Immune consequences of exposure to β-pinene oxidation aerosols: adult versus gestational murine models

Abstract

Background

While studies demonstrating the adverse effects of air pollution on human health are accumulating, studies on secondary organic aerosol (SOA) are scarce. However, SOA accounts for a significant portion of airborne particulate matter. In particular, pinene biogenic SOA contributes predominantly to SOA loading in the outdoor atmosphere of natural and urban areas and are also emitted indoors because of the presence of terpenes in numerous consumer products. Our aim was to study the immune consequences of acute exposure to β-pinene ozonolysis gaseous and SOA products in mice. This reaction was generated in an atmospheric simulation chamber, and the mice were exposed to the particulate and gaseous products, to the gaseous products only, or to synthetic air 2 h per day for 3 days in real time in a whole-body inhalation chamber. Exposures were performed in adulthood or in utero. Since some adverse effects only occur in individuals weakened by existing immune activation, such as low-grade inflammation, the immune response was measured in the steady state or in a state of moderate systemic inflammation induced by lipopolysaccharide administration.

Results

Exposure of healthy adult mice caused minor immunosuppression in the lungs. However, in adult mice weakened by moderate systemic inflammation, the same exposure conditions revealed that mice exposed to the β-pinene ozonolysis particulate and gaseous products presented deficient pulmonary and systemic immune responses, including excessive recruitment of B lymphocytes, CD4⁺ T lymphocytes, CD11b⁺ dendritic cells, inflammatory monocytes and neutrophils in the lungs and defective recruitment of regulatory T cells in the spleen. In offspring exposed to β-pinene ozonolysis products in utero, the LPS-induced upregulation of Ccl2, Cxcl10 and Icam1 mRNA levels in the lungs and the activation of dendritic cells in the spleen were excessive in female mice. The male offspring developed a normal response to moderate systemic inflammation, except for impaired activation of CD4⁺ T cells and increased activation of CD103⁺ dendritic cells in the spleen.

Conclusion

In mice, pulmonary and systemic immune reactions in response to moderate systemic inflammation are dysregulated by exposure to common secondary oxidation products, highlighting interest in the role of these neglected atmospheric compounds in immune disease development and susceptibility to infections.

Background

The World Health Organization estimates that ambient air pollution, in both cities and rural areas, caused 4.2 million premature deaths worldwide per year in 2019. Some 37% of outdoor air pollution-related premature deaths are due to ischaemic heart disease and stroke, 18% and 23% of deaths are due to chronic obstructive pulmonary disease and acute lower respiratory infections, respectively, and 11% of deaths are due to cancer within the respiratory tract [1]. Moreover, household air pollution was responsible for an estimated 3.2 million deaths per year in 2020, including over 237 000 deaths of children under the age of 5. The combined effects of ambient air pollution and household air pollution are associated with 6.7 million premature deaths annually [2]. Indeed, the adverse effects of air pollution have been demonstrated by many scientific studies, particularly meta-analyses. A 10 µg.m^-³ increase in short-term PM_2.5 exposure was significantly associated with 0.70%, 0.86%, 0.38% and 0.96% increases in cardiovascular mortality, respiratory mortality, cardiovascular morbidity, and respiratory morbidity, respectively. The diseases with significant associations included stroke, ischaemic heart disease, heart failure, arrhythmia, chronic obstructive pulmonary disease, pneumonia and allergic rhinitis. The pooled estimates per 10 µg.m^-³ increase in long-term PM_2.5 exposure were 15.1%, 11.9% and 21.0% increases in cardiovascular, stroke and lung cancer mortality and 17.4%, 11.0% and 4.9% increases in cardiovascular, hypertension and lung cancer incidence, respectively. Adverse changes in blood pressure, heart rate variability, systemic inflammation, blood lipids, lung function and airway inflammation were observed in response to either short-term or long-term PM_2.5 exposure or both [3]. A meta-analysis revealed a significant association between PM_2.5 exposure and an increased risk of acute exacerbation of idiopathic pulmonary fibrosis [4]. Furthermore, exposure to PM_2.5 is associated with an increased incidence of central nervous system diseases in humans, such as Alzheimer’s disease, Parkinson’s disease, dementia, and high levels of PM_2.5 exposure during the first three years after birth are associated with increased global autism spectrum disorder in children [5]. Maternal exposure to PM_2.5 increases the risk of gestational diabetes mellitus and various health outcomes in newborns, including birth defects, preterm birth and low birthweight [6, 7]. This highlights the direct and indirect impacts of PM_2.5 on global health.

PM_2.5 includes approximately 50 inorganic species and potentially hundreds of thousands of organic compounds [8]. Organic aerosols are highly variable compounds, ranging from simple hydrocarbons to highly oxidized compounds [9]. Directly emitted hydrocarbon carbonaceous particles are referred to as primary organic aerosols, and organic aerosols formed via the atmospheric oxidation of volatile and semivolatile organic compounds are referred to as secondary organic aerosol (SOA). The main atmospheric oxidants allowing oxidation reactions are ozone (O₃), nitrate (NO₃) and hydroxyl radicals (OH). SOA is formed when gas-phase volatile or semivolatile organic compounds react with atmospheric oxidants to form products that condense into the aerosol phase, where they can undergo further reactions. SOA is categorized into two parts on the basis of its sources: biogenic SOA originates from biogenic precursors (e.g., terrestrial vegetation, grassland, peatlands and forest), and anthropogenic SOA is formed from anthropogenic sources (e.g., biomass burning, coal combustion, transportation, solvent utilization and industry).

SOA accounts for a significant fraction of PM_2.5, which ranges from 65 to 95% between urban and remote regions [10]. Notably, biogenic SOA, particularly pinene biogenic SOA, is ubiquitous. Monoterpenes, such as α-/β-pinene, are mostly emitted by coniferous plants and most flowers and fruits. α-/β- pinene SOA is the predominant SOA formed in forest aerosols [11, 12] and is also the dominant SOA present in urban areas throughout the world. Lanzafame et al. measured SOA tracers for one year in Paris (France) [13]. They reported that pinene biogenic SOA represented all or the majority of secondary organic carbon from March to December 2015, which were the warmest months. This finding was consistent with previous studies showing that during warm months, high solar fluxes and oxidant concentrations (i.e., O₃ and OH) could facilitate high biogenic secondary organic carbon formation [14, 15]. Moreover, during the coldest months, from October to the end of December 2015, pinene biogenic secondary organic carbon still accounted for a significant fraction of the total secondary organic carbon mass. The dominance of pinene biogenic SOA has also been reported in 2 recent studies performed in China. Hong et al. analysed PM_2.5-bound SOA tracers in aerosol samples collected during summer and winter in southeastern China [16]. They measured isoprene, α-/β-pinene, β-caryophyllene, and toluene SOA and demonstrated the predominance of α-/β- pinene SOA: they accounted for 70.1% of the SOA in winter and 45.8% of the SOA in summer. Similarly, Yang et al. determined 14 typical PM_2.5-bound SOA tracers in three cities in the Yangtze River Delta region in the winters of 2014 and the summers of 2015 [17]. Among all the SOA tracers, α-/β-pinene SOA tracers contributed 55.9%, followed by isoprene SOA tracers (33.7%), anthropogenic benzene SOA tracers (6.4%), and β-caryophyllene SOA tracers (4.0%). Therefore, pinene SOA contributes predominantly to aerosol loading in the outdoor atmosphere of urban areas.

