Related Factors with Vascular Dementia: A Two-Sample Mendelian Randomization Study

Abstract

Pathogenesis of vascular dementia (VD) is still unclear, there are currently no effective prevention and treatment methods. We applied Mendelian randomization (MR) using summary statistics from large-scale GWAS of metabolites and VD to reveal the causal effect of metabolites on the VD. One set of genetics instrument was used for analysis, derived from publicly available genetic summary data. Which was 32 single-nucleotide polymorphisms robustly associated with metabolites. Inverse-variance weighted, weighted median method, MR-Egger regression, and MR Pleiotropy RESidual Sum and Outlier test were used for MR analyses. Strong evidence for a positive effect of metabolites, which means N6-threonylcarbamoyladenosine (t⁶A) on VD was found in inverse-variance weighted (odds ratios [OR]: 0.667, 95% confidence interval [CI]: 0.548–0.812, p < 0.001), MR-Egger (OR: 0.647, 95% CI: 0.458–0.913, p = 0.019), and weighted median (OR: 0.650, 95% CI: 0.466–0.908, p = 0.012). The MR analysis indicated that metabolites (t⁶A) may be causally associated with a positive effect on VD.

Introduction

Vascular dementia (VD) accounts for about 20% of dementia types and is the second largest type of dementia after Alzheimer’s disease (AD).¹ Most research on dementia focuses on AD. However, in recent years, the incidence rate of VD has increased year by year, reaching 1.1%–3.0%, and the prevalence rate increases with age.² VD is a major syndrome characterized by cognitive impairment caused by risk factors for cerebrovascular diseases, including obvious or inconspicuous cerebrovascular diseases.³ Its main clinical features include low computational ability, emotional loss, decreased comprehension, and decline or even disappearance of emotional, cognitive, speech, and memory functions.⁴ The memory, understanding, and expression abilities of most patients will be affected for a long time.⁵ The pathogenesis of VD is relatively complex, and most scholars believe that the pathogenesis of VD may be related to cellular inflammatory response and oxidative stress.⁶ The pathogenesis of VD is still unclear, and there is currently no effective prevention and treatment method.⁷ By analyzing the changes in metabolites during the occurrence and development of VD, we have the potential to discover key metabolic pathways related to disease progression and screen for biomarkers with clinical application potential. However, due to the high dimensionality and diversity of metabolomics data, processing and analyzing large-scale metabolite profile data has become quite complex. Traditional biostatistics methods may not fully reveal potential patterns and correlations in the data, and more precise and effective data mining and statistical analysis methods are needed.⁸ This study provides an effective and convenient method to map the genetic structure of metabolites, laying the foundation for a deeper understanding of the role of genes in common diseases and therapeutic targets.

Mendelian randomization (MR) is an epidemiological research method that infers causal relationships between diseases and the effects of confounding factors under specific exposure conditions.⁹ It can improve the accuracy of causal inference by utilizing genetic variation to solve the problem of reverse causality and interference factors.¹⁰ MR has been applied in various disease fields and has made significant progress.¹¹ However, the instrumental variables in MR research must meet the three core assumptions of association, independence, and exclusivity.¹² Even if the core hypothesis is established, the application of MR research in causal inference may be influenced by issues such as pleiotropy, population stratification, and developmental compensation, resulting in bias or confusion. Therefore, these issues need to be detected and corrected through comprehensive methods and data.¹³ This study utilized the advantages of MR analysis and used single-nucleotide polymorphism (SNP) as an instrumental variable to screen for metabolite phenotypes with causal effects on VD.¹⁴ Because the MR analysis results represent the long-term stable genetic effects of metabolite phenotype,¹⁵ this will provide a good research approach for exploring the relationship between metabolite phenotypes with long-term genetic effects and VD and for in-depth analysis of the mechanism of metabolite action in the occurrence, development, and treatment of VD. By comprehensively and accurately analyzing the metabolic characteristics in VD, we hope to provide a more precise, stable, and novel perspective for a deeper understanding of the pathogenesis of VD and to provide the scientific basis for the development of personalized long-term treatment strategies in the future.

Materials and Methods

An overview of the study design is presented in Figure 1.

FIG. 1.

