Abstract
Background Atrial fibrillation (AF) is a prevalent clinical condition worldwide, with a high global incidence that significantly impacts disease burden and mortality rates. Single nucleotide polymorphisms in ABCB1, ABCG2 and CYP3A5 are common, but the clinical outcomes are poorly understood. This study examines the association between the genetic variations of ABCB1, ABCG2 and CYP3A5 and the risk of AF in a Taiwanese population.
Methods This case-control study recruited 216 AF patients from two hospitals in Taiwan between 2021 and 2023. Control groups were matched by age (± one year), gender, and AF-related variables from the Taiwan Biobank. Logistic regression analyzed the association between three genetic variants and AF risk.
Results A significant association was noted between ABCG2 rs2231142 and AF risk. Those with ABCG2 rs2231142 G/T and T/T genotypes had a 1.91-fold (95% CI = 1.04-3.53) increased risk of AF compared to those with the G/G genotype. This association was particularly pronounced in males in those carrying ABCG2 rs2231143 T/T genotype having a 4.47-fold (95% CI = 1.02-19.67) increased risk after adjusting for covariates. There were no overall significant associations between AF risk and the polymorphisms of ABCB1 rs4148738 and rs1128503, nor CYP3A5 rs776746.
Conclusion A robust risk association between the ABCG2 rs2231142 T allele and AF in Asian populations, particularly in male adults, suggests that genetic testing for this polymorphism could be integrated into risk assessment models for AF.
Introduction
Atrial fibrillation (AF) is recognized globally as a prevalent clinical condition that elevates the risk of systemic thromboembolism, heart failure, stroke, dementia and mortality, significantly impacting disease burden and mortality rates worldwide. 1,2 Current estimates indicate that approximately 33 million people are affected by AF globally, demonstrating its extensive reach and the urgency for comprehensive understanding and effective management strategies. 3 The prevalence of AF differs markedly between genders, with rates of 1.4% in males and 0.7% in females. Additionally, incidence rates are reported at 1.68 per 1000 males and 0.76 per 1000 females, highlighting a substantial gender disparity in AF occurrence. 1,2 Age also plays a crucial role, with the Global Burden of Disease Study 2017 (GBD 2017) revealing that AF prevalence increases with age—from 0.07% among those aged 30 to 34, to 8.18% for those aged 95 and above.4
Reports have also suggested potential disparities in the incidence and characteristics of AF across different racial and ethnic groups. 5-7 In Taiwan, heart disease, primarily driven by AF, ranks as the leading cause of death in individuals aged 65 and above and the second leading cause of death overall, following cancer, according to the 2021 national mortality statistics from the Ministry of Health and Welfare. 2 Extensive research has been directed towards unraveling the mechanisms underlying AF. Predominantly, AF is seen as a consequence of diseases such as hypertension, ischemic and structural heart disease, which contribute to the electrical and structural remodeling of the atria. 8-10 Despite the known risk factors such as advancing age, heart disease, elevated blood pressure and excessive alcohol consumption, a significant number of cases occur without identifiable risk factors, hinting at a potential genetic predisposition to AF. This gap in understanding limits the development of effective treatments and highlights the necessity for further investigation, particularly in the realm of precision medicine. 2,11
Over 60% of AF variability is attributed to genetic factors heightening AF susceptibility, influencing AF development and maintenance. 12-14 Genetic variations in the genes ABCB1 (adenosine triphosphate (ATP) binding cassette subfamily B member 1), ABCG2 (ATP-binding cassette subfamily G member 2), and CYP3A5 (cytochrome P450 family 3 subfamily A member 5), responsible for encoding P-glycoprotein (P-gp), breast cancer resistance protein (BCRP) and cytochrome P450 proteins (CYP450) respectively, play a significant role in altering transporter function. These genetic variations impact crucial aspects of AF, including cardiac electrical conduction, ion channel function and structural remodeling. 15,16 Studies have found specific single nucleotide polymorphisms (SNPs) such as rs1128503 and rs4148738 in ABCB1, which modify P-gp expression; furthermore, ABCG2 rs2231142 G>T genotype has been linked to decreased BCRP protein expression, 17,18 while CYP3A5 is crucial in metabolizing anticoagulants and stroke-prevention drugs, and this process significantly impacts the achievement of appropriate anticoagulation levels in patients with AF. 19
