Plasma 8-Hydroxy-2′-Deoxyguanosine, a Potential Valuable Biomarker for Atrial Fibrosis Is Influenced by Polymorphism of DNA Methylation Gene

Xiafeng Peng; Shixin Wang; Jing Wang; Weizhu Ju; Gang Yang; Kai Gu; Hailei Liu; Zidun Wang; Xiaohong Jiang; Mingfang Li; Hongwu Chen; Jiaojiao Shi; Minglong Chen

doi:10.1253/circj.CJ-22-0694

Abstract

Background: Previous studies revealed a relationship between 8-hydroxy-2′-deoxyguanosine (8-OHdG) and the occurrence/recurrence of atrial fibrillation (AF). This 2-part study aimed to validate whether DNA damage related to 8-OHdG is associated with left atrial (LA) fibrosis in AF patients quantified by voltage mapping (Part I), and to identify the underlying genetic components regulating the 8-OHdG level (Part II).

Methods and Results: Plasma 8-OHdG determination, DNA extraction, and genotyping were conducted before catheter ablation. LA voltage mapping was performed under sinus rhythm. According to the percentage of low voltage area (LVA), patients were categorized as stage I (<5%), stage II (5–10%), stage III (10–20%), and stage IV (>20%). Part I included 209 AF patients. The 8-OHdG level showed an upward trend together with advanced LVA stage (stage I 8.1 [6.1, 10.5] ng/mL, stage II 8.5 [5.7, 14.1] ng/mL, stage III 14.3 [12.1, 16.5] ng/mL, stage IV 13.9 [10.5, 16.0] ng/mL, P<0.000). Part II included 175 of the 209 patients from Part I. Gene-set analysis based on genome-wide association study summary data identified that the gene set named ‘DNA methylation on cytosine’ was the only genetic component significantly associated with 8-OHdG concentration.

Conclusions: Higher 8-OHdG levels may predict more advanced LVA of the LA in AF patients. DNA methylation is the putative genetic component underlying oxidative DNA damage in AF patients.

Atrial fibrillation (AF) is the most common cardiac arrhythmia and can induce severe complications such as stroke and heart failure.¹ Atrial fibrosis is confirmed to be relevant to AF progression and prognosis,²^–⁴ and preoperative evaluation of fibrotic degree before ablation contributes to selection of the best candidates for ablation and effective use of medical resources.⁵ In clinical practice prediction of the fibrotic degree is usually based on the morphology of the P/p wave on ECG and the left atrial (LA) dimension assessed with echocardiography, but the specificity and sensitivity are inferior to that of voltage mapping. Therefore, we still need more reliable and convenient methods to resolve this issue.

Oxidative stress plays a vital role in the pathological process of many diseases and also in the electrical and contractile impairment of atrial cardiomyocytes in AF.⁶ Tascanov et al found that total oxidant status and DNA damage levels were significantly higher in patients with paroxysmal AF (PAF) than in healthy controls.⁷ In addition, the metabolites of oxidative stress, such as circulating mitochondrial DNA (mtDNA), are valuable for staging AF.⁸

8-hydroxy-2′-deoxyguanosine (8-OHdG) is a major byproduct of oxidative DNA damage. It is formed when hydroxyl radicals interact with the double bond at the C-8 position of the guanine base⁹ and its stable property make it useful for evaluating oxidative DNA damage in various diseases.¹⁰ Toyama et al found that urinary 8-OHdG/creatinine levels are significantly higher in AF patients and are relieved with restoration of sinus rhythm.¹¹ Furthermore, 8-OHdG has shown an ability to predict AF recurrence and postoperative AF.¹² However, all these studies have only described a clinical phenomenon, and studies exploring the relationship between the 8-OHdG levels and pathologic change in the atrium are still absent, Therefore, this study had 2 aims: in Part I to examine the plasma 8-OHdG levels in AF patients with differing severity of LA fibrosis, as assessed by the proportion of low voltage area (LVA) acquired through electroanatomic bipolar voltage mapping; in Part II, inspired by the regulatory role of single nucleotide polymorphism (SNP) for 8-OHdG,¹³ we performed a genome-wide association study (GWAS) and derived gene-based/gene-set analyses to identify the enriched gene network related to the plasma 8-OHdG concentration.

Methods

A flowchart of the study is presented in Figure 1.

Figure 1.

Flowchart of the overall study. AF, atrial fibrillation; 8-OHdG, 8-hydroxy-2′-deoxyguanosine; LVA, low voltage area; SNP, single nucleotide polymorphism.

Study Population

A total of 209 patients undergoing initial AF ablation were consecutively enrolled. PAF was defined if AF terminated within 7 days; non-PAF was defined when AF duration lasted >7 days. After admission, transthoracic echocardiography was performed before AF ablation. Patients with the following conditions or had been diagnosed with the following diseases were excluded: ablation history, chronic heart failure with left ventricular ejection fraction <50% or acute heart failure, history of acute coronary syndrome, valvular heart disease, cardiomyopathy, myocarditis, atrial amyloidosis, LA diameter (LAd) >55 mm, obstructive sleep apnea, thyroid function abnormality, severe chronic obstructive pulmonary disease, acute or terminal liver or kidney disease, rheumatoid disease, malignant tumors, or acute active infection. The detailed definitions of comorbidities are given in the Supplementary Methods. All study participants signed an informed consent form, and the study conformed with the ethical guidelines of the Declaration of Helsinki and was approved by the Human Research Ethics Committee of the First Affiliated Hospital of Nanjing Medical University (reference no. 2022-SR-256).

8-OHdG Measurement, DNA Extraction and Genotyping Array

Fasting venous blood samples of all patients were collected in the morning before ablation for the measurement of plasma 8-OHdG, and the extraction and genotyping of genomic DNA. Detailed methods are available in the Supplementary Methods.

