Assessment on the lung injury of mice posed by airborne PM2.5 collected from developing area in China and associated molecular mechanisms: from histopathology to integrated transcriptome analysis

Some epidemiological investigations have revealed airborne ne particulate matter (PM 2.5 ) could induce adverse effects on respiratory system of human. However, the experimental evidences of the harmful effect of PM 2.5 from mid-scale city in China and associated molecular mechanisms was still scarce. In this study, we aimed to evaluate the adverse effect on the lung system of PM 2.5 collected from mid-city of China and elucidate the underlying molecular mechanism through mRNA-seq and microRNA-seq integrated analysis. We exposed male mice for 8 weeks (298.52 ± ) using a whole-body exposure system. Micro-CT and histopathological analyses were performed to determine the morphology and histopathology changes of lung tissues induced by PM 2.5 . Transcriptome (both mRNA and microRNA) sequencing and the immunohistochemistry assay were performed to reveal the associated underlying mechanisms. The contents of PAHs absorbed in the PM 2.5 , as well as the pearson correlation index between it and the target genes and microRNA were examined.

asthma [5], and chronic obstructive pulmonary disease (COPD) [6]. Otherwise, patients with lung-diseases would be two or three times the risk of death than normal population [7]. Hence, World Health Organization cites PM 2.5 pollution as fth leading risk factor for mortality and economic loss. In China, ambient PM 2.5 was responsible for 1.2 million premature death as the forth healthy risk factor [8,9]. From the public health view of point, it was imperative to comprehensively elucidate the causal relationship between PM 2.5 pollution and lung injury.
There have been several publications about the adverse effects of PM 2.5 to respiratory system using mice as model animal with oropharyngeal aspiration (OPA) or intranasal administration. Balb/c mice exposed to PM 2.5 (2.5-20 µg for each mice) through OPA for 21 days showed obvious collagen deposition around small airway [10].Acute OPA of PM 2.5 extract to mice may induce greater lung neutrophilia and in ammation [11]. C57BL/6 mice exposed to PM 2.5 (3 mg/kg) by OPA can adversely affect prenatal lung development in the offspring [12]. C57BL/6J mice showed signi cant in ammation and incipient brosis symptom after directly intranasal administration with PM 2.5 (100 µg/day) for 4 weeks [13]. The obvious brous cap thickness observed from apoe-/-mice exposed to PM 2.5 by intranasal instillation [14]. Generally, direct administration by PM 2.5 suspension may under-or overestimate the healthy risk. There were only few investigations about the wholebody inhalational exposure. Yuan et al. (2020) found that C57BL/6 mice whole-body exposed to PM 2.5 (59.77 µg/m 3 ) sampled from Beijing, China exhibited severe lung injury and brosis [11]. They also demonstrated that treatment with local PM 2.5 polluted air at Beijing, China would strikingly induce lung oxidative stress and injury in mice for 3 weeks and 6 months, respectively [15,16]. Zhou et al. (2019) exposed C57BL/6 mice to PM 2.5 collected from Shijiazhuang, China, and found that signi cant increase in circulating white blood cells and in ammation in lungs of mice from exposure groups [17]. All the investigations were preformed to estimate the adverse effects on lung by PM 2.5 in Beijing-Tianjin-Hebei region, China. The underlying mechanism have not been comprehensively elucidated.
Evidences have unveiled that air pollution was much more serious in economically developing levels cities than the economically developed mega-cities in China [18][19][20]. However, the assessment on the healthy risk of PM 2.5 pollution in this mid-scale cities was scarce. An increase of 10 µg/m 3 PM 2.5 was associated with 15-27% increase in lung cancer mortality [21]. Moreover, the annual mean concentration of PM 2.5 in mid-scale cities in China was marked higher than the limitation suggested by WHO as 10 µg/m 3 [19,22]. Hence, its warranted to investigate the consequences posed by PM 2.5 in this region to pulmonary system using "real-world" exposure system.
High-throughput sequencing is a precise method to measure global gene expression pro les by RNAsequencing including mRNA and microRNA. Recently, this tool has been used to illuminate the associated molecular mechanism of cytotoxicity induced by PM 2.5 [23,24]. There was only one publication about the potential mechanism of lung injury posed by airborne PM 2.5 at Beijing, China using transcriptome analysis [25].
Meanwhile, some evidences suggested that microRNA deregulation was likely to be associated with the respiratory diseased induced by PM [26,27]. Hence, the integrated analysis between mRNA and microRNA to elucidate the underlying mechanisms of the lung injury posed by PM 2.5 was imperative.
In summary, we for the rst time established a whole-body exposure system to investigate the adverse effects posed by PM 2.5 sampled from Baoji (33.35°-35.06° N, 106.18°-108.03° E), an inland city located in the mid-west part of China, using mice for 40 days. The objectives are to evaluate: (1) histopathological changes of lung tissue; (2) morphological changes of lung; (3) underlying mechanisms using mRNA-seq, microRNA-seq, and immunohistochemical analysis; (4) the potential biomarkers to lung injury posed by PM 2.5 ; (4) the dominant components in PM 2.5 associated with lung injury.