Understanding the relationship between these genetic variants and the risk of developing AF is essential as it can provide insight into underlying molecular mechanisms and potentially guide the development of targeted therapies in the era of precision medicine. 20 In recent years, precision medicine has garnered increasing attention, emphasizing the significant relationship between diseases and genes. 21 Each individual possesses a unique genome that can influence their disease susceptibility, treatment response and drug metabolism, 22 and through genetic testing and analysis, researchers can establish correlations between genetic variants and particular diseases. 23 This knowledge can predict an individual’s disease risk, facilitate accurate disease diagnosis, and aid in selecting the most suitable treatment options; 24 however, the current understanding of the relationship between these gene variants and the risk of AF remains unclear. In this study, we aimed to investigate the association between genetic variants of ABCB1 (rs1128503, rs4148738), ABCG2 (rs2231142) and CYP3A5 (rs776746), and the risk of AF.
Materials and methods
Study subjects
This is an ongoing case-control study. The case patients diagnosed as AF by board-certified cardiologists were recruited from Kaohsiung Medical University Hospital (KMUH) and Kaohsiung Municipal Ta-Tung Hospital (KMTTH) between January 2021 and January 2023. The study subjects included individuals aged 20-85 years and willing to cooperate with blood-draw procedures. Patients undergoing cancer treatment, those with impaired renal function (eGFR<15 ml/min/1.73 m2) and those who could not cooperate were excluded from the study. After the written informed consent was received, a designated researcher administered a standardized and structured questionnaire and reviewed the medical charts to collect age, gender, history of hypertension and diabetes and blood test data such as glycated hemoglobin (HbA1C), serum creatinine, triglyceride (TG) and low-density lipoprotein cholesterol (LDL)). A total of 216 AF patients were enrolled in the study period.
Potential healthy controls were selected from the Taiwan Biobank (TWB) cohort, a community-based cohort comprising healthy adults recruited from all over Taiwan. The TWB cohort, established in 2005, aimed to collect genetic and medical information from 200,000 ethnic Taiwanese individuals aged 20 years or above without a history of cancer, where this cohort underwent comprehensive data collection including questionnaires, physical examinations and biochemical tests on blood and urine samples, and has continued to collect biological samples to-date, including DNA, blood plasma and urine from study subjects, which can be provided to researchers in Taiwan for study purposes at a minimal cost. As of 2020, the TWB cohort had enrolled 122,068 potential study adults aged 30-75 years with baseline data. The Institutional Review Board (IRB) of KMUH approved this study.
Among the 216 AF patients, 185 subjects have provided blood samples for genotyping results. In addition, in order to match the AF patients of similar age with healthy controls from TWB cohort, we further excluded 44 patients aged over 76 years and left 141 eligible cases in the study (Figure 1). Then, the case patients were matched with the controls (1: 1 ratio) based on age within a range of ±1 year and the same gender from TWB cohort with specific genotypes of ABCB1 (rs1128503, rs4148738), ABCG2 (rs2231142), and CYP3A5 (rs776746) (n = 108,829). The final study subjects were 141 AF cases and 141 healthy controls.
Genotyping of ABCB1 (rs1128503, rs4148738), ABCG2 (rs2231142) and CYP3A5 (rs776746) polymorphisms from the case study group
The QIAamp DNA Mini Kit (Qiagen, Germany) was used to extract genomic DNA from peripheral blood samples collected from patients in the case study group, following the manufacturer’s instructions, while TaqMan probes designed explicitly for standardized SNPs assays (Thermo Fisher Scientific, Inc) were employed to assess single nucleotide polymorphic variations in the ABCB1 (rs1128503, rs4148738), ABCG2 (rs2231142) and CYP3A5 (rs776746) genes. Probe IDs and SNPs regions are listed in Table S1.
