Differential Lung Proteome Profiles Upon Pulmonary Exposure to Traffic-related Particulate Matter


 Background: Exposure to particulate matter (PM) pollution has direct impacts on the respiratory organs, yet the molecular alterations underlying PM-induced pulmonary injury remain unclear. In this study, we investigated the effect of PM on lung tissues from our previously reported rat model with whole-body exposure to traffic-related PM1 pollutants and compared it with rats exposed to high-efficiency particulate air-filtered gaseous pollutants and clean air control for 3 and 6 months. Lung function and histological examinations as well as quantitative proteomics analysis and functional validation were performed. Results: The rats in 6-month PM1-exposed group showed significant decline in lung function by decreased forced expiratory flow and forced expiratory volume, but the histological analysis revealed an earlier lung damage evidenced by increased congestion and macrophage infiltration in 3-month PM1-exposed rat lungs. The lung tissue proteomics analysis identified 2,673 proteins which highlighted dysregulations on proteins involved in oxidative stress, cellular metabolisms, calcium signaling, inflammatory responses, and actin dynamics. The presence of fine particles specifically enhanced the oxidative stress and inflammatory reactions under sub-chronic exposure of traffic-related PM1 and suppressed the glucose metabolism and actin cytoskeleton signaling which might lead to repair failure and thus lung function decline after chronic exposure of traffic-related PM1. A detailed pathogenic mechanism was proposed to depict the temporal and dynamic molecular regulations associated with PM1-induced lung injury.Conclusion: Our study explored the earlier lung injury prior to lung function decline and proposed several proteins as potential molecular features for traffic-related PM1-induced lung injury.


Background
Ambient air pollution contributes substantially to major disease burden as well as mortality globally, with at least 4.2 million reported deaths in 2016 (1). Among various pollutants in ambient air pollution, exposure to particulate matter (PM), particularly in the ne (PM 2.5 ) and ultra ne ranges (PM 0.1 ), is considered as a key risk factor for many adverse health consequences of which acute or chronic respiratory complaints are prominently reported due to a direct deterioration in the organs upon inhalation of air pollution (2, 3). Short-or long-term exposure to ne PM air pollution would signi cantly reduce the lung function with increased pulmonary oxidative stress and persistent in ammation (4)(5)(6)(7). Multiple signaling pathways involved in the transcriptional and/or translational activations of AMP-activated protein kinase (AMPK)/signal transducer and activator of transcription (STAT)-1 (8), epithelial growth factor receptor (EGFR)/mitogen-activated protein kinase (MAPK)/ nuclear factor κB (NF-κB) (9), and transforming growth factor (TGF)-β/Smad (10) that promote the release of pro-in ammatory cytokines were proposed on lung epithelium cells or rodent models as potential pathogenic mechanisms of the PMinduced pulmonary toxicity. However, the detailed pathogenesis behind remains to be fully investigated.
Mass spectrometry (MS)-based proteomics analysis that enables unbiased identi cation and quanti cation of thousands of proteins has been applied to study the aberrant molecular pro les upon PM-induced damages on primary skin keratinocytes (11), trophoblast cells (12), lung epithelial cells (13), rat lung (14) and rat brain (15). Several dysregulated proteins involved in mitochondria dysfunction, energy metabolism and ER stress were discovered as potential biomarkers for PM-induced organ damage. To exploit the impact of PM on organ injury, we have previously established the rat model with whole-body exposure system and reported the central neurotoxicity induced by sub-chronic or chronic exposure of tra c-related ne PM (PM 1 ), even though it was below the WHO air quality guideline (16). In this study, we aimed to investigate the effects of tra c-related PM 1 in lungs from the same rat model.
Speci cally, we elaborated the lung function and pathological changes upon sub-chronic or chronic exposure of tra c-related PM 1 as well as systematically elucidated the dysregulated molecules and signaling pathways by using quantitative proteomics analysis to construct the molecular mechanisms underlying tra c-related PM 1 -induced lung injury.

