Title: Associations of B‐Type Natriuretic Peptide and Its Coding Gene Promoter Methylation With Functional Outcome of Acute Ischemic Stroke: A Mediation Analysis
Abstract: HomeJournal of the American Heart AssociationVol. 9, No. 18Associations of B‐Type Natriuretic Peptide and Its Coding Gene Promoter Methylation With Functional Outcome of Acute Ischemic Stroke: A Mediation Analysis Open AccessResearch ArticlePDF/EPUBAboutView PDFView EPUBSections ToolsAdd to favoritesDownload citationsTrack citations ShareShare onFacebookTwitterLinked InMendeleyRedditDiggEmail Jump toOpen AccessResearch ArticlePDF/EPUBAssociations of B‐Type Natriuretic Peptide and Its Coding Gene Promoter Methylation With Functional Outcome of Acute Ischemic Stroke: A Mediation Analysis Aili Wang, MD, PhD, Mingzhi Zhang, MD, PhD, Yi Ding, MD, Xingbo Mo, MD, PhD, Chongke Zhong, MD, PhD, Zhengbao Zhu, MD, Daoxia Guo, MD, Xiaowei Zheng, MD, Tan Xu, MD, PhD, Yan Liu, MD, Yonghong Zhang, and MD, PhD, and Hao PengMD, PhD Aili WangAili Wang , Department of Epidemiology, , School of Public Health and Jiangsu Key Laboratory of Preventive and Translational Medicine for Geriatric Diseases, , Medical College of Soochow University, , Suzhou, , China, , Mingzhi ZhangMingzhi Zhang * Correspondence to: Hao Peng, MD, PhD, Department of Epidemiology, School of Public Health and Jiangsu Key Laboratory of Preventive and Translational Medicine for Geriatric Diseases, Medical College of Soochow University, 199 Renai Road, Industrial Park District, Suzhou, China 215123. E-mail: E-mail Address: [email protected] and Mingzhi Zhang, MD, PhD, Department of Biostatistics, School of Public Health, Medical College of Soochow University, 199 Renai Road, Industrial Park, Suzhou, China 215123. E-mail: E-mail Address: [email protected] , Department of Epidemiology, , School of Public Health and Jiangsu Key Laboratory of Preventive and Translational Medicine for Geriatric Diseases, , Medical College of Soochow University, , Suzhou, , China, , Yi DingYi Ding , Department of Epidemiology, , School of Public Health and Jiangsu Key Laboratory of Preventive and Translational Medicine for Geriatric Diseases, , Medical College of Soochow University, , Suzhou, , China, , Xingbo MoXingbo Mo , Department of Epidemiology, , School of Public Health and Jiangsu Key Laboratory of Preventive and Translational Medicine for Geriatric Diseases, , Medical College of Soochow University, , Suzhou, , China, , Chongke ZhongChongke Zhong https://orcid.org/0000-0003-2314-6814 , Department of Epidemiology, , School of Public Health and Jiangsu Key Laboratory of Preventive and Translational Medicine for Geriatric Diseases, , Medical College of Soochow University, , Suzhou, , China, , Zhengbao ZhuZhengbao Zhu https://orcid.org/0000-0003-4795-535X , Department of Epidemiology, , School of Public Health and Jiangsu Key Laboratory of Preventive and Translational Medicine for Geriatric Diseases, , Medical College of Soochow University, , Suzhou, , China, , Daoxia GuoDaoxia Guo , Department of Epidemiology, , School of Public Health and Jiangsu Key Laboratory of Preventive and Translational Medicine for Geriatric Diseases, , Medical College of Soochow University, , Suzhou, , China, , Xiaowei ZhengXiaowei Zheng , Department of Epidemiology, , School of Public Health and Jiangsu Key Laboratory of Preventive and Translational Medicine for Geriatric Diseases, , Medical College of Soochow University, , Suzhou, , China, , Tan XuTan Xu , Department of Epidemiology, , School of Public Health and Jiangsu Key Laboratory of Preventive and Translational Medicine for Geriatric Diseases, , Medical College of Soochow University, , Suzhou, , China, , Yan LiuYan Liu , Genesky Biotechnologies Inc, , Shanghai, , China, , Yonghong ZhangYonghong Zhang , Department of Epidemiology, , School of Public Health and Jiangsu Key Laboratory of Preventive and Translational Medicine for Geriatric Diseases, , Medical College of Soochow University, , Suzhou, , China, , and Hao PengHao Peng * Correspondence to: Hao Peng, MD, PhD, Department of Epidemiology, School of Public Health and Jiangsu Key Laboratory of Preventive and Translational Medicine for Geriatric Diseases, Medical College of Soochow University, 199 Renai Road, Industrial Park District, Suzhou, China 215123. E-mail: E-mail Address: [email protected] and Mingzhi Zhang, MD, PhD, Department of Biostatistics, School of Public Health, Medical College of Soochow University, 199 Renai Road, Industrial Park, Suzhou, China 215123. E-mail: E-mail Address: [email protected] https://orcid.org/0000-0001-6277-662X , Department of Epidemiology, , School of Public Health and Jiangsu Key Laboratory of Preventive and Translational Medicine for Geriatric Diseases, , Medical College of Soochow University, , Suzhou, , China, Originally published2 Sep 2020https://doi.org/10.1161/JAHA.120.017499Journal of the American Heart Association. 