Preplanned Studies: Acute Effects of Personal Ozone Exposure on Biomarkers of Inflammation, Oxidative Stress, and Mitochondrial Oxidative Damage — Shanghai Municipality, China, May–October 2016

It remains inconclusive whether short-term ozone exposure can cause an inflammatory response and oxidative damage in the circulatory system. We conducted a longitudinal panel study of 43 non-smoking college students with four rounds of visits in Shanghai Municipality, China, from May to October 2016. Personal real-time ozone exposure was monitored for 3 days per visit, followed by blood sample collection. We measured 10 inflammatory biomarkers, 2 oxidative stress biomarkers, and 2 indicators of mitochondrial oxidative damage in the blood. Linear mixed-effect (LME) models were used to analyze their associations with ozone in R software (V 3.5.2, R Foundation for Statistical Computing, Vienna, Austria) and a P-value <0.05 was considered statistically significant. During the study period, mean ozone exposure levels ranged from 18.0 ppb to 22.7 ppb at different lag periods. Each 10-ppb increase in ozone exposure was greatly associated with a 4.86% increase in tumor necrosis factor alpha (TNF-α), 3.14% in soluble intercellular adhesionmolecule-1 (sICAM-1), and 0.89% in malondialdehyde (MDA), as well as a decrease of 0.23 in the average methylation (%5mC) of the mitochondria displacement loop (D-loop) region. Our results suggest that even short-term exposure to low-level ozone may lead to inflammation, lipid peroxidation, and mitochondrial oxidative damage.

For each visit, personal ozone exposure was monitored in real time for 3 days during daytime (from 8∶00 to 18∶00) using Personal Ozone Monitors (POMs, 2B Technologies, US). In addition, we used HOBO data loggers (Onset Computer Corporation, Pocasset, MA, US) to monitor temperature and relative humidity at the individual level. For each individual, we collected blood samples immediately after ozone monitoring. Serum levels of 10 inflammatory biomarkers were measured using the MILLIPLEX MAP Human Cytokine/Chemokine Kit (Millipore Corp., Billerica, MA, US), including interleukin-6 (IL-6), interleukin-8 (IL-8), interleukin-10 (IL-10), interleukin-17 alpha (IL-17α), TNF-α, monocyte chemoattractant protein-1 (MCP-1), granulocyte macrophage colony stimulating factor (GM-CSF), sICAM-1, soluble vascular cell adhesion molecule-1 (sVCAM-1), and vascular endothelial growth factor (VEGF). Overall, 2 markers of oxidative stress, including 8-isoprostaglandin F2 alpha (8-iso-PGF2α) and MDA, were measured in serum using enzyme linked immunosorbent assay (Oxford Medical, UK). We also measured the relative mitochondrial DNA copy number (RMtDNAcn) and the methylation of the mitochondria D-loop region in blood to indicate mitochondrial oxidative damage, according to previous methods (1). The study was approved by the Institutional Review Board of the School of Public Health, Fudan University (No. 2014-07-0523). All participants signed informed consent at enrollment.

We applied LME models to analyze the associations between personal ozone exposure and these biomarkers, with each participant’s identification number incorporated as random intercepts. Data on 10 inflammatory biomarkers, 2 oxidative stress biomarkers, and RMtDNAcn were naturally log-transformed before statistical analyses. To capture the lagged effects of ozone, we considered exposure windows of 0–2 hours, 3–5 hours, 6–8 hours, 0–8 hours, 1 day, and 2 days preceding the blood sample collection. We also included: 1) individual characteristics, including age, sex, body mass index (BMI); 2) two natural cubic smooth functions of temperature and relative humidity with 3 degrees of freedom (dfs) for both; and 3) a natural cubic smooth function of the day within the study period with 3 dfs. We further performed a sensitivity analysis by adjusting for fine particulate matter, sulfur dioxide, nitrogen dioxide, and carbon monoxide, separately.

Three asthmatic patients were excluded, leaving 40 participants (30 females and 10 males) for the current analysis. Their mean age was 24 years, and their mean BMI was 21 kg/m². As shown in Supplementary Table S1, the average personal ozone exposure at different lags ranged from 17.8±20.5 ppb to 22.7±17.4 ppb, which was far below the current ambient air quality standard in China (i.e., 8-hour maximum: 160 μg/m³, equivalent to 75 ppb). The mean levels of 14 biomarkers varied appreciably, and detailed descriptive data can be found in Table 1.

Among the 10 inflammatory biomarkers, we observed significant increases in TNF-α, sICAM-1, sVCAM-1, and VEGF with ozone exposure (Figure 1). However, we did not find any significant associations of ozone with IL-6, IL-8, IL-10, IL-17α, MCP-1, GM-CSF (Supplementary Figure S1). We also found that ozone exposure showed a positive and statistically significant association with MDA, while the association with 8-iso-PGF2α was insignificant (Figure 2). In addition, we observed a significant increase in RMtDNAcn but a significant decrease in the D-loop methylation with ozone exposure (Figure 2).

Figure 2. 

Percentage changes in 8-iso-PGF2α (A), MDA (B), and RMtDNAcn (C), and absolute change in mitochondria D-loop methylation levels (D) associated with a 10-ppb increase in personal ozone exposure at different lag periods in 40 college students in Shanghai Municipality, China, May–October 2016.

Note: X-axis denotes to the different lag periods; Y-axis denotes the corresponding percentage/absolute change in biomarkers. Abbreviations: 8-iso-PGF2α=8-isoprostaglandin F2 alpha; MDA=malonaldehyde; RMtDNAcn=relative mitochondrial DNA copy number.

The lag patterns of these significant associations were similar, except for the association with MDA. These effects mostly occurred within 2 hours of exposure, then were gradually attenuated over longer lags, and lost statistical significance at a lag of 2 days. At lag 0–2 hours, a 10-ppb increase in ozone concentrations was associated with the following increases: 4.86% [95% confidence interval (CI): 1.39%–8.45%] in TNF-α; 3.14% (95% CI: 0.87%–5.46%) in sICAM-1; 2.23% (95% CI: 0.09%–4.40%) in sVCAM-1; 10.26% (95% CI: 0.65%–20.79%) in VEGF; and 30.51% (95% CI: 0.31%–69.81%) in RMtDNAcn. A decrease of 0.23 (95% CI: 0.01%–0.45%) was found in the D-loop methylation (%5mC). The significant association with MDA was observed at lag 3–5 hours only, with a 0.89% (95% CI: 0.14%–1.65%) increase in MDA per 10-ppb increase in ozone exposure at this lag. After adjusting for other air pollutants, the associations of ozone with TNF-α, sICAM-1, MDA, and the D-loop methylation were almost unchanged, while the associations with sVCAM-1, VEGF, and RMtDNAcn were unstable (Supplementary Table S2).