Lung function and histological examination
We examined the lung function of each rat by measuring the forced expiratory ow at 25-75% of the pulmonary volume (FEF  ) and forced expiratory volume at 20 ms (FEV 20 ). As shown in Fig. 1, there was no signi cant differences in FEF  and FEV 20 among rats in the 3-month exposure groups. After 6month exposure rats in PM 1 group showed signi cant decrease in FEF  and FEV 20 in comparison to both GAS and CTL rats. We also evaluated the degree of lung injury based on the presence and severity of congestion, hemorrhage, immune cell in ltration, and thickness of the alveolar wall using the lung histological analysis. The 3-month PM 1 -exposed lungs already exhibited higher levels of congestion and macrophage in ltration, indicating a signi cant lung injury in PM 1 rats compared to GAS and CTL rats.
Similar results were observed after chronic exposure to PM 1 (Table 1). Fig. 2 showed the histological images demonstrating the increased thickness of airway wall and disruption of alveolar and airways integrity, along with abundant immune cell accumulation within the peribronchial area in both 3-month and 6-month PM 1 -exposed lungs. However, no signi cant difference in lung jury was observed between rats under 3-month and 6-month exposures.

Quantitative proteomics analysis of rat lungs
In order to elucidate the molecular mechanisms underlying tra c related air pollution (TRAP)-induced lung injury, we applied the tandem mass tag (TMT)-based quantitative proteomics analysis on the lung tissues from the 6 exposure groups. As shown in Fig. 3A, ve rats were randomly selected from each group to collect the lung tissue lysate and perform gel-assisted digestion with trypsin individually.
Peptides from the same group of rats were pooled and labeled with one of the TMT tags. All the TMTlabeled peptides were combined for high-pH reversed phase (RP) StageTip fractionation and each fraction was analyzed in duplicate by LC-MS/MS, followed by proteome identi cation and quanti cation using Proteome Discoverer. A total of 2,673 proteins were con dently identi ed (p<0.05, FDR<1%) of which 2,562 proteins were quanti ed (Fig. 3A).
In the present study, we arranged our analyses in three directions and considered proteins with 1.3-fold change (log2 ratio >0.38 or <-0.38) in abundance in two comparisons as differentially expressed proteins (DEPs). (1) We aimed to observe the sub-chronic effect of TRAP by comparing the protein expressions in 3-month (3M) exposure of GAS and PM 1 groups to the clean-air CTL group which generated 218 and 179 DEPs in the 3M-GAS and 3M-PM 1 groups, respectively. (2) The proteins associated with progressive lung injury when exposed to GAS and PM 1 were studied by comparing the protein expressions between 6-and 3-month exposures on each GAS and PM 1 groups. A total of 408 and 413 progression-associated DEPs were identi ed in GAS and PM 1 groups, respectively. (3) We elucidated the particle-speci c regulations by comparing the PM 1 to GAS under sub-chronic (3-month) and chronic (6-month) exposures, which resulted in 119 and 103 DEPs in 3M and 6M comparisons, respectively (Fig. 3B). The DEPs were listed in Additional le 1-3: Table S1-3 and were further analyzed by using Gene Ontology (GO) and IPA (Ingenuity pathway analysis) to delineate the dysregulated cellular functions and pathways in lungs.
Dysregulated cellular functions and pathways in rat lungs under sub-chronic and chronic exposure to GAS and PM 1 pollutants As shown in Additional le 4: Figure S1, 3-month exposure to tra c related air pollution (TRAP)-related gaseous and PM 1 pollutants induced alterations in proteins involved in metabolism-and acute phase response-related biological processes. Upregulation of arginine metabolic process and fatty acid betaoxidation, as well as down-regulation of endopeptidase inhibitor activity and negative regulation of mRNA metabolic process were enriched in 3M-GAS group (Additional le 4: Figure S1A). Similarly, subchronic exposure to PM 1 up-regulated triglyceride catabolism, reactive oxygen species (ROS) metabolism, long-chain fatty acid metabolism, and down-regulated glycolytic process within lung tissues (Additional le 4: Figure S1B). It is noted that 3M-GAS group exhibited down-regulation of immune-related functions, including negative regulation of CaN-NFAT signaling cascade, blood coagulation and complement activation (classical pathway). Meanwhile the 3M-PM 1 group regulated processes related to tissue damage and wound healing, such as ATP biosynthesis, ROS metabolism, regulation of wound healing and DNA geometric change. The DEPs in both 3M-GAS and 3M-PM 1 groups were enriched as mitochondrial membrane proteins. Additionally, the 3M-PM 1 group enriched proteins localized in sarcoplasmic reticulum (Additional le 4: Figure S1). Pathway analysis of the DEPs suggested common up-regulation of oxidative phosphorylation and inositol phosphate metabolism and down-regulation of sirtuin signaling pathway in both 3M-PM 1 and 3M-GAS groups (Fig. 4). The glycolysis and gluconeogenesis pathways were enriched in both groups as well, but showing inhibition in only 3M-GAS group. In ammation related pathways, including acute phase response signaling and mTOR signaling, were activated exclusively in 3M-PM 1 group, while complement system and acute phase response signaling were inhibited in 3M-GAS group. Altogether, our results indicated that exposure to TRAP induced early metabolic changes. Sub-chronic PM 1 exposure promoted more acute phase responses within the lung tissue, while exposure to gaseous pollutant showed more suppression on the complement system.
We next analyzed the injury progression-related biological process within the lung tissue. As shown in Additional le 5: Figure S2, we observed several tissue development and wound healing related processes in both GAS-and PM 1 -exposed rats. The DEPs involved in in ammatory response and metabolism processes were also enriched. The cellular component mapping pointed out that DEPs located in extracellular vesicle and plasma membrane were upregulated, while cytosolic proteins were prominently downregulated in both groups. The pathway analysis result in Fig. 4A showed exclusive inhibition of inositol phosphate metabolism pathways in lung tissues upon prolonged exposure to GAS pollutant. The glycolysis I pathway was up-regulated in GAS group but down-regulated in PM 1 group. In ammatory pathways including acute phase response signaling and complement system were exclusively enriched in PM 1 group, yet the net effect is not clear. In addition, we observed some common pathways including activated G Beta Gamma and calcium signaling as well as inhibited protein kinase A, EIF2 signaling and LXR/RXR activation in both GAS and PM 1 groups. Interestingly, exposure to GAS pollutants uniquely enriched progressive up-regulation of pathways related to actin cytoskeleton and Rho family GTPases signaling which involve in cell migration, muscle contraction and potentially tissue repair.
Regarding the particle-speci c effects, the GO analyses suggested that the ne particles in the present study prominently up-regulated the endopeptidase inhibitor activity and humoral immune response in the sub-chronic exposure stage, while blood coagulation was continuously activated upon prolonged exposure (Additional le 6: Figure S3). Furthermore, the particles down-regulated proteins located in chromosome under sub-chronic exposure, while later at 6 months of exposure, downregulated proteins localized within the A band of the muscle ber were noticed (Additional le 6: Figure S3). Pathway analysis indicated that during sub-chronic exposure, the particles in the TRAP speci cally induced greater in ammatory reactions through LXR/RXR activation, acute phase response signaling, and complement response (Fig. 4B). Upon prolonged exposure, the particles inhibited glycolysis and gluconeogenesis as well as ILK and actin cytoskeleton signaling when compared to the GAS group. In summary, from 3 to 6 months of PM 1 exposure, we observed an early induction of in ammatory reactions and a progressive reduction in the glucose metabolism and cell movement functions.

