Health effects from freshly emitted versus oxidatively or photochemically aged air pollutants

Abstract

Epidemiology studies over the past five decades have provided convincing evidence that exposure to air pollution is associated with multiple adverse health outcomes, including increased mortality. Air pollution is a complex mixture of particles, vapors and gases emitted from natural and anthropogenic sources as well as formed through photochemical transformation processes. In metropolitan areas, air pollutants from combustion emissions feature a blend of emitted particles, oxides of carbon, sulfur and nitrogen, volatile organic compounds, and secondary reaction products, such as ozone, nitrogen dioxide, and secondary organic aerosols. Because many of the primary and transformed pollutants track together, their relative contributions to health outcomes are difficult to disentangle. Aside from the criteria pollutants O₃ and NO₂ and some of the simpler aldehydes (e.g. formaldehyde and acrolein), other products from photochemical processes are a particularly vexing class of chemicals to investigate since they comprise a dynamic ill-defined complex mixture in both particulate and gas phases. The purpose of this review was to describe and compare health effects of freshly emitted versus oxidatively or photochemically aged air pollutants. In some cases, (e.g. single volatile organic compounds) photochemical transformation resulted in marked enhancements in toxicity through formation of both known and unidentified reaction products, while in other examples (e.g. aging of automobile emissions) the potentiation of effect was variable. The variation in experimental design, aging system employed, concentration and type of starting agent, and toxicity endpoints make comparisons between different studies exceedingly difficult. A more systematic approach with a greater emphasis on higher throughput screening and computational toxicology is needed to fully answer under what conditions oxidatively- or photochemically-transformed pollutants elicit greater health effects than primary emissions.

Graphical Abstract

1. Introduction

The chemical composition of Earth’s lower atmosphere is extremely complex. Particles and gases are emitted from many natural processes, such as wind erosion of arid soil, breaking of ocean waves, volcanic activity, forest fires, and the physiological functions of plants and animals. In addition, human activities emit numerous gaseous and particulate pollutants from motor vehicles, power plants, farming operations, mining activities, construction equipment, commercial cooking, home heating, chemical manufacturing, and a variety of industrial processes. From the United States alone, anthropogenic sources emit six million tons of fine particulate matter (PM_2.5), eighteen million tons of volatile organic compounds (VOCs), and an additional eighteen million tons of nitrogen oxides (NO_X) into the atmosphere annually (Rao, Tooly, and Drukenbrod, 2013). The emissions from any given source are typically a mixture of hundreds of different chemical species (Simon et al., 2010).

The combined output from these natural and anthropogenic sources results in tens of thousands of freshly-emitted pollutant species, or primary pollutants, in the lower atmosphere. Once emitted, many air pollutants undergo atmospheric reactions that yield additional chemical species, gases, and particles, known as secondary pollutants. In a sequence of oxidative reactions, the diverse array of freshly-emitted VOCs react with three main atmospheric oxidants, hydroxyl radical (OH˙), ozone (O₃), and nitrate radical (NO₃˙), to produce hundreds of thousands of intermediates and products, such as aldehydes, peroxides, organic nitrates, and organic acid (Aumont, Szopa, and Madronich, 2005). In addition, shortwave radiation from the sun can initiate photochemical reactions that break apart freshly-emitted pollutants to form other products. The combination of oxidative and photochemical reactions is known to form harmful pollutants, including O₃. The oxidation of VOCs often produces compounds with lower volatility than the freshly-emitted VOC. Within a few hours, these products form secondary organic aerosols (SOA) by shifting from the gas phase to the particulate phase (Kroll and Seinfeld, 2008). The mechanisms of SOA formation are poorly understood and are currently the subject of active research. The overall process by which freshly-emitted pollutants are chemically and physically transformed in the atmosphere is herein referred to as aging. Importantly, it is unclear whether and how atmospheric transformation reactions influence the toxicity of emissions and if this varies based on the type of emission.

It is well established that exposure to air pollution is associated with increased morbidity and mortality, and such evidence has been the rationale for the implementation of the Clean Air Act (CAA) in the U.S. Depending on the pollutant, health effects include reduction of host defenses, increased cancer risk, exacerbation of many types of pulmonary and cardiovascular diseases, and premature mortality (Landrigan et al., 2018). The contribution of secondary pollutants to health impacts is well known for some products, such as O₃, but not for many other reaction products or complex mixtures, including SOA. Due to the complexity of the aging process and the multitude of pollutants formed, it is not practical to study the toxicological effects of each individual reaction product. Rather, the approach has been to use various types of photochemical “smog” chambers to simulate the atmospheric aging of a freshly-emitted pollutant and then to assess the toxicity of the resultant mixture. When comparing results across multiple studies, it is important to remember that the different chamber conditions used to simulate aging (e.g., oxidant concentrations and reaction times) can produce different mixtures of pollutants even if the parent compounds are identical.

While epidemiological studies can show associations between air pollution exposure and increased health effects, it is difficult to identify whether these are due to primary pollutants, transformed species, or a combination of both mixtures. Thus, the health impacts from exposure to air pollutants aged through photochemical or oxidative transformation have more often been examined using experimental exposure scenarios in humans, animals, and cell-based toxicity assays. Although these types of systems have been studied for many years, there are few reports of controlled human exposure studies, primarily because of the uncertainty (and potential toxicity) in the subsequent product formation. The controlled human exposure studies that have been conducted measured relatively non-invasive outcomes such as eye irritation, nasal inflammation, neurobehavioral effects, and biomarkers in blood. The most common approaches to assessing the toxicity of aged mixtures have been using inhalation studies in animals or using cell-based systems. Most of the animal toxicological studies examine cardiopulmonary effects such as inflammatory and oxidative stress mediators in the lung, blood, and aortic tissues, as well as carcinogenicity and reproductive endpoints. In vitro studies report on mutagenicity, cytotoxicity, and inflammatory markers (e.g. IL-8, COX-2, IL-6, TNF-alpha,), as well as global genomic approaches to identify novel pathway induction.

The purpose of this review is to describe and compare the potential adverse health effects following exposure to fresh versus aged atmospheres and to identify research gaps for future investigation. We conducted a comprehensive literature search aimed at identifying all relevant research papers. To provide a framework, the various atmospheres were categorized into four sections: single VOC; VOC mixtures; SOA; and combustion-generated emissions. The first three categories were subdivided into photochemical or oxidative transformation, while the final combustion category was further sub-divided by emission source, including automobile, diesel, and coal or wood combustion. Our results indicate that for most atmospheres, aging increases the potency of the emissions, resulting in more significant health effects than the precursor compounds. The results for automobile exhaust, however, were mixed, with several studies reporting that fresh emissions were more potent than the resulting aged mixtures.

2. Approach

We aimed to capture all published studies that examined health effects of aged air pollution when directly compared to freshly emitted, primary products within the same study. To frame the research question and guide the literature search, we defined a PECO (Population, Exposure, Comparator, Outcome) statement:

• Population: The study population of any controlled human exposure or animal toxicological study at any lifestage, or any in vitro model.

• Exposure: Short- or long-term exposure by any method to fresh, primary air pollutants and the same pollutants aged for any duration through photochemical or oxidative transformations.

• Comparator: A direct comparison between freshly emitted air pollutants and their aged counterpart.

• Outcome: Any health-related endpoints.