2020;9:e017499AbstractBackgroundThe prognostic role of B‐type natriuretic peptide (BNP) in stroke has been suggested, but limited studies have shown mixed results and unknown underlying mechanisms. DNA methylation, a molecular modification that alters gene expression, may represent a candidate mechanism for this purpose. We aimed to examine the associations of BNP and methylation of its coding gene (natriuretic peptide B [NPPB]) with the functional outcome in a large sample of patients with acute ischemic stroke from CATIS (China Antihypertensive Trial in Acute Ischemic Stroke).Methods and ResultsLeveraging participants from CATIS with available specimens, serum proBNP (equimolarly produced with BNP) was measured in 3216 patients (mean age, 62 years; 64% men), and peripheral blood DNA methylation of the NPPB promoter was quantified by targeted bisulfite sequencing in 806 patients (mean age, 62 years; 54% men). The functional outcome was defined as an ordered modified Rankin Scale score assessed at 14 days or hospital discharge after stroke onset. Mediation analysis was conducted to test the potential mediating effect of proBNP on the relationship between NPPB methylation and functional outcome. The results showed that a higher level of proBNP was significantly associated with a higher risk of having a poorer functional outcome (odds ratio [OR], 1.14; P=0.006). Every 5% of hypermethylation at 2 (Chr1:11919160 [OR, 0.93; P=0.022] and Chr1:11918989 [OR, 0.92; P=0.032]) of 11 CpG loci assayed was associated with 7% and 8% lower risk, respectively, of having a poor functional outcome. In addition, proBNP was negatively correlated to hypermethylation at 1 CpG (Chr1:11918989 [β=−0.029; P=0.009]) and mediated approximately 7.69% (95% CI, 2.50%–13.82%) of the association between this CpG methylation and the functional outcome.ConclusionsHypermethylation at the NPPB promoter is associated with the functional outcome after ischemic stroke, at least partially by suppressing BNP expression or excretion.Nonstandard Abbreviations and AcronymsAISacute ischemic strokeBNPB‐type natriuretic peptideCATISChina Antihypertensive Trial in Acute Ischemic StrokeKEGGKyoto Encyclopedia of Genes and GenomesmRSmodified Rankin ScaleNIHSSNational Institutes of Health Stroke ScaleNPPBnatriuretic peptide BClinical PerspectiveWhat Is New?Hypermethylation at the natriuretic peptide B (NPPB) promoter is associated with functional outcome after ischemic stroke.Hypermethylation at the NPPB promoter is associated with lower circulating levels of pro–B‐type natriuretic peptide.What Are the Clinical Implications?Hypermethylation at the NPPB promoter may be a marker of poor functional outcome after ischemic stroke.Hypermethylation at the NPPB promoter may be potential drug target for ischemic stroke.B‐type natriuretic peptide (BNP) is a circulating cardiac hormone synthesized and released as a precursor protein, mainly from the ventricles, in response to increased myocardial wall stress related to volume or pressure overload.1 It has been globally endorsed in clinical guidelines as a biomarker to aid in the diagnosis of heart failure and to monitor disease progression.2, 3, 4 In addition to heart diseases, BNP was elevated in patients with acute ischemic stroke (AIS)5, 6 and associated with the poor functional outcome of AIS,7, 8, 9, 10, 11 although not consistently.12, 13 Its expression in the brain implies a possible role for BNP in neurologic function.14 Administration of BNP has been demonstrated to improve cerebral blood flow and to reduce inflammation in brain injury models of mice, as manifested by reduced neurodegeneration