For SNPs genotyping, a reaction mixture containing 10 μL was prepared, containing 10 ng of genomic DNA, 2X TaqMan® Master Mix, 20X Assay Working Stock probe and nuclease-free water. A negative control (no template) was included in the experiment. The cycling conditions adhered to the manufacturer’s standard protocol as recommended where real-time PCR was conducted on all samples using the QuantStudio 3 instrument (Applied Biosystems) and QuantStudio™ Design & Analysis Software was employed to analyze the data and identify homozygous and heterozygous alleles. To ensure the reliability of the genotyping assays, approximately 9.4% of patients (n = 15/141) were randomly selected for the genotyping of each polymorphism; this additional step being performed to validate the consistency and accuracy of the genotyping results. The reaction of PCR for direct sequencing is shown in Supplemental Material and Table S2 that was also performed to confirm the genotyping results (Figure S1). For TWB cohorts, we extracted the genotype information of ABCB1 (rs1128503, rs4148738), ABCG2 (rs2231142), and CYP3A5 (rs776746) genes from Genome-Wide Genotyping: Affymetrix Axiom TWB and TWB 2.0 database, in which genotype analysis was either from the Axiom Genome-Wide TWB array or Axiom Genome-Wide TWB 2.0 array. 25
Statistical analysis
The frequencies of ABCB1 (rs1128503), ABCG2 (rs2231142) and CYP3A5 (rs776746) genotypes were assessed for Hardy-Weinberg disequilibrium in the TWB-1 control group (n = 141). A comparison of demographic characteristics and genotypes between the case group (n = 141) and the TWB-1 control group (n = 141) was performed. Logistic regression models were employed to investigate the impact of ABCB1 (rs1128503, rs4148738), ABCG2 (rs2231142), and CYP3A5 (rs776746) genotypes on AF risk, both before and after adjusting for covariates, including age (continuous variables), gender, diabetes, hypertension, HbA1C, serum creatinine, TG and LDL. Odds ratios (OR), confidence intervals (CI) and p-values were calculated and presented. If both ABCB1 (rs1128503, rs4148738), ABCG2 (rs2231142) and CYP3A5 (rs776746) genotypes showed significant differences between case patients and controls, the combined effect of both genes was further examined. These analyses were also stratified by gender. All statistical analyses were conducted using IBM SPSS Statistics 25 software.
Results
Characteristics of the Study Population
The demographic characteristics of 141 AF patients and the 141 TWB-1 controls are shown in Table 1. Among these 141 AF patients, 48 (34.0%) were female and 93 (66.0%) were male, while the percentages of gender, age group (40-60 and > 60 years) and educational levels were identical in AF patients and TWB-1 controls. Additionally, between the two groups of AF patients and TWB-1 controls, the aforementioned variables (gender, age groups and other variables including diabetes, hypertension, HbA1c, serum creatinine, TG and LDL) were not significantly different (all p > 0.05). Of the 141 AF patients, 24.8% had diabetes and 35.5% had hypertension; additionally, the HbA1c level (6.2%) was higher than the normal range of 4.0-5.6%.
Allele frequencies in TWB cohort
The prevalence of ABCB1 rs1128503 A/A, A/G, and G/G was 41.8%, 46.8% and 11.3% respectively, for TWB-1 (n = 141). For ABCB1 rs4148738 T/T, C/T, and C/C, the prevalence was 34.0%, 41.8% and 24.1% respectively, for TWB-1. ABCG2 rs2231142 G/G, G/T, and T/T were 59.6%, 33.3% and 7.1% respectively, in TWB-1. CYP3A5 rs776746 C/C, T/C, and T/T were 46.1%, 44.0% and 9.9% respectively, in TWB-1 (Table S3). All frequencies of ABCB1 (rs1128503, rs4148738), ABCG2 (rs2231142), and CYP3A5 (rs776746) genotypes in TWB-1 were consistent with the statistics of the Hardy-Weinberg equilibrium, which were insignificant (p > 0.05) (Table S3).