LA Voltage Mapping

LA voltage mapping was performed after circumferential pulmonary vein isolation under sinus rhythm. Detailed procedures are available in the Supplementary Methods. The LVA was defined as an area with a bipolar peak-to-peak voltage amplitude <0.5 mV¹⁴ in >3 adjacent low voltage points with a space difference of 0.5 cm. The size of the LVA was shown as a percentage of LVA to the LA surface area (excluding the pulmonary vein antral region, LA appendage orifice, and mitral valve). According to the LVA percentage, patients were categorized as stage I (<5%, n=45), stage II (5–10%, n=56), stage III (10–20%, n=76), and stage IV (>20%, n=32), with further subdivision into non-severe LVA (stages I and II, n=101) and severe LVA (stages III and IV, n=108) groups.

Quality Control of Genotype Data and Principal Components Analysis

Quality controls of the genotype data, including SNP and sample quality control, were conducted by ‘PLINK’ software.¹⁵ The principle components analysis was accomplished by GCTA software to assess population stratification.¹⁶ Details are given in the Supplementary Methods. Finally, 298,109 high-quality autosomal SNPs and 175 independent individuals were used to perform the GWAS (Figure 1).

GWAS, Gene-Based, and Gene-Set Analyses

Applying ‘PLINK’, a linear regression model was used in the GWAS for testing the associations between SNPs and plasma 8-OHdG concentration. We included age, body mass index (BMI), LAd, and PC1–PC5 as covariates in the linear regression. The Manhattan plot was visualized by the R package ‘CMplot’, and the genome-wide significance threshold was assessed with the formula P=0.05/n, where n is the number of independent SNPs.¹⁷ Therefore, the P value threshold for genome-wide significant SNPs was approximately 1.68×10⁻⁷.

Because of the small sample size of our study, the effects of individual markers may have been too weak to detect, so to increase power, we used ‘MAGMA’, a method analyzing multiple genetic markers together to determine their joint effect, to conduct gene-based and gene-set analyses based on the GWAS summary data.¹⁸ Input SNPs were mapped to 11,630 genes and 32,681 gene sets; therefore, the significant thresholds for gene-based and gene-set analyses were P<4.30×10⁻⁶ and 1.53×10⁻⁶, respectively.

Statistical Analysis

Continuous, normally distributed variables are expressed as mean±SD. Continuous, non-normally distributed variables are expressed as medians and interquartile ranges. Comparison between groups was carried out using Student’s t-test, the Mann-Whitney U test, or the Kruskal-Wallis test. Categorical variables are presented as frequencies with percentages (%), and comparison between groups was carried out using Fisher’s exact test or Pearson’s chi-squared test. Univariate and multivariate logistic regression analyses were used to determine the factors related to severe LVA, assessed by calculating the odds ratio (OR) and 95% confidence interval (CI), and only factors with a P value <0.2 were involved in the univariate and multivariate models. Receiver operating characteristic (ROC) curve analysis was used to evaluate diagnostic accuracy. All analyses were two-tailed, and P values <0.05 were considered statistically significant. The statistical analyses were performed using SPSS version 26.0 (SPSS Inc., Chicago, IL, USA) and PLINK version 1.9 (Shaun Purcell).

Results

Baseline Characteristics

In total, 209 patients (64.0 [55.0, 69.5] years, 67.9% male, 48.3% non-PAF) were included and their clinical characteristics are summarized in Table 1. Patients with more advanced stages of LVA were older and were more likely to be female with non-PAF. Moreover, they had significantly larger LAd and right atrial diameter (RAd), and higher levels of N-terminal prohormone of B-type natriuretic peptide (NT-proBNP). The BMI, comorbidities and echocardiographic parameters of the ventricle were not significantly different among the 4 groups of LVA stages.

Table 1. Baseline Characteristic of the Patients Included for 8-OHdG-LVA Analysis

	Total study population (n=209)	Stage I (n=45)	Stage II (n=56)	Stage III (n=76)	Stage IV (n=32)	P value
Age (years)	64.0 [55.0, 69.5]	59.0 [49.5, 68.5]	64.0 [53.5, 70.0]	63.0 [56.3, 68.0]	68.0 [58.3, 70.8]	0.044*
Sex, male (n, %)	142 (67.9)	36 (80.0)	43 (76.8)	52 (68.4)	11 (34.4)	0.000*
BMI (kg/m²)	25.0 [23.7, 26.5]	24.8 [23.5, 25.4]	25.0 [23.8, 26.8]	25.0 [23.8, 27.4]	25.0 [22.1, 27.1]	0.249
Hypertension (n, %)	112 (53.6)	24 (53.3)	31 (55.4)	40 (52.6)	17 (53.1)	0.992
Diabetes (n, %)	35 (16.7)	4 (8.9)	11 (19.6)	13 (17.1)	7 (21.9)	0.401
CAD (n, %)	32 (15.3)	3 (6.7)	14 (25.0)	12 (15.8)	3 (9.4)	0.057
Stroke (n, %)	5 (2.4)	0 (0.0)	3 (5.4)	2 (2.6)	0 (0.0)	0.366
Nonparoxysmal AF (n, %)	101 (48.3)	15 (33.3)	29 (51.8)	36 (47.4)	21 (65.6)	0.042*
LAd (mm)	40.9 [38.0, 44.0]	39.0 [35.5, 40.4]	40.9 [37.3, 44.0]	41.0 [39.0, 44.0]	44.0 [40.3, 46.0]	0.000*
RAd (mm)	39.0 [35.0, 40.0]	35.0 [32.0, 40.0]	38.0 [34.0, 40.0]	39.0 [36.0, 41.0]	40.0 [38.0, 41.0]	0.001*
LVDd (mm)	47.5 [45.0, 50.0]	47.0 [45.0, 49.0]	48.0 [46.0, 51.0]	48.0 [45.3, 50.0]	47.0 [45.0, 48.8]	0.407
RVDd (mm)	34.0 [32.0, 36.0]	33.0 [31.0, 36.0]	34.0 [32.0, 36.0]	34.0 [33.0, 35.0]	34.0 [32.0, 36.0]	0.454
LVEF (%)	63.3 [61.9, 64.4]	63.3 [62.0, 64.4]	63.0 [61.7, 64.0]	63.4 [61.7, 64.6]	63.0 [61.4, 65.2]	0.659
NT-proBNP (ng/L)	446.0 [189.0, 843.0]	317.0 [124.5, 712.0]	351.5 [156.8, 730.3]	466.5 [201.3, 785.0]	875.0 [555.9, 1,605.4]	0.000*

