Identifying serum metabolite biomarkers for autoimmune diseases: a two-sample mendelian randomization and meta-analysis

Abstract

Background: Extensive evidence suggests a link between alterations in serum metabolite composition and various autoimmune diseases (ADs). Nevertheless, the causal relationship underlying these correlations and their potential utility as dependable biomarkers for early AD detection remain uncertain.

Objective: The objective of this study was to employ a two-sample Mendelian randomization (MR) approach to ascertain the causal relationship between serum metabolites and ADs. Additionally, a meta-analysis incorporating data from diverse samples was conducted to enhance the validation of this causal effect.

Materials and methods: A two-sample MR analysis was performed to investigate the association between 486 human serum metabolites and six prevalent autoimmune diseases: systemic lupus erythematosus (SLE), rheumatoid arthritis (RA), inflammatory bowel disease (IBD), dermatomyositis (DM), type 1 diabetes (T1D), and celiac disease (CeD). The inverse variance weighted (IVW) model was employed as the primary analytical technique for the two-sample MR analysis, aiming to identify blood metabolites linked with autoimmune diseases. Independent outcome samples were utilized for further validation of significant blood metabolites. Additional sensitivity analyses, including heterogeneity test, horizontal pleiotropy test, and retention rate analysis, were conducted. The results from these analyses were subsequently meta-integrated. Finally, metabolic pathway analysis was performed using the KEGG and Small Molecule Pathway Databases (SMPD).

Results: Following the discovery and replication phases, eight metabolites were identified as causally associated with various autoimmune diseases, encompassing five lipid metabolism types: 1-oleoylglycerophosphoethanolamine, 1-arachidonoylglycerophosphoethanolamine, 1-myristoylglycerophosphocholine, arachidonate (20:4 n6), and glycerol. The meta-analysis indicated that three out of these eight metabolites exhibited a protective effect, while the remaining five were designated as pathogenic factors. The robustness of these associations was further confirmed through sensitivity analysis. Moreover, an investigation into metabolic pathways revealed a significant correlation between galactose metabolism and autoimmune diseases.

Conclusion: This study revealed a causal relationship between lipid metabolites and ADs, providing novel insights into the mechanism of AD development mediated by serum metabolites and possible biomarkers for early diagnosis.

Keywords: Mendelian randomization; autoimmune diseases; biomarkers; meta-analysis; serum metabolites.

Figure 1

The overview of the research workflow.

Figure 2

Heatmap of mendelian randomized association effect values for known metabolites with significant differences in phenotypic risk for six autoimmune diseases (from fixed effects IVW analysis). (^*: P<0.05; ^**: P<0.01; ^***: P<0.001).

Figure 3

Scatter plot showing the genetic associations of seven metabolites on the risk of 6 AD phenotypes. (A) 1-oleoylglycerophosphoethanolamine on CeD, (B) betaine on DM, (C) 1-arachidonoylglycerophosphoethanolamine on IBD, (D) arachidonate (20:4n6) on IBD, (E) 1-myristoylglycerophosphocholine on RA, (F) glycerol on RA, (G) 2-methoxyacetaminophen sulfate on SLE, (H) glycerol 2-phosphate on T1D.

Figure 4

The funnel plot represents IVs for each significant causal association between metabolites and OA phenotypes. (A) 1-oleoylglycerophosphoethanolamine on CeD, (B) betaine on DM, (C) 1-arachidonoylglycerophosphoethanolamine on IBD, (D) arachidonate (20:4n6) on IBD, (E) 1-myristoylglycerophosphocholine on RA, (F) glycerol on RA, (G) 2-methoxyacetaminophen sulfate on SLE, (H) glycerol 2-phosphate on T1D.

Figure 5

Meta-analysis of significantly associated (IVW derived P< 0.05) between metabolites and CeD (A), DM (B), IBD (C–D), RA (E–F), SLE (G), and T1D (H). 95% CI, 95% confidence interval; OR, odds ratio.