Proposed molecular mechanisms underlying PM 1 -induced lung injury
Based on the functional analysis results, we proposed a detailed molecular mechanism in Fig. 5. The expression changes of DEPs were displayed by using the boxes that surrounded the protein gene symbol.
Boxes on the upper side of the protein gene symbol indicate the protein expression in PM 1 (upper) and GAS (lower) groups. Boxes on the left panel of upper side showed the protein ratios in 3M-PM 1 /CTL and 3M-GAS/CTL, while those on the right panel represented a progression from 3-to 6-month exposures (PM 1 -6M/3M and GAS-6M/3M, respectively). Upregulated proteins were marked in red, downregulated proteins in green, while unchanged proteins were shown in white color. We also indicated the particle-speci c regulations in both 3 (left panel) and 6 months (right panel) of exposures in boxes in the right side of the protein gene symbol (3M-PM 1 /GAS and 6M-PM 1 /GAS, respectively). Protein expressions enhanced by particles was presented with fuchsia color, while suppression was in cyan color.
Firstly, we highlighted the dysregulated proteins and pathways involved in acute phase response signaling during the sub-chronic exposure to tra c-related GAS and PM 1 ; Ras/Erk to NF-IL6 pathway, Ras/Pi3k/Akt to Nf-κB pathway, Il-6 to Stat3 pathway, and Tcf transcriptional regulations showed differential expressions in both 3M-GAS and 3M-PM 1 rats (Fig. 5, pathways in blue color). It is noted that proteins responding to oxidative stress, including Ttr (compared to 3M-CTL), Cp, and Hmox2 (compared to 3M-GAS), as well as the in ammatory related proteins (Itih3, Serpina3, Fga and Hpx) within the acute phase response pathways were up-regulated in 3M-PM 1 group. In addition, the presence of particles speci cally up-regulated complement proteins (C3, C4, C5 and C1q) in comparison to 3M-GAS group.
The mTOR complex which responds to stress and regulates cellular metabolism, cell growth and survival was dysregulated under sub-chronic PM 1 exposure (Fig. 4A, Fig. 5, pathway in purple color). In response to PM exposure, mTOR has been reported to be activated in macrophage (17) or suppressed in airway epithelium (18), where it both showed protective effect against lung injury by attenuation of in ammatory responses and cell death. In the present study, we observed up-regulated mTOR pathway, in particular the mTORC2 complex with upregulation of mTor and down-regulation of Mapkap1 and inhibitory Tsc2 under the sub-chronic exposure to PM 1 , suggesting potential protection and homeostasis effect to minimize lung injury. Furthermore, the up-regulated Ras/Raf/Mek/Erk signaling promoted Eif2 signaling through up-regulation of Ppp1c and Eif2b3 proteins (Fig. 5, pathways in grey color). However, this pathway was gradually down-regulated under prolonged exposure to both GAS and PM 1 with concurrent decrease in the 40S and 60S ribosomal protein complexes which resulted in reduced protein synthesis and enhanced apoptosis (19).
Secondly, we highlighted the dysregulation of calcium signaling cascade in PM 1 -induced injury progression ( Fig. 5, pathways in orange color). In our data, multiple signallings that evoke the regulation of intracellular calcium ion (Ca 2+ ) showed differential expressions. The inositol 1,4,5-trisphosphate (IP3) pathway was triggered by progressive elevation of G protein couple receptor (represented as Gβ4)/phospholipase C (Plcb2) signaling to release the Ca 2+ from endoplasmic reticulum to mitochondria and lysosome which further regulated metabolic processes. The progressive up-regulation of sarco/endoplasmic reticulum Ca 2+ -ATPase (shown as Atp2a1) and down-regulation of ryanodine receptor (Ryr1) suggested a re-storage and inhibited release of Ca 2+ from sarcoplasmic reticulum to cytoplasm in the cell. In addition, we observed the upregulations of intracellular calcium-binding Camk2g (calcium/calmodulin-dependent protein kinase type II subunit gamma) as well as inhibitory Chp1 (Calcineurin B homologous protein 1) and Akap5 (A-kinase anchor protein 5) proteins that suppress the calcium-dependent calcineurin phosphatase activity for calcium homeostasis control. According to these results, we speculated an accumulation of intracellular Ca 2+ in rat lungs which might promote lymphocyte activation during injury progression.
Thirdly, proteins involved in RhoA/Rock-, Rac-and Cdc42-mediated signalings for controlling the muscle contraction, actin dynamics, and cell migration were also impaired during the progression of lung injury (Fig. 5, pathways in green color). Interestingly, we observed an unbalanced activation of these three pathways in the PM 1 rats. The RhoA/Rock pathway activated via Gb4/Gai2 and mTORC2 ultimately upregulated Acta2, myosin 11 (Myh11), and tropomyosin (Tpm2, 5, 12) complex which increased the contractility in lungs of both GAS and PM 1 rats. However, the Rac-and Cdc42 pathways were activated only in GAS rats through up-regulation of Cd14, Arhgef, Apc, Acta2 and down-regulation of inhibitory Tmsb4 and Ssh3 proteins which promoted actin polymerization and stabilization and subsequent cell migration. These results suggested an activation of myosin-axis contraction but absence of actin dynamics under the prolonged exposure of tra c-related PM 1 .

