEX5, PMS2, POLG, POMT1, PRF1, RPE65, SLC12A3, SLC17A5, SLC3A1, TSFM
C0231246	Failure to gain weight	1.93E-09	1E-06	ABCC8, ALDOB, ASL, ASS1, CFTR, DPM1, FASTKD2, GALT, GBA, GBE1, LAMB3, MC2R, PCCB, PEX5, POLG, PRF1, RYR1, SLC17A5, SLC3A1
C2315100	Pediatric failure to thrive	1.93E-09	1E-06	ABCC8, ALDOB, ASL, ASS1, CFTR, DPM1, FASTKD2, GALT, GBA, GBE1, LAMB3, MC2R, PCCB, PEX5, POLG, PRF1, RYR1, SLC17A5, SLC3A1
C0009421	Comatose	3.45E-09	1.57E-06	ABCC8, ALDOB, ASL, ASS1, DBT, MC2R, PCCB, POLG, PRF1
C0042963	Vomiting	1.34E-08	5.42E-06	ABCC8, ALDOB, ASL, ASS1, BTD, DBT, GALT, HSD3B2, PCCB, POLG, TRMU
C0026827	Muscle hypotonia	2.36E-08	8.59E-06	ADSL, AR, BTD, COQ4, DBT, DPM1, FASTKD2, GAA, GBA, GBE1, PEX5, PLOD1, PMS2, POLG, POMT1, PRF1, RPE65, RYR1, SLC17A5, SLC3A1, TG, TRMU
b. By gene ontology cellular component
Gene set	Description	P-value	FDR	Genes
GO:0005759	Mitochondrial matrix	3.56E-04	0.061	BTD, DARS2, DBT, FASTKD2, LIPT1, MCCC2, PCCB, POLG, TARS2, TSFM
GO:0045177	Apical part of cell	1.60E-03	0.099	ABCC6, CBLIF, CFTR, OTOG, PROM1, SLC12A3, SLC34A3, USH2A
GO:0009295	Nucleoid	2.09E-03	0.099	DBT, FASTKD2, POLG
GO:0005774	Vacuolar membrane	2.30E-03	0.099	ABCC6, CFTR, CTSA, GAA, GBA, HLA-DRB1, SLC17A5, SLC3A1
c. By gene ontology molecular function
Gene set	Description	P-value	FDR	Genes
GO:0016798	Hydrolase activity, acting on glycosyl bonds	8.07E-06	0.002	CTSA, GAA, GBA, GBE1, HEXB, MUTYH, OTOG
GO:0016757	Transferase activity, transferring glycosyl groups	5.45E-05	0.008	ALG1, APRT, DPM1, FUT1, GBE1, GYS2, HEXB, PLOD1, POMT1
GO:0016705	Oxidoreductase activity, acting on paired donors, with incorporation or reduction of molecular oxygen	4.37E-04	0.041	CYP4F22, FMO3, P3H1, PAH, PLOD1, TYR
d. By hallmark
Gene set	Description	P-value	FDR	Genes
None
e. By human phenotype ontology
Gene set	Description	P-value	FDR	Genes
HP:0001939	Abnormality of metabolism/homeostasis	3.60E-08	1.68E-04	ABCA4, ABCC6, ABCC8, ALDOB, ALG1, APRT, AR, ASL, ASS1, ATM, BLM, BTD, CAPN3, CBLIF, CFTR, COQ4, CTSA, CYP4F22, DBT, DCTN1, DHDDS, DPM1, FANCI, FASTKD2, FMO3, GAA, GALT, GBA, GBE1, GYS2, HEXB, HLA-DRB1, HSD3B2, IL17RC, LAMB3, LIPT1, MC2R, MCCC2, MUTYH, PAH, PCCB, PDE6B, PEPD, PEX5, PLOD1, PMS2, POLG, POMT1, PRF1, PROM1, RAD50, RPE65, RYR1, SLC12A3, SLC17A5, SLC34A3, SLC3A1, TARS2, TNFRSF13B, TRMU, TSFM, USH2A
