Skip to main content
Download PDF
Download ePub

Longitudinal association of peripheral blood DNA methylation with liver fat content: distinguishing between predictors and biomarkers

Lipids in Health and Disease volume 23, Article number: 309 (2024) Cite this article

A Correction to this article was published on 17 December 2024

This article has been updated

Abstract

Background

Alterations in DNA methylation (DNAm) have been observed in patients with fatty liver, but whether they are cause or consequence remains unknown. The study aimed to investigate longitudinal association of epigenome-wide DNAm with liver fat content (LFC) in Chinese participants, and explore their temporal relationships.

Methods

Data were obtained from 2 waves over a four-year time period of the Shanghai Changfeng Study (discovery, n = 407 and replication, n = 126). LFC and peripheral blood DNAm were repeatedly measured using quantitative hepatic ultrasonography and the 850 K Illumina EPIC BeadChip, respectively. Longitudinal and cross-sectional epigenome-wide association studies (EWASs) were conducted with linear mixed model and linear regression model, respectively. Meta-analysis was performed using METAL. Cross-lagged panel analysis (CLPA) was carried out to infer temporal relationships between the significant CpGs and LFC.

Results

Longitudinal EWAS identified cg11024682 (SREBF1), cg06500161 (ABCG1), cg16740586 (ABCG1), cg15659943 (ABCA1) and cg00163198 (SNX19) significantly associated with LFC with P < 1e-7. Another 6 of the 22 previously reported CpGs were replicated in the present longitudinal EWAS. CLPA showed longitudinal effects of cg11024682 (SREBF1) (β = 0.14 [0.06, 0.23]), cg16740586 (ABCG1) (β = 0.17 [0.08, 0.25]), cg06500161 (ABCG1) (β = 0.12 [0.03, 0.20]), cg17901584 (DHCR24) (β = -0.10 [-0.18, -0.02]), cg00574958 (CPT1A) (β = -0.09 [-0.17, -0.01]), cg08309687 (LINC00649) (β = -0.11 [-0.19, -0.03]), and cg27243685 (ABCG1) (β = 0.09 [0.01, 0.18]) on subsequent LFC. The effects were attenuated when further adjusting for body mass index. High levels of LFC led to alterations in DNAm of cg15659943 (ABCA1) (β = 0.13 [0.04, 0.21]), cg07162647 (β = -0.11 [-0.19, -0.03]), cg06500161 (ABCG1) (β = 0.10 [0.02, 0.18]), and cg27243685 (ABCG1) (β = 0.10 [0.02, 0.18]).

Conclusions

Blood DNAm at SREBF1, ABCG1, DHCR24, CPT1A, and LINC00649 may be predictors of subsequent LFC change. The effects of DNAm at SREBF1 and ABCG1 on LFC were partially influenced by obesity. The findings have potential implications in understanding disease pathogenesis and highlight the potential of DNAm for early detection or intervention of fatty liver.

Background

The global prevalence of metabolic dysfunction-associated steatotic liver disease (MASLD) has increased 50.4% over the past 3 decades, and it is estimated to be 38.2% recently [1]. The incidence rate of MASLD was reported to be highest in Mainland China [2]. The presence of MASLD is influenced by geographical region, race, genomic variation and lifestyle factors [3]. For example, genetic variant rs58542926 C > T of TM6SF2 was reported associated with the full spectrum of MASLD but lower blood lipid levels [4]. Further studies revealed that TM6SF2 mainly localizes to the smooth endoplasmic reticulum, and prevents MASLD by promoting bulk lipidation of apolipoproteins B-containing lipoproteins and very low-density lipoprotein exportation from liver [5, 6].

Epigenetic regulations, especially DNA methylation (DNAm), can also explain part of the variance of MASLD. Using liver biopsies, epigenome-wide association study (EWAS) of MASLD reported that differences in DNAm can distinguish patients with advanced vs. mild MASLD [7, 8], and DNAm alterations associated with MASLD were partially reversible by bariatric surgery [9]. Nevertheless, liver biopsy is invasive, costly, and difficult to standardize and repeat. There have been growing interests in identifying biomarkers from non-invasive imaging techniques and peripheral blood to predict the onset and stratify the severity of MASLD. Using peripheral blood DNA and 450k BeadChip, the largest EWAS of MASLD enrolled 3400 European participants and identified a total of 22 CpGs associated with hepatic fat, such as CpGs annotated to ABCG1, SREBF1, SLC7A11 and so on [10]. An EWAS of peripheral leukocytes from 35 Chinese MASLD patients and 30 healthy controls revealed that hypomethylation of ACSL4 and CPT1C was associated with MASLD [11].

However, prior studies mainly focused on cross-sectional associations of DNAm with MASLD or hepatic fat, and were limited by smaller size of detected probes-mostly less than 450,000. Ma et al. investigated potential causal association between the CpGs and MASLD using Mendelian randomization (MR) analyses, and observed a causal relationship between hypomethylated cg08309687 (LINC00649) and increased hepatic fat [10]. The relatively small-scale GWAS used might have led to insufficient statistical power in their MR analyses. Thus, this research was designed to investigate longitudinal associations between DNAm at > 700,000 CpGs in peripheral blood and liver fat content (LFC) in a Chinese cohort, and to assess temporal directional relationships between DNAm of the significant CpGs and LFC using cross-lagged panel analysis. It is hypothesized that DNAm alterations of different CpGs might be driver or consequence of LFC change through various pathways.

Methods

Study population and design

The study population and design are presented in Fig. 1. Shanghai Changfeng Study focuses on the risk factors and management of chronic cardiometabolic traits among Chinese residents aged ≥ 45 years in Changfeng community, Shanghai, China [12]. A total of 6595 residents were enrolled at baseline between 2009 and 2012, and 3343 of them were revisited between 2014 and 2017. The discovery cohort included 407 participants with complete DNAm data, among whom 358 and 405 had LFC data at baseline and at follow-up, respectively. Another 126 participants with complete DNAm data were included in the replication cohort, among whom 97 and 125 had LFC data at baseline and at follow-up, respectively. In the longitudinal analyses, the discovery and the replication cohorts included 358 and 96 participants without missing data, respectively.