Data are presented as median [IQR] or n (%). *P<0.05. 8-OHdG, 8-hydroxy-2′-deoxyguanosine; AF, atrial fibrillation; BMI, body mass index; CAD, coronary artery disease; LAd, left atrial diameter; LVA, low voltage area; LVDd, left ventricular end diastolic diameter; LVEF, left ventricular ejection fraction; NT-proBNP, N-terminal prohormone of B-type natriuretic peptide; RAd, right atrial diameter; RVDd, right ventricular end diastolic diameter.

Relation Between Plasma 8-OHdG Level and LVA Size

Plasma 8-OHdG levels determined before ablation are shown in Figure 2. Stage III patients showed higher plasma 8-OHdG levels than stages I (14.3 [12.1, 16.5] ng/mL vs. 8.1 [6.1, 10.5] ng/mL, P<0.000) and II (14.3 [12.1, 16.5] ng/mL vs. 8.5 [5.7, 14.1] ng/mL, P<0.000). Similar results were observed for stage IV compared with stages I (13.9 [10.5, 16.0] ng/mL vs. 8.1 [6.1, 10.5] ng/mL, P<0.000) and II (13.9 [10.5, 16.0] ng/mL vs. 8.5 [5.7, 14.1] ng/mL, P<0.000). However, the levels between both stages I and II (8.1 [6.1, 10.5] ng/mL vs. 8.5 [5.7, 14.1] ng/mL, P=0.692), and stages III and IV (14.3 [12.1, 16.5] ng/mL vs. 13.9 [10.5, 16.0] ng/mL, P=0.558) showed no significant differences. We further divided the study population into a non-severe LVA group (n=101, stage I and II) and severe LVA group (n=108, stage III and IV). Characteristics with a P value <0.2 in Table 1, including age, sex, coronary artery disease (CAD), AF type, LAd, RAd, NT-proBNP level, and 8-OHdG level, were included in the univariate logistic regression analysis (Table 2), and age, sex, LAd, RAd, NT-proBNP, and 8-OHdG were found to be associated with severe LVA (all P<0.05). All these factors were further included in the multivariate logistic regression analysis (Table 3), and both sex and 8-OHdG level could predict the presence of severe LVA independent of other factors (P<0.05).

Figure 2.

Relationship of plasma 8-OHdG level and LA fibrosis quantitatively evaluated by LVA size. Stage III and IV patients showed significantly higher plasma 8-OHdG levels compared with stages I and II. The differences in the plasma 8-OHdG levels between both stage I and II, and stages III and IV were not significant. *P<0.05, compared with stage I; **P<0.05, compared with stage II. 8-OHdG, 8-hydroxy-2′-deoxyguanosine; LA, left atrial; LVA, low voltage area.

Table 2. Results of Univariate Logistic Regression Analysis of Characteristics Related to Severe LVA

Characteristic	Coefficient	OR	95% CI	P value
Age (years)	0.032	1.032	1.005–1.060	0.022*
Sex, male	−0.942	0.390	0.212–0.716	0.002*
CAD	−0.227	0.797	0.375–1.695	0.797
Nonparoxysmal AF	0.370	1.448	0.839–2.498	0.184
LAd (mm)	0.095	1.099	1.032–1.171	0.003*
RAd (mm)	0.109	1.116	1.044–1.192	0.001*
NT-proBNP (ng/L)	0.000	1.000	1.000–1.001	0.047*
8-OHdG (ng/mL)	0.239	1.269	1.177–1.369	0.000*

*P<0.05. CI, confidence interval; OR, odds ratio. Other abbreviations as in Table 1.

Table 3. Results of Multivariate Logistic Regression Analysis of Characteristics Related to Severe LVA

Characteristics	Coefficient	OR	95% CI	P value
Age (years)	0.025	1.025	0.990–1.062	0.166
Sex, male	−1.076	0.341	0.162–0.717	0.005*
CAD	−0.153	0.858	0.334–2.200	0.749
Nonparoxysmal AF	−0.017	0.983	0.447–2.161	0.966
LAd (mm)	0.058	1.060	0.956–1.174	0.268
RAd (mm)	0.067	1.070	0.961–1.191	0.219
NT-proBNP (ng/L)	0.000	1.000	0.999–1.000	0.645
8-OHdG (ng/mL)	0.228	1.256	1.159–1.360	0.000*

*P<0.05. Abbreviations as in Tables 1,2.

Diagnostic Value of 8-OHdG for Severe LVA

ROC curve analysis showed that a plasma 8-OHdG level >11.35 ng/mL could predict the presence of severe LVA with an area under the curve (AUC) of 0.764, sensitivity of 78.0%, and specificity of 74.0% (P<0.00) (Figure 3).

Figure 3.

ROC curve of plasma 8-OHdG level in predicting the risk of severe LVA. An 8-OHdG level >11.35 ng/mL could predict the risk of severe LVA with a sensitivity of 78.0% and specificity of 74.0% (P<0.000). Area under the curve was 0.764. 8-OHdG, 8-hydroxy-2′-deoxyguanosine; LVA, low voltage area; ROC, receiver operating characteristic; SE, sensitivity; SP, specificity.