HP:0004360	Abnormality of acid–base homeostasis	1.64E-07	2.87E-04	ABCC8, ALDOB, ASL, ASS1, BTD, COQ4, DBT, FASTKD2, GALT, GYS2, HSD3B2, LIPT1, MCCC2, PAH, PCCB, POLG, RYR1, SLC12A3, SLC3A1, TARS2, TRMU, TSFM
HP:0001438	Abnormality of abdomen morphology	1.84E-07	2.87E-04	ABCC8, ALDOB, ALG1, ASL, ASS1, ATM, BTD, CFTR, CTSA, DHDDS, DPM1, FASTKD2, FMO3, GAA, GALT, GBA, GBE1, HEXB, HLA-DRB1, PCCB, PEPD, PEX5, PMS2, POLG, PRF1, SLC17A5, SLC34A3, TNFRSF13B, TRMU, TSFM
HP:0001941	Acidosis	2.55E-07	2.94E-04	ABCC8, ALDOB, ASL, ASS1, BTD, COQ4, DBT, FASTKD2, GALT, GYS2, HSD3B2, LIPT1, MCCC2, PAH, PCCB, POLG, RYR1, SLC3A1, TARS2, TRMU, TSFM
HP:0003271	Visceromegaly	3.14E-07	2.94E-04	ABCC8, ALDOB, ALG1, ASL, ASS1, ATM, BTD, CFTR, CTSA, DHDDS, DPM1, FASTKD2, FMO3, GAA, GALT, GBA, GBE1, HEXB, HLA-DRB1, PCCB, PEPD, PEX5, POLG, PRF1, SLC17A5, TNFRSF13B, TRMU, TSFM
HP:0410042	Abnormal liver morphology	1.64E-06	1.28E-03	ABCC8, ALDOB, ALG1, ASL, ASS1, ATM, BTD, CFTR, CTSA, DHDDS, DPM1, FASTKD2, GAA, GALT, GBA, GBE1, GYS2, HEXB, HLA-DRB1, IL17RC, LIPT1, PCCB, PEPD, PEX5, POLG, PRF1, SLC17A5, TARS2, TNFRSF13B, TRMU, TSFM
HP:0001259	Coma	2.25E-06	1.46E-03	ABCC8, ALDOB, ASL, ASS1, BTD, DBT, MC2R, MCCC2, PCCB, POLG, PRF1
HP:0002240	Hepatomegaly	2.61E-06	1.46E-03	ABCC8, ALDOB, ALG1, ASL, ASS1, BTD, CFTR, DHDDS, DPM1, FASTKD2, GAA, GALT, GBA, HLA-DRB1, PCCB, PEPD, PEX5, POLG, PRF1, SLC17A5, TNFRSF13B, TRMU, TSFM
HP:0002012	Abnormality of the abdominal organs	2.80E-06	1.46E-03	ABCC8, ALDOB, ALG1, ASL, ASS1, ATM, BTD, CFTR, CTSA, DBT, DHDDS, DPM1, FANCI, FASTKD2, FMO3, GAA, GALT, GBA, GBE1, GYS2, HEXB, HLA-DRB1, IL17RC, LIPT1, MMP21, PCCB, PEPD, PEX5, PMS2, POLG, PRF1, RAD50, SLC17A5, TARS2, TG, TNFRSF13B, TRMU, TSFM
HP:0001392	Abnormality of the liver	5.14E-06	2.37E-03	ABCC8, ALDOB, ALG1, ASL, ASS1, ATM, BTD, CFTR, CTSA, DHDDS, DPM1, FANCI, FASTKD2, GAA, GALT, GBA, GBE1, GYS2, HEXB, HLA-DRB1, IL17RC, LIPT1, PCCB, PEPD, PEX5, PMS2, POLG, PRF1, SLC17A5, TARS2, TG, TNFRSF13B, TRMU, TSFM