Moreover, pinene SOA is also an emerging concern in indoor air. In addition to the well-established role of the ozonolysis of monoterpenes in the production of atmospheric SOA [18], ozonolysis-initiated indoor SOA formation could also be a dominant indoor particle source [19]. First, the indoor environment is typically insulated from outdoor O₃ because of O₃ losses to the surfaces of ventilation systems, building envelope components, interior walls and furnishings, human skin, and clothing [20]. Second, numerous consumer products can emit significant quantities of terpenes, and the resulting pinene SOA production has been reported. By deliberately introducing known concentrations of O₃ and either a selected terpene or a terpene-based cleaner, Weschler et al. demonstrated that reactions between O₃ and various terpenes in indoor environments can significantly increase the number and mass concentration of submicronic particles [21]. Under conditions representative of typical indoor settings (without supplemental O₃), average increases in mass concentrations in the range of 2.5–5.5 µg.m^− 3 were measured. Sarwar et al. demonstrated that homogeneous reactions between O₃ and terpenes from various consumer products, e.g., a lime-scented liquid air freshener, a pinescented solid air freshener, a lemon-scented general purpose cleaner, a wood floor cleaner, and a perfume, can lead to increases in fine particle mass concentrations [22]. These authors reported that fine particle formation/growth can occur following the application of such products in indoor environments. Singer et al. investigated the formation of SOA resulting from household product use [23]. An orange oil-based degreaser and a pine oil-based general-purpose cleaner were used for surface cleaning applications. A plug-in scented-oil air freshener was used for several days. Cleaning products were applied realistically with quantities scaled to simulate residential use rates. This use of terpenoid-containing cleaning products or air fresheners combined with indoor O₃ produced substantial levels of secondary air pollutants, leading to substantial fine particle concentrations (more than 100 µg.m^− 3) in some experiments. Nematollahi et al. measured volatile organic compounds emitted from a range of fragranced baby products, including baby hair shampoos, body washes, lotions, creams, ointments, oils, hair sprays, and fragrance [24]. α- and β-pinene are among the most common volatile organic compounds emitted from baby products. Taken together, these studies provide evidence that pinene SOA is also emitted in polluted indoor air.

Therefore, there is accumulating evidence that oxidation reactions of β-pinene occur in both outdoor and indoor air, but the immunological consequences of inhaling the byproducts of these reactions are unknown. Therefore, the aim of this study was to assess the immune changes induced by air pollutants formed by ozonolysis of β-pinene. To avoid the bias of physicochemical transformation of SOA and integrate all the byproducts, we developed a unique setup of real-time generation coupled with direct inhalation of β-pinene ozonolysis products by mice. Our hypothesis was that acute exposure to b-pinene ozonolysis products could induce immune dysregulations specific to the individual’s susceptibility relative to the exposure window or prior activation of the immune system. Mice exposed acutely to β-pinene ozonolysis products were subjected to either steady-state or systemic inflammatory conditions via low-dose i.p. lipopolysaccharide (LPS) administration, in order to decipher the effects both in healthy individuals and those with existing immune activation, such as patients with immune-mediated diseases [25, 26]. The effects of in utero acute exposure to β-pinene ozonolysis products were also assessed in male and female offspring. The pulmonary and systemic responses were analysed.

Methods

β-Pinene ozonolysis parameters

A homemade cubic Atmospheric Simulation Chamber (ASC) in polymethyl methacrylate and Teflon with a volume of 200 L was used. The reaction between β-pinene and O₃ led to the formation of SOA within the ASC. β-Pinene (Thermo Scientific, 99%) at a 1440 parts per billion in volume (ppbv) concentration was vaporized under a flow of synthetic air (Messer, 5.0; hydrocarbon free, H₂O < 5 parts per million in volume (ppmv), room temperature) via a 10 µl microsyringe (Hamilton). O₃ at a 1000 ppbv concentration was introduced into the ASC from a corona discharge generator (C-Lasky-DTI; Air Tree Europe GmbH), which transformed dioxygen contained in synthetic air into O₃. During the experimental development phase, the SOA formed in the ASC were monitored by scanning mobility particle size (SMPS, Model 3081 Long DMA, TSI Inc., CPC Model 3775, TSI) [27]. The aerosol mass concentration was calculated assuming an aerosol mass density of 1.0 g.cm^− 3. The particle number concentration and size distribution were measured between 20 and 680 nm using a 50 s scan time and a time delay of 16 s.

Mouse exposure protocol

Female C57BL/6 mice (aged 7 weeks) and pregnant C57BL/6 mice (aged 9 weeks) were purchased from Janvier Labs (Le Genest Saint Isle, France) and housed under conventional conditions. The animal protocol was approved by the Ethical Committee on Animal Experimentation (APAFIS #35551-2022020814293415) and was in agreement with the European directive 2010/63/EU for the protection of animals used for scientific purposes.

The mice were exposed for 3 consecutive days. Every day, the ASC was first cleaned for 30 min with synthetic air. β-Pinene (in the liquid state) was then injected into the ASC through a septum and vaporized for 15 min. O₃ was then introduced to start the reaction. After 40 min, the maximum SOA concentration was reached, and two pumps working at 1 mL/min were connected at the exit of the two exposure chambers for 2 h (InExpose whole-body inhalation chamber, Emka Technologies, Paris, France) to expose the mice to the reaction mixture with or without the SOA (Fig. 1A). One group of mice was exposed to the whole reaction mixture comprising both particulate and gaseous β-pinene oxidation products. One group of mice was exposed to the same mixture filtered through an aerosol filter (TSI HEPA capsule filter) and therefore exposed to the gaseous β-pinene oxidation products only. The third group was exposed to synthetic air only. After each exposure, the ASC was disconnected from the exposure chamber and flushed with synthetic air for 30 min. LPS (250 µg/kg pc, (Sigma‒Aldrich, Merck, Darmstadt, Germany)) was administered intraperitoneally to trigger systemic inflammation 16 h before sacrifice. The mice were euthanized the morning following the third exposure day. Bronchoalveolar lavage fluid (BAL), lung, spleen and liver samples were collected, kept on ice and processed for flow cytometry analysis, or immediately frozen and stored at -80 °C.

Measurement of lung function

Measurements of respiratory parameters were assessed on anaesthetized animals. A deep general anaesthesia was obtained by intraperitoneal injection of 150 µl of a solution containing 0.2–0.3 mg xylazine base (RompunR, Bayer Healthcare) and 1-1.5 mg ketamine hydrochloride (ImalgeneR, Mérial, Lyon). Then, the anaesthetized mice were tracheotomized and connected via an 18G cannula to a flexiVent FX system operated by flexiWare software v7.7 (SCIREQ Inc., Montreal, QC, Canada). After connecting the mouse to the system, hyperventilation was applied to prevent mouse spontaneous breathing. No mouse paralysis was needed. The animals were ventilated at a respiratory rate of 150 breaths/min, and multiple mechanical properties of the subjects’ respiratory system were assessed at baseline, i.e., before the construction of a full-range pressure‒volume (PV) curve [28]. The PV loop assesses the distensibility of the respiratory system at rest over the entire inspiratory capacity. The deflation arm of this curve is fitted with the exponential function described by Salazar and Knowles [29]. Static compliance (Cst) and the parameters A (estimate of inspiratory capacity) and K (shape constant) can be extracted from the Salazar–Knowles equation. Static compliance (Cst) reflects the intrinsic elastic properties of the respiratory system (i.e., the lung + chest wall). After acquisition of all respiratory parameters, the mouse was disconnected from the system, spontaneous breathing was observed before euthanasia by cervical dislocation.

