Causal effect of gut microbiota on DNA methylation phenotypic age acceleration: a two-sample Mendelian randomization study

The causal relationship between gut microbiota and DNA methylation phenotypic age acceleration remains unclear. This study aims to examine the causal effect of gut microbiota on the acceleration of DNA methylation phenotypic age using Mendelian randomization. A total of 212 gut microbiota were included in this study, and their 16S rRNA sequencing data were obtained from the Genome-wide Association Study (GWAS) database. The GWAS data corresponding to DNA methylation phenotypic age acceleration were selected as the outcome variable. Two-sample Mendelian randomization (TSMR) was conducted using R software. During the analysis process, careful consideration was given to address potential biases arising from linkage disequilibrium and weak instrumental variables. The results from inverse-variance weighting (IVW) analysis revealed significant associations (P < 0.05) between single nucleotide polymorphisms (SNPs) corresponding to 16 gut microbiota species and DNA methylation phenotypic age acceleration. Out of the total, 12 gut microbiota species exhibited consistent and robust causal effects. Among them, 7 displayed a significant positive correlation with the outcome while 5 species showed a significant negative correlation with the outcome. This study utilized Mendelian randomization to unravel the intricate causal effects of various gut microbiota species on DNA methylation phenotypic age acceleration.

The human gastrointestinal tract harbors an immensely large microbial community, encompassing an estimated range of 1000-1150 bacterial species collectively known as the gut microbiota (GM) 1 .In recent years, gut microbiota has emerged as a prominent focus of medical research and has been substantiated to be intricately associated with immune function, metabolism, and the development of various diseases 2,3 .The composition of GM undergoes dynamic changes from infancy to adulthood and throughout the aging process in human individuals 4 .Therefore, gaining a comprehensive understanding of the profound association between gut microbiota and individual aging holds significant importance in the realms of anti-aging interventions and the prevention of age-related diseases.
The epigenetic clock, developed by Levin et al. 5 in 2018, is a tool that utilizes gene methylation patterns to infer an individual's biological age.Existing research has demonstrated that biological age, as measured by the epigenetic clock, outperforms chronological age in assessing an individual's true aging status and predicting their lifespan 6 .DNA methylation phenotypic age acceleration, also known as epigenetic clock acceleration, serves as a biomarker reflecting an individual's aging status.It is commonly used to refer to the difference between an individual's biological age and chronological age 7 .Currently, it has been firmly established that GM is significantly associated with organismal aging and several age-related diseases [8][9][10] .However, the causal effects of GM on DNA methylation phenotypic age acceleration remain unclear.
Mendelian randomization (MR) is an epidemiological research method that utilizes genetic variation as instrumental variables (IVs) to investigate the causal associations between exposures and outcomes 11 .MR has gained widespread application in medical research in recent years due to its ability to mitigate confounding factors and reverse causality, which are often encountered in traditional epidemiological studies.Moreover, MR offers advantages such as cost-effectiveness, efficiency, and increased control over variables compared to randomized controlled trials (RCTs) 12,13 .This study aims to investigate the causal impact of gut microbiota on DNA methylation phenotypic age acceleration using the analysis method of Mendelian randomization.

Data sources and software preparation
The 212 gut microbiota datasets utilized in this study were obtained from GWAS database (http:// gwas.mrcieu.ac.uk/ datas ets/).These data were derived from the sequencing information of the 16S rRNA of the gut microbiota in a total of 18,340 samples across 24 cohorts, as conducted by Kurilshikov et al 14 .The data on DNA methylation phenotypic age acceleration originates from a GWAS dataset based on epigenetic aging, specifically encoded as ebi-a-GCST90014292 15 .The data analysis was performed using the R software (version 4.1.3)and the TwoSa-mpleMR package.

Selection of instrumental variables for GM
The SNP selection for the gut microbiota followed the criterion of P < 1*10 -5 .To ensure adherence to Mendel's Second Law of Independent Assortment, which represents the principle of free combination, in this study, the criteria for linkage disequilibrium were set as follows: r 2 < 0.001 and kb > 10,000.The SNPs filtered based on the aforementioned criteria will be used as IVs for subsequent analysis.