Discussion

Common genetic variants and POTS susceptibility

This study presents a unique approach to a systemic evaluation of the etiology and molecular mechanisms of POTS. The application of GWAS to POTS has encountered challenges, primarily due to the disorder’s extensive phenotypic heterogeneity. This heterogeneity poses a significant obstacle for GWAS, which typically depends on a well-defined, uniform phenotype to effectively identify common genetic variants linked to a specific condition. A major challenge in GWAS for POTS is accurately characterizing its diverse phenotypes. The clinical complexity of POTS makes it difficult to distinguish between potential subtypes and to define a consistent phenotype that truly represents the disorder. Given the substantial phenotypic diversity of POTS, the GWAS approach was unable to identify any loci of genome-wide significance. Nevertheless, genes that showed nominal significance in gene-based association tests exhibited a highly significant enrichment in several gene sets important to POTS physiobiology. This finding underscores the role of common genetic variants in influencing POTS susceptibility and provides insights into its pathophysiology (Table 2).

GO cellular component cell–cell junction (GO:0005911) and synaptic membrane (GO:0097060)

These gene sets are integral to neuronal communication, which is crucial for the proper functioning of the ANS. Genes associated with cell–cell junctions play a role in maintaining the structural and functional integrity of synapses [1], the points of communication between neurons. Synaptic membrane genes are involved in neurotransmitter release and reuptake [74], which are critical for signal transmission in the ANS. Common genetic variations in genes associated with these processes can influence autonomic responses, a hallmark of POTS.

Neuronal cell body (GO:0043025) and axon part (GO:0033267): Genes associated with the neuronal cell body and axon are crucial for the health and function of neurons. Axonal genes play a role in the transmission of electrical signals along the nerve fiber [39]. Changed function in these cellular components by genetic variants can lead to impaired transmission of autonomic signals, contributing to the risk of orthostatic intolerance and tachycardia in POTS. Additionally, there is increasing evidence to suggest that a significant number of patients with POTS experience small fiber neuropathy (SFN), an autoimmune disorder that specifically targets and damages the small fibers responsible for conducting autonomic and pain signals [33, 36]. This further underscores the importance of understanding the genetic and cellular mechanisms underlying neuronal function and integrity.

Transporter complex (GO:1,990,351): This gene set is involved in the transport of various molecules across cellular membranes, including neurotransmitters [5]. In the context of POTS, the regulation of neurotransmitters such as norepinephrine is particularly relevant. Dysregulation in neurotransmitter transport can lead to imbalances in sympathetic nervous system activity, a critical aspect of POTS pathophysiology [85].

GO molecular function cell adhesion molecule binding (GO:0050839)

Genes involved in cell adhesion molecule binding play a crucial role in the interaction and adhesion of cells to their surrounding extracellular matrix and to other cells [10]. This is particularly important in the cardiovascular system, where endothelial cell integrity is essential for maintaining vascular function. In POTS, the dysregulation of this function could lead to compromised blood vessel reactivity and integrity, influencing blood flow dynamics.

Actin binding (GO:0003779): Actin is a fundamental component of the cellular cytoskeleton and is critical in various cellular processes, including maintenance of cell shape, cell movement, and muscle contraction [95]. Actin-binding genes are essential for the proper functioning of muscle cells, including cardiac [46] and smooth muscle cells that line blood vessels [50]. In POTS, abnormalities in actin binding could impact cardiac muscle function and vascular tone regulation, both of which are vital for maintaining stable blood pressure and heart rate.

Motor activity (GO:0003774): This gene set is associated with the generation of force and movement within cells, a function that is crucial in muscle cells, including the heart [73]. In the context of POTS, motor activity genes could influence how heart and vascular muscles respond to autonomic signals, especially in adjusting heart rate and vascular tone in response to orthostatic stress.

Early estrogen response (HALLMARK_ESTROGEN_RESPONSE_EARLY)

POTS is observed to be more common in women, with a ratio of as much as five female cases to one male case [59]. However, the link with sex is not well understood. There is a recognized association between female hormones, notably estrogen, and changes in blood volume and vascular function [38]. This gene set comprises genes that are responsive to estrogen in the early phase of its action [65]. These early estrogen response genes could potentially play a role in POTS, given the higher prevalence of the condition in women. The potential effects include: (1) autonomic regulation and cardiovascular effects—estrogen is known to influence autonomic regulation and cardiovascular function [55], which are both key aspects in the pathophysiology of POTS; (2) extended thoracic hypovolemia—estrogen can affect fluid retention and blood vessel constriction [15, 26], potentially influencing the degree of hypovolemia and the strain on the autonomic nervous system; (3) autoimmune responses—estrogen can modulate immune responses [62], which might intersect with autoimmune processes targeting the autonomic nervous system in POTS; furthermore, female patients have a higher prevalence of autoimmune disorders compared with male patients [70]; (4) inflammatory mechanisms—estrogen has both proinflammatory and antiinflammatory effects, depending on the context [61]; the early estrogen response genes might play a role in the inflammatory underpinnings of POTS; (5) autonomic neuropathies and sympathetic denervation—estrogen influences nerve function and repair [60], and its early response genes could be involved in the development or compensation of autonomic neuropathies in POTS; and (6) impaired norepinephrine reuptake—estrogen can modulate the expression and function of neurotransmitter transporters, possibly impacting norepinephrine reuptake mechanisms [97]. Clinically, we observed a case series of three transgender female patients transitioning to male sex whose POTS symptoms significantly improved after the addition of exogenous testosterone [9]. Additionally, both published [70] and our unpublished data have observed that female patients with POTS experience a worsening of symptoms around their menstrual periods.

