Assessing the causal association between human blood metabolites and the risk of gout

The occurrence of gout is closely related to metabolism, but there is still a lack of evidence on the causal role of metabolites in promoting or preventing gout.


| INTRODUCTION
Gout is one of the common and complex inflammatory diseases caused by monosodium urate crystals deposited in the joints after prolonged hyperuricaemia. 1 The prevalence of gout is approximately less than 1%-6.8%,with an incidence of 0.58-2.89per 1000 person-years. 2At the same time, the incidence of gout is increasing dramatically worldwide. 3According to epidemiological studies, gout is commonly comorbid with hypertension, diabetes, and nephropathy, resulting in increased mortality and heavy economic burden. 4As such, early recognition and prevention of gout are imperative.
6][7] For instance, a study using UPLC-Q/TOFMS technology identified 63 serum differential metabolites in patients with gout, including leucine, 2-Hydroxypurine, taurodeoxycholic acid, etc. 8 Blood uric acid is the primary serological marker for the diagnosis of gout, and it is similarly elevated in patients with hyperuricaemia.A study found that L-isoleucine, L-lysine and L-alanine were potential markers for differentiating gout from hyperuricaemia when patients had similar blood uric acid levels. 9These studies suggested that certain metabolites are associated with the progression of gout.However, a majority of studies tended to provide only associations of metabolites with disease occurrence, but the causality remained unclear.Therefore, there is an urgent need for a comprehensive and thorough analysis to uncover the causality of metabolites in the gout mechanism.
As a popular method of genetic epidemiology study design, Mendelian randomization (MR) explores the causal inference between exposure and outcome. 10As known, randomized controlled trials (RCTs) are the gold standard of clinical evidence.When RCTs are not feasible, the MR method is a key alternative strategy supplying credible evidence of a causal relationship between exposure and disease risk. 11MR uses single nucleotide polymorphisms (SNPs) as instrumental variables to infer causality.Considering that genetic variants were randomly assigned during gametogenesis, and genotypes were typically unaffected by the external environment, a rigorously designed MR could largely avoid confounding and reverse causality.
Given that the causal impacts of blood metabolites on gout were poorly understood, we used a two-sample MR analysis based on genome-wide association study (GWAS) statistics to explore the causal relationship between a wide range of these blood metabolites and the risk of gout.The results of this study would not only contribute to the understanding of the pathophysiological mechanisms of gout, but also could provide a reliable basis for the development of feasible clinical strategies for the early diagnosis, prevention and treatment of gout.

| Study design
We systematically assessed the causal relationship between 486 human blood metabolites and gout risk using a two-sample MR design.A convincing MR design should strictly be in accordance with three basic assumptions: (1) the SNPs should be significantly associated with the exposures (blood metabolites in this study); (2) the SNPs should not be related to confounders; (3) the SNPs should exert effects on the outcome (gout in this work) through the exposures.Statistical analyses were performed by the 'TwoSampleMR' MR package (version 0.5.6) in the R program (version 4.2.2).

| GWAS data for human blood metabolites
We obtain the genetic information of blood metabolites from a GWAS conducted by Shin et al. 12 Specifically, a total of 7824 European descents were included in this GWAS and approximately 2.1 million SNPs were tested for 486 blood metabolites.According to the Kyoto Encyclopaedia of Genes and Genomes (KEGG) database, 13 there were 309 metabolites chemically identified and assigned to seven broad metabolic groups, including amino acid, carbohydrate, cofactors and vitamin, lipid, nucleotide, peptide and xenobiotic metabolism.Another 177 metabolites were classified as 'unknown' as their chemical nature had not been conclusively determined.Genetic variants for each blood metabolite were obtained from the Metabolomics GWAS server (http:// metab olomi cs.helmh oltz.muenchen.de/gwas/).

| GWAS data for gout
The GWAS data for gout were obtained from the FinnGen Database 14 (Round 8) (FinnGen http:// www.finng en.fi/ en), in which a total of 429,209 Finns were involved, and the gout phenotype included 7461 cases and 221,323 controls.According to the FinnGen consortium, the patients with gout were diagnosed based on the International Classification of Diseases (ICD)-8, ICD-9 and ICD-10.All the statistics we used are freely available in public databases and therefore do not require ethical approval.

| Instruments selection
A series of steps for selecting eligible genetic variants associated with metabolites were performed.First, considering the limited number of SNPs reaching genome-wide significance, we relaxed the association threshold using P < 5 × 10 −5 [pairwise linkage disequilibrium (LD) r 2 < 0.1 within a 500 kb distance] to obtain top independent SNPs, which was similar to the study of Yang et al. 15 This method was widely used in previous MR studies. 16,17Meanwhile, to avoid bias owing to the employment of weak instruments, F statistics were calculated for each SNP to measure the statistical strength as previously described. 16SNPs with F < 10 were recognized as weak instruments and were then discarded to ensure all the SNPs conferred sufficient variance for corresponding metabolites. 18We then extracted the exposure SNPs from the outcome data and excluded those associated with the outcome (P < 5× 10 −5 ).For SNPs absent in the outcome, proxied SNPs were identified in high LD (r 2 > 0.8) based on the European reference panel of the 1000 Genomes Project.For those absent and no appropriate proxies identified, we discarded them.Harmonization was then conducted to align the alleles of the exposure-and outcome-SNPs, and discard palindromic SNPs with intermediate effect allele frequencies (EAF >0.42) or SNPs with incompatible alleles (e.g.A/G vs. A/C).Finally, metabolites with more than two SNPs were kept for MR analysis. 19

