Chemical Clustering Analysis of Ambient and Emission Source Particulate Matter Reveals Compositional Determinants of Pulmonary Toxicity Responses

Comparative toxicological studies of heterogeneous particulate matter (PM) samples are needed to evaluate the influence of particle chemistry on pulmonary toxicity outcomes. Here, groups of mice were exposed by oropharyngeal aspiration of a 100 μg dose of one of seven PM samples, including three coarse and two fine ambient air PM, and 2 fine emission source PM. Acute inflammatory and lung injury markers in the bronchoalveolar lavage fluid (BALF) were assessed. A weighted chemical correlation network analysis (WCCNA) clustered PM chemical constituents into four modules based on comodulation within samples. These modules and their components were then correlated with lung toxicity end points. One module represented the highest levels of zinc, lead, copper, and tin, and was strongly correlated with BALF neutrophils, macrophage inflammatory protein-2, and several markers of lung injury. A second module represented the highest levels of several toxic transition metals including magnesium, nickel, vanadium, and cobalt, and was strongly correlated with pro-inflammatory interleukin-6, in addition to neutrophils, albumin, and lactate dehydrogenase. A third module, represented by high levels of elemental carbon, nitrate, sulfate, and phosphate, was correlated with pro-inflammatory tumor necrosis factor-α (TNF-α), in addition to BALF protein and other lung injury markers. The final module consisted of 7 elements associated with the 3 coarse crustal PM samples, and these individual elements exhibited moderate correlations with BALF neutrophils and TNF-α. Toxic transition metals produced the greatest effects on lung toxicity, followed by anions and carbon species. These studies demonstrated that chemical and toxicological assessments of heterogeneous samples of PM produce clusters of chemical constituents that can be correlated with separate toxicological outcomes.