Workflow of MR study revealing causality from metabolites on VD. IVW, inverse-variance weighted; MR, Mendelian randomization; MR-PRESSO, MR Pleiotropy RESidual Sum and Outlier; SNPs, single-nucleotide polymorphisms; VD, vascular dementia.

Data sources for metabolites, metabolite ratios, and VD

The GWAS summary data for 1400 metabolites is sourced from the Protein Data Bank in Europe database (http://ftp.ebi.ac.uk/pub/databases/gwas/summary_statistics/) The ID range of the data is GCST90199621-GCST90201020, which includes 1091 blood metabolites and 309 metabolite ratio data. The metabolite phenotype-related gene set is sourced from GeneCards: The Human Gene Database (https://www.genecards.org/). The summary-level GWAS data correlated with 1091 blood metabolites and 309 metabolite ratios were obtained from an MR analysis of the GWAS of North America ancestry participants from the Canadian Longitudinal Study on Aging cohort. Ninety-four effector genes and 48 metabolite ratios were identified for 109 metabolites. Using MR, more than 20 metabolites and ratios were identified, which are associated with 12 diseases, including whey salts α-hydroxy isovalerate and ergot. The details of the participant characteristics of the metabolome prioritize Gen Consortium studies have been reported by Chen.¹⁶

In terms of genetic variation selection related to VD, genetic association estimates for the outcomes were obtained from the FinnGen consortium. The information of these studies is VD (GWAS ID: finn-b-F5_VASCDEM). The relevant data have been publicly published and can be found on the website (https://gwas.mrcieu.ac.uk/datasets/) download from the article.¹⁷ Totally, there were 881 VD cases and 211,508 control cases. Overall, 16,380,457 SNPs were extracted. The subjects were diagnosed with VD through autopsy or clinical diagnostic criteria. To avoid bias due to population stratification, all the genetic data we selected were sourced from the North America population.

MR analysis

Using the MR analysis method to explore the causal relationship between metabolites and VD. The MR framework utilizes genetic variations as instrumental variables (IVs) to infer causal relationships by simulating the random allocation of alleles during meiosis. To ensure the robustness of MR analysis, we have established strict selection criteria for instrumental variables for each metabolite. Select SNPs (p < 1e-5) strongly associated with specific metabolites as instrumental variables. To ensure the independence of each instrumental variable, a clustering method based on linkage disequilibrium (LD) was used (LD r ² threshold = 0.001, window size = 10,000 kb) to remove SNPs with strong LD. Use the F-statistic to quantify the strength of each SNP as an instrumental variable, where F > 10 indicates a lower likelihood of weak instrumental variable bias. Using five different statistical methods, namely MR Egger, weighted median, inverse-variance weighted (IVW), simple mode, and weighted mode, evaluate the causal effects of 1400 metabolites on VD and calculate the corresponding odds ratios (OR). Among them, IVW results are considered the main indicator for evaluating causal effects, and p < 0.05 is considered statistically significant. Using Cochran’s Q statistic to evaluate heterogeneity in MR estimation, a p < 0.05 indicates significant heterogeneity. The intercept term in MR-Egger regression was used to evaluate directional pleiotropy, and p > 0.05 indicates the absence of horizontal pleiotropy and conforms to MR principles. Strict quality control procedures are used to screen metabolites for subsequent analysis, with screening conditions including IVW statistical results (p < 0.05), consistency of causal effects evaluated by five MR statistical methods (quantified based on the positive and negative consistency of OR values), and pleiotropy indicators (p > 0.05). The above steps are all executed using the R package “TwoSampleMR” (version 0.5.7).

Filtering of tool variables

We screened GWAS data to include SNPs significantly associated with the disease. Based on the MR analysis results of disease level, p value threshold of SNPs with correlation was set to p < 5 × 10⁻⁸, the parameter r ² threshold is set to 0.001, and the Clumping distance is 10,000 kb. All SNPs extracted beta, SE, and p from summary data of VD GWAS with R; if not available, using proxy SNPs (R ² > 0.8). In order to ensure a strong correlation between instrumental variables and endogenous variables and avoid weak instrumental bias, we calculated R ² for each SNP, representing the proportion of variance explained by the instrumental variable SNP; F-statistic, used to evaluate the strength of instrumental variables.