Exclusion of weak instruments
To ensure the accuracy of the study results and the validity of the instrumental variable assumptions for Mendelian randomization, weak IVs were identified and removed based on the calculated F-statistic.The F-statistic is a statistical measure that reflects the strength of the association between IVs and the exposure factor.It is calculated using the formula: F = (β/SE) 2 .SNPs are regarded as weak IVs and are excluded from the analysis if F < 10.

Data analysis
Mendelian randomization analysis was conducted using the TwoSampleMR package in R software, employing the inverse-variance weighting (IVW) method.The regression results of IVW are used to determine whether there is a significant causal effect and directionality between the exposure and the outcome.To mitigate potential biases introduced by individual statistical methods, the weighted median (WM) and MR-Egger were employed as complementary analyses to the IVW.If the directions of the regression coefficients in the three aforementioned analyses are inconsistent, it indicates an unstable causal effect.In this step of the analysis, if the P-value of the IVW < 0.05, it is considered as a significant association between the exposure and the outcome.

Heterogeneity and horizontal pleiotropy testing
In this study, Cochran's Q test and MR-Egger intercept test were employed to assess the heterogeneity and horizontal pleiotropy of the results, respectively.Heterogeneity represents the variability of causal effect estimates among each SNP.If heterogeneity is significant, it suggests an unstable causal effect between the exposure and the outcome.On the other hand, horizontal pleiotropy refers to the possibility of SNPs influencing the outcome through factors other than the exposure.If horizontal pleiotropy is significant, it indicates a violation of the exclusivity assumption in MR analysis.In both of the aforementioned tests, P-value < 0.05 is considered statistically significant.

Ethics approval and consent to participate
This analysis of publicly available data does not require ethical approval.

Data and detailed information
The GM data obtained from the GWAS database is used as the exposure variable, while DNA methylation phenotypic age acceleration is considered the outcome variable in this study.The GM data consists of a total of 212 sub-datasets, which cannot be fully presented in Table 1.For detailed information regarding the GM data, please refer to Supplementary Material 1 (Suppl.1).

IVs selection, linkage disequilibrium, and weak IVs exclusion
The following criteria were used to screen for SNPs: P < 1*10 -5 ; linkage disequilibrium parameters: r 2 < 0.001 and kb > 10,000.IVs with an F-statistic < 10 were excluded from the analysis.The F-statistics of each SNP and more detailed information can be found in Supplementary Material 2 (Suppl.2), with F-statistics ranging from 14.90981 to 35.41665 for all SNPs.

Heterogeneity and horizontal pleiotropy tests
Heterogeneity and horizontal pleiotropy tests were conducted using Cochran's Q test and MR-Egger intercept test, respectively.If there is significant heterogeneity and horizontal pleiotropy (P < 0.05) observed for the SNPs corresponding to the GM in relation to the outcome, it indicates that the causal effect is not established.A total of 12 GM species showed relatively stable causal effects and passed the heterogeneity and horizontal pleiotropy tests  www.nature.com/scientificreports/(Table 3).As shown in Figs. 2 and 3, among the aforementioned 12 GM species, 7 species exhibited a significant positive correlation with the outcome (Fig. 2), while 5 species showed a significant negative correlation with the outcome (Fig. 3). Figure 4 presents a heatmap, generated using the complexheatmap package in R (version 4.1.3),displaying the sorted β-values of the 12 GM species in relation to DNA methylation phenotypic age acceleration.