| MR analysis
A preliminary analysis was performed using the random effects inverse variance weighting (IVW) method, with p < 0.05 representing a significant causal relationship between metabolites and gout.IVW is the primary method in MR studies that combines all Wald ratios for each SNP to produce summary estimates.The IVW method has significantly higher statistical power than other MR methods but is also prone to pleiotropy bias even when there was only one invalid IV included. 20Therefore, we used the IVW method as the primary analysis to explore metabolites having a preliminary association with gout.Two other MR models, including the weighted median (WM) and MR-Egger method, were used.The WM method assumes that at least half of the instruments were valid, 21 while MR-Egger regression provides consistent estimates accounting for pleiotropy when all the instruments are invalid. 22To determine the significance, we considered that blood metabolites associated with gout at P < 1.03 × 10 −4 (Bonferroni correction: 0.05/486 metabolites) were potential candidate metabolites involved in the development of gout.Besides, those associated with gout at 1.03 × 10 −4 < p < 0.05 using the IVW method were considered as suggestive significant metabolites.In addition, we also considered that at least two of the three MR methods yielded associations of P < 0.05 to establish convincing MR causal relationships.

| Sensitivity analysis
Despite the advantages of the MR design, the causal associations could still be biased when existing of certain invalid IVs, and thus, we used a series of sensitivity analyses to evaluate the robustness of the MR results. 22We tested heterogeneity using Cochran's Q test and we considered the existence of heterogeneity with a Q-derived p < 0.05. 23esides, we tested horizontal pleiotropy utilizing Egger's intercept and a detected p < 0.05 for MR-Egger intercept indicated the existence of pleiotropy. 24To determine whether the yielded causal estimates were driven by any individual SNPs, we performed a leave-one-out (LOO) analysis by discarding each exposure-related SNP in turn to replicate the IVW analysis.Based on the complementary MR models and sensitivity analyses, we considered that potentially eligible candidate metabolites involved in the development of gout must meet: (1) consistent direction and magnitude across the three MR methods: (2) no heterogeneity or pleiotropy detected; (3) no high impact sites identified in the LOO analysis.

| Confounding analysis
Except for a range of statistical methods performed in sensitivity analyses to assess any violations of MR assumptions, we also used the Phenoscanner website (http:// www.pheno scann er.medsc hl.cam.ac.uk/ ) to see if certain SNPs were associated with other phenotypes serving as potential risk factors of gout, including body mass index (BMI), 25 smoking, 26 and diabetes. 27Once SNPs were identified to be associated with potential confounders mentioned above at a threshold of p < 5 × 10 −5 , IVW was repeated after removing these SNPs manually to verify the robustness of the results.

| Metabolic pathway analysis
Metabolites screened by MR methods were included for metabolic pathway analysis based on the KEGG database using MetaboAnalyst 5.0 (https:// www.metab oanal yst.ca/ ), a free online website for simplified metabolomics data analysis.

| RESULTS
With rigorous SNPs filtration, 486 metabolites were retained for MR analysis (four metabolites with fewer than three SNPs were excluded) (Additional file Table S1).The number of SNPs for each metabolite ranged from 3 to 507.F-statistics were all greater than 10, indicating that no weak instruments were used and the data after coordinated alignment are shown in (Additional file Table S2).

| MR analysis
The IVW method initially identified 33 metabolites significantly associated with gout (p < 0.05; Figures 1, 2).Among them, the chemical nature of 16 metabolites was unknown, and the other 17 metabolites were chemically identified and assigned to amino acid, carbohydrate, lipid, nucleotide, peptide, xenobiotics, cofactors and vitamins metabolism.Further sensitivity analysis identified the remaining 14 metabolites meeting the criteria for eligible candidate metabolites involved in the development of gout (Table 1).

| Metabolic pathway analysis
We identified a potential metabolic pathway associated with the pathogenesis of gout (Table 2), with 1-Methylxanthine involved in Caffeine metabolism (p = 0.02).

| DISCUSSION
Our results confirmed causal relationships between 14 metabolites and gout.The increase of nine metabolites: piperine, ibuprofen, 2-aminobutyrate, inosine, gammaglutamylleucine, bradykinin des-arg (9), 3-methylxanthine, isovalerate and 1-linoleoylglycerol (1-monolinolein) had adverse effects on gout, while the increase of the other five metabolites, including pro-hydroxy-pro, benzoate, undecanoate (11:0), 1-methylxanthine and hexadecanedioate, played a protective role in the development of gout.Among them, the blood metabolite hexadecanedioate represented the most significant association with gout.We also found that 1-methylxanthine, which is involved in metabolic pathways, may play a critical role in the development of gout.As far as we know, this is the F I G U R E 1 Forest plot for the causal effect of metabolites on the risk of Gout derived from inverse variance weighted (IVW).CI, confidence interval; OR, odds ratio.
first study using an MR design to assess the causal relationship between human blood metabolites and the risk of gout.
The high prevalence and teratogenicity of gout have brought a heavy burden to patients and their families, which makes screening and prevention of the disease much more important.Most of the causes of gout have not been clarified, and only a few patients are caused by enzyme defects (such as Hypoxanthine-guanine phosphate nucleotide converting enzyme deletion). 28A study examined serum samples from patients with gout and healthy controls and found that 16 metabolites, including lipids F I G U R E 2 scatter plot for the significant Mendelian randomization (MR) association (P < 0.05) between metabolites and Gout.SNP, single nucleotide polymorphism.