Fig. 1
figure 1

Study population and design

Full size image

Measurement of LFC and BMI

Participants received face-to-face interviews and provided basic information including age and sex. Current smokers were defined as smoking ≥ 1 cigarette/day for one or more years, and past smokers were defined as smokers that discontinued smoking for more than 6 months. LFC was measured by trained interviewers using a quantitative hepatic ultrasonography and calculated as LFC (%) = 62.592*standardized ultrasound hepatic/renal ratio + 168.076*standardized ultrasound hepatic attenuation rate-27.863 [13]. Body mass index (BMI) was obtained by dividing body weight by height squared (kg/m2).

Blood DNA methylation profiling, quality control, and data preprocessing

Genomic DNA of the study participants was extracted from peripheral blood using a DNeasy Blood and Tissue Kit (Qiagen, Hilden, Germany), followed by estimation of purity and concentration using NanoDrop One spectrophotometer and Qubit 3.0 Fluorometer (Thermo Scientific, Waltham, MA, USA). An EZ DNA Methylation Kit (Zymo Research, Irvine, CA, USA) was used to perform bisulfite conversion of DNA from each sample. Methylation profiles of DNA were measured with Illumina Infinium Methylation EPIC BeadChip (Illumina, San Diego, CA, USA).

Data quality control and preprocessing was conducted with the R package ChAMP separately in the discovery and the replication cohorts [14]. Probes were filtered out with the following criteria: detection P > 0.01, or > 3 beads in ≥ 5% samples, or non-CpG, or SNP-related, or multi-hit, or non-autosomal. Multidimensional Scaling plot, density plot, and dendrogram plot generated using the function champ.QC were used to further assess data quality. The final clean methylation working dataset contained 711,268 (discovery) and 740,701 (replication) autosomal CpG probes.

Beta value (β, 0–1 scale) was used to represent DNAm level of the CpGs, and normalization of the β value matrix was conducted with the BMIQ method [15]. Batch effects were estimated by using surrogate variable analysis, and were corrected by applying Combat [16] in ChAMP. CpG sites were mapped to human reference GRCh37 and annotated using IlluminaHumanMethylationEPICanno.ilm10b4.hg19. Leukocyte composition of each sample was estimated using the R package EpiDISH [17].

DNA methylation analysis

Longitudinal EWAS of LFC changes

For the 358 and the 96 participants in the two cohorts, DNAm and LFC were repeatedly measured. Longitudinal EWAS was performed using linear mixed models (LMM) in R package limma [18]. Methylation β value and LFC value were entered as the response and the independent variable, adjusting for age, sex, smoking and estimated leukocyte composition at baseline and at follow-up (model 1). Participant ID was entered as a random effect.

In literature reporting longitudinal EWAS of BMI, longitudinally changes in methylation and changes in BMI were used as independent and response variables [19]. Therefore, the association between longitudinally changes in DNAm and LFC was analyzed using multilinear regression, adjusting for age, sex, follow up time, and smoking status and estimated leukocyte composition at baseline and at follow-up (model 1).

Cross-sectional EWAS of LFC

Cross-sectional EWAS was performed to analyze the associations between DNAm and LFC at baseline and at follow-up, respectively. A multiple linear regression model in limma was used, adjusting age, sex, smoking and estimated leukocyte composition (model 1).

For both longitudinal and cross-sectional analysis, BMI was additionally adjusted in model 2.

Replication study

In the replication cohort, longitudinal and cross-sectional EWAS was conducted, adjusting for the same covariates as those in the discovery cohort in model 1 and model 2.

EWAS meta-analysis

In longitudinal and cross-sectional EWAS, inverse-variance weighted fixed-effects meta-analysis that combined EWAS results from the discovery and the replication cohorts was conducted, using the computationally efficient software METAL [20].

Epigenome-wide significant level was set at P < 1e-7 with Bonferroni correction.

Look up of the 22 known CpGs in Changfeng population

Twenty-two CpGs were reported cross-sectionally associated with hepatic fat in previous EWAS in European ancestry [10]. The CpGs were looked up in the present longitudinal EWAS meta-analysis results in model 1 to see whether they were also associated with LFC changes in individuals of Chinese ancestry. Associations between the 22 CpGs and LFC in cross-sectional EWAS meta-analyses at baseline and at follow-up respectively in model 1 are also presented. The significance level was set at 2.3e-3 (0.05/22) using Bonferroni correction.

Methylation quantitative trait loci (meQTL) of CpGs

To explore the potential genetic basis of the identified and replicated CpGs, clumped meQTLs of the markers were looked up in the EPIGEN MeQTL Database [21].

Functional enrichment analysis

Twenty CpGs that were most significantly associated with LFC in the longitudinal meta-analysis in model 1 were collected for the functional enrichment analysis. The CpGs were first annotated to their closest genes with the R package IlluminaHumanMethylationEPICanno.ilm10b4.hg19 [22]. Then, the Gene Ontology (GO) [23], Reactome pathway [24], and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway [25] enrichment of the annotated genes were conducted using the R package clusterProfiler v4.8.1 [26] (one-tailed hypergeometric test).

Cross-lagged panel analysis (CLPA)

CLPA is widely used to examine directional influences between intercorrelated and longitudinally changing variables. To test the longitudinal bidirectional relationship between LFC and DNAm, CLPA using longitudinal data of LFC and DNAm was performed, combining the two cohorts. Prior to analysis, LFC values were adjusted for age, sex, and smoking status by linear regression and residuals were Z-transformed; methylation β values were similarly processed but further adjusting for batch effects and estimated leukocyte composition (model 1). BMI was additionally adjusted in model 2. The analysis simultaneously estimated two cross-lagged path coefficients, which represent the effect of baseline LFC (or DNAm) on future DNAm (or LFC) at follow-up. Besides, temporal relationships between DNAm of the CpGs and BMI in model 1 are also presented. R package Lavaan with structural equation model was used to perform CLPA [27]. Model fitting was validated by a goodness-of-fit index > 0.90, a comparative fit index > 0.95, and a standardized root-mean-square residual < 0.08 [28].

