The Exposome and Health

The exposome and health: Where chemistry meets biology. Vermeulen R, Schymanski EL, Barabási AL, Miller GW. Science. 2020 Jan 24;367(6476):392-396. doi: 10.1126/science.aay3164. PMID: 31974245

What is the Exposome?

• all environmental, i.e., nongenetic, drivers of health and disease
• exposures to chemicals, dietary constituents, and physical and psychosocial factors, as well as the biological responses to these exposures (including effects of genetic modifiers and resident microbiota, e.g., different people may metabolize drugs differently or differ in their susceptibility to PTSD)

The Exposome Represents Potentially Modifiable Health Risk

• genetics account for less than half of the risk of heart disease
• modifiable metabolic, environmental, occupational, and behavioral risks contribute to ~60% of deaths worldwide, according to the Global Burden of Disease (GBD) project
• 9 million deaths per year (16% of deaths worldwide) have been attributed to environmental pollution alone, and that is likely an underestimate

This review focuses on characterizing the chemical component of the exposome and how it relates to human health.

Measuring Chemicals en Masse

High-resolution mass spectrometry (HRMS) can measure thousands of chemicals per sample. Compounds are identified by comparison to a list (exact mass of compound, and if available, experimental fragment data or “fingerprint”):

• limited availability of experimental data on metabolites has hindered identifying all compounds in a sample
• computational metabolite prediction gives many FP and FN

Areas of Current Focus and Improvement

• automated text mining of literature and patents for chemical entities/data
• grouping correlated exposures; representing correlations as edges in a network of compounds
• EWAS (analogous to GWAS but environment instead of genome); at least partly exploratory rather than hypothesis-driven:
  • – analyzes a panel of environmental factors vs. a phenotype of interest; one study found pesticide exposure significantly associated with type 2 diabetes
  • – a goal is to enable calculating an exposome risk score (ERS) analogous to a polygenic risk score (PRS) for a disease
  • – having to predefine the panel is a limit on pure exploration
  • – will need much larger sample sets, consortium efforts
  • – does not address nonadditive interactions among exposures
• statistical methods to detect nonadditive interactions
• defining a baseline; what is a healthy or “normal” exposome?

Informative Exposome Study Designs

• biological samples should be linked to exposures (E), biological effects (B), and preferably, the health phenotype of interest (P)
• although challenging due to temporal lag and other offsets that could obscure causality, the meet-in-the middle (MITM) approach analyzes the entire E-B-P chain rather than E-B and B-P steps separately
• MITM studies using HRMS data have identified individual and combinatorial effects of chemicals:
  • HELIX: early-life exposure to environmental chemicals → childhood lung function
  • EXPOsOMICs: air pollution → linoleic acid metabolism → adult-onset asthma and cardiovascular disease

Important Next Steps for Exposome Research

• improve chemical and biological screening technology for higher throughput and lower costs
• continue to develop chemical and spectral data resources, e.g. NORMAN and Global Natural Product Social Molecular Networking (GNPS)
• cooperate internationally to scale up studies for greater statistical power
• continue to develop and support cheminformatic and bioinformatic tools, including network theory and network medicine
• ensure adequate protection against false positives by insisting on independent replication and using methods to establish causation (Mendelian randomization, within-sibling comparisons, exposure-negative and outcome-negative controls)