DOI: https://doi.org/10.21203/rs.3.rs-2103290/v1
Background: Sarcopenia is progressive loss of skeletal muscle mass and strength that can lead to physical impairment, poor quality of life, and death. DNA methylation is being studied as a hallmark with a crucial influence on aging and sarcopenia. However, studies have limitations in that they depended on a small sample size, and did not distinguish between those with sufficient muscle mass and those with insufficient muscle mass among the older people. Therefore, extensive studies on DNA methylation in older people with sarcopenia are needed.
Methods: We obtained Korean Genome and Epidemiology Study (KOGES) data conducted between 2009–2010 for analysis. We compared the demographic data of people with high muscle mass index (MMI) and those of people with low MMI. Furthermore, we conducted a DNA methylation study and investigated the effects of epigenetic factors on sarcopenia by identifying differentially methylated regions (DMRs). The pathfindR package of R software was used to perform DMR enrichment analysis to evaluate the relationship between identified DMRs and MMI according to gender.
Results: Muscle loss according to age was clearly revealed in men, but in women, the age difference according to MMI was not significant in demographic study. The enrichment analysis of DMRs showed that in the male group, human T-cell leukemia virus 1 infection showed the highest association, followed by allograft rejection, graft-versus-host disease, type 1 diabetes mellitus, and autoimmune thyroid disease. On the other hand, cell cycle showed the highest association, followed by ubiquitin-mediated proteolysis and the MAPK signaling pathway in women group. In men, many DMRs related to autoimmune were found, and in women, the ubiquitin-proteasome system-related DMRs played an important role.
Conclusions: The present study results provide differences according to gender in the epigenetic study of sarcopenia and provide an insight in the direction of further sarcopenia research.
Aging refers to a decrease in physical ability and physiological function with increasing age, increasing disease risk and mortality rate.(1) Humans gradually lose 0.1–0.5% of skeletal muscle every year from the age of 30, and the loss accelerates from the age of 65.(2) The strength also decreases along with the muscle mass decrease. This age-related loss of muscle mass and strength is called sarcopenia.(3) Sarcopenia is associated with restricted mobility and poor quality of life, consequently, it has attracted increasing research interest.(4) European Working Group on Sarcopenia in Older People (EWGSOP) also referred clinical definition of sarcopenia as progressive loss of skeletal muscle mass and strength that can lead to physical impairment, poor quality of life, and death.(5) Sarcopenia is a multifactorial pathogenesis that is associated with various factors related to aging, such as oxidative stress, chronic inflammation, and lifestyle factors.(6) Among them, DNA methylation is being studied as a hallmark with a crucial influence on aging, and the relationship between epigenetic changes and sarcopenia associated with such aging has also recently attracted a lot of attention.(7)
Epigenetics are modifications of DNA that control gene expression by preventing transcription factors from binding to DNA or recruiting proteins implicated in gene regulation.(8) DNA methylation is the most well-studied epigenetic alteration in the context of aging.(9) According to recent studies, age-related DNA methylation has been reported to alter muscle methylome. They found epigenetic features such as age-associated CpG methylation changes between young and old muscle samples.(10–12) However, the above studies have limitations in that they depended on a small sample size (n = 10–50), and did not distinguish between those with sufficient muscle mass and those with insufficient muscle mass among the older people. Therefore, extensive studies on DNA methylation in older people with sarcopenia are needed.
The epidemiological evidence on the risk of sarcopenia in elderly men and women has been controversial.(13–15) According to Payette et al., absolute muscle loss was greater in men than in women, regardless of the initial muscle mass.16 However, according to Batsis et al., the prevalence of sarcopenia in women is low, but the mortality rate is gradually increasing.(17) There is also a cohort study that studied the difference in sarcopenia between men and women with 200 subjects, but the overall number of subjects was insufficient, and studies on the mechanism of the difference were insufficient.(18)
Therefore, we conducted a DNA methylation study for aging sarcopenia using the Korean Genome and Epidemiology Study (KOGES) data for men and women separately. We compared the methylation data of people with high muscle mass index (MMI) and those of people with low MMI. The present study aimed to investigate the effects of epigenetic factors on sarcopenia by identifying differentially methylated regions (DMRs) in men and women.