Statistical analysis

Numerical variables are presented as mean (standard deviation, SD) for symmetrical distributions, and median [quartile 1, quartile 3] for asymmetric distributions. Differences in normally and nonnormally distributed data were analyzed using Student’s t test and the Mann-Whitney U test, respectively. Sex and smoking status are presented as n (%), and the chi-square test was used to compare proportions. All the statistical analyses were performed in the R statistical environment version 4.2.2. All reported p values are two-sided with a significance level of 0.05.

Results

Basic characteristics

The discovery cohort and the replication cohort had mean ages of 61.2 (SD: 7.29 years) and 61.9 (SD: 8.37 years) at baseline, respectively. Approximately 58% and 47% of the participants were female, respectively. The replication cohort had a slightly longer follow-up time than did the discovery cohort (4.39 years vs. 4.10 years). Baseline age, smoking status, BMI, LFC, and longitudinal change of LFC were comparable between the discovery and the replication cohorts (Table 1 and Additional File 1 Supplementary Fig. 1).

Table 1 Basic characteristics of the study participants
Full size table

Longitudinal EWAS

Longitudinal EWAS of LFC using LMM in the discovery cohort revealed that hypermethylation of cg11024682 (SREBF1), cg06500161 (ABCG1), and cg16740586 (ABCG1) was significantly associated with increased LFC after adjusting for age, sex, and smoking status (Fig. 2A; Table 2), which was successfully replicated in the replication cohort (P < 0.05 with the same direction of effects). Longitudinal EWAS of LFC using LMM in the replication cohort identified no significant associations (Fig. 2B). The fixed-effects meta-analysis combining EWAS results from the discovery and the replication cohorts revealed 5 significantly associated CpGs (P < 1e-7 with the same effect directions in the two cohorts), among which hypermethylation of cg15659943 (ABCA1) and cg00163198 (SNX19) was also associated with increased LFC (Fig. 2C; Table 2). Further adjusting for BMI attenuated the strength of the associations but the 5 CpGs remained near significantly associated with LFC (Table 2).

Fig. 2
figure 2

Manhattan plot of the longitudinal association of peripheral blood-derived DNA methylation with LFC using LMM. Plot was generated using results from the sex-, age-, and smoking-adjusted model in the discovery cohort (A), the replication cohort (B), and meta-analysis (C)

Full size image
Table 2 Significant CpGs in longitudinal EWAS and cross-sectional EWAS of LFC
Full size table

Analysis of the association between longitudinal change of LFC and that of DNAm found that hypomethylation of cg09154567 annotated to SULT4A1 was associated with increased LFC in the replication cohort, but not significant in the discovery cohort or the meta-analysis (Additional File 1 Supplementary Fig. 2 and Table 2).

Functional enrichment

Functional enrichment analysis was conducted for the top 20 CpGs associated with LFC according to the longitudinal meta-analysis results (Table 3). Specifically, the enrichment of genes where the top CpGs located in terms of GO, KEGG, and Reactome pathways was checked. As a result, the following biological processes were significantly enriched: lipid storage and transport (Fig. 3A-B), NR1H2 and NR1H3-mediated signaling (Fig. 3C), ABC transporters and cholesteral metabolism (Fig. 3D), and some other lipid-related functions and pathways.

Table 3 Top 20 CpGs that most significantly associated with LFC in the longitudinal meta-analysis in model 1
Full size table
Fig. 3
figure 3

Functional enrichment of the top 20 CpGs that most significantly associated with LFC. (A, B) Enrichment of the annotated genes in GO, where the x-axis represents the negative log p value of enrichment significance (A) and the number of genes overlapped with the functional term (B). (C, D) Enrichment of the annotated genes in Reactome pathways (C) and KEGG pathways (D)

Full size image

Cross-sectional EWAS

Cross-sectional EWAS at baseline and at follow-up in the two cohorts was performed, respectively. EWAS in the meta-analysis at follow-up found that hypomethylation of cg07162647 and hypermethylation of cg11024682 (SREBF1) and cg06500161 (ABCG1) were cross-sectionally associated with higher LFC (Additional File 1 Supplementary Fig. 3 and Table 2).

As shown in Additional File 1 Supplementary Table 1, the significant CpGs identified in the longitudinal analyses were also cross-sectionally associated with LFC with consistent directionality and nominal P < 0.05 in the meta-analyses at baseline and at follow-up, respectively.

Further adjustment of BMI did not materially change the associations (Table 2 and Additional File 1 Supplementary Table 1). A total of 8 meQTLs were found for the 6 CpGs identified in Changfeng population (Additional File 1 Supplementary Table 2).

Look up of the 22 previously known CpGs

Previous EWAS of hepatic fat in 1,496 participants from the Framingham Heart Study identified 58 significant CpGs with an FDR < 0.05, 22 of which were successfully replicated in 1,904 participants from another three cohorts in a sex- and age-adjusted model [10]. Eight of the 22 CpGs remained significant in the longitudinal EWAS using LMM in model 1 after Bonferroni correction, and were also nominally significant in the meta-analyses in cross-sectional EWAS at baseline and at follow-up (Table 4). This suggested that alterations in DNAm detected in peripheral blood could be verified in different ethnic groups, and was associated not only with cross-sectional LFC variation but also with longitudinal LFC changes. A total of 9 meQTLs were found for the 6 CpGs (except for cg11024682 and cg06500161) validated in Changfeng population (Additional File 1 Supplementary Table 3).