Substance-related disorders (DisGeNET C0236969): this correlation carries two implications

Firstly, POTS may share a common genetic susceptibility with substance-related disorders; secondly, this gene set might be linked to POTS due to the role of certain substances in modulating the autonomic nervous system and cardiovascular responses. Dysautonomia can be exacerbated or triggered by substance exposure. The underlying mechanisms may include: (1) the autonomic nervous system may be influenced by various medications commonly utilized in clinical practice [23]. For example, β-adrenergic receptors are activated by some bronchodilators for asthma management. Amphetamines, such as those prescribed for attention deficit hyperactivity disorder, or consuming caffeine, can lead to an increase in the release of the sympathetic neurotransmitter norepinephrine. Tricyclic antidepressants can inhibit the reuptake of norepinephrine, thus increasing its availability in the synaptic cleft [21]. (2) Common substances can exert direct or indirect effects on the cardiovascular system, such as caffeine, alcohol, nicotine, and antidepressants [32]. Calcium channel blockers may cause peripheral vasodilation and reduce venous return [88], thus exacerbating the hypovolemic state often seen in POTS. β-blockers may influence myocardial contractility or heart rate, contributing to the dysregulation of cardiovascular function. (3) Substances can also alter the body’s response to stress, a factor that is often implicated in the exacerbation of POTS symptoms [42]. The dysregulation of stress hormones and the sympathetic nervous system can lead to increased heart rate and blood pressure variability.

These gene sets offer a window into the complex interplay of common genetic variants and their potential role in predisposing individuals to POTS. The exploration of GWAS gene sets in the context of POTS not only enhances our understanding of the genetic basis of the syndrome, but also opens new pathways for personalized and preventive healthcare strategies.

Rare functional variants and POTS heterogeneity

Compared with the results of our GWAS study, our WES study emphasizes the importance of rare coding variants in the pathogenesis of POTS. Two complementary analyses were employed in this study: the burden analysis of rare variants and the identification of P/LP variants. The burden analysis entails assessing the cumulative impact of rare functional variants in the individuals with POTS compared with the control group. The primary focus is to determine whether there is a higher prevalence of functional rare variants in patients with POTS, as opposed to common variants identified in the association study. This analysis does not necessarily prioritize the predicted pathogenicity of each variant. Instead, it focuses on evaluating the overall burden of these functional rare variants in the genome, providing a comprehensive overview of the genetic landscape. Conversely, the analysis of P/LP rare variants involves identifying deleterious variants, particularly those classified by ClinVar. This can help establish a direct link between specific genetic changes and POTS, leading to a better understanding of the molecular mechanisms of the disease and potentially guiding targeted treatments.

Insights gained by burden analysis of VOIs

Using 2719 unrelated European control subjects, this study identified 55 genes associated with POTS with genome-wide significance by burden analysis of rare coding variants. The 55 genes identified in this study highlight both known and also unveil novel knowledge of POTS heterogeneity.

Muscular dysfunction in POTS

The ORA analysis in this study emphasized the importance of possible muscular dysfunction in POTS, with genes involved in muscle function and muscular diseases enriched with highly statistical significance (Table 4a,e). Altogether, 32 out of the 55 genes are related to muscular dysfunction. The affected muscular function may not be limited to myocardium and vascular smooth muscle. For instance, the calf muscle pump generates pressure gradient between the thigh and the lower leg veins, and is the major force for return of venous blood from the lower extremities to the heart [81]. Decreased calf muscle pump activity (HP:0003690 limb muscle weakness) may thus contribute to the venous pooling in lower extremities in some patients with POTS [93]. It is worth noting that no muscle dysfunction has been observed in these patients with POTS, suggesting that any potential involvement of muscular mechanisms may be subclinical in terms of skeletal muscle dysfunction.