Association of ABCB1, ABCG2 and CYP3A5 polymorphisms
Table 2 shows the association of ABCB1 (rs1128503, rs4148738), ABCG2 (rs2231142), and CYP3A5 (rs776746) genotypes with AF risk before and after adjusting for other covariates (including diabetes, hypertension, HbA1c, serum creatinine, TG and LDL). Genotypes of ABCG2 rs2231142 G/T and T/T exhibited a significant risk of developing AF compared to G/G genotypes after adjusting for other covariates (AOR = 1.912, 95% CI = 1.035-3.53), although no significant difference was observed in the other three SNPs after adjustment.
Categorized by sex, the significant effect of ABCG2 rs2231143 remained present in males but not females, likely due to the small sample size of female AF patients (n = 48) (Table 3). Males with the combined genotypes of ABCG2 rs2231143 G/T and T/T had a significant risk of AF when compared with males carrying ABCG2 rs2231143 G/G genotype (crude OR=2.005, 95% CI=1.118-3.596; AOR=2.179, 95% CI=0.894-5.307). There was also a trend toward statistical significance for the association between male subjects and AF (p = 0.021 and 0.041 before and after adjusting for other covariates), suggesting a potential risk factor associated with this haplotype in males.
Discussion
This hospital-based case-control study underscores the importance of the ABCG2 rs2231142 G/T and T/T genotypes, which exhibit a higher risk of developing AF than those with the G/G genotype, particularly among male patients. However, no associations were identified between the polymorphisms of ABCB1 rs4148738 and rs1128503, and CYP3A5 rs776746 genotypes and AF risk.
ABCG2 belongs to a superfamily of 48 human ATP-binding cassette transporters, which typically facilitate the transport of various substrates and rely on ATP-binding for activation. The ABCG2 rs2231142 variant results in a substitution of glutamine with lysine (Q141K) in exon 5. Previous studies have reported that individuals carrying the ABCG2 rs2231142 variant (G/T and T/T) exhibit significantly lower levels of ABCG2 protein within red blood cells as well as reduced ABCG2 adenosine triphosphatase (ATPase) activity compared to those with the G/G genotype.
The presence of the ABCG2 rs2231142 T allele leads to an approximately 50% reduction in ABCG2 protein expression. 26 Zhang et al. found that variations in the ABCG2 rs2231142 variant could impact the serum levels of pro-inflammatory and pro-angiogenic markers in chronic inflammatory arterial conditions, hinting at a correlation between inflammation and the ABCG2 rs2231142 variant. 27 These findings suggest that ABCG2 rs2231142 could play a crucial physiological role as an excreter of environmental and endogenous toxins, and it might influence the multidrug transporter function in human erythrocytes, corresponding to pharmacologically relevant genetic variations. 28 Furthermore, recent studies have emphasized that the pharmacogenetic implications of these findings could connote the efficacy of direct oral anticoagulants (DOACs) being enhanced, as genetic variants influence drug metabolism and response. 29,30
In recent years, DOACs such as apixaban, dabigatran, edoxaban and rivaroxaban have replaced vitamin K antagonists for treating AF due to their ease of use and better effectiveness. In a rat model of pharmacokinetic interaction, almonertinib significantly increased systemic exposure to apixaban and rivaroxaban by inhibiting ABCB1 and ABCG2. 31 Previous studies have shown variable results regarding the relationship between genotypes and DOACs in AF patients. Gulilat et al. reported that the ABCG2 rs2231142 T allele predicted higher concentrations of apixaban in Caucasian AF patients, 32 but was associated with a lower concentration and lower oral clearance of apixaban in Japanese AF patients. 33,34 Additionally, no correlations were observed between the ABCG2 rs2231142 and rivaroxaban concentrations in Japanese AF patients. 35 Nevertheless, we observed a significant increase in the risk of AF associated with the ABCG2 rs2231142 T allele in patients treated with rivaroxaban (Table S4), indicating that this SNPs might affect the drug transporters and metabolism of DOACs, making additional studies necessary to clarify the role of this SNPs in DOACs pharmacokinetic interactions.