Table 4 Longitudinal and cross-sectional associations of 22 previously known CpGs with LFC in Changfeng population
Full size table

Cross-lagged panel analysis

CLPA was performed to investigate the longitudinal directional relationships between LFC and DNAm of the 12 identified and replicated CpGs. In model 1 adjusting for age, sex, and smoking status, baseline hypermethylation at cg11024682 (SREBF1) (β = 0.14 [0.06, 0.23]), cg16740586 (ABCG1) (β = 0.17 [0.08, 0.25]), cg06500161 (ABCG1) (β = 0.12 [0.03, 0.20]), cg27243685 (ABCG1) (β = 0.09 [0.01, 0.18]) and hypomethylation at cg17901584 (DHCR24) (β = -0.10 [-0.18, -0.02]), cg00574958 (CPT1A) (β = -0.09 [-0.17, -0.01]), cg08309687 (LINC00649) (β = -0.11 [-0.19, -0.03]) predicted increased subsequent LFC; higher LFC led to hypermethylated cg15659943 (ABCA1) (β = 0.13 [0.04, 0.21]), cg06500161 (ABCG1) (β = 0.10 [0.02, 0.18]), cg27243685 (ABCG1) (β = 0.10 [0.02, 0.18]) and hypomethylation at cg07162647 (β = -0.11 [-0.19, -0.03]) (Fig. 4).

Fig. 4
figure 4

Cross-lagged path coefficients in longitudinal analysis of LFC and DNA methylation measured at baseline and at follow-up in model 1 adjusted for age, sex, and smoking status

Full size image

Among the 12 identified and replicated CpGs, cg15659943 (ABCA1) (β = 0.07 [0.02, 0.11]) and cg11024682 (SREBF1) (β = 0.06 [0.01, 0.10]) showed longitudinal effects on subsequent increased BMI, and BMI was temporally associated with hypomethylation of cg00574958 (CPT1A) (β = -0.12 [-0.20, -0.05]) and cg08309687 (LINC00649) (β = -0.09 [-0.15, -0.03]) (Additional File 1 Supplementary Fig. 4).

In model 2 further adjusting for BMI, the temporal effects of cg11024682 (SREBF1), cg16740586 (ABCG1) and cg06500161 (ABCG1) on LFC were almost halved with p values of 0.076, 0.032, and 0.165, which indicated that the effects were partially influenced by increased obesity (Additional File 1 Supplementary Fig. 5). Temporal associations of other CpGs and LFC were not greatly changed (Additional File 1 Supplementary Fig. 5).

Discussion

The present research identified 5 CpGs associated with longitudinal change of LFC, another 1 CpG cross-sectionally associated with LFC, and replicated another 6 previously reported LFC-associated CpGs in longitudinal EWAS in the sex-, age-, and smoking-adjusted model. Among the significant CpGs, 9 showed longitudinal directional associations with LFC, and 8 remained significant after further adjusting for BMI in model 2. The findings are notable in not only identified/replicated DNAm of CpGs longitudinally associated with LFC in Chinese participants, but also distinguished between predictors and biomarkers.

Several cross-sectional EWASs have observed altered DNAm patterns collected from liver biopsies or peripheral blood samples of MASLD patients [29], but longitudinal EWASs of LFC are lacking. In accordance with previous studies, this study likewise showed that genes involved in lipid metabolism and biosynthesis pathways appeared to be the most prominent markers. Cg15659943, at which DNAm is longitudinally associated with baseline LFC, is annotated to ABCA1. A mild but significant effect of baseline LFC on subsequent hypermethylation at cg06500161 and cg27243685 mapping to ABCG1 was observed. ABCA1 and ABCG1 are members of the ABC transporter superfamily that are essential for the biogenesis of high-density lipoprotein (HDL) and reverse cholesterol transport. ABCA1 functions as an efflux pump of cholesterol from peripheral cells to apolipoprotein A-I and generates the nascent HDL, and ABCG1 facilitates subsequent cholesterol efflux to HDL for further maturation [30, 31]. These implicated that LFC alteration might regulate gene expression involved in reverse cholesterol transport via epigenetic mechanisms.

Cg00163198 is a novel EPIC array marker, and has been described to be related to incident type 2 diabetes [32] and cardiovascular health factors including BMI, blood pressure, glucose, and cholesterol levels [33]. It lies on the intronic region of the Sorting nexin 19 (SNX19) gene, which was found to be associated with liver enzymes, lipids, and body fat in a GWAS [34]. SNX19 enables the binding of lipids and phosphatidylinositol, and plays important roles in insulin secretion [35]. CpG cg00163198 located in an active regulatory region and was significantly correlated with an eQTL of SNX19 in the adipose tissue (eQTL P = 8.78 × 10− 22 by GTEx; meQTL P = 1.93 × 10− 13 by Pan-meQTL) [36]. This indicates that the regulation between cg00163198 and SNX19 is mediated by the eQTL and ultimately affects gene expression.

Evidence on causal relationships between DNAm and LFC or fatty liver is scarce to date. The present study did not conduct MR analysis due to the lack of large GWASs on LFC in Asian populations and insufficient statistical power. Previous studies applying CLPA reported that DNAm is a biomarker rather than a predictor or cause of BMI [37, 38], which was in accordance with the findings of MR [19] and longitudinal enrichment analysis [39]. Using CLPA in a cohort with repeatedly measured data, this study demonstrated that baseline hypermethylation at cg11024682 (SREBF1) and the three CpGs (cg16740586, cg06500161, and cg27243685) investigated mapping to ABCG1 predicted increased subsequent LFC levels. DNAm at cg11024682 (SREBF1) was reported associated with hepatic fat [10], BMI [19, 40], type 2 diabetes [41], blood triglyceride levels [42, 43], serum liver enzyme levels [44], and cardiovascular health [33]. Methylation at CpGs mapping to ABCG1 was reported associated with increased blood lipids, glucose, insulin, and increased risk of diabetes [45, 46]. DNAm of cg27243685 was reported associated with hepatic fat [10, 47]. Animal studies demonstrated that Abcg1 played a pivotal role in preventing hepatic fat accumulation in mice challenged with high-fat diet [48].