A comprehensive literature search was conducted by employing key word searches within PubMed and Google Scholar for studies published through August 2019. Key words were based on three categories: air pollution, aging, and health effects. Air pollution terms included VOCs commonly found in outdoor air, SOA, or pollutants from combustion emissions. Aging was represented by terms related to photochemical (e.g. sunlight or UV irradiation) or oxidative reactions (e.g., hydroxyl radical, ozone, nitrate radical). Health effects studies were identified with terms associated with human, animal, and in vitro studies. Additional relevant studies were later identified from the reference lists of relevant articles or referrals from individual scientists.

Our final literature search resulted in 47 studies for inclusion. These studies were divided into four sections: single VOCs, VOC mixtures, SOA, or combustion emissions. The combustion emission section is further divided into automobile emissions, diesel emissions, and coal and wood smoke. While many studies included multiple treatment groups, we focus our report of the results to those that met our study criteria and allowed for a direct comparison of the primary, freshly emitted pollutant to its aged counterpart.

There were two main experimental approaches to studying the health effects of fresh versus aged pollutants in the atmosphere. The first approach, using photochemical transformations, examined individual or mixed compounds + NO_X in the presence of sunlight or artificial UV-irradiation to produce O₃, which can then react with the parent compound to create additional reaction products. The second method, oxidation, combined pollutants directly with O₃ to form reaction products.

3. Results

3.1. Single VOCs

Four studies met our criteria for photochemical transformation of single VOCs (Table 1), all of which included in vitro assays, and one of which also included an animal toxicological assay. Two of these studies examined pulmonary effects, one examined mutagenicity, and one carcinogenicity. Overall, the studies reported greater effects after exposure to the aged versus primary VOC. In addition, we identified eight studies that examined the effects of oxidative transformations of single VOCs (Table 1); one controlled human exposure study that examined ocular effects, six animal toxicology studies examining respiratory tract effects, and one in vitro study examining pulmonary effects. Four of these studies reported statistically significant effects associated with exposure to the aged versus primary VOC, while the other four studies reported mixed or null results with no difference between effects from fresh compared to aged VOCs.

Table 1.

Summary of studies that examined health effects from fresh versus aged single volatile organic compounds (VOCs).

Single VOCs
Transformation Process	Primary Pollutant(s)	Products (if known)	Model System	Aging Duration	Exposure Duration	Outcomes Assessed	Class with Main Effects	Reference
Sunlight	1,3-butadiene + NO isoprene + NO	BD: acrolein, acetaldehyde, formaldehyde ISO: methacrolein, methyl vinyl ketone, formaldehyde	Human lung epithelial cells (A549)		5 h	• Toxicity • Inflammatory gene expression	Aged	Doyle et al. 2004
Sunlight	1,3-butadiene + NO	BD: acrolein, acetaldehyde, formaldehyde, furan, ozone	Human lung epithelial cells (A549)		5 h	• Toxicity • Inflammatory gene expression	Aged	Doyle et al. 2007
UV-irradiation	benzene, toluene, ethylbenzene, o-, m-, and p-xylene, and 1,2,4- and 1,3,5-trimethylbenzene, m-cresol, napthalene (+ NOx)	peroxy acid, aldehyde, alcohol	Salmonella TA100	4.5 h	1, 2, 4, 8, or 16 h	• Mutagenicity	Aged	Riedel et al. 2018
UV-irradiation	benzo[a]pyrene		Mouse fibroblast cell line (NIH3T3)	8–60 h	2 h	• Carcinogenicity	Aged	Teranishi et al. 2010
Oxidation with O3	limonene, isoprene		Humans (male)		20 min	• Eye blink frequency	Aged	Kleno & Wolkoff 2004
Oxidation with O3	a-pinene, limonene, isoprene	Identified products did not account for observed effects; <10% O3 remained	BALB/c mice	16–26 sec	30 min	• Respiratory rate	Aged	Wolkoff et al. 2000
Oxidation with O3	a-pinene, limonene, isoprene		BALB/c mice		60 min	• Sensory irritation • Pulmonary irritation • Airflow limitation	Aged	Rohr et al. 2002
Oxidation with O3	isoprene + NO2	formaldehyde, formic acid, acetic acid, methacrolein, methylvinyl ketone	BALB/c mice	30 sec	30 min	• Respiratory rate • Bronchial constriction • Pulmonary irritation	Aged	Wilkins et al. 2001
Oxidation with O3	limonene	NR	BALB/cA mice (male)	16 sec	1 h/day for 10 days	• Respiratory rate • Airflow limitation • Airway inflammation	Aged (respiratory rate and airflow limitation)	Wolkoff et al. 2012
Oxidation with O3	limonene	NR; 95% degradation of O3	BALB/cJ mice (female) + ovalbumin	16 sec	5 d/week for 2 weeks, then once weekly for 12 weeks	• Airway inflammation • Allergic sensitization	Aged (allergic sensitization)	Hansen et al. 2013
Oxidation with O3	limonene	NR; increase in particle number from <1000/cm−3 to 1.3–1.5 ×107/cm−3	BALB/cJ mice (female) ± ovalbumin	16 sec	3 days	• Sensory & pulmonary irritation • Allergic sensitization	Null	Hansen et al. 2016
Oxidation with O3	limonene	NR	Human lung epithelial cells (A549); differentiated epithelial tissue (MucilAir™)	15–30 min (?)	1 h; 4 h; 1 h/day, 5 days per week for 4 weeks	• Inflammatory cytokine production • Cell proliferation	Mixed	Anderson et al. 2013

3.1.1. Photochemical transformation of single VOCs

Aged olefins increased inflammatory mediators in vitro (Doyle et al., 2004; Doyle et al., 2007). Using a sunlight model system, Doyle et al. exposed human epithelial cells to the reaction products of 1,3-butadiene (BD) or isoprene (ISO) + NO + sunlight and found significant increases in cytotoxicity and IL-8 expression compared to exposure to the parent compounds (Doyle et al., 2004). In a subsequent study, in an effort to identify the specific chemical species responsible for this effect, Doyle et al. compared the reaction products of aged BD to a synthetic mixture of the detected and quantified photochemical products (acrolein, O₃, and formaldehyde), as well as to the equivalent concentration of O₃ generated within the photochemical mixture (Doyle et al., 2007). Exposure to aged BD resulted in an increase in IL-8 protein release compared to the synthetic blend of primary products, suggesting the presence of additional unidentified reactant products.

Mutagenicity assays in Salmonella typhimurium TA100 showed that irradiation of single VOCs, NO_X, and ammonium sulfate seed aerosol resulted in mutagenic atmospheres for eight of the ten VOCs examined (Riedel et al., 2018). Specifically, after aging, benzene, toluene, ethylbenzene, o-, m-, and p-xylene, and 1,2,4- and 1,3,5-trimethylbenzene resulted in mutagenic atmospheres, while m-cresol and napthalene did not. All non-irradiated, fresh VOCs were non-mutagenic at the concentrations tested. The potential chemical species that were responsible for the observed mutagenicity included peroxy acid, aldehyde, and alcohol, among others.

The carcinogenicity of aged benzo[a]pyrene (B[a]P), a semi-volatile organic compound, was evaluated in NIH3T3 cells (a mouse fibroblast cell line) that were inoculated subcutaneously into the left leg of nude mice (Teranishi et al., 2010). Treatment with aged B[a]P, but not fresh, inhibited apoptosis of these cells and resulted in increased production of reactive oxygen species. The surviving treated cells had the ability to form colonies in soft agar and tumors in nude mice. Since primary B[a]P at these concentrations did not affect apoptosis, the carcinogenesis results are attributable to products arising from irradiated B[a]P.

