The tissue-specific effects of glucose-lowering drug targets on aging mediated through DNA methylation: a multi-omics genetic study

The tissue-specific effects of glucose-lowering drug targets on aging mediated through DNA methylation: a multi-omics genetic study | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article The tissue-specific effects of glucose-lowering drug targets on aging mediated through DNA methylation: a multi-omics genetic study Yuqi Sun, Haonan Zheng, Min Ma, Rongrong Gu, Manqing Wang, Si Fang, and 4 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7798894/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 9 You are reading this latest preprint version Abstract Background DNA methylation plays a key role in mediating the anti-aging effects of glucose-lowering drugs. This study aims to systematically explore the potential anti-aging effects of target genes of FDA-approved glucose-lowering drugs and the underlying epigenetic mediators. Methods We conducted a two-sample Mendelian randomization (MR) study to investigate the putative causal relationships between the gene expression levels of glucose-lowering drug targets and 10 aging-related phenotypes, followed by a two-step MR to estimate the mediation effect of DNA methylation. Drug candidates were selected according to the latest review of clinical drug use for type 2 diabetes, and their target genes were obtained from the DGIdb database. Tissue-specific cis- expression quantitative trait loci (eQTLs) from GTEx consortium were selected as genetic instruments to proxy the expression level of drug-target genes. Glycemic phenotypes were used as positive controls to validate the instruments. The cis- and trans- methylation QTLs of Cytosine-phosphate-Guanine sites near the drug target genes were obtained from GoDMC consortium. Additionally, we performed enrichment analyses focused on tissue specificity and aging pathways to further corroborate our findings. Results We obtained 194 target genes interacted with 36 FDA-approved anti-diabetic drugs, of which the tissue-specific eQTLs were used to proxy the drug target effects. MR showed strong evidence that 9 interacting genes of 6 glucose-lowering drugs showed anti-aging potential on one or more aging-related phenotypes mediated by DNA methylation: EHMT2 , HSPA4 , IGF2BP2 , IRS1 , LPL , NDUFAF1 , NDUFS3 , SLC22A3 and TCF7L2 . These genes were distributed in 17 tissues, especially in the central nervous system, suggesting a potential neural component in their anti-aging effects. For instance, expression of EHMT2 in several brain basal ganglia regions, which the gene interacted with Tolazamide, showed a protective effect on frailty (odds ratio[OR] in caudate = 1.02, 95%CI = 1.01–1.04, FDR adjusted P = 1.69×10 − 2 ; OR in putamen = 1.02, 95%CI = 1.01–1.03, P FDR =3.37×10 − 2 , OR in nucleus accumbens = 1.02, 95%CI = 1.01–1.04, P FDR =3.37×10 − 2 ). These associations were externally validated by searching literature evidence in existing EWAS and TWAS studies, as well as evidence from enricment analyses. Conclusions This study prioritizes nine glucose-lowering genes as anti-aging drug targets in specific tissues and prioritizes their epigenetic regulation through DNA methylation for future drug development. Aging DNA methylation Anti-aging drugs Anti-diabetic drugs Mendelian randomization Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction Aging is a complicated process characterized by the progressive accumulation of molecular damage, where epigenetic alteration is considered one of the primary hallmarks and an unambiguous contributor [ 1 ]. Since the human DNA methylation (DNAm) landscape has been associated with plenty of aging-related phenotypes [ 2 ], specific therapeutic interventions on the epigenetic status yield tangible benefits for healthspan. Given the inconsistency of DNAm patterns across different human tissues, recent studies have emphasized the importance of developing tissue-specific interventions [ 3 ]. A retrospective review of randomized controlled trials (RCTs) summarized 8 categories of promising anti-aging compounds that act via various pathways, including epigenetic alteration [ 4 ]. Two of them are glucose-lowering drugs: metformin and glucagon-like peptide-1 receptor agonists, thus attracting people’s attention to the anti-aging potential of all types of glucose-lowering drugs. RCTs involving glucose-lowering drugs are usually designed to evaluate their effects on age-associated diseases rather than aging [ 5 ]. Those RCTs were designed to recruit high-risk participants (e.g., those with cardiovascular diseases or dementia), thus providing limited evidence of extending healthspan or decelerating aging in the general population [ 5 ]. A small RCT, Metformin in Longevity Study (N = 14), provided preliminary evidence that metformin can lead to transcriptomic changes in pathways affecting aging in skeletal muscle and subcutaneous adipose tissues [ 6 ]. An ongoing RCT, Targeting Aging with Metformin [ 7 ], aims to explore the effect of metformin among 3,000 ethnically diverse, non-diabetic participants aged 65–80, but the results have not been released yet. Besides, only a few clinical trials explored the beneficial effects of other glucose-lowering agents on aging, including acarbose (N = 10, NCT02953093), canagliflozin (N = 30, NCT06301529), and dapagliflozin (N = 20, NCT04401904), all of which have very limited sample sizes. Given the limitations of conventional observational studies and inconclusive results from RCTs, human genetics could help shed new light on the anti-aging nature of glucose-lowering drug targets, which may further facilitate the development of new lifespan-extending medications [ 8 ]. Previous Mendelian randomization (MR) studies indicated potentially protective effects of targets for metformin, sodium-glucose cotransporter 2 inhibitors, and sulfonylureas on aging-related diseases, including cardiovascular diseases, cancer, dementia, inflammation, and heart failure [ 9 – 12 ]. In addition, metformin might promote healthy aging via targeting GPD1 and AMPKγ2 ( PRKAG2 ) [ 10 ]. Nevertheless, it remains unclear whether glucose-lowering drugs protect against aging itself rather than aging-related diseases, and whether they function through epigenetic alterations. The aim of this study is to explore the target-specific effects of glucose-lowering drug targets on aging biomarkers using two-sample MR. To provide mechanistic insights, we further used two-step MR to investigate whether the effect of glucose-lowering drug targets on aging is mediated through DNAm. Methods Study design Figure 1 presents the study design and workflow, including a schematic representation detailing the research procedure adopted and a graphical abstract outlining the data resource and core assumptions of the study. Selection of genetic instruments for glucose-lowering drug targets This study considered nine major categories of glucose-lowering drugs used in first-line clinical practice [ 13 ]: Glucagon-Like peptide-1 receptor agonists, biguanides, sulfonylureas, meglitinides, α-glucosidase inhibitors, dipeptidyl peptidase-4 inhibitors, sodium-glucose cotransporter 2 inhibitors, thiazolidinediones, and Lck/Yes-related Novel protein tyrosine kinase activators. A total of 36 FDA-approved drugs were selected as candidates. Since the mechanism by which these drugs act is not fully understood, we searched the listed glucose-lowering drugs in the Drug-Gene Interaction database (DGIdb) v.5.0.7 [ 14 ], and reserved 194 related genes as potential targets. Using information implemented in DrugBank, those gene targets with clear mechanisms of action were prioritized as Tier 1 druggable genes, where those genes exhibiting drug-gene interactions were classified as Tier 2 (electronic supplementary material [ESM] Table 1). We obtained tissue-specific cis- acting expression quantitative trait loci ( cis- eQTLs) of each glucose-lowering drug target gene from Genotype-Tissue Expression (GTEx) v8 Project (15,201 samples of 49 tissues from 838 donors), and used them to select tissue-specific genetic instruments [ 15 ]. A stringent selection process was performed to identify tissue-specific genetic variants for each target, requiring a SNP a) being associated with expression levels of corresponding genes ( P < 5 × 10 − 8 ) in the cis -acting regions (i.e. ±500 kb window of each gene probe); b) passing a stringent linkage disequilibrium (LD) clumping test using the panel of European ancestry with clumping r 2 of less than 0.001 and clumping window of 10,000 kb; c) having an F-statistic > 10, which indicating good instrument strength; d) having minor allele frequency ≥ 0.01. We also applied Steiger filtering to identify the genetic variants explaining more variations of the exposures than the outcomes, thereby addressing potential reverse causality [ 16 ]. In addition, we considered five glycemic traits: Hemoglobin A1c, fasting glucose, fasting insulin, 2-hour glucose, and type 2 diabetes (T2D) as positive controls to validate the reliability of the instruments. The associations of genetic susceptibility to the targets with glycemic traits were estimated using two-sample MR. Only targets relevant to at least one of these traits (false discovery rate, FDR < 0.05) were reserved for further analyses. Selection of aging proxy indicators A variety of aging biomarkers and phenotypes have been proposed for the identification and evaluation of longevity interventions [ 17 ]. Referring to the latest aging biomarker consensus and previous MR studies about aging [ 17 , 18 ], we selected ten phenotypes as aging proxy indicators: father’s age at death, mother’s age at death, combined parental age at death, Hannum Age, Intrinsic Horvath Age, PhenoAge, Grimage, leukocyte telomere length, healthspan, and frailty. The details of genome-wide association studies (GWAS) on each phenotype are described in ESM Method 1. The main characteristics of each GWAS are presented in Supplementary Table 1. Selection of DNAm mediators The GoDMC study published a database of > 270,000 independent DNAm quantitative trait loci (mQTLs), involving 32,851 European participants [ 19 ]. Putative genetic variants associated with DNAm at 420,509 CpG sites were identified from blood samples, with genotype data imputed to the 1000 Genomes Project reference panel [ 20 ], and DNAm profiles quantified from bisulfite-converted genomic whole-blood DNA using the Infinium HumanMethylation BeadChip (HumanMethylation450 or EPIC arrays). We localized the genome-wide DNAm datasets and the genetic location of all CpG sites available from GoDMC to establish a retrieval library, using a standardized pipeline to convert all data to hg19/GRCh37 build to ensure consistency. For genetic instruments of CpG sites, we used the same standard applied in eQTL instrument selection. CpG sites associated with expression levels of the drug-related genes as identified by MR results, and located within ± 500kb of the gene, are considered cis -regulated by drug-related targets; otherwise, they are classified as trans -regulated. Selection of positive controls To ensure our findings reflect specific pharmacological effects rather than non-specific processes such as drug metabolism, we restricted our analyses to target genes that exhibited a significant impact on glycemic traits. GWAS of five indexes which reflect the change of blood glucose are used as outcomes in positive controls to select valid instruments for drug targets, including HbA1c level (N = 146,806), fasting glucose level (N = 200,622), fasting insulin level (N = 151,013), 2-hour glucose level (N = 63,396) from MAGIC consortium [ 21 ] and T2D (N = 1,812,017, N cases = 242,283, N controls = 1,569,734) from the T2DGGI study [ 22 ]. Participants were of European descent, and detailed information on these GWAS is summarized in Supplementary Table 2. Statistical analysis Genetic correlation analysis We used LD-score regression to calculate the genetic correlations and SNP heritability of the 10 aging traits [ 23 – 25 ], based on the European-ancestry LD reference panel. In line with recommendations [ 25 ], low-quality (INFO < 0.9) and rare (minor allele frequency < 0.01) SNPs, as well as those located in the Major Histocompatibility Complex, were discarded before merging the summary statistics. We further clustered those ten traits into four groups according to their genetic correlations. Mendelian randomization analyses We firstly conducted a two-sample MR to identify the glucose lowering drug target genes whose expression level causally affect the aging outcomes. A two-step MR was further conducted to assess whether DNAm could mediate the effect. The first step is to assess how each target gene’s expression level influences DNAm of CpG sites across the whole genome. The second step was carried out on those DNAm that passed the first step (MR P FDR < 0.05). For these DNAm, we evaluated whether they affect aging-related outcomes. Mediation analysis was then performed to identify the mediation proportion of the DNAm of CpG sites in each target gene expression and aging outcome pair. The product of coefficients method and the delta method were applied to calculate the indirect effect and the standard error of the indirect effect, respectively, for each of the mediator [ 26 ]. For all the MR analysis listed above, we utilized Wald’s ratio method for exposures with only one instrument, and the inverse variance weighted (IVW) method for exposures with two or more instruments [ 27 ]. To ensure the robustness of associations, we further performed sensitivity analyses using weighted median and weighted mode methods for exposures with three or more instruments. For instruments that are not available in the outcome GWAS data sets during the MR analysis, proxy genetic variants in high LD with this eQTL/mQTL ( r 2 > 0.8 in the 1000 Genomes data in the European population [ 20 ]) were used as substitutes. Validation of MR assumptions MR analyses require three key assumptions: (1) relevance: the genetic predictors have robust associations with the exposure; (2) exchangeability: there are no common causes (confounders) between the instrument and outcome; (3) exclusion restriction: the instruments are only related to the outcome through the exposure being studied [ 28 ]. Findings were reported according to the STROBE-MR (Strengthening the Reporting of Mendelian Randomization Studies) guidelines (ESM Method 2) [ 29 ]. To assess the relevance assumption, we examined the strength of each genetic variant using F-statistics. Only variants meet that this standard will be considered valid instrumental variables. The exchangeability assumption cannot be definitively proven, but associations with known confounders can be tested and demonstrated [ 30 ]. We used GWAS in European participants to avoid population mixture. These GWAS were adjusted for the top principal components or through the use of adjusted linear mixed models to minimize the confounding by population stratification [ 31 ]. We also conducted LD check analysis to evaluate the LD structures of causal genes or CpGs with robust MR evidence (FDR < 0.05) and mitigate bias resulting from confounding by LD [ 32 ]. We estimated the pairwise LD between each eQTL/mQTL and mQTL/aging-associated GWAS variant in the cis- region. We utilized the threshold of LD r 2 > 0.8 as evidence of LD checked for the whole study [ 32 ]. Only results that pass the LD check will be included in further analysis. To assess the exclusion restriction assumption, we applied two sensitivity tests: weighted median and weighted mode methods, to minimize potential bias and improve confidence in our causal interpretation when more than 3 instruments were available. Enrichment analyses We performed tissue-enrichment analyses for all prospective drug target genes identified through the MR analyses. Gene-level RNA-Seq expression values, measured in transcripts per million (TPM), were obtained from the GTEx V8 release. The tissue-specific genes are defined using the simplified algorithm from the Human Protein Atlas (ESM Method 2) [ 33 ]. We also queried Enrichr to identify the tissue-specific aging-associated expression variation and relevant aging pathways for the drug target genes [ 34 – 36 ]. Triangulation of evidence Public databases of epigenome-wide association study (EWAS) and transcriptome-wide association study (TWAS) were reviewed to support our discoveries. We searched the EWAS Catalog (retrieval version: data last updated on March 22, 2024) [ 37 ] on the EWAS Open platform, to explore DNAm associations for our drug target genes. We also searched the TWAS Atlas (retrieval version: data last updated on May 15, 2022) [ 38 ], to identify disease associations with gene transcription in whole blood and multiple tissues, collecting tissue-specific TWAS results to further support our findings. Results Summary of instrument and outcome selection We searched drug-gene interaction (DGI) for 36 glucose-lowering drugs (Fig. 2 , ESM Table 1). A total of 194 drug-related genes were selected as candidates; 160 of them were found in the DGI database with 1,619 corresponding cis -eQTLs from 49 types of human tissues identified as