The temporal effects of DNAm of CpGs at SREBF1 and ABCG1 on subsequent LFC were reduced after adjusting for BMI. This was consistent with the findings from the cross-sectional EWAS in another study [10] and was not unexpected, as obesity, diabetes, and lipid metabolism are established correlated with each other and are related with MASLD. BMI was positively associated with LFC in Changfeng population with Spearman r = 0.32 (P < 2.2e-16), and increased obesity might lie on the mechanistic pathway from DNAm of CpGs at SREBF1 and ABCG1 to increased LFC. In CLPA, BMI showed weak association with cg11024682 (SREBF1) and non-significant associations with the three CpGs (cg16740586, cg06500161, and cg27243685) of ABCG1. Together, the results indicated that the effects of DNAm of SREBF1 and ABCG1 were mediated by increased BMI as well as other mechanisms, for example, diabetes and lipid metabolism.

Results of CLPA illustrated that hypomethylation at cg08309687 (LINC00649) has a strong prospective effect on increasing LFC, which was in accordance with previous findings from MR analysis [10]. There were four meQTLs at LINC00649 for cg08309687 according to the EPIGEN MeQTL Database. A three-way association by Ma et al. reported that 8% of the association of cg08309687 and hepatic fat was mediated by the expression of TMEM50B [10]. In addition, hypermethylation at cg17901584 (DHCR24) and cg00574958 (CPT1A) were putatively predictors for decreased LFC. DHCR24 encodes 24-dehydrocholesterol reductase that is involved in the final step of cholesterol synthesis, converting desmosterol into cholesterol [49]. Cg17901584 was also reported associated with metabolic traits such as BMI [19], blood lipids [43], and type 2 diabetes [50].

Strengths and limitations

The strength of this study lies in its prospective nature with repeatedly measured data, and it adds value to EWAS of fatty liver by distinguishing between predictors and biomarkers. This study has potential limitations. First, the modest sample size might result in limited statistical power to detect CpGs with small effect sizes. Second, the study used DNA from peripheral blood as liver tissues were unavailable. However, previous studies demonstrated that quantification of DNAm from peripheral blood or plasma cell-free DNA released by dying hepatocytes could reflect the severity of MASLD and could be suggested as an alternative approach with high accuracy to stratify MASLD [11, 51]. Third, experiments are needed to explore the mechanisms involved.

Conclusions

In summary, differences in peripheral blood DNA methylation occur in interindividual variation and intraindividual longitudinal change of LFC. Blood DNAm at SREBF1, ABCG1, DHCR24, CPT1A, and LINC00649 may be predictors of subsequent LFC change. The effects of DNAm at SREBF1 and ABCG1 were partially influenced by obesity, which implicates the importance of controlling body weight in preventing fatty liver. For clinical perspective, reversible alterations in DNA methylation result from various exposures and lead to changes in gene expression, distinguishing between predictors and biomarkers may be helpful in developing preventive and therapeutic strategies for fatty liver disease.

Data availability

Summary statistics of the EWAS of LFC in the present study is available from the corresponding authors upon reasonable request.

Change history

  • 15 December 2024

    The original online version of this article was revised: the authors would like to correct the grant number in the Funding section. The Funding currently reads: This work was supported by Shanghai Municipal Science and Technology Major Project (Grant No. 2017SHZDZX01), Chinese Academy of Sciences Young Team Program for Stable Support of Basic Research (Grant No. YSBR-077), the National Natural Science Foundation of China (Grant No. 32200472), and Outstanding Resident Clinical Postdoctoral Program of Zhongshan Hospital Affiliated to Fudan University (Grant No. 2023ZYYS-003). The Funding should read: This work was supported by Shanghai Municipal Science and Technology Major Project (Grant No. 2023SHZDZX02), Chinese Academy of Sciences Young Team Program for Stable Support of Basic Research (Grant No. YSBR-077), the National Natural Science Foundation of China (Grant No. 32200472), and Outstanding Resident Clinical Postdoctoral Program of Zhongshan Hospital Affiliated to Fudan University (Grant No. 2023ZYYS-003).

  • 17 December 2024

    A Correction to this paper has been published: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12944-024-02405-5

Abbreviations

DNAm:

DNA Methylation

LFC:

Liver Fat Content

EWAS:

Epigenome-Wide Association Study

CLPA:

Cross-Lagged Panel Analysis

MASLD:

Metabolic Dysfunction-Associated Steatotic Liver Disease

MR:

Mendelian Randomization

GWAS:

Genome-Wide Association Study

BMI:

Body Mass Index

LMM:

Linear Mixed Model

MeQTL:

Methylation Quantitative Trait Loci

GO:

Gene Ontology

KEGG:

Kyoto Encyclopedia of Genes and Genomes

SD:

Standard Deviation

HDL:

High-Density Lipoprotein

References

  1. Younossi ZM, Golabi P, Paik JM, Henry A, Van Dongen C, Henry L. The global epidemiology of nonalcoholic fatty liver disease (NAFLD) and nonalcoholic steatohepatitis (NASH): a systematic review. Hepatology. 2023;77:1335–47.

    Article  PubMed  Google Scholar 

  2. Le MH, Le DM, Baez TC, Wu Y, Ito T, Lee EY, Lee K, Stave CD, Henry L, Barnett SD, et al. Global incidence of non-alcoholic fatty liver disease: a systematic review and meta-analysis of 63 studies and 1,201,807 persons. J Hepatol. 2023;79:287–95.

    Article  PubMed  Google Scholar 

  3. Devarbhavi H, Asrani SK, Arab JP, Nartey YA, Pose E, Kamath PS. Global burden of liver disease: 2023 update. J Hepatol. 2023;79:516–37.

    Article  PubMed  Google Scholar 

  4. Luo F, Oldoni F, Das A. TM6SF2: a Novel Genetic Player in nonalcoholic fatty liver and Cardiovascular Disease. Hepatol Commun. 2022;6:448–60.

    Article  CAS  PubMed  Google Scholar 

  5. Luo F, Smagris E, Martin SA, Vale G, McDonald JG, Fletcher JA, Burgess SC, Hobbs HH, Cohen JC. Hepatic TM6SF2 is required for Lipidation of VLDL in a Pre-golgi Compartment in mice and rats. Cell Mol Gastroenterol Hepatol. 2022;13:879–99.