The epidemiological and genomic data sets used in this investigation were collected from a survey conducted between 2009–2010 by the Korea Center for Disease and Prevention's Ansan and Ansung cohort of the KoGES. The study was approved by the institutional review board (IRB) of Gyeongsang National University (approval number: GNUIRB-2019-04-010-013). Informed consent was obtained from all participants and all methods were performed in accordance with the relevant guidelines and regulations. The number of baseline participants was 9,351, who lived in Ansan or Ansung located in Gyeonggi Province, South Korea. Participants’ age in this cohort ranged from 40 to 69 years old. Among the participants, 446 subjects with a DNA methylation profile were included in the study. We divided the subjects into male and female groups; 226 males and 220 females. After the subjects were further classified as top 30% and the bottom 30% based on MMI in each group we were left with; men (n = 115) and women (n = 93). And then, we examined the demographics of the subjects and performed an epigenetic analysis based on their DNA methylation profile. Based on the analysis, 8 males and 13 females with missing values were excluded. Finally, 107 males and 80 females were investigated in this study. The study population selection flowchart is presented in Figure S1.
During each follow-up visit, all the participants went to a community clinic for clinical examinations. The weight and height of the subjects were estimated. MMI was calculated by dividing the muscle mass by the square of the height. The body mass index (BMI) was calculated by dividing the weight in kilograms by the square of the height in meters. The circumference of the waist and hips were also measured. For other surveys, the status of drinking and smoking and history of hypertension, diabetes, allergy, myocardial infarction, congestive heart failure, coronary artery disease, hyperlipidemia, asthma, peripheral vascular disease, various tumors, cerebrovascular disease, degenerative arthritis, and rheumatoid arthritis were taken.
The subjects’ KoGES DNA methylation data were obtained using an Infinium HumanMethylation 450K beadChip (Illumina, Inc., San Diego, CA, USA) containing over 485,000 CpG probes. Each CpG probe has a beta value indicating the DNA methylation level. [(Methylated reads) / (Unmethylated reads) + (Methylated reads)] was used to calculate the beta value, which ranged from 0 to 1; more methylation of CpG results in a value closer to 1. There were 389,321 CpGs left after filtering for epigenome-wide association analysis. Then, using R, we combined the CpG locations and annotation data from the Illumina Human Methylation EPIC manifest package provided by Bioconductor (https://www.bioconductor.org). The pathfindR package of R software was used to perform DMR enrichment analysis. Circos plot was used to show heatmaps of methylated regions.
Categorical data is expressed as n (%), while continuous data is expressed as mean ± standard deviation. If the normality assumption was met, an unpaired t-test was used for continuous variables to discover significant differences in baseline characteristics or clinical features between the under 0.3 and over 0.7 quantiles of MMI. Also, Wilcoxon's rank-sum test was used. The Chi-squared test for categorical variables was used to examine differences in proportions between the under 0.3 and over 0.7 quantiles of MMI. Fisher's exact test was used if the Chi-squared test assumption was not met.
To identify genes from methylation CpG site annotation results between muscle mass traits groups, an unpaired T-test was used. A volcano plot was used to illustrate methylated CpG sites. The following criteria were used to differentiate each CpG site: p < 10^-2 and | fold change | > 1.2.
R software version 4.1.2 was used to conduct all statistical analyses. (R Core Team. R Foundation for Statistical Computing, Vienna, Austria, 2020). The p < 0.05 was set as the significance level.