PM 2.5 -induced lung injury in mice
Micro-CT were performed to determine the potential effect on the lung morphology induced by PM 2.5 in vivo after the exposure before the histopathology assay. The tomograms showed higher density and con uent opacity of the mice lung tissues from exposure groups ( Figure. 1d and e), indicating the remarkable pulmonary in ammation induced by PM 2.5 , while no morphological changes were observed in the control ( Figure. 1a and   b). Moreover, we created the 3D reconstruction of the lung tissue of the mice to estimate the pulmonary function injury posed by PM 2.5 . As shown in Figure.1c and f, the effective pulmonary function from exposure group markedly shrunk compared to the control group. The same results about the higher lung density and deteriorated PF were found by previous publication through intranasally administrating [10,13,28]. However, the adverse effects found in this work were much more obvious than the previous results, while the exposure concentration (289.52 µg/m 3 ) was much lower that it (equal to 400-1543 µg/m 3 ) in the previous work [10,13], indicating exposure way may markedly impact the assessment on the toxicity of PM 2.5 on the respiratory system. Meanwhile, the PM 2.5 particles used previously were obtained from Shanghai, China, one of the megacities located in Yangtze River Delta, which suggested that the hazard level of the PM 2.5 collected from the developing area in this work may be higher than the developed city.
As shown in Figure. 2, obvious alveolar intervals thicken, in ammatory cell in ltration, and alveolar structure damage ( Figure. 2b, e, h) were observed from the pulmonary alveoli from exposure group compared with the vacuolated thin-walled of alveolar cavity as the intact structure shown in the control ( Figure. 2a, d, g). The marked bronchiolar epithelial hyperplasia and endoluminal in ammatory cells in ltration were shown in the lung of PM 2.5 -treated mice ( Figure. 2c, f, i), while simple ciliated columnar epithelium of bronchia in lung was characterized as normal pulmonary structure observed in control mice ( Figure. 2a, d, g). The results found in this work were consistent to the previous investigations [15,17,27,29]. Noted that the pulmonary histopathological injury was irreversible, the long-term exposure to PM 2.5 may pose severe threat on the respiratory system to human beings since the pollution is persistent. Since the prevalence of pulmonary brosis (PF), a critical lung interstitial disease, was signi cantly associated with long-term exposure to PM 2.5 [1], we determined lung brosis in mice from control and exposure groups using Masson trichrome stain. Obvious lung brosis increase was observed from exposure group ( Figure. 2k) compared with the control (Figure. 2j), and the statistical results showed 2.56-fold (p < 0.05) increase between the exposure and control group ( Figure. 2l).
To further elucidate the immune response activated through PM 2.5 exposure in mice respiratory system, the relative abundances of leucocytes CD45+, as well as the epithelial cells (CD45-/CD31-/CD326+) in lung tissue of mice were determined using ow cytometry analysis. As shown in Figure. 3, the relative abundance of leucocytes in exposure group were signi cantly higher than it in the control ( Figure. 3a-c). However, there was no statistical difference of the epithelial cell abundances between the PM 2.5 -treated and control groups ( Figure. 3d-f). The results indicated that immune system response may be the dominate mechanism for the effects by PM 2.5 , which was parallel to previous works [30,31]. However, the bronchiolar epithelial hyperplasia cannot be manifested using the relative abundance of the cells, which may ascribe to the interference of the overwhelming quantity advantage of the leucocytes. Further experiments (e.g., single-cell sequencing) will be performed to elucidate the underlying mechanism of the difference occurred in epithelial cells.

PM 2.5 -induced changes of gene expression pro le in mice lung
To investigate the underlying molecular mechanism of the lung injury posed by PM 2.5 , RNA-sequencing was performed to analyze the whole-genome expression pro ling changes of the mice lung tissues with or without exposed to PM 2.5 . As shown in the Figure (Table S3), which was parallel to the previous study about the mice exposed to PM 2.5 collected from Beijing, China [25]. Meanwhile, the top 10 up-and downregulated genes were listed in Table S4.