Enriched Gene Sets Affecting the Plasma 8-OHdG Level

GWAS According to Plasma 8-OHdG Level We conducted a GWAS in 175 of the 209 patients for 298,109 autosomal SNP loci. The Manhattan plot shown in Figure 4 revealed no genome-wide significant association signal (P<1.68×10⁻⁷).

Figure 4.

Genome-wide association study based on plasma 8-OHdG levels. The dotted line represents the significant threshold (P=1.68×10⁻⁷). Chr, chromosome.

Gene-Based and Gene-Set Analyses The gene-based analysis conducted by MAGMA did not show any genes reaching genome-wide significance (P<4.30×10⁻⁶). The complete results of the analysis are presented in Supplementary Table 1. We further applied gene-set analysis by MAGMA, which revealed that the significantly enriched gene set ‘DNA methylation on cytosine’ (P=9.5×10⁻⁷) was the most significantly associated with the plasma 8-OHdG level, and also the only one achieving genome-wide significance (P<1.53×10⁻⁶) (Table 4). The complete results of this analysis are presented in Supplementary Table 2. The ‘DNA methylation on cytosine’ gene set contained euchromatic histone lysine methyltransferase 2 (EHMT2), DNA methyltransferase 1 (DNMT1), DNMT3A, DNMT3B, DNMT3L, sperm-tail PG-rich repeat containing 4 (STPG4) genes (searched in Molecular Signatures Database, https://www.gsea-msigdb.org/gsea/msigdb/index.jsp). Accordingly, we listed all the SNPs located in these 6 genes in our GWAS summary data (Supplementary Table 3). Furthermore, 8 of the SNPs shown in Supplementary Table 3, with a P value <0.05 and reported in Pubmed, the NCBI dbSNP database, or GWAS catalog, are listed in Supplementary Table 4.

Table 4. Top 10 Significant Gene Sets of Gene-Based Analysis by MAGMA

Database	Gene set	No. of genes	β	SE	P value
GO bp	DNA methylation on cytosine	6	1.513	0.317	9.50E-07
GO bp	Cellular response to cocaine	8	1.193	0.263	2.90E-06
GO bp	Cellular response to alkaloid	32	0.644	0.151	1.01E-05
GO bp	Response to antibiotic	34	0.632	0.160	4.05E-05
GSE40666	WT vs. STAT1 KO CD8 T cell with IFNα STIM 90 min dn	135	0.291	0.080	1.38E-04
HPO	Tongue thrusting	7	1.366	0.379	1.59E-04
GSE39820	Control vs. TGFβ3 IL6 IL23A CD4 T cell up	136	0.282	0.078	1.60E-04
GO bp	Fatty acid derivative catabolic process	11	0.958	0.273	2.21E-04
WP	MTHFR deficiency	21	0.616	0.177	2.50E-04
GO bp	Regulation of inositol phosphate biosynthetic process	13	0.854	0.252	3.51E-04

bp, biological process; GO, gene ontology; GSE, gene expression data set; HPO, human phenotype ontology; SE, standard error; WP, WikiPathways.

Discussion

Main Findings

In this study, we observed that 8-OHdG levels were significantly increased in patients with severe LVA (≥10%) compared with those with non-severe LVA (<10%). Moreover, a plasma 8-OHdG level >11.35 ng/mL showed diagnostic value for distinguishing severe LVA. Our investigations of the genomic aspect further implied that genetic polymorphism of the ‘DNA methylation cytosine’ gene set was the potential regulator of 8-OHdG levels in AF patients.

Variation Pattern and Significance of 8-OHdG for Progressive Atrial Fibrosis in AF

Inspired by the reported close connection between oxidative DNA damage and AF,⁶^–⁸^,¹⁹^–²¹ we explored the potential relationship between oxidative DNA damage evaluated by plasma 8-OHdG and the LA fibrotic burden assessed by voltage mapping. We found that the plasma 8-OHdG levels in patients with stage III or IV LVA (≥10%) were significantly higher than in stage I or II patients (LVA <10%), although the differences between stages III and IV, and stages I and II, were not statistically significant. Therefore, we further divided the patients into non-severe LVA (stages I and II) and severe LVA (stages III and IV) subgroups.

This variation pattern of plasma 8-OHdG levels in our study might reflect distinct levels of oxidative DNA damage, or even oxidative stress, among the different stages of LVA. In fact, AF patients with severe LVA might be a population with more advanced “atrial cardiomyopathy”.²² Recently, an expert consensus has identified and depicted “atrial cardiomyopathy” that contributes to atrial arrhythmogenicity and thrombogenesis, with atrial fibrosis being one of the crucial characteristics.²² As a central mediator for atrial remodeling in AF, special emphasis is given to the role of aggravated oxidative stress in atrial cardiomyopathy.²³ Our previous study confirmed that in RA biopsies from patients with idiopathic isolated fibrotic atrial cardiomyopathy, there was a hypermetabolic state with significant enrichment of oxidative stress-related proteins, mitochondrial damage, and profound fibrosis.²⁴

Moreover, developing research has directly unmasked the role of oxidative DNA damage in cardiac fibrosis. In reactive oxygen species (ROS)-mediated mtDNA damage, Wang et al found that cardiac overexpression of DNA repair enzyme 8-oxoguanine DNA glycosylase 1 (OGG1) reduced the mtDNA content of 7,8-dihydro-8-oxoguanine (8-oxo-dG) in stressed hearts, and this protection of mtDNA was associated with mitigated cardiac fibrosis.²⁵ For oxidative damage of nuclear DNA, pathological studies by Frustaci et al suggested serious oxidative damage of myocyte DNA in cocaine-related cardiomyopathy, which resulted in extensive myocardial fibrosis.²⁶

