[Multi-omics Mendelian randomization study on the causality between non-ionizing radiation and facial aging].

Objective: To investigate the causality between non-ionizing radiation and facial aging, and to identify potential genes associated with facial aging. Methods: This study employed a method of analysis based on multiple Mendelian randomization (MR). Genome-wide association study data of non-ionizing radiation (FinnGen database, n=218 281) and facial aging (UK Biobank database, n=423 999) were retrieved. Single nucleotide polymorphisms (SNPs) were used as instrumental variables, with a significance threshold (P<5×10^-6) applied and further linkage disequilibrium analysis performed to select SNPs associated with non-ionizing radiation. Two-sample MR (TSMR) analysis was conducted to assess the causality between non-ionizing radiation and facial aging, using inverse variance weighting (IVW) method as the primary analytical method and supplementing with MR-Egger regression, weighted median, weighted mode, and simple mode methods for validation. For the selected non-ionizing radiation-associated SNPs, heterogeneity was tested by Cochran Q test, horizontal pleiotropy was assessed by the MR-Egger intercept test and MR-PRESSO test, and robustness was evaluated via leave-one-out analysis. Multivariable MR (MVMR) analysis was performed to adjust for confounding factors affecting facial aging including smoking frequency, blood alcohol concentration, exercise frequency, body mass index, and systolic and diastolic blood pressure. Summary-data-based MR (SMR) analysis using expression quantitative trait loci (eQTL) data was conducted to screen candidate genes of facial aging, which were then validated by TSMR analysis. Protein quantitative trait loci (pQTL) and methylation quantitative trait loci (mQTL) data were analyzed by TSMR analysis to examine the causal role of MED1 gene with facial aging from multi-omics aspect. The genetic association of MED1 gene with facial aging was verified by colocalization analysis (posterior probability H4>50%). Results: Twenty non-ionizing radiation-related SNPs that reached the significance threshold were screened out, with F values being all >10. IVW analysis demonstrated a positive causality between non-ionizing radiation and facial aging (with odds ratio of 1.02, with 95% confidence interval of 1.01-1.02, P<0.05). The analysis results of MR-Egger regression, weighted median, simple mode method, and weighted mode method (with odds ratios of 1.02, 1.02, 1.01, and 1.01, respectively, with 95% confidence intervals of 1.01-1.03, 1.01-1.02, 0.99-1.02, respectively, P<0.05) were consistent with IVW method. For these 20 non-ionizing radiation-related SNPs, Cochran Q test under IVW method and MR-Egger showed no significant heterogeneity (with Q values of 23.20 and 22.59, respectively, P>0.05); the MR-Egger intercept test (with intercept absolute value of 0.01, with standard error of 0.01, P>0.05) and MR-PRESSO test (P>0.05) indicated no horizontal pleiotropy. Leave-one-out analysis further confirmed that no individual SNP had a significant effect on the results. After correction of confounding factors such as systolic blood pressure, diastolic blood pressure, smoking frequency, blood alcohol concentration, body mass index, and exercise frequency, MVMR analysis showed that non-ionizing radiation remained a risk factor for facial aging (with odds ratios of 1.01, 1.01, 1.02, 1.02, 1.01, and 1.04, respectively, with 95% confidence intervals of 1.01-1.02, 1.01-1.02, 1.01-1.02, 1.01-1.02, 1.00-1.01, and 1.03-1.05, respectively, all P values <0.05). SMR analysis identified 12 potential facial aging-related genes (SENP7, CCND1, LTBP2, IKZF3, MED1, ORMDL3, ZBTB7B, LOX, NEBL, EXOSC6, PSMA4, and EIF2B2, with odds ratios of 1.01, 1.03, 1.04, 0.99, 1.04, 1.01, 1.06, 0.88, 1.01, 0.99, 1.04, and 0.99, respectively, all P values <0.05). Subsequent TSMR analysis retained 6 risk genes (ZBTB7B, SENP7, NEBL, MED1, PSMA4, and ORMDL3, with odds ratios of 1.04, 1.01, 1.00, 1.02, 1.03, and 1.01, respectively, with 95% confidence intervals of 1.02-1.05, 1.00-1.01, 1.00-1.01, 1.01-1.03, 1.01-1.04, and 1.00-1.01, respectively, all P values <0.05) for facial aging and 4 protective genes (LOX, EIF2B2, EXOSC6, and IKZF3, with odds ratios of 0.92, 0.99, 0.99, and 0.99, respectively, with 95% confidence intervals of 0.90-0.94, 0.99-0.99, 0.99-1.00, and 0.99-1.00, respectively, all P values <0.05). TSMR analysis based on pQTL data showed the MED1 protein was positively associated with facial aging (with odds ratio of 1.04, P<0.05), which was consistent with the causal direction observed in eQTL-based SMR and TSMR analyses. TSMR analysis based on mQTL data indicated MED1 gene methylation (with probes of cg15445000 and cg03013999) had a protective effect on facial aging (with odds ratios of 0.99 and 0.99, respectively, both P values <0.05). Colocalization analysis yielded a posterior probability H4=58.4%, suggesting that MED1 gene and facial aging likely shared the same causal genetic variant. Conclusions: Through multi-omics MR analyses, it has confirmed that there is a causality between non-ionizing radiation and facial aging, which remained highly significant after correcting for potential confounders such as smoking frequency, blood alcohol concentration, exercise frequency, and the others. Clearly, 10 genes including SENP7, NEBL, EIF2B2, PSMA4, EXOSC6, IKZF3, ORMDL3, ZBTB7B, LOX, and MED1, particularly the MED1, may be involved in the process of facial aging.