Muscular function relies on coordinated activity between muscle fibers and the metabolic and regulatory machineries [66]. The structural components of muscle cells that may be affected by rare coding variants include (Table 4b): (1) contractile fiber (GO:0043292), sarcolemma (GO:0042383), and sarcoplasm (GO:0016528). The genes with rare coding variants include AHNAK nucleoprotein (AHNAK); calcium voltage-gated channel subunit alpha1 D (CACNA1D); cardiomyopathy associated 5 (CMYA5); myosin heavy chain 7B (MYH7B); nebulin (NEB); nebulin related anchoring protein (NRAP); obscurin, cytoskeletal calmodulin and titin-interacting RhoGEF (OBSCN); plectin (PLEC); ryanodine receptor 1 (RYR1); spectrin repeat containing nuclear envelope protein 2 (SYNE2); titin (TTN); and xin actin binding repeat containing 2 (XIRP2). (2) Extracellular matrix (GO:0031012). The genes with rare coding variants are agrin (AGRN), cartilage intermediate layer protein (CILP), collagen type XII alpha 1 chain (COL12A1), collagen type XXVII alpha 1 chain (COL27A1), collagen type VII alpha 1 chain (COL7A1), filaggrin (FLG), heparan sulfate proteoglycan 2 (HSPG2), laminin subunit alpha 5 (LAMA5), and usherin (USH2A). (3) Cell-substrate junction (GO:0030055). The related genes with rare coding variants are AHNAK, Rho GTPase activating protein 22 (ARHGAP22), FAT atypical cadherin 1 (FAT1), heparan sulfate proteoglycan 2 (HSPG2), NRAP, PLEC, SYNE2, and XIRP2. The molecular functions of these genes are related to the dynein motor to generate force, cytoskeletal actinin /ankyrin/actin binding, and ATPase activity for providing energy (Table 4c).

Microtubule dysfunction in POTS

Among the 32 genes that are related to muscular dysfunction, four dynein axonemal heavy chain (DNAH) genes (DNAH1, DNAH2, DNAH3, DNAH10, and the SYNE2 gene) involve microtubule function. Axonemal dynein produces force to move other proteins and cell materials by microtubules within cilia [101]. Dysfunction in endothelial cilia contributes to aberrant fluid-sensing and results in vascular disorders, including hypertension [69]. In addition, an intact microtubule network is necessary for proper subcellular structure and function [107]. Aberrant growth of cardiomyocyte microtubules contribute to contractile dysfunction [98]. Targeting at microtubules may improve cardiomyocyte function in human heart failure [12]. SYNE2 encodes nuclear envelope spectrin-repeat protein (Nesprin)-2, functioning as intracellular scaffolds and linkers to establish nuclear-cytoskeletal connections by binding cytoplasmic F-actin, in addition to its role as a microtubule scaffold [80]. Mutations of SYNE2 may lead to structural and adaptive signaling defects in mechanically stressed tissues such as muscle, and cause Emery–Dreifuss muscular dystrophy (EDMD5) [110]. Besides the above genes, two additional genes, kinetochore associated 1 (KNTC1) and RP1 like 1 (RP1L1) also encode proteins of the microtubule complex (GO:0005874). Notably, there has been no observed contractile dysfunction in these patients with POTS, implying that any potential engagement of microtubule mechanisms may manifest subclinically concerning contractile function.

Genes reported of association with blood pressure

According to the GWAS catalog, 15 of the 55 genes have been reported of association with blood pressure regulation (https://www.ebi.ac.uk, accessed on 5 September 2021), including 7 genes related to muscular function (ARHGAP22, CACNA1D, DNAH2, DNAH3, PLEC, SACS, TTN) with 8 other genes contributing (ARID1B, BAHCC1, CSMD1, LRP2, NUP160, PKD1, RP1L1, ZFHX3). The genes involved in muscular function may be related to POTS by their roles involving myocardium or vascular smooth muscle function. For example, the two DNAH genes (DNAH2[40, 45] and DNAH3 [34]) are associated with systolic blood pressure, while DNAH3 is also reported of association with diastolic blood pressure [34]. The association of DNAH2 and DNAH3 with blood pressure may be related to their roles in cardiomyocyte function [98] (for systolic blood pressure) and microtubule function in vascular smooth muscle contraction [109]. However, clinically, no contractile dysfunction has been demonstrated in POTS thus far, suggesting the need for further investigation into the underlying mechanisms.

The LDL receptor related protein 2 gene (LRP2) encodes the endocytic receptor megalin, which has regulatory effects on the renin-angiotensin system activity in the kidney [94], in addition to its key roles in renal proximal tubular function [17].

Genes reported of association with heart rate

Among the 55 genes, 10 genes have been reported of association with heart rate, including 4 genes related to muscular function (CACNA1D, COL12A1, PLEC, TTN) and 6 other contributing genes (CELSR1, CSMD1, DAB2IP, EPHB4, RP1L1, ZFHX3) (https://www.ebi.ac.uk, accessed on 5 September 2021). Among the ten genes, the genes CACNA1D, CSMD1, PLEC, RP1L1, TTN, and ZFHX3 are also associated with blood pressure.