3.1.2. Oxidative transformation of single VOCs

A controlled human exposure study evaluated the health effects of limonene and ISO oxidation products by measuring changes in eye blink frequency as a measure of ocular irritation (Kleno and Wolkoff, 2004). Aged limonene, compared to fresh, increased eye blink frequency and irritant response in adult men.

High dose acute exposures of ISO, limonene, and α-pinene have been shown to cause sensory and pulmonary irritation in animal toxicological studies when combined with O₃. For example, Wolkoff et al. (Wolkoff et al., 2000) exposed mice to aged α-pinene, limonene, or ISO. They found that the reacted atmosphere caused a reduction in respiratory rate of 30%, 33%, and 51%, respectively, compared to the primary compounds. Using the same reaction apparatus and VOCs but with a longer exposure time (60 min vs 30 min), Rohr et al. reported that exposure to the oxidation products, but not to the primary terpenes, resulted in prominent sensory irritation and airflow limitation, with reductions in respiratory rate of 30%, 37%, and 64% and in airflow of 23%, 28%, and 40% for the products of α-pinene, limonene, and ISO, respectively (Rohr et al., 2002). In a similar fashion, Wilkins et al. (Wilkins et al., 2001) exposed mice to fresh or aged ISO with or without NO₂. The aged atmosphere with NO₂ caused a 58% reduction in respiratory rate, which was largely driven by the reaction of ISO with O₃ (52% reduction). Further analysis found that the irritant effect of aged limonene and ISO was dependent on mixing time and relative humidity, with maximum irritation being observed under conditions of low humidity and short reaction times (Wilkins et al., 2003). It was also concluded that the expected reaction products, including formaldehyde, methylvinyl ketone, and methacrolein, were not solely responsible for these effects, and that formation of unidentified products during the oxidative aging process were also likely involved. Wolkoff et al. found that repeated exposure of mice to aged limonene, but not to fresh limonene or O₃ alone, resulted in sensory irritation and limited airflow. There was no change in the presence of inflammatory cells in the bronchoalveolar fluid (BALF) in any treatment groups (Wolkoff et al., 2012).

Hansen et al. exposed ovalbumin-sensitized mice to limonene, O₃, and their reaction products to determine whether inhalation of any of the compounds induced airway inflammation or allergic sensitization when mice were challenged with ovalbumin (Hansen et al., 2013). After a 14-week exposure protocol, a significant increase in ovalbumin-specific IgE was found with exposure to aged limonene as opposed to exposure to either O₃ alone, limonene alone, or ovalbumin alone. In the BALF, there were no differences in effects from exposure to aged or fresh limonene, suggesting that neither compound results in airway inflammation. In a related study, Hansen et al. investigated whether these atmospheres could acutely affect lung irritant responses and other markers of allergic lung disease at or around allergen challenge (Hansen et al., 2016). For three consecutive days, naïve or ovalbumin-sensitized mice were exposed to fresh or aged limonene. There were no statistically significant differences between the two treatments for any endpoints. The authors concluded that aged limonene reaction products could cause sensory and pulmonary irritation in naïve mice but that allergic mice were protected from this effect possibly via a scavenging effect of mucus in the airways.

Mixed effects were reported in an in vitro analysis that utilized both an epithelial cell line (A549 cells) and highly differentiated epithelial tissue (MucilAir) in order to study the relative toxicity of fresh or aged limonene (Anderson et al., 2013). A 4-h exposure to aged limonene resulted in increased IL-8 cytokine production and decreased cell proliferation at 12-h post-exposure in A549 cells compared to exposure to fresh limonene. In contrast, following a 1-h exposure, at 24-h post-exposure, aged limonene resulted in decreased production of the cytokine MCP-1 compared to fresh limonene. In the MucilAir cells, repeated exposure to aged limonene, but not to fresh, increased IL-6 and IL-8 after 3 weeks of exposure, while for MCP-1, exposure to aged limonene decreased production at weeks 1 and 2 compared to fresh limonene.

3.2. VOC mixtures

We identified four studies that examined the photochemical transformation of VOC mixtures (Table 2). These in vitro studies reported that the aging process increased cytotoxicity, mutagenicity, and inflammatory cytokines. We found two studies that examined the effects of oxidative transformation of VOC mixtures (Table 2). Neither of these human studies reported any differences in responses between primary and aged VOC mixtures.

Table 2.

Summary of studies that examined health effects from fresh versus aged volatile organic compound (VOC) mixtures.

VOC mixtures
Transformation Process	Primary Pollutant(s)	Products (if known)	Model System	Aging Duration	Exposure Duration	Outcomes Assessed	Main Effects	Reference
Sunlight	55-component hydrocarbon mixture + NOx	O3, formaldehyde, toluene, PAN, NO2	Human lung epithelial cells (A549)	Sunrise to sunset	5 h	• Inflammatory cytokine production (IL-8 mRNA)	Aged	Sexton et al. 2004
Sunlight	SynUrb54 + NOx (± seed particle)		Human lung epithelial cells (A549)	14 h; 12 h	4 h (gas-only); 30 min (PM-only)	• Cytotoxicity • Inflammatory cytokine release (IL-8)	Aged	Ebersviller et al. 2012
Sunlight	55-component hydrocarbon mixture + NOx + NO2	O3, formaldehyde, peroxyacetyl nitrate	Human lung epithelial cells (A549)	Morning to evening	4 h	• Gene expression (microarray) • Cytotoxicity	Aged	Rager et al. 2011
UV-irradiation	Simulated atmospheres containing NO, ammonium sulfate, gasoline, α-pinene, isoprene, PM2.5, O3, and NO2	Carbonyl compounds (eg, formaldehyde, acetaldehyde, acetone)	Salmonella (TA98, TA100, TA104)	12 h	0–14 h	• Cytotoxicity • Mutagenicity	Aged	Zavala et al. 2018
Oxidation with O3	23 VOCs (e.g. limonene, α-pinene, 1,2,4-Trimethylbenzene)	aldehydes, hydrogen peroxide, organic acids, secondary organic aerosols, ultrafine particles	Healthy humans (women)		135 min	• Neurobehavioral performance • Salivary cortisol • Lung function	Null	Fiedler et al. 2005
Oxidation with O3	23 VOCs (e.g. limonene, α-pinene, 1,2,4-Trimethylbenzene)	aldehydes, hydrogen peroxide, organic acids, secondary organic aerosols, ultrafine particles	Healthy humans (women)		135 min	• Nasal inflammation	Null	Laumbach et al. 2005

3.2.1. Photochemical transformation of VOC mixtures

In vitro exposures to urban-like atmospheres increased cellular production and release of inflammatory mediators. To gain understanding of the health effects of complex mixtures, a 55-component hydrocarbon mixture was developed to represent a typical urban atmosphere. Sexton et al. examined the effects of the fresh or aged mixture on IL-8 expression levels in A549 cells (Sexton et al., 2004). Cells exposed to the sunlight-aged mixture had significantly greater IL-8 expression. These effects were attributed to the reaction products since IL-8 expression was 2–3 times higher than in cells exposed to the same concentrations of O₃ and NOx formed in the reacted atmosphere. The results indicated that gas phase components other than the initial reaction mixture or O₃ contributed to the inflammatory responses induced by exposure to representative urban smog mixture.

