A comprehensive phenome wide analysis of the role of neutrophils in health and disease

Abstract

Neutrophil release of cytoplasmic granules containing antimicrobial agents is a critical component of innate immunity. Neutrophils are widely implicated in tissue inflammation however the extent of the neutrophil contribution to human health and disease is incompletely characterized. To explore this further, we leveraged publicly available genetic data to conduct a Mendelian randomization phenome-wide association study (MR-PheWAS) of neutrophil traits and 14,983 outcomes. Genetic proxies for neutrophil count, granularity, and serum myeloperoxidase were linked to 145 outcomes. Higher neutrophil count was associated with lower body weight, reduced obesity risk, and increased vascular activation markers but not with atherosclerosis. Elevated neutrophil count was robustly linked to Alzheimer's disease and neutrophil granularity with gut microbiota abundance and dental pathology. Our findings reveal the diverse roles of neutrophils extending beyond pathogen defense and underscore the potential for MR-PheWAS in identifying novel neutrophil-related pathophysiology.

Keywords: Mendelian randomization; genome-wide association study; neutrophil; phenome-wide association study.

Fig. 1.

MR-PheWAS methodology. (A) Directed acyclic graph depicting the principle of MR and underlying IV assumptions. Solid arrows indicate causal links and dashed arrows indicate potential violations of MR assumptions. Assumption 1 is that the IV is associated with the exposure of interest (relevance), which can be assessed using the F statistic. Assumption 2 is that the IV is associated with the outcome only through the exposure (independence). With control for population structure in the source GWAS, Mendel's law implies that IVs will not be directly associated with confounders. Assumption 3 is that the IV influences the outcome only through the exposure and not through alternative pathways (exclusion restriction). Primary MR analyses typically assume no pleiotropy, however, sensitivity analyses such as MR Egger calculate the effect estimate adjusted for any directional pleiotropy. (B) Details of GWAS from which exposure instruments were constructed. (C) MR-PheWAS workflow. Three separate PheWASs were undertaken to examine the causal effect of neutrophil traits (exposures) on a range of outcomes. IVs for the exposures were selected, data for these variants were extracted from outcome GWAS summary statistics and then harmonized to ensure that variant effects on the exposure and outcome corresponded to the same allele. Two-sample MR was undertaken, and results interrogated using a range of sensitivity analyses. All analyses were undertaken using the TwoSampleMR R package. eQTL, expression quantitative trait loci; FDR, false discovery rate.

Fig. 2.

Heritability and genetic correlations for neutrophil traits. (A) Heritability for each neutrophil trait calculated using linkage disequilibrium score regression. (B) Genetic correlation between neutrophil traits. Genetic correlations (rg) calculated using bivariate linkage disequilibrium score regression.

Fig. 3.

Phenome-wide associations with neutrophil characteristics. Manhattan plots for NEUT, SSC, and MPO. Hematological parameters excluded to aid visualization. The dashed gray line indicates false discovery rate–adjusted P value of 0.05.

Fig. 4.

MR of NEUT associations with body mass. Forest plots with MR effect estimates for body mass index, weight, glucose, and obesity. IVW (primary analysis) is shown in in purple, MR Egger is shown in blue, the weighted median is shown in green, and the weighted mode is shown in red. Error bars signify 95% CIs.

Fig. 5.

MR of neutrophil trait associations with the gut microbiome. Forest plots with MR effect estimates for abundance of gut microbiota and serum acetate. IVW (primary analysis) is shown in in purple, MR Egger is shown in blue, the weighted median is shown in green, and the weighted mode is shown in red. Error bars signify 95% confidence intervals.

Fig. 6.

MR of neutrophil trait associations with AD. Forest plots with MR effect estimates for AD, brain cortical thickness, and serum lactate. IVW (primary analysis) is shown in in purple, MR Egger is shown in blue, the weighted median is shown in green, and the weighted mode is shown in red. Error bars signify 95% CIs.