Functional validation
Among the discovered DEPs, we selectively validated the expression levels of in ammatory related proteins, C3, Chp1, and Serpina3, in rat lungs by using western blotting. The representative images were shown in Additional le 7: Figure S4. The statistical result in Fig. 6A showed that C3 were signi cantly upregulated in 3M-PM 1 group compared to 3M-GAS group while no signi cant difference was observed in 6M exposure groups. Serpina3 showed signi cantly higher expressions in 3M-PM 1 group and slightly higher expression in 6M-PM 1 groups in comparison to the corresponding GAS groups (Fig. 6B). The inhibitory Chp1 protein exhibited up-regulation in only 3M-GAS group. Additionally, we investigated the oxidative stress and in ammatory pro les within lung tissue by measuring 8-isoprostane and IL-6 using ELISA assays, respectively. The data revealed a higher level of oxidative stress in both 3M-GAS and 3M-PM1 groups while only 6M-GAS group showed higher oxidative stress compared to CTL group (Fig. 6D). Interestingly, the rat lungs exhibited higher level of IL-6 in 3M-PM1 group in comparison to GAS and CTL groups while there was no signi cant difference in IL-6 level between 6M-PM 1 and 6M-GAS groups (Fig.   6E). These results suggested an earlier oxidative stress and in ammatory reaction in 3-month PM 1exposed rats which was in consistent with the proteomics data.