BAL fluid and cell collection

The mice that were assessed for lung function were immediately exsanguinated, and a BAL was performed with 500 µl of ice-cold PBS for supernatant analysis and 1 additional ml for cytological analysis. After centrifugation at 400× g for 6 min at 4 °C, the supernatant of the 500 µl lavage mixture was collected and frozen for further analysis. Pelleted cells from the total 1500 ml lavage were suspended for total leukocyte number determination.

Cytokine/chemokine profiling

A mouse XL Cytokine Luminex Performance 45-plex Fixed Panel (Thermo Fisher Scientific, Illkirch, France) was used. The concentrations of 45 cytokines/chemokines in the BAL fluid and liver lysate supernatants were analysed simultaneously following the manufacturer’s recommendations in a Luminex 200 Instrument system (Thermo Fisher Scientific, Illkirch, France). Data were processed using Bio-Plex Manager software (Bio-Rad, Marnes-la-Coquette, France). For liver, cytokines/chemokines quantification was expressed according to the total quantity of proteins determined using the DC™ protein assay kit (Bio-Rad, Marnes-la-Coquette, France). A calibration kit was run to standardize fluorescence signal before each experiment and a validation kit ensuring performance of fluidics and optics systems was run monthly.

Flow cytometry

The left lobe of the lung was mashed with a sterile blade and then digested with collagenase (Collagenase Type VI 17104–019 Gibco by Life Technologies, Carlsbad, California, United States) at 37 °C. After 15 min of digestion, the lungs were homogenized with an 18G needle and further digested for 15 min. After centrifugation at 1700 rpm for 6 min at 4 °C, the pellets were resuspended in 30% Percoll solution (Percoll TM GE Healthcare 17–0891-01, Chicago, IL, United States) and centrifuged at 2000 rpm for 15 min. Total spleen cells were also isolated from crushed spleens and centrifuged at 1700 rpm for 6 min at 4 °C. The lung and spleen pellets were resuspended in red blood cell (RBC) lysis buffer for 5 min at room temperature to remove erythrocytes. The RBC lysis reaction was stopped with 2% FBS in PBS (Gibco by Life Technologies, Carlsbad, California, United States). After centrifugation at 1700 rpm for 6 min at 4 °C, pulmonary and spleen cells were resuspended in 2% FBS in PBS, enumerated and used for flow cytometry.

BAL, lung and spleen total cells were incubated with the appropriate panel of antibodies for 30 min in PBS supplemented with 2% FCS. Conjugated antibodies were used against mouse CD5 (ref130–102–574, FITC-conjugated), PBS57-loaded CD1 d Tetramer (NIH facility, PE-conjugated), NK1.1 (ref 130–109–963, PerCp-Cy5.5–conjugated), CD4 (ref 130–102–411, PE-Cy7-conjugated), CD25 (ref 130–102–550, APC-conjugated), CD69 (ref 561–238, Alexa700-conjugated), TCRγδ (ref 130–104–016, APC-Vio770 conjugated), TCR-β (ref 130–104–815, V450-conjugated), CD8 (ref 130–109–252, V500-conjugated), CD45 (ref BLE103140, BV605-conjugated), I-Ab (ref 130–102–168, FITC-conjugated), F4/80 (ref 130–102–422, PE conjugated), CD103 (ref 563–637, PerCP-Cy5.5-conjugated), CD11c (ref 558–079, PE Cy7-conjugated), CD86 (ref 560–581, Alexa-700 conjugated), Ly6G (ref 560–600, APC-H7 conjugated), CD11b (ref 560–455, V45O conjugated), CD45 (ref 130–402–512, V500 conjugated), Ly6C (ref BLE128036, BV605-conjugated) (BD Biosciences, Franklin Lakes, United States; Biolegend, San Diego, United States and Myltenyi Biotech, Paris, France) and CCR2 (ref FAB 5538 A, R&D systems, APC conjugated). Data were acquired on a LSR Fortessa (BD Biosciences, Franklin Lakes, United States) and analysed with FlowJo™ software v7.6.5 (Stanford, CA, USA). The gating strategy has been previously described [30, 31].

Gene expression analysis

Total lung and liver mRNAs were extracted via a Nucleospin RNA kit (Macherey-Nagel, Hoerdt, France). Reverse transcription‒polymerase chain reaction (RT‒PCR) was performed via a high-capacity cDNA reverse transcription kit and SYBR Green PCR master mix on a StepOne™ Real-Time PCR System (Thermo Fisher Scientific, Illkirch, France). The primer sequences were designed via Primer Express 3 and are available upon request. Melting curve analyses were performed for each sample and gene. The relative expression of each target gene was normalized to the relative expression of the Polr2a housekeeping gene. The quantification of target gene expression was based on the comparative cycle threshold (Ct) value. Fold changes in the expression of target genes were analysed via the 2 − ΔΔCt method.

Statistics

The results are expressed as the mean ± standard error of the mean. The statistical significance of differences between experimental groups was calculated via the Mann–Whitney nonparametric U test (GraphPad Prism software v8, USA). Statistical significance was defined as p < 0.05. For all experiments, *p < 0.05, **p < 0.01, and ***p < 0.005.

Results

Development and characterization of a mouse model of real-time exposure to β-pinene oxidation products

An exposure protocol was developed allowing the exposure of mice in real time either to synthetic air or gaseous β-pinene oxidation products (blue frame) or to the SOA and gaseous products generated by β-pinene oxidation (pink frame) (Fig. 1A). The objectives were as follows: (1) No longer have residual O₃ at the end of the reaction to avoid direct deleterious effects of O₃. (2) to expose the mice to an average of 500 µg. m^-3 of SOA. After the development of the experimental conditions, three further experiments were performed to control the repeatability of the protocol. Pumping at 2 L.min^-1 for 2 h was performed, and the evolution of the SOA mass concentrations during these simulations was as follows. The maximum SOA concentrations formed were 611, 507, and 716 µg.m^-3. The times to reach the maximum SOA concentrations were 41, 43, and 40 min, respectively. The SOA concentrations after 2 h of pumping were 300, 294, and 374 µg, respectively. m^-3 (Fig. 1B). Therefore, the mice were exposed for 2 h to mean SOA concentrations ranging from 611.3 ± 69.8 at the start of exposure to 322.7 ± 34.2 at the end of exposure. The variability between experiments (both before and after the mouse experiments) was about 11%, which is an expected and acceptable rate, taking into account in particular the accuracy of SPMS measurements. A representative assessment of the aerosol number concentration and size distribution is presented in Fig. 1C. Moreover, according to the previous studies performed on the composition of b-pinene ozonolysis products, the concentrations of the main gaseous products expected in the atmospheric simulation chamber were calculated from their formation yield and the initial b-pinene and ozone concentrations, and presented in Table 1 [32,33,34].