Discussion
Numerous studies have demonstrated a strong correlation between changes in GM and the aging process.However, there is a lack of definitive conclusions regarding the causal relationship between the two in the majority of these studies 16 .In the GM of healthy adults, the abundance of Bacteroidetes and Firmicutes phyla is generally higher, while Actinobacteria and Proteobacteria phyla are comparatively less abundant 8 , 17 .During the aging process, there is a gradual increase in the abundance of Bacteroidetes and Proteobacteria phyla, while the diversity of the gut microbiota tends to decrease 18 .Due to the complexity and diversity of the GM itself, identifying specific GM alterations associated with aging becomes challenging.Hence, this study aimed to explore the causal relationship between the GM and DNA methylation phenotypic age acceleration using MR analysis approach from the perspective of epigenetic clocks.
In this study, we identified seven specific GM species that exhibited significant positive causal effects on DNA methylation phenotypic age acceleration.These species were Christensenellaceae, Peptococcaceae, Dorea, Lachnospiraceae UCG001, Lachnospiraceae UCG008, Tyzzerella3 and Firmicutes.A total of five GM species exhibited significant negative causal effects on DNA methylation phenotypic age acceleration.These species were Coriobacteriia, Coriobacteriaceae, Lactobacillus, Ruminococcus torques group and Coriobacteriales.Among them, Coriobacteriia, Coriobacteriaceae, and Coriobacteriales belong to the phylum Actinobacteria, while Lactobacillus and Ruminococcus torques group belong to the phylum Firmicutes.Therefore, based on the results of this study, it can be concluded that the causal effects of GM on DNA methylation phenotypic age acceleration are complex.In 2022, Kumar et al. 19 utilized D-galactose to establish an animal model for accelerated aging and intervened by administering Lactobacillus through dietary supplementation.The findings revealed that Lactobacillus exhibited antioxidant potential in ameliorating the accelerated aging model.These results are consistent with our research, demonstrating a negative causal effect between the abundance of Lactobacillus and DNA methylation phenotypic age acceleration.In 2023, Liu et al. 20 investigated the relationship between GM and longevity using Mendelian randomization.The results demonstrated a positive correlation between the abundance of Coriobacteriaceae and increased odds of longevity, which is consistent with the findings of our study.
In recent years, researchers have put forth the idea of fecal microbiota transplantation (FMT) in the context of longevity.FMT involves the transfer of gut microbiota from healthy and long-lived individuals to patients, aiming to potentially delay aging and promote longevity 21,22 .Although some studies have suggested a potential association between GM and the aging process [23][24][25][26] , the consistency of results among different studies is relatively poor.Therefore, further research is needed to comprehensively understand the impact of different GM species on the epigenetic clock.The findings of this study suggest that the causal effects of GM on the aging process are intricate, with different microbial taxa exerting distinct influences on DNA methylation phenotypic age acceleration.In conclusion, the findings of this study provide valuable insights for the clinical application of FMT and personalized treatments.
This study has several limitations.Firstly, the majority of the GWAS data used in this study were based on European individuals, which may introduce geographical and ethnic biases; Secondly, this study did not investigate the specific mechanisms through which different GM species influence the epigenetic clock.Further exploration is required using larger sample size and laboratory data to address these limitations.

Conclusion
This study utilized Mendelian randomization to uncover the complex causal effects of different gut microbiota species on DNA methylation phenotypic age acceleration.

Figure 3 .
Figure 3. Scatter plots of 5 gut microbiota with negative direction.Footnote for Fig. 2 and Fig. 3: The positive or negative slope of the straight lines represents the direction of the causal effect, and three different colors are used to represent three different methods: light blue for IVW (Inverse Variance Weighted), dark blue for MR-Egger, and green for Weighted Median.

Figure 4 .
Figure 4. Heat map shows the beta ranking of 12 gut microbiota.Footnote: In the heatmap, the colors and shades represent the causal direction and effect size of the gut microbiota on the acceleration of the epigenetic clock, respectively.The dendrogram in the figure represents an unsupervised clustering analysis.

Table 1 .
Detailed information of the dataset used in this study.As shown in Fig.1, the analysis results of IVW indicate a potential causal association (P < 0.05) between the abundance of 16 GM species and DNA methylation phenotypic age acceleration.If the regression coefficients' directions from the Weighted median, MR-Egger and IVW methods are not consistent, it is considered an indication of an unstable causal effect and should be excluded from the final results.The detailed results of IVW and MR-Egger for the 16 GM species, including P-values and beta values are recorded in Table2.Supplementary Material 3 (Suppl.3) provides more detailed information about the 16 GM species.

Table 2 .
Detailed analysis results of IVW and MR-Egger of 16 gut microbiota.ID.

Table 3 .
Detailed information of 12 gut microbiota with causal effect on the outcome.Footnote: The β values here specifically denote the β of the IVW method.