instruments (ESM Table 2). Among these, the expression of 23 genes showed a causal effect on at least one glycemic phenotype in tissue-specific conditions at the FDR threshold of P < 0.05 and were regarded as drug target genes for this study (Supplementary Fig. 1, ESM Table 3). By calculating genetic correlations (rg) and subsequently performing hierarchical clustering based on the magnitude of these correlations, we found that the 10 aging phenotypes from four distinctive clusters of high genetic similarity (|rg| ≥ 0.2; P < 5× 10 − 3 ) (Supplementary Fig. 2, ESM Table 4). Thus, we categorized them into four groups: telomere length, parental longevity, healthspan and epigenetic age acceleration (EAA), to represent aging from different aspects. Tissue-specific associations of gene expression levels of glucose-lowering drug targets with aging outcomes Through the two-sample MR, ten target genes ( EHMT2 , HSPA4 , IGF2BP2 , IRS1 , LPL , NDUFAF1 , NDUFS3 , SLC22A3 , TCF7L2 , and KCNJ11 ) of sixteen drugs (acarbose, acetohexamide, canagliflozin, chlorpropamide, dapagliflozin, gliclazide, glimepiride, glipizide, gliquidone, glyburide, metformin, nateglinide, pioglitazone, repaglinide, tolazamide and tolbutamide) displayed strong evidence for their tissue-specific expressions to be causally related to aging phenotypes (Fig. 3 , ESM Table 5). We found the expression levels of EHMT2 , HSPA4 , and NDUFS3 in multiple brain regions, including caudate basal ganglia, cerebellar hemisphere, nucleus accumbens basal ganglia, and putamen basal ganglia, were associated with frailty and leukocyte telomere length. LPL and SLC22A3 expression in the tibial nerve was associated with parental age (odds ratio [OR] = 1.06, 95%CI = 1.03–1.09, P FDR =3.36×10 − 3 ) and health span (OR = 1.07, 95%CI = 1.02–1.11, P FDR =3.37×10 − 2 ). In addition, SLC22A3 also exhibited a pan-tissue effect, which showed associations with healthspan in multiple tissues, including the left ventricle (OR = 1.05, 95%CI = 1.02–1.08, P FDR =3.22×10 − 2 ), liver (OR = 0.96, 95%CI = 0.95–0.98, P FDR =6.29×10 − 4 ), subcutaneous adipose tissue (OR = 1.06, 95%CI = 1.02–1.10, P FDR =6.29×10 − 4 ), ovary (OR = 1.03, 95%CI = 1.02–1.05, P FDR =1.03×10 − 2 ), and testis (OR = 1.06, 95%CI = 1.04–1.09, P FDR =1.42×10 − 4 ) in addition to tibial nerve. Mediation analysis In Step 1 of the two-step MR, associations of 3,526 pairs of target genes and CpG sites were tested, with 3,230 showing at least marginal MR evidence, including 2,064 cis- acting associations and 1,166 trans- acting associations (ESM Table 6, Fig. 4 ). In Step 2, the 5,742 genetic variants robustly associated with DNAm were used as instruments to proxy 1,158 CpG sites. We then tested the associations between DNAm of CpG sites and aging traits (ESM Tables 7 and 8, Supplementary Fig. 3). Eventually, nine target genes ( EHMT2 , HSPA4 , IGF2BP2 , IRS1 , LPL , NDUFAF1 , NDUFS3 , SLC22A3 and TCF7L2 ) of eight drugs: acarbose, canagliflozin, dapagliflozin, glimepiride, metformin, pioglitazone, repaglinide and tolazamide, were associated with aging outcomes mediated by DNAm of 1,058 CpG sites (ESM Table 9, Fig. 5 , Supplementary Fig. 4). 755 of 1,058 (71.4%) of the mediating DNAm are cis- regulated by their corresponding drug target genes, except EHMT2 exhibits both cis- and trans- regulated DNAm. The target genes mediated through DNAm distribute in 16 tissues of five systems, including nervous system (caudate, nucleus accumbens, putamen and cerebellar hemisphere and tibial nerve), cardiovascular system (artery aorta and left ventricle), immune and endocrine system (cultured fibroblasts, EBV-transformed lymphocytes, spleen, adipose tissues, liver and thyroid), respiratory system (lung) and reproductive system (testis and ovary). Triangulation of external evidence Results extracted from the previous EWAS study supported the mediation effect of relevant target genes and aging via epigenetic regulation. DNAm of five genes: EHMT2 , HSPA4 , IGF2BP2 , SLC22A3 , and TCF7L2 , showed associations with aging-related outcomes in EWAS, which in line with our findings (ESM Table 10). For TWAS findings reported in TWAS Atlas, only a limited number of studies reported associations with aging-related outcomes. Thus, the top eight aging-related diseases, as well as their common risk factors, were included to proxy aging, in accordance with the definition of healthspan phenotype [ 39 ]. All ten prioritized aging associated genes also showed relevant TWAS evidence: EHMT2 , HSPA4, IGF2BP2 , IRS1 , LPL , NDUFAF1 , NDUFS3 , SLC22A3 and TCF7L2 (ESM Table 11). Tissue enrichment analysis showed no evidence of preferential expression of drug target genes in tissues corresponding to MR results (Supplementary Table 3). However, results from Enrichment analysis from Enrichr indicated that these genes are members of several metabolism, mitochondrial functions, insulin/IGF-1 signaling pathways and other aging-associated pathways [ 40 , 41 ], with tissue-specific regulatory relevance of aging further supported by external evidence from GEO datasets [ 41 ]. Discussion This drug target genetic study comprehensively investigated the tissue-specific effect of glucose-lowering drug targets on a range of aging phenotypes mediating through DNAm. We extended the recent discoveries of anti-aging targets and identified DNAm at multiple CpG sites as potential mediators in the aging process, highlighting the importance of exploring epigenetic regulatory mechanisms in aging-related therapies. Our study showed potentials of eight relevant anti-diabetic drugs as prospective candidates to resist aging. The tissue-specific expression of nine drug-targeting or drug-interacting genes: EHMT2 , HSPA4 , IGF2BP2 , IRS1 , LPL , NDUFAF1 , NDUFS3 , SLC22A3 , and TCF7L2 as targets for future research on decelerating aging with a focus on their epigenetic regulations through DNAm. Among the eight anti-diabetic drugs, acarbose [ 42 ], canagliflozin [ 43 ], dapagliflozin [ 44 ] and metformin [ 45 , 46 ], and pioglitazone [ 47 ] have previously shown the potentials to extend lifespan or decrease the risk of several aging-related diseases in animal models. However, their evidence in humans was largely based on observational studies [ 6 , 48 – 51 ], leaving it unclear whether these associations are driven by residual confounding, reverse causality and immortal time bias [ 52 ]. Besides, most of these studies focused on the protective effects of the anti-diabetic drugs in high-risk populations from certain aging-related diseases (e.g. heart failure) rather than aging directly. Our results provided evidence to support the conservatism of the anti-aging effects of the relevant drug targets in the general population. Among the drugs we identified, glimepiride, repaglinide and tolazamide showed their potential of extending lifespan that were not reported before. These drugs could be considered as drug candidates for future clinical research. The tissue-specific effects of gene expression we observed align well with previous studies of age-related diseases. For instance, low-dose pioglitazone has been shown to extend lifespan and improve cognitive function in aged mice by modulating metabolic homeostasis [ 53 – 55 ], though no direct evidence has been reported in humans. In this study, we identified the mitochondrial protein HSPA4 , as a potential core gene in pioglitazone’s regulatory pathway of decelerating aging in brain tissues. Notably, HSPA4 expression has also been found to be downregulated in human front cortex as well as in several mouse brain tissues [ 41 ], consistent with the tissue specificity highlighted by our MR results. In addition, a previous study found that acarbose has a pronounced lifespan-extending effect in male mice [ 42 ]. Here, we found one of its target genes, LPL , and its expression in the tibial nerves is the only exposure related to parental longevity, particularly the father’s life span. Similarly, SLC22A3 expression in testis has been found to be related to combined parental age at death, which aligned with the gender difference in metformin effect on aging [ 46 ]. This may contribute as a starting point to further investigate the gender difference in the anti-aging effect of drugs. We uncovered potential anti-aging roles of target genes for two sulfonylurea drugs ( EHMT2 for tolazamide and IRS1 for glimepiride), and a meglitinide class drug ( IGF2BP2 for repaglinide). Specifically, EHMT2 is a well-recognized histone methyltransferase to have the ability of regulating DNAm [ 56 ]. It is also a member of the Senescence Associated Secretory Phenotype (SASP) pathway [ 56 ], and has been observed to be downregulated in aging mouse hippocampus. Results from model organisms suggest IRS1 is involved in the IGF1/GH1 axis that could regulate aging [ 57 ], with loss of expression extending lifespan in mice [ 58 ] and Drosophila [ 59 ]. IGF2BP2 has been identified as a key factor to activate stem cell aging [ 60 ] and brain aging [ 61 ] in mouse models, and also recognized as an epigenetic regulatory factor in epitranscriptomics [ 62 ]. This warrants further investigation to clarify its roles in human aging, including its involvement in the insulin/IGF signaling pathways and the crosstalk between RNA and DNA modification. Metformin has been shown in multiple studies to exert lifespan-extending effects [ 48 ]. Here we identified four novel target genes: NDUFS3 , NDUFAF1 , SLC22A3 , and TCF7L2 , mainly involved in mitochondrial and energy metabolism pathways [ 40 ], that may underlie Metformin’s novel anti-aging mechanisms and merit further investigation. Notably, three of them: EHMT2 , HSPA4 and NDUFS3 , showed a brain-specific effect on aging, which were mediated by methylation variation across multiple CpG sites. This underscores the key role of epigenetic changes in the connecting central nervous system regulation and systemic aging. The strengths of this study included the widely used pipelines for instrument selection and a high resolution of tissue-specific results. In addition, we included a comprehensive set of aging outcomes, a wide range of drug target selection, and an emphasis on epigenetic regulation of aging. We also integrated external multi-omics evidence to enhance the interpretations of our findings, aiming to provide more informative and predictive results compared with previous drug-target MR studies on aging and offering implications for future targeted drug development [ 11 , 18 , 63 , 64 ]. This study still has limitations. First, the eQTL data used in this study are derived from a relatively small sample size of 838 individuals, with only one instrument available for each target gene per tissue. In addition, solely relying on lead cis-eQTLs for target genes in each tissue may limit the detection of potential bias [ 30 , 65 ]. Second, some target genes, such as TCF7L2 , have qualified instruments in only a few tissues. This led to the incomplete exploration of the tissue-specific function of these genes. Third, the data for DNAm, the EAA group of aging indicators, and leukocyte telomere length were derived from blood samples, introducing bias due to sample incoherence. Fourth, for drugs like acarbose [ 42 ] and canagliflozin [ 43 ], the utilization of sex-stratified data, including sex-stratified eQTL data and sex-stratified GWASs of aging phenotype, may help elucidate gender-specific differences in epigenetic regulations [ 66 ]. Fifth, not all aging hallmarks listed in the consensus have publicly available GWAS datasets [ 1 ]. Selection bias is inevitable in our study, which may contribute to the results that no gene was found to affect the epigenetic age acceleration group. More GWAS meta-analyses are required for better investigation. Sixth, several CpG sites showed a mediation proportion exceeding 100%. One of the three following conditions accounts for the phenomenon: (1) there are other mediators with a negative portion mediated; (2) the DNAm of each CpG site affects one another; (3) there are interactions between the DNAm of each CpG site [ 67 , 68 ]. Thus, the observed indirect effect requires further investigation. Seventh, the intervention of these drugs could have varying effects depending on age groups, ethnic backgrounds, and disease states. We need to assess the specific effects of these drug-gene expressions across diverse populations to understand the generalizability of our results in the near future. Finally, MR can only proxy the drug target effect rather than the direct effect of the drug on aging. Therefore, the evidence provided by this study cannot be used for any clinical investigations before formal clinical trials. Conclusion In conclusion, our study identified 9 target genes of 6 glucose-lowering drugs showing robust effects on longevity through DNA methylation. We identified both previously recognized and novel candidate genes that may contribute to lifespan extension by modulating aging-associated pathways, such as SASP signaling, mitochondrial and lipid metabolism, insulin signaling, and epitranscriptomic regulation. Several effects were tissue-specific, particularly within Five drug interacting genes exhibited a tissue-specific enrichment in the nervous system, emphasizing the neural regulation in systemic aging. Previous studies provide strong support for these findings. Further clinical and experimental research is required to validate and explore the feasibility and effectiveness of targeting these genes and drugs for aging prevention. Abbreviations DNAm DNA methylation EAA Epigenetic age acceleration EWAS Epigenome-wide association study FDR False discovery rate eQTL Expression quantitative trait locus GWAS Genome-Wide Association Study LD Linkage disequilibrium mQTL DNAm quantitative trait locus MR Mendelian randomization RCT Randomized controlled trial TWAS Transcriptome-wide association study T2D Type 2 diabetes Declarations Availability of data and materials The GWAS summary statistics for the ten primary aging outcomes are available from the GWAS Catalog database (www.ebi.ac.uk/gwas/). The GWAS summary statistics for glycemic traits are contributed by MAGIC investigators (www.magicinvestigators.org) and T2DGGI study (http://www.diagram-consortium.org/downloads.html). The analytical script of the MR analyses conducted in this study is available via the GitHub repository of the ‘TwoSampleMR’ R package (github.com/MRCIEU/TwoSampleMR/). The analytical script of the genetic correlation analyses conducted in this study is available via the GitHub repository of the ‘ldsc’ R package (github.com/mglev1n/ldscr/). The code used during the current study is available upon reasonable request from the corresponding author. All the other data generated or analysed during this study are included in this published article and the ESM files. Competing Interest J.Z and Y.B.S is the editorial board member of BMC medicine. Other authors declare that they have no competing interests. Funding This study was supported by grants from the Noncommunicable Chronic Diseases-National Science and Technology Major Project (2024ZD0531500, 2024ZD0531502), Students’ Innovation Training Program, Shanghai Jiao Tong University, School of Medicine, China (1824955Y) and grants from the National Key Research and Development Program of China (2022YFC2505203). Y.Q.S is supported by Roche University Research Fund, Chinese Society of Biochemistry and Molecular Biology, China (20240513). Z.J. and Q.Y. are supported by the National Natural Science Foundation of China (32570728 and 32500519). Y.B. is supported by the Shanghai Municipal Education Commission–Gaofeng Clinical Medicine Grant Support (20161307 and 20152508 Round 2). Contribution Statements Y.Q.S., H.Z. and J.Z. conceived and designed the study. Y.Q.S drafted the manuscript. H.Z performed data analysis. M.M., S.F., Y.B.S., Q.Y. and J.Z. contributed to reviewing and editing the manuscript. Y.Q.S. and H.Z. contributed to visualization. Y.Q.S., H.Z., R.G. and M.W. contributed to data management and related paperwork. Y.Q.S. and H.Z. accessed and verified the data reported in the study. J.Z. is the guarantor of the findings. All the authors approved the final version to be published. Acknowledgments We thank all of the investigators and participants who provided these data to support this study. We thank Huiling Zhao, Haoyu Liu and Xueyan Wu for the help of device acquisition and instructions. Author’s relationship and activities Y.Q.S., H.Z., R.G. and M.W. are members of the College Innovation Training Program in Shanghai Jiao Tong University, School of Medicine. All authors declare that there are no relationships or activities that might bias, or be perceived to bias this work. References Lopez-Otin C, Blasco MA, Partridge L, Serrano M, Kroemer G. Hallmarks of aging: An expanding universe. Cell. 2023;186(2):243–78. Gladyshev VN, Kritchevsky SB, Clarke SG, Cuervo AM, Fiehn O, de Magalhaes JP, Mau T, Maes M, Moritz R, Niedernhofer LJ, et al. Molecular Damage in Aging. Nat Aging. 2021;1(12):1096–106. Seale K, Horvath S, Teschendorff A, Eynon N, Voisin S. Making sense of the ageing methylome. Nat Rev Genet. 2022;23(10):585–605. Dor Y, Cedar H. Principles of DNA methylation and their implications for biology and medicine. Lancet. 