The same group (Ebersviller et al., 2012) exposed A549 cells to a 54-component synthetic hydrocarbon mixture (SynUrb54) and NO_X with or without the presence of a seed particle to generate secondary aerosol (i.e. aerosolized mineral oil (MOA)) in a gas-only or particle-only exposure system. The mixture without the seed particles took place in the summer with a longer daylight exposure time (14 h sunlight). The study found that aging of the mixture potentiated cell cytotoxicity and inflammatory cytokine release. The experiment was repeated in September under different sunlight conditions (12 h sunlight) with the addition of MOA. Again, the sunlight-aged mixture produced a greater response compared to the original starting material, however, in this case effects could be seen for both gas-only and particle-only exposures. Additional experiments by Rager et al. (Rager et al., 2011) identified gene expression changes associated with in vitro exposure to the same VOC and NO_X mixture, but with different aging times. Cells were exposed for 4 h to the mixture either in the morning or early evening, thereby allowing the reaction to continue throughout the day. Transcriptional changes were assessed using microarrays 9 h after exposure. The morning exposure upregulated expression levels of 19 genes, whereas extending the reaction throughout the day and conducting a 4 h evening exposure increased production of IL-8 and changed expression of 709 genes. In addition to inflammation, genomic pathways associated with cancer and cardiopulmonary diseases were also altered with the evening exposure. Overall, significantly greater gene expression changes were seen with exposure to pollutants with a longer photochemically aged atmosphere. Furthermore, cytotoxicity was also higher from exposure to the longer aged atmosphere.

The mutagenicity and cytotoxicity of simulated atmospheres were examined in three Salmonella strains at the air-agar interface (Zavala et al., 2018). None of the primary chemical components in the atmospheres were found to be mutagenic or cytotoxic, including NO, ammonium sulfate, gasoline, α-pinene, isoprene, PM_2.5, O₃, and NO₂. In contrast, the photochemically aged atmospheres were cytotoxic and reported as direct-acting base-substitution mutagens with similar potency in TA100 and TA104.

3.2.2. Oxidative transformation of VOC mixtures

Human volunteers were exposed to relatively low levels of 23 VOCs, including limonene, α-pinene, and 1,2,4-trimethylbenzene, with or without the presence of O₃ (Fiedler et al., 2005; Laumbach et al., 2005). Volunteers completed a 37-symptom questionnaire before, during, and after exposure, and nasal lavage was performed pre- and post-exposures. No significant differences in symptoms or markers of nasal inflammation between any exposures were found (Laumbach et al., 2005). Additionally, no differences were reported in neurobehavioral performance, salivary cortisol, or lung function between the fresh or aged atmospheres and filtered air control (Fiedler et al., 2005).

3.3. Secondary organic aerosols (SOA)

We identified only one study that examined the health effects of freshly emitted versus aged SOA (Table 3). Using an in vitro model, this study reported that aging increased toxicity.

Table 3.

Summary of studies that examined health effects from fresh versus aged secondary organic aerosols (SOA).

SOA
Transformation Process	Primary Pollutant(s)	Products (if known)	Model System	Aging Duration	Exposure Duration	Outcomes Assessed	Main Effects	Reference
UV-irradiation	Naphthalene, α-pinene		Human lung epithelial cells (A549)		1–6 h	• Cell viability • Gene expression	Aged (for gene expression, napththalene only)	Chowdhury et al. 2018

3.3.1. Photochemical transformation of VOCs to form SOA

Photochemically aged naphthalene and α-pinene-derived SOA were found to be more toxic than the fresh SOA when assessed via exposure to the air-liquid interface of lung epithelial cells (A549)(Chowdhury et al., 2018). In addition, aged compared to fresh naphthalene-derived SOA had higher peroxide levels and induced an upregulation of genes related to protective gene signaling, inflammation gene signaling, and cell death.

3.4. Combustion

We identified seventeen studies (one controlled human exposure, 14 animal toxicology, and two in vitro studies) that examined the effects of photochemical transformations on automobile (leaded or unleaded gasoline) exhaust (Table 4). Aged exhaust resulted in more potent effects in seven of the studies, while fresh exhaust was more potent in four studies, and mixed or null results were reported for the remaining. For diesel exhaust, one animal toxicology study examined the effects of photochemical transformation, while four examined the effects of oxidation. For coal emissions, one study (in vitro) examined the effects of photochemical transformation, while three (animal toxicology) the effects of oxidation. Finally, we found two studies that assessed the effects of photochemical transformation of wood combustion emissions in vitro. For diesel, coal, and wood smoke emissions, all studies reported more significant effects associated with exposure to the aged mixtures.

Table 4.

Summary of studies that examined health effects from fresh versus aged combustion emissions.