For 8-OHdG, previous studies revealed significantly increased levels of 8-OHdG with extended AF duration, and sinus rhythm restoration improved this enhanced oxidative stress status.⁷^,¹¹^,¹² Furthermore, increased levels of this biomarker showed a correlation with AF recurrence after pulmonary vein isolation and with postoperative AF. Therefore, it seems there is continuous involvement of 8-OHdG in the development, progression and relapse of AF. The previous studies⁷^,¹¹^,¹² adopted a temporal pattern-based stratification of AF, whereas our substrate-based classification of AF patients might reflect the severity of atrial cardiomyopathy more faithfully, and contribute more to clarifying the role of oxidative DNA damage in fibrotic remodeling in AF.²⁷

However, we also found that plasma 8-OHdG levels did not differ significantly between stages I and II, or stages III and IV. A similar variation pattern of serum autoantibodies against M2-muscarinic receptors was observed in a recent study.²⁸ This phenomenon may indicate that the progression of atrial fibrosis in AF is not constant, with different fibrotic promoters being selectively activated at different stages and to differing extents. Moreover, there was even an inclination towards a decrease in 8-OHdG levels in stage IV compared with stage III. Similar to our findings, Wiersma et al⁸ revealed a significant increase in cell-free circulating mtDNA levels in PAF patients but a gradual decrease in persistent AF and longstanding-persistent AF patients. They attributed it mainly to exhaustion of mitochondrial function in the terminal stages of AF. Because mtDNA might be a more predominant contributor to the plasma 8-OHdG level than nuclear DNA,²⁹ the exhaustion of mitochondrial function in the terminal phases may greatly affect the coeval plasma 8-OHdG level.

Diagnostic Capability of 8-OHdG for Severe LVA and Their Interrelationship

The present study set a cutoff value for severe LVA at 10%. Similarly, Vlachos et al defined the threshold of significant LVA as 10%, and it showed a significant correlation with AF recurrence³⁰ and P wave vector magnitude,³¹ indicating its significance in reflecting AF prognosis and impaired electrical activity in the LA. The approximate scar threshold and similar proportion of cases with significant LVA in both our study and Vlachos et al³⁰ indicated that the 10% cutoff is representative of AF atrial remodeling.

In our multivariate logistic regression analysis, we found that only sex and plasma 8-OHdG level were associated with severe LVA. Female sex has been widely confirmed to be related to a higher burden of atrial fibrosis,³² but age and the ‘classical’ comorbidities (e.g., hypertension, diabetes) show no association with the extent of atrial fibrosis.³³ This detrimental role of female sex in atrial fibrosis may be attributed to the effect of menopause-related estrogen deficiency in exacerbating oxidative stress during cardiac fibrotic remodeling.³⁴^,³⁵

The capability of 8-OHdG to diagnose severe LVA was also shown through the ROC curve analysis, with an AUC of 76.4%, sensitivity of 78.0%, and specificity of 74.0%, suggesting relatively good diagnostic ability of the plasma 8-OHdG level. Nonetheless, there was still a wide overlapping range of 8-OHdG levels in the non-severe and severe LVA groups, which limited its diagnostic power. This limitation suggests the requirement of combined indexes for the diagnosis of advanced LVA stage. For instance, our study of soluble suppression of tumorigenicity 2 (sST2), a well-known profibrotic biomarker, suggested its significance in predicting new abnormal substrate during redo ablation for AF patients.³⁶ Combined consideration of sST2 and 8-OHdG may optimize the identification of atrial fibrosis.

Interaction Between ROS and DNA Methylation

DNA methylation mediated by DNA methyltransferases (DNMT) is one of the major mechanisms of epigenetic regulation. It refers to a covalent transfer of a methyl group from S-adenosyl methionine to the C5 of cytosine residues in the cytosine-paired-with-guanine (CpG) dinucleotide sequences. It has been found that methylation of CpG islands in the promoter region leads to epigenetic silencing of regulatory genes.³⁷ Three active DNMTs have been described in mammals: DNMT1 is required for the maintenance of methylation through generations, while DNMT3A and DNMT3B are mainly involved in the establishment of de novo DNA methylation patterns, and the activity of de novo methyltransferases are enhanced by DNMT3L.³⁸

In our gene-set analysis by MAGMA, we identified that ‘DNA methylation on cytosine’ was the only genome-wide significant gene set, and a total of 6 genes (EHMT2, DNMT1, DNMT3A, DNMT3B, DNMT3L, and STPG4) were embodied in this set. This finding suggested that during the progression of LVA in the LA of AF patients, there might be involvement of DNA methylation in oxidative DNA damage. In fact, there is an extensive interconnection between ROS and DNA methylation in cardiovascular diseases. ROS affect DNA methylation through various means. (1) Direct modification of DNA bases with modified production of 8-OHdG and 5-hydroxymethylcytosine; 8-OHdG could impede methylation in the new DNA strand,³⁹ and 5-hydroxymethylcytosine can inhibit the inheritance of methylation patterns by interfering with DNMT1.⁴⁰ (2) Affecting either the activity or expression of DNMTs. For instance, hydrogen peroxide inhibits the activity of DNMTs through S-adenosyl methionine,⁴¹ and ROS can also increase DNMTs expression via hypoxia-inducible factor 1.⁴² (3) Inducing the binding of DNMT-containing complexes. It has been confirmed that ROS promote the recruitment of DNMT1-involved complex to induce the hypermethylation of the E-cadherin promoter.⁴³ (4) Modulating methylation through ten-eleven translocation family proteins.⁴⁴ Epigenetic regulation-mediated oxidative stress has also been revealed. For instance, methylation of the superoxide dismutase gene is implicated in atherosclerosis,⁴⁵ and pulmonary arterial hypertension,⁴⁶ with redox signaling potentially being the disrupted downstream pathway. Promoter demethylation of mitochondrial adaptor p66Shc induced by hyperglycemia could enhance mitochondrial oxidative stress in the progression of diabetic vascular complications.⁴⁷ In summary, considering the tight connection between oxidative stress and DNA methylation, it is reasonable to infer a potential role of epigenetic gene variants in oxidative DNA damage.