Fig. 1

Setup developed to expose mice to β-pinene oxidation products in real time. (A) Experimental scheme of generation and exposure to β-pinene oxidation products overall (pink area) or without SOA (blue area). (B) SOA mass concentrations measured via SMPS analyses before and after the pumping of the atmospheric simulation chamber for 2 h at 2 L.min^− 1. (C) Representative aerosol concentration and size distribution measured by the SPMS

Table 1 Main gaseous products formed from the β-pinene ozonolysis in the atmospheric simulation chamber

Pulmonary response to acute direct exposure in adulthood to β-pinene oxidation products

First, we aimed to explore the pulmonary response to acute exposure to β-pinene oxidation products. The mice were exposed to both the particulate and gaseous products of β-pinene oxidation (SOA + GP group), only the gaseous products of β-pinene oxidation (GP group), or synthetic air (air group) 2 h per day for 3 consecutive days (Fig. 2A). Lung function was assessed via invasive measurements of pulmonary parameters. Exposure to SOA and GPs or GPs did not significantly affect any of the investigated parameters, suggesting that acute exposure to SOA and GPs or GPs had no effect on lung function (Fig. 2B and data not shown). In BAL fluid, the protein levels of 45 cytokines and chemokines were measured (Fig. 2C). Compared with air exposure, acute exposure to SOA and GPs increased IL9, CXCL5 (also called LIX) and TIMP1 secretion and decreased LIF secretion (Air vs. SOA + GP: IL9, p = 0.01; CXCL5, p = 0.01; TIMP1, p = 0.002; LIF, p = 0.02). The cell count in the lung did not vary (Fig. 2D). The frequencies of B cells, T cells, dendritic cells, alveolar macrophages, interstitial macrophages and neutrophils did not significantly differ between the GP and SOA + GP groups (Fig. 2E). The abundance of inflammatory monocytes and CD8⁺ T cells was greater in the mice exposed to GPs than in the mice exposed to air (Air vs. GP: p = 0.04 and p = 0.004, respectively; Fig. 2F-G). Moreover, the frequencies of CD4⁺ T cells and CD103⁺ dendritic cells were lower in the mice exposed to GPs (Air vs. GP: CD4⁺ T cells, p = 0.002; CD103⁺ dendritic cells, p = 0.04) and to SOA and GPs (Air vs. SOA + GP: CD4⁺ T cells, p = 0.05; CD103⁺ dendritic cells, p = 0.002) than in the air-exposed mice (Fig. 2G-H). The mRNA levels of Il6, Il27, and Ccl19 were lower in the lungs of the SOA + GP group than in those of the Air group (Air vs. SOA + GP: p = 0.004, 0.04, and 0.002, respectively; Fig. 2I). Taken together, these results show that acute exposure to β-pinene oxidation products moderately affects the number of immune cells as well as their function, as reflected by the significant dysregulation of some inflammatory cytokines at the transcriptional level in the lung and at the protein level in the BAL fluid.

Fig. 2

Pulmonary response to acute direct exposure in adulthood to β-pinene oxidation products. (A) Experimental design. n = 10/group. (B) Inspiratory capacity function. (C) Mouse XL cytokine panel assay in BAL fluid. (D) Total cell counts in the lungs. (E-H) Immune cell frequency in the lungs. Parent population was CD45⁺CD3⁺TCRb⁺ cells for CD4⁺, CD8⁺, Reg T cells and CD45⁺CD5⁺NK1.1⁺ cells for NKT like T cells. (I) mRNA expression levels of cytokines in the lungs. A representative of two independent experiments is shown. The data represent the means ± SEMs. * p < 0.05, ** p < 0.01

Pulmonary and systemic responses to moderate systemic inflammation after acute direct exposure in adulthood to β-pinene oxidation products

We next sought to assess whether adverse effects induced by β-pinene oxidation products could be revealed in susceptible hosts, which presented mild systemic inflammation. Moderate systemic inflammation was induced by i.p. administration of LPS (Fig. 3A). The immune profile of the lung was first analysed. In the Air + LPS group, ICAM1 production in BAL fluid was lower than that in the Air group (p = 0.004), and ICAM1 production was significantly greater in the SOA + GP + LPS group than in the Air + LPS group (p = 0.04, Fig. 3B). The increased IL1β and CCL3 secretion induced by LPS administration (Air + PBS vs. Air + LPS: p = 0.01 and p = 0.002, respectively) was exacerbated by SOA and GP exposure (Air + LPS vs. SOA + GP + LPS: p = 0.03 for both). In the lung compartment, the frequencies of B cells, T cells, dendritic cells, and inflammatory monocytes were significantly lower in the Air + LPS group than in the Air + PBS group (B cells, p = 0.002; T cells, p = 0.002; dendritic cells, p = 0.009; inflammatory monocytes, p = 0.03) and significantly greater in the SOA + GP + LPS group than in the Air + LPS group (B cells, p = 0.002; T cells, p = 0.04; dendritic cells, p = 0.01; inflammatory monocytes, p = 0.004; Fig. 3C-D). Moreover, under moderate systemic inflammation conditions, the abundance of neutrophils was significantly greater in the SOA + GP + LPS group than in the Air + LPS group (p = 0.01). T-cell populations were further explored (Fig. 3E). The CD4⁺ T-cell frequency was significantly inhibited by LPS administration (Air + PBS vs. Air + LPS: p = 0.002), and this inhibition was significantly reversed in the lungs of the SOA + GP + LPS group (Air + LPS vs. SOA + GP + LPS: p = 0.023). The NKT-like cell frequency was significantly increased by LPS administration (Air + PBS vs. Air + LPS: p = 0.004), and this recruitment was significantly inhibited in the lungs of the SOA + GP + LPS group (Air + LPS vs. SOA + GP + LPS: p = 0.02). When the 2 main populations of dendritic cells were analysed, a significantly increased abundance of CD11b⁺ dendritic cells was observed in response to SOA and GPs in mice with moderate systemic inflammation (Air + LPS vs. SOA + GP + LPS: p = 0.009, Fig. 3F). Finally, qPCR quantification of inflammatory cytokines and chemokines in the lungs revealed that the levels of CCl2/Mcp1, Ccl3/Mip1a, Ccl19/Mip3b, Cxcl1/Kc and Icam1 mRNAs were increased by LPS administration (Air + PBS vs. Air + LPS: CCl2/Mcp1: p = 0.002; Ccl3/Mip1a: p = 0.002; Ccl19/Mip3b: p = 0.002; Cxcl1/Kc: p = 0.002; Icam1: p = 0.03) but decreased by SOA and GP exposure (Air + LPS vs. SOA + GP + LPS: CCl2/Mcp1: p = 0.009; Ccl3/Mip1a: p = 0.009; Ccl19/Mip3b: p = 0.002; Cxcl1/Kc: p = 0.01; Icam1: p = 0.009; Fig. 3G). The expression of Cxcl10 was increased by LPS administration (Air + PBS vs. Air + LPS: p = 0.002) and decreased by both GP or SOA + GP exposure (Air + LPS vs. GP + LPS: p = 0.02, Air + LPS vs. SOA + GP + LPS: p = 0.002). These results show that under moderate systemic inflammation conditions, exposure to β-pinene oxidation gaseous products did not significantly disturb mouse lungs, whereas SOA and GP inhalation significantly disturbed most of the immune response induced by LPS administration. These findings support the adverse effects of β-pinene oxidation products on the pulmonary response under inflammatory conditions.

