Assessment of Benzo(a)pyrene-equivalent Carcinogenicity and Mutagenicity of Residential Indoor versus Outdoor Polycyclic Aromatic Hydrocarbons Exposing Young Children in New York City

The application of benzo(a)pyrene (BaP)-toxic equivalent factor to polycyclic aromatic hydrocarbons (PAH) concentrations can provide a more accurate risk assessment from environmental exposure to PAH. We hypothesized that BaP-equivalent toxicity determined following residential air monitoring among young urban children may vary by season. Residential indoor and outdoor air levels of PAH measured over two-weeks in a cohort of 5–6 year old children (n = 260) in New York City were normalized to the cancer and mutagen potency equivalent factor of BaP (BaP = 1). Data are presented as carcinogenic equivalents (BaP-TEQ) and mutagenic equivalents (BaP-MEQ) for the sum of 8 PAH (Σ8PAH; MW ≥ 228) and individual PAH and compared across heating versus nonheating seasons. Results show that heating compared to nonheating season was associated significantly with higher (BaP-TEQ)Σ8PAH and (BaP-MEQ)Σ8PAH both indoors and outdoors (p < 0.001). Outdoor (BaP-TEQ)Σ8PAH and (BaP-MEQ)Σ8PAH were significantly higher than the corresponding indoor measures during the heating season (p < 0.01). These findings suggest that at levels encountered in New York City air, especially during the heating season, residential exposure to PAH may pose an increased risk of cancer and mutation.


Introduction
Many polycyclic aromatic hydrocarbons (PAH) are suspected or known carcinogens and mutagens [1][2][3]. Benzo(a)pyrene (BaP), believed to be the most toxic PAH, has been well-characterized toxicologically. However, less information is available for most of the other PAH. In most risk assessments, many individual PAH have been considered to be of equivalent toxicity as BaP [4]. This approach could result in the overestimation of cancer and mutagen potency of individual PAH because most PAH are considerably less toxic than BaP when analyzed in the same assay systems (e.g., mouse skin, human cell mutagenicity assays) [1][2][3]5].
Several approaches have been developed to obtain a more accurate assessment of potential risk of exposure to a complex mixture of PAH using toxic equivalency factors based on BaP [1][2][3][6][7][8]. One is the carcinogenic equivalency factors (TEF) that can vary at either low or high dose settings. For example, TEF of dibenz(ah)anthracene (DahA) is around 5 at low dose and close to 1 at higher dose based on local tumors induced by subcutaneous injection into mice [9]. Mutagenicity of individual PAH relative to BaP also has been assessed using the mutagenic equivalency factor (MEF) proposed by Durant et al. [1,2]. Mutagenic activity, while not as uniformly associated with cancer [10][11][12], may have implications for other non-cancerous adverse health effects, such as pulmonary diseases [13,14].
TEF and MEF values in combination with measured air concentrations have been used for the calculation of carcinogenic equivalents (TEQ, expressed in ng/m 3 ) and mutagenic equivalents (MEQ, expressed in ng/m 3 ) in environmental samples [15][16][17]. To date, these studies have yielded important reassessments of the contributions of exposure to PAH, including those derived from traffic emissions, on lung cancer risk [17]. However, most pediatric cohort research that addresses the adverse health effects of exposure to air pollution, and PAH specifically, have compared levels of airborne PAH, or their metabolites, with clinical outcomes [18]. Use of the TEQ or MEQ may lead to a more accurate assessment of potential health risk in a pediatric cohort.
We hypothesized that BaP-equivalent toxicity, when assessed using residential monitoring in a pediatric urban cohort, varies by season due to higher emissions from heating sources and more frequent stagnant meteorological conditions in the winter. Our approach was to calculate BaP-TEQ and PAH compound with its TEF for cancer potency relative to BaP [3] and MEF relative to BaP [1][2] Each air monitoring result was assessed and flagged if there are any issues of sampling conditions such as tube disconnection from the pump, late-takedown, pump failure, switch error, and any other human errors. Once flagged, air monitoring data were given a quality assurance (QA) score of 1 (0: highest quality) and further examined for additional score for erroneous length of sampling time, erroneous flow rate of pump, and missing documentation. If a final QA score is ≥ 3, the data was excluded, sampling was redone. Flagged data were included for analysis if they passed a quality control test (QA  2), as described [25]. Five failed the quality control test.
Mean recovery of deuterated surrogate standards was 97.9% (17% Standard deviation, SD) and 102.6% (15%, SD) for d10 anthracene and d14-p-terphenyl, respectively in all batches except for one. In one batch of measures, the mean recovery efficiency exceeded 130% in some samples (attributed to evaporation during storage) and adjustment was made downward by the multiplier 100/ (mean recovery) and included for the data analysis. The limit of detection (LODs) for 8 individual PAH was 0.03 ng/m 3 .
Statistical analysis. Descriptive statistics were used to describe overall BaP-TEQ and BaP-MEQ concentrations for individual PAH and  8 PAH indoors and outdoors. Due to the non-normal distributions of individual PAH and the sum of 8 PAH ( 8 PAH) concentrations, Mann-Whitney and Wilcoxon signed ranks test were conducted. Heating season was defined as any sampling that was initiated October 1st through April 30st as described [21]. Indoor and outdoor comparisons were based on the subset of homes in which both indoor and outdoor air concentrations were measured simultaneously. Analyses were conducted using SPSS software (SPSS; Chicago, IL, version 17).