In this study, the waist and hip circumference, age, and BMI according to MMI were compared among the groups. In addition, the disease history, hypertension, diabetes, allergy, myocardial infarction, congestive heart failure, coronary artery disease, hyperlipidemia, asthma, peripheral vascular disease, various tumors, cerebrovascular disease, degenerative arthritis, and rheumatoid arthritis were compared. Based on the results, muscle loss according to age was clearly revealed in men, but in women, the age difference according to MMI was not significant. BMI and circumstance of waist and hip showed significant differences in both males and females (Table 1).
| Men | Women | |||||||
|---|---|---|---|---|---|---|---|---|
| Characteristic | MMI 70% over (n = 60) | MMI 30% less (n = 47) | p.val | sig. | MMI 70% over (n = 38) | MMI 30% less (n = 42) | p.val | sig |
| Age | 47.6 ± 5.6 | 52.6 ± 9.0 | 0.009 | ** | 51.5 ± 7.0 | 51.8 ± 9.3 | 0.912 | |
| BMI | 27.2 ± 2.5 | 21.2 ± 2.2 | < .001 | *** | 28.7 ± 3.1 | 21.7 ± 2.0 | < .001 | *** |
| Waist | 89.1 ± 6.3 | 77.9 ± 6.4 | < .001 | *** | 89.9 ± 8.9 | 74.0 ± 7.2 | < .001 | *** |
| Hip | 98.6 ± 5.0 | 89.4 ± 5.4 | < .001 | *** | 98.5 ± 5.8 | 89.5 ± 4.7 | < .001 | *** |
| Drink | 0.251 | 0.339 | ||||||
| Never | 10 (16.7%) | 8 (17.0%) | 32 (84.2%) | 32 (76.2%) | ||||
| Former drinker | 6 (10.0%) | 10 (21.3%) | 0 ( 0.0%) | 3 ( 7.1%) | ||||
| Current drinker | 44 (73.3%) | 29 (61.7%) | 6 (15.8%) | 7 (16.7%) | ||||
| Smoke | 0.191 | 0.222 | ||||||
| Never smoked | 15 (25.0%) | 6 (12.8%) | 36 (94.7%) | 42 (100.0%) | ||||
| Former smokers | 14 (23.3%) | 13 (27.7%) | 1 ( 2.6%) | 0 ( 0.0%) | ||||
| Current smokers (Sometimes) | 1 ( 1.7%) | 4 ( 8.5%) | 1 ( 2.6%) | 0 ( 0.0%) | ||||
| Current smokers (often) | 30 (50.0%) | 24 (51.1%) | 0 (0.0%) | 0 (0.0%) | ||||
| History of Diabetes | 1 | 0.844 | ||||||
| No | 56 (93.3%) | 44 (93.6%) | 31 (81.6%) | 36 (85.7%) | ||||
| Yes | 4 ( 6.7%) | 3 ( 6.4%) | 7 (18.4%) | 6 (14.3%) | ||||
| History of Allergy | 0.629 | 1 | ||||||
| No | 57 (95.0%) | 46 (97.9%) | 36 (94.7%) | 40 (95.2%) | ||||
| Yes | 3 ( 5.0%) | 1 ( 2.1%) | 2 ( 5.3%) | 2 ( 4.8%) | ||||
| History of Myocardial infarction | 1 | 0.475 | ||||||
| No | 60 (100.0%) | 47 (100.0%) | 37 (97.4%) | 42 (100.0%) | ||||
| Yes | 0 (50.0%) | 0 (50.0%) | 1 ( 2.6%) | 0 ( 0.0%) | ||||
| History of Congestive heart failure | ||||||||
| No | 60 (100.0%) | 47 (100.0%) | 38 (100.0%) | 42 (100.0%) | ||||
| Yes | ||||||||
| History of Coronary artery disease | 0.503 | 1 | ||||||
| No | 58 (96.7%) | 47 (100.0%) | 38 (100.0%) | 42 (100.0%) | ||||