2018;392(10149):777–86. Guarente L, Sinclair DA, Kroemer G. Human trials exploring anti-aging medicines. Cell Metab. 2024;36(2):354–76. Kulkarni AS, Brutsaert EF, Anghel V, Zhang K, Bloomgarden N, Pollak M, Mar JC, Hawkins M, Crandall JP, Barzilai N. Metformin regulates metabolic and nonmetabolic pathways in skeletal muscle and subcutaneous adipose tissues of older adults. Aging Cell 2018, 17(2). Padki MM, Stambler I. Targeting Aging with Metformin (TAME). In: Encyclopedia of Gerontology and Population Aging. edn. Edited by Gu D, Dupre ME. Cham: Springer International Publishing; 2021: 4908–4910. Ochoa D, Karim M, Ghoussaini M, Hulcoop DG, McDonagh EM, Dunham I. Human genetics evidence supports two-thirds of the 2021 FDA-approved drugs. Nat Rev Drug Discov. 2022;21(8):551. Zheng J, Xu M, Walker V, Yuan J, Korologou-Linden R, Robinson J, Huang P, Burgess S, Au Yeung SL, Luo S, et al. Evaluating the efficacy and mechanism of metformin targets on reducing Alzheimer's disease risk in the general population: a Mendelian randomisation study. Diabetologia. 2022;65(10):1664–75. Luo S, Wong ICK, Chui CSL, Zheng J, Huang Y, Schooling CM, Yeung SLA. Effects of putative metformin targets on phenotypic age and leukocyte telomere length: a mendelian randomisation study using data from the UK Biobank. Lancet Healthy Longev. 2023;4(7):e337–44. Chen Z, Wu X, Yang Q, Zhao H, Ying H, Liu H, Wang C, Zheng R, Lin H, Wang S, et al. The Effect of SGLT2 Inhibition on Brain-related Phenotypes and Aging: A Drug Target Mendelian Randomization Study. J Clin Endocrinol Metab. 2025;110(4):1096–104. Guo W, Zhao L, Huang W, Chen J, Zhong T, Yan S, Hu W, Zeng F, Peng C, Yan H. Sodium-glucose cotransporter 2 inhibitors, inflammation, and heart failure: a two-sample Mendelian randomization study. Cardiovasc Diabetol. 2024;23(1):118. DeMarsilis A, Reddy N, Boutari C, Filippaios A, Sternthal E, Katsiki N, Mantzoros C. Pharmacotherapy of type 2 diabetes: An update and future directions. Metabolism. 2022;137:155332. Freshour SL, Kiwala S, Cotto KC, Coffman AC, McMichael JF, Song JJ, Griffith M, Griffith OL, Wagner AH. Integration of the Drug-Gene Interaction Database (DGIdb 4.0) with open crowdsource efforts. Nucleic Acids Res. 2021;49(D1):D1144–51. Consortium GT. The GTEx Consortium atlas of genetic regulatory effects across human tissues. Science. 2020;369(6509):1318–30. Hemani G, Tilling K, Davey Smith G. Orienting the causal relationship between imprecisely measured traits using GWAS summary data. PLoS Genet. 2017;13(11):e1007081. Moqri M, Herzog C, Poganik JR, Biomarkers of Aging C, Justice J, Belsky DW, Higgins-Chen A, Moskalev A, Fuellen G, Cohen AA, et al. Biomarkers of aging for the identification and evaluation of longevity interventions. Cell. 2023;186(18):3758–75. Mao R, Li J, Xiao W. Identification of prospective aging drug targets via Mendelian randomization analysis. Aging Cell. 2024;23(7):e14171. Min JL, Hemani G, Hannon E, Dekkers KF, Castillo-Fernandez J, Luijk R, Carnero-Montoro E, Lawson DJ, Burrows K, Suderman M, et al. Genomic and phenotypic insights from an atlas of genetic effects on DNA methylation. Nat Genet. 2021;53(9):1311–21. Genomes Project C, Abecasis GR, Auton A, Brooks LD, DePristo MA, Durbin RM, Handsaker RE, Kang HM, Marth GT, McVean GA. An integrated map of genetic variation from 1,092 human genomes. Nature. 2012;491(7422):56–65. Chen J, Spracklen CN, Marenne G, Varshney A, Corbin LJ, Luan J, Willems SM, Wu Y, Zhang X, Horikoshi M, et al. The trans-ancestral genomic architecture of glycemic traits. Nat Genet. 2021;53(6):840–60. Suzuki K, Hatzikotoulas K, Southam L, Taylor HJ, Yin X, Lorenz KM, Mandla R, Huerta-Chagoya A, Melloni GEM, Kanoni S, et al. Genetic drivers of heterogeneity in type 2 diabetes pathophysiology. Nature. 2024;627(8003):347–57. Bulik-Sullivan B, Finucane HK, Anttila V, Gusev A, Day FR, Loh PR, ReproGen C, Psychiatric Genomics C, Genetic Consortium for Anorexia Nervosa of the Wellcome Trust Case Control, Duncan C. An atlas of genetic correlations across human diseases and traits. Nat Genet. 2015;47(11):1236–41. Bulik-Sullivan BK, Loh PR, Finucane HK, Ripke S, Yang J, Schizophrenia Working Group of the Psychiatric, Genomics C, Patterson N, Daly MJ, Price AL, Neale BM. LD Score regression distinguishes confounding from polygenicity in genome-wide association studies. Nat Genet. 2015;47(3):291–5. Finucane HK, Bulik-Sullivan B, Gusev A, Trynka G, Reshef Y, Loh PR, Anttila V, Xu H, Zang C, Farh K, et al. Partitioning heritability by functional annotation using genome-wide association summary statistics. Nat Genet. 2015;47(11):1228–35. Vabistsevits M, Davey Smith G, Sanderson E, Richardson TG, Lloyd-Lewis B, Richmond RC. Deciphering how early life adiposity influences breast cancer risk using Mendelian randomization. Commun Biol. 2022;5(1):337. Burgess S, Zuber V, Valdes-Marquez E, Sun BB, Hopewell JC. Mendelian randomization with fine-mapped genetic data: Choosing from large numbers of correlated instrumental variables. Genet Epidemiol. 2017;41(8):714–25. Zheng J, Baird D, Borges MC, Bowden J, Hemani G, Haycock P, Evans DM, Smith GD. Recent Developments in Mendelian Randomization Studies. Curr Epidemiol Rep. 2017;4(4):330–45. Skrivankova VW, Richmond RC, Woolf BAR, Yarmolinsky J, Davies NM, Swanson SA, VanderWeele TJ, Higgins JPT, Timpson NJ, Dimou N, et al. Strengthening the Reporting of Observational Studies in Epidemiology Using Mendelian Randomization: The STROBE-MR Statement. JAMA. 2021;326(16):1614–21. Sanderson E, Glymour MM, Holmes MV, Kang H, Morrison J, Munafo MR, Palmer T, Schooling CM, Wallace C, Zhao Q et al. Mendelian randomization. Nat Rev Methods Primers 2022, 2. Brumpton B, Sanderson E, Heilbron K, Hartwig FP, Harrison S, Vie GA, Cho Y, Howe LD, Hughes A, Boomsma DI, et al. Avoiding dynastic, assortative mating, and population stratification biases in Mendelian randomization through within-family analyses. Nat Commun. 2020;11(1):3519. Zheng J, Haberland V, Baird D, Walker V, Haycock PC, Hurle MR, Gutteridge A, Erola P, Liu Y, Luo S, et al. Phenome-wide Mendelian randomization mapping the influence of the plasma proteome on complex diseases. Nat Genet. 2020;52(10):1122–31. Uhlen M, Fagerberg L, Hallstrom BM, Lindskog C, Oksvold P, Mardinoglu A, Sivertsson A, Kampf C, Sjostedt E, Asplund A et al. Proteomics. Tissue-based map of the human proteome. Science 2015, 347(6220):1260419. Chen EY, Tan CM, Kou Y, Duan Q, Wang Z, Meirelles GV, Clark NR. Ma'ayan A: Enrichr: interactive and collaborative HTML5 gene list enrichment analysis tool. BMC Bioinformatics. 2013;14:128. Kuleshov MV, Jones MR, Rouillard AD, Fernandez NF, Duan Q, Wang Z, Koplev S, Jenkins SL, Jagodnik KM, Lachmann A, et al. Enrichr: a comprehensive gene set enrichment analysis web server 2016 update. Nucleic Acids Res. 2016;44(W1):W90–97. Xie Z, Bailey A, Kuleshov MV, Clarke DJB, Evangelista JE, Jenkins SL, Lachmann A, Wojciechowicz ML, Kropiwnicki E, Jagodnik KM, et al. Gene Set Knowledge Discovery with Enrichr. Curr Protoc. 2021;1(3):e90. Battram T, Yousefi P, Crawford G, Prince C, Sheikhali Babaei M, Sharp G, Hatcher C, Vega-Salas MJ, Khodabakhsh S, Whitehurst O, et al. The EWAS Catalog: a database of epigenome-wide association studies. Wellcome Open Res. 2022;7:41. Lu M, Zhang Y, Yang F, Mai J, Gao Q, Xu X, Kang H, Hou L, Shang Y, Qain Q, et al. TWAS Atlas: a curated knowledgebase of transcriptome-wide association studies. Nucleic Acids Res. 2023;51(D1):D1179–87. Zenin A, Tsepilov Y, Sharapov S, Getmantsev E, Menshikov LI, Fedichev PO, Aulchenko Y. Identification of 12 genetic loci associated with human healthspan. Commun Biol. 2019;2:41. Martens M, Ammar A, Riutta A, Waagmeester A, Slenter DN, Hanspers K, Digles RAM, Lopes D, Ehrhart EN. WikiPathways: connecting communities. Nucleic Acids Res. 2021;49(D1):D613–21. Edgar R, Domrachev M, Lash AE. Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. Nucleic Acids Res. 2002;30(1):207–10. Harrison DE, Strong R, Alavez S, Astle CM, DiGiovanni J, Fernandez E, Flurkey K, Garratt M, Gelfond JAL, Javors MA, et al. Acarbose improves health and lifespan in aging HET3 mice. Aging Cell. 2019;18(2):e12898. Jiang E, Dinesh A, Jadhav S, Miller RA, Garcia GG. Canagliflozin shares common mTOR and MAPK signaling mechanisms with other lifespan extension treatments. Life Sci. 2023;328:121904. Shihab EM, Kadhim HM, Shahooth SS. Dapagliflozin mitigates oxidative stress, inflammatory, and histopathological markers of aging in mice. J Med Life. 2024;17(2):157–63. Ramsey KM, Mills KF, Satoh A, Imai S. Age-associated loss of Sirt1-mediated enhancement of glucose-stimulated insulin secretion in beta cell-specific Sirt1-overexpressing (BESTO) mice. Aging Cell. 2008;7(1):78–88. Martin-Montalvo A, Mercken EM, Mitchell SJ, Palacios HH, Mote PL, Scheibye-Knudsen M, Gomes AP, Ward TM, Minor RK, Blouin MJ, et al. Metformin improves healthspan and lifespan in mice. Nat Commun. 2013;4:2192. Jafari M, Khodayari B, Felgner J, Bussel II, Rose MR, Mueller LD. Pioglitazone: an anti-diabetic compound with anti-aging properties. Biogerontology. 2007;8(6):639–51. Mohammed I, Hollenberg MD, Ding H, Triggle CR. A Critical Review of the Evidence That Metformin Is a Putative Anti-Aging Drug That Enhances Healthspan and Extends Lifespan. Front Endocrinol (Lausanne). 2021;12:718942. Stevenson-Hoare J, Leonenko G, Escott-Price V. Comparison of long-term effects of metformin on longevity between people with type 2 diabetes and matched non-diabetic controls. BMC Public Health. 2023;23(1):804. La Grotta R, Frige C, Matacchione G, Olivieri F, de Candia P, Ceriello A, Prattichizzo F. Repurposing SGLT-2 Inhibitors to Target Aging: Available Evidence and Molecular Mechanisms. Int J Mol Sci 2022, 23(20). Yen CL, Wu CY, Tsai CY, Lee CC, Li YJ, Peng WS, Liu JR, Liu YC, Jenq CC, Yang HY, et al. Pioglitazone reduces cardiovascular events and dementia but increases bone fracture in elderly patients with type 2 diabetes mellitus: a national cohort study. Aging. 2023;15(7):2721–33. Suissa S. Immortal time bias in pharmaco-epidemiology. Am J Epidemiol. 2008;167(4):492–9. Shen D, Li H, Zhou R, Liu MJ, Yu H, Wu DF. Pioglitazone attenuates aging-related disorders in aged apolipoprotein E deficient mice. Exp Gerontol. 2018;102:101–8. Xu L, Ma X, Verma N, Perie L, Pendse J, Shamloo S, Marie Josephson A, Wang D, Qiu J, Guo M, et al. PPARgamma agonists delay age-associated metabolic disease and extend longevity. Aging Cell. 2020;19(11):e13267. Seok H, Lee M, Shin E, Yun MR, Lee YH, Moon JH, Kim E, Lee PH, Lee BW, Kang ES, et al. Low-dose pioglitazone can ameliorate learning and memory impairment in a mouse model of dementia by increasing LRP1 expression in the hippocampus. Sci Rep. 2019;9(1):4414. Tachibana M, Sugimoto K, Fukushima T, Shinkai Y. Set domain-containing protein, G9a, is a novel lysine-preferring mammalian histone methyltransferase with hyperactivity and specific selectivity to lysines 9 and 27 of histone H3. J Biol Chem. 2001;276(27):25309–17. de Magalhaes JP, Abidi Z, Dos Santos GA, Avelar RA, Barardo D, Chatsirisupachai K, Clark P, De-Souza EA, Johnson EJ, Lopes I, et al. Human Ageing Genomic Resources: updates on key databases in ageing research. Nucleic Acids Res. 2024;52(D1):D900–8. Selman C, Lingard S, Choudhury AI, Batterham RL, Claret M, Clements M, Ramadani F, Okkenhaug K, Schuster E, Blanc E, et al. Evidence for lifespan extension and delayed age-related biomarkers in insulin receptor substrate 1 null mice. FASEB J. 2008;22(3):807–18. Clancy DJ, Gems D, Harshman LG, Oldham S, Stocker H, Hafen E, Leevers SJ, Partridge L. Extension of life-span by loss of CHICO, a Drosophila insulin receptor substrate protein. Science. 2001;292(5514):104–6. Suo M, Rommelfanger MK, Chen Y, Amro EM, Han B, Chen Z, Szafranski K, Chakkarappan SR, Boehm BO, MacLean AL, et al. Age-dependent effects of Igf2bp2 on gene regulation, function, and aging of hematopoietic stem cells in mice. Blood. 2022;139(17):2653–65. Li M, Guo H, Carey M, Huang C. Transcriptional and epigenetic dysregulation impairs generation of proliferative neural stem and progenitor cells during brain aging. Nat Aging. 2024;4(1):62–79. Huang H, Weng H, Sun W, Qin X, Shi H, Wu H, Zhao BS, Mesquita A, Liu C, Yuan CL, et al. Recognition of RNA N(6)-methyladenosine by IGF2BP proteins enhances mRNA stability and translation. Nat Cell Biol. 2018;20(3):285–95. Chen H, Tang X, Su W, Li S, Yang R, Cheng H, Zhang G, Zhou X. Causal effects of lipid-lowering therapies on aging-related outcomes and risk of cancers: a drug-target Mendelian randomization study. Aging. 2023;15(24):15228–42. Martin L, Boutwell BB, Messerlian C, Adams CD. Mendelian randomization reveals apolipoprotein B shortens healthspan and possibly increases risk for Alzheimer's disease. Commun Biol. 2024;7(1):230. Palmer TM, Lawlor DA, Harbord RM, Sheehan NA, Tobias JH, Timpson NJ, Davey Smith G, Sterne JA. Using multiple genetic variants as instrumental variables for modifiable risk factors. Stat Methods Med Res. 2012;21(3):223–42. Oliva M, Munoz-Aguirre M, Kim-Hellmuth S, Wucher V, Gewirtz ADH, Cotter DJ, Parsana P, Kasela S, Balliu B, Vinuela A et al. The impact of sex on gene expression across human tissues. Science 2020, 369(6509). VanderWeele TJ. Mediation Analysis: A Practitioner's Guide. Annu Rev Public Health. 2016;37:17–32. VanderWeele TJ, Vansteelandt S. Mediation Analysis with Multiple Mediators. Epidemiol Methods. 2014;2(1):95–115. Additional Declarations Competing interest reported. Jie Zheng and Yangbo Sun currently serve as members of the Editorial Board of BMC Medicine Supplementary Files ESM.docx ESMTables.xlsx SupplementaryTable1.xlsx SupplementaryTable2.xlsx SupplementaryTable3.xlsx SupplementaryFigure1.pdf Supplementary Figure 1 Forest plot of the top MR results: the effects of drug target gene expression levels on glycemic phenotypes. Outcomes include 2-hour glucose (A), HbA1c (B), fasting glucose (C), fasting insulin (D), and type 2 diabetes (E). The x-axis represents the change in the outcome per standard deviation (SD) increase in the expression level of each drug target gene. Statistical significance was defined as P FDR < 0.05. Detailed MR estimates are provided in ESM Table 3. SupplementaryFigure2.pdf Supplementary Figure 2 Genetic correlations between ten aging outcomes. Diagonal values show the observed-scale SNP heritability of each aging-related phenotype and off-diagonal circles show the genetic correlation among these traits. Blank squares are those did not pass multiple testing correction threshold (FDR > 0.05). The bottom dendrogram shows the hierarchical relationship between traits, based on the magnitude of their genetic correlations. The black box highlights the 4 clusters of aging-related phenotypes used in follow-up analyses: telomere length, parental longevity, healthspan and epigenetic age acceleration (EAA). HR, hazard ratio. Detailed results are provided in ESM Table 4. SupplementaryFigure3.pdf Supplementary Figure 3 Genomic distribution of mQTL and the target genes. The bar chart (A) on the top represents the position of target gene candidates and the count of CpG sites associated with them. The scatter plot (B) below represents the genomic positions of each mQTL and its corresponding CpG site. Detailed results are provided in ESM Table 6. SupplementaryFigure4.pdf Supplementary Figure 4 Forest plot of all MR results of the expression level of drug target genes on ten aging phenotypes, including combined parental age at death (A), father’s age at death (B), frailty (C), healthspan (D), leukocyte telomere length (E). The horizontal axis represents the effect size of exposure on outcomes. HR, hazard ratio. Statistical significance was defined as P FDR < 0.05. Detailed MR estimations are provided in ESM Table 5. Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 15 Jan, 2026 Reviews received at journal 14 Jan, 2026 Reviews received at journal 02 Jan, 2026 Reviewers agreed at journal 18 Dec, 2025 Reviewers agreed at journal 12 Nov, 2025 Reviewers invited by journal 08 Oct, 2025 Editor assigned by journal 08 Oct, 2025 Submission checks completed at journal 08 Oct, 2025 First submitted to journal 07 Oct, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7798894","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":531678326,"identity":"d33c47fd-e875-4efa-9659-bd70a2ebbca2","order_by":0,"name":"Yuqi Sun","email":"","orcid":"","institution":"Shanghai Jiao Tong University","correspondingAuthor":false,"prefix":"","firstName":"Yuqi","middleName":"","lastName":"Sun","suffix":""},{"id":531678327,"identity":"781c453e-c690-49d4-af36-52981e8365ff","order_by":1,"name":"Haonan Zheng","email":"","orcid":"","institution":"Shanghai Jiao Tong University","correspondingAuthor":false,"prefix":"","firstName":"Haonan","middleName":"","lastName":"Zheng","suffix":""},{"id":531678328,"identity":"a89bb348-0383-484f-9fbb-738c56ff04ee","order_by":2,"name":"Min Ma","email":"","orcid":"","institution":"Ruijin Hospital","correspondingAuthor":false,"prefix":"","firstName":"Min","middleName":"","lastName":"Ma","suffix":""},{"id":531678329,"identity":"9d2c6469-6a72-4482-acc1-5d54511f00d4","order_by":3,"name":"Rongrong Gu","email":"","orcid":"","institution":"Shanghai Jiao Tong University","correspondingAuthor":false,"prefix":"","firstName":"Rongrong","middleName":"","lastName":"Gu","suffix":""},{"id":531678330,"identity":"0c985605-ac78-4c31-9ee2-9c7d35cd74a0","order_by":4,"name":"Manqing Wang","email":"","orcid":"","institution":"Shanghai Jiao Tong University","correspondingAuthor":false,"prefix":"","firstName":"Manqing","middleName":"","lastName":"Wang","suffix":""},{"id":531678331,"identity":"e538a708-b02c-40cf-99cc-2d083045edbb","order_by":5,"name":"Si Fang","email":"","orcid":"","institution":"University of Bristol","correspondingAuthor":false,"prefix":"","firstName":"Si","middleName":"","lastName":"Fang","suffix":""},{"id":531678332,"identity":"55e9fa13-9e29-475e-871a-594f91b8fbf8","order_by":6,"name":"Yangbo Sun","email":"","orcid":"","institution":"University of Tennessee Health Science Center","correspondingAuthor":false,"prefix":"","firstName":"Yangbo","middleName":"","lastName":"Sun","suffix":""},{"id":531678333,"identity":"52d972ff-c518-4d15-be88-04ede256e2e4","order_by":7,"name":"Qian Yang","email":"","orcid":"","institution":"Ruijin Hospital","correspondingAuthor":false,"prefix":"","firstName":"Qian","middleName":"","lastName":"Yang","suffix":""},{"id":531678334,"identity":"f52dc025-3eaf-4ccc-9ca9-04a3ea029fee","order_by":8,"name":"Yufang Bi","email":"","orcid":"","institution":"Ruijin Hospital","correspondingAuthor":false,"prefix":"","firstName":"Yufang","middleName":"","lastName":"Bi","suffix":""},{"id":531678335,"identity":"0f21c11e-2bfe-4c07-ac8b-3ac0c8a58737","order_by":9,"name":"Jie Zheng","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAzElEQVRIie3QsQ6CMBCA4SMkdWnCeizwCiUuPk6NAxM74xESJh+ARJ9D1zYMLuhMggObr8BgVHQxYYC6OfSfbvmSuwOw2f4ynXX+4xkA1MakysWKuctfyKLAlLlrAm4IwkwTNpzFx+ysoE8P4O1omjikKSqRJ3t1kc62bgGvapq4oGmDApMSauE6RQsC5TRhA6nuUsRoTPjwZEIlpTlB0Dn4pKL3Ynq4hWMzQ8LydOsHEiLWUdenbeCVMwRQfQf12XQ2j8aDzWaz2Ua9ADooR0k5SFv4AAAAAElFTkSuQmCC","orcid":"","institution":"Ruijin Hospital","correspondingAuthor":true,"prefix":"","firstName":"Jie","middleName":"","lastName":"Zheng","suffix":""}],"badges":[],"createdAt":"2025-10-07 11:08:53","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7798894/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7798894/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":94026297,"identity":"694999a7-f538-44ce-8708-541c275ec6ea","added_by":"auto","created_at":"2025-10-21 13:33:51","extension":"doc","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":270336,"visible":true,"origin":"","legend":"","description":"","filename":"Manuscript929final.doc","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/764ad7ce649086f2e68cbae2.doc"},{"id":94025974,"identity":"b370190a-14a4-45ac-acde-3b80df4c949a","added_by":"auto","created_at":"2025-10-21 13:25:51","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2444093,"visible":true,"origin":"","legend":"","description":"","filename":"ESMTables.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/b2de7b708cb79785e9138934.xlsx"},{"id":94026300,"identity":"4c510452-1dda-41d9-be06-558dcf2634dd","added_by":"auto","created_at":"2025-10-21 13:33:51","extension":"xlsx","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":12034,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTable1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/95ed38a617a0a16254034d2b.xlsx"},{"id":94026303,"identity":"05827d09-6607-4f2d-8ea5-25ef0eb0d64e","added_by":"auto","created_at":"2025-10-21 13:33:51","extension":"xlsx","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":11535,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTable2.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/5666531a1f18d2154b084724.xlsx"},{"id":94026304,"identity":"26ceeb1f-84eb-4b5d-acf4-7795147c1aba","added_by":"auto","created_at":"2025-10-21 13:33:51","extension":"xlsx","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":10482,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTable3.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/6d906c0f7ebccffa5547b9a6.xlsx"},{"id":94025968,"identity":"33f0ae8f-b7ec-4c80-bb02-6bffb1a5fcc2","added_by":"auto","created_at":"2025-10-21 13:25:51","extension":"json","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":12690,"visible":true,"origin":"","legend":"","description":"","filename":"7adbf4feab16466c933aefdf12aaafab.json","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/e644612bc623a785976a7abf.json"},{"id":94026309,"identity":"c692b0b4-e8c3-46ad-87f8-ff93fb051f48","added_by":"auto","created_at":"2025-10-21 13:33:51","extension":"docx","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":69635,"visible":true,"origin":"","legend":"","description":"","filename":"ESM.docx","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/69368f823e1ad510dc0f2e88.docx"},{"id":94025989,"identity":"e149fae0-098a-46bb-b075-3e9db4439d2b","added_by":"auto","created_at":"2025-10-21 13:25:51","extension":"xml","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":178936,"visible":true,"origin":"","legend":"","description":"","filename":"7adbf4feab16466c933aefdf12aaafab1enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/4418881131833b698293f1c8.xml"},{"id":94026307,"identity":"55175c4d-9df8-46d6-81f0-62d07ddd0cf6","added_by":"auto","created_at":"2025-10-21 13:33:51","extension":"pdf","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":549983,"visible":true,"origin":"","legend":"","description":"","filename":"Figure1.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/5a2ee19052815958166bcd29.pdf"},{"id":94025991,"identity":"89f8234f-6dd1-411d-87b3-c76bf4c2fafe","added_by":"auto","created_at":"2025-10-21 13:25:52","extension":"pdf","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":214651,"visible":true,"origin":"","legend":"","description":"","filename":"Figure2.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/357736bf57819689d2e5d077.pdf"},{"id":94027047,"identity":"619aab81-692e-4185-b781-8e323679c094","added_by":"auto","created_at":"2025-10-21 13:41:51","extension":"pdf","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":77414,"visible":true,"origin":"","legend":"","description":"","filename":"Figure3.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/b1dd2a0e4e112894eb5ff1f7.pdf"},{"id":94025979,"identity":"4445c5bd-fabd-4546-960c-fcb442b28ba2","added_by":"auto","created_at":"2025-10-21 13:25:51","extension":"pdf","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":233514,"visible":true,"origin":"","legend":"","description":"","filename":"Figure4.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/aba2f558c5b9f0a5235eac25.pdf"},{"id":94025981,"identity":"a27c1482-e291-4cbe-8cbb-b4eff16481c5","added_by":"auto","created_at":"2025-10-21 13:25:51","extension":"pdf","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":10659,"visible":true,"origin":"","legend":"","description":"","filename":"Figure5.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/06b3a8bc0688d36c4bc8361a.pdf"},{"id":94025985,"identity":"91ceb486-93f1-40c5-97e6-e3a8e601cd82","added_by":"auto","created_at":"2025-10-21 13:25:51","extension":"pdf","order_by":22,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":14651,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryFigure1.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/6b5c2084454d349536bc25a4.pdf"},{"id":94026310,"identity":"710ae3ab-08d9-4965-83b9-1e38591f4b1c","added_by":"auto","created_at":"2025-10-21 13:33:51","extension":"pdf","order_by":23,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":404186,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryFigure2.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/a7e241279de815db2b35f77e.pdf"},{"id":94025990,"identity":"1da7a749-b674-4599-882c-419661ea995b","added_by":"auto","created_at":"2025-10-21 13:25:52","extension":"pdf","order_by":24,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":565282,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryFigure3.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/e2c7633600c52dab16fde707.pdf"},{"id":94027046,"identity":"fc783648-8781-4eeb-8696-d1e7921fd75a","added_by":"auto","created_at":"2025-10-21 13:41:51","extension":"pdf","order_by":25,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":8769,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryFigure4.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/79ca9085a241d9ed371d6ba8.pdf"},{"id":94027048,"identity":"86952a04-82cb-4929-9d18-4ece297f0cf5","added_by":"auto","created_at":"2025-10-21 13:41:51","extension":"xml","order_by":26,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":177181,"visible":true,"origin":"","legend":"","description":"","filename":"7adbf4feab16466c933aefdf12aaafab1structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/10c7d2255bbe406e620dbcb7.xml"},{"id":94025986,"identity":"0569e36b-cbe1-44d9-a95c-6ac1bd503731","added_by":"auto","created_at":"2025-10-21 13:25:51","extension":"html","order_by":27,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":191469,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/d5748cb518374405679d0b78.html"},{"id":94025958,"identity":"8fb58b0f-aa04-43ab-812c-8512535884b1","added_by":"auto","created_at":"2025-10-21 13:25:51","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":117715,"visible":true,"origin":"","legend":"\u003cp\u003eThe study design and workflow. A schematic representation detailing the research procedure adopted and a graphical abstract outlining the core assumptions of this study.\u003c/p\u003e","description":"","filename":"Figure1.png","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/76c9ad8d183afed0bcec73da.png"},{"id":94025959,"identity":"a2e9cd0e-11cd-4f85-9caf-c5898924de0f","added_by":"auto","created_at":"2025-10-21 13:25:51","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":79510,"visible":true,"origin":"","legend":"\u003cp\u003eCategory map of first-line glucose-lowering drugs and their interacted genes. Genes that showed a causal relationship with aging phenotypes are marked with red diamonds. The bar chart in the upper left corner shows the distribution of target gene categories. In total, this study collected 36 FDA-approved glucose-lowering drugs across 9 categories: 14 α-glucosidase inhibitors, 65 thiazolidinediones, 85 sulfonylureas, 14 SGLT-2 inhibitors, 22 meglitinides, 2 LYN kinase agonists, GLP-1 receptor agonists, DPP-4 inhibitors, and 91 biguanides.\u003c/p\u003e","description":"","filename":"Figure2.png","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/9022a0b610a4195ac4a364f9.png"},{"id":94025961,"identity":"f54c2222-ddf8-4e91-84a5-8d37ebb8295c","added_by":"auto","created_at":"2025-10-21 13:25:51","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":144630,"visible":true,"origin":"","legend":"\u003cp\u003eCircular plot showing all MR estimations of drug target genes’ expression levels on aging phenotypes. Z-scores were calculated as the effect size divided by its standard error (β/SE) to represent the strength and direction of each association. The innermost ring indicates the tissue of origin for each gene's expression, while the outermost ring shows the name of the drug associated with each gene. EAA, epigenetic age acceleration. HR, hazard ratio. Detailed MR results are provided in ESM Table 5.\u003c/p\u003e","description":"","filename":"Figure3.png","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/40963cdd2e44d9ffc0b5a4d3.png"},{"id":94027043,"identity":"ba68e7df-a77c-4459-ab82-39b85f21ac7a","added_by":"auto","created_at":"2025-10-21 13:41:51","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":136081,"visible":true,"origin":"","legend":"\u003cp\u003eCircular plot showing CpG sites distribution associated with the target genes in relevant tissues. Each spot represents a single CpG site. The y-axis shows the −log₁₀-transformed P\u003csub\u003eFDR\u003c/sub\u003e-values indicating the statistical significance of the association between genes and CpG site methylation levels in each tissue. Detailed MR results are provided in ESM Table 6.\u003c/p\u003e","description":"","filename":"Figure4.png","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/834f603f5d53bc0ca05b2142.png"},{"id":94025962,"identity":"ad3224a6-78c4-478c-9a58-3d4aee02283e","added_by":"auto","created_at":"2025-10-21 13:25:51","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":85932,"visible":true,"origin":"","legend":"\u003cp\u003eForest plot showing MR results of tissue-specific expression levels of drug target genes on aging phenotypes, along with their associated mediating CpG sites. 5 genes were selected as representative examples based on prior knowledge and are indicated at the top of each panel. Thetop half of each chart displays the effect of gene expression on aging outcomes. The x-axis represents the change in outcome per standard deviation (SD) increase in the expression level of each drug target gene. The bottom half of each chart shows the top cis-CpG site mediators (up to 10 per gene) , selected based on the lowest p-values of the mediation effect. The x-axis represents the change in the outcome per standard deviation (SD) increase in the methylation level of each CpG site. Detailed estimations are provided in ESM Table 5, 7, 8 and 9.\u003c/p\u003e","description":"","filename":"Figure5.png","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/13f8022fc206cf95685f196b.png"},{"id":94028194,"identity":"dab4dbdf-2baf-4024-be37-7c9c49fb4bfc","added_by":"auto","created_at":"2025-10-21 13:57:53","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1314213,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/102025db-0505-4356-a71a-d3ee6ed505f4.pdf"},{"id":94026296,"identity":"75972c84-209d-4fd1-a208-359fb6ce08ff","added_by":"auto","created_at":"2025-10-21 13:33:51","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":69635,"visible":true,"origin":"","legend":"","description":"","filename":"ESM.docx","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/a02a6551127a7fb720b44bba.docx"},{"id":94025975,"identity":"ddce0ae9-dde2-422a-958f-77ea035911aa","added_by":"auto","created_at":"2025-10-21 13:25:51","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":2444093,"visible":true,"origin":"","legend":"","description":"","filename":"ESMTables.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/a9eb96fd9ad95ef0dc18e1a0.xlsx"},{"id":94027483,"identity":"b9485474-149b-4f6d-8d7e-4c5b258b81fe","added_by":"auto","created_at":"2025-10-21 13:49:51","extension":"xlsx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":12034,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTable1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/1ed419292a56a34814658482.xlsx"},{"id":94026301,"identity":"1cb1d2dc-9e8f-4057-8eb6-c17f95058d94","added_by":"auto","created_at":"2025-10-21 13:33:51","extension":"xlsx","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":11535,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTable2.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/78a645b2796e2fc7533d60c8.xlsx"},{"id":94025966,"identity":"071dab1e-e9e8-45e2-969e-599ab9abff95","added_by":"auto","created_at":"2025-10-21 13:25:51","extension":"xlsx","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":10482,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cbr\u003e\u003c/p\u003e","description":"","filename":"SupplementaryTable3.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/efc5368526a2c96262280509.xlsx"},{"id":94025971,"identity":"5db25244-fd9f-4824-837e-e893e2381ac4","added_by":"auto","created_at":"2025-10-21 13:25:51","extension":"pdf","order_by":6,"title":"","display":"","copyAsset":false,"role":"supplement","size":14651,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSupplementary Figure 1\u003c/strong\u003e Forest plot of the top MR results: the effects of drug target gene expression levels on glycemic phenotypes. Outcomes include 2-hour glucose (A), HbA1c (B), fasting glucose (C), fasting insulin (D), and type 2 diabetes (E). The x-axis represents the change in the outcome per standard deviation (SD) increase in the expression level of each drug target gene. Statistical significance was defined as P\u003csub\u003eFDR\u003c/sub\u003e \u0026lt; 0.05.\u0026nbsp; Detailed MR estimates are provided in ESM Table 3.\u003c/p\u003e","description":"","filename":"SupplementaryFigure1.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/e45e3f8f4592ebe9835621ba.pdf"},{"id":94026305,"identity":"688fa216-c6a4-4845-86e8-eb4ae48d0ca3","added_by":"auto","created_at":"2025-10-21 13:33:51","extension":"pdf","order_by":7,"title":"","display":"","copyAsset":false,"role":"supplement","size":404186,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSupplementary Figure 2\u003c/strong\u003e Genetic correlations between ten aging outcomes. Diagonal values show the observed-scale SNP heritability of each aging-related phenotype and off-diagonal circles show the genetic correlation among these traits. Blank squares are those did not pass multiple testing correction threshold (FDR \u0026gt; 0.05). The bottom dendrogram shows the hierarchical relationship between traits, based on the magnitude of their genetic correlations. The black box highlights the 4 clusters of aging-related phenotypes used in follow-up analyses: telomere length, parental longevity, healthspan and epigenetic age acceleration (EAA). HR, hazard ratio. Detailed results are provided in ESM Table 4.