Statistical analysis

All analyses were performed using the packages TwoSampleMR (version 0.4.25) and MRPRESSO (version 1.0) in R (version 3.6.2) packages. This MR study mainly uses the IVW method to explore the causal relationship between various influencing factors and VD. To ensure the robustness of the statistical results, we simultaneously used a weighted median estimator (WME) and MR-Egger regression based on Egger regression to perform sensitivity analysis on the statistical results. IVW is considered the standard method for summarizing data in MR. This method uses the Wald ratio method to estimate the causal effects of each included tool SNP and then performs a weighted summary analysis. The weighted median estimation method only requires that at least 50% of the weights contributed by genetic variation are valid for statistical calculation. MR-Egger regression can find and correct the problem of multiple validity, which requires that the included instrumental variables meet the InSDEI hypothesis (instrument strength independent of direct effect, InSDEI), that is, the correlation between tool exposure and tool result is assumed to be independent. In this MR analysis, the odds ratio (OR) is taken as the effect size, and the 95% confidence interval (CI) is taken. When p < 0.05, the difference is considered statistically significant.

Sensitivity and pleiotropy analysis

In terms of sensitivity, this study used IVW and MR-Egger regression to calculate Cochran’s Q statistic, respectively. If p > 0.05, it indicates no significant heterogeneity. At the same time, we also excluded and included SNPs one by one by leaving one method to observe whether it had an impact on the analysis results and drew a forest map. If a certain SNP was excluded and p > 0.05 was obtained, it is considered that SNPs will not have a significant impact on the results. In terms of multiple validity, we also used the intercept term of MR-Egger regression, MR Pleiotropy RESidual Sum and Outlier (MR-PRESSO) to test the level of multiple validity included in the SNP. In MR-Egger regression, if the intercept tends to 0, it can be considered that there is no horizontal pleiotropy. In MR-PRESSO test, not only the level of included instrumental variables can be calculated, but also the effect amount between exposure and outcome can be calculated after removing the outlier, and the results before and after correction can be tested.

Results

Mendelian randomization analysis

Before conducting MR analysis, we screened instrumental variables according to strict quality control procedures to ensure that the selected variables have sufficient effectiveness. The effectiveness of instrumental variables was evaluated through F-statistics, and the results showed that all instrumental variables had F-values >10, indicating a strong correlation between these SNPs and exposure. The results of using them as instrumental variables in MR analysis are reliable. MR analysis was conducted on an instrumental variable dataset containing 1400 metabolites, with the IVW method as the primary MR analysis method to explore the causal relationship between metabolites and VD. Five MR methods guided by instrumental variables were used to obtain estimated causal effects, OR values, p values, and corresponding confidence intervals for each metabolite on VD. These results provide important clues for understanding the causal relationship between metabolites and VD. In addition, pleiotropy and heterogeneity tests are used to test the reliability and robustness of MR analysis. Finally, based on pleiotropy testing (p > 0.05), IVW method (p < 0.05), and the consistency of five MR analysis methods in explaining causal effects (positive and negative OR values), we comprehensively evaluated the causal effects of 1400 metabolites on VD. The results showed that the plasma expression levels and ratios of one metabolite had a causal effect on VD, with 32 plasma metabolite phenotypes having OR values <.

Obtaining and differential analysis of metabolite-related gene sets

Through searching the GeneCards database, we successfully obtained a set of genes related to metabolite phenotypes. This process involved searching for metabolite phenotype names and ultimately summarizing the target genes in the GeneCards database, forming the metabolite phenotype-related gene set for our further research. We conducted differential analysis on the metabolite-related gene set obtained above using VD transcriptome data. This differential analysis provides important clues for subsequent functional research and biological interpretation, revealing the expression changes of metabolite-related genes in the development of VD and laying the foundation for a deeper understanding of the molecular mechanisms of VD. Strong evidence for a positive effect of metabolites, which means t⁶A on VD was found in IVW (OR: 0.667, 95% CI: 0.548–0.812, p < 0.001), MR-Egger (OR: 0.647, 95% CI: 0.458–0.913, p = 0.019), and weighted median (OR: 0.650, 95% CI: 0.466–0.908, p = 0.012) (Table 1 and Fig. 2).

FIG. 2.

(A) Circle diagram of MR study revealing causality from metabolites on VD. (B) R-plot of MR study revealing causality from metabolites on VD.