T A B L E 1 Sensitivity analysis for the causal association between blood metabolites and Gout.
Metabolites and amino acids, changed significantly in patients with gout, which may be biomarkers of gout. 29Previous studies have identified various changes in blood metabolites such as disturbances in lipid metabolism, carbohydrate metabolism, amino acid metabolism and energy metabolism in gout patients. 29Despite a growing body of research on the relationship between metabolic disturbance and gout, there is still insufficient evidence to establish a causal relationship between them.To further investigate the causal relationship between metabolites and gout, this work utilized an MR framework to systematically and comprehensively assess the causal relationship between human blood metabolites and gout.

| Adverse effects of metabolites on gout
Our study identified several metabolites that had adverse effects on gout, among which the potential mechanisms of some metabolites could be traced from previous studies.Inosine is converted to hypoxanthine under the action of purine nucleoside phosphorylase, which is the intermediate product of purine metabolism, and the final product is uric acid, which leads to the occurrence of gout. 30It has been studied that a stable hyperuricemia model can be formed after intraperitoneal injection of inosine in mice, and the serum level of uric acid was four times higher than that of the blank group. 31This further proves that inosine is involved in and leads to gout, which is consistent with our results.Previously study reported that 3-methylxanthine produced xanthine and uric acid during oxidation, suggesting that 3-methylxanthine may also be involved in the pathogenesis of gout through purine metabolism. 324][35] Some experiments have proved that piperine can inhibit the inflammation induced by the crystallization of monosodium urate. 36A previous study also found that ibuprofen can inhibit the activity of xanthine oxidase and reduce the occurrence of inflammation. 37This is contrary to the result that the higher level of piperine and ibuprofen predicted a higher risk of gout, which warrants further exploration in the future.Few studies have been found on the relationship between the remaining metabolites and gout.Therefore, this MR analysis has a reference significance for their causal relationship with gout.

| Positive effects of metabolites on gout
Our study also identified several metabolites exerting a positive effect on gout.After Bonferroni correction, we found that hexadecanedioate was closely related to the pathogenesis of gout.Hexadecanedioate acid is a longchain dicarboxylic acid produced by ω-oxidation of fatty acids and then metabolized by β-oxidation in the peroxisome. 38Hexadecanedioate may reduce uric acid production by inhibiting the β-oxidation pathway of fatty acids, which typically occurs in mitochondria.It has been suggested that hexadecanedioate can affect the potential and permeability of mitochondrial membranes, 39 which might reduce the uric acid synthesis and consequently lead to a slowdown in the metabolism of fatty acids and a decrease in metabolites such as pyruvate and acetyl coenzyme A. Meanwhile, a previous study showed that hexadecanedioate has the properties of reducing blood lipids, anti-obesity, and anti-diabetes. 40Given that diabetes and obesity might induce gout, there might be the possibility that hexadecanedioate prevents gout through anti-diabetes and anti-obesity, which deserves further exploration under specific experimental conditions.Thus, hexadecanedioate may be a potential diagnostic indicator and therapeutic target for gout.Benzoate, also known as benzoic acid or benzenecarboxylate was also identified in our study as a protective metabolite for gout.A recent study found that benzoic acid has anti-inflammatory activity, 41 which may be the reason why it has a positive effect on gout.

| Strengths and limitations
Our study has certain innovations.First, 486 blood metabolites were included in this study for MR analysis, which is the most comprehensive and systematic study to date to explore the causal relationship of circulating metabolites with gout.Second, under the MR framework, this study uses various models and thorough sensitivity analyses to assess the causality and evaluate the reliability of the MR results, which largely avoids reverse causality and residual confounding that are commonly inevitable in traditional observational studies.However, several limitations should be noted in our study.First, all the GWAS data included in this work were derived from European populations, and this is to some extent ethnically restrictive, and future studies are needed to further evaluate the generality of our results in different populations.Secondly, some of the identified metabolites involved in gout development were chemically unknown, making conclusive interpretation unavailable.

| Conclusion
In conclusion, we have determined the causal relationship between 14 metabolites and gout by MR analysis, which provided preliminary evidence for the role of these 14 metabolites in the progression of gout.Of these, hexadecanedioate is most closely associated with gout.These metabolites may be helpful in clinical screening and prevention of high-risk groups of gout, as well as candidate molecules for future mechanism exploration.
Significant metabolic pathways involved in the pathogenesis of Gout.