\u003c/p\u003e","description":"","filename":"SupplementaryFigure2.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/35a0e3189dfd08ce96fb1470.pdf"},{"id":94025976,"identity":"0efa251e-c0ee-4549-995c-c9175e8fb7bc","added_by":"auto","created_at":"2025-10-21 13:25:51","extension":"pdf","order_by":8,"title":"","display":"","copyAsset":false,"role":"supplement","size":565282,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSupplementary Figure 3 \u003c/strong\u003eGenomic distribution of mQTL and the target genes. The bar chart (A) on the top represents the position of target gene candidates and the count of CpG sites associated with them. The scatter plot (B) below represents the genomic positions of each mQTL and its corresponding CpG site. Detailed results are provided in ESM Table 6.\u003c/p\u003e","description":"","filename":"SupplementaryFigure3.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/7fc0f5ee0dab2d621e524d19.pdf"},{"id":94027045,"identity":"a3623b84-751d-4760-a78f-7566bde486a1","added_by":"auto","created_at":"2025-10-21 13:41:51","extension":"pdf","order_by":9,"title":"","display":"","copyAsset":false,"role":"supplement","size":8769,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSupplementary Figure 4 \u003c/strong\u003eForest plot of all MR results of the expression level of drug target genes on ten aging phenotypes, including combined parental age at death (A), father’s age at death (B), frailty (C), healthspan (D), leukocyte telomere length (E). The horizontal axis represents the effect size of exposure on outcomes. HR, hazard ratio. Statistical significance was defined as P\u003csub\u003eFDR\u003c/sub\u003e \u0026lt; 0.05. Detailed MR estimations are provided in ESM Table 5.\u003c/p\u003e","description":"","filename":"SupplementaryFigure4.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7798894/v1/95a35d9c2552f01d1074e692.pdf"}],"financialInterests":"Competing interest reported. Jie Zheng and Yangbo Sun currently serve as members of the Editorial Board of BMC Medicine","formattedTitle":"The tissue-specific effects of glucose-lowering drug targets on aging mediated through DNA methylation: a multi-omics genetic study","fulltext":[{"header":"Introduction","content":"\u003cp\u003eAging is a complicated process characterized by the progressive accumulation of molecular damage, where epigenetic alteration is considered one of the primary hallmarks and an unambiguous contributor [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. Since the human DNA methylation (DNAm) landscape has been associated with plenty of aging-related phenotypes [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e], specific therapeutic interventions on the epigenetic status yield tangible benefits for healthspan. Given the inconsistency of DNAm patterns across different human tissues, recent studies have emphasized the importance of developing tissue-specific interventions [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. A retrospective review of randomized controlled trials (RCTs) summarized 8 categories of promising anti-aging compounds that act via various pathways, including epigenetic alteration [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. Two of them are glucose-lowering drugs: metformin and glucagon-like peptide-1 receptor agonists, thus attracting people\u0026rsquo;s attention to the anti-aging potential of all types of glucose-lowering drugs.\u003c/p\u003e\u003cp\u003eRCTs involving glucose-lowering drugs are usually designed to evaluate their effects on age-associated diseases rather than aging [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. Those RCTs were designed to recruit high-risk participants (e.g., those with cardiovascular diseases or dementia), thus providing limited evidence of extending healthspan or decelerating aging in the general population [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. A small RCT, Metformin in Longevity Study (N\u0026thinsp;=\u0026thinsp;14), provided preliminary evidence that metformin can lead to transcriptomic changes in pathways affecting aging in skeletal muscle and subcutaneous adipose tissues [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. An ongoing RCT, Targeting Aging with Metformin [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e], aims to explore the effect of metformin among 3,000 ethnically diverse, non-diabetic participants aged 65\u0026ndash;80, but the results have not been released yet. Besides, only a few clinical trials explored the beneficial effects of other glucose-lowering agents on aging, including acarbose (N\u0026thinsp;=\u0026thinsp;10, NCT02953093), canagliflozin (N\u0026thinsp;=\u0026thinsp;30, NCT06301529), and dapagliflozin (N\u0026thinsp;=\u0026thinsp;20, NCT04401904), all of which have very limited sample sizes.\u003c/p\u003e\u003cp\u003eGiven the limitations of conventional observational studies and inconclusive results from RCTs, human genetics could help shed new light on the anti-aging nature of glucose-lowering drug targets, which may further facilitate the development of new lifespan-extending medications [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. Previous Mendelian randomization (MR) studies indicated potentially protective effects of targets for metformin, sodium-glucose cotransporter 2 inhibitors, and sulfonylureas on aging-related diseases, including cardiovascular diseases, cancer, dementia, inflammation, and heart failure [\u003cspan additionalcitationids=\"CR10 CR11\" citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. In addition, metformin might promote healthy aging via targeting GPD1 and AMPKγ2 (\u003cem\u003ePRKAG2\u003c/em\u003e) [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. Nevertheless, it remains unclear whether glucose-lowering drugs protect against aging itself rather than aging-related diseases, and whether they function through epigenetic alterations.\u003c/p\u003e\u003cp\u003eThe aim of this study is to explore the target-specific effects of glucose-lowering drug targets on aging biomarkers using two-sample MR. To provide mechanistic insights, we further used two-step MR to investigate whether the effect of glucose-lowering drug targets on aging is mediated through DNAm.\u003c/p\u003e"},{"header":"Methods","content":"\n\u003ch3\u003eStudy design\u003c/h3\u003e\n\u003cp\u003eFigure\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e presents the study design and workflow, including a schematic representation detailing the research procedure adopted and a graphical abstract outlining the data resource and core assumptions of the study.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003eSelection of genetic instruments for glucose-lowering drug targets\u003c/h2\u003e\u003cp\u003eThis study considered nine major categories of glucose-lowering drugs used in first-line clinical practice [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]: Glucagon-Like peptide-1 receptor agonists, biguanides, sulfonylureas, meglitinides, α-glucosidase inhibitors, dipeptidyl peptidase-4 inhibitors, sodium-glucose cotransporter 2 inhibitors, thiazolidinediones, and Lck/Yes-related Novel protein tyrosine kinase activators. A total of 36 FDA-approved drugs were selected as candidates. Since the mechanism by which these drugs act is not fully understood, we searched the listed glucose-lowering drugs in the Drug-Gene Interaction database (DGIdb) v.5.0.7 [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e], and reserved 194 related genes as potential targets. Using information implemented in DrugBank, those gene targets with clear mechanisms of action were prioritized as Tier 1 druggable genes, where those genes exhibiting drug-gene interactions were classified as Tier 2 (electronic supplementary material [ESM] Table\u0026nbsp;1).\u003c/p\u003e\u003cp\u003eWe obtained tissue-specific \u003cem\u003ecis-\u003c/em\u003eacting expression quantitative trait loci (\u003cem\u003ecis-\u003c/em\u003eeQTLs) of each glucose-lowering drug target gene from Genotype-Tissue Expression (GTEx) v8 Project (15,201 samples of 49 tissues from 838 donors), and used them to select tissue-specific genetic instruments [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. A stringent selection process was performed to identify tissue-specific genetic variants for each target, requiring a SNP a) being associated with expression levels of corresponding genes (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;5 \u0026times; 10\u003csup\u003e\u0026minus;\u0026thinsp;8\u003c/sup\u003e) in the \u003cem\u003ecis\u003c/em\u003e-acting regions (i.e. \u0026plusmn;500 kb window of each gene probe); b) passing a stringent linkage disequilibrium (LD) clumping test using the panel of European ancestry with clumping \u003cem\u003er\u003c/em\u003e\u003csup\u003e2\u003c/sup\u003e of less than 0.001 and clumping window of 10,000 kb; c) having an F-statistic\u0026thinsp;\u0026gt;\u0026thinsp;10, which indicating good instrument strength; d) having minor allele frequency\u0026thinsp;\u0026ge;\u0026thinsp;0.01. We also applied Steiger filtering to identify the genetic variants explaining more variations of the exposures than the outcomes, thereby addressing potential reverse causality [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. In addition, we considered five glycemic traits: Hemoglobin A1c, fasting glucose, fasting insulin, 2-hour glucose, and type 2 diabetes (T2D) as positive controls to validate the reliability of the instruments. The associations of genetic susceptibility to the targets with glycemic traits were estimated using two-sample MR. Only targets relevant to at least one of these traits (false discovery rate, FDR\u0026thinsp;\u0026lt;\u0026thinsp;0.05) were reserved for further analyses.\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eSelection of aging proxy indicators\u003c/h3\u003e\n\u003cp\u003eA variety of aging biomarkers and phenotypes have been proposed for the identification and evaluation of longevity interventions [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. Referring to the latest aging biomarker consensus and previous MR studies about aging [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e], we selected ten phenotypes as aging proxy indicators: father\u0026rsquo;s age at death, mother\u0026rsquo;s age at death, combined parental age at death, Hannum Age, Intrinsic Horvath Age, PhenoAge, Grimage, leukocyte telomere length, healthspan, and frailty. The details of genome-wide association studies (GWAS) on each phenotype are described in ESM Method 1. The main characteristics of each GWAS are presented in Supplementary Table\u0026nbsp;1.\u003c/p\u003e\n\u003ch3\u003eSelection of DNAm mediators\u003c/h3\u003e\n\u003cp\u003eThe GoDMC study published a database of \u0026gt;\u0026thinsp;270,000 independent DNAm quantitative trait loci (mQTLs), involving 32,851 European participants [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. Putative genetic variants associated with DNAm at 420,509 CpG sites were identified from blood samples, with genotype data imputed to the 1000 Genomes Project reference panel [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e], and DNAm profiles quantified from bisulfite-converted genomic whole-blood DNA using the Infinium HumanMethylation BeadChip (HumanMethylation450 or EPIC arrays). We localized the genome-wide DNAm datasets and the genetic location of all CpG sites available from GoDMC to establish a retrieval library, using a standardized pipeline to convert all data to hg19/GRCh37 build to ensure consistency. For genetic instruments of CpG sites, we used the same standard applied in eQTL instrument selection. CpG sites associated with expression levels of the drug-related genes as identified by MR results, and located within \u0026plusmn;\u0026thinsp;500kb of the gene, are considered \u003cem\u003ecis\u003c/em\u003e-regulated by drug-related targets; otherwise, they are classified as \u003cem\u003etrans\u003c/em\u003e-regulated.\u003c/p\u003e\n\u003ch3\u003eSelection of positive controls\u003c/h3\u003e\n\u003cp\u003eTo ensure our findings reflect specific pharmacological effects rather than non-specific processes such as drug metabolism, we restricted our analyses to target genes that exhibited a significant impact on glycemic traits. GWAS of five indexes which reflect the change of blood glucose are used as outcomes in positive controls to select valid instruments for drug targets, including HbA1c level (N\u0026thinsp;=\u0026thinsp;146,806), fasting glucose level (N\u0026thinsp;=\u0026thinsp;200,622), fasting insulin level (N\u0026thinsp;=\u0026thinsp;151,013), 2-hour glucose level (N\u0026thinsp;=\u0026thinsp;63,396) from MAGIC consortium [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e] and T2D (N\u0026thinsp;=\u0026thinsp;1,812,017, N cases\u0026thinsp;=\u0026thinsp;242,283, N controls\u0026thinsp;=\u0026thinsp;1,569,734) from the T2DGGI study [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. Participants were of European descent, and detailed information on these GWAS is summarized in Supplementary Table\u0026nbsp;2.\u003c/p\u003e\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e\u003ch2\u003eStatistical analysis\u003c/h2\u003e\u003cdiv id=\"Sec8\" class=\"Section3\"\u003e\u003ch2\u003eGenetic correlation analysis\u003c/h2\u003e\u003cp\u003eWe used LD-score regression to calculate the genetic correlations and SNP heritability of the 10 aging traits [\u003cspan additionalcitationids=\"CR24\" citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e], based on the European-ancestry LD reference panel. In line with recommendations [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e], low-quality (INFO\u0026thinsp;\u0026lt;\u0026thinsp;0.9) and rare (minor allele frequency\u0026thinsp;\u0026lt;\u0026thinsp;0.01) SNPs, as well as those located in the Major Histocompatibility Complex, were discarded before merging the summary statistics. We further clustered those ten traits into four groups according to their genetic correlations.\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\n\u003ch3\u003eMendelian randomization analyses\u003c/h3\u003e\n\u003cp\u003eWe firstly conducted a two-sample MR to identify the glucose lowering drug target genes whose expression level causally affect the aging outcomes. A two-step MR was further conducted to assess whether DNAm could mediate the effect. The first step is to assess how each target gene\u0026rsquo;s expression level influences DNAm of CpG sites across the whole genome. The second step was carried out on those DNAm that passed the first step (MR P\u003csub\u003eFDR\u003c/sub\u003e \u0026lt; 0.05). For these DNAm, we evaluated whether they affect aging-related outcomes. Mediation analysis was then performed to identify the mediation proportion of the DNAm of CpG sites in each target gene expression and aging outcome pair. The product of coefficients method and the delta method were applied to calculate the indirect effect and the standard error of the indirect effect, respectively, for each of the mediator [\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eFor all the MR analysis listed above, we utilized Wald\u0026rsquo;s ratio method for exposures with only one instrument, and the inverse variance weighted (IVW) method for exposures with two or more instruments [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]. To ensure the robustness of associations, we further performed sensitivity analyses using weighted median and weighted mode methods for exposures with three or more instruments. For instruments that are not available in the outcome GWAS data sets during the MR analysis, proxy genetic variants in high LD with this eQTL/mQTL (\u003cem\u003er\u003c/em\u003e\u003csup\u003e\u003cem\u003e2\u003c/em\u003e\u003c/sup\u003e\u0026thinsp;\u0026gt;\u0026thinsp;0.8 in the 1000 Genomes data in the European population [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]) were used as substitutes.\u003c/p\u003e\n\u003ch3\u003eValidation of MR assumptions\u003c/h3\u003e\n\u003cp\u003eMR analyses require three key assumptions: (1) relevance: the genetic predictors have robust associations with the exposure; (2) exchangeability: there are no common causes (confounders) between the instrument and outcome; (3) exclusion restriction: the instruments are only related to the outcome through the exposure being studied [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]. Findings were reported according to the STROBE-MR (Strengthening the Reporting of Mendelian Randomization Studies) guidelines (ESM Method 2) [\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eTo assess the relevance assumption, we examined the strength of each genetic variant using F-statistics. Only variants meet that this standard will be considered valid instrumental variables.