Table 1.

Analysis of the Correlation between t6A and VD by 2SMR

	Exposure	Outcome	Method	Nsnp	OR	Lower 95% CI	Upper 95% CI	p value
1	T6a levels	Vascular dementia	MR Egger	32	0.647	−0.780	−0.091	0.019
2	T6a levels	Vascular dementia	Weighted median	32	0.650	−0.765	−0.097	0.012
3	T6a levels	Vascular dementia	Inverse-variance weighted	32	0.667	−0.601	−0.208	0.000
4	T6a levels	Vascular dementia	Simple mode	32	0.709	−0.845	0.156	0.187
5	T6a levels	Vascular dementia	Weighted mode	32	0.647	−0.769	−0.102	0.016

MR, Mendelian randomization; OR, odds ratio; VD, vascular dementia.

Discussion

We conducted a comprehensive and in-depth exploration using MR methods by utilizing abundant plasma metabolite phenotype data (including 1400 metabolites) from the Protein Data Bank in Europe database as exposure factors, combined with VD data from the FinnGen Alliance database as outcomes. Our goal is to reveal the potential causal effects of plasma metabolites on the occurrence of VD and further screen for metabolite phenotypes with clear causal effects on VD. Based on the results of MR analysis, we summarized the gene sets related to these metabolite phenotypes and conducted subsequent differential analysis and clinical feature analysis. This series of work not only helps to understand the specific role of metabolite phenotype in the occurrence of VD but also helps to reveal the relationship between these metabolites and the micro background, and clinical and pathological characteristics of VD patients. The most important point is that the MR analysis results show long-term genetic effects, which means that the phenotypes of these metabolites may not undergo significant changes with changes in external conditions, providing a long-term and stable molecular target research basis for the metabolomics research of VD. Our research further delves into the impact of metabolites on VD patients. We found that the plasma expression levels and ratios of one metabolite (t6A) have a causal effect on VD, with 32 plasma metabolite phenotypes having OR values <1. This differential analysis result provides important clues for subsequent functional research and biological explanation, revealing the expression changes of metabolite-related genes in the development process of VD, and laying the foundation for a deeper understanding of the molecular mechanism of VD.

MR can be seen as an approximate randomized controlled trial.¹⁸ This method adopts the Mendelian random law principle and follows the principle of random allocation of alleles during gamete formation. Therefore, the genetic effects of gene phenotypes are not affected by classical confounding factors such as lifestyle and reverse causality, as in these cases, the phenotype level is influenced by the presence of disease.¹⁹ Therefore, genetic polymorphism is believed to alter gene expression and its impact on diseases, providing indirect evidence for the causal relationship between genetic polymorphism and disease.²⁰ MR is a useful method to infer the etiology of cardiovascular disease, diabetes, Parkinson’s disease, and other diseases.²¹ It may provide another perspective on the etiology of VD, as its onset time is often difficult to determine clinically.

Although significant progress has been made in the research of VD in the past few decades, there are still some shortcomings that affect our comprehensive understanding of this pathological and physiological process in depth and breadth. The heterogeneity of VD leads to the instability of research results.²² The heterogeneity of VD is manifested by cellular phenotype and genetic variation, which makes the pathogenic mechanism of VD more complex and diverse. In the early stages of the disease, the typical symptom of patients is memory impairment, usually dominated by recent memory impairment.²³ In the middle stage of the disease, patients may experience a certain degree of cognitive impairment and continue to worsen over time.²⁴ The main manifestation of this disease in the late stage is personality disorder, which ultimately leads to the inability to take care of oneself, and most patients die from complications.