and improved functional outcome.15 BNP seems to be implicated in neuroprotection following brain injury from AIS, but the underlying molecular mechanisms are not clear.Genetic polymorphisms in the coding gene of BNP (natriuretic peptide B [NPPB]) have been associated not only with blood BNP concentrations16, 17 but also with determinants of AIS pathogenesis and prognosis, for example, hypertension18, 19 and diabetes mellitus.20 As an interface between the fixed genome and dynamic environment, epigenetic factors such as DNA methylation in the NPPB gene may affect its function and subsequent BNP synthesis and excretion. DNA methylation status can change dramatically, and dysregulated DNA methylation has previously been related to poor tissue outcome after cerebral ischemic injury in mice.21, 22 However, the epigenetic markers of functional outcome of AIS, including DNA methylation of the NPPB gene, have scarcely been studied in humans.The objectives of this study were to examine whether BNP and its coding gene promoter methylation levels at the acute phase could predict functional outcome and whether BNP mediated the relationship between NPPB promoter methylation and functional outcome in patients with AIS, using data from CATIS (China Antihypertensive Trial in Acute Ischemic Stroke).23MethodsThe data that support the findings of this study are available from the corresponding author on reasonable request.Study PatientsCATIS is a multicenter randomized clinical trial (ClinicalTrials.gov identifier NCT01840072) designed to test whether moderate lowering of blood pressure in the acute phase after the onset of an AIS reduces the risks of death and major disability at 14 days or hospital discharge. The study design, intervention, and survey methods of CATIS have been described previously.23 In brief, 4071 patients with first‐ever ischemic stroke were recruited into CATIS. All surviving patients were reexamined at 14 days or hospital discharge after stroke onset. The study protocols were approved by the institutional review boards at Soochow University in China and Tulane University in the United States and by the ethics committees of all participating hospitals. Written informed consent was obtained from all study participants or their immediate family members.Figure 1 describes the selection of study participants in the current analysis. After excluding 855 patients because of lack of data on proBNP, 3216 patients were included in the analysis of the association between proBNP and functional outcome of AIS. In total, 806 patients with eligible DNA samples for DNA methylation quantification were included in the analysis of the association between NPPB promoter methylation and functional outcome of AIS. Of those, 704 patients with available data on both proBNP and DNA methylation were included in the analysis of the mediating effect of proBNP on the association between NPPB promoter methylation and functional outcome of AIS.Download figureDownload PowerPointFigure 1. Flowchart of selection of study participants.BNP indicates B‐type natriuretic peptide; CATIS, China Antihypertensive Trial in Acute Ischemic Stroke; and NPPB, natriuretic peptide B.Measurements of proBNPDespite being equimolarly cleaved from its precursor protein (1–108 amino acids), the biologically inert proBNP (1–76 amino acids; ie, N‐terminal proBNP) has a longer half‐life and is more stable in the circulation than BNP (77–108 amino acids).24 Therefore, we used proBNP concentrations to approximately reflect BNP excretion in this study. Serum proBNP concentrations were measured using ELISA (Biomedica Medizinprodukte), according to the manufacturer's guidelines, in fasting blood samples drawn within 24 hours of hospital admission. Intra‐ and interassay coefficients of variation were <4% and <3%, respectively. All samples were processed at the Jiangsu Key Laboratory of Preventive and Translational Medicine for Geriatric Diseases in a duplicate assay by laboratory technicians blinded to the clinical characteristics of the study participants.Quantification of NPPB Promoter MethylationDNA methylation levels in the promoter region of NPPB were quantified by targeted bisulfite sequencing, as described previously,25 using genomic DNA isolated from peripheral blood