AH-INT-197

Title

Environmental exposure testing in the AIRWAVE study

Summary of Proposal

The aims of this project are to 1) to assess a broad range of environmental contaminants in human blood samples collected from the AIRWAVE cohort 2) describe the socio-demographic, geographical and dietary determinants of contaminant levels 3) link contaminants to health, including inflammatory profiles and cognition.

Analysis will be conducted on serum/plasma samples that were previously selected for endogenous metabolomic profiling at the National Phenome Centre (NPC). Parent plates of samples (up to 6,000 samples) that were used for previous aliquoting of NPC LC-MS assays (one freeze thaw cycle) will be transferred to laboratory of Leon Barron at the MRC-Centre for Environment and Health. Dr Barron has developed non-targeted screening assays based on liquid chromatography- and gas chromatography mass spectrometry for simultaneous detection of >200 chemicals, including pesticides, PFAS, pharmaceuticals, plastic-related and industrial chemicals in environmental samples. These assays will be adapted for analysis of blood samples and will aim to analyse at least 1,000 samples for sufficiently powered health analysis.

The following analyses are planned:

• Summary statistics of exposure levels by sex, age, education level, job role, police force region, body mass, diet score<
• Principal component /clustering analysis to define “exposome” profiles for analysis with sex, age, education level, job role, police force region, body mass, diet score
• Analysis of associations of contaminants with cognitive scores from computerised testing. Priority will be given for analysis of potentially neurotoxic compounds such as nicotinamide pesticides.
• Analysis of associations of contaminants with Olink inflammatory protein profiles. This work will build on toxicological work in the Horizon Europe Endomix project, investigating immunotoxic properties of endocrine disrupting chemicals. Machine learning methods such as O2-PLS will be applied to identify shared variance between exposome and inflammatory datasets.