Fig. 3

Pulmonary response to moderate systemic inflammation after acute direct exposure in adulthood to β-pinene oxidation products. (A) Experimental design. n = 10/group. (B) Mouse XL cytokine panel assay in BAL fluid. (C-F) Immune cell frequency in the lungs. Parent population was CD45⁺CD3⁺TCRβ⁺ cells for CD4⁺, CD8⁺, Reg T cells and CD45⁺CD5⁺NK1.1⁺ cells for NKT like T cells. (G) mRNA expression levels of cytokines in the lungs. A representative of two independent experiments is shown. The data represent the means ± SEMs. * p < 0.05, ** p < 0.01

Second, the systemic immune response was analysed. Immunophenotyping of the spleen revealed that the B-cell frequency was greater in the Air + LPS group than in the Air + PBS group (p = 0.02), and a significant increase in B-cell frequency was detected in both the GP + LPS group and the SOA + GP + LPS group compared with the Air + LPS group (p = 0.04 and p = 0.01, respectively; Fig. 4A). The abundance of T cells was reduced by LPS administration (Air + PBS vs. Air + LPS: p = 0.002) and enhanced by SOA and GP exposure under systemic inflammation conditions (Air + LPS vs. SOA + GP + LPS: p = 0.004). The population of inflammatory monocytes did not significantly differ (Fig. 4B). Neutrophils, CD4⁺ T cells, CD8⁺ T cells and NKT-like cells were impaired by LPS-induced systemic inflammation (Air + PBS vs. Air + LPS: neutrophils: p = 0.002; CD4⁺ T cells: p = 0.002; CD8⁺ T cells: p = 0.002; NKT-like cells: p = 0.002) but not by GP or SOA and GP inhalation (Fig. 4A and C). The increased recruitment of regulatory T cells induced by LPS administration (Air + PBS vs. Air + LPS: p = 0.002) was partly reversed by SOA and GP inhalation (Air + LPS vs. SOA + GP + LPS: p = 0.002). Furthermore, the protein and mRNA levels of inflammatory cytokines and chemokines in the liver were measured. The decreased secretion of EGF and IL9 induced by LPS administration (Air + PBS vs. Air + LPS: EGF p = 0.004, IL9 p = 0.009) was reversed by SOA and GP exposure (Air + LPS vs. SOA + GP + LPS: EGF p = 0.01, IL9 p = 0.03), as was the increased secretion of ICAM1 (p = 0.04 in the Air + PBS vs. Air + LPS group, p = 0.004 in the SOA + GP + LPS vs. Air + LPS group; Fig. 4D). Cxcl1/Kc and Csf1 mRNA expression was greater in the Air + LPS group than in the Air + PBS group (Cxcl1: p = 0.002, Csf1: p = 0.04; Fig. 4E). Compared with those in the Air + LPS group, the levels of Csf1 and Il13 were greater in the GP + LPS group (Csf1: p = 0.01, Il13: p = 0.009). Compared with those in the Air + LPS group, Cxcl1/Kc and Cxcl10 transcript levels were significantly lower, and Il13 transcript levels were significantly greater in the SOA + GP + LPS group (Cxcl1: p = 0.04, Cxcl10: p = 0.03, Il13: p = 0.01). Therefore, at the systemic and pulmonary levels, the immune response induced by LPS administration was disrupted in mice exposed to particulate and gaseous β-pinene oxidation products, indicating that exposure to these pollutants led to inappropriate or exaggerated immune responses in the host.

Fig. 4

Systemic response to moderate systemic inflammation after acute direct exposure in adulthood to β-pinene oxidation products. (A-C) Immune cell frequency in the spleen. Parent population was CD45⁺CD3⁺TCRβ⁺ cells for CD4⁺, CD8⁺, Reg T cells and CD45⁺CD5⁺NK1.1⁺ cells for NKT like T cells. (D) Mouse XL cytokine panel assay in the liver. (E) mRNA expression levels of cytokines in the liver. A pool of two independent experiments is shown. The data represent the means ± SEMs. * p < 0.05, ** p < 0.01

Pulmonary and systemic responses to moderate systemic inflammation in female offspring exposed to β-pinene oxidation products in utero

We then reproduced the exposure protocol in pregnant mice, and we sought to evaluate the effects of this exposure on the offspring in adulthood (Scheme 1). Pregnant mice were exposed to both particulate and gaseous products of β-pinene oxidation, only gaseous products of β-pinene oxidation, or synthetic air 2 h per day for 3 days from gestational day 14 to gestational day 16. We deliberately chose not to expose the mice during the first part of gestation. Since the first part of gestation is the most vulnerable period, this would have increased the risk of gestational abnormalities, and these reproductive abnormalities would have biased the immune abnormalities that are the subject of this study. The exposure was also limited to 2 h per day in order to reduce the mouse stress. A low dose of LPS was administered to male and female offspring at the age of 8 weeks, and the pulmonary and systemic immune responses were analysed 16 h later. In female offspring, lung immunophenotyping revealed that LPS administration led to a decreased frequency of B cells and NKT-like cells and an increased frequency of dendritic cells, alveolar and interstitial macrophages, neutrophils and regulatory T cells (Air + PBS vs. Air + LPS: B cells: p = 0.03; NKT-like cells: p = 0.003; dendritic cells: p = 0.0007; alveolar macrophages: p = 0.001; interstitial macrophages: p = 0.008; neutrophils: p = 0.0007; regulatory T cells: p = 0.04; Fig. 5A). Compared with those in the lungs of the LPS-challenged female mice exposed to air in utero, a decreased number of B cells and an increased number of CD103⁺ dendritic cells were observed in the lungs of the LPS-challenged female mice exposed to GPs in utero (Air + LPS vs. GP + LPS: B cells: p = 0.003; CD103⁺ dendritic cells: p = 0.005; Fig. 5A-D). Under systemic inflammatory conditions, in contrast with mice exposed to air in utero, no variation in immune populations was detected in mice exposed to SOA and GPs in utero. Analysis of the activation markers of T cells and dendritic cells revealed significantly greater CD69 expression in CD8⁺ T cells in the SOA + GP + LPS group than in those in the Air + LPS group (p = 0.004, Fig. 5 Extended Data A-D). Lung function was not affected under any conditions, as demonstrated by flexiVent analysis (Fig. 1 Extended Data E). However, at the transcriptional level, the upregulation of Ccl2/Mcp1, Cxcl10, and Icam1, which was induced by moderate systemic inflammatory conditions (Air + PBS vs. Air + LPS: Ccl2: p = 0.0007; Cxcl10: p = 0.0007; Icam1: p = 0.0007), was exacerbated in the mice exposed to SOA and GPs in utero (Air + LPS vs. SOA + GP + LPS: Ccl2: p = 0.01; Cxcl10: p = 0.04; Icam1: p = 0.02; Fig. 5E).