Residential indoor and outdoor BaP-equivalents
The BaP-equivalent (BaP-TEQ and BaP-MEQ) concentrations calculated for  8 PAH and individual PAH measured in this study are shown in Table 1. The levels of indoor (BaP-TEQ) ∑8PAH and (BaP-MEQ) ∑8PAH ranged from 0.098-8.348 ng/m 3 and 0.069-19.72 ng/m 3 , respectively. For all samples studied, the largest contribution of individual PAH to (BaP-TEQ) ∑8PAH and (BaP-MEQ) ∑8PAH was made by BaP, followed by DahA for (BaP-TEQ) ∑8PAH and IP and BghiP for (BaP-MEQ) ∑8PAH (Table 1). In outdoor air, the contribution of BbFA to (BaP-MEQ) ∑8PAH was substantially elevated when compared to indoor air, making it the dominant compound contributing to (BaP-MEQ) ∑8PAH .

Seasonal variations in BaP-equivalents
Heating compared to nonheating season was associated significantly with higher (BaP-TEQ) ∑8PAH and (BaP-MEQ) ∑8PAH both indoors and outdoors (Figure 1-a and Figure 1-b; p < 0.001, Mann-Whitney test). This pattern was apparent when the individual 8 PAH were assessed (p < 0.001).
In addition, during the heating season, outdoor (BaP-TEQ) ∑8PAH and (BaP-MEQ) ∑8PAH were significantly higher than the corresponding indoor measures ( Table 2; p < 0.01, Wilcoxon signed ranks test). A similar pattern was not observed during the nonheating season ( Table 2; p > 0.05, Wilcoxon signed ranks test). Similarly, most individual 8 PAH (except BaP and IP) were higher outdoors compared to indoors during the heating season (p < 0.05). BbFA and Chry were significantly higher outdoors than indoors regardless of season (p < 0.05, Wilcoxon signed ranks test).