DAB2IP associated with heart rate [41] encodes a Ras GTPase-activating protein. In addition to its role as a tumor suppressor [106], DAB2IP protein functions as a scaffold protein and modulates different signal cascades associated with cell proliferation, survival, and apoptosis [108]. Through the DAB2IP-ASK1-JNK signaling pathway, DAB2IP plays important roles in the function and apoptosis of vascular endothelial cells [111].

CELSR1 encodes a member of the flamingo subfamily of the cadherin superfamily [19], with important roles in neuronal morphogenesis [29]. Mutations of this gene have been reported with correlation with neural tube defects [51]. CELSR1 was reported with association with heart rate in patients with heart failure by a previous GWAS [24]. Concerning the potential roles of CELSR1 in regulating heart rate and in POTS, vestibular hair cells of the inner ear convert mechanical stimuli into neural activity, thus to control balance, blood pressure, and heart rate [99]. CELSR1 coordinates the planar polarity organization of vestibular hair cells in inner ear development [22]. Meanwhile, we have observed patients who still have vestibular dysfunction clinically, even without a history of head trauma or concussion.

Cardiac insufficiency resulting from genetic mutations

In addition to the knowledge gained from the gene set enrichment analysis, 12 of the 32 genes related to muscular dysfunction (NEB, PLEC, XIRP2, TTN, CACNA1D, CMYA5, FAT1, HSPG2, MYH7B, NRAP, OBSCN, SYNE2) have also been reported of association with cardiomyopathy according to the HGMD professional dataset [90] 2021.1 release. A total of 4 of the 12 genes (NEB, PLEC, XIRP2, TTN) and 9 other genes (ARID1B, CACNA1A, CELSR1, KMT2C, LRP2, AHNAK, COL7A1, LAMA5, RYR1) are also related to congenital heart disease. For instance, the TTN gene encodes the giant muscle filament titin of striated muscle. TTN is associated with familial hypertrophic cardiomyopathy [83] and familial dilated cardiomyopathy [31], as well as a specific form of cardiomyopathy characterized by arrhythmia, i.e., arrhythmogenic right ventricular cardiomyopathy (ARVC) [96]. MYH7B encodes the major contractile protein in heart and vascular smooth muscle and is directly involved in muscle contraction [18]. These findings highlight a subset of patients with POTS with rare coding variants from genes related to inherited cardiomyopathy, congenital heart defects, or congenital channelopathy (e.g., RYR1 [4], CACNA1D [3]). The POTS symptoms in these patients may be attributed to cardiac insufficiency resulting from genetic mutations, without necessarily involving subclinical or inconspicuous structural or functional changes.

Psychiatric and neurodevelopmental disorders in POTS

It is not uncommon for patients with POTS to experience psychological issues such as depression and anxiety [79]. There is a potential bidirectional relationship between POTS and psychological distress, whereas the exact role of psychiatric and psychological factors in the development of POTS remains a topic of ongoing research. From the 55 genes we identified, 4 have been linked to anxiety disorder, 8 to schizophrenia, and 4 to autism spectrum disorder (ASD) (the GWAS catalog https://www.ebi.ac.uk, accessed on 5 September 2021). As per HGMD, 14 genes are linked to schizophrenia, and notably, 43 out of the 55 genes are related to ASD (Supplementary Table 5). Autonomic dysfunction is common in ASD [13]. The findings of our study imply that individuals diagnosed with POTS may also have concurrent atypical psychiatric or neurodevelopmental disorders. Owens et al. have documented a correlation between dysautonomia and ASD [68]. Moreover, in clinical settings, we have observed a number of patients with POTS with ASD.

Insights gained by ClinVar P/LP variants

In our WES study, we identified 92 heterozygous P/LP variants in 87 different genes classified by ClinVar. Many of these genes are associated with autosomal recessive predisposition; therefore, patients do not typically manifest obvious genetic syndromes when these variants are present in a heterozygous state. Among these genes, the otogelin gene (OTOG) has also been identified in the gene-based GWAS study on common genetic variants. OTOG encodes a protein that is primarily associated with the acellular membranes of the inner ear and plays a crucial role in auditory and vestibular functions [87]. The LP variant NP_001278992.1:p.Gly2238Ser causes a rare genetic deafness with autosomal recessive inheritance (https://www.ncbi.nlm.nih.gov/clinvar/variation/930161/). While OTOG is primarily associated with the inner ear, there is some evidence to suggest that ANS dysfunction can be linked to inner ear disorders [104]. Disruptions in the vestibular system can lead to balance and coordination problems, which may indirectly affect ANS regulation in some individuals. Furthermore, several intriguing genes offer additional insights into the pathogenesis of POTS.