T⁶A is a necessary posttranscriptional modification located at position 37 of tRNA responsible for decoding the ANN (N is A, U, C, or G) codon.²⁵ It plays an important role in regulating translation efficiency and intracellular protein homeostasis and is widely present in almost all cellular life systems.²⁶ The biosynthesis of t⁶A in different biological systems is achieved through a two-step continuous chemical reaction involving members of the TsaC/Sua5 protein family, TsaD/Kae1/Qri7 protein family, and other regulatory proteins.²⁷ In some organisms, t⁶A can be further transformed into ct⁶A and ht⁶A with multiple derivative modifications, playing a more refined regulatory role in cellular life.²⁸ However, the composition and molecular regulatory mechanisms of t⁶A and ct⁶A modifying enzymes in different biological systems exhibit distinct characteristics of conservation and diversity, and their molecular evolutionary and regulatory mechanisms are still unclear.²⁹ t⁶A modification can promote tRNA to enter ribosome accurately and participate in protein biosynthesis. t⁶A modification plays an important role in protein biosynthesis, and the deletion of t⁶A modification may be closely related to the occurrence of diseases.³⁰ More and more evidence shows that RNA modification is an important regulator of complex physiological and pathological processes such as aging, stress response, and epigenetics.³¹ However, whether small RNAs and their modifications change in dementia is still unclear. Some studies have shown that the expression of tRNA is reduced in AD, but our study confirmed that tRNA is closely related to VD, but the specific mechanism is not clear, and further research is needed.

Our research findings emphasize the important role of metabolites in the occurrence and development of VD and explore how metabolite phenotypes based on MR analysis affect the occurrence and development process of VD, ultimately affecting the prognosis of VD patients. This provides a new perspective for long-term and in-depth understanding of the molecular mechanisms of VD. In addition, our research also provides strong support for the development of metabolite-based VD treatment strategies and personalized medicine. Finally, we also focused on exploring the possible impact of genetic variation on VD metabolic pathways, which provides new methods and approaches for further VD metabolomics research and molecular target studies of antimetabolic therapy. It should be pointed out that although our study is based on MR analysis results and has made significant progress in revealing the relationship between long-term stable metabolite phenotypes and VD, more validation and in-depth research are needed before translating these relationships into practical clinical applications. In future work, we will further explore the biological mechanisms of genetic variations related to these metabolites in VD in order to better guide the diagnosis, treatment, and prognosis management of VD.

MR has been applied in scientific research for many years. With the update and progress of technology, its use has become more convenient, its scientificity has become more rigorous, and its system has become more mature. The update of databases has also brought about more discoveries, even cognitive updates. Perhaps with the development of this field, there will be more new discoveries or viewpoints in the future that break past conclusions. The selected association factors in this study were high-risk factors determined through patient follow-up and clinical experience and then compared with VD to determine correlation. There may be other deeper pathogenic factors that we are not aware of, so more research is needed in this field.

Limitations

Complexity and computational demand

MR analyses involve complex computational analyses and the integration of large datasets, which require significant computational resources and expertise. This complexity can be a barrier for some researchers and may limit the accessibility of this approach.

Data quality and availability

The accuracy of MR analyses predictions heavily depends on the quality and completeness of the underlying data. Incomplete or biased datasets can lead to incorrect predictions and interpretations, which may hinder the identification of true therapeutic targets.

Validation requirements

Predictions made by MR analyses approaches need to be experimentally validated, which can be time-consuming and resource-intensive. While these predictions provide valuable insights, they must be confirmed through laboratory experiments and clinical trials to ensure their reliability and applicability.

Limited dynamic information

MR analyses primarily use static data to create interaction networks. This approach does not always capture the dynamic nature of biological systems, such as temporal changes in gene expression or protein activity, which can be crucial for understanding disease progression and drug effects.

Conclusion

This study suggests that metabolites (t⁶A) may be causally associated with a positive effect on VD. Further studies are warranted to elucidate potential mechanisms.

Footnotes

Data Availability Statement

The original contributions presented in the study are included in the article, further inquiries can be directed to the corresponding author.

Authors’ Contributions

All the authors contributed equally to this project. S.C. and B.C. designed the study and data acquisition. S.C. and X.F. were responsible for the data analysis and interpretation. X.F. revised the article critically. S.C. and B.C. drafted the article. M.L. was responsible for statistical analysis. All the authors actively participated in drafting the article and approved the final version.

Author Disclosure Statement

All authors declare that there are no conflicts of interest in this work.

Funding Information

This work was supported by the Natural Science Foundation of Henan Province (222300420247), the Postdoctoral Foundation of Henan Province (312145), Central Government Guides Local Projects (Z20221343023), and the Medical Science and Technology Research Plan of Henan Province (LHGJ20221031).