Investigation of SNPs Involved in ‘DNA Methylation on Cytosine’

Based on the selected SNPs shown in Supplementary Table 3, we subsequently chose those with a P value <0.05 and also reported in Pubmed, the NCBI dbSNP database, or GWAS catalog. Finally, 8 SNPs were listed for detailed investigation (Supplementary Table 4). Taken together, these SNPs widely participate in the progression of neoplastic, hematological, autoimmune, and neuropsychiatric diseases, and are significantly associated with some hematological, biochemical, and physical parameters. The underlying mechanism may be attributed to the effect of these SNPs on the expression or activity of DNMTs, which subsequently regulate the methylation of target genes to facilitate disease progression or participate in trait regulation. However, there is a lack of direct evidence supporting the involvement of these particular SNPs in oxidative DNA damage. Therefore, there remains a need to clarify the role of epigenetic gene variants in oxidative DNA damage.

Other Gene Sets Related to Plasma 8-OHdG Levels

In addition, fatty acid and 1 carbon cycle metabolism have been widely considered to be interrelated with oxidative stress,⁴⁸ or even oxidative DNA damage,⁴⁹^,⁵⁰ and there has been gradual recognition of the role of 1,2,3,4,5,6-cyclohexanehexol as an antioxidant.⁵¹ This evidence may partly explain the suggested significance of ‘Fatty acid derivative catabolic process’, ‘MTHFR deficiency’, and ‘Regulation of inositol phosphate biosynthetic process’ on plasma 8-OHdG levels (Table 4). Collectively, our analysis in-silico of the GWAS summarized data suggested the potential effect of DNA methylation in oxidative DNA damage characterized by plasma 8-OHdG, as well as the possible participation of fatty acid, 1 carbon cycle, and inositol phosphate metabolism in this process.

Study Limitations

There are several limitations to this study. For Part I, first, investigating multiple oxidative DNA damage indicators and antioxidants will contribute to a comprehensive evaluation of DNA damage. Second, we did not perform RA voltage mapping, by AF is gradually being understood as a biatrial remodeling disease. Third, we could not identify the specific source of 8-OHdG and clarify the causation between elevated 8-OHdG levels and progressive LVA of the LA, which needs a deeper dive into experimental studies. Fourth, although bipolar voltage amplitude has been widely accepted for fibrotic substrate identification in clinical practice, LVA could not be completely equated with fibrosis due to the lack of sufficient pathological evidence.

For Part II, first, we had a small, monoracial sample from a single center. Although we used a multimethod strategy including gene-based and gene-set analyses to increase the power, the underpowered sample size could still elicit false-positive findings. Second, more research needs to be conducted to reveal the actual regulatory mechanism of SNPs on epigenetic-regulatory genes in oxidative DNA damage. Third, there are still plenty of enriched but untapped gene sets in our analysis, making our interpretation inadequate.

Conclusions

The elevated levels of an oxidative DNA damage marker, 8-OHdG, were related to more advanced LVA of the LA in AF patients. In-silico gene enrichment analysis based on GWAS summary data recognized the SNPs in ‘DNA methylation’ as the most significantly enriched genetic component, which acted as the potential regulator of oxidative DNA damage in AF. Therefore, the intervention of DNA methylation may improve oxidative DNA damage, thus slowing down the progression of LVA of the LA in AF, and might be a therapeutic target of atrial fibrosis in AF.

Acknowledgments

This work was funded by the National Natural Science Foundation of China [grant number: 82170322], the Special Foundation for Clinical Science and Technology of Jiangsu Province [grant number: BE2017754], and the Department of Science and Technology of Guangdong Province [grant number: 2019B020230004].

Disclosures

The authors declare that there are no conflicts of interest.

IRB Information

The Human Research Ethics Committee of the First Affiliated Hospital of Nanjing Medical University approved this study (reference no. 2022-SR-256).

Data Availability Statement

Due to the nature of this research, participants of this study did not agree for their data to be shared publicly, so supporting data is not available.

Supplementary Files

Please find supplementary file(s);