Discussion and Conclusions
Our objective was to estimate the potential carcinogenic and mutagenic risks of residential exposure to PAH for a cohort of inner city young children based on BaP-equivalent concentration (BaP-TEQ and BaP-MEQ). We found a significant effect of heating season on BaP-TEQ and BaP-MEQ possibly due to higher emissions from heating sources and more frequent stagnant meteorological conditions in the winter. While an effect of heating season has been documented in small studies [15,16], this is the first paper to examine seasonal differences in carcinogenic and mutagenic risks based on residential exposure of a large cohort of urban children.
The indoor (BaP-TEQ) ∑8PAH levels observed in this study were considerably lower than those reported for other homes impacted heavily by industrial and traffic emissions in Ohura et al. [15]. (BaP-TEQ) ∑8PAH risk levels depend not only on concentrations of individual 8 PAH, but also the composition of PAH mixtures affected by varying emission sources. While  8 PAH concentrations measured in the Ohura study were 1.3-2.2 times higher than those measured in this study, (BaP-TEQ) ∑8PAH levels were 2-3.5 times higher. Consistent with Ohura et al. [15], we observed that BaP was the dominant compound contributing to (BaP-TEQ) ∑8PAH , accounting for 45% of indoor (BaP-TEQ) ∑8PAH and 35% of outdoor (BaP-TEQ) ∑8PAH .
While the contributions of the sum of IP and BghiP to (BaP-TEQ) ∑8PAH both indoors and outdoors were only 11-12%, their contributions to (BaP-MEQ) ∑8PAH were much higher considerably (45-48%). Considering that these compounds are considered tracers of vehicular emissions [26], exposure to traffic emissions may impact the mutagenicity risk to a greater extent than the carcinogenic risk.
Both indoors and outdoors, higher (BaP-TEQ) ∑8PAH and (BaP-MEQ) ∑8PAH risks were observed in the heating season, compared to the nonheating season. The heating season may be associated with (1) increased use of fossil fuel combustion for residential heating (2) reduced air exchange rates (AERs) [27], (3) reduced PAH transformation through photochemical/chemical reaction due to a lower temperature and ozone concentration [28], (4) gas/particle partitioning in favor of the particulate phase with lower temperature and (5) frequent stagnant meteorological conditions such as a lower mixing height. Presumably some combination of these characteristics of heating season led to changes in either concentrations or relative composition that was pertinent to the calculation of these measures.
Several studies reported that heavier 5-7 ring PAH have higher airborne measures outdoors compared to indoors, due to the presence of major outdoor emission sources (i.e., traffic sources, industry, and power generation etc) of those compounds [15,29]. Heavier PAH concentrations indoors are strongly affected not only by the outdoor concentrations but also by outdoor-to-indoor AERs that were shown to be lower during winter than summer when residential windows usually are shut [15,27,29]. Similar trends were obtained for (BaP-TEQ) ∑8PAH and (BaP-MEQ) ∑8PAH during the heating season. This result suggests that children are subjected to higher carcinogenic and mutagenic risks derived from PAH when they play outside more so than inside the home during the heating season. When compared to other urban outdoor values measured in the winter season [15][16][17], outdoor (BaP-TEQ) ∑8PAH and (BaP-MEQ) ∑8PAH observed in this study were 2-5 times lower during the heating season.
These risks measured and calculated here raise significant concerns for public health. For example, applying the World Health Organization (WHO) suggested unit risk of 8.7  10 -5 (ng/m 3 ) -1 for lifetime (70 years) PAH exposure [30], the corresponding lifetime lung cancer risks were found to be 4.2  10 -5 (= 0.478 ng/m 3  8.7  10 -5 (ng/m 3 ) -1 ) and 5.1  10 -5 (= 0.590 ng/m 3  8.7  10 -5 (ng/m 3 ) -1 ) when the average indoor (BaP-TEQ) ∑8PAH and (BaP-MEQ) ∑8PAH concentration were used, respectively. Hence, if 1,000,000 people were exposed to 0.478 ng/m 3 of indoor (BaP-TEQ) ∑8PAH for 70 years, then 42 people may develop lung cancer. An excess lung cancer risk from lifetime exposure to (BaP-MEQ) ∑8PAH concentration would be 51 cases among one million individuals exposed. The estimated cancer risks from air pollution in NYC did not exceed a health-based guideline (8.7  10 -5 ) calculated based on the maximum permissible level of 1 ng/m 3 of BaP [30]. However, it should be noted that the risk estimates presented are very uncertain, and could be understood only as a crude estimation of cancer risk from the PAH inhalation.
We acknowledge study limitations. The potential risk of PAH exposure based on TEQ or MEQ may be underestimated if the interaction of some PAHs are synergistic rather than additive. Chemical degradation of PAH by ambient oxidants (ozone, hydroxyl radical, or nitrogen oxides) in the atmosphere as well as on the filters during sampling could also underestimate the potential risks due to reduced measured PAH levels, as shown in several studies [28,[31][32][33][34][35][36]. For example, BaP concentrations collected without denuder that can remove atmospheric oxidants such as ozone and OH radical can be underestimated by more than 200% of the measured value at high ozone levels in summer [28,31]. Also PAH collected on the filters may be decomposed through heterogeneous chemical reactions with ozone during an extended sampling period [28,[34][35][36]. Furthermore, nitrated/oxygenated PAH compounds have not been measured in this study, underestimating the full carcinogenic and mutagenic potential of PAH exposure. Those PAH compounds formed by photochemical/chemical reactions are known to be more toxic than their parent PAH based on tumorigenicity of PAH in a newborn mouse assay [8,37,38]. Young children can be exposed to PAH through other routes besides inhalation. These include ingesting food, nondietary ingestion of dust or soil through hand-to-mouth activity, or dermal contact with soil polluted by PAH [39,40]. Although inhalation is an important pathway for inner-city children because of high levels of PAH measured in indoor and outdoor air, dietary ingestion and non-dietary ingestion pathways are thought to be more important for young children's exposure to heavier PAHs [39,40]. Thus, the values reported in this study may need to be considered as the lower limit of estimated potential PAH health risk resulted from inhalation of air. Further investigations are needed whether BaP-equivalent levels are associated with any observed health outcomes (i.e., respiratory or allergic symptoms etc.) within the cohort.
In conclusion, we found that heating season is an important contributor to the potential risk of PAH exposure. This finding has implications for the design of environmental health studies that focus on air pollution exposure and young children living in the inner city.