P/LP variants with dominant effects

Among the 92 heterozygous P/LP variants, 3 have been reported of dominant genetic effects, including USP48 (ubiquitin specific peptidase 48)/NP_115612.4:p.Gly406Arg causing deafness, autosomal dominant 85; CAPN3 (calpain 3)/NP_000061.1:p.Arg490Trp causing muscular dystrophy, limb-girdle, autosomal dominant 4; POLG (DNA polymerase gamma, catalytic subunit)/NP_002684.1:p.Trp748Ser causing progressive external ophthalmoplegia with mitochondrial DNA deletions, autosomal dominant 1. The co-occurrence of these P/LP variants with POTS could be coincidental. However, CAPN3 encodes a muscle-specific component of the calpain protease, which is a muscle-specific member of the calpain large subunit family, and exhibits a specific binding affinity for the protein titin [67]. The variant causing muscular dystrophy can lead to muscle weakness and mobility issues, contributing to POTS by promoting deconditioning and muscle pump dysfunction. POLG encodes the catalytic subunit of mitochondrial DNA polymerase, a critical enzyme responsible for replicating mitochondrial DNA [47]. POLG plays a pivotal role in maintaining the integrity and proper functioning of mitochondrial DNA, which is essential for the production of energy within cells. Mitochondrial dysfunction can affect multiple physiological processes, including those related to the autonomic nervous system and cardiovascular regulation [43]. However, POLG-related disorders typically do not lead to heart problems [78], and both parents of the proband with the POLG variant do not have POTS. Besides these P/LP variants, the myosin heavy chain 7 (MYH7, related to hypertrophic cardiomyopathy) variant NP_000248.2:p.Arg787Cys at exon21 is classified as DM by HGMD and likely pathogenic by InterVar, but with conflicting interpretations of pathogenicity by ClinVar. MYH7 encodes the beta (or slow) heavy chain subunit of cardiac myosin. This specific heavy chain is primarily expressed in the normal human ventricle, as well as in skeletal muscle tissues rich in slow-twitch type I muscle fibers[100]. Its mutation can affect myocardial contractility.

Insights gained from enriched gene sets with P/LP variants

ORA analysis of the 87 genes with P/LP variants identified several gene sets of statistical significance. Significant DisGeNET gene sets include hepatomegaly (C0019209), epilepsy (C0014544), cerebellar Ataxia (C0007758), seizures (C0036572), failure to gain weight (C0231246), pediatric failure to thrive (C2315100), comatose (C0009421), vomiting (C0042963), and muscle hypotonia (C0026827). Hepatomegaly may be related to splanchnic redistribution of blood, contributing to thoracic hypovolemia in POTS. Epilepsy and seizures often cause autonomic nervous system dysfunction [20]. Cerebellar ataxia, affecting balance and coordination, may contribute to orthostatic intolerance in POTS. Moreover, the association of POTS with gene sets linked to clinical diagnoses such as coma might suggest that certain genetic mutations have a profound impact on neurological functions. Muscle hypotonia can contribute to POTS by promoting deconditioning and muscle pump dysfunction.

Gene sets of GO cellular component include mitochondrial matrix (GO:0005759), and apical part of cell (GO:0045177). Dysfunction in the mitochondrial matrix can lead to energy deficits, which are implicated in dysautonomia and may impact muscle function, including the heart and vascular system, thus contributing to POTS [43]. The apical part of a cell is important in cellular polarization and signalling [11]. In endothelial cells, dysfunction in the apical part could affect vascular tone and blood flow regulation.

Gene sets of GO molecular function include hydrolase activity, acting on glycosyl bonds (GO:0016798); transferase activity, transferring glycosyl groups (GO:0016757); and oxidoreductase activity, acting on paired donors, with incorporation or reduction of molecular oxygen (GO:0016705). Hydrolases that act on glycosyl bonds are involved in the breakdown of carbohydrates and glycoproteins [105]. Impaired carbohydrate metabolism could affect energy availability, potentially influencing the energy-dependent processes of the autonomic nervous system. Glycoproteins play roles in cell signaling and immune responses [77]. Abnormalities in glycoprotein breakdown could contribute to dysregulated immune responses, potentially relevant in autoimmune etiologies of POTS. Glycosylation is important in cell signaling and immune function [89]. Aberrations here could contribute to autoimmune responses or dysregulation of the autonomic nervous system, both implicated in POTS. Oxidoreductase enzymes play a central role in oxidative phosphorylation and energy production in cells, and are closely related to mitochondrial function. These enzymes also play roles in oxidative stress, which has been implicated in various pathologies, including inflammation and autoimmunity.