References

, Wang

, Qu

, et al. Protective effect of hyperbaric oxygen therapy on cognitive function in patients with vascular dementia. Cell Transplant, 2019; 28(8):1071–1075; doi: 10.1177/0963689719853540

, He

, Huang

. The relationship between serum amyloid a level and cognitive dysfunction in patients with vascular dementia: Preliminary findings. Biomed Res Int, 2021; 2021:6676144; doi: 10.1155/2021/6676144

Tkacheva

, Mkhitaryan

, Kolykhalov

, et al. [Treatment of cognitive, behavioral and mental disorders in patients with vascular dementia: Results of a multicenter, randomized, double-blind, placebo-controlled clinical trial] (Article in Russian). Zh Nevrol Psikhiatr Im S S Korsakova, 2023; 123(7):41–49; doi: 10.1711/6/jnevro202312307141

Pérez-Sáez

, Justo-Henriques

, Alves Apóstolo

. Multicenter randomized controlled trial of the effects of individual reminiscence therapy on cognition, depression and quality of life: Analysis of a sample of older adults with Alzheimer’s disease and vascular dementia. Clin Neuropsychol, 2022; 36(7):1975–1996; doi: 10.1080/13854046.2021.1871962

Pauls

MMH

, Binnie

, Benjamin

, et al. The PASTIS trial: Testing tadalafil for possible use in vascular cognitive impairment. Alzheimers Dement, 2022; 18(12):2393–2402; doi: 10.1002/alz.12559

Lim

, Lee

, Kang

, et al. Efficacy and safety of oxiracetam in patients with vascular cognitive impairment: A multicenter, randomized, double-blinded, placebo-controlled, phase IV clinical trial. Contemp Clin Trials, 2023; 126:107108; doi: 10.1016/j.cct.2023.107108

, Wang

, Jiang

, et al. [Acupuncture and moxibustion for vascular dementia and its effect on serum VEGF and AChE.] (Article in Chinese). Zhongguo Zhen Jiu, 2021; 41(8):851–854; doi: 10.1370/3/j.0255-2930.20200816-0001

Leijenaar

, Groeneveld

, Klaassen

, et al. Methylphenidate and galantamine in patients with vascular cognitive impairment-the proof-of-principle Study STREAM-VCI. Alzheimers Res Ther, 2020; 12(1):10; doi: 10.1186/s13195-019-0567-z

Zhang

, Wang

, Zhao

, et al. Effects of blood pressure and antihypertensive drugs on osteoarthritis: A Mendelian Randomized Study. Aging Clin Exp Res, 2023; 35(11):2437–2444; doi: 10.1007/s40520-023-02530-8

10.

Wang

, Zhang

, Liu

, et al. Association of serum uric acid with metabolic syndrome and its components: A mendelian randomization analysis. Biomed Res Int, 2020; 2020:6238693; doi: 10.1155/2020/6238693

11.

Young

, Warren

, Mook-Kanamori

, et al. Genetically determined serum calcium levels and markers of ventricular repolarization: A Mendelian Randomization Study in the UK Biobank. Circ Genom Precis Med, 2021; 14(3):e003231; doi: 10.1161/CIRCGEN.120.003231

12.

Siedlinski

, Jozefczuk

, Xu

, et al. White blood cells and blood pressure: A Mendelian Randomization Study. Circulation, 2020; 141(16):1307–1317; doi: 10.1161/CIRCULATIONAHA.119.045102

13.

Sun

, Gao

, Wu

, et al. The causal relationship between gut microbiota and type 2 diabetes: A Two-Sample Mendelian Randomized Study. Front Public Health, 2023; 11:1255059; doi: 10.3389/fpubh.2023.1255059

14.

Yang

, Li

, Ding

, et al. Effect of smoking-related features and 731 immune cell phenotypes on esophageal cancer: A Two-Sample and Mediated Mendelian Randomized Study. Front Immunol, 2024; 15:1336817; doi: 10.3389/fimmu.2024.1336817

15.

, Xu

, et al. Effect of n-3 polyunsaturated fatty acids on ischemic heart disease and cardiometabolic risk factors: A Two-Sample Mendelian Randomization Study. BMC Cardiovasc Disord, 2021; 21(1):532; doi: 10.1186/s12872-021-02342-6

16.

Chen

, Lu

, Pettersson-Kymmer

, et al. Genomic atlas of the plasma metabolome prioritizes metabolites implicated in human diseases. Nat Genet, 2023; 55(1):44–53; doi: 10.1038/s41588-022-01270-1

17.