mononuclear cells. Briefly, based on the genomic coordinates of the NPPB promoter in Genome Reference Consortium Human Build 37, we carefully designed the primers to detect the maximum CpG loci within the CpG islands. The targeted sequence (Chr1:11918953–11919190, reverse strand) and primers for sequencing are schematically illustrated in Figure 2. Following primer validation, genomic DNA was treated with bisulfite using the EZ DNA Methylation‐Gold Kit (Zymo Research), according to the manufacturer's protocol, which converts unmethylated cytosine into uracil and leaves methylated cytosine unchanged. The treated samples were amplified, bar‐coded, and sequenced by Illumina Hiseq 2000 (Illumina) using the paired‐end sequencing protocol, according to the manufacturer's guidelines. Methylation level at each CpG locus was calculated as the percentage of the methylated alleles over the sum of methylated and unmethylated alleles. For quality control, the samples with a bisulfite conversion rate <98% and the cytosine sites with average coverage <20× were filtered out.Download figureDownload PowerPointFigure 2. Schematic illustration of the targeted sequence and primers for targeted bisulfite sequencing.Red represents the CpG loci assayed in the NPPB gene promoter (−198 to +39 bp from TSS). NPPB indicates natriuretic peptide B; and TSS, transcriptional start siteAssessment of Functional Outcome of AISFor patients with AIS, the modified Rankin Scale (mRS) was recommended to assess functional outcome after stroke.26 It ranged from 0 to 6, with 0 indicating no symptoms, 5 indicating severe disability (ie, bedridden, incontinent, or requiring constant nursing care and attention), and 6 indicating death. The mRS evaluation at 14 days after stroke onset or at hospital discharge was administered by trained neurologic physicians blinded to baseline characteristics. An ordered 7‐level categorical score of the mRS at 14 days or hospital discharge was the functional outcome of AIS in this study.Assessment of Potential Covariates at AdmissionData on demographic characteristics (age, sex), lifestyle risk factors (current smoking, current drinking), and medical history (coronary heart disease, hypertension, diabetes mellitus, hyperlipidemia) were collected at admission using a standard questionnaire. Fasting glucose and lipids at admission were examined at every participating hospital. Ischemic stroke was classified as thrombotic, embolic, and lacunar subtypes according to the symptoms and imaging data.27 Stroke severity at baseline was evaluated by trained neurologists using the National Institutes of Health Stroke Scale (NIHSS). Trained research staff measured participants' body weight (in kilograms) and height (in centimeters) with participants wearing light clothes and no shoes. Body mass index was calculated by dividing weight in kilograms by the square of height in meters (kg/m2). Three blood pressure measurements were performed at admission by trained nurses, according to a common protocol adapted from procedures recommended by the American Heart Association.28Statistical AnalysisTo carefully evaluate the roles of BNP and its coding gene methylation in the neurologic recovery of AIS, we examined the associations of proBNP and NPPB promoter methylation at admission with functional outcome, followed by a mediation analysis among them. Log10 transformation was applied to maximize normal distribution of proBNP, and the generated values (log10‐proBNP) were used in downstream analyses. All statistical analyses were performed using SAS statistical software (v9.4; SAS Institute).Association AnalysisTo examine the association between proBNP at admission and functional outcome of AIS, we constructed an ordered logistic regression model in which mRS score was the dependent variable and proBNP (continuous log10‐proBNP or categorical quartiles) was the independent variable, adjusting for potential covariates at admission including age, sex, stroke subtype, NIHSS score, hours from onset to hospitalization, current smoking, current drinking, body mass index, systolic blood pressure, disease history (hypertension, diabetes mellitus, dyslipidemia, coronary heart disease), and treatment group. The model fit was assessed using a Hosmer–Lemeshow