Discussion
Epidemiological studies reported that prolonged exposure of PM is associated with the declined lung function (20,21), yet the detailed molecular mechanism remain unclear. Thus, we aimed to investigate how the tra c-related PM 1 impacts on lungs in our previously reported rat model (16). In the present study, we observed statistically signi cant decline of lung function with decreased FEF  and FEV 20 after 6 months of pulmonary exposure to tra c-related PM 1 . On the other hand, the histological examination of the lung tissues showed a signi cant lung injury under sub-chronic (3-month) exposure of PM 1 , as characterized by increased level of congestion and macrophage in ltration in the lung tissues (Table 1). These results suggested that the presence of PM caused earlier lung damages prior to lung function decline under the pulmonary exposure to tra c-related PM 1 , even though the overall air quality was under the WHO guideline. Other study in rats exposed to biomass fuel and motor-vehicle exhaust pollutants also showed increased leukocyte counts in bronchoalveolar lavage and accumulated in ammatory cells within the airway walls as early as one to three months of exposure, while the signi cant reduction of lung functions were only observed at 7 months of exposure (7).
In order to elucidate the underlying mechanism, we systematically studied the perturbed function and pathways underlying lung injury upon sub-chronic and chronic exposure of tra c-related PM 1 by using MS-based quantitative proteomics analysis (Fig. 3, 4 and Additional le 4-6: Figure S1-3). The functional analysis revealed early dysregulations on lipid, glucose and protein metabolism, acute phase response signaling and complement system which may directly or indirectly contribute to the lung injury under the sub-chronic exposure of PM 1 . While the injury progressed, the metabolic pathways were down-regulated while the tissue repair and wound healing related functions and pathways were activated, especially in rats in the GAS group. Nevertheless, the rats in 6M-PM 1 group showed signi cant lung function decline by decreased FEV 20 and FEF 25-75 which might be due to the failure on triggering wound healing-related pathways. Our data also suggested that the presence of ultra ne-sized particles speci cally enhanced the in ammatory reactions during sub-chronic exposure as well as inhibited the glucose metabolism and actin cytoskeleton signaling after chronic exposure. The temporal and dynamic proteome changes observed in the present study highlighted a complex regulation at the cellular and molecular levels in lungs caused by tra c-related PM 1 pollutants.
Fine and ultra ne PMs have been reported to induce lung diseases through generation of ROS and oxidative stress as well as activation of innate and adaptive immunity, leading to cell barrier and tissue damage (22,23). Several known promoters of in ammatory response, such as nuclear factor-κB (NF-κB), activation protein-1 (AP-1), nuclear factor erythroid 2 related factor 2 (Nrf2), and CREB-binding proteins (CBPs), are activated by oxidative stress (24)(25)(26). In our study, we observed signi cantly higher levels of oxidative stress and IL-6 in together with up-regulation of in ammation-related pathways including acute phase response and complement system in rats exposed to PM 1 for 3 months. As a key component in innate immune system, both pathogen infection and tissue damage would trigger the complement system which further promotes chemotaxis (27), activates neutrophil and macrophage for chemokine secretion (28), and exacerbates acute lung injury through autophagy-mediated alveolar macrophage apoptosis (29). Walters et al. reported that PM 2.5 -treated mice underwent airway hyper-responsiveness resulted from the activation of C3 (30). In adults with age above 65 years in China, short-term exposure to PM 2.5 resulted in a signi cant increase in serum complement C3 and in ammatory reaction (31). The ultra ne-sized PM (PM 0.1 ) enhanced pulmonary in ammation by rapid in ux of neutrophils and proin ammatory cytokine secretion (32). Moreover, smaller size of PM induced higher IL-6 release from A549 cells compared to coarse particles (33), pointing out the threat of smaller particles, which can penetrate deep into the respiratory tract and absorbed by the blood stream (34). These ndings support that the prompt attack of the ne particles in our rat model elevated the oxidative stress in lung and promoted the higher in ltration of immune cells, contributing to activation of acute phase signaling pathways and complement system and thus the signi cant lung injury in 3M-PM 1 rats (Table 1 and Fig. 2C). However, it is noted that proteins involved in complement system and the acute phase response signaling did not show signi cant changes upon chronic PM 1 exposure in our rat model. Thus, we speculated that the acute phase response signaling and complement activation may be the early events mediating pulmonary in ammation upon PM 1 exposure. These early in ammatory responses to PM 1 could be characterized as unique molecular features in sub-chronic stage of lung injury.