Scheme 1

Experimental design for gestational exposure to β-pinene oxidation products

Fig. 5

Pulmonary response to moderate systemic inflammation in female offspring exposed to β-pinene oxidation products in utero. (A-D). Immune cell frequency in the lungs. Parent population was CD45⁺CD3⁺TCRβ⁺ cells for CD4⁺, CD8⁺, Reg T cells and CD45⁺CD5⁺NK1.1⁺ cells for NKT like T cells. (E) mRNA expression levels of cytokines in the lungs. A pool of two independent experiments is shown. n = 10/group. The data represent the means ± SEMs. * p < 0.05, ** p < 0.01, *** p < 0.001

We next analysed the systemic response of female offspring. In the spleen, an increase in neutrophils and a decrease in inflammatory monocytes and CD103⁺ dendritic cells were observed in the Air + LPS group compared with the Air + PBS group (neutrophils: p = 0.01; inflammatory monocytes: p = 0.003; CD103⁺ dendritic cells: p = 0.0003; Fig. 6A-D). Under moderate systemic inflammation conditions, splenic dendritic cells were less abundant in the mice exposed in utero to GPs or to SOA and GPs than in the mice exposed in utero to air (GP + LPS vs. Air + LPS: p = 0.01, SOA + GP + LPS vs. Air + LPS: p = 0.03). The frequency of inflammatory monocytes was lower in the mice exposed in utero to GPs or to SOA and GPs than in the mice exposed in utero to air (GP + LPS vs. Air + LPS: p = 0.03, SOA + GP + LPS vs. Air + LPS: p = 0.04). The analysis of activation markers revealed that LPS administration increased CD69 expression in CD4⁺ and CD8⁺ T cells (Air + LPS vs. Air + PBS: p = 0.0007 and p = 0.02, respectively) and CD25 expression in CD8⁺ T cells (Air + LPS vs. Air + PBS: p = 0.02, Fig. 6E-F). LPS administration also increased CD86 and MHC class II (Iab) in total dendritic cells (Air + LPS vs. Air + PBS: p = 0.0007 and p = 0.001, Fig. 6G-H). Moreover, increased CD86 and Iab in total dendritic cells (p = 0.005 and p = 0.01, respectively) and Iab in CD11b⁺ dendritic cells (p = 0.007) were detected in the SOA + GP + LPS group compared with the Air + LPS group. Furthermore, decreased Il10, Il13, and Csf3 mRNA expression was detected in the liver in the SOA + GP + LPS group compared with the Air + LPS group (Il10: p = 0.02; Il13: p = 0.009; Csf3: p = 0.01; Fig. 6I). Taken together, these findings showed that in utero exposure to β-pinene oxidation products did not induce drastic perturbations in the pulmonary and systemic response to LPS in female offspring. The main disturbances are decreased recruitment and increased activation of splenic dendritic cells, decreased abundance of splenic inflammatory monocytes and decreased expression of hepatic Il10, Il13 and Csf3.

Fig. 6

Systemic response to moderate systemic inflammation in female offspring exposed to β-pinene oxidation products in utero. (A-D) Immune cell frequency in the spleen. Parent population was CD45⁺CD3⁺TCRβ⁺ cells for CD4⁺, CD8⁺, Reg T cells and CD45⁺CD5⁺NK1.1⁺ cells for NKT like T cells. (E-F) CD69 and CD25 activation markers on T cells. (G-H) CD86 and Iab activation markers on dendritic cells. (I) mRNA expression levels of cytokines in the liver. A pool of two independent experiments is shown. n = 10/group. The data represent the means ± SEMs. * p < 0.05, ** p < 0.01, *** p < 0.001

Pulmonary and systemic responses to systemic inflammation in male offspring exposed to β-pinene oxidation products in utero

We next assessed the same parameters in male offspring. The response to systemic inflammation in the male lung was characterized by a decreased abundance of B cells, T cells and CD103⁺ dendritic cells and an increased abundance of alveolar macrophages, neutrophils, inflammatory monocytes, regulatory T cells, and CD11b⁺ dendritic cells (Air + PBS vs. Air + LPS: B cells: p = 0.002; T cells: p = 0.002; CD103⁺ dendritic cells: p = 0.03; alveolar macrophages: p = 0.0007; neutrophils: p = 0.0007; inflammatory monocytes: p = 0.03; regulatory T cells: p = 0.002; and CD11b⁺ dendritic cells: p = 0.004; Fig. 7A-D). No significant difference in pulmonary immune populations was found between the SOA + GP + LPS group and the Air + LPS group. A decreased number of total dendritic cells and increased numbers of inflammatory monocytes and CD103⁺ dendritic cells were observed in the GP + LPS group compared with the Air + LPS group (dendritic cells: p = 0.04; inflammatory monocytes: p = 0.003; and CD103⁺ dendritic cells: p = 0.04). The analysis of activation markers revealed that LPS administration increased CD69 and CD25 expression in total conventional (Air + PBS vs. Air + LPS: CD69: p = 0.002, CD25: p = 0.002), CD4⁺ (CD69: p = 0.006, CD25: p = 0.004), and CD8⁺ T cells (CD69: p = 0.002, CD25: p = 0.04). The only significant variation induced by exposure to GPs or SOA and GPs in utero was increased CD86 expression in total dendritic cells (Air + LPS vs. GP + LPS: p = 0.001, Air + LPS vs. SOA + GP + LPS: p = 0.005; Fig. 7 Extended Data A-D). As in females, lung function was not affected under any conditions (Fig. 7 Extended Data E). The gene expression quantification of cytokines that were dysregulated in female offspring (see above) revealed the upregulation of Ccl2/Mcp1 and Cxcl10 in male lungs in response to LPS administration (Air + PBS vs. Air + LPS: Ccl2: p = 0.0007, Cxcl10: p = 0.0007), but unlike in female mice, this upregulation remained identical regardless of the type of in utero exposure (Fig. 7E).

Fig. 7

Pulmonary response to moderate systemic inflammation in male offspring exposed to β-pinene oxidation products in utero. (A-D) Frequencies of major immune cells in the lungs. Parent population was CD45⁺CD3⁺TCRβ⁺ cells for CD4⁺, CD8⁺, Reg T cells and CD45⁺CD5⁺NK1.1⁺ cells for NKT like T cells. (E) mRNA expression levels of cytokines in the lungs. A pool of two independent experiments is shown. n = 10/group. The data represent the means ± SEMs. * p < 0.05, ** p < 0.01, *** p < 0.001

At the systemic level, i.e., in the spleen, LPS challenge increased the frequencies of B cells, neutrophils, CD4⁺ T cells and regulatory T cells and decreased the frequencies of CD8⁺ T cells and CD103⁺ dendritic cells (Air + PBS vs. Air + LPS: B cells: p = 0.0007; neutrophils: p = 0.005; CD4⁺ T cells: p = 0.002; regulatory T cells: p = 0.0007; CD8⁺ T cells: p = 0.0007; and CD103⁺ dendritic cells: p = 0.001; Fig. 8A‒D). Compared with this response, increased B cells and decreased CD4⁺ T cells and CD11b⁺ dendritic cells were observed in the spleens of the mice exposed to GPs in utero (Air + LPS vs. GP + LPS: B cells: p = 0.0003; CD4⁺ T cells: p = 0.04; CD11b⁺ dendritic cells: p = 0.05). Compared with those in the air + PBS group, CD69 and CD25 were upregulated in total conventional (CD69: p = 0.0007, CD25: p = 0.0007), CD4⁺ (CD69: p = 0.0007, CD25: p = 0.0007) and CD8⁺ T cells (CD69: p = 0.002, CD25: p = 0.0007) in the Air + LPS group (Fig. 8E-F). CD25 expression in conventional and CD4⁺ T cells was lower in the SOA + GP + LPS group than in the Air + LPS group (conventional T cells: p = 0.05; CD4⁺ T cells: p = 0.05; Fig. 8F). CD86 was upregulated in total and CD103⁺ dendritic cells in the Air + LPS group compared with those in the Air + PBS group (p = 0.008 and 0.001, respectively; Fig. 8G). Consistently, Iab was upregulated in total and CD103⁺ dendritic cells in the Air + LPS group compared with those in the Air + PBS group (p = 0.001 and 0.003, respectively; Fig. 8H). A further significant increase in CD86 in CD103⁺ splenic dendritic cells occurred in the SOA + GP + LPS group compared with the Air + LPS group (p = 0.05, Fig. 8G). The expression of cytokines that were deregulated in female livers was also quantified in male livers; however, their expression did not vary in males (Fig. 8I). Overall, in males, with the exception of the dysregulated abundance of inflammatory monocytes and CD103⁺ dendritic cells in the lung and B, CD4⁺ T and CD11b⁺ dendritic cells in the spleen caused by in utero exposure to GPs and the impaired activation of splenic CD4⁺ T cells and CD103⁺ dendritic cells caused by in utero exposure to SOA and GPs, in utero exposure to β-pinene oxidation products did not prevent an efficient pulmonary and systemic response to systemic inflammation in males.