| Yes | 2 ( 3.3%) | 0 ( 0.0%) | 1 (100.0%) | 0 ( 0.0%) | ||||
| History of Hyperlipidemia | 0.503 | 1 | ||||||
| No | 58 (96.7%) | 47 (100.0%) | 38 (100.0%) | 41 (97.6%) | ||||
| Yes | 2 ( 3.3%) | 0 ( 0.0%) | 0 ( 0.0%) | 1 ( 2.4%) | ||||
| History of Asthma | 1 | 1 | ||||||
| No | 58 (96.7%) | 45 (95.7%) | 37 (97.4%) | 41 (97.6%) | ||||
| Yes | 2 ( 3.3%) | 2 ( 4.3%) | 1 ( 2.6%) | 1 ( 2.4%) | ||||
| History of Peripheral vascular disease | 1 | 0.475 | ||||||
| No | 60 (100.0%) | 47 (100.0%) | 37 (97.4%) | 42 (100.0%) | ||||
| Yes | 0 ( 0.0%) | 1 (100.0%) | 1 ( 2.6%) | 0 ( 0.0%) | ||||
| History of Various tumors | 0.439 | 1 | ||||||
| No | 60 (100.0%) | 46 (97.9%) | 38 (100.0%) | 41 (97.6%) | ||||
| Yes | 0 ( 0.0%) | 1 ( 2.1%) | 0 ( 0.0%) | 1 ( 2.4%) | ||||
| History of Cerebrovascular disease | 0.31 | 0.475 | ||||||
| No | 60 (100.0%) | 47 (100.0%) | 37 (97.4%) | 42 (100.0%) | ||||
| Yes | 2 (22.2%) | 7 (77.8%) | 1 ( 2.6%) | 0 ( 0.0%) | ||||
| History of Degenerative arthritis | 0.652 | 0.276 | ||||||
| No | 58 (96.7%) | 44 (93.6%) | 29 (76.3%) | 37 (88.1%) | ||||
| Yes | 2 ( 3.3%) | 3 ( 6.4%) | 9 (23.7%) | 5 (11.9%) | ||||
| History of Rheumatoid arthritis | 0.693 | 0.117 | ||||||
| No | 56 (93.3%) | 45 (95.7%) | 38 (100.0%) | 38 (90.5%) | ||||
| Yes | 4 ( 6.7%) | 2 ( 4.3%) | 0 ( 0.0%) | 4 ( 9.5%) | ||||
| Time of physical activity during the day (Hours) | 0.028 | * | 0.287 | |||||
| Steady state | 2.1 ± 2.1 | 3.0 ± 2.2 | 0.032 | * | 2.5 ± 2.4 | 2.0 ± 1.9 | 0.481 | |
| Sedentary state | 6.0 ± 2.3 | 5.3 ± 2.6 | 0.136 | 5.0 ± 2.8 | 4.5 ± 2.5 | 0.314 | ||
| Light activity | 4.4 ± 2.6 | 3.8 ± 2.5 | 0.242 | 4.3 ± 2.2 | 4.7 ± 2.6 | 0.545 | ||
| Moderate activity | 1.9 ± 2.5 | 1.5 ± 2.1 | 0.538 | 1.3 ± 1.8 | 1.6 ± 2.3 | 0.921 | ||
| Hard activity | 1.7 ± 2.8 | 2.3 ± 3.2 | 0.182 | 3.2 ± 3.6 | 1.8 ± 3.3 | 0.044 | * | |
| *p < 0.05, **p < 0.01, ***p < 0.001 | ||||||||
In order to identify the DNA methylation pattern according to muscle mass in men and women, DNA methylation analysis was performed between people with an MMI of 30% or more and 30% or less in each group. The DNA methylation matrices of the two groups classified by sex are represented by volcano plots (Fig. 1). In the male group, a total of 54 DMRs were observed; 14 hypermethylated regions and 40 hypomethylated regions (Fig. 1A). In the female group, a total of 3868 DMRs were observed; 1864 hypermethylated regions and 2004 hypomethylated regions (Fig. 1B). A heatmap for each group showing the methylation changes with MMI was also generated (Figure S2).