Previous studies have shown that the loss-of-function ABCG2 rs2231142 T allele is associated with hyperuricemia, which might contribute to vascular damage and cardiovascular diseases such as atrial fibrillation, coronary artery disease, heart failure and stroke. 36-38 Similar observations were noted in an earlier prospective cohort study conducted in Taiwan, where hyperuricemia emerged as an independent risk factor for mortality from all causes, total cardiovascular disease and ischemic stroke. 39 In other studies, it is also suggested that the effect size of the ABCG2 rs2231142 is more significant in males. 36,40,41 Our current data showed a similar interaction, as we found a significant association between the rs2231142 T allele and AF risk in males, which was consistent with these Asian population studies, 36-41 and although previous studies have demonstrated that estrogen levels could affect ABCG2 expression, 42,43 this might explain why we observed the association with atrial fibrillation risk predominantly in males rather than females.
Currently, no studies have examined the AF risk linked to the ABCG2 rs2231142 genotype, but research on other SNPs regarding AF risk has been conducted. Genetic factors contributing to AF are gaining more attention, with estimates of heritability reaching 22% in the UK Biobank and 62% in a twin study conducted in Denmark. 14,44 Daniel et al. were the first to report the Genome-wide association studies (GWAS) for AF, where the upstream region of the transcription factor gene PITX2 on 4q25 was significantly associated with AF. 45 Several other genetic loci related to AF have been identified, including variants in or near ZFHX3 (Zinc Finger Homeobox 3), PITX2 (paired like homeodomain 2), and MYOZ1 (myozenin 1), which show the most associations. 46
Our research further highlights the significant relationship between the ABCG2 rs2231142 T allele and increased AF risk, particularly in males. This suggests that integrating genetic testing for this polymorphism into risk assessment models for AF could enhance precision in AF management strategies, especially in populations with a high prevalence of this risk allele. Past studies have primarily focused on pharmacogenomics or the risk of cardiovascular and hyperuricemia. 36-41 The association between polymorphisms of ABG2 rs2231142 genotypes and AF subjects might result from the influence of hyperuricemia, obesity, hypertension and dyslipidemia, as well as sex differences caused by estrogen levels, or both. 47 In the study conducted by Doshi et al., 48 ABCG2 expression was examined in male obesity mice. The findings indicated elevated levels of ABCG2 protein, suggesting a potential association between increased ABCG2 expression and obesity-related AF, although the mechanism of the influence of the ABCG2 rs2231142 locus on the risk effect of developing AF still needs to be investigated.
Table S5 shows the similar variant allele frequencies of ABCB1, ABCG2 and CYP3A5 genotypes in East Asians from different databases and from the datasets of TWB, suggesting the robustness of target genotype frequencies in the public-domain databases. The 1000 Genomes Project (1000Genomes) and the Genome Aggregation Database (gnomAD) are public websites that respectively encompass data from more than 2,500 individuals representing global populations and aggregate over 800,000 exome and whole-genome sequences from seven ethnic groups, sharing summary data with the scientific community (https://www.internationalgenome.org/ and https://reurl.cc/RWnZzg, accessed on 2024/4/18). The TWB is an ongoing prospective study that involves over 200,000 individuals aged 20–70 in Taiwan.