PM 2.5 -induced changes of microRNA expression pro le in mice lung
According to the microRNA-sequencing analysis, a total of 70 microRNA (53 up-regulation and 17 downregulation) were identi ed in comparison of control and exposure groups, and the fold changes were shown in Figure. S5. As shown in Table S5, the top 20 KEGG pathway enrichment determined by KEGG annotation including GnRH signaling pathway, HIF-1 signaling pathway, and MAPK signaling pathway were identi ed. The similar results of the KEGG pathway (i.e., HIF-1 signaling pathway, Insulin signaling pathway, and MAPK signaling pathway) were found by previous investigation about the PM 2.5 collected from Shijiazhuang, Hebei, China using microRNA microarray analysis [27]. The top 10 up-and down-regulated microRNA were listed in Table S6.

Integrated analysis between the gene expression and microRNA
To further elucidate the molecular mechanisms, the integrated analysis of mRNA and microRNA expression data were performed. A total of 3024 mRNA (1625 up-regulated and 1399 down-regulated) and 100 microRNA (59 up-regulated and 41 down-regulated) were identi ed in comparison of control and exposure groups (Figure.  S6b and Table S7), which was consistent to the KEGG annotations associated with mRNA expression levels, but was robuster through integrated analysis. GO annotations found that the biology process (BP) including immune system process, regulation of biological quality, and regulation of immune system process were signi cantly enriched ( Figure. S6c).
To further validate the causal relationship between the enriched KEGG pathway and the lung injury posed by PM 2.5 . Nine representative pathways involved into the immune system, including B cell receptor signaling pathway, cell adhesion molecules (CAMs), antigen processing and presentation, were chosen ( Figure. 4a). The input genes associated with the enriched KEGG pathway annotated by integrated analysis were showed in Table S7. As showed in Figure. 4b, we selected 27 genes to explore the mechanism for the pathway in PM 2.5exposed lung tissue. Among these genes (showed in Table 1), Cd72, Cd81, Cd19, Pik3cd, and Ppp3cc were involved into B cell receptor signaling pathway, while Cd80, Cd22, H2-M2, H2-T24, and H2-T3 were involved into cell adhesion molecules (CAMs). Tapbp and Klrc1, Prf1, and Jak3 and Tnfrsf13c were involved into antigen processing and presentation, graft-versus-host disease, and primary immunode ciency, respectively. Angpt2, Pdgfra, P2ry1, and Ig were associated with Rap1 signaling pathway. Ptpn7 and Fgf22 were related to MAPK signaling pathway. Rasal1 and Lamc1 were involved into Ras signaling pathway and small cell lung cancer, respectively. Gene Met, Ralgds, and Cd8b1 were identi ed involved into multiple pathways. The associated microRNAs (total of 30) were mapped in Figure.4b and Table 1. The RT-qPCR analysis was performed to validate the results from transcriptomics. As showed in Table 1, most of the genes and microRNAs we selected showed similar results to those of the transcriptomics. Gene Cd72, Cd19, Pik3cd involved into B cell receptor signaling pathway were signi cantly up-regulated from exposure group, while gene Cd81 was markedly down-regulated ( Figure. 4c). The mRNA transcript levels of gene Cd22 and H2-M2 related to CAMs strikingly increased, and the levels of H2-T24 decreased markedly ( Figure. 4c). In addition, the associated microRNA involved into B cell receptor signaling pathway including 203-5p, 7a-5p, and 92a-1-5p were signi cantly up-regulated in PM 2.5 -exposed group. The obvious down-regulations of microRNA 149-5p, 328-3p, 466i-5p, and 24-3p were observed in exposure group ( Figure. 4d). Among the CAMs, we found that the microRNA 674-3p, 486a-3p, and 1247-5p increased remarkably, while 486a-3p was signi cantly downregulated in exposure group ( Figure. 4d).
The protein expression of CD19, CD81, and PIK3CD involved into the B cell receptor signaling pathway was examined to con rm the changes of the associated genes by immunohistochemistry. As showed in Figure.

