Mediating Role of Blood Metabolites in the Relationship Between Immune Cell Traits and Heart Failure: A Mendelian Randomization and Mediation Analysis

Abstract

Background: Observational studies have shown a significant association between immune cells and heart failure (HF). Nevertheless, the precise biological mechanisms underlying this association remain unclear.

Methods: To investigate the causative relationships and underlying mechanisms between immune cell traits and adult HF, 3 main methods of Mendelian randomization were used: 2-sample Mendelian randomization, multivariable Mendelian randomization with controlling for several factors affecting HF, and mediation analysis. Results from the inverse variance-weighted model indicated that genetic predispositions for human leukocyte antigen-type DR (HLA DR) on CD33dim HLA DR+ CD11b+ (odds ratio, 0.967 [95% CI, 0.939-0.996]; P=0.028) may be associated with a reduced risk of HF. Although the association between HF and HLA DR on CD33 dim HLA DR+ CD11b+ did not withstand multiple-testing correction, the Mendelian randomization results (P_IVW <0.05) decrease the likelihood that the observational results are due to chance.

Results: Our 2-step mediation analysis demonstrated that genetic predispositions for HLA DR on CD33dim HLA DR+ CD11b+ (odds ratio,1.085 [95% CI, 1.020-1.155]; P=0.010) was associated with increased levels of the metabolite Octadecanedioate, while genetic predispositions for Octadecanedioate levels (odds ratio, 0.917 [95% CI, 0.849-0.991]; P=0.028) was associated with a reduced risk of HF. Moreover, our results also demonstrated that the association between HLA DR on CD33dim HLA DR+ CD11b+ and HF was possibly mediated by Octadecanedioate levels, with a mediation proportion of 21.4% [95% CI, 43.7 -0.998].

Conclusions: These findings underscore the importance of HLA DR on CD33dim HLA DR+ CD11b+ in the development of HF, with Octadecanedioate levels acting as a possible mediator in this pathway.

Keywords: HF; Mendelian randomization; blood metabolites; immune cell; mediation analysis.

Figure 1. Diagrams illustrating the associations investigated in this study.

A, Total effect of immune cells on blood metabolites. (B) Total effect of blood metabolites on HF. (C, D) Total effect between immune cells and HF. (C′) the direct effect (C′=C–A×B). GWAS indicates genome‐wide association study; HF, heart failure; IVW, inverse variance‐weighted; MR, Mendelian randomization; MVMR multivariable Mendelian randomization; and SNP, single‐nucleotide polymorphism.

Figure 2. Results of Mendelian randomization analysis (immune cells on HF).

HF indicates heart failure; HLA DR, human leukocyte antigen‐type DR; MDSC AC, Myeloid‐Derived Suppressor Cells AC; nsnp, number of single nucleotide polymorphism; OR, odds ratio; pval, P value; and Treg, Regulatory T cells.

Figure 3. Results of mendelian randomization analysis (blood metabolites on HF).

HF indicates heart failure; nsnp, number of single nucleotide polymorphism; OR, odds ratio; and pval, P value.

Figure 4. Forest plot to visualize the causal effects of Octadecanedioate levels with HLA DR on CD33dim HLA DR+ CD11b+ and HF.

HF indicates heart failure; HLA DR, human leukocyte antigen‐type DR; nsnp, number of single nucleotide polymorphism; OR, odds ratio; and pval, P value.

Figure 5. Mediation analysis.

A, Displays scatterplot results depicting the relationship between HLA DR on CD33dim HLA DR+ CD11b+ and Octadecanedioate levels in 5 different MR analyses. (B) Funnel plots (HLA DR on CD33dim HLA DR+ CD11b+ on Octadecanedioate levels). (C) MR leave‐one‐out sensitivity analysis. Leave‐one‐out analysis is used to evaluate whether any SNP is driving the causal effect (HLA DR on CD33dim HLA DR+ CD11b+ on Octadecanedioate levels). (D) Displays scatterplot results depicting the relationship between Octadecanedioate levels and HF in 5 different MR analyses. (E) Funnel plots of Octadecanedioate levels on HF. (F) MR leave‐one‐out sensitivity analysis. Leave‐one‐out analysis is used to evaluate whether any SNP is driving the causal effect (Octadecanedioate levels on HF). (G) Displays scatterplot results depicting the relationship between HLA DR on CD33dim HLA DR+ CD11b+ and HF in 5 different MR analyses. (H) Funnel plots (HLA DR on CD33dim HLA DR+ CD11b+ and HF), (I) MR leave‐one‐out sensitivity analysis. Leave‐one‐out analysis is used to evaluate whether any SNP is driving the causal effect (HLA DR on CD33dim HLA DR+ CD11b+ and HF). HLA DR indicates human leukocyte antigen‐type DR; MR, Mendelian randomization; and SNP, single‐nucleotide polymorphism.