Combustion emissions
Transformation Process	Primary Pollutant(s)	Products (if known)	Model System	Aging Duration	Exposure Duration	Outcomes Assessed	Main Effects	Reference
UV-irradiation	Automobile exhaust		Humans	2 h		• Eye irritation	Aged	Schuck & Rogers 1959
UV-irradiation	Automobile exhaust		Sprague-Dawley rats; mice; guinea pigs		2–6 h	• Mortality • Respiratory rate • Spontaneous activity • Metabolic enzymes • Carboxyhemoglobin levels	Aged (respiratory rate; spontaneous activity; carboxyhemoglobin)	Murphy et al. 1963
UV-irradiation	Automobile exhaust		Swiss, A/j, C57/B, LAF1, 101 mice; Sprague-Dawley rats; Golden-Syrian hamsters; Harley, NIH guinea pigs		7 days/week for 23 months	• Mortality • Morbidity • Growth • Immunology • Blood gas analysis • Oxygen consumption • Pneumonia severity • Pathology • Hematology • Voluntary activity • Reproduction • Bone lead • Pulmonary function	Aged (pneumonia, voluntary activity, reproduction; hematology) Fresh (bone lead)	Heuter et al. 1966
UV-irradiation	Automobile exhaust ± SO2		Beagles		18 months	• Pulmonary function • Upper airway absorption	Null	Vaughan et al. 1969
UV-irradiation	Automobile exhaust ± SO2		Beagles		36, 61 months	• Pulmonary function	Mixed	Lewis et al. 1974
UV-irradiation	Automobile exhaust + Mt. Saint Helens Ash (seed aerosol)		Sprague-Dawley rats (male)		5 h	• Breathing pattern • Bronchoalveolar lavage	Null	Diaz et al. 2012
UV-irradiation	Traffic tunnel plenum particles ± SOA		Sprague-Dawley rats (male)		5 h/day for 2, 4, or 12 days	• Breathing pattern • Bronchoalveolar lavage • Complete blood count • In vivo chemiluminescence of heart and lung	Fresh (4 days)	Diaz et al. 2013
UV-irradiation	Automobile exhaust ± SOx		Beagles (female)		68 months	• Collagen synthesis enzyme • Tissue collagen	Aged (enzyme)	Orthoefer et al. 1976
UV-irradiation	Automobile exhaust ± SOx		Beagles (female)		68 months	• Air space enlargement • Hyperplasia	Aged (air space enlargement) Fresh (hyperplasia)	Hyde et al. 1978
UV-irradiation	Automobile exhaust + Mt. Saint Helens Ash (seed aerosol)	nitric oxide depletion, nitrogen dioxide formation, ozone accumulation, SOA formation	Sprague-Dawley rats (male)		6 h	• In vivo chemiluminescence of heart and lung	Aged (lung only)	Papapostolou et al. 2011
UV-irradiation	Automobile exhaust		Beagles (female)		4 yr	• Blood rheology	Null	Bloch et al. 1973
UV-irradiation	Automobile exhaust		Beagles (female)		5 yr	• Electrocardiographic and vectorcardiographic abnormalities	Null	Bloch et al. 1972
UV-irradiation	Traffic tunnel plenum particles ± SOA		Sprague-Dawley rats (male)		5 h/day,4 days/week for 3 weeks	• Blood pressure • Heart rate	Fresh (blood pressure)	Lamoureux et al. 2013
UV-irradiation	Automobile exhaust		A/j; C57/B; LAF1 mice		15 months	• Bone lead	Fresh	Lutmer et al. 1967
Sunlight	Ambient PM2.5 from central Los Angeles		BV-2 microglia	Collected 6–9am (fresh) or 12–4pm (aged)	24 h	• Inflammatory response • Oxidative activity	Fresh	Lovett et al. 2018
UV-irradiation	Automobile exhaust		Salmonella typhimurium (TA100, TA98)		4–10 h	• Mutagenicity	Aged	Kleindienst et al 1992
Sunlight	Diesel exhaust particles		F344 rats; BALB/c mice (male)	4–5 h	Instillation	• Bronchoalveolar lavage • Oxidative stress	Aged (oxidative stress)	Zielinska et al. 2010
Oxidation with O3	Diesel exhaust particles		Sprague-Dawley rats (male); human bronchial epithelial cells		24 h (cells); Instillation	• Lung inflammation and injury • Bronchoalveolar lavage	Aged	Madden et al. 2000
Oxidation with O3	Diesel exhaust particles		BALB/c mice (male)	10 h	Instillation	• Gene expression in brain and lung (neurological and immunological biomarkers) • Histology in brain and lung	Mostly null (aged induced higher expression of BDNF)	Win-Shwe et al. 2013
Oxidation with O3	Diesel exhaust		BALB/c mice (male, female (gestational))		5 h/day, 5 days/week for 1 or 3 months (males); 1 month (gestational)	• Novel object recognition • Memory-related gene expression in hippocampus • Oxidative stress • Maternal behavior (gestational) • Maternal-related gene expression in hypothalamus (gestational) • Immunohistochemistry (estrogen and oxytocin receptors)	Aged	Win-Shwe et al. 2014
Oxidation with O3	Diesel exhaust		BALB/c mice		5 h/day, 5 days/week until postnatal day 21	• Adult social behavior • Social behavior-related gene expression in hypothalamus • Inflammatory and oxidative stress marker genes in hypothalamus • Neuronal excitability in hypothalamus	Aged (social novelty discrimination; OTR mRNA)	Win-Shwe et al. 2016
Oxidation with O3	Coal-fired power plant emissions ± SOA		Sprague-Dawley rats (male)		6 h	• Bronchoalveolar lavage • Complete blood count • Histopathology	Aged (+ SOA, brochoalveolar lavage)	Godleski et al. 2011
Oxidation with O3	Coal-fired power plant emissions ± SOA		Sprague-Dawley rats (male)		6 h	• In vivo chemiluminescence of heart and lung • TBARS	Aged (+ SOA, IVCL)	Lemos et al. 2011
Oxidation with O3	Coal-fired power plant emissions ± SOA		Sprague-Dawley rats (male)		6 h	• Breathing pattern	Aged	Diaz et al. 2011
UV-irradiation	Coal-pyrolyzed products + NOx		Salmonella typhimurium (TA100, TA98)	5 h		• Mutagenicity	Aged	Hirayama et al. 1983
UV-irradiation	Wood combustion emissions ± NOx		Salmonella typhimurium (TA100, TA98)		1–10 h	• Mutagenicity	Aged	Kleindienst et al. 1986
Sunlight and/or O3	Wood combustion emissions + NO2		Salmonella typhimurium (TA98)	4 h		• Mutagenicity	Aged	Kamens et al. 1984

3.4.1. Photochemical or oxidative transformation of (gasoline fueled) vehicle exhaust

Gasoline-fueled automobile exhaust aged by UV irradiation induced eye irritation in humans. Schuck and Rogers (Schuck and Rogers, 1959) reported increased eye irritation with irradiated automobile exhaust compared to non-irradiated exhaust and that the degree of eye irritation increased with increasing irradiation intensity. This irritant effect was also found to have a 24 h half-life indicating the presence of a stable irritant.

Respiratory effects of fresh versus aged exhaust were examined in the 1960s and early 1970s and then again more recently within the last ten years. A 4 h exposure to aged exhaust in guinea pigs resulted in an increase in total expiratory flow, accompanied by a decrease in respiratory rate and a small increase in tidal volume, while fresh exhaust resulted in little effect (Murphy et al., 1963). In contrast, 20-month exposure to either fresh or aged exhaust did not result in any changes in pulmonary function (Hueter et al., 1966). Some pulmonary effects were also observed in beagles after long-term exposure to aged auto exhaust. Vaughan et al. (Vaughan, Jennelle, and Lewis, 1969) exposed beagles for 18 months to aged or fresh auto exhaust with or without oxides of sulfur (SOx). No effects were observed for pulmonary function or upper airway absorption. It was suggested these null findings were due to nasal scrubbing efficiency in dogs. Lewis et al. (Lewis et al., 1974) examined pulmonary function in the same beagles after 36 and 61 months of exposure to lower levels. After 61 months, a higher diffusing capacity was found in dogs exposed to fresh auto exhaust, while aged exhaust exposure with or without SO_x induced a higher total expiratory resistance.

More recently, Diaz et al. (Diaz et al., 2012) reported no changes in breathing patterns or BALF in Sprague-Dawley rats exposed for 5 h to fresh versus aged diluted car exhaust + Mt. Saint Helens Ash (as seed aerosol; primary and secondary gases removed via a nonselective diffusion denuder). In a separate study, Diaz et al. (Diaz et al., 2013) exposed rats for 5 h per day for 2, 4, or 12 days to traffic tunnel plenum particles that were either fresh or photochemically aged. After 4 days, the fresh, but not aged, particles induced a rapid, shallow breathing pattern suggestive of an irritant response. In addition, exposure to fresh compared to aged particles resulted in increased basophils in BALF and complete blood count (CBC) and decreased monocytes in CBC. No changes were observed for in vivo chemoluminescence (IVCL), a measure of oxidative stress, in the heart or lung after 1 day of exposure. In a similar study, 6 hours of exposure to aged automobile exhaust resulted in increased IVCL in the lung but not the heart (Papapostolou et al., 2011).

Several of the earlier studies investigated the ability of aged exhaust to promote the development of pulmonary disease in dogs and rodents. Following 68 months of exposure, Orthoefer et al. (Orthoefer et al., 1976) reported a pronounced increase in prolyl hydroxylase, a key enzyme in the synthesis of collagen, in the lungs of dogs exposed to aged exhaust compared to fresh emissions, although no changes in tissue collagen concentrations were detected. Pulmonary morphological changes were reported in dogs exposed to exhaust with or without SOx (Hyde et al., 1978). Enlargement of alveolar air spaces, a characteristic feature of emphysema, was reported in dogs exposed to aged exhaust. In contrast, exposure to fresh exhaust resulted in the highest degree of hyperplasia of non-ciliated bronchiolar cells (Hyde et al., 1978).

Murphy et al. (Murphy et al., 1963) injected mice with Alpha-Naphthylthiourea (ANTU) to produce pulmonary edema, in an effort to determine the effect of fresh and aged auto exhaust in a disease model. ANTU was administered by intraperitoneal injection to groups of male mice 30 min before a 6 h exposure to either UV-irradiated and non-irradiated auto exhaust. After exposure, mortality from pulmonary edema and pleural effusion was similar in dogs exposed to either fresh or aged exhaust.