Discussion
Some evidences have suggested that the adverse effects on lung tissue posed by PM 2.5 maybe associated with the immune and in ammatory response [11,15,32]. Hence, B cell receptor signaling pathway, as the important pathway responsible for immune response [33], enriched by integrated analysis was foreseeable, which was also illuminated by previous study using RNA-sequencing [25]. The transcriptional levels of the target genes involved in this pathway were also signi cantly up-or down-regulated by PM 2.5 . Among them, Cd72, type II transmembrane protein coding gene belonging to C-type lectin family [34], was elucidated play an important role in controlling the magnitude of B cell responses [35], and then responsible for the immune system homeostasis regulation [36]. The overexpression of Cd72 found in this work from PM 2.5 -exposed group may promote B cell survival and proliferation, and enhance release of CD23, and then activate the immune response [37][38][39]. In addition, anti-CD72 monoclonal antibodies (mAbs) can activate CD19 tyrosine phosphorylation [40]. The similar result was observed in this work (the up-regulation of Cd19 induced by PM 2.5 ).
It was reported that mutation of gene pik3cd was associated with the prevalence of systemic lupus erythematosus, which is a typical immune system disease [41]. Hence, the marked increase of the mRNA level of pik3cd in lung tissue from exposure group might related to the immune response induced by PM 2.5 .
Meanwhile, the associate microRNA of Cd72 149-5p and 328-3p were signi cantly down-regulated by PM 2.5 . It has reported that the two microRNA may be responsible for the lung in ammatory and brotic pathology in mice [42]. The microRNA (203-5p and 7a-5p) related to Pik3cd were markedly up-regulated in mice lung from exposure group. Based on the KEGG and GO annotations, the microRNA 203-5p was involved into the B cell receptor signaling pathway ( Figure. S7a). Therefore, the increase of the expression levels of gene Cd72, Cd19, and Pik3cd involved into B cell receptor signaling pathway may contribute to the lung injury including in ammation and brosis induced by PM 2.5 .
Surprisingly, we rstly found the most signi cantly enriched KEGG pathway of DEGs through integrated analysis was cell adhesion molecules (CAMs), which was identi ed to mediate the process of cell recruitment and homing, and play an important role in the in ammatory process [43]. Once the in ammatory process of pulmonary immune system was triggered by PM 2.5 , the up-regulation of adhesion molecules genes occurred in endothelial and immune cells to mediate leukocyte adhesion and then migrating to in ammation sites [44]. From the results of RT-qPCR quanti cation and transcriptomics analysis, the signi cant upswing of the expression levels of genes Cd22 and H2-M2 were observed from PM 2.5 -treated group. The gene Cd22, a immunoglobulin superfamily cell-surface molecule that serves as an adhesion receptor for sialic acid-bearing ligands [45], was elucidated to active B cells and regulate antigen receptor signaling in vitro [46]. Similarly, the associated microRNA 3110-5p for Cd22 were signi cantly stimulated from exposure group. Thus, our nding indicated that CAMs may be a key pathway responsible for the adverse effects posed by PM 2.5 to respiratory system. According to the KEGG and GO annotations ( Figure. S7b), the genes Cd22 and H2-M2 can be used as feasible biomarkers.
Furthermore, the target genes involved in the insigni cant (P > 0.05) enriched KEGG pathways (i.e., antigen processing and presentation, graft-versus-host disease, primary immunode ciency, rap1 signaling pathway, MAPK signaling pathway, ras signaling pathway, small cell lung cancer, and PI3K-Akt signaling pathway) were also signi cantly up-or down-regulated by PM 2.5 , suggesting these pathways related to immune response of lung tissue should also be concerned for the further investigation.
A number of investigations have demonstrated that long-term exposure to PAHs increased risk of developing lung cancer [47][48][49]. CCA was performed using concentrations of PAHs as environmental factor. As showed in PAHs mixtures [50]. DBP has been evaluated 100 times carcinogenic potency than BaP [50]. The potency value of dibenzo[a, h]anthracene (DBA), which was estimated to be associated with human cancer [51], was reported to be up to 10 times than Bap [52]. To further explore the covariation between the PAHs and genetic indicators, Pearson's index was determined. Obvious covariable relationship was observed between the target genes including Cd72, Cd19, Pik3cd, Cd22, H2-M24, and associated microRNA, and the PAHs detected from PM 2.5 ( Figure. S9, p < 0.05). This result suggested the target genetic indicators we selected showed a signi cant covariation with environmental pollutants, and then they may be used as the biomarkers to indicate the healthy risk posed by these types of pollutants.
From the public health point of view, PM 2.5 samples used in this work were collected from Baoji, the mid-scale city located at the developing area. The main components of PAHs detected from the PM 2.5 were similar to the mega-city (e.g., Beijing, Nanjing, et.al), while the concentrations of them, especially for Bap, DBA, DBP were statistically higher than the developed area [53][54][55]. Noted the concentrations of PM 2.5 in all regions in China were very high than the threshold proposed by WHO (35 µg/mL) or more permissive limits adopted by China (75 µg/mL) in spite of the general decrease trends observed recent years [56]. Meanwhile, energy structure was much different between the mid-scale city and the mega-city in China [57], which directly result in the variation of the occurrence levels of the organic pollutants (PAHs) derived from the fossil fuel combustion [58]. Combined with the obviously adverse effects of the PM 2.5 collected from Baoji city on the respiratory system found in this work, the healthy risk posed by PM 2.5