\u003c/p\u003e\u003cp\u003eThe exchangeability assumption cannot be definitively proven, but associations with known confounders can be tested and demonstrated [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. We used GWAS in European participants to avoid population mixture. These GWAS were adjusted for the top principal components or through the use of adjusted linear mixed models to minimize the confounding by population stratification [\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]. We also conducted LD check analysis to evaluate the LD structures of causal genes or CpGs with robust MR evidence (FDR\u0026thinsp;\u0026lt;\u0026thinsp;0.05) and mitigate bias resulting from confounding by LD [\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e]. We estimated the pairwise LD between each eQTL/mQTL and mQTL/aging-associated GWAS variant in the \u003cem\u003ecis-\u003c/em\u003eregion. We utilized the threshold of LD \u003cem\u003er\u003c/em\u003e\u003csup\u003e\u003cem\u003e2\u003c/em\u003e\u003c/sup\u003e\u0026thinsp;\u0026gt;\u0026thinsp;0.8 as evidence of LD checked for the whole study [\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e]. Only results that pass the LD check will be included in further analysis.\u003c/p\u003e\u003cp\u003eTo assess the exclusion restriction assumption, we applied two sensitivity tests: weighted median and weighted mode methods, to minimize potential bias and improve confidence in our causal interpretation when more than 3 instruments were available.\u003c/p\u003e\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\u003ch2\u003eEnrichment analyses\u003c/h2\u003e\u003cp\u003eWe performed tissue-enrichment analyses for all prospective drug target genes identified through the MR analyses. Gene-level RNA-Seq expression values, measured in transcripts per million (TPM), were obtained from the GTEx V8 release. The tissue-specific genes are defined using the simplified algorithm from the Human Protein Atlas (ESM Method 2) [\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e]. We also queried Enrichr to identify the tissue-specific aging-associated expression variation and relevant aging pathways for the drug target genes [\u003cspan additionalcitationids=\"CR35\" citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e].\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\u003ch2\u003eTriangulation of evidence\u003c/h2\u003e\u003cp\u003ePublic databases of epigenome-wide association study (EWAS) and transcriptome-wide association study (TWAS) were reviewed to support our discoveries. We searched the EWAS Catalog (retrieval version: data last updated on March 22, 2024) [\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e] on the EWAS Open platform, to explore DNAm associations for our drug target genes. We also searched the TWAS Atlas (retrieval version: data last updated on May 15, 2022) [\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e], to identify disease associations with gene transcription in whole blood and multiple tissues, collecting tissue-specific TWAS results to further support our findings.\u003c/p\u003e\u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\u003ch2\u003eSummary of instrument and outcome selection\u003c/h2\u003e\u003cp\u003eWe searched drug-gene interaction (DGI) for 36 glucose-lowering drugs (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, ESM Table\u0026nbsp;1). A total of 194 drug-related genes were selected as candidates; 160 of them were found in the DGI database with 1,619 corresponding \u003cem\u003ecis\u003c/em\u003e-eQTLs from 49 types of human tissues identified as instruments (ESM Table\u0026nbsp;2). Among these, the expression of 23 genes showed a causal effect on at least one glycemic phenotype in tissue-specific conditions at the FDR threshold of P\u0026thinsp;\u0026lt;\u0026thinsp;0.05 and were regarded as drug target genes for this study (Supplementary Fig.\u0026nbsp;1, ESM Table\u0026nbsp;3).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eBy calculating genetic correlations (rg) and subsequently performing hierarchical clustering based on the magnitude of these correlations, we found that the 10 aging phenotypes from four distinctive clusters of high genetic similarity (|rg| \u0026ge; 0.2; P\u0026thinsp;\u0026lt;\u0026thinsp;5\u0026times; 10\u003csup\u003e\u0026minus;\u0026thinsp;3\u003c/sup\u003e) (Supplementary Fig.\u0026nbsp;2, ESM Table\u0026nbsp;4). Thus, we categorized them into four groups: telomere length, parental longevity, healthspan and epigenetic age acceleration (EAA), to represent aging from different aspects.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\u003ch2\u003eTissue-specific associations of gene expression levels of glucose-lowering drug targets with aging outcomes\u003c/h2\u003e\u003cp\u003eThrough the two-sample MR, ten target genes (\u003cem\u003eEHMT2\u003c/em\u003e, \u003cem\u003eHSPA4\u003c/em\u003e, \u003cem\u003eIGF2BP2\u003c/em\u003e, \u003cem\u003eIRS1\u003c/em\u003e, \u003cem\u003eLPL\u003c/em\u003e, \u003cem\u003eNDUFAF1\u003c/em\u003e, \u003cem\u003eNDUFS3\u003c/em\u003e, \u003cem\u003eSLC22A3\u003c/em\u003e, \u003cem\u003eTCF7L2\u003c/em\u003e, and \u003cem\u003eKCNJ11\u003c/em\u003e) of sixteen drugs (acarbose, acetohexamide, canagliflozin, chlorpropamide, dapagliflozin, gliclazide, glimepiride, glipizide, gliquidone, glyburide, metformin, nateglinide, pioglitazone, repaglinide, tolazamide and tolbutamide) displayed strong evidence for their tissue-specific expressions to be causally related to aging phenotypes (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, ESM Table\u0026nbsp;5). We found the expression levels of \u003cem\u003eEHMT2\u003c/em\u003e, \u003cem\u003eHSPA4\u003c/em\u003e, and \u003cem\u003eNDUFS3\u003c/em\u003e in multiple brain regions, including caudate basal ganglia, cerebellar hemisphere, nucleus accumbens basal ganglia, and putamen basal ganglia, were associated with frailty and leukocyte telomere length. \u003cem\u003eLPL\u003c/em\u003e and \u003cem\u003eSLC22A3\u003c/em\u003e expression in the tibial nerve was associated with parental age (odds ratio [OR]\u0026thinsp;=\u0026thinsp;1.06, 95%CI\u0026thinsp;=\u0026thinsp;1.03\u0026ndash;1.09, P\u003csub\u003eFDR\u003c/sub\u003e=3.36\u0026times;10\u003csup\u003e\u0026minus;\u0026thinsp;3\u003c/sup\u003e) and health span (OR\u0026thinsp;=\u0026thinsp;1.07, 95%CI\u0026thinsp;=\u0026thinsp;1.02\u0026ndash;1.11, P\u003csub\u003eFDR\u003c/sub\u003e=3.37\u0026times;10\u003csup\u003e\u0026minus;\u0026thinsp;2\u003c/sup\u003e). In addition, \u003cem\u003eSLC22A3\u003c/em\u003e also exhibited a pan-tissue effect, which showed associations with healthspan in multiple tissues, including the left ventricle (OR\u0026thinsp;=\u0026thinsp;1.05, 95%CI\u0026thinsp;=\u0026thinsp;1.02\u0026ndash;1.08, P\u003csub\u003eFDR\u003c/sub\u003e=3.22\u0026times;10\u003csup\u003e\u0026minus;\u0026thinsp;2\u003c/sup\u003e), liver (OR\u0026thinsp;=\u0026thinsp;0.96, 95%CI\u0026thinsp;=\u0026thinsp;0.95\u0026ndash;0.98, P\u003csub\u003eFDR\u003c/sub\u003e=6.29\u0026times;10\u003csup\u003e\u0026minus;\u0026thinsp;4\u003c/sup\u003e), subcutaneous adipose tissue (OR\u0026thinsp;=\u0026thinsp;1.06, 95%CI\u0026thinsp;=\u0026thinsp;1.02\u0026ndash;1.10, P\u003csub\u003eFDR\u003c/sub\u003e=6.29\u0026times;10\u003csup\u003e\u0026minus;\u0026thinsp;4\u003c/sup\u003e), ovary (OR\u0026thinsp;=\u0026thinsp;1.03, 95%CI\u0026thinsp;=\u0026thinsp;1.02\u0026ndash;1.05, P\u003csub\u003eFDR\u003c/sub\u003e=1.03\u0026times;10\u003csup\u003e\u0026minus;\u0026thinsp;2\u003c/sup\u003e), and testis (OR\u0026thinsp;=\u0026thinsp;1.06, 95%CI\u0026thinsp;=\u0026thinsp;1.04\u0026ndash;1.09, P\u003csub\u003eFDR\u003c/sub\u003e=1.42\u0026times;10\u003csup\u003e\u0026minus;\u0026thinsp;4\u003c/sup\u003e) in addition to tibial nerve.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec16\" class=\"Section2\"\u003e\u003ch2\u003eMediation analysis\u003c/h2\u003e\u003cp\u003eIn Step 1 of the two-step MR, associations of 3,526 pairs of target genes and CpG sites were tested, with 3,230 showing at least marginal MR evidence, including 2,064 \u003cem\u003ecis-\u003c/em\u003eacting associations and 1,166 \u003cem\u003etrans-\u003c/em\u003eacting associations (ESM Table\u0026nbsp;6, Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). In Step 2, the 5,742 genetic variants robustly associated with DNAm were used as instruments to proxy 1,158 CpG sites. We then tested the associations between DNAm of CpG sites and aging traits (ESM Tables\u0026nbsp;7 and 8, Supplementary Fig.\u0026nbsp;3). Eventually, nine target genes (\u003cem\u003eEHMT2\u003c/em\u003e, \u003cem\u003eHSPA4\u003c/em\u003e, \u003cem\u003eIGF2BP2\u003c/em\u003e, \u003cem\u003eIRS1\u003c/em\u003e, \u003cem\u003eLPL\u003c/em\u003e, \u003cem\u003eNDUFAF1\u003c/em\u003e, \u003cem\u003eNDUFS3\u003c/em\u003e, \u003cem\u003eSLC22A3\u003c/em\u003e and \u003cem\u003eTCF7L2\u003c/em\u003e) of eight drugs: acarbose, canagliflozin, dapagliflozin, glimepiride, metformin, pioglitazone, repaglinide and tolazamide, were associated with aging outcomes mediated by DNAm of 1,058 CpG sites (ESM Table\u0026nbsp;9, Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e, Supplementary Fig.\u0026nbsp;4). 755 of 1,058 (71.4%) of the mediating DNAm are \u003cem\u003ecis-\u003c/em\u003eregulated by their corresponding drug target genes, except \u003cem\u003eEHMT2\u003c/em\u003e exhibits both \u003cem\u003ecis-\u003c/em\u003e and \u003cem\u003etrans-\u003c/em\u003eregulated DNAm. The target genes mediated through DNAm distribute in 16 tissues of five systems, including nervous system (caudate, nucleus accumbens, putamen and cerebellar hemisphere and tibial nerve), cardiovascular system (artery aorta and left ventricle), immune and endocrine system (cultured fibroblasts, EBV-transformed lymphocytes, spleen, adipose tissues, liver and thyroid), respiratory system (lung) and reproductive system (testis and ovary).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec17\" class=\"Section2\"\u003e\u003ch2\u003eTriangulation of external evidence\u003c/h2\u003e\u003cp\u003eResults extracted from the previous EWAS study supported the mediation effect of relevant target genes and aging via epigenetic regulation. DNAm of five genes: \u003cem\u003eEHMT2\u003c/em\u003e, \u003cem\u003eHSPA4\u003c/em\u003e, \u003cem\u003eIGF2BP2\u003c/em\u003e, \u003cem\u003eSLC22A3\u003c/em\u003e, and \u003cem\u003eTCF7L2\u003c/em\u003e, showed associations with aging-related outcomes in EWAS, which in line with our findings (ESM Table\u0026nbsp;10). For TWAS findings reported in TWAS Atlas, only a limited number of studies reported associations with aging-related outcomes. Thus, the top eight aging-related diseases, as well as their common risk factors, were included to proxy aging, in accordance with the definition of healthspan phenotype [\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e]. All ten prioritized aging associated genes also showed relevant TWAS evidence: \u003cem\u003eEHMT2\u003c/em\u003e, \u003cem\u003eHSPA4, IGF2BP2\u003c/em\u003e, \u003cem\u003eIRS1\u003c/em\u003e, \u003cem\u003eLPL\u003c/em\u003e, \u003cem\u003eNDUFAF1\u003c/em\u003e, \u003cem\u003eNDUFS3\u003c/em\u003e, \u003cem\u003eSLC22A3\u003c/em\u003e and \u003cem\u003eTCF7L2\u003c/em\u003e (ESM Table\u0026nbsp;11). Tissue enrichment analysis showed no evidence of preferential expression of drug target genes in tissues corresponding to MR results (Supplementary Table\u0026nbsp;3). However, results from Enrichment analysis from Enrichr indicated that these genes are members of several metabolism, mitochondrial functions, insulin/IGF-1 signaling pathways and other aging-associated pathways [\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e, \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e], with tissue-specific regulatory relevance of aging further supported by external evidence from GEO datasets [\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e].\u003c/p\u003e\u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eThis drug target genetic study comprehensively investigated the tissue-specific effect of glucose-lowering drug targets on a range of aging phenotypes mediating through DNAm. We extended the recent discoveries of anti-aging targets and identified DNAm at multiple CpG sites as potential mediators in the aging process, highlighting the importance of exploring epigenetic regulatory mechanisms in aging-related therapies. Our study showed potentials of eight relevant anti-diabetic drugs as prospective candidates to resist aging. The tissue-specific expression of nine drug-targeting or drug-interacting genes: \u003cem\u003eEHMT2\u003c/em\u003e, \u003cem\u003eHSPA4\u003c/em\u003e, \u003cem\u003eIGF2BP2\u003c/em\u003e, \u003cem\u003eIRS1\u003c/em\u003e, \u003cem\u003eLPL\u003c/em\u003e, \u003cem\u003eNDUFAF1\u003c/em\u003e, \u003cem\u003eNDUFS3\u003c/em\u003e, \u003cem\u003eSLC22A3\u003c/em\u003e, and \u003cem\u003eTCF7L2\u003c/em\u003e as targets for future research on decelerating aging with a focus on their epigenetic regulations through DNAm.\u003c/p\u003e\u003cp\u003eAmong the eight anti-diabetic drugs, acarbose [\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e], canagliflozin [\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e], dapagliflozin [\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e] and metformin [\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e, \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e], and pioglitazone [\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e] have previously shown the potentials to extend lifespan or decrease the risk of several aging-related diseases in animal models. However, their evidence in humans was largely based on observational studies [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan additionalcitationids=\"CR49 CR50\" citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e], leaving it unclear whether these associations are driven by residual confounding, reverse causality and immortal time bias [\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e]. Besides, most of these studies focused on the protective effects of the anti-diabetic drugs in high-risk populations from certain aging-related diseases (e.g. heart failure) rather than aging directly. Our results provided evidence to support the conservatism of the anti-aging effects of the relevant drug targets in the general population. Among the drugs we identified, glimepiride, repaglinide and tolazamide showed their potential of extending lifespan that were not reported before. These drugs could be considered as drug candidates for future clinical research.\u003c/p\u003e\u003cp\u003eThe tissue-specific effects of gene expression we observed align well with previous studies of age-related diseases. For instance, low-dose pioglitazone has been shown to extend lifespan and improve cognitive function in aged mice by modulating metabolic homeostasis [\u003cspan additionalcitationids=\"CR54\" citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e], though no direct evidence has been reported in humans. In this study, we identified the mitochondrial protein \u003cem\u003eHSPA4\u003c/em\u003e, as a potential core gene in pioglitazone\u0026rsquo;s regulatory pathway of decelerating aging in brain tissues. Notably, \u003cem\u003eHSPA4\u003c/em\u003e expression has also been found to be downregulated in human front cortex as well as in several mouse brain tissues [\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e], consistent with the tissue specificity highlighted by our MR results. In addition, a previous study found that acarbose has a pronounced lifespan-extending effect in male mice [\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e]. Here, we found one of its target genes, \u003cem\u003eLPL\u003c/em\u003e, and its expression in the tibial nerves is the only exposure related to parental longevity, particularly the father\u0026rsquo;s life span. Similarly, \u003cem\u003eSLC22A3\u003c/em\u003e expression in testis has been found to be related to combined parental age at death, which aligned with the gender difference in metformin effect on aging [\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e]. This may contribute as a starting point to further investigate the gender difference in the anti-aging effect of drugs.