The in ltration of PM in lung is known to disrupt the cell membrane integrity and subsequently increase the intracellular calcium ion (Ca 2+ ) concentration. The calcium signaling affects a broad spectrum of cellular functions such as motility, metabolism, cell growth, proliferation and apoptosis (35) and the dysregulation of intracellular Ca 2+ has been reported to associate with PM-induced oxidative stress and in ammation in human lung broblast cells (36), pulmonary artery endothelium cells (37), and mouse lungs (38). Proteins that control e ux and in ux of calcium stand important roles in calcium homeostasis. In cystic brosis, the decreased expression and activity of SERCA were observed to increase susceptibility of airway epithelium cells to oxidant gas exposure and cell death (39,40). In our model, we observed dysregulation of IP3-, SERCA-and calmodulin/calcineurin-mediated calcium signalings. Among the involved DEPs, SERCA (as Atp2a1, sarcoplasmic/endoplasmic reticulum calcium ATPase 1) was inhibited in both 3M-PM 1 and 3M-GAS groups, probably as an early response to TRAP. It was later upregulated in both 6M-PM 1 and 6M-GAS rats in together with the lower expression of Ryr1, which could be an attempt to limit the Ca 2+ leak from sarco/endoplasmic reticulum and balance the elevated cytoplasmic Ca 2+ concentration. The high intracellular Ca 2+ concentration were also reported to promote NFATc signaling and subsequent activation of lymphocyte (41), demonstrating its role as proin ammatory mediator. Despite of the biological signi cance of calcium signaling in lung injury, how these dysregulated calcium signalings coordinately contribute to PM-induced oxidative stress, metabolism, and in ammation remain unclear. More cellular and molecular studies are required to address these issues.
Furthermore, the accumulation of Ca 2+ in the cytoplasm may serve as a signal for initiation of wound healing in the injured rat lungs (42). Wound healing and tissue repair are a multi-step process composed of wound sensing and blocking, plasma membrane restorage, and cytoskeleton remodeling. Although the detailed wound healing mechanism is not clear yet, it is reported to be tightly controlled by Rho family GTPase pathways, including RhoA/Rock, Rac and Cdc42 signalings (43)(44)(45). In our model, the Rhoa/Rock-mediated actomyosin contraction was activated in response to the prolonged exposure to PM 1 by the upregulated Ga4, Itgb1, Cd14 proteins and the increased levels of intracellular Ca 2+ and mTORC2 complex, suggesting a repair initiation for membrane resealing and wound closure. However, the Rac and Cdc42 signalings were not activated in PM 1 rat, making it unable to trigger actin polymerization and stabilization for cytoskeleton remodeling (46) and thus limit the tissue repair potential and ultimately lead to the declined lung function. The concomitant up-regulation of these Rho family GTPase-mediated pathways was exclusively observed in the GAS-exposed rats, highlighting the role of PM in disrupting the repair potential. On the other hand, the progressive upregulation of the actin, myosin, and tropomyosin protein complexes in the PM 1 rats may represent the increased quantity of smooth muscle as observed on the thickening of the airway wall, or a possible brosis formation within the lung tissue (47).
Although not included in Fig. 5, both GAS and PM 1 pollutants stimulated alterations on metabolic pathways, notably on oxidative phosphorylation, lipid (inositol phosphate superpathways) and glucose (glycolysis/gluconeogenesis) metabolism. Increased oxidative phosphorylation was observed in both 3M-PM 1 and 3M-GAS groups which may imply impaired mitochondrial functions, higher oxidative stress (48) and probably a compensation of metabolic shifting from glycolysis to pentose phosphate pathway (49). Glycolysis was reported to worsen infection related-pulmonary brosis (50), and inhibition of glycolysis would attenuate the injury by suppressing in ammation and apoptosis (51,52). Concurrently, our model showed negative regulation of glycolysis/gluconeogenesis in chronic PM 1 -exposed rats, suggesting a potential mechanism to limit lung injury. Regarding the lipid metabolism, a recent metabolomics study reported that an organic component of PM 2.5 , benzo[a]pyrene, would induce lung injury by altering lipid metabolism and phospholipase A2 activity (53). Downregulation of lipid metabolism and upregulation of glucose metabolism mediated by autophagy is also reported to involve in alveolar repair after bleomycin-induced injury (54), while in another study, inhibition of lipid synthesis was shown to exacerbate bleomycin-induced lung brosis (55). These controversial results indicated that the interplay between lipid and glucose metabolisms induced by PM remains not well characterized. The composition of PM pollutants and the exposure times could be potential factors to cause such inconsistency and should be examined in more details.
Overall, we studied the impact of tra c-related PM 1 on lung injury using our previously established rat model for mimicking the dynamic day-to-day exposure on human. We elaborated the lung function and histological changes as well as elucidated the dysregulated molecular mechanisms upon the sub-chronic and chronic exposures of tra c-related PM 1 . Although the tra c-related air pollution did not reach a threatening level as suggested by WHO, we observed a signi cant histological and molecular changes in rats' lungs after three months of tra c-related PM 1 exposure. High levels of oxidative stress and in ammation were con rmed as well. We highlighted dysregulated acute phase response signaling, complement system and intracellular metabolisms as early molecular features in PM 1-exposed rats of which we proposed Serpina3, Atp2a1 and complement proteins as molecular predictors for early PM 1induced lung injury. After 6-months accumulation of PM 1 exposure, the rats showed signi cant lung function decline which might be due to the failure in triggering the actin dynamics-related tissue repair mechanisms. Our study utilized rat lung tissues to unravel the pulmonary phenotypes and molecular alterations, however, future study to detect these candidate proteins in the peripheral blood of high-risk human population should be conducted to evaluate their potential as non-invasive biomarker candidates.