Larsson

, Woolf

, Gill

. Plasma caffeine levels and risk of Alzheimer’s Disease and Parkinson’s disease: Mendelian Randomization Study. Nutrients, 2022; 14(9):1697; doi: 10.3390/nu14091697

18.

Zhu

, Ding

. Association between gut microbiota and seven gastrointestinal diseases: A Mendelian Randomized Study. J Gene Med, 2024; 26(1):e3623; doi: 10.1002/jgm.3623

19.

Zhou

, Zhang

, Fang

, et al. Relationship between Parkinson’s disease and cardio-cerebrovascular diseases: A Mendelian Randomized Study. Sci Rep, 2023; 13(1):20428; doi: 10.1038/s41598-023-47708-2

20.

Zhao

, Guo

, Tang

, et al. Evidence for a causal relationship between psoriasis and cutaneous melanoma: A Bidirectional Two-Sample Mendelian Randomized Study. Front Immunol, 2023; 14:1201167; doi: 10.3389/fimmu.2023.1201167

21.

Zhang

, Yang

, Li

. Causality between sarcopenia-related traits and major depressive disorder: A bi-directional, Two-Sample Mendelian Randomized Study. Medicine (Baltimore), 2023; 102(40):e35071; doi: 10.1097/MD.0000000000035071

22.

Huang

, Yin

, Li

, et al. Effects of acupuncture on vascular cognitive impairment with No dementia: A randomized controlled trial. J Alzheimers Dis, 2021; 81(4):1391–1401; doi: 10.3233/JAD-201353

23.

Hoevenaar-Blom

, Richard

, Moll van Charante

, et al. Targeting vascular risk factors to reduce dementia incidence in old age: Extended Follow-up of the Prevention of Dementia by Intensive Vascular Care (preDIVA) randomized clinical trial. JAMA Neurol, 2021; 78(12):1527–1528; doi: 10.1001/jamaneurol.2021.3542

24.

Han

, Li

, Jiang

, et al. Effect of early acupuncture on cognitive function in patients with vascular dementia after cerebral infarction] (Article in Chinese). Zhongguo Zhen Jiu, 2021; 41(9):979–983; doi: 10.1370/3/j.0255-2930.20210117-k0008

25.

Zheng

, Su

, Duan

, et al. Molecular basis of A. thaliana KEOPS complex in biosynthesizing tRNA t6A. Nucleic Acids Res, 2024; 52(8):4523–4540; doi: 10.1093/nar/gkae179

26.

Wang

, Zhou

, Mao

, et al. Commonality and diversity in tRNA substrate recognition in t6A biogenesis by eukaryotic KEOPSs. Nucleic Acids Res, 2022; 50(4):2223–2239; doi: 10.1093/nar/gkac056

27.

Jin

, Zhang

, Yu

, et al. Structure-function analysis of an ancient TsaD-TsaC-SUA5-TcdA modular enzyme reveals a prototype of tRNA t6A and ct6A synthetases. Nucleic Acids Res, 2023; 51(16):8711–8729; doi: 10.1093/nar/gkad587

28.

Zhou

, Wang

, Zeng

, et al. Molecular basis for t6A modification in human mitochondria. Nucleic Acids Res, 2020; 48(6):3181–3194; doi: 10.1093/nar/gkaa093

29.

Missoury

, Plancqueel

, Li de la Sierra-Gallay

, et al. The structure of the TsaB/TsaD/TsaE complex reveals an unexpected mechanism for the bacterial t6A tRNA-modification. Nucleic Acids Res, 2019; 47(17):9464–9465; doi: 10.1093/nar/gkz719

30.

Luthra

, Paranagama

, Swinehart

, et al. Conformational communication mediates the reset step in t6A biosynthesis. Nucleic Acids Res, 2019; 47(12):6551–6567; doi: 10.1093/nar/gkz439

31.

Missoury

, Plancqueel

, Li de la Sierra-Gallay

, et al. The structure of the TsaB/TsaD/TsaE complex reveals an unexpected mechanism for the bacterial t6A tRNA-modification. Nucleic Acids Res, 2018; 46(11):5850–5860; doi: 10.1093/nar/gky323