The distributions of genotype and allelic frequencies in Taiwan and other populations such as American, European, African, South Asian, Central Asian and East Asian are summarized in Table S6. The distributions of allelic frequencies in other populations were taken from the database of NCBI (https://www.ncbi.nlm.nih.gov/, assessed on 2024/4/18). The ABCB1 rs1128503 G allele and rs4148738 T allele have a higher prevalence among Africans, with frequencies of 80% and 77% respectively, while the ABCG2 T allele is more common in American, East Asian and Taiwanese populations consistent with the findings reported by Alrajeh et al. 49 Their reports noted that the frequency of the allele associated with reduced ABCG2 function is exceptionally high among Asians while the frequency of the CYP3A5 C allele is lower in African populations, with a frequency of 31%; interestingly, the prevalence of the ABCG2 T allele is higher in East Asian, Taiwanese and American populations (29.8%, 31.5% and 22.5%, respectively) compared to European, South Asian and African populations (9.8%, 9% and 2.6%, respectively). These differences may be attributed to factors such as database sources, study designs and ethnicities.
Conclusion
In conclusion, the findings suggest a strong association between the ABCG2 rs2231142 T allele and a higher risk of AF among Taiwanese male subjects, with a 4.5-fold higher risk compared to controls, and a 1.9-fold risk when combining ABCG2 genotypes in the overall population. However, no significant associations were identified between the polymorphisms of ABCB1 rs4148738 and rs1128503, as well as CYP3A5 rs776746 genotypes and AF risk. Further studies with larger sample sizes are necessary to investigate the interaction of the ABCG2 rs2231142 genotype with AF risk.
Data Availability
The data are not publicly available due to the nature of information that could compromise the privacy of research participants.
Sources of Funding
This work was supported by grants from Kaohsiung Medical University Hospital (KMUH108-8R71), Kaohsiung Medical University Research Center Grant (KMU-TC108A01), from the Integrative Project of National Health Research Institutes and Kaohsiung Medical University (NHRIKMU-111-I005), Ministry of Science and Technology (MOST 108-2638-B-037-001-MY2).
Disclosures
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Authors’ contributions
Tzu-Yen Lin was responsible for Methodology, Validation, Investigation, Data curation; Tzu-Yu Pan was responsible for Methodology, Validation, Formal analysis, Resources, Data curation, Writing – original draft and Writing-Reviewing and Editing; Wei-Chung Tsai were responsible for Data Curation and Resources, Writing-Review and editing; Ming-Tsang Wu was responsible for Conceptualization, Writing-Review and editing, Supervision, Project administration, Funding acquisition.
Availability of data and materials
The data are not publicly available due to the nature of information that could compromise the privacy of research participants.
Ethics approval and consent to participate
The study design was approved by the institutional review board of Kaohsiung Medical University Hospital (KMUHIRB-G(I)-20200040).
Acknowledgment
Not applicable.
Nonstandard Abbreviations and Acronyms
- 1000Genomes
- 1000 Genomes Project
- AF
- Atrial fibrillation
- ABCB1
- Adenosine triphosphate binding cassette subfamily B member 1
- ABCG2
- ATP-binding cassette subfamily G member 2
- ATPase
- Adenosine triphosphatase
- BCRP
- Breast cancer resistance protein
- CYP3A5
- Cytochrome P450 family 3 subfamily A member 5
- CYP450
- Cytochrome P450 proteins
- DOACs
- Direct oral anticoagulants
- GBD 2017
- Global Burden of Disease Study 2017
- gnomAD
- Genome Aggregation Database
- GWAS
- Genome-Wide Association Studies
- HbA1C
- Glycated hemoglobin
- IRB
- Institutional Review Board
- KMUH
- Kaohsiung Medical University Hospital
- KMTTH
- Kaohsiung Municipal Ta-Tung Hospital
- LDL
- Low-density lipoprotein cholesterol
- MYOZ1
- Myozenin 1
- P-gp
- P-glycoprotein
- PITX2
- Paired like homeodomain 2
- SNPs
- Single Nucleotide Polymorphisms
- TG
- Triglyceride
- TWB
- Taiwan Biobank
- ZFHX3
- Zinc finger homeobox 3