Conclusion
Our study systematically explored the phenotypes and pathogenic mechanisms in rat lungs upon subchronic and chronic tra c-related PM 1 exposures. The in-depth quantitative tissue proteomics analysis explored detailed molecular mechanisms involved in the progression of lung injury which eventually led to disturbance of lung functions. According to these ndings, we proposed several potential proteins associated with early lung damages in response to tra c-related PM 1 which might be used to screen the subject more susceptible to PM 1 air pollution. Rat model with whole-body exposure to tra c related air pollution (TRAP)

Methods
The rat model was previously published in (16) to mimic the TRAP exposure in human. A whole-body exposure system was developed, and the ambient air was continuously sampled by an omnidirectional PM inlet located on the roof of the animal housing followed by exposure into the animal cages. All procedures were performed compliance with the animal and ethics review committee of the Laboratory Animal Center at Taipei Medical University (Taipei, Taiwan).
Each rat was randomly assigned into three groups for exposure within two different periods: (1) three and six months of exposures to whole air from TRAP (3M-PM 1 and 6M-PM 1 groups, respectively); (2) three and six months of exposures to high-e ciency particulate air (HEPA) ltered TRAP (tra c-related gaseous pollutants, shorted as 3M-GAS and 6M-GAS groups, respectively); and (3) three and six months of exposure to conditioned clean air (3M-CTL and 6M-CTL group, respectively). The PM 1 and GAS groups were placed in an urban region nearby a major highway and expressway in New Taipei City, Taiwan. The CTL group was housed in a speci c pathogen free I level of Laboratory Animal Center (Taipei, Taiwan).
As indicated in (16)

Lung function examination
After the designated time period of exposure, each rat from every group was examined for the lung function by using a Forced Pulmonary Maneuver System (Buxco Research Systems, Wilmington, NC, USA) following the manufacturer's protocols. Brie y, an endotracheal tube was attached to an airway port, and the distal end of the catheter was passed through a small opening located near the airway port.
The catheter was then connected to the system. Forced expiratory ow at 25-75% of the pulmonary volume (FEF  ) and forced expiratory volume at 20 ms (FEV 20 ) were measured. For each test, at least three acceptable measurements were conducted to obtain a reliable mean for all numeric parameters. All procedures were performed under su cient anesthesia.

