Vehicular Particulate Matter (PM) Characteristics Impact Vascular Outcomes Following Inhalation

Abstract

Roadside proximity and exposure to mixed vehicular emissions (MVE) have been linked to adverse pulmonary and vascular outcomes. However, because of the complex nature of the contribution of particulate matter (PM) vs. gases, it is difficult to decipher the precise causative factors regarding PM and the copollutant gaseous fraction. To this end, C57BL/6 and apolipoprotein E knock out mice (ApoE^−/−) were exposed to either filtered air (FA), fine particulate (FP), FP+gases (FP+G), ultrafine particulate (UFP) or UFP+gases (UFP+G). Two different timeframes were employed: 1-day (acute) or 30-day (subchronic) exposures. Examined biological endpoints included aortic vasoreactivity, aortic lesion quantification, and aortic mRNA expression. Impairments in vasorelaxation were observed following acute exposure to FP+G in C57BL/6 animals and FP, UFP and UFP+G in ApoE^−/− animals. These effects were completely abrogated or markedly reduced following subchronic exposure. Aortic lesion quantification in ApoE^−/− animals indicated a significant increase in atheroma size in the UFP, FP and FP+G exposed groups. Additionally, ApoE^−/− mice demonstrated a significant fold increase in TNFα expression following FP+G exposure and ET-1 following UFP exposure. Interestingly, C57BL/6 aortic gene expression varied widely across exposure groups. TNFα decreased significantly following FP exposure and CCL-5 decreased in the UFP, FP and FP+G exposed groups. Conversely, ET-1, CCL-2 and CXCL-1 were all significantly upregulated in the FP+G group. These findings suggest that gas-particle interactions may play a role in vascular toxicity, but the contribution of surface area is not clear.

1. Introduction

Exposure to air pollution is associated with greater risk for vascular morbidity and mortality[1–5]. Epidemiological studies have identified that the size and shape of PM are linked to pathological outcome severity[6–8] and numerous reports highlight that near-roadway emissions may have a more potent impact of cardiovascular health[1, 9]. Toxicological studies suggest that particulate matter and gaseous components of freshly generated vehicle exhaust may physically associate to interactively promote cardiopulmonary toxicity[9–11]. We have recently noted that smaller ultrafine particulates (UFP), with a higher surface area per mass basis than fine particles (FP), exhibited potentiated pulmonary and neurological outcomes in combination with copollutant vehicular gases[11]. Acute inhalation exposure to UFP and UFP with gases (UFP+G) resulted in increased IL-1β, IL-6, Tgf-β, and TNF-α mRNA in the hippocampus of ApoE^−/− mice. Additionally, exposure to UFP+G particles resulted in increased CXCL1 and TNF-α protein in bronchoalveolar lavage fluid. In general, inflammatory responses were the most pronounced in the UPF+G group, suggesting potential surface area-dependent interactions of gases and particles as a primary driver of toxicity. Using this exposure paradigm, we sought to address whether systemic vascular toxicity was similarly driven by gas-particle interactions.

Cardiovascular effects of vehicular emission inhalation have been substantiated in the literature, particularly for diesel exhaust[2, 12–15]. Mechanisms of increased risk to vascular disease generated by PM inhalation largely involve pulmonary and systemic inflammation and subsequent endothelial dysfunction. Atherosclerosis-prone ApoE knockout mice (ApoE^−/−) exposed to subchronic inhalation of diesel exhaust (7 weeks) exhibited a significant increase in levels of alveolar macrophages and alterations to atherosclerotic plaques including enhanced oxidative stress and DNA oxidation[12]. However, overall plaque size was not increased. Similar studies of inhaled whole diesel emissions also noted a lack of plaque growth, despite evidence of increased inflammation, plaque complexity and vasomotor effects[12, 16]. A study of oropharyngeal exposure to just diesel particulates, in the absence of copollutant gases such as carbon monoxide and nitrogen dioxide, did show pronounced enhancement of lesions size and complexity[17]. Additionally, exposure to ambient PM, which contains a portion of diesel exhaust PM, has routinely promoted plaque growth in vulnerable mice[18, 19].