https://doi.org/10.1253/circj.CJ-22-0694

References

1. Watanabe R, Nagashima K, Wakamatsu Y, Otsuka N, Yokoyama K, Matsumoto N, et al. Different determinants of the recurrence of atrial fibrillation and adverse clinical events in the mid-term period after atrial fibrillation ablation. Circ J 2022; 86: 233–242.
2. Lin Y, Yang B, Garcia FC, Ju W, Zhang F, Chen H, et al. Comparison of left atrial electrophysiologic abnormalities during sinus rhythm in patients with different type of atrial fibrillation. J Interv Card Electrophysiol 2014; 39: 57–67.
3. Teh AW, Kistler PM, Lee G, Medi C, Heck PM, Spence SJ, et al. Electroanatomic remodeling of the left atrium in paroxysmal and persistent atrial fibrillation patients without structural heart disease. J Cardiovasc Electrophysiol 2012; 23: 232–238.
4. Takahashi Y, Yamaguchi T, Fukui A, Otsubo T, Hirota K, Kawano Y, et al. Impact of renal dysfunction on left atrial structural remodeling and recurrence after catheter ablation for atrial fibrillation: A propensity score matching analysis. Circ J 2020; 84: 1254–1260.
5. Yang B, Jiang C, Lin Y, Yang G, Chu H, Cai H, et al. STABLE-SR (Electrophysiological Substrate Ablation in the Left Atrium During Sinus Rhythm) for the treatment of nonparoxysmal atrial fibrillation: A prospective, multicenter randomized clinical trial. Circ Arrhythm Electrophysiol 2017; 10: e005405.
6. Zhang D, Hu X, Li J, Liu J, Baks-Te Bulte L, Wiersma M, et al. DNA damage-induced PARP1 activation confers cardiomyocyte dysfunction through NAD(+) depletion in experimental atrial fibrillation. Nat Commun 2019; 10: 1307.
7. Tascanov MB, Tanriverdi Z, Gungoren F, Besli F, Erkus ME, Altiparmak IH, et al. Relationships between paroxysmal atrial fibrillation, total oxidant status, and DNA damage. Rev Port Cardiol (Engl Ed) 2021; 40: 5–10.
8. Wiersma M, van Marion DMS, Bouman EJ, Li J, Zhang D, Ramos KS, et al. Cell-free circulating mitochondrial DNA: A potential blood-based marker for atrial fibrillation. Cells 2020; 9: 1159.
9. Kroese LJ, Scheffer PG. 8-hydroxy-2′-deoxyguanosine and cardiovascular disease: A systematic review. Curr Atheroscler Rep 2014; 16: 452.
10. Wu LL, Chiou CC, Chang PY, Wu JT. Urinary 8-OHdG: A marker of oxidative stress to DNA and a risk factor for cancer, atherosclerosis and diabetics. Clin Chim Acta 2004; 339: 1–9.
11. Toyama K, Yamabe H, Uemura T, Nagayoshi Y, Morihisa K, Koyama J, et al. Analysis of oxidative stress expressed by urinary level of 8-hydroxy-2′-deoxyguanosine and biopyrrin in atrial fibrillation: Effect of sinus rhythm restoration. Int J Cardiol 2013; 168: 80–85.
12. Li J, Zhang D, Ramos KS, Baks L, Wiersma M, Lanters EAH, et al. Blood-based 8-hydroxy-2′-deoxyguanosine level: A potential diagnostic biomarker for atrial fibrillation. Heart Rhythm 2021; 18: 271–277.
13. Goncalves AC, Alves R, Baldeiras I, Cortesao E, Carda JP, Branco CC, et al. Genetic variants involved in oxidative stress, base excision repair, DNA methylation, and folate metabolism pathways influence myeloid neoplasias susceptibility and prognosis. Mol Carcinog 2017; 56: 130–148.
14. Yamaguchi T, Tsuchiya T, Nakahara S, Fukui A, Nagamoto Y, Murotani K, et al. Efficacy of left atrial voltage-based catheter ablation of persistent atrial fibrillation. J Cardiovasc Electrophysiol 2016; 27: 1055–1063.
15. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, et al. PLINK: A tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet 2007; 81: 559–575.
16. Yang J, Lee SH, Goddard ME, Visscher PM. GCTA: A tool for genome-wide complex trait analysis. Am J Hum Genet 2011; 88: 76–82.
17. Lu HF, Chou PH, Lin GH, Chou WH, Wang ST, Adikusuma W, et al. Pharmacogenomics study for raloxifene in postmenopausal female with osteoporosis. Dis Markers 2020; 2020: 8855423.
18. de Leeuw CA, Mooij JM, Heskes T, Posthuma D. MAGMA: Generalized gene-set analysis of GWAS data. PLoS Comput Biol 2015; 11: e1004219.
19. Tsuboi M, Hisatome I, Morisaki T, Tanaka M, Tomikura Y, Takeda S, et al. Mitochondrial DNA deletion associated with the reduction of adenine nucleotides in human atrium and atrial fibrillation. Eur J Clin Invest 2001; 31: 489–496.
20. Lai LP, Tsai CC, Su MJ, Lin JL, Chen YS, Tseng YZ, et al. Atrial fibrillation is associated with accumulation of aging-related common type mitochondrial DNA deletion mutation in human atrial tissue. Chest 2003; 123: 539–544.
21. Lin PH, Lee SH, Su CP, Wei YH. Oxidative damage to mitochondrial DNA in atrial muscle of patients with atrial fibrillation. Free Radic Biol Med 2003; 35: 1310–1318.
22. Goette A, Kalman JM, Aguinaga L, Akar J, Cabrera JA, Chen SA, et al. EHRA/HRS/APHRS/SOLAECE expert consensus on atrial cardiomyopathies: Definition, characterization, and clinical implication. Europace 2016; 18: 1455–1490.
23. Goette A, Lendeckel U. Atrial cardiomyopathy: Pathophysiology and clinical consequences. Cells 2021; 10: 2605.
24. Cui C, Jiang X, Ju W, Wang J, Wang D, Sun Z et al. Atrial remodeling and metabolic dysfunction in idiopathic isolated fibrotic atrial cardiomyopathy. Int J Cardiol 2018; 265: 155–161.
25. Wang J, Wang Q, Watson LJ, Jones SP, Epstein PN. Cardiac overexpression of 8-oxoguanine DNA glycosylase 1 protects mitochondrial DNA and reduces cardiac fibrosis following transaortic constriction. Am J Physiol Heart Circ Physiol 2011; 301: H2073–H2080.
26. Frustaci A, Russo MA, Morgante E, Scopelliti F, Aquilano K, Ciriolo MR, et al. Oxidative myocardial damage in human cocaine-related cardiomyopathy. Eur J Heart Fail 2015; 17: 283–290.