    Article  PubMed  Google Scholar 

  6. Xia M, Zeng H, Wang S, Tang H, Gao X. Insights into contribution of genetic variants towards the susceptibility of MAFLD revealed by the NMR-based lipoprotein profiling. J Hepatol. 2021;74:974–7.

    Article  CAS  PubMed  Google Scholar 

  7. Murphy SK, Yang H, Moylan CA, Pang H, Dellinger A, Abdelmalek MF, Garrett ME, Ashley-Koch A, Suzuki A, Tillmann HL, et al. Relationship between methylome and transcriptome in patients with nonalcoholic fatty liver disease. Gastroenterology. 2013;145:1076–87.

    Article  CAS  PubMed  Google Scholar 

  8. Zeybel M, Hardy T, Robinson SM, Fox C, Anstee QM, Ness T, Masson S, Mathers JC, French J, White S, Mann J. Differential DNA methylation of genes involved in fibrosis progression in non-alcoholic fatty liver disease and alcoholic liver disease. Clin Epigenetics. 2015;7:25.

    Article  PubMed  PubMed Central  Google Scholar 

  9. Ahrens M, Ammerpohl O, von Schönfels W, Kolarova J, Bens S, Itzel T, Teufel A, Herrmann A, Brosch M, Hinrichsen H, et al. DNA methylation analysis in nonalcoholic fatty liver disease suggests distinct disease-specific and remodeling signatures after bariatric surgery. Cell Metab. 2013;18:296–302.

    Article  CAS  PubMed  Google Scholar 

  10. Ma J, Nano J, Ding J, Zheng Y, Hennein R, Liu C, Speliotes EK, Huan T, Song C, Mendelson MM, et al. A peripheral blood DNA methylation signature of hepatic Fat reveals a potential causal pathway for nonalcoholic fatty liver disease. Diabetes. 2019;68:1073–83.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  11. Wu J, Zhang R, Shen F, Yang R, Zhou D, Cao H, Chen G, Pan Q, Fan J. Altered DNA methylation sites in Peripheral Blood leukocytes from patients with simple steatosis and Nonalcoholic Steatohepatitis (NASH). Med Sci Monit. 2018;24:6946–67.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  12. Gao X, Hofman A, Hu Y, Lin H, Zhu C, Jeekel J, Jin X, Wang J, Gao J, Yin Y, Zhao N. The Shanghai Changfeng Study: a community-based prospective cohort study of chronic diseases among middle-aged and elderly: objectives and design. Eur J Epidemiol. 2010;25:885–93.

    Article  PubMed  Google Scholar 

  13. Xia MF, Yan HM, He WY, Li XM, Li CL, Yao XZ, Li RK, Zeng MS, Gao X. Standardized ultrasound hepatic/renal ratio and hepatic attenuation rate to quantify liver fat content: an improvement method. Obes (Silver Spring). 2012;20:444–52.

    Article  CAS  Google Scholar 

  14. Tian Y, Morris TJ, Webster AP, Yang Z, Beck S, Feber A, Teschendorff AE. ChAMP: updated methylation analysis pipeline for Illumina BeadChips. Bioinformatics. 2017;33:3982–4.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  15. Teschendorff AE, Marabita F, Lechner M, Bartlett T, Tegner J, Gomez-Cabrero D, Beck S. A beta-mixture quantile normalization method for correcting probe design bias in Illumina Infinium 450 k DNA methylation data. Bioinformatics. 2013;29:189–96.

    Article  CAS  PubMed  Google Scholar 

  16. Johnson WE, Li C, Rabinovic A. Adjusting batch effects in microarray expression data using empirical Bayes methods. Biostatistics. 2007;8:118–27.

    Article  PubMed  Google Scholar 

  17. Zheng SC, Breeze CE, Beck S, Dong D, Zhu T, Ma L, Ye W, Zhang G, Teschendorff AE. EpiDISH web server: epigenetic dissection of Intra-sample-heterogeneity with online GUI. Bioinformatics. 2019;36:1950–1.

    Article  PubMed  PubMed Central  Google Scholar 

  18. Smyth GK. Linear models and empirical bayes methods for assessing differential expression in microarray experiments. Stat Appl Genet Mol Biol. 2004;3:Article3.

    Article  PubMed  Google Scholar 

  19. Wahl S, Drong A, Lehne B, Loh M, Scott WR, Kunze S, Tsai PC, Ried JS, Zhang W, Yang Y, et al. Epigenome-wide association study of body mass index, and the adverse outcomes of adiposity. Nature. 2017;541:81–6.

    Article  CAS  PubMed  Google Scholar 

  20. Willer CJ, Li Y, Abecasis GR. METAL: fast and efficient meta-analysis of genomewide association scans. Bioinformatics. 2010;26:2190–1.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  21. Villicaña S, Castillo-Fernandez J, Hannon E, Christiansen C, Tsai PC, Maddock J, Kuh D, Suderman M, Power C, Relton C, et al. Genetic impacts on DNA methylation help elucidate regulatory genomic processes. Genome Biol. 2023;24:176.

    Article  PubMed  PubMed Central  Google Scholar 

  22. KD H. IlluminaHumanMethylationEPICanno. ilm10b4. hg19: Annotation for Illumina’s EPIC methylation arrays. 2017. R package version 06 0.

  23. Consortium GO. The Gene Ontology (GO) database and informatics resource. Nucleic Acids Res. 2004;32:D258–61.

    Article  Google Scholar 

  24. Fabregat A, Jupe S, Matthews L, Sidiropoulos K, Gillespie M, Garapati P, Haw R, Jassal B, Korninger F, May B. The reactome pathway knowledgebase. Nucleic Acids Res. 2018;46:D649–55.

    Article  CAS  PubMed  Google Scholar 

  25. Kanehisa M, Goto S. KEGG: kyoto encyclopedia of genes and genomes. Nucleic Acids Res. 2000;28:27–30.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  26. Yu G, Wang LG, Han Y, He QY. clusterProfiler: an R package for comparing biological themes among gene clusters. Omics. 2012;16:284–7.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  27. Rosseel Y. Lavaan: an R Package for Structural equation modeling. J Stat Softw. 2012;48:1–36.