Fig. 8

Systemic response to moderate systemic inflammation in male offspring exposed to β-pinene oxidation products in utero. (A-D) Frequencies of major immune cells in the spleen. Parent population was CD45⁺CD3⁺TCRβ⁺ cells for CD4⁺, CD8⁺, Reg T cells and CD45⁺CD5⁺NK1.1⁺ cells for NKT like T cells. (E-F) CD69 and CD25 activation markers on T cells. (G-H) CD86 and Iab activation markers on dendritic cells. (I) mRNA expression levels in the liver. A pool of two independent experiments is shown. n = 10/group. The data represent the means ± SEMs. * p < 0.05, ** p < 0.01, *** p < 0.001

Discussion

In indoor and outdoor air, reactions between volatile organic compounds and oxidants produce particulate and gaseous oxidation products of extreme diversity. SOA accounts for a significant portion of airborne particulate matter. However, the contributions of particulate and gaseous products of secondary oxidation reactions to the known hazardous effects of air pollution remain poorly understood. We therefore developed a setup that allows the exposure of mice to secondary oxidation products synthesized in real time via inhalation. The particulate and gaseous fractions were combined in an experimental group, the purpose of which was therefore to simulate human exposure because humans are indeed exposed to a mixture of particulate and gaseous fractions. This choice brings our experiments closer to real life and, to our knowledge, represents the first exposure of this type. Some studies carried out thus far have been based on exposure to SOA sampled and then administered after storage to mice. For example, Niu et al. simulated the reactions of O₃ and limonene with and without the presence of NH₃. The SOA was sampled, stored at -4 °C, extracted with distilled deionized water and then administered to the mice. One intranasal instillation of 5 µg/mouse limonene-derived SOA induced pulmonary oxidative stress and inflammation [35, 36]. Other studies have generated real-time SOA such as our study, but have deliberately chosen to remove the gases for inhalation by mice to focus only on the effects of the SOA. The first known laboratory inhalation exposure to SOA was conducted in 2010 by Jacob D. McDonald et al. [37]. They studied the cardiopulmonary response in rats and mice exposed to SOA derived from α-pinene oxidation without gas-phase reaction products. Seven days of exposure via nose-only inhalation at a concentration of 200 µg.m^− 3 elicited only mild vascular to no pulmonary response. They confirmed in a similar study that SOA derived from α-pinene or toluene modulated the expression of vascular factors associated with the progression of cardiovascular disease in male atherosclerotic apolipoprotein E null mice [38]. Notably, Win-Shwe et al. generated diesel exhaust SOA by mixing diesel exhaust with O₃ at 0.6 ppm. They showed that an acute single intranasal instillation of diesel exhaust SOA (50 µg/mouse) induced an inflammatory response in the lungs but not in the brains of adult mice [39]. They subsequently performed 2 studies in which an inhalation chamber was used. The first study reported an impaired discrimination ability between familiar and novel objects in male mice exposed to diesel exhaust SOA for three months, and maternal behavior appeared to decrease in female mice exposed to diesel exhaust SOA for one month before pregnancy [40]. The second study indicated that early-life exposure of BALB/c mice to diesel exhaust SOA may affect late-onset hypothalamic expression of social behavior-related genes, trigger neurotoxicity and impair social behavior in males [41]. Diesel exhaust SOA exposure in utero and during the neonatal period may affect olfactory-based spatial learning behavior [42] and induce spatial learning impairment in neonatal mice [43]. Perinatal exposure to diesel exhaust-derived SOA induced autism-like behavior in rats [44]. Overall, knowledge of the health effects of secondary oxidation remains very limited, and our study focusing on the combined effects of particulate and gaseous β-pinene oxidation products is therefore particularly innovative.

In our experimental protocol, the oxidation of β-pinene resulted in the formation of SOA at an average concentration of 500 µg. m^− 3 and delivered 2 h per day for 3 days. Owing to complex chemical transformation reactions, the involvement of numerous start and end products, and the variable physicochemical characteristics of existing compounds in the atmosphere, direct estimation of SOA in air is not possible. However, SOA can be estimated via indirect methods, which have allowed the determination of SOA concentrations of 2.9 µg.m^− 3 in the USA [45], 9 µg. m^− 3 [46], and 5.17 to 10.7 µg. m^− 3 in China [47]. More specifically, at a suburban site in central India, the concentrations of the 4 monoterpene SOA tracers were 1.82, 2.81, 4.68, and 6.88 µg.m^− 3 [48]. On the basis of these data, our protocol simulated 2 h per day for 3 days an exposure which is approximately 100 times greater than the chronic daily exposure in humans. Similar concentrations are commonly used in murine studies of the impact of atmospheric particles [49,50,51]. It should be noted that, since inhalation was performed in a whole-body exposure chamber to avoid restraint of pregnant mice, the associated limitation is that some of the aerosols may have settled on the ground or fur before being inhaled by the mice, which may have led to a reduction in the inhalation exposure dose and possibly low oral exposure.

The composition of gaseous products and SOA formed from the ozonolysis of β-pinene has been previously reported. The O₃ reaction with alkenes proceeds via the addition of O₃ to the C = C double bond, leading to the formation of an ozonide, which then decomposes into two stable volatile organic compounds (nopinone or formaldehyde) as well as two Criegee intermediates. The stabilization of Criegee intermediates leads to the formation of other oxygenated organic compounds, such as pinic acid, norpinic acid, pinalic-3-acid, pinalic-4-acid, norpinalic acid, 4-OH-pinalic-3-acid, 3-ketopinone, 3-hydroxynopinone, carbon monoxide, and carbon dioxide [32,33,34]. In addition to the gas-phase oxidation products, the ozonolysis of β-pinene is known to lead to the formation of SOA due to the condensation of low-volatility products. The analysis of SOA, performed via liquid chromatography‒mass spectrometry, revealed the formation of oligomers [52]. The chemical composition of SOA is very complex due to the number of species that constitute the oligomers; nevertheless, some of them are highly oxygenated organic molecules [52].