We performed pathway enrichment analysis on the DMRs using pathfindR. In addition, a bubble chart of enrichment results, upset plots of enriched terms, gene-concept network, and aggregated term scores were used to visualize the enrichment analysis results. In the male group, human T-cell leukemia virus 1 infection showed the highest association, followed by allograft rejection, graft-versus-host disease, type 1 diabetes mellitus, and autoimmune thyroid disease (Fig. 2A). The upset plot shows the top related genes, and among them, HLA-DQB1, HLA-DQA1, and HLA-A were included in the most highly related pathways (Fig. 3A). Gene-concept network and aggregated term scores also showed related terms and genes (Figure S3A, S3B).
In the female group, the cell cycle showed the highest association, followed by ubiquitin-mediated proteolysis and the MAPK signaling pathway (Fig. 2B). The upset plot also represented and showed that the top related genes are associated with high significance according to which terms they belong (Fig. 3B). The relationship between the top 10 terms and related DMRs was visualized through the gene-concept network. Among them, ubiquitin-mediated proteolysis, endocytosis, NF-kappa B signaling pathway, and MAPK signaling pathway were at the center of the network diagram (Figure S4A). Aggregated term scores also showed term scores between the top 30% and bottom 30% based on MMI (Figure S4B).
The Circos plot has heatmaps indicating fold changes, and peaks indicating − log10(p-value). The tracks located outside the Circos plot are the correlation and t-test analysis results for 107 men and 80 women each at baseline. The tracks outside of the Circos plot are the baseline correlation and t-test analysis results for 107 men and 80 women. In both male and female circos plots, the most significant region was the short arm of chromosome 9. Also, in the case of women, there were regions of high significance in chromosomes 17 and 19 (Fig. 4).
The present study, an epigenetic nationwide cohort study using KOGES data, compared the elderly with relatively sufficient muscle mass and those with relatively insufficient muscle mass based on the MMI for each gender. The demographic study of men showed a significant decrease in MMI with age. However, in the case of women, there was no significant difference in MMI according to age. According to Yanping Du et al., males had a higher incidence of sarcopenia than females using the criteria of the Asian Working Group for Sarcopenia.(19) In the physical activity category of men, a steady state was found to be significantly higher in the lower 30% of the MMI group. On the other hand, in the case of women, hard activity was significantly higher in the MMI top 30% group. The history of the disease did not show any significant results in both men and women.
Sarcopenia is characterized by somatic changes with age and is described as skeletal muscle loss.(20) In addition, differences in the incidence of sarcopenia between men and women have been studied. In particular, in the case of men, the decrease in muscle mass and strength is more prominent than in women.(21) According to Gallagher et al., total apendicular muscle mass loss according to age in men and women was 15% and 11%.(22) Over a three-year period, older men lost twice as much absolute knee extensor strength as women in the Health ABC trial.(23) Notably, a decrease in physical function in females is more evident in people aged 65–74 years.(24) In addition, women exhibited lower functional impairment, fracture risk, and loss of independence than men, according to the clinical trial.(25)
In the present study, we examined the change in the methylation profile according to muscle mass loss in males and females. In the enrichment analysis, genes related to autoimmune-related disease pathways such as allograft rejection, graft-versus-host disease, type 1 diabetes mellitus, and autoimmune thyroid disease were significantly present in the male group DMRs. Studies on the relationship between sarcopenia and autoimmune diseases have been ongoing. Among them, research on rheumatoid arthritis, a representative autoimmune disease, is the most active. Cross-sectional studies of RA and sarcopenia revealed a higher incidence of sarcopenia in RA patients.25–28 According to Hyo Jin An et al., sarcopenia prevalence ranged from 10.1 to 45.1 percent, with a median of 29.1 percent in the RA studies.(30) In addition, the association between sarcopenia and inflammatory bowel disease, and type 1 diabetes mellitus has also been studied.(31, 32) However, there have been few studies on the epigenetic effect on sarcopenia in men compared to women in relation to autoimmune diseases.