During the irradiated auto exhaust studies in the 1960s, Hueter et al. (Hueter et al., 1966) reported a spontaneous outbreak of pneumonia in guinea pigs after two weeks of exposure to fresh and aged auto exhaust with animals exposed to the irradiated auto exhaust being more susceptible to infection as measured by gross lung pathology, decreased body weight, increased lung weight, and increased mortality compared to those exposed to fresh exhaust.

Hematological and biochemical effects of aged automobile pollutants were mixed. Rats exposed to aged or fresh auto exhaust for 6 h showed little or no alterations in metabolic enzymes in the serum (Murphy et al., 1963). Carboxyhemoglobin levels were greater in rats exposed to aged exhaust than those exposed to freshly emitted auto exhaust, although both atmospheres contained comparable CO concentrations, suggesting the irradiated atmosphere altered the rodents’ capacity to absorb CO (Murphy et al., 1963). Hueter et al. (Hueter et al., 1966) reported increased white blood cell numbers in mice following long-term exposure to aged auto exhaust. No significant differences were observed in blood viscosity, hematocrit, or methemoglobin concentrations measured daily for 4 years in beagles exposed to fresh or aged auto exhaust (Bloch et al., 1973).

There were no reported differences in long-term cardiovascular effects in female beagles exposed to either fresh or aged auto exhaust atmospheres (Bloch et al., 1973; Bloch et al., 1972). Dogs were exposed for up to 5 years to fresh or aged auto exhaust with or without SO_x. Electrocardiogram (ECG) and vectorcardiogram (VCG) measurements were made at rest and during exercise. Venous right ventricle, pulmonary, and systemic arterial pressure were also measured. Animals exposed for 5 years to all exhaust types had a higher incidence than controls of ECG and VCG abnormalities and myocardial infarctions, but there were no differences between fresh and aged exhaust.

Murphy et al. (Murphy et al., 1963) reported decreased spontaneous running activity after acute exposure to aged versus fresh auto exhaust in mice. Hueter et al. reported that after 10 days of exposure, activity levels in mice exposed to fresh exhaust returned to control levels, but this effect took 23 days for mice exposed to aged auto exhaust (Hueter et al., 1966).

Recent studies tested the effect of relatively low concentrations of fresh and aged traffic-derived particles using a photochemical reaction chamber proximal to a traffic tunnel, which had both diesel and gasoline fueled vehicles (Lamoureux et al., 2013). Blood pressure (systolic and diastolic) and heart rate were assessed in rats exposed for 3 weeks. Diastolic blood pressure increased significantly following exposure to fresh compared to aged particles. No significant changes were observed for heart rate for any of the exposure conditions.

Hueter et al. (Hueter et al. 1966) showed that mice exposed to irradiated automobile exhaust had decreased parental fertility and fecundity along with decreased infant survival compared to fresh exhaust.

Two studies on long-term (15–23 months) exposure to fresh and aged automobile exhaust reported that mice exposed to fresh compared to aged exhaust had higher bone lead concentrations (Hueter et al., 1966; Lutmer, Busch, and Miller, 1967) (lead has since been removed from gasoline).

The oxidative and proinflammatory activity of fresh and aged PM_2.5 was assessed by collecting samples from a central Los Angeles site either in the morning (representative of fresh primary traffic emissions) or in the afternoon (photochemically oxidized emissions) (Lovett et al., 2018). This study found a higher inflammatory response from the morning-collected PM_2.5, as assessed by an upregulation of gene expression of the proinflammatory cytokines IL-1β, IL-6, and MCP-1 in BV-2 microglial cells (Lovett et al., 2018).

Aged gas-phase auto exhaust (albeit from a different era) increased mutagenicity in a variant of the Ames test (Kleindienst et al., 1992). TA100 and TA98 strains of Salmonella typhimurium were exposed to aged and fresh gas-phase emissions, as well as extracts from the particulate phase. Exposure to aged gas-phase emissions increased mutagenic activity in both strains compared to exposure to fresh emissions, while particulates did not show mutagenic activity in either strain or exposure condition.

3.4.2. Photochemical or oxidative transformation of diesel emissions

Zielinska et al. (Zielinska et al., 2010) conducted an experiment in the large Europhore (European photoreactor) outdoor stimulation chamber in Valencia, Spain. Diesel exhaust particle samples were collected, extracted, and intratracheally instilled in mice and rats. Oxidized glutathione increased following exposure to aged compared to fresh diesel exhaust particles. In general, the treatment group that contained hydroxyl radicals in the aged exhaust showed the greatest toxicity.

The health effects of diesel exhaust particles (DEP) exposed to O₃ in a cell-free system for up to 48 h were examined by Madden et al. (Madden et al., 2000). DEP incorporated the O₃ (labeled with ¹⁸O) in a linear fashion. PGE₂, a prostaglandin involved in inflammation, was significantly increased in human bronchial epithelial cells exposed to DEP aged for 48 h; however, this effect was not observed with DEP aged for less than 24 h. The pulmonary toxicity of aged compared to fresh DEP was investigated after intratracheal instillation in Sprague-Dawley rats. Aged DEP exposed to 0.1 ppm O₃ was more potent in increasing neutrophilia, lavage total protein, and LDH activity compared to fresh DEP, while exposure to a higher O₃ concentration (1.0 ppm) led to a decreased bioactivity of the particles. In contrast, carbon black particles, low in organic content relative to DEP, did not exhibit an increase in any of the bioactivities examined after exposure to 0.1 ppm O₃.

More recently, Win-Shwe et al. (Win-Shwe et al., 2013) treated DEP with higher concentrations of O₃ and exposed mice to the fresh or aged particles intranasally. There were no statistically significant changes in neurological or immunological biomarkers in the brain and no effects on histology in either the brain or lung. In the lung, expression of BDNF was higher in the group exposed to aged but not to fresh DEP. This group expanded their studies by exposing male BALB/c mice for one or three months to fresh or aged diesel exhaust generated from mixing 0.6 ppm O₃ into the exhaust (Win-Shwe et al., 2014). Aged but not fresh exhaust significantly impaired novel object recognition following both 1- and 3-month exposures. In addition, after both exposure times, aged but not fresh exhaust induced increased expression of the NMDA receptor subunit NR1 in the hippocampus, while subunit NR2 showed decreased expression in the aged group only after 3 months of exposure. There was no difference in NR3 mRNA levels, neither were there differences in plasma levels of 8OHdG, a marker of DNA oxidative stress. Gestational exposure to aged but not fresh diesel exhaust resulted in decreased estrogen receptor-α and oxytocin receptor in the hypothalamus, as well as reduced maternal performance, such as nest building and crouching (Win-Shwe et al., 2014). Using a similar exposure scenario, Win-Shwe et al. further examined whether early life exposure to aged diesel exhaust results in changes in adult social behavior (Win-Shwe et al., 2016). BALB/c mice were exposed to aged or fresh diesel exhaust from gestational day 14 to postnatal day 21. At 13 weeks, mice were analyzed for changes in social behavior using a series of behavioral tests. Early life exposures to aged diesel exhaust impaired social novelty discrimination, but there were no other significant differences in social behavior between aged and fresh exhaust. Furthermore, exposure to aged compared to fresh exhaust resulted in decreased expression of oxytocin receptor, but not ERα, in the hypothalamus. There were no statistically significant differences between the two treatments in hypothalamic glutamate levels or in the expression of COX2 or HO-1, markers of inflammation and oxidative stress.