Gasoline engine emissions (GEE) effects on cardiopulmonary dysfunction are less studied than the effects of diesel emissions. Vascular mRNA expression of MMP-2/9, endothelin (ET)-1, TIMP-2, and ROS were significantly elevated in an ApoE^−/− animal exposure model following 1 or 7 days of GEE inhalation[20]. In a more recent study, no changes to pro-inflammatory or oxidative stress were observed through GEE exposure using a multi-cellular in vitro approach, however repeated exposures increased TNF-α and HMOX-1 mRNA expression[21]. Gasoline emissions, generally speaking, contain a greater proportion of gaseous species compared to PM, while conventional diesel emissions have a higher proportion of PM emissions. Initial studies combining diesel (with high PM) and GEE (with high volatile organic gaseous composition) to model an urban background show a high degree of interaction in terms of vascular lipid peroxidation[22].

While alternative fuel sources and various additives are emerging, and improvements have been made to limit the release of diesel emissions, gaseous components of these fuels remain a concern as they may interact with novel fuel particulates and background PM. Emerging fuels such as biobutanol or blends with diesel and other similar fuels result in different PM/gas concentrations and ratios[23]. However, soot nanostructure and PM morphology are not affected by alcohol fumigation or engine load[24]. Additionally, the sum of inorganic ions from biodiesel vehicular emissions is significantly higher compared to traditional diesel and gasoline combustion products[25]. Assessment of bronchoalveolar lavage fluid in a study of acute diesel and biodiesel exposure revealed an increase in neutrophil influx following inhalation of either 600 or 1200 μg PM/m³ biodiesel derived from sewage methyl esters[26]. Moreover, the addition of cerium oxide nanoparticles to diesel exhaust, for increased fuel burning efficiency, has been reported to increase in the gaseous components and UFP amount emitted[27, 28]. Alterations caused to the emission profile have been reported to induce adverse pulmonary and systemic effects at levels where the gas phase components are below the expected impact[29]. Moreover, deciphering molecular interactions between air pollution particles with various material components and morphologies remains an important issue for effective protection of human health and environmental policy development.

A recently published study from our group has provided toxicological evidence suggesting that PM surface area and adsorptive gaseous species interact to influence both pulmonary and neuroinflammatory outcomes[11]. Moreover, our previous research has identified a link between gaseous and particulate components and cardiovascular toxicity[9, 10]. However, there is limited understanding regarding whether vascular toxicity is influenced by particulate size and shape or chemical composition. To assess this, we used our established exposure paradigm in combination with an animal model of atherosclerosis[11]. ApoE^−/− and C57BL/6 mice were exposed to carbonaceous (engine-derived) UFP or FP with or without the vapor phase of mixed vehicular emissions (MVE). We hypothesized that greater surface area on UFP would allow for greater gas absorption onto the surface, thereby causing an increased inflammatory reaction in the lung and subsequent downstream vascular involvement.

2. Methods

2.1. Animals

Male C57BL/6J mice were purchased from Jackson Laboratories (Bar Harbor, Maine) at 6–8 weeks old and housed at Lovelace Respiratory Research Institute (Albuquerque, NM). Mice were allowed to acclimate for one week and the institutional IACUC protocol was followed in addition to the PHS Policy on Human Care and Use of Laboratory Animals. A second animal model was employed to examine the effect of these inhaled mixtures in a susceptible vascular model. Apolipoprotein E-deficient mice (ApoE^−/−), which develop high cholesterol when fed a high-fat chow diet, were also purchased at 6–8 weeks of age from Jackson laboratory. Animals were pair-housed and maintained in a controlled environment (30–60% relative humidity, 20–24 °C), kept on a 12h light:dark cycle, and had unrestricted access to water. Food was removed during the daily exposures. C57BL/6 animals were fed a normal chow diet (standard Harlan Global Certified Rodent Chow) and ApoE^−/− animals were fed a high-fat diet (Harlan 21% fat and 0.2% cholesterol). Both animal models were randomized and exposed to one of five treatment groups described below. Animal weights were monitored weekly throughout the study and all animals were fasted overnight prior to euthanization. Animals were euthanized using an IP injection of Euthasol. Tissues were rapidly dissected and either flash frozen and stored at −80 °C until processed or used directly for bioassays.