Our results first revealed that immune response disturbances were induced by adult exposure to secondary oxidation products of β-pinene, which occurred moderately under basal conditions and exacerbated under inflammatory conditions. The pulmonary and systemic responses to systemic inflammation in mice subjected to acute exposure to secondary oxidation products of β-pinene differ from those in mice that breathed unpolluted air. In particular, excessive recruitment of B lymphocytes, CD4⁺ T lymphocytes, CD11b⁺ dendritic cells, neutrophils and inflammatory monocytes occurred in the lung, whereas a defect in NKT recruitment was observed. At the transcriptomic level, the cytokine response of the lung was reduced, which may be due to the establishment of a response intended to counter the excessive recruitment of inflammatory cells.

The majority of these effects were observed only in the group exposed to the mixture of particles and gases. Two principal hypotheses can then be formulated: either these effects are mainly due to the action of SOA, or they are due to a synergistic effect taking place only in the presence of gaseous and particulate fractions of secondary oxidation products. A major limitation of our experimental protocol is that it does not include a group exposed to SOA only and therefore does not allow us to determine the respective contributions of the gaseous and particulate fractions or to prove an additive, opposite or synergistic effect between the two fractions.

Furthermore, our results focused on the generational effects of acute exposure to secondary oxidation products of β-pinene. In both male and female offspring mice, no functional abnormalities were observed by flexiVent exploration, indicating that no serious toxic effects were induced. However, in female offspring, a repeated trend toward T-cell overactivation was associated with a significant increase in CD69 expression in CD8⁺ T cells, suggesting that the pulmonary response of these mice subjected to the pollutant mixture in utero could be altered. The exacerbation of the overexpression of CCl2, Cxcl10 and Icam1 in the lungs supports this hypothesis. In the spleen and liver, other alterations in the immune response, such as defects in the recruitment of inflammatory monocytes, overactivation of dendritic cells in the spleen, and defects in the expression of Il10, Il13 and Csf3 in the liver, have been observed. These results suggest that maternal exposure leads to a defect in the maturation of the immune system in female offspring, which leads to greater susceptibility to diseases in adulthood. This hypothesis is part of the DOHAD concept, which postulates that the early-life environment and events have long-term effects on shaping health and disease vulnerability later in life [53]. Some studies have shown that air pollution can contribute to individuals’ susceptibility to noncommunicable diseases. In particular, in mice, in utero exposure to ultrafine particulate matter, diesel exhaust air pollution or simulated complex urban air pollution caused offspring pulmonary immunosuppression [54], increased offspring susceptibility to heart failure [55], or disturbed offspring gut maturation and microbiota [56], respectively. In humans, PM_2.5 exposure during pregnancy increases the risk of asthma, wheezing [57], and autism spectrum disorder [58]. Another meta-analysis revealed a link between outdoor air pollution exposure during pregnancy and leukemia risk [59]. Therefore, in view of our results and the literature, it would be interesting to study whether gestational exposure to secondary oxidation products could lead to pathologies involving the pulmonary and systemic immune systems, such as asthma and pulmonary infections, in offspring.

Moreover, our results showed that in male offspring exposed to secondary oxidation products, the response to LPS continues to be strong but does not show any variation specific to the type of gestational exposure. The effects observed in female offspring are therefore sex specific, which is consistent with other effects of air pollution reported as sex specific [60]. Accordingly, D Rousseau-Raillard reported that in utero nose-only exposure to diesel engine exhaust predisposed female rabbit offspring to cardiometabolic disorders in a sex-specific manner [61]. In contrast, prenatal diesel exhaust particle exposure alters metabolic, behavioral, and neuroinflammatory responses to a high-fat diet in adult males but not females [62]. In humans, prenatal PM_2.5 exposure in late pregnancy is associated with impaired early childhood lung function and early childhood asthma development, which are more evident in boys [63, 64].

Notably, few alterations in the immune response were detected in mice exposed only to gaseous products. It can be speculated that the response induced by gaseous products was opposite to that induced by SOA and that the combination of opposing effects resulted in the absence of a response in the mice exposed to all the β-pinene oxidation products. Formaldehyde is a gaseous product of the ozonolysis of monoterpenes that can influence the induced immune response because its immunomodulatory properties have been reported. In C57BL/6 mice, exposure to 2 mg/kg bw formaldehyde for 1 week or one month increased the serum levels of several Th1, Th2 and Th17 cytokines [65]. The exposure of mice to formaldehyde (3 mg.m^− 3, 8 h per day for 12 days) in a whole-body chamber increased the ROS level of the splenic cells but did not affect the number of lymphocytes, monocytes or neutrophils in the blood or delayed-type hypersensitivity responses [66]. Wen et al. exposed C57BL/6 and BALB/c mice to 0.5 or 3.0 mg.m^− 3 formaldehyde with or without ovalbumin for 6 h/day over 25 consecutive days [67]. Th2-type allergic responses were induced in both nonsensitized BALB/c and C57BL/6 mice and were more prominent in BALB/c mice. In addition, exposure to 3 mg.m^− 3 formaldehyde in sensitized C57BL/6 mice suppressed the development of ovalbumin-induced allergic responses, whereas exposure to 0.5 mg.m^− 3 formaldehyde exacerbated allergic responses to ovalbumin. In contrast to formaldehyde, other gaseous products of β-pinene ozonolysis have not been studied in mice. An in vitro study was conducted in BEAS-2B bronchial epithelial cells on α-pinene SOA. The α-pinene SOA was generated, collected on a filter, and stored at -20 °C, after which the methanol extracts were tested for their ability to induce a cytotoxic response and cellular oxidative stress. Pinic acid, pinonic acid, and 3-methyl-1,2,3-butanetricarboxylic acid, which contributed ∼57% of the α-pinene SOA mass, were tested individually and in combination, and they did not induce significant toxicity, suggesting that the 3 aforementioned compounds are not critically toxic to bronchial cells. However, the paucity of studies does not exclude the possibility that pinic acid or other gaseous compounds involved in the ozonolysis of β-pinene may also have immunomodulatory properties.

Conclusions

In conclusion, in this study, mice were exposed by inhalation to particulate and gaseous products of b-pinene ozonolysis synthetized in real time. Exposure to gaseous products only slightly modified the pulmonary and systemic immune response, in all tested experimental conditions. By contrast, the immune response was disturbed by the most physiological exposure, combining both particulate and gaseous products, suggesting a major contributory role of particulate products. The disturbances were observed after acute exposure in adulthood. In healthy individuals, acute exposure to the mixture of particulate and gaseous products reduced the pulmonary frequency of CD4⁺ T lymphocytes and CD103⁺ dendritic cells. In individuals with existing immune activation, it induced a disrupted response to LPS, including excessive pulmonary recruitment of B lymphocytes, CD4⁺ lymphocytes, CD11b⁺ dendritic cells, inflammatory monocytes and neutrophils. Immune disturbances were also observed, to a lesser extent, in female offspring of exposed dams. Upregulation of chemokine expression in lung and the activation of dendritic cells in spleen in response to LPS were exaggerated. These perturbations are unlikely to cause adverse effects in healthy individuals, but it cannot be ruled out that they could contribute to adverse effects in diseased individuals, such as increased susceptibility to infection in immunocompromised patients. Further studies are needed to determine whether and to what extent secondary oxidation reactions contribute to the development of pathologies related to immune system dysfunction.