goodness‐of‐fit test. To examine the association between NPPB promoter methylation at admission and functional outcome of AIS, we constructed similar ordered logistic regression models with DNA methylation at each CpG locus as the independent variable. Multiple testing was controlled by adjusting for the total number of CpG loci tested, and a false discovery rate–adjusted P value (ie, q value) of <0.2 was considered nominally significant.Mediation AnalysisThis analysis focuses on CpG sites at which DNA methylation showed nominally significant associations with both proBNP and functional outcome of AIS. To test whether proBNP mediates the association between NPPB promoter methylation and functional outcome, we constructed a causal mediation model by fitting a series of conditional regression models. Specifically, we tested the following conditions: The relationship between DNA methylation (X) and mRS (Y) (ordered logistic regression model Y=βTotX, βTot: total effect);The relationship between DNA methylation (X) and log10‐proBNP (M) (quantile regression model M=β1X);The relationship between log10‐proBNP (M) and mRS (Y) after controlling for DNA methylation (X) (Y=β2M+βDirX,βDir: direct effect);We then calculated the mediating effect (βMed=β1×β2) and the proportion of mediation (βMed/βTot×100%). Mediation analysis was performed using R package "mediation,"29 adjusted for the covariates listed above. The 95% CI of the mediating effect was estimated by Monte Carlo CIs.Secondary AnalysisBased on the associations of each single CpG methylation with functional outcome, we tested the joint association of DNA methylation at multiple CpG probes in the NPPB gene with functional outcome using the weighted truncated product method, as described previously.30 This method combines P values of all CpGs that reach a preselected threshold (eg, raw P<0.1 in this study). The regression coefficient of each individual CpG methylation was included as weight in the weighted truncated product method statistic. Gene Ontology function and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were conducted for the NPPB gene by DAVID online analysis (https://david.ncifcrf.gov/). The regulatory network involving the NPPB gene was also constructed by GeneMANIA (http://genemania.org/).ResultsBaseline CharacteristicsOf the 4071 patients with AIS in CATIS, 3216 patients (mean age, 62 years; 64% male) with available proBNP and 806 patients (mean age, 62 years; 54% male) with measurements of NPPB promoter methylation were included in the current study. Their clinical characteristics at admission are shown in Table 1. In both subsets of study participants, approximately 79%, 18%, 7%, and 10% of patients had hypertension, diabetes mellitus, hyperlipidemia, and coronary heart disease, respectively.Table 1. Clinical Characteristics of Acute Ischemic Stroke Patients at AdmissionCharacteristicsPatients With Available proBNPPatients With Available NPPB MethylationNo. of participants3216806Age, y62.5±10.862.6±12.2Sex, male (%)2048 (63.68)432 (53.60)Current smoking, n (%)1187 (36.91)278 (34.49)Current drinking, n (%)986 (30.66)211 (26.18)Disease history, n (%)Hypertension2541 (79.01)629 (78.04)Diabetes mellitus562 (17.48)153 (18.98)Hyperlipidemia226 (7.03)54 (6.70)Coronary heart disease341 (10.60)84 (10.42)Ischemic stroke subtype, n (%)Thrombotic2448 (76.12)766 (95.04)Embolic165 (5.13)40 (4.96)Lacunar686 (21.33)…Body mass index, kg/m224.90±3.1325.12±3.39Systolic blood pressure, mm Hg166.5±17.0168.3±16.8Diastolic blood pressure, mm Hg96.6±11.097.0±10.7Fasting glucose, mmol/L6.69±2.736.79±2.84Total cholesterol, mmol/L5.08±1.155.12±1.17Triglycerides, mmol/L1.84±2.821.92±5.10LDL cholesterol, mmol/L2.95±0.952.94±1.00HDL cholesterol, mmol/L1.30±0.451.30±0.41Hours from onset to hospitalization15.10±13.0013.95±12.93NIHSS score, points5.8±4.87.1±5.2proBNP, pg/mL, median (IQR)142.4 (67.1–327.3)156.9 (71.4–374.5)Log10‐transformed proBNP2.17±0.552.20±0.57John Wiley & Sons, LtdAll results are expressed as mean±SD unless otherwise noted. BNP indicates B‐type natriuretic peptide; HDL, high‐density lipoprotein; IQR, interquartile range; LDL, low‐density lipoprotein; NIHSS, National Institutes of Health Stroke Scale; and NPPB, natriuretic peptide B.Association Between proBNP at Admission