Comprehensive Blood Metabolome and Exposome Analysis, Annotation, and Interpretation in E-Waste Workers

Abstract

Background: Electronic and electrical waste (e-waste) production has emerged to be of global environmental public health concern. E-waste workers, who are frequently exposed to hazardous chemicals through occupational activities, face considerable health risks. Methods: To investigate the metabolic and exposomic changes in these workers, we analyzed whole blood samples from 100 male e-waste workers and 49 controls from the GEOHealth II project (2017-2018 in Accra, Ghana) using LC-MS/MS. A specialized computational workflow was established for exposomics data analysis, incorporating two curated reference libraries for metabolome and exposome profiling. Two feature detection algorithms, asari and centWave, were applied. Results: In comparison to centWave, asari showed better sensitivity in detecting MS features, particularly at trace levels. Principal component analysis demonstrated distinct metabolic profiles between e-waste workers and controls, revealing significant disruptions in key metabolic pathways, including steroid hormone biosynthesis, drug metabolism, bile acid biosynthesis, vitamin metabolism, and prostaglandin biosynthesis. Correlation analyses linked metal exposures to alterations in hundreds to thousands of metabolic features. Functional enrichment analysis highlighted significant perturbations in pathways related to liver function, vitamin metabolism, linoleate metabolism, and dynorphin signaling, with the latter being observed for the first time in e-waste workers. Conclusions: This study provides new insights into the biological impact of prolonged metal exposure in e-waste workers.

Keywords: e-waste; exposomics; mass spectrometry; metabolomics; metals.

Figure 1

Correlation patterns between various study factors, with a particular emphasis on the relationships between metal concentrations and other participant characteristics. BMI: body mass index. Metal elements are highlighted in bold. Se, selenium; Tb, terbium; Cu, copper; Mg, magnesium; Ca, calcium; Sr, strontium; Rb, rubidium; Tl, thallium; Zn, zinc; Mn, manganese; Cd, cadmium; Fe, iron; La, lanthanum; Ce, cerium; Eu, europium; Nd, neodymium; Pb, lead; Y, yttrium.

Figure 2

Workflow for exposomics data analysis implemented in this study.

Figure 3

Sensitivity evaluation of feature detection for asari and centWave. (A) Distribution of all MS1 features in the C18 ESI⁻ mode detected by asari and centWave at different intensity levels. (B) A comparison of trace-level features (intensity < 10,000) detected by asari and centWave.

Figure 4

Summary of statistical analysis and compound identification results. (A) Statistical analysis of MS1 features across different modes using the centWave and asari algorithms. (B) Compound identification results from different chemical universes in various modes, based on MS1 features obtained from either centWave or asari. The modes include C18 ESI⁻, C18 ESI⁺, HILIC ESI⁻, and HILIC ESI⁺. Statistical significance was determined using a two-sided t-test (p-values adjusted by false discovery ratio (FDR), setting the threshold as 0.05).

Figure 5

Summary of exposome categories identified in control and e-waste workers. PFAS, per- and polyfluoroalkyl substances. PMT, persistent, mobile, and toxic substances.

Figure 6

Correlation analysis results. (A) Summary of Pearson’s correlation analysis between metal concentrations and MS features detected by the asari and centWave algorithms across different analytical modes. (B) Linear regression analysis of MS features in the C18 ESI⁻ mode and their associations with copper (Cu) concentration groups. “Non-significant” refers to MS features with no significant associations. “Positive” indicates features showing a significant positive association with metal concentration, while “Negative” refers to features with a significant inverse association. (C) Summary of compounds from both metabolome and exposome chemical databases that were found to be significantly correlated with various metals. Se, selenium; Tb, terbium; Cu, copper; Mg, magnesium; Ca, calcium; Sr, strontium; Rb, rubidium; Tl, thallium; Zn, zinc; Mn, manganese; Cd, cadmium; Fe, iron; La, lanthanum; Ce, cerium; Eu, europium; Nd, neodymium; Pb, lead; Y, yttrium.

Figure 7

Results from functional enrichment analysis. (A) The bubble plot summarizing all functional enrichment results across all metals. Pathways are sorted by p-values, with the most significant pathways (lowest p-values) listed at the top. (B) The top 10 pathways from panel (A) representing the most significantly enriched pathways based on the comparison of metal levels. (C) Scatter plot of functional enrichment analysis from the C18 dataset, comparing e-waste workers to controls. The enrichment factor of a pathway is the ratio of significant pathway hits in the user-uploaded data to the expected number of hits for that pathway. All analyses were performed based on the MS features detected using the asari algorithm.

Figure 8

Dose–response analysis results. (A) Distribution of benchmark doses (BMDs) for the metal copper (Cu). The BMDs of MS features associated with Cu exhibit a normal distribution, with most BMDs concentrated within the Log10 (BMD) range of 2.90–2.95. (B) Distribution of benchmark doses (BMDs) for the metal lanthanum (La). In contrast, the BMDs of MS features associated with La displayed a significantly skewed distribution, with most BMD values falling within the initial range of 0–0.1 on the Log10 (BMD) scale (first two histogram intervals). Red arrows, Log10 (BMD) intervals with the most MS features. The dose–response curves for 7-Hexyl-2-oxepanone (C) and LysoPC (P-16:0/0:0) (D) illustrate distinct dose–response trends for Cu and La, respectively, highlighting the differences in MS feature response patterns to each metal.