Histological evaluation
Lung tissue were collected and washed with ice-cold PBS, followed by xation with 10% neutral buffered formalin, embedded in para n, sectioned, and stained with hematoxylin and eosin (H&E). Histological examinations were conducted under light microscopy by a histopathologist in a blinded manner. The degree of lung injury was scored according to the four criteria: (1) alveolar congestion, (2) hemorrhage, (3) immune cell in ltration, and (4) thickness of the alveolar wall (56). Congestion and thickness of the alveolar wall were graded by a ve-points scale as follows: 0 for minimal (little) damage, 1 for mild damage, 2 for moderate damage, 3 for severe damage, and 4 for maximal damage. Hemorrhage was graded as follows: 0 for no red blood cells (RBC) outside of blood vessels, 1 for few interstitial RBC, 2 for few RBC in some alveoli, 3 for moderate number of RBC in some alveoli, 4 for many RBC in most alveoli, 5 for large numbers of RBC in all alveoli. In ltration of macrophage was graded as follows: 0 for nonerare, 1 for 1-10% of alveoli/saccules contain macrophages, 2 for 10-25%, 3 for 25-75%, 4 for >75% (57).

Tissue lysate collection
Lung tissue from each rat was grounded in liquid nitrogen, collected into an microcentrifuge tube, and lysed with Cellytic MT cell lysis reagent, protease inhibitors, and EDTA in the volume ratio of 98:1:1. The tissue lysate was homogenized by using Minilys® personal homogenizer (Bertin, Rockville, MD, USA) in high speed mode for 15 s twice and the homogenate was centrifuged at 13,000 rpm at 4°C for 10 min to collect the clear supernatant as lung tissue lysate. The lung tissue lysate from each rat was assayed by using BCA protein assay kit to determine the protein concentration.
Gel-assisted digestion, tandem mass tag (TMT) labeling and high pH reversed phage (RP) StageTip fractionation Fifty micrograms of lung tissue proteins were aliquoted from each of the 5 rats from the 6 exposure groups for our reported gel-assisted digestion with trypsin individually (58). The resulting peptides were vacuum-dried and resuspended in 100 mM TEABC for BCA protein assay. Ten micrograms of peptides were aliquoted from each of the 5 rats in one exposure group to generate a pooled peptide sample . The   pooled peptides from 3M-CTL, 3M-PM 1 and 3M-GAS groups were labeled with TMT126, TMT127 and   TMT128 respectively, whereas the 6M-CTL, 6M-PM 1 and 6M-GAS were labeled with TMT129, TMT130 and TMT131 respectively. The TMT-labeled peptides from each group were pooled for RP StageTip fractionation following the protocol in (59). Peptides were eluted sequentially by using 11.1%, 14.5%, peptide and protein levels for con dent identi cation. Identi ed peptides in high con dence with at least 6 amino acids were accepted. For proteome quanti cation, only unique peptides were included to estimate the protein abundance which was further normalized by the total peptide abundance. Proteins with 1.3-fold changes in abundance (log2 ratio >0.38 or <-0.38) were considered as differentially expressed proteins (DEPs).

Functional analysis
The DEPs from each exposure group were submitted to Ingenuity Pathway Analysis (IPA) (60) for pathway enrichment analysis and Cytoscape (version 3.

Declarations
Ethics approval and consent to participate Animals were treated humanely and all procedures were performed compliance with the animal and ethics review committee of the Laboratory Animal Center at Taipei Medical University (Taipei, Taiwan).

Consent for publication
Not applicable Availability of data and materials The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests   Tables   Due to technical limitations Table 1 is available as a download in the Supplementary Files. Figure 1 Lung function examination. Forced expiratory ow at 25-75% of the pulmonary volume (FEF25-75) and forced expiratory volume at 20 ms (FEV20) were measured for each rat with at least three acceptable measurements. Signi cant reduction of FEF25-75 and FEV20 were observed in rats from 6M-PM1 group. * p<0.05     Functional validation of selected proteins, oxidative stress and IL-6 in lung tissues. A, B and C showed the statistical results of western blot analysis of C3, Serpina3, and Chp1 proteins in lung tissues, respectively.

Figures
The expression of each protein was normalized to the control reference running in every western blot analysis. D and E showed the ELISA results of 8-isoprostane and IL-6, representing the status of oxidative stress and in ammation in lung tissues respectively. The quantities of 8-isoprostane and IL-6 were normalized to total protein lysate before comparison. * p<0.05.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. Table1.pdf OnlineSupplementaryMaterialsall.pdf