\u003c/p\u003e\u003cp\u003eWe uncovered potential anti-aging roles of target genes for two sulfonylurea drugs (\u003cem\u003eEHMT2\u003c/em\u003e for tolazamide and \u003cem\u003eIRS1\u003c/em\u003e for glimepiride), and a meglitinide class drug (\u003cem\u003eIGF2BP2\u003c/em\u003e for repaglinide). Specifically, \u003cem\u003eEHMT2\u003c/em\u003e is a well-recognized histone methyltransferase to have the ability of regulating DNAm [\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e]. It is also a member of the Senescence Associated Secretory Phenotype (SASP) pathway [\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e], and has been observed to be downregulated in aging mouse hippocampus. Results from model organisms suggest \u003cem\u003eIRS1\u003c/em\u003e is involved in the IGF1/GH1 axis that could regulate aging [\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e], with loss of expression extending lifespan in mice [\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e] and \u003cem\u003eDrosophila\u003c/em\u003e [\u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e59\u003c/span\u003e]. \u003cem\u003eIGF2BP2\u003c/em\u003e has been identified as a key factor to activate stem cell aging [\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e] and brain aging [\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e] in mouse models, and also recognized as an epigenetic regulatory factor in epitranscriptomics [\u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e]. This warrants further investigation to clarify its roles in human aging, including its involvement in the insulin/IGF signaling pathways and the crosstalk between RNA and DNA modification. Metformin has been shown in multiple studies to exert lifespan-extending effects [\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e]. Here we identified four novel target genes: \u003cem\u003eNDUFS3\u003c/em\u003e, \u003cem\u003eNDUFAF1\u003c/em\u003e, \u003cem\u003eSLC22A3\u003c/em\u003e, and \u003cem\u003eTCF7L2\u003c/em\u003e, mainly involved in mitochondrial and energy metabolism pathways [\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e], that may underlie Metformin\u0026rsquo;s novel anti-aging mechanisms and merit further investigation. Notably, three of them: \u003cem\u003eEHMT2\u003c/em\u003e, \u003cem\u003eHSPA4\u003c/em\u003e and \u003cem\u003eNDUFS3\u003c/em\u003e, showed a brain-specific effect on aging, which were mediated by methylation variation across multiple CpG sites. This underscores the key role of epigenetic changes in the connecting central nervous system regulation and systemic aging.\u003c/p\u003e\u003cp\u003eThe strengths of this study included the widely used pipelines for instrument selection and a high resolution of tissue-specific results. In addition, we included a comprehensive set of aging outcomes, a wide range of drug target selection, and an emphasis on epigenetic regulation of aging. We also integrated external multi-omics evidence to enhance the interpretations of our findings, aiming to provide more informative and predictive results compared with previous drug-target MR studies on aging and offering implications for future targeted drug development [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e, \u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e, \u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e64\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eThis study still has limitations. First, the eQTL data used in this study are derived from a relatively small sample size of 838 individuals, with only one instrument available for each target gene per tissue. In addition, solely relying on lead cis-eQTLs for target genes in each tissue may limit the detection of potential bias [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e, \u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e65\u003c/span\u003e]. Second, some target genes, such as \u003cem\u003eTCF7L2\u003c/em\u003e, have qualified instruments in only a few tissues. This led to the incomplete exploration of the tissue-specific function of these genes. Third, the data for DNAm, the EAA group of aging indicators, and leukocyte telomere length were derived from blood samples, introducing bias due to sample incoherence. Fourth, for drugs like acarbose [\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e] and canagliflozin [\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e], the utilization of sex-stratified data, including sex-stratified eQTL data and sex-stratified GWASs of aging phenotype, may help elucidate gender-specific differences in epigenetic regulations [\u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e66\u003c/span\u003e]. Fifth, not all aging hallmarks listed in the consensus have publicly available GWAS datasets [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. Selection bias is inevitable in our study, which may contribute to the results that no gene was found to affect the epigenetic age acceleration group. More GWAS meta-analyses are required for better investigation. Sixth, several CpG sites showed a mediation proportion exceeding 100%. One of the three following conditions accounts for the phenomenon: (1) there are other mediators with a negative portion mediated; (2) the DNAm of each CpG site affects one another; (3) there are interactions between the DNAm of each CpG site [\u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e67\u003c/span\u003e, \u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e68\u003c/span\u003e]. Thus, the observed indirect effect requires further investigation. Seventh, the intervention of these drugs could have varying effects depending on age groups, ethnic backgrounds, and disease states. We need to assess the specific effects of these drug-gene expressions across diverse populations to understand the generalizability of our results in the near future. Finally, MR can only proxy the drug target effect rather than the direct effect of the drug on aging. Therefore, the evidence provided by this study cannot be used for any clinical investigations before formal clinical trials.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eIn conclusion, our study identified 9 target genes of 6 glucose-lowering drugs showing robust effects on longevity through DNA methylation. We identified both previously recognized and novel candidate genes that may contribute to lifespan extension by modulating aging-associated pathways, such as SASP signaling, mitochondrial and lipid metabolism, insulin signaling, and epitranscriptomic regulation. Several effects were tissue-specific, particularly within Five drug interacting genes exhibited a tissue-specific enrichment in the nervous system, emphasizing the neural regulation in systemic aging. Previous studies provide strong support for these findings. Further clinical and experimental research is required to validate and explore the feasibility and effectiveness of targeting these genes and drugs for aging prevention.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cdiv class=\"DefinitionList\"\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eDNAm\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eDNA methylation\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eEAA\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eEpigenetic age acceleration\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eEWAS\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eEpigenome-wide association study\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eFDR\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eFalse discovery rate\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eeQTL\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eExpression quantitative trait locus\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eGWAS\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eGenome-Wide Association Study\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eLD\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eLinkage disequilibrium\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003emQTL\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eDNAm quantitative trait locus\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eMR\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eMendelian randomization\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eRCT\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eRandomized controlled trial\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eTWAS\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eTranscriptome-wide association study\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eT2D\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eType 2 diabetes\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003eAvailability of data and materials\u003c/p\u003e\n\u003cp\u003eThe GWAS summary statistics for the ten primary aging outcomes are available from the GWAS Catalog database (www.ebi.ac.uk/gwas/). The GWAS summary statistics for glycemic traits are contributed by MAGIC investigators (www.magicinvestigators.org) and T2DGGI study (http://www.diagram-consortium.org/downloads.html). The analytical script of the MR analyses conducted in this study is available via the GitHub repository of the \u0026lsquo;TwoSampleMR\u0026rsquo; R package (github.com/MRCIEU/TwoSampleMR/). The analytical script of the genetic correlation analyses conducted in this study is available via the GitHub repository of the \u0026lsquo;ldsc\u0026rsquo; R package (github.com/mglev1n/ldscr/). The code used during the current study is available upon reasonable request from the corresponding author. All the other data generated or analysed during this study are included in this published article and the ESM files.\u003c/p\u003e\n\u003cp\u003eCompeting Interest\u003c/p\u003e\n\u003cp\u003eJ.Z and Y.B.S is the editorial board member of BMC medicine. Other authors declare that they have no competing interests.\u003c/p\u003e\n\u003cp\u003eFunding\u003c/p\u003e\n\u003cp\u003eThis study was supported by grants from the Noncommunicable Chronic Diseases-National Science and Technology Major Project (2024ZD0531500, 2024ZD0531502), Students\u0026rsquo; Innovation Training Program, Shanghai Jiao Tong University, School of Medicine, China (1824955Y) and grants from the National Key Research and Development Program of China (2022YFC2505203). Y.Q.S is supported by Roche University Research Fund, Chinese Society of Biochemistry and Molecular Biology, China (20240513). Z.J. and Q.Y. are supported by the National Natural Science Foundation of China (32570728 and 32500519). Y.B. is supported by the Shanghai Municipal Education Commission\u0026ndash;Gaofeng Clinical Medicine Grant Support (20161307 and 20152508 Round 2).\u003c/p\u003e\n\u003cp\u003eContribution Statements\u003c/p\u003e\n\u003cp\u003eY.Q.S., H.Z. and J.Z. conceived and designed the study. Y.Q.S drafted the manuscript. H.Z performed data analysis. M.M., S.F., Y.B.S., Q.Y. and J.Z. contributed to reviewing and editing the manuscript. Y.Q.S. and H.Z. contributed to visualization. Y.Q.S., H.Z., R.G. and M.W. contributed to data management and related paperwork. Y.Q.S. and H.Z. accessed and verified the data reported in the study. J.Z. is the guarantor of the findings. All the authors approved the final version to be published.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAcknowledgments\u003c/p\u003e\n\u003cp\u003eWe thank all of the investigators and participants who provided these data to support this study. We thank Huiling Zhao, Haoyu Liu and Xueyan Wu for the help of device acquisition and instructions.\u003c/p\u003e\n\u003cp\u003eAuthor\u0026rsquo;s relationship and activities\u003c/p\u003e\n\u003cp\u003eY.Q.S., H.Z., R.G. and M.W. are members of the College Innovation Training Program in Shanghai Jiao Tong University, School of Medicine. All authors declare that there are no relationships or activities that might bias, or be perceived to bias this work.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eLopez-Otin C, Blasco MA, Partridge L, Serrano M, Kroemer G. Hallmarks of aging: An expanding universe. Cell. 2023;186(2):243\u0026ndash;78.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGladyshev VN, Kritchevsky SB, Clarke SG, Cuervo AM, Fiehn O, de Magalhaes JP, Mau T, Maes M, Moritz R, Niedernhofer LJ, et al. Molecular Damage in Aging. Nat Aging. 2021;1(12):1096\u0026ndash;106.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSeale K, Horvath S, Teschendorff A, Eynon N, Voisin S. Making sense of the ageing methylome. Nat Rev Genet. 2022;23(10):585\u0026ndash;605.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDor Y, Cedar H. Principles of DNA methylation and their implications for biology and medicine. Lancet. 2018;392(10149):777\u0026ndash;86.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGuarente L, Sinclair DA, Kroemer G. Human trials exploring anti-aging medicines. Cell Metab. 2024;36(2):354\u0026ndash;76.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKulkarni AS, Brutsaert EF, Anghel V, Zhang K, Bloomgarden N, Pollak M, Mar JC, Hawkins M, Crandall JP, Barzilai N. Metformin regulates metabolic and nonmetabolic pathways in skeletal muscle and subcutaneous adipose tissues of older adults. Aging Cell 2018, 17(2).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePadki MM, Stambler I. Targeting Aging with Metformin (TAME). In: \u003cem\u003eEncyclopedia of Gerontology and Population Aging.\u003c/em\u003e edn. Edited by Gu D, Dupre ME. Cham: Springer International Publishing; 2021: 4908\u0026ndash;4910.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eOchoa D, Karim M, Ghoussaini M, Hulcoop DG, McDonagh EM, Dunham I. Human genetics evidence supports two-thirds of the 2021 FDA-approved drugs. Nat Rev Drug Discov. 2022;21(8):551.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZheng J, Xu M, Walker V, Yuan J, Korologou-Linden R, Robinson J, Huang P, Burgess S, Au Yeung SL, Luo S, et al. Evaluating the efficacy and mechanism of metformin targets on reducing Alzheimer's disease risk in the general population: a Mendelian randomisation study. Diabetologia. 2022;65(10):1664\u0026ndash;75.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLuo S, Wong ICK, Chui CSL, Zheng J, Huang Y, Schooling CM, Yeung SLA. Effects of putative metformin targets on phenotypic age and leukocyte telomere length: a mendelian randomisation study using data from the UK Biobank. Lancet Healthy Longev. 2023;4(7):e337\u0026ndash;44.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChen Z, Wu X, Yang Q, Zhao H, Ying H, Liu H, Wang C, Zheng R, Lin H, Wang S, et al. The Effect of SGLT2 Inhibition on Brain-related Phenotypes and Aging: A Drug Target Mendelian Randomization Study. J Clin Endocrinol Metab. 2025;110(4):1096\u0026ndash;104.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGuo W, Zhao L, Huang W, Chen J, Zhong T, Yan S, Hu W, Zeng F, Peng C, Yan H. Sodium-glucose cotransporter 2 inhibitors, inflammation, and heart failure: a two-sample Mendelian randomization study. Cardiovasc Diabetol. 2024;23(1):118.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDeMarsilis A, Reddy N, Boutari C, Filippaios A, Sternthal E, Katsiki N, Mantzoros C. Pharmacotherapy of type 2 diabetes: An update and future directions. Metabolism. 2022;137:155332.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFreshour SL, Kiwala S, Cotto KC, Coffman AC, McMichael JF, Song JJ, Griffith M, Griffith OL, Wagner AH. Integration of the Drug-Gene Interaction Database (DGIdb 4.0) with open crowdsource efforts. Nucleic Acids Res. 2021;49(D1):D1144\u0026ndash;51.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eConsortium GT. The GTEx Consortium atlas of genetic regulatory effects across human tissues. Science. 2020;369(6509):1318\u0026ndash;30.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHemani G, Tilling K, Davey Smith G. Orienting the causal relationship between imprecisely measured traits using GWAS summary data. PLoS Genet. 2017;13(11):e1007081.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMoqri M, Herzog C, Poganik JR, Biomarkers of Aging C, Justice J, Belsky DW, Higgins-Chen A, Moskalev A, Fuellen G, Cohen AA, et al. Biomarkers of aging for the identification and evaluation of longevity interventions. Cell. 2023;186(18):3758\u0026ndash;75.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMao R, Li J, Xiao W. Identification of prospective aging drug targets via Mendelian randomization analysis. Aging Cell. 2024;23(7):e14171.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMin JL, Hemani G, Hannon E, Dekkers KF, Castillo-Fernandez J, Luijk R, Carnero-Montoro E, Lawson DJ, Burrows K, Suderman M, et al. Genomic and phenotypic insights from an atlas of genetic effects on DNA methylation. Nat Genet. 2021;53(9):1311\u0026ndash;21.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGenomes Project C, Abecasis GR, Auton A, Brooks LD, DePristo MA, Durbin RM, Handsaker RE, Kang HM, Marth GT, McVean GA. An integrated map of genetic variation from 1,092 human genomes. Nature. 2012;491(7422):56\u0026ndash;65.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChen J, Spracklen CN, Marenne G, Varshney A, Corbin LJ, Luan J, Willems SM, Wu Y, Zhang X, Horikoshi M, et al. The trans-ancestral genomic architecture of glycemic traits. Nat Genet. 2021;53(6):840\u0026ndash;60.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSuzuki K, Hatzikotoulas K, Southam L, Taylor HJ, Yin X, Lorenz KM, Mandla R, Huerta-Chagoya A, Melloni GEM, Kanoni S, et al. Genetic drivers of heterogeneity in type 2 diabetes pathophysiology. Nature. 2024;627(8003):347\u0026ndash;57.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBulik-Sullivan B, Finucane HK, Anttila V, Gusev A, Day FR, Loh PR, ReproGen C, Psychiatric Genomics C, Genetic Consortium for Anorexia Nervosa of the Wellcome Trust Case Control, Duncan C. An atlas of genetic correlations across human diseases and traits. Nat Genet. 2015;47(11):1236\u0026ndash;41.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBulik-Sullivan BK, Loh PR, Finucane HK, Ripke S, Yang J, Schizophrenia Working Group of the Psychiatric, Genomics C, Patterson N, Daly MJ, Price AL, Neale BM. LD Score regression distinguishes confounding from polygenicity in genome-wide association studies. Nat Genet. 