and the Functional Outcome of AISAbout 58% of the 3216 patients with available proBNP experienced an unfavorable functional outcome including disability that was slight (n=773, 24.04%), moderate (n=462, 14.37%), moderately severe (n=437, 13.59%), or severe (n=156, 4.85%) or death (n=30, 0.93%) at 14 days or hospital discharge (Table 2). Table 3 shows the association between proBNP at admission and the functional outcome. Ordered logistic regression using log10‐proBNP as the independent variable showed that a higher level of proBNP at admission was significantly associated with a higher risk of having a higher mRS score at 14 days or hospital discharge (odds ratio [OR], 1.14; P=0.006), independent of age, sex, stroke subtype, NIHSS score, hours from onset to hospitalization, current smoking, current drinking, body mass index, blood pressure, disease history (hypertension, diabetes mellitus, dyslipidemia, coronary heart disease), and treatment group at admission. Regression using categorical proBNP quartiles as the independent variable revealed a similar result with the same direction. Compared with patients with the lowest level of proBNP, those with the highest level of proBNP had a 32% higher risk of having a higher mRS score (OR, 1.32; P=0.004).Table 2. Distribution of Functional Outcome at 14 days or Hospital Discharge After Stroke OnsetmRS ScorePatients With Available proBNP (n=3216)Patients With Available NPPB Methylation (n=806)n%n%0 (No symptoms)2888.96435.331 (No significant disability despite symptoms)107033.2721526.672 (Slight disability)77324.0420325.193 (Moderate disability)46214.3712315.264 (Moderately severe disability)43713.5914317.745 (Severe disability)1564.85678.316 (Death)300.93121.49John Wiley & Sons, LtdBNP indicates B‐type natriuretic peptide; and NPPB indicates natriuretic peptide B.Table 3. Association of proBNP at Admission With the Functional Outcome at 14 Days or Hospital Discharge After Stroke Onset (n=3216)proBNP, pg/mLUnadjustedAdjusted†OR (95% CI)*P ValueOR (95% CI)*P ValueLog10‐transformed proBNP1.18 (1.05–1.32)0.0041.14 (1.01–1.33)0.006Quartile 1 (≤67.1)1.00 (reference)—1.00 (reference)—Quartile 2 (67.2–142.4)1.16 (0.97–1.38)0.0931.16 (0.95–1.40)0.118Quartile 3 (142.5–327.2)1.28 (1.07–1.52)0.0061.30 (1.06–1.54)0.004Quartile 4 (≥327.3)1.32 (1.11–1.57)0.0021.32 (1.11–1.59)0.004John Wiley & Sons, LtdBNP indicates B‐type natriuretic peptide; and OR, odds ratio.*Odds of having a higher mRS score at 14 d or hospital discharge.†Adjusted for age, sex, stroke subtype, National Institutes of Health Stroke Scale score, hours from onset to hospitalization, current smoking, current drinking, body mass index, systolic blood pressure, disease history (hypertension, diabetes mellitus, dyslipidemia, coronary heart disease), and treatment group at admission. A Hosmer–Lemeshow goodness‐of‐fit test was applied to examine the model fit (χ2=68.30, P=0.077 for continuous log10‐transformed proBNP).Association Between NPPB Promoter Methylation and the Functional Outcome of AISAbout 68% of the 806 patients with available NPPB methylation data experienced an unfavorable functional outcome including disability that was slight (n=203, 25.19%), moderate (n=123, 15.26%), moderately severe (n=143, 17.74%), or severe disability (n=67, 8.31%) or death (n=12, 1.49%) at 14 days or hospital discharge (Table 2). DNA methylation levels at almost all CpG loci seemed to decrease with increased mRS score, but DNA methylation at only 2 CpG sites (CpG1 located in Chr1:11919160 [β=−0.524, P=0.027] and CpG11 located in Chr1:11918989 [β= −0.606, P=0.032]) preserved a statistically significant trend (Figure 3). The ordered logistic regression using mRS score as the dependent variable revealed similar results (Table 4). After multivariate adjustment, almost all CpGs methylation presented a negative association with odds of having a higher mRS score. Of these, every 5% of hypermethylation at the same CpG sites (CpG1: OR, 0.93 [P=0.022]; CpG11: OR, 0.92 [P=0.032]) was associated with 7% and 8% lower risk of having a higher mRS score, which indicates poorer functional outcome at 14 days or hospital discharge. These associations persisted as nominally significant after adjusting for multiple testing (all false discovery rate–adjusted