Patients with sarcopenia are at an increased risk of hip fracture due to falls because of weakened muscle strength and poor balance.(33) According to Kang et al, the immunological and inflammatory reactions in the lungs were increased following a bone fracture.(34) Therefore, according to this study, in men, hip fracture due to sarcopenia and pulmonary inflammation due to immune changes in the lungs seem to have a positive correlation. Alternatively, since both sarcopenia and pneumonia are immune-related diseases, the incidence could be higher in patients with autoimmune diseases. Further studies on the co-occurrence rate of sarcopenia and pneumonia in patients with immune diseases are needed.
In the enrichment analysis of the female group, DMRs related to ubiquitin-mediated proteolysis, MAPK signaling pathway, and NF-kappa B signaling pathway were significant. MAPK protein affects skeletal muscle by several signaling modules, such as ERK1/2, p38 MAPK, and ERK5.(35) The ubiquitin-proteasome system is a key regulator of protein breakdown, allowing regulatory and structural proteins to be degraded selectively.(36) In addition, NF-kappa B can induce skeletal muscle protein degradation by upregulating the expression of several proteins in the ubiquitin-proteasome system.(37) Muscle degeneration which is induced by the ubiquitin-proteasome system, have a strong influence on aging sarcopenia;(38) catabolic effects in skeletal muscle by MuRF-1 and MAFbx, which are representative factors in the ubiquitin-proteasome system, may be the main cause of sarcopenia.(39) Therefore, the results of this female epigenetic study suggest that protein degradation through the signaling pathway-like ubiquitin-proteasome system is an important cause of muscle mass loss in the pathogenesis of sarcopenia.
In both male and female circos plots in this study, the most significant region was found in the short arm of chromosome 9. According to a study by Ikeuchi, distal myopathy with rimmed vacuoles, which is a neuromuscular disorder and causes skeletal muscle weakness, showed that chromosome 9 is affected by the gene loci.(40) Also, in another study, rs7022373, the most significant SNP in the pathogenesis of sarcopenia, is located on chromosome 9.(41) However, in epigenetic studies, there are only a few studies on chromosomes that are highly related to sarcopenia, therefore, further studies are needed.
The present study has several limitations. First, we measured the MMI of the subjects, but not the grip strength used in the diagnosis of sarcopenia at the same time. Second, the age range of the subjects ranged from 40 to 69, which is included in the middle-aged. However, prevention of sarcopenia in middle age is necessary for the elderly’s quality of life. Third, there were no common results between DMRs and genotypes. Further studies with extended cohorts are necessary.
In the present study, we found that epigenetic features were different in males and females with sarcopenia using methylation profiles from the KOGES data. In men, many DMRs related to autoimmune were found, and in women, the ubiquitin-proteasome system-related DMRs played an important role. Also, the most significant methylation region was present on chromosome 9 in both males and females. Therefore, the present study results provide differences according to gender in the epigenetic study of sarcopenia and provide an insight in the direction of further sarcopenia research.
EWGSOP: European Working Group on Sarcopenia in Older People
KOGES: Korean Genome and Epidemiology Study
MMI: Muscle mass index
DMRs: Differentially methylated regions
Ethics approval and consent to participate
This study was approved by the institutional review board (IRB) of Gyeongsang National University (approval number: GNUIRB-2019-04-010-013). Informed consent was obtained from all participants and all methods were performed in accordance with the relevant guidelines and regulations.
Consent for publication
Not applicable
Availability of data and material
The datasets used and/or analyzed during the current study are available from the corresponding authors on reasonable request.
Competing interests
The authors have no conflicts of interest to disclose.
Funding
This study was conducted with support from the Rural Development Administration (PJ014155052019).
Authors' contributions
Sangyeob Lee, Jeong-An Gim, and Jun-Il Yoo contributed to the study concept and survey design. Seung Chan Kim and Kyung-Wan Baek helped with data management and cleaning. Seung Chan Kim performed statistical analysis and interpretation. Sangyeob Lee and Jun-Il Yoo helped with clinical perspectives. Sangyeob Lee, Jeong-An Gim, Jun-Il Yoo contribute to the preparation of the manuscript. Jun-Il Yoo helped with critical revision and feedback.
Acknowledgments
Not applicable