3.4.3. Photochemical or oxidative transformation of coal and wood smoke

The toxicological effects of primary and aged coal emissions were evaluated as part of the Toxicological Evaluation of Realistic Emissions of Source Aerosols (TERESA) project. The coal emission experiments were conducted at three US coal-fired power plants with different coal types and emission controls. Exposure scenarios included primary stack emissions from three different coal-fired power plants and different aged atmospheres creating the following scenarios: fresh, fresh + aged, fresh + aged + SOA, fresh + neutralized aged + SOA. In each study, male Sprague-Dawley rats were exposed to the test compounds for 6 hours. Godleski et al., (Godleski et al., 2011) reported on the pulmonary cellular responses, CBC, and histopathological findings after exposure to all test atmospheres. Increases in total cells and macrophages in the BALF were observed 24 h after a 6 h exposure to fresh + aged + SOA and fresh + neutralized aged + SOA, but not to fresh or fresh + aged. Neutrophils increased after exposure to the fresh atmosphere only. No statistically significant differences were observed for CBC or histopathology, suggesting that the inflammatory response was mild. To investigate oxidative stress, IVCL was immediately measured in the heart and lungs and tissues were assayed for thiobarbituric reactive substances (TBARS) (Lemos et al., 2011). When data from all the plants were combined, fresh + neutralized aged + SOA and fresh + aged + SOA increased IVCL in the lungs and heart, respectively. No other significant effects on combined data were reported. Continuous respiratory data were obtained during exposures to the test atmospheres using whole body plethysmography chambers (Diaz et al., 2011). Increases in respiratory frequency and decreases in inspiratory time were found with the fresh + aged exposure. Inhalation of fresh and fresh + neutralized aged + SOA resulted in decreased expiratory flow. Other effects were variable. In a multivariate analysis, effects from fresh + neutralized aged + SOA were associated with elemental carbon.

Coal pyrolysis is a process used to produce coal-refined oil. Hirayama et al (Hirayama et al., 1983) investigated the mutagenic effects of irradiated coal-pyrolyzed products. Briefly, coal-pyrolyzed products were collected and subsequently fractionated into 4 different solvents with increasing polarity: aliphatic hydrocarbons, neutral polycyclic aromatic compounds, nitrogen polycyclic aromatic compounds, and hydroxy polycyclic aromatic compounds. Each fraction was then irradiated with NO_X for 5 h. Salmonella typhimurium, strains TA100 and TA98, in the absence and presence of the metabolic enzymes (i.e. S9), were exposed to the fresh material and aged mixtures. In general, the aged mixtures demonstrated greater mutagenicity than the fresh. The aged fraction containing hydroxy polycyclic aromatic compounds was the most potent mutagenic fraction tested in the absence and presence of S9, followed by the aged nitrogen polycyclic aromatic compounds fraction, suggesting that the majority of the mutagenic activity was in the more polar fractions of the extracts.

Kleindienst et al. (Kleindienst et al., 1986) irradiated diluted mixtures of wood combustion emissions with or without additional NO_X. The gas-phase and PM components were tested separately for mutagenic activity by exposing Salmonella typhimurium, strains TA100 and TA98, to the fresh and aged mixtures. Overall, the aged gas-phase products from wood smoke with or without NO_X greatly increased mutagenic activity compared to the primary emissions. In general, there was little difference in the mutagenicity of fresh compared to aged PM mixtures, with the exception of the TA98 strain exposed to the aged mixture with added NO_X, which increased mutagenicity. In a related study, wood combustion emissions were aged with O₃, NO₂, or both in the presence or absence of sunlight (Kamens et al., 1984). O₃ aged emissions with NO₂ resulted in greater mutagenicity in Salmonella TA98 compared to emissions with NO₂ alone. Exposure to sunlight did not change the mutagenicity of the mixtures containing NO₂.

4. Discussion

The Earth’s atmosphere is a complex mixture of particles and gases in constant dynamic composition. The starting materials can originate from anthropogenic or biogenic sources and reactions are continually occurring. Secondary pollutant formation is a well-recognized process that leads to increased levels of O₃, NO₂, organic aerosol, and a multitude of chemicals, some of which are identified as air toxics. Importantly, air pollution is still a major contributor to morbidity and mortality across the world and these effects are growing worse in countries where populations and combustion emissions related to energy and transportation are increasing (Brook et al., 2010; Landrigan et al., 2018). Thus, it is important to understand the reactions and products that most contribute to toxicity. The purpose of this review was to identify and synthesize studies that compared the health effects from fresh, primary air pollutants to pollutants that have been aged, via either photochemical or oxidative transformation processes. We identified a total of 47 papers that met these criteria through investigations of VOCs, SOA, and combustion emissions. A broad range of outcomes related to ocular, respiratory, cardiovascular, immunological, hepatological, hematological, reproductive, mutagenic, carcinogenic, and other effects (i.e. body/organ weights, morbidity/mortality, bone lead levels, and spontaneous activity) were reported in controlled human exposure studies, animal studies, or in vitro assays.

The most frequently studied single VOC was limonene, followed by isoprene. Four animal toxicological studies reported that exposure to aged compared to fresh single VOC resulted in significant effects on respiratory rate, however, three studies that examined airway inflammation reported no differences between the two. Other studies of single VOCs reported that aged products compared to the primary VOC increased toxicity, mutagenicity, carcinogenicity, and ocular irritation. Several of these studies characterized the products generated by aging, although these differed based on study design. Given that the atmosphere is composed of a multitude of VOC mixtures from various anthropogenic and biogenic sources, it is also important to examine the effects of environmentally relevant VOC mixtures. Transformation of VOC mixtures is complex and produces many different reaction products (see Rohr, 2013 for a comprehensive review). The three studies we identified examining cytotoxicity reported that aged VOC mixtures were more cytotoxic, and two studies found that aged VOC mixtures increased the release/production of inflammatory cytokines when compared to the fresh mixtures. However, the two controlled human exposure studies of women that examined oxidized VOC mixtures reported no change in effects between aged and fresh mixtures on neurobehavioral performance, lung function, cortisol levels, or nasal inflammation. Notably, neither of these studies detected any significant effects from exposure to either of the VOC mixtures when compared to filtered air. It is possible that the irritation threshold for these specific mixtures was below the concentrations used in the studies.

SOA are a significant contributor to atmospheric particulate mass. While their formation and atmospheric evolution is complex and involves several unknown precursors, oxidative or photochemical transformation of VOCs accounts for a substantial proportion of SOA. Given that SOA are formed as a product, it is challenging to disentangle the effects of aging. We identified only one study that allowed for a comparison between fresh and aged SOA. While our review criteria limited our study selection to include comparisons between the parent compound(s) and subsequent aged pollutants, several studies have examined the effects of aged SOA compared to filtered air. For example, two in vitro studies reported no cytotoxicity associated with exposure to aged SOA compared to air controls, while the effects on inflammatory cytokines was mixed (Gaschen et al., 2010; Jang, Ghio, and Cao, 2006). A controlled human exposure study suggested that aged SOA increases oxidative stress (Kipen et al., 2011). Other studies have examined the pulmonary and cardiovascular toxicity of SOA comparing the effects of neutral versus acidic conditions (Lund et al., 2013; McDonald et al., 2010; McDonald et al., 2012).