    Article  Google Scholar 

  28. Ryu E. Model fit evaluation in multilevel structural equation models. Front Psychol. 2014;5:81.

    Article  PubMed  PubMed Central  Google Scholar 

  29. Hyun J, Jung Y. DNA methylation in nonalcoholic fatty liver disease. Int J Mol Sci 2020, 21.

  30. Yu XH, Tang CK. ABCA1, ABCG1, and Cholesterol Homeostasis. Adv Exp Med Biol. 2022;1377:95–107.

    Article  CAS  PubMed  Google Scholar 

  31. Matsuo M. ABCA1 and ABCG1 as potential therapeutic targets for the prevention of atherosclerosis. J Pharmacol Sci. 2022;148:197–203.

    Article  CAS  PubMed  Google Scholar 

  32. Hillary RF, McCartney DL, Smith HM, Bernabeu E, Gadd DA, Chybowska AD, Cheng Y, Murphy L, Wrobel N, Campbell A, et al. Blood-based epigenome-wide analyses of 19 common disease states: a longitudinal, population-based linked cohort study of 18,413 Scottish individuals. PLoS Med. 2023;20:e1004247.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  33. Zheng Y, Joyce BT, Hwang SJ, Ma J, Liu L, Allen NB, Krefman AE, Wang J, Gao T, Nannini DR, et al. Association of Cardiovascular Health through Young Adulthood with genome-wide DNA methylation patterns in midlife: the CARDIA Study. Circulation. 2022;146:94–109.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  34. Martin S, Cule M, Basty N, Tyrrell J, Beaumont RN, Wood AR, Frayling TM, Sorokin E, Whitcher B, Liu Y, et al. Genetic evidence for different adiposity phenotypes and their opposing influences on ectopic Fat and Risk of Cardiometabolic Disease. Diabetes. 2021;70:1843–56.

    Article  CAS  PubMed  Google Scholar 

  35. Harashima S, Horiuchi T, Wang Y, Notkins AL, Seino Y, Inagaki N. Sorting nexin 19 regulates the number of dense core vesicles in pancreatic β-cells. J Diabetes Investig. 2012;3:52–61.

    Article  CAS  PubMed  Google Scholar 

  36. Lonsdale J, Thomas J, Salvatore M, Phillips R, Lo E, Shad S, Hasz R, Walters G, Garcia F, Young N. The genotype-tissue expression (GTEx) project. Nat Genet. 2013;45:580–5.

    Article  CAS  Google Scholar 

  37. Reed ZE, Suderman MJ, Relton CL, Davis OSP, Hemani G. The association of DNA methylation with body mass index: distinguishing between predictors and biomarkers. Clin Epigenetics. 2020;12:50.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  38. Sun D, Zhang T, Su S, Hao G, Chen T, Li QZ, Bazzano L, He J, Wang X, Li S, Chen W. Body Mass Index drives changes in DNA methylation: a longitudinal study. Circ Res. 2019;125:824–33.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  39. Vehmeijer FOL, Küpers LK, Sharp GC, Salas LA, Lent S, Jima DD, Tindula G, Reese S, Qi C, Gruzieva O, et al. DNA methylation and body mass index from birth to adolescence: meta-analyses of epigenome-wide association studies. Genome Med. 2020;12:105.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  40. Mendelson MM, Marioni RE, Joehanes R, Liu C, Hedman ÅK, Aslibekyan S, Demerath EW, Guan W, Zhi D, Yao C, et al. Association of Body Mass Index with DNA methylation and gene expression in blood cells and relations to Cardiometabolic Disease: a mendelian randomization Approach. PLoS Med. 2017;14:e1002215.

    Article  PubMed  PubMed Central  Google Scholar 

  41. Walaszczyk E, Luijten M, Spijkerman AMW, Bonder MJ, Lutgers HL, Snieder H, Wolffenbuttel BHR, van Vliet-Ostaptchouk JV. DNA methylation markers associated with type 2 diabetes, fasting glucose and HbA(1c) levels: a systematic review and replication in a case-control sample of the lifelines study. Diabetologia. 2018;61:354–68.

    Article  CAS  PubMed  Google Scholar 

  42. Braun KVE, Dhana K, de Vries PS, Voortman T, van Meurs JBJ, Uitterlinden AG, Hofman A, Hu FB, Franco OH, Dehghan A. Epigenome-wide association study (EWAS) on lipids: the Rotterdam Study. Clin Epigenetics. 2017;9:15.

    Article  PubMed  PubMed Central  Google Scholar 

  43. Wu Z, Chen L, Hong X, Si J, Cao W, Yu C, Huang T, Sun D, Liao C, Pang Y, et al. Temporal associations between leukocytes DNA methylation and blood lipids: a longitudinal study. Clin Epigenetics. 2022;14:132.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  44. Nano J, Ghanbari M, Wang W, de Vries PS, Dhana K, Muka T, Uitterlinden AG, van Meurs JBJ, Hofman A, Franco OH, et al. Epigenome-Wide Association Study Identifies Methylation Sites Associated with Liver enzymes and Hepatic Steatosis. Gastroenterology. 2017;153:1096–e11061092.

    Article  CAS  PubMed  Google Scholar 

  45. Dayeh T, Tuomi T, Almgren P, Perfilyev A, Jansson PA, de Mello VD, Pihlajamäki J, Vaag A, Groop L, Nilsson E, Ling C. DNA methylation of loci within ABCG1 and PHOSPHO1 in blood DNA is associated with future type 2 diabetes risk. Epigenetics. 2016;11:482–8.

    Article  PubMed  PubMed Central  Google Scholar 

  46. Hidalgo B, Irvin MR, Sha J, Zhi D, Aslibekyan S, Absher D, Tiwari HK, Kabagambe EK, Ordovas JM, Arnett DK. Epigenome-wide association study of fasting measures of glucose, insulin, and HOMA-IR in the Genetics of lipid lowering drugs and Diet Network study. Diabetes. 2014;63:801–7.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  47. Ochoa-Rosales C, Portilla-Fernandez E, Nano J, Wilson R, Lehne B, Mishra PP, Gao X, Ghanbari M, Rueda-Ochoa OL, Juvinao-Quintero D, et al. Epigenetic link between statin therapy and type 2 diabetes. Diabetes Care. 2020;43:875–84.