P<0.2).Download figureDownload PowerPointFigure 3. Heat map illustrating the average methylation levels at all CpG loci in the NPPB promoter according to mRS score at 14 days or hospital discharge.Color depth shows the value of the corresponding methylation level. BNP, B‐type natriuretic peptide; mRS indicates the modified Rankin Scale; and NPPB, natriuretic peptide B.Table 4. Association of CpG Methylation in NPPB Promoter With the Functional Outcome at 14 days or Hospital Discharge After Stroke Onset (n=806)CpG LociGenomic Position (GRCh37)Relative to TSS, bpMethylation Level, %UnadjustedAdjusted†OR (95% CI)*P ValueOR (95% CI)*P ValueCpG1Chr1:11919160−16826.25±9.800.93 (0.87–0.99)‡0.027‡0.93 (0.86–0.98)‡0.022‡CpG2Chr1:11919144−15228.15±9.940.99 (0.93–1.05)0.7880.98 (0.93–1.07)0.684CpG3Chr1:11919141−14927.05±9.500.98 (0.92–1.04)0.4660.96 (0.90–1.05)0.439CpG4Chr1:11919135−14329.19±10.270.98 (0.92–1.04)0.4760.97 (0.92–1.04)0.330CpG5Chr1:11919133−14134.85±10.380.96 (0.91–1.02)0.2280.97 (0.92–1.03)0.126CpG6Chr1:11919117−12531.31±11.980.99 (0.94–1.04)0.7191.00 (0.96–1.08)0.848CpG7Chr1:11919096−10433.65±12.660.97 (0.92–1.02)0.2310.99 (0.93–1.06)0.747CpG8Chr1:11919047−5511.05±7.650.99 (0.91–1.07)0.7290.98 (0.90–1.05)0.851CpG9Chr1:11919019−277.21±5.431.05 (0.94–1.17)0.4131.05 (0.94–1.18)0.376CpG10Chr1:11919011−198.26±6.510.98 (0.89–1.08)0.6440.99 (0.89–1.08)0.787CpG11Chr1:11918989331.63±11.750.94 (0.89–0.99)‡0.023‡0.92 (0.88–0.99)‡0.032‡John Wiley & Sons, LtdGRCh37 indicates Genome Reference Consortium Human Build 37; NPPB, natriuretic peptide B; OR, odds ratio; and TSS, transcriptional start site.*Odds of having a higher modified Rankin Scale score at 14 d or hospital discharge associated with per5% increment of corresponding CpG methylation.†Adjusted for age, sex, stroke subtype, National Institutes of Health Stroke Scale score, hours from onset to hospitalization, current smoking, current drinking, body mass index, systolic blood pressure, disease history (hypertension, diabetes mellitus, dyslipidemia, coronary heart disease), and treatment group at admission. A Hosmer–Lemeshow goodness‐of‐fit test was applied to examine the model fit (χ2=88.28, P=0.002 for CpG1; χ2=51.46, P=0.534 for CpG11).‡Values in bold indicate a false discovery rate–adjusted P<0.2.Correlation Between NPPB Promoter Methylation and proBNPFigure 4 illustrates the correlation between log10‐proBNP and each CpG methylation in the promoter region of the NPPB gene. DNA methylation at almost all 11 CpG loci assayed appeared to be negatively correlated with proBNP, but only 1 CpG methylation reached statistical significance, with a correlation coefficient of −0.09 (CpG11, P=0.016). These results indicated that hypermethylation of the NPPB gene promoter may be associated with a lower level of proBNP.Download figureDownload PowerPointFigure 4. Spearman correlation matrix of DNA methylation of all CpG loci assayed in the NPPB promoter and proBNP (after log10‐transformation) at admission.The areas of circles and color depth show the absolute value of corresponding correlation coefficients. BNP indicates B‐type natriuretic peptide; and NPPB, natriuretic peptide B.Mediating Effect of proBNP on the Association Between NPPB Promoter Methylation and the Functional Outcome of AISOur mediation analysis focused on the only CpG locus (CpG11 located in Chr1:11918989) at which DNA methylation was identified as significantly associated with both proBNP and functional outcome. The mediating effect of proBNP on the association between this CpG methylation and functional outcome is schematically illustrated in Figure 5. We found that that proBNP accounted for ≈7.69% (95% CI, 2.50%–13.82%) of the association between DNA methylation at this CpG site and functional outcome.Download figureDownload PowerPointFigure 5. Schematic illustration of the mediating effect of proBNP on the association between DNA methylation at CpG11 (located in Chr:11918989) in the NPPB promoter and the functional outcome at 14 days or hospital discharge in 704 patients with ischemic stroke.In this model, DNA methylation at CpG11 (per 5% increment) was the exposure variable, proBNP (after log10‐transformation) was the mediator variable, and mRS was the outcome variable, adjusting for age, sex, stroke subtyp