Combustion emissions are another class of complex mixtures containing particles, semi-volatile matter, and gases. The particulate matter fraction can contain thousands of known organic chemicals such as alkanes, aromatic hydrocarbons, and carboxylic acids, as well as unidentified compounds. Combustion emissions have been linked to numerous adverse health effects including respiratory and cardiovascular disease, reproductive effects, and cancer, however, it is unclear if atmospheric transformation reactions influence the potency of these emissions. We identified a total of 21 studies that examined the effects of fresh versus aged automobile exhaust, 10 of which were conducted between 1959 and 1978. Notably, the studies conducted in the 1960–70s were with vehicles that employed less stringent emission controls and were fueled by leaded gasoline. Although these studies contributed to the establishment of the early air quality standards, their relevance to contemporary urban air pollution is less clear. Of the recent studies conducted since 2010, three found effects from fresh exhaust, including increased blood pressure, inflammation, and negative effects on breathing patterns, four found effects from aged exhaust, including and increased oxidative stress in the lungs and memory impairment, and one reported no difference between fresh and aged exhaust. Interestingly, the earlier studies also reported mixed results across various biological endpoints and models, illustrating the difficulty in understanding and quantifying the health impact of dynamic air pollution mixtures.

In summary, for single and mixed VOCs, nearly all studies reported that the aging process induced significant effects and/or increased the severity of effects when compared to exposure to the untransformed primary pollutant. In contrast, for automobile (gasoline) emissions, four studies on a diversity of endpoints, including respiratory effects, blood pressure, and bone lead in animal studies, as well as inflammatory response and oxidative activity in microglial cells, identified fresh emissions as more potent than their aged counterpart. The difference between single/mixed VOCs and automobile emissions may relate to differences in the complexity of the mixture as well as reaction dynamics. For VOCs, the photo-oxidation reactions are relatively stable and can result in the production of a variety of products with known toxicity, including, for example, ozone, formaldehye, and acrolein. In contrast, for automobile emissions, competing gases, vapors, and particles can potentially antagonize and quench the reactions resulting in variable mixtures with potentially reduced toxicity. For example, two of the studies that reported effects from exposure to primary automobile exhaust used a paradigm consisting of either primary or aged traffic-derived particles. For the aged particle group, ozone was continuously added to the chamber to titrate nitric oxide, in addition to the ozone that was formed during irradiation (Papapostolou et al., 2013). Ozone may react with organic compounds on the surface of particles, potentially changing both the composition and reactivity of the aged particles compared to the unaltered primary particles (Lamoureux et al., 2013). Furthermore, primary traffic-derived particles include a larger mass fraction of non-polar water-insoluble organic compounds compared to oxidized aged particles which form more polar compounds (Lovett et al., 2018; Papapostolou et al., 2013). The non-polar components, such as polycyclic aromatic hydrocarbons, may more readily permeate the hydrophobic lipid-bilayer of cell membranes, resulting in greater overall toxicity (Lovett et al., 2018). In addition, Lovett et al. found a greater concentration of metals in the fresh, primary traffic-derived particles, which may further contribute to the observed response of greater oxidative stress from exposure to fresh compared to aged particles.

4.1. Future research needs

While most of the studies identified through the literature search and reviewed in this article provide good evidence that formation of aged atmospheres may result in greater health impacts, there is still appreciable uncertainty as to which sources and aging conditions lead to the most adverse health effects, and at what concentration and duration of exposure. Formation of O₃ through the photochemical reaction of nitrogen oxides and VOCs is the best example of a secondary air pollutant with well documented health impacts. In addition to this criteria pollutant, the potential health impact of other chemicals whose activity may be under-appreciated within the complex mix of secondary pollutants requires further investigation. Experiments to date have identified some conditions and products (within mixtures) that are more toxic than the starting emissions. The variation in experimental design, aging system employed, concentration and type of starting agent, and toxicity endpoints, however, make comparisons between different studies exceedingly difficult. Despite this caveat, the results confirm that photochemical aging of anthropogenic and biogenic emissions can lead to a wide array of adverse health impacts. In the case of O₃ and PM, these pollutants are well regulated under the Clean Air Act, however, the long list of complex organic compounds that may also impart biological activity is generally less well studied.

There is little doubt that some of the variation reported between studies arises from differences in the starting mixture, light source, and wavelength (be it natural or synthetic), chamber size and materials, ambient conditions, duration of aging time and a host of other controlled or uncontrolled conditions. To the extent possible, a harmonization of the materials, methods, and chemical analyses utilized in these types of investigations would allow for better comparison across studies. As a separate though related issue, the studies described here spanned multiple exposure regimes ranging from a matter of minutes to years (in the case of some of the earlier vehicle exhaust studies). While longer durations of exposure are regarded as being worse for chronic disease outcomes, utilization of relevant biomarkers, integration of acute and chronic responses and a more complete understanding of reversable and/or adaptive responses are important factors in understanding the health hazards of these exposures. There is also a pressing need to identify assays that are the most sensitive and predictive of potential health endpoints caused by inhalation exposure to complex mixtures. In general, in vitro exposures were more apt to show effects in, for example, mediator release or gene expression changes, compared to whole animal or human studies which focused on alterations in pathophysiology. This is not surprising given that the delivered dose of pollutant at the target site was likely greater than during inhalation exposures over the entire surface are of the respiratory tract. Furthermore, changes at the molecular or biochemical level may be more sensitive indicators of effect than “clinical” alterations in tissue morphology or function. While there is a reasonably rich database for the production or expression of some common biomarkers such as lactate dehydrogenase (LDH), hemoxygenase, and IL-8, a more holistic assessment of these and other measurements under various exposure scenarios is required to confirm their utility and application within the risk assessment paradigm. Induction of oxidative stress as a generic initiating event is for example a promising approach as this is thought to be a principal mechanism (or key initiating event) by which air pollutants cause injury and adverse responses (Buckley and Farraj, 2015).

With this in mind, a more systematic methodology utilizing high throughput screening, whole genome assessment, and computational toxicology will more ably address under what conditions and chemistries photochemically-transformed pollutants elicit greater health effects than primary emissions. This process has been employed for thousands of chemicals identified in the Tox 21 program (Sand et al., 2017), as well as for emerging contaminants such as inhaled nanoparticles (Nymark et al., 2018), and shows great promise for the screening of these and other materials. Significant progress has been made in developing “fit for purpose” testing protocols and identifying benchmark doses which are used to derive points of departure (POD) for a variety of molecular cellular and physiological responses (Zhang et al., 2018). These tools are also applied to metabolic, toxicokinetc and dynamic mathematical models that provide extrapolation across exposure durations and species (Zhang et al., 2018). Given time, resources and establishment of reproducible methods for exposing cells to particles, gases and volatile componenets at the air liquid interphase, this approach could be adapted for many of the common air toxics as well as relevant mixtures. Finally, the application could be extended by developing quantitative relationships between stochastic chemical models that predict half-life and degradation of emitted chemicals in the atmosphere, the formation of secondary and tertiary intermediates and their subsequent biological activity.

Although these future research needs are ambitious, and will take years more to come to fruition, the understanding of the basic chemistry and health effects associated with air pollution exposure is well established, and while the myriad scenarios cannot be tested through an iterative experimental process in animals, careful planning and laying of groundwork through selected test cases could result in a reliable framework for computational risk assessment. In order to accomplish this, however, the in vitro high throughput screening tools need to be further validated against relevant physiological effects such as changes in cardiopulmonary function, alterations in immune activity, carcinogenesis, and a host of other effects. A natural extension of this process would also be to expand potential effects to specific disease phenotypes such as hypertension, diabetes, and asthma since those individuals are thought to be more “at risk” to the effects of air pollution. Finally, the entire platform has not been adequately explored with regard to mixtures which may result in additive, synergistic, or antagonistic effects involving multiple pathways (Dominici et al., 2010). This is particularly important since the pollutant landscape is forever changing due to emerging technologies and alterations in climate, which together with other factors, such as increased emissions from wildland fires, can strongly influence the chemistry and resultant health impacts in local and regional airsheds.