Conclusions and perspective

Leveraging our expertise in omics and the analysis of heterogeneous phenotypes, this study marks an important step forward in understanding the complex etiologies of POTS, a condition with significant phenotypic heterogeneity and elusive genetic underpinnings. With convincing statistical significance, we have illuminated the role of both common and rare genetic variants in POTS development. We have identified several gene sets through GWAS, notably linked to cell–cell junctions, synaptic membranes, transporter complexes, and early estrogen responses. Our WES analysis brings into focus specific genes and molecular mechanisms including muscular and microtubule dysfunction, autonomic nervous system regulation, and mitochondrial activity. This enhanced genetic understanding opens new avenues for developing personalized treatment strategies, tailored to the unique genetic makeup of individual patients with POTS. Meanwhile, the study’s findings regarding the relationship between POTS-related genes and psychiatric and neurodevelopmental disorders underscore the importance of addressing psychological aspects in the management of POTS.

The burden analysis of VOIs in the WES study identified 55 genes with statistical significance and 87 genes with P/LP variants (including 3 genes with dominant genetic variants). Due to the limitations imposed by the sample size and phenotypic heterogeneity, the GWAS study achieved statistical significance for several gene sets rather than for individual genes. Nonetheless, common variants from several plausible candidate genes might exert regulatory effects as modifiers in the pathophysiology of POTS and merit further investigation. For instance, common variants in the glycoprotein alpha-galactosyltransferase 1 gene (GGTA1), the 3-oxoacid CoA-transferase 2 gene (OXCT2), and the 3-oxoacid CoA-transferase 2 pseudogene 1 gene (OXCT2P1) have shown nominal statistical significance in association with POTS. These variants are linked to gene expression in the heart atrial appendage, as per the Genotype-Tissue Expression (GTEx) project data [57] (Supplementary Table 6), implying a direct role in heart rate regulation [2], a key aspect of POTS pathogenesis. The genetic insights not only enhance our knowledge of POTS pathogenesis, but also hold promise for developing more effective, individualized treatment strategies, ultimately improving patient outcomes in this challenging and multifaceted condition.

This study has limitations. The limited number of subjects in our GWAS study may lead to missed genetic loci and false positives. To address these issues related to statistical power, we advocate for the use of aggregated methodologies, such as gene-set analysis and polygenic risk scores (PRS). In these analyses, a more relaxed statistical threshold can be applied. To specifically address the issue of potential false positives arising from the relaxed threshold, gene-set analysis enhances the reliability of our findings by focusing on groups of genes that share biological functions, reducing the likelihood of spurious associations that might appear significant due to random chance in smaller datasets [16]. This method consolidates weaker signals across multiple genes, providing a more robust signal than individual gene analysis could. Similarly, polygenic risk scores (PRS) compile the effects of numerous variants to estimate an individual’s genetic predisposition to a disease, thereby diluting the impact of any single spurious genetic variant and increasing the overall accuracy of genetic assessments [52]. An exemplary implementation is the use of PRS involving millions of SNPs for complex traits, which has demonstrated excellent performance [64]. These approaches may thus enhance the robustness of findings in studies with small sample sizes. Nonetheless, replication remains crucial, and validation in another cohort would strengthen the evidence for our findings. It is also important to note that for burden analysis of rare variants, using controls from public databases, despite being defined as subjects of European origin, could introduce bias in the representation of population-specific variants. To address this issue, we enhanced the robustness of our comparisons by including 2719 unrelated European non-POTS control subjects, all of whom were sequenced by WES at CAG.

Supplementary Information

Below is the link to the electronic supplementary material.

Funding

This study was funded in part by donation from the Esther Feigenbaum Foundation, the Siemer Family Foundation, by an Endowed Chair in Genomic Research (HH), and by an Institutional Development Award to the Center for Applied Genomics from the Children’s Hospital of Philadelphia.

Data availability

The data that support the findings of this study are available on request from the corresponding author.

Declarations

Conflict of interest

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Ethics approval and consent to participate

Informed consent was obtained from all subjects, or if subjects are under 18 years, from a parent and/or legal guardian with assent from the child if 7 years or older. The Institutional Review Board (IRB) of CHOP approved this study.

Consent for publication

Not applicable.