27. Potpara TS, Lip GYH, Blomstrom-Lundqvist C, Boriani G, Van Gelder IC, Heidbuchel H, et al. The 4S-AF scheme (Stroke Risk; Symptoms; Severity of Burden; Substrate): A novel approach to in-depth characterization (rather than classification) of atrial fibrillation. Thromb Haemost 2021; 121: 270–278.
28. Gurses KM, Yalcin MU, Kocyigit D, Kesikli SA, Canpolat U, Yorgun H, et al. M2-muscarinic acetylcholine receptor autoantibody levels predict left atrial fibrosis severity in paroxysmal lone atrial fibrillation patients undergoing cryoablation. Europace 2015; 17: 239–246.
29. Ermakov AV, Konkova MS, Kostyuk SV, Izevskaya VL, Baranova A, Veiko NN. Oxidized extracellular DNA as a stress signal in human cells. Oxid Med Cell Longev 2013; 2013: 649747.
30. Vlachos K, Efremidis M, Letsas KP, Bazoukis G, Martin R, Kalafateli M, et al. Low-voltage areas detected by high-density electroanatomical mapping predict recurrence after ablation for paroxysmal atrial fibrillation. J Cardiovasc Electrophysiol 2017; 28: 1393–1402.
31. Nakatani Y, Sakamoto T, Yamaguchi Y, Tsujino Y, Kataoka N, Kinugawa K. P-wave vector magnitude predicts the left atrial low-voltage area in patients with paroxysmal atrial fibrillation. J Electrocardiol 2020; 59: 35–40.
32. Akoum N, Mahnkopf C, Kholmovski EG, Brachmann J, Marrouche NF. Age and sex differences in atrial fibrosis among patients with atrial fibrillation. Europace 2018; 20: 1086–1092.
33. Kottkamp H. Human atrial fibrillation substrate: Towards a specific fibrotic atrial cardiomyopathy. Eur Heart J 2013; 34: 2731–2738.
34. Jessup JA, Zhang L, Chen AF, Presley TD, Kim-Shapiro DB, Chappell MC, et al. Neuronal nitric oxide synthase inhibition improves diastolic function and reduces oxidative stress in ovariectomized mRen2.Lewis rats. Menopause 2011; 18: 698–708.
35. Groban L, Yamaleyeva LM, Westwood BM, Houle TT, Lin M, Kitzman DW, et al. Progressive diastolic dysfunction in the female mRen(2). Lewis rat: Influence of salt and ovarian hormones. J Gerontol A Biol Sci Med Sci 2008; 63: 3–11.
36. Liu H, Wang K, Lin Y, Liang X, Zhao S, Li M, et al. Role of sST2 in predicting recurrence of atrial fibrillation after radiofrequency catheter ablation. Pacing Clin Electrophysiol 2020; 43: 1235–1241.
37. Wang Y, Hao Y, Zhang H, Xu L, Ding N, Wang R, et al. DNA Hypomethylation of miR-30a mediated the protection of hypoxia postconditioning against aged cardiomyocytes hypoxia/reoxygenation injury through inhibiting autophagy. Circ J 2020; 84: 616–625.
38. Alex AM, Saradalekshmi KR, Shilen N, Suresh PA, Banerjee M. Genetic association of DNMT variants can play a critical role in defining the methylation patterns in autism. IUBMB Life 2019; 71: 901–907.
39. Turk PW, Laayoun A, Smith SS, Weitzman SA. DNA adduct 8-hydroxyl-2′-deoxyguanosine (8-hydroxyguanine) affects function of human DNA methyltransferase. Carcinogenesis 1995; 16: 1253–1255.
40. Branco MR, Ficz G, Reik W. Uncovering the role of 5-hydroxymethylcytosine in the epigenome. Nat Rev Genet 2011; 13: 7–13.
41. Kietzmann T, Petry A, Shvetsova A, Gerhold JM, Gorlach A. The epigenetic landscape related to reactive oxygen species formation in the cardiovascular system. Br J Pharmacol 2017; 174: 1533–1554.
42. Watson CJ, Collier P, Tea I, Neary R, Watson JA, Robinson C, et al. Hypoxia-induced epigenetic modifications are associated with cardiac tissue fibrosis and the development of a myofibroblast-like phenotype. Hum Mol Genet 2014; 23: 2176–2188.
43. Lim SO, Gu JM, Kim MS, Kim HS, Park YN, Park CK, et al. Epigenetic changes induced by reactive oxygen species in hepatocellular carcinoma: Methylation of the E-cadherin promoter. Gastroenterology 2008; 135: 2128–2140, 2140.e1–e8.
44. Coulter JB, O’Driscoll CM, Bressler JP. Hydroquinone increases 5-hydroxymethylcytosine formation through ten eleven translocation 1 (TET1) 5-methylcytosine dioxygenase. J Biol Chem 2013; 288: 28792–28800.
45. Laukkanen MO, Mannermaa S, Hiltunen MO, Aittomaki S, Airenne K, Janne J, et al. Local hypomethylation in atherosclerosis found in rabbit ec-sod gene. Arterioscler Thromb Vasc Biol 1999; 19: 2171–2178.
46. Bonnet S, Michelakis ED, Porter CJ, Andrade-Navarro MA, Thebaud B, Bonnet S, et al. An abnormal mitochondrial-hypoxia inducible factor-1alpha-Kv channel pathway disrupts oxygen sensing and triggers pulmonary arterial hypertension in fawn hooded rats: Similarities to human pulmonary arterial hypertension. Circulation 2006; 113: 2630–2641.
47. Paneni F, Mocharla P, Akhmedov A, Costantino S, Osto E, Volpe M, et al. Gene silencing of the mitochondrial adaptor p66(Shc) suppresses vascular hyperglycemic memory in diabetes. Circ Res 2012; 111: 278–289.
48. Assies J, Mocking RJ, Lok A, Ruhe HG, Pouwer F, Schene AH. Effects of oxidative stress on fatty acid- and one-carbon-metabolism in psychiatric and cardiovascular disease comorbidity. Acta Psychiatr Scand 2014; 130: 163–180.
49. Sakai C, Ishida M, Ohba H, Yamashita H, Uchida H, Yoshizumi M, et al. Fish oil omega-3 polyunsaturated fatty acids attenuate oxidative stress-induced DNA damage in vascular endothelial cells. PLoS One 2017; 12: e0187934.
50. Vijaya Lakshmi SV, Naushad SM, Seshagiri Rao D, Kutala VK. Oxidative stress is associated with genetic polymorphisms in one-carbon metabolism in coronary artery disease. Cell Biochem Biophys 2013; 67: 353–361.
51. Vucenik I, Shamsuddin AM. Protection against cancer by dietary IP6 and inositol. Nutr Cancer 2006; 55: 109–125.

Corresponding author

Register with J-STAGE for free!