    Article  CAS  PubMed  Google Scholar 

  48. Kennedy MA, Barrera GC, Nakamura K, Baldán Á, Tarr P, Fishbein MC, Frank J, Francone OL, Edwards PA. ABCG1 has a critical role in mediating cholesterol efflux to HDL and preventing cellular lipid accumulation. Cell Metabol. 2005;1:121–31.

    Article  CAS  Google Scholar 

  49. Zerenturk EJ, Sharpe LJ, Ikonen E, Brown AJ. Desmosterol and DHCR24: unexpected new directions for a terminal step in cholesterol synthesis. Prog Lipid Res. 2013;52:666–80.

    Article  CAS  PubMed  Google Scholar 

  50. Fraszczyk E, Spijkerman AMW, Zhang Y, Brandmaier S, Day FR, Zhou L, Wackers P, Dollé MET, Bloks VW, Gào X, et al. Epigenome-wide association study of incident type 2 diabetes: a meta-analysis of five prospective European cohorts. Diabetologia. 2022;65:763–76.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  51. Hardy T, Zeybel M, Day CP, Dipper C, Masson S, McPherson S, Henderson E, Tiniakos D, White S, French J, et al. Plasma DNA methylation: a potential biomarker for stratification of liver fibrosis in non-alcoholic fatty liver disease. Gut. 2017;66:1321–8.

    Article  CAS  PubMed  Google Scholar 

Download references

Acknowledgements

Not applicable.

Funding

This work was supported by Shanghai Municipal Science and Technology Major Project (Grant No. 2023SHZDZX02), Chinese Academy of Sciences Young Team Program for Stable Support of Basic Research (Grant No. YSBR-077), the National Natural Science Foundation of China (Grant No. 32200472), and Outstanding Resident Clinical Postdoctoral Program of Zhongshan Hospital Affiliated to Fudan University (Grant No. 2023ZYYS-003).

Author information

Author notes

  1. Hailuan Zeng and Wenran Li contributed equally to this work.

Authors and Affiliations

  1. Department of Endocrinology and Metabolism, Zhongshan Hospital, Fudan University, NO. 180 Fenglin Road, Shanghai, 200032, China

    Hailuan Zeng, Mingfeng Xia, Huandong Lin & Xin Gao

  2. Fudan Institute for Metabolic Diseases, Shanghai, China

    Hailuan Zeng, Mingfeng Xia, Huandong Lin & Xin Gao

  3. Human Phenome Institute, Fudan University, Shanghai, China

    Hailuan Zeng & Xin Gao

  4. CAS Key Laboratory of Computational Biology, Shanghai Institute of Nutrition and Health, University of Chinese Academy of Sciences, Chinese Academy of Sciences, Shanghai, China

    Wenran Li, Jieyu Ge & Sijia Wang

  5. Department of Geriatrics, Zhongshan Hospital, Fudan University, Shanghai, China

    Hui Ma & Lingyan Chen

  6. Department of Laboratory Medicine, Zhongshan Hospital, Fudan University, Shanghai, China

    Baishen Pan

  7. Taizhou Institute of Health Sciences, Fudan University, Taizhou, Jiangsu, China

    Sijia Wang

  8. Center for Excellence in Animal Evolution and Genetics, Chinese Academy of Sciences, Kunming, China

    Sijia Wang

Authors
  1. Hailuan Zeng
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Wenran Li
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Mingfeng Xia
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Jieyu Ge
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Hui Ma
    View author publications

    You can also search for this author inPubMed Google Scholar

  6. Lingyan Chen
    View author publications

    You can also search for this author inPubMed Google Scholar

  7. Baishen Pan
    View author publications

    You can also search for this author inPubMed Google Scholar

  8. Huandong Lin
    View author publications

    You can also search for this author inPubMed Google Scholar

  9. Sijia Wang
    View author publications

    You can also search for this author inPubMed Google Scholar

  10. Xin Gao
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

Study conceptualization, project administration, supervision, writing-review & editing: XG, SW and HL. Data analysis, visualization, interpretation, and writing-original draft: HZ and WL. Data curation: HL, HM, LC, MX, JG and BP. Funding acquisition: XG, SW, WL and HZ.

Corresponding authors

Correspondence to Huandong Lin, Sijia Wang or Xin Gao.

Ethics declarations

Ethics approval and consent to participate

All participants provided written informed consent, with ethics approval granted by the ethics committee of Zhongshan Hospital affiliated to Fudan University.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

The original online version of this article was revised: the authors would like to correct the grant number in the Funding section. The Funding currently reads: This work was supported by Shanghai Municipal Science and Technology Major Project (Grant No. 2017SHZDZX01), Chinese Academy of Sciences Young Team Program for Stable Support of Basic Research (Grant No. YSBR-077), the National Natural Science Foundation of China (Grant No. 32200472), and Outstanding Resident Clinical Postdoctoral Program of Zhongshan Hospital Affiliated to Fudan University (Grant No. 2023ZYYS-003). The Funding should read: This work was supported by Shanghai Municipal Science and Technology Major Project (Grant No. 2023SHZDZX02), Chinese Academy of Sciences Young Team Program for Stable Support of Basic Research (Grant No. YSBR-077), the National Natural Science Foundation of China (Grant No. 32200472), and Outstanding Resident Clinical Postdoctoral Program of Zhongshan Hospital Affiliated to Fudan University (Grant No. 2023ZYYS-003).

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Zeng, H., Li, W., Xia, M. et al. Longitudinal association of peripheral blood DNA methylation with liver fat content: distinguishing between predictors and biomarkers. Lipids Health Dis 23, 309 (2024). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12944-024-02304-9

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12944-024-02304-9

Keywords

Lipids in Health and Disease

ISSN: 1476-511X