2015;47(3):291\u0026ndash;5.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFinucane HK, Bulik-Sullivan B, Gusev A, Trynka G, Reshef Y, Loh PR, Anttila V, Xu H, Zang C, Farh K, et al. Partitioning heritability by functional annotation using genome-wide association summary statistics. Nat Genet. 2015;47(11):1228\u0026ndash;35.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eVabistsevits M, Davey Smith G, Sanderson E, Richardson TG, Lloyd-Lewis B, Richmond RC. Deciphering how early life adiposity influences breast cancer risk using Mendelian randomization. Commun Biol. 2022;5(1):337.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBurgess S, Zuber V, Valdes-Marquez E, Sun BB, Hopewell JC. Mendelian randomization with fine-mapped genetic data: Choosing from large numbers of correlated instrumental variables. Genet Epidemiol. 2017;41(8):714\u0026ndash;25.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZheng J, Baird D, Borges MC, Bowden J, Hemani G, Haycock P, Evans DM, Smith GD. Recent Developments in Mendelian Randomization Studies. Curr Epidemiol Rep. 2017;4(4):330\u0026ndash;45.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSkrivankova VW, Richmond RC, Woolf BAR, Yarmolinsky J, Davies NM, Swanson SA, VanderWeele TJ, Higgins JPT, Timpson NJ, Dimou N, et al. Strengthening the Reporting of Observational Studies in Epidemiology Using Mendelian Randomization: The STROBE-MR Statement. JAMA. 2021;326(16):1614\u0026ndash;21.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSanderson E, Glymour MM, Holmes MV, Kang H, Morrison J, Munafo MR, Palmer T, Schooling CM, Wallace C, Zhao Q et al. Mendelian randomization. Nat Rev Methods Primers 2022, 2.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBrumpton B, Sanderson E, Heilbron K, Hartwig FP, Harrison S, Vie GA, Cho Y, Howe LD, Hughes A, Boomsma DI, et al. Avoiding dynastic, assortative mating, and population stratification biases in Mendelian randomization through within-family analyses. Nat Commun. 2020;11(1):3519.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZheng J, Haberland V, Baird D, Walker V, Haycock PC, Hurle MR, Gutteridge A, Erola P, Liu Y, Luo S, et al. Phenome-wide Mendelian randomization mapping the influence of the plasma proteome on complex diseases. Nat Genet. 2020;52(10):1122\u0026ndash;31.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eUhlen M, Fagerberg L, Hallstrom BM, Lindskog C, Oksvold P, Mardinoglu A, Sivertsson A, Kampf C, Sjostedt E, Asplund A et al. Proteomics. Tissue-based map of the human proteome. \u003cem\u003eScience\u003c/em\u003e 2015, 347(6220):1260419.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChen EY, Tan CM, Kou Y, Duan Q, Wang Z, Meirelles GV, Clark NR. Ma'ayan A: Enrichr: interactive and collaborative HTML5 gene list enrichment analysis tool. BMC Bioinformatics. 2013;14:128.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKuleshov MV, Jones MR, Rouillard AD, Fernandez NF, Duan Q, Wang Z, Koplev S, Jenkins SL, Jagodnik KM, Lachmann A, et al. Enrichr: a comprehensive gene set enrichment analysis web server 2016 update. Nucleic Acids Res. 2016;44(W1):W90\u0026ndash;97.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eXie Z, Bailey A, Kuleshov MV, Clarke DJB, Evangelista JE, Jenkins SL, Lachmann A, Wojciechowicz ML, Kropiwnicki E, Jagodnik KM, et al. Gene Set Knowledge Discovery with Enrichr. Curr Protoc. 2021;1(3):e90.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBattram T, Yousefi P, Crawford G, Prince C, Sheikhali Babaei M, Sharp G, Hatcher C, Vega-Salas MJ, Khodabakhsh S, Whitehurst O, et al. The EWAS Catalog: a database of epigenome-wide association studies. Wellcome Open Res. 2022;7:41.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLu M, Zhang Y, Yang F, Mai J, Gao Q, Xu X, Kang H, Hou L, Shang Y, Qain Q, et al. TWAS Atlas: a curated knowledgebase of transcriptome-wide association studies. Nucleic Acids Res. 2023;51(D1):D1179\u0026ndash;87.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZenin A, Tsepilov Y, Sharapov S, Getmantsev E, Menshikov LI, Fedichev PO, Aulchenko Y. Identification of 12 genetic loci associated with human healthspan. Commun Biol. 2019;2:41.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMartens M, Ammar A, Riutta A, Waagmeester A, Slenter DN, Hanspers K, Digles RAM, Lopes D, Ehrhart EN. WikiPathways: connecting communities. Nucleic Acids Res. 2021;49(D1):D613\u0026ndash;21.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eEdgar R, Domrachev M, Lash AE. Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. Nucleic Acids Res. 2002;30(1):207\u0026ndash;10.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHarrison DE, Strong R, Alavez S, Astle CM, DiGiovanni J, Fernandez E, Flurkey K, Garratt M, Gelfond JAL, Javors MA, et al. Acarbose improves health and lifespan in aging HET3 mice. Aging Cell. 2019;18(2):e12898.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJiang E, Dinesh A, Jadhav S, Miller RA, Garcia GG. Canagliflozin shares common mTOR and MAPK signaling mechanisms with other lifespan extension treatments. Life Sci. 2023;328:121904.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eShihab EM, Kadhim HM, Shahooth SS. Dapagliflozin mitigates oxidative stress, inflammatory, and histopathological markers of aging in mice. J Med Life. 2024;17(2):157\u0026ndash;63.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRamsey KM, Mills KF, Satoh A, Imai S. Age-associated loss of Sirt1-mediated enhancement of glucose-stimulated insulin secretion in beta cell-specific Sirt1-overexpressing (BESTO) mice. Aging Cell. 2008;7(1):78\u0026ndash;88.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMartin-Montalvo A, Mercken EM, Mitchell SJ, Palacios HH, Mote PL, Scheibye-Knudsen M, Gomes AP, Ward TM, Minor RK, Blouin MJ, et al. Metformin improves healthspan and lifespan in mice. Nat Commun. 2013;4:2192.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJafari M, Khodayari B, Felgner J, Bussel II, Rose MR, Mueller LD. Pioglitazone: an anti-diabetic compound with anti-aging properties. Biogerontology. 2007;8(6):639\u0026ndash;51.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMohammed I, Hollenberg MD, Ding H, Triggle CR. A Critical Review of the Evidence That Metformin Is a Putative Anti-Aging Drug That Enhances Healthspan and Extends Lifespan. Front Endocrinol (Lausanne). 2021;12:718942.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eStevenson-Hoare J, Leonenko G, Escott-Price V. Comparison of long-term effects of metformin on longevity between people with type 2 diabetes and matched non-diabetic controls. BMC Public Health. 2023;23(1):804.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLa Grotta R, Frige C, Matacchione G, Olivieri F, de Candia P, Ceriello A, Prattichizzo F. Repurposing SGLT-2 Inhibitors to Target Aging: Available Evidence and Molecular Mechanisms. Int J Mol Sci 2022, 23(20).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYen CL, Wu CY, Tsai CY, Lee CC, Li YJ, Peng WS, Liu JR, Liu YC, Jenq CC, Yang HY, et al. Pioglitazone reduces cardiovascular events and dementia but increases bone fracture in elderly patients with type 2 diabetes mellitus: a national cohort study. Aging. 2023;15(7):2721\u0026ndash;33.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSuissa S. Immortal time bias in pharmaco-epidemiology. Am J Epidemiol. 2008;167(4):492\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eShen D, Li H, Zhou R, Liu MJ, Yu H, Wu DF. Pioglitazone attenuates aging-related disorders in aged apolipoprotein E deficient mice. Exp Gerontol. 2018;102:101\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eXu L, Ma X, Verma N, Perie L, Pendse J, Shamloo S, Marie Josephson A, Wang D, Qiu J, Guo M, et al. PPARgamma agonists delay age-associated metabolic disease and extend longevity. Aging Cell. 2020;19(11):e13267.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSeok H, Lee M, Shin E, Yun MR, Lee YH, Moon JH, Kim E, Lee PH, Lee BW, Kang ES, et al. Low-dose pioglitazone can ameliorate learning and memory impairment in a mouse model of dementia by increasing LRP1 expression in the hippocampus. Sci Rep. 2019;9(1):4414.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTachibana M, Sugimoto K, Fukushima T, Shinkai Y. Set domain-containing protein, G9a, is a novel lysine-preferring mammalian histone methyltransferase with hyperactivity and specific selectivity to lysines 9 and 27 of histone H3. J Biol Chem. 2001;276(27):25309\u0026ndash;17.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ede Magalhaes JP, Abidi Z, Dos Santos GA, Avelar RA, Barardo D, Chatsirisupachai K, Clark P, De-Souza EA, Johnson EJ, Lopes I, et al. Human Ageing Genomic Resources: updates on key databases in ageing research. Nucleic Acids Res. 2024;52(D1):D900\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSelman C, Lingard S, Choudhury AI, Batterham RL, Claret M, Clements M, Ramadani F, Okkenhaug K, Schuster E, Blanc E, et al. Evidence for lifespan extension and delayed age-related biomarkers in insulin receptor substrate 1 null mice. FASEB J. 2008;22(3):807\u0026ndash;18.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eClancy DJ, Gems D, Harshman LG, Oldham S, Stocker H, Hafen E, Leevers SJ, Partridge L. Extension of life-span by loss of CHICO, a Drosophila insulin receptor substrate protein. Science. 2001;292(5514):104\u0026ndash;6.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSuo M, Rommelfanger MK, Chen Y, Amro EM, Han B, Chen Z, Szafranski K, Chakkarappan SR, Boehm BO, MacLean AL, et al. Age-dependent effects of Igf2bp2 on gene regulation, function, and aging of hematopoietic stem cells in mice. Blood. 2022;139(17):2653\u0026ndash;65.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLi M, Guo H, Carey M, Huang C. Transcriptional and epigenetic dysregulation impairs generation of proliferative neural stem and progenitor cells during brain aging. Nat Aging. 2024;4(1):62\u0026ndash;79.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHuang H, Weng H, Sun W, Qin X, Shi H, Wu H, Zhao BS, Mesquita A, Liu C, Yuan CL, et al. Recognition of RNA N(6)-methyladenosine by IGF2BP proteins enhances mRNA stability and translation. Nat Cell Biol. 2018;20(3):285\u0026ndash;95.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChen H, Tang X, Su W, Li S, Yang R, Cheng H, Zhang G, Zhou X. Causal effects of lipid-lowering therapies on aging-related outcomes and risk of cancers: a drug-target Mendelian randomization study. Aging. 2023;15(24):15228\u0026ndash;42.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMartin L, Boutwell BB, Messerlian C, Adams CD. Mendelian randomization reveals apolipoprotein B shortens healthspan and possibly increases risk for Alzheimer's disease. Commun Biol. 2024;7(1):230.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePalmer TM, Lawlor DA, Harbord RM, Sheehan NA, Tobias JH, Timpson NJ, Davey Smith G, Sterne JA. Using multiple genetic variants as instrumental variables for modifiable risk factors. Stat Methods Med Res. 2012;21(3):223\u0026ndash;42.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eOliva M, Munoz-Aguirre M, Kim-Hellmuth S, Wucher V, Gewirtz ADH, Cotter DJ, Parsana P, Kasela S, Balliu B, Vinuela A et al. The impact of sex on gene expression across human tissues. \u003cem\u003eScience\u003c/em\u003e 2020, 369(6509).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eVanderWeele TJ. Mediation Analysis: A Practitioner's Guide. Annu Rev Public Health. 2016;37:17\u0026ndash;32.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eVanderWeele TJ, Vansteelandt S. Mediation Analysis with Multiple Mediators. Epidemiol Methods. 2014;2(1):95\u0026ndash;115.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-medicine","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bmed","sideBox":"Learn more about [BMC Medicine](http://bmcmedicine.biomedcentral.com/)","snPcode":"12916","submissionUrl":"https://submission.nature.com/new-submission/12916/3","title":"BMC Medicine","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Aging, DNA methylation, Anti-aging drugs, Anti-diabetic drugs, Mendelian randomization","lastPublishedDoi":"10.21203/rs.3.rs-7798894/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7798894/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDNA methylation plays a key role in mediating the anti-aging effects of glucose-lowering drugs. This study aims to systematically explore the potential anti-aging effects of target genes of FDA-approved glucose-lowering drugs and the underlying epigenetic mediators.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe conducted a two-sample Mendelian randomization (MR) study to investigate the putative causal relationships between the gene expression levels of glucose-lowering drug targets and 10 aging-related phenotypes, followed by a two-step MR to estimate the mediation effect of DNA methylation. Drug candidates were selected according to the latest review of clinical drug use for type 2 diabetes, and their target genes were obtained from the DGIdb database. Tissue-specific \u003cem\u003ecis-\u003c/em\u003eexpression quantitative trait loci (eQTLs) from GTEx consortium were selected as genetic instruments to proxy the expression level of drug-target genes. Glycemic phenotypes were used as positive controls to validate the instruments. The \u003cem\u003ecis-\u003c/em\u003e and \u003cem\u003etrans-\u003c/em\u003emethylation QTLs of Cytosine-phosphate-Guanine sites near the drug target genes were obtained from GoDMC consortium. Additionally, we performed enrichment analyses focused on tissue specificity and aging pathways to further corroborate our findings.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe obtained 194 target genes interacted with 36 FDA-approved anti-diabetic drugs, of which the tissue-specific eQTLs were used to proxy the drug target effects. MR showed strong evidence that 9 interacting genes of 6 glucose-lowering drugs showed anti-aging potential on one or more aging-related phenotypes mediated by DNA methylation: \u003cem\u003eEHMT2\u003c/em\u003e, \u003cem\u003eHSPA4\u003c/em\u003e, \u003cem\u003eIGF2BP2\u003c/em\u003e, \u003cem\u003eIRS1\u003c/em\u003e, \u003cem\u003eLPL\u003c/em\u003e, \u003cem\u003eNDUFAF1\u003c/em\u003e, \u003cem\u003eNDUFS3\u003c/em\u003e, \u003cem\u003eSLC22A3\u003c/em\u003e and \u003cem\u003eTCF7L2\u003c/em\u003e. These genes were distributed in 17 tissues, especially in the central nervous system, suggesting a potential neural component in their anti-aging effects. For instance, expression of \u003cem\u003eEHMT2\u003c/em\u003e in several brain basal ganglia regions, which the gene interacted with Tolazamide, showed a protective effect on frailty (odds ratio[OR] in caudate = 1.02, 95%CI = 1.01–1.04, FDR adjusted P = 1.69×10\u003csup\u003e− 2\u003c/sup\u003e; OR in putamen = 1.02, 95%CI = 1.01–1.03, P\u003csub\u003eFDR\u003c/sub\u003e=3.37×10\u003csup\u003e− 2\u003c/sup\u003e, OR in nucleus accumbens = 1.02, 95%CI = 1.01–1.04, P\u003csub\u003eFDR\u003c/sub\u003e=3.37×10\u003csup\u003e− 2\u003c/sup\u003e). These associations were externally validated by searching literature evidence in existing EWAS and TWAS studies, as well as evidence from enricment analyses.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study prioritizes nine glucose-lowering genes as anti-aging drug targets in specific tissues and prioritizes their epigenetic regulation through DNA methylation for future drug development.\u003c/p\u003e","manuscriptTitle":"The tissue-specific effects of glucose-lowering drug targets on aging mediated through DNA methylation: a multi-omics genetic study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-10-21 13:25:46","doi":"10.21203/rs.3.rs-7798894/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-01-15T17:03:28+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-01-15T04:59:23+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-01-02T16:03:10+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"62073494815367515159356066281419251085","date":"2025-12-19T03:31:56+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"179586703990179026866191843140823354564","date":"2025-11-13T02:31:31+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-10-08T16:35:15+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-10-08T04:55:33+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-10-08T04:46:44+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Medicine","date":"2025-10-07T10:56:04+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-medicine","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bmed","sideBox":"Learn more about [BMC Medicine](http://bmcmedicine.biomedcentral.com/)","snPcode":"12916","submissionUrl":"https://submission.nature.com/new-submission/12916/3","title":"BMC Medicine","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"3c9cac93-243b-4f88-8c13-7c37f887e52a","owner":[],"postedDate":"October 21st, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-04-09T16:40:59+00:00","versionOfRecord":[],"versionCreatedAt":"2025-10-21 13:25:46","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7798894","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7798894","identity":"rs-7798894","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}