2.2. Study design and MVE production

To assess the contribution of PM vs. gases, the exposure paradigm involved acute (6 h) and subchronic (30 days/6h per day) 300 μg PM/m³ exposures (described below) and five groups depicted in Figure 1: filtered air controls (FA), fine particulate (FP), ultrafine particulate (UFP), fine particulate and gases (FP+G), and ultrafine particulate and gases (UFP+G) using 8 mice per strain per group. FP and UFP were isolated by denuding the gases from MVE (described below). Target levels for PM and gases were derived based off of previous publications with methods and system schematics described in further detail[9, 22]. Roadway MVE was generated based on a model of combined diesel engine emissions (DEE) and gasoline engine emissions (GEE)[30, 31]. DEE was produced from a single-cylinder, 5500-watt Yanmar diesel-engine generator. Heavy sulfur fuel and Number 2 Diesel Certification Fuel (Philips Chemical Company) and 40 weight motor oil (Rotella T, Shell) was used for the combustion process. Exhaust was subsequently diluted with FA until the desired concentration was reached. GEE exhaust was generated from a 1996 General Motors 4.3 L V6 gasoline engine with a stock exhaust system, muffler and catalyst. A dynamometer (Model Alpha 240, Zöllner, Kiel, Germany) was linked to an interface (Type DTC-1, Dyne Systems Co., LLC, Germantown, WI) and software to record emissions output (Cell Assistant, Dyne Systems Co.). Oil filter (Duraguard PF52, AC Delco, Detroit, MI) and crankcase oil (10 W-30, Pennzoil, Houston, TX) were replaced every 122h (~ 3,000 driving mi.) during engine operation. MVE was created by combining a dilution of GEE and DEE to reach a target of 300 ug PM/m³ as previously described[11]. Animals were exposed for 1 day (Figure 2) or 30 days (Figure 3) to either FA, FP, FP +G, UFP, or UFP+G.

Fig. 1. Study Design.

C57BL/6 and ApoE null mice were exposed acutely (6 hours) and subchronically (30 days/6 hours per day) to either filtered air (FA), fine particulate (FP), FP+gases (FP+G), ultrafine particulate (UFP) or UFP+gases (UFP+G) using 8 mice per strain per group.

Fig. 2. Acetylcholine-mediated vasorelaxation following 1 day (6 hours) inhaled exposure.

A) C57BL/6 animals exposed to FP and FP+G as compared to FA B) C57BL/6 animals exposed to UFP or UFP +G as compared to FA C) ApoE^−/− animals exposed to FP and FP +G as compared to FA D) ApoE^−/− animals exposed to UFP or UFP +G as compared to FA. Data are represented as mean ± SEM. Asterisks indicate significant difference from FA control group by two-way ANOVA using an Uncorrected Fisher’s Least Squared Differences test and are color matched to group of significance (*p<0.05, **p<0.01, ***p<0.001; n=6–8 per group).

Fig. 3. Acetylcholine-mediated vasorelaxation following 30 days inhaled exposure.

A) C57BL/6 animals exposed to FP and FP+G as compared to FA B) C57BL/6 animals exposed to UFP or UFP+G as compared to FA C) ApoE^−/− animals exposed to FP and FP+G as compared to FA D) ApoE^−/− animals exposed to UFP or UFP+G as compared to FA. Data are represented as mean ± SEM. Asterisks indicate significant difference from FA control group by two-way ANOVA using an Uncorrected Fisher’s Least Squared Differences test and are color matched to group of significance (*p<0.05; n=6–8 per group).

2.3. Denuding the gases

UFP+G was generated using the MVE directly from a 1-m³ mixing chamber containing DEE and GEE. UFP atmosphere was generated by filling a 1-m³ mixing chamber with MVE and subsequently sending it through a Harvard denuder[30]. This process ensures gaseous removal at an 80–90% efficiency rate. FP+G and FP without gases were generated using deposits from the exhaust line of the diesel engine and generated using a similar method as previously described[10, 11]. Deposits were packed in a delivery cup of a Wright Dust Feeder (CH Technologies, USA) and passed through to the inhalation exposure chamber using HEPA-filtered compressed air. The output was then passed through a cyclone with a cut-point of 2.5 μm (URG, Inc. Chapel Hill, NC) for large particle removal. HEPA filtered gases from 1 m³ mixing chambers with MVE were pumped into the exposure chamber for FP+G. For FP, HEPA filtered MVE gases were not added to the exposure chamber.

2.4. Atmospheric monitoring and characterization

Atmospheric characterization was conducted as previously described[22, 30, 31]. Concentrations of PM, carbon monoxide, and nitrogen oxide gases for each of the atmospheres during exposures are listed in Table 1. Gases were analyzed by chemiluminescence (NO_x) and infrared spectroscopy. Gravimetric analysis using 47-mm glass fiber filters was used to analyze particulate mass concentration (GE Whatman, Pittsburgh, PA). Particle size distribution within the chamber was measured with a Fast Mobility Particle Sizer (FMPS, TSI, St. Paul, MN) for the ~10–500 nm size particles and an Aerodynamic Particle Sizer (TSI, St Paul, MN) to measure the 0.5–20 μm size range.

Table 1:

Test Atmosphere concentrations of particulate matter (PM), CO, and NO_x

Atmosphere	Exposure	PM (μg/m3)	NOX (ppm)	CO (ppm)
UFP	1 day	371.3 ± 15.6	7.84 ± 4.14	16.77 ± 1.10
UFP + G	1 day	327.3 ± 35.6	33.11 ± 5.35	27.93 ± 4.05
FP	1 day	315.3 ± 50.7	0.10 ± 0.00*	11.60 ± 0.00*
FP + G	1 day	350.3 ± 47.4	20.17 ± 23.26	73.85 ± 107.76
UFP	30 days	306.3 ± 86.2	4.24 ± 2.80	33.53 ± 40.66
UFP + G	30 days	324.2 ± 75.7	20.80 ± 11.51	107.5 ± 126.5
FP	30 days	308.4 ± 116.3	0.88 ± 2.17	22.03 ± 24.66
FP +G	30 days	318.7 ± 91.7	12.31 ± 10.27	49.36 ± 56.46

^*only one measurement was captured for gas concentrations in the FP chamber per day. This table is a modified reproduction from Tyler et al. 2016 [11] and BioMed Central is the original publisher.

Additionally, important characteristics of the relationships between particulates and semivolatile gases have been described[32].

2.5. Aortic vasoreactivity

To determine the impact of UFP and FP in combination with gaseous co-pollutants on vascular endothelial function, we analyzed acetylcholine-induced vasorelaxation responses in naïve C57BL/6 aortas exposed to serum from C57BL/6 animals and ApoE^−/− animals. Thoracic aorta rings were isolated and cleaned of fat and connective tissue under a dissecting microscope. Thoracic aorta ring segments (2–3 mm in length) were mounted on a 4-chamber force-transducer myograph (610M; Danish Myo Technology A/S, Aarhus, Denmark) and submerged in physiological saline solution (119.0 mM NaCl, 25 mM NaHCO₃, 5.5 mM glucose, 4.7 mM KCl, 1.2 mM MgSO₄, 1.2 mM KH₂PO₄, 0.025 mM EDTA, 2.5 mM CaCl₂) bubbled with carbogen at 37 °C with 21% O₂–5% CO₂ balance N₂ and rings were equilibrated for 30 min at 2 g tension. Normalization as per the DMT protocol was applied to the rings in 2 mN increments until 10 mN over the course of 5 min intervals. Vessel viability was confirmed using a KPSS incubation twice (64.9 mM NaCl, 25.0 mM NaHCO₃, 5.5 mM glucose, 58.9 mM KCl, 1.2 mM MgSO₄, 1.2 mM KH₂PO₄, 0.025 mM EDTA, 2.5 mM CaCl₂). Following an additional 30-min equilibration in PSS, vessels were treated with 1% serum from previously exposed mice[33, 34]. Serum caused contraction of aortic rings, which stabilized over the course of 15 min. Acetylcholine (Ach) was then incrementally added (10⁻⁹ to 10⁻⁴ mM) to the bath to induce vasorelaxation. All data were recorded through LabChart software (AD instruments) and calculated as a percentage of serum-induced contraction and normalized to a baseline tension of 10mN.

2.6. Aortic lesion quantification

Briefly, aortas were dissected from previously-exposed ApoE^−/− mice, cut into 5 mm rings, mounted in O.C.T compound (Fisher Scientific) and subsequently flash frozen. Tissue cryosections (6–8 sections per mouse) at 5 μm were fixed on glass slides and stored at −80^ο until further use. Oil Red O (0.5% in propylene glycol; Polyscientific, Baywood, NY) was prepared as per the manufacturer’s instructions (300 mg of Oil Red O powder was added to 100 mL of 99% isopropanol). Three parts of Oil Red O stock solution was mixed with 2 parts deionized water, subsequently filtered and set aside for further use. Slides were placed in propylene glycol for 2 min and further incubated with Oil Red O solution for 6 min. Slides were then rinsed three times in distilled water. A hematoxylin solution was then incubated with the slides for 1 min. Slides were then washed 3 times with distilled water. A coverslip was subsequently placed over the tissue on the slide using an aqueous mounting medium. Lesions were then visualized and photographed using an Olympus microscope and imaging system. Quantification of staining was performed using the sum of all individual areas measured throughout the aorta (6–8 sections) and executed using Image J (NIH, Bethesda, MD).

2.7. Aortic mRNA gene expression analysis

Transcription analysis was performed using reverse transcription polymerase chain reaction (RT-PCR). Briefly, RNA was extracted from homogenized aortas using a Qiagen RNeasy Mini Kit (Germantown, MD) and cDNA was subsequently synthesized using 500 ng of high-quality RNA, random hexamers (Applied Biosystems) and SuperScript III (Invitrogen) on a ThermoCycler (Model # PTC-200; MJ Research). Gene expression was examined on 50 ng transcribed cDNA using the TaqMan primers: Tnf-α (tumor necrosis factor alpha; Mm00443258_m1), Ccl5 (chemokine ligand 5; Mm01302427_m1), Et-1 (endothelin 1; mM00438656_m1), Ccl2 (chemokine ligand 2; Mm00441242_m1), Vcam (vascular cell adhesion molecule 1; Mm01320970_m1), and Cxcl1 (chemokine C-X-C motif ligand 1; Mm04207460_m1) (Life Technologies) on a StepOne real-time PCR system (Applied Biosystems). Data were log transformed for normalization. Hypoxanthine guanine phosphoribosyltransferase (Hprt1) and Tbp (TATA-box binding protein) were used as internal references. Relative gene expression was calculated using the 2^−ΔΔct comparative threshold method with filtered air-treated mice serving as the reference group. Samples were run in duplicate, using only CT values under 35 for analysis. Data are expressed as mean fold change.

2.8. Statistical analysis

Myography results were analyzed using a two-way analysis of variance (ANOVA) and acetylcholine (ACh) concentration and exposure as the 2 variables using Uncorrected Fisher’s Least Squared Differences as a post-hoc test (GraphPad Prism, v 6.0). Other statistical analyses were conducted using a one-way ANOVA and Dunnett’s test posthoc for multiple comparisons. Data are represented as mean ± SEM unless otherwise stated and a p-value of ≤ 0.05 was considered significant.

3. Results

3.1. Force-tension myography

Serum from C57BL/6 animals exposed to FP+G for 1-day elicited a significant impairment in vasorelaxation response (Figure 2A) as compared to FA controls, while no change in vasorelaxation was observed from serum collected from C57BL/6 animals exposed to FP, UFP, or UFP+G (Figure 2B). In contrast, serum from ApoE^−/− animals elicited significant impairments to vasorelaxation following 1-day exposure to FP and UFP, as well as UFP+G (Figure 2C, D). Although a trend was seen for impairment of relaxation response in ApoE^−/− animals following FP+G exposure, this did not reach statistical significance. No significant alterations were observed in C57BL/6 mice exposed for 30-days to any of the treatments and minor impairments were observed in ApoE^−/− animals exposed for 30-days to either FP or UFP+G (Figure 3). The significant impact of acute UFP exposure on ApoE^−/− animals was completely abrogated following subchronic 30-day exposure.

3.2. Aortic lipid staining

Atherosclerotic lesions in the aortic leaflet region of ApoE^−/− mice following 30-day exposure to the four permutations as well as FA control were examined to investigate alterations in gross lesion area. Oil Red O staining was present in varying degrees in all ApoE^−/− 30-day exposed mice (Fig. 4A–E). Quantification in 30-day exposed high-fat diet fed ApoE^−/− mice indicated that exposure to UFP (p=0.025), FP (p=0.005) and FP+G (p≤0.0001) significantly increased lesion size compared to FA controls (Figure 4F). However, while moderately elevated, lesions from the UFP+G group were not significantly different compared to FA controls (Figure 4F).

Fig. 4. Oil-Red O staining of atherosclerotic lesions in ApoE^−/− mice after 30 days of inhalation exposure.

A) Filtered air (FA) control B) UFP C) UFP+G D) FP E) FP+G F) Lesion quantification. Quantification of staining was performed using the sum of all individual areas measured throughout the valve (6–8 sections). Data are represented as mean ± SEM. Asterisks indicate significant difference from FA control group by two-way ANOVA followed by Dunnett’s posthoc test (*p<0.05, **p<0.01; n=6–8 per group).

3.3. Aortic mRNA expression

Gene expression levels of inflammatory cytokines and markers of vascular toxicity were analyzed for alteration following 30-day inhalation exposure compared to FA control. Transcriptional responses varied the most between exposure groups in mice on a wildtype background (Figure 5), however upregulation of inflammatory and vasoconstrictive markers was observed in aortas of exposed ApoE^−/− mice (Figure 6). Interestingly, TNFα levels were significantly decreased in the FP-exposed C57BL/6 group (p=0.0093). In addition, CCL-5 mRNA levels were significantly lower following exposure to UFP (p=0.0209), FP (p=0.0044), and FP+G (p=0.0022) compared to mice exposed to filtered air only. Conversely, CCL-2 mRNA expression was significantly upregulated in the FP+G group (p=0.0012), as were ET-1 and CXCL-1 (p=0.0454 and p= 0.0167, respectively). No change in aortic VCAM-1 mRNA expression was observed among any of the C57BL/6 exposure groups. Among 30-day exposed ApoE^−/− mice, there were no significant differences in the mRNA expression of CCL-2, VCAM-1, or CXCL-1 across all groups. TNFα was significantly elevated (p=0.003) in mice exposed to FP+G, while ET-1 was elevated in the UFP group (p=0.0410).

Fig. 5. C57BL/6 aortic gene expression following 30-days of inhaled exposure.

TNF-α mRNA expression was decreased following 30 days of exposure to FP while CCL-5 mRNA expression was significantly decreased in C57BL/6 exposed to UFP and to a greater extent FP and FP +G. ET-1 expression was moderately increased in the FP +G exposure group only while CCL-2 was significantly increased in that group. No difference was observed in the mRNA expression of VCAM-1 while CXCL-1was elevated in the FP +G group. Data are represented as mean ± SEM. Asterisks indicate significant difference compared to FA control by one-way ANOVA with Dunnett’s post-hoc test (*p<0.05; **p<0.01; n= 6–8 per group).

Fig. 6. aortic gene expression assessed using quantitative real-time analysis following 30-days of inhaled exposure.

ApoE^−/− TNF-α mRNA expression was increased following 30 days of exposure to FP +G while no difference was seen in CCL5 mRNA expression in any exposure group. ET-1 mRNA expression was increased in animals exposed to UFP. No difference was observed in gene expression of CCL-2, VCAM1, or CXCL-1 among any of the groups. Data are represented as mean ± SEM. Asterisks indicate significant difference compared to FA control by one-way ANOVA with Dunnett’s post-hoc test (*p<0.05; ***p<0.001; n= 6–8 per group).

4. Discussion

The present study sought to identify cardiovascular outcomes associated with different fractions of vehicle engine-derived emissions. We had previously shown that engine-derived UFP combined with gases were more potent at inducing both pulmonary and neuroinflammatory outcomes compared to PM or gases alone, from which we hypothesized that the smaller PM – with a greater surface area per mass concentration – would have a greater interaction with volatile and semivolatile gases in the whole emissions, thereby conferring greater toxicity[22]. While this occurred for neuroinflammatory and pulmonary responses[11], this trend was not seen in the present suite of vascular toxicity outcomes. For the four contrasting exposure atmospheres, all appeared to promote vascular atheroma development. Vasorelaxation, on the other hand, was more substantially impaired by PM-only groups in the ApoE^−/− mice, with the copollutant gases partially mitigating this effect. Fine particles with gases appeared to have the most consistent effect of vascular inflammation. But overall, there was not a clear tendency for UFP+G to drive enhanced vascular toxicity as was observed for lung and neuroinflammation[11].

Exposure to PM contained in traffic pollution has adverse effects on the cardiovascular system[2, 9, 10, 35–37] and near-roadway exposures are associated with greater cardiopulmonary toxicity than ambient PM exposure[38–40]. Epidemiological studies report that exposure to PM from a mobile source results in higher risk of cardiovascular and respiratory sequelae[1, 10, 41, 42]. These observations are likely due in part to the higher concentrations of engine emissions arising from the source, which become more dilute as they diffuse into the air. However, it is also possible that PM toxicity in near-roadway atmospheres is potentiated due to gases and PM physically interacting in the fresh emissions. A substantial body of evidence indicates that the smaller the PM, the greater its toxicity[18, 43]. We have previously reported that combining engine exhaust particulates with gaseous components leads to greater systemic vascular toxicity and increased neuroinflammatory outcomes[10, 11]. Therefore, we hypothesized that PM would interact with gaseous co-pollutants in a surface area-dependent manner to enhance vascular toxicity.

Previous studies have shown that exposure to air pollution components leads to compositional changes in the systemic circulation and that this serum activity can increase vascular inflammation and impair vasodilation[34, 44, 45]. Potentially, this compositional change occurs due to inflammatory activation of peptidases in the lung, leading to shed peptide fragments that retain bioactivity[33, 46]. The present study used serum collected from exposed mice to modify vasodilatory responses in aortic rings from naïve mice. Circulating components in the serum from C57BL/6 mice acutely exposed to FP elicited clear impairments in ACh-mediated vasorelaxation, even at a highly dilute concentration. While serum from C57BL/6 showed no alterations in bioactivity when these animals were exposed to UFP combined with gas, serum from ApoE^−/− animals exposed acutely to FP and especially UFP in the presence of gaseous co-pollutants led to a clear impairment in vasorelaxation. These findings are supported by previous research indicating that combined gas and PM mixtures drive vascular toxicity to a greater extent in ApoE^−/− animals[10]. Impairments in vasorelaxation resulting from acute exposure to FP+G in C57BL/6 animals were abolished following subchronic exposure, suggesting some resolution, acclimation or compensatory mechanism. It is well-known that longer exposure to air pollutants results in attenuation of an inflammatory stimulus over time. Several studies have noted more robust inflammatory responses to acute air pollution rather than longer-term exposures[47–49]. Although not measured in the present study, ApoE^−/− mouse serum has been documented to have high serum lipids. Many studies have demonstrated ApoE^−/− mice on a high-fat chow have cholesterol levels >10 times that of wildtype animals[50, 51]. Despite this difference from C57BL/6 serum, there were no differences in vasodilation in the naïve rings treated with serum from the two mouse strains. Serum compositional changes from the exposures, however, had a substantial impact on vasodilation. It is also possible that vessels from different vascular beds that were not tested within the scope of this study may respond differently to serum compositional changes.

Atherosclerotic lesions in ApoE^−/− mice fed a high-fat diet were increased in size following exposure to FP+G, FP, and UFP. These findings are consistent with numerous other reports showing that PM exposure progresses lesion size[19, 52]. The lack of significant progression in the UFP+G group, however, is also consistent with previous studies of diesel exhaust exposures, where compositional changes in the plaques were observed, but the net plaque size was unchanged[13]. The present study was statistically underpowered to better resolve differences in atherosclerotic progression between pollutant atmospheres, thus we are only able to conclude that PM from vehicular emissions has a consistent and predictable effect on lesion size. A further limitation of our study is the examination of atheroma size at a single anatomical location, as we used the aortic tissue to assess transcriptional changes.

Furthermore, the significant alterations observed in ApoE^−/− animals after acute exposure were markedly reduced or completely abrogated in the 30-day subchronic exposure paradigm. These data indicate that long-term exposure of both C57BL/6 and ApoE^−/− animals to smaller PM may initiate internal protective mechanisms against vascular vulnerability. In addition to the vascular system, compensatory mechanisms may be initiated upon subchronic exposure in other target tissues as well. Indeed, a previous publication from our laboratory suggests that exposure to FP and UFP alone and in the presence of gaseous copollutants elicits neuroinflammatory outcomes following acute exposure, an effect that is mitigated after subchronic exposure in C57BL/6 animals[11]. However, increased expression of proinflammatory cytokines in the hippocampus were elevated in subchronically exposed ApoE^−/− animals.

Analysis of C57BL/6 aortas and ApoE^−/− animals exposed to each PM with and without gases showed varying results with regard to vascular transcriptional effects. C57BL/6 animals exposed to FP+G exhibited increased mRNA expression of CCL1, CCL2, and ET-1 while a decrease was seen for CCL5. Moreover, C57BL/6 animals exposed to FP without gases had decreased expression of TNFα and CCL5. CCL5 was also decreased in C57BL/6 mice exposed to UFP. ApoE^−/− animals exhibited a significant upregulation of TNFα in the group subjected to inhalation of FP+G. Interestingly, UFP+G did not alter the expression of any marker in our inflammatory panel in either animal model. It is possible that fine and ultrafine particles may have diverging pathways of disease etiology with regard to short versus long-term exposure, or that particulate size may play a role in the degree of toxicity depending on the target tissue. It’s also plausible that freshly-generated PM induces disparate vascular outcomes than historical PM. Several studies have shown that UFP leads to increased adverse cardiovascular and pulmonary outcomes in comparison to larger PM[41, 53, 54]. In our study paradigm, UFP were derived using the motor vehicle emissions directly from a mixing chamber with combined DEE and GEE, which models freshly generated PM. In contrast, FP are generated using the deposits from the exhaust line of a diesel engine, representing historically generated or aged PM. Freshly-generated PM has been shown to potentiate cardiovascular outcomes to a greater extent than historical PM[10, 54]. However, fresh vs. aged PM interactions with gases remain unclear.

As a well-known, potent vasoconstrictor, ET-1 has been reported as responsive to inhaled toxicant exposures in numerous studies[20, 55, 56]. It was generally elevated in aortas from C57BL/6 mice exposed to most atmospheres except FP+G, while for ApoE^−/− mice it was only elevated in the FP+G group. ET-1 is not only vasoconstrictive but has been shown to enhance growth of vascular smooth muscle and promote atheroma development[57]. ET-1 may also be released from macrophages. Elevated ET-1 mRNA in the FP+G group of ApoE^−/− mice may simply be due to an influx of macrophages and a change in the population of cell types assayed.

The present study was undertaken to examine potential differential vascular toxicity from four contrasting exposures, with the underlying hypothesis that UFP +G would represent the most potently toxic atmosphere owing to the higher surface area of PM to facilitate gas-particulate adsorptive interactions and deeper pulmonary and potentially cellular delivery of the gaseous components. Despite several previous studies showing potentiation of toxicity of the combined gas-PM mixture[9–11], the present investigation of vascular outcomes suggested relatively non-specific effects occurred for all atmospheres. Our results generally support previous work that inhaled vehicular emissions drive deleterious outcomes of vascular health and function. Additional studies are needed to delineate the underlying molecular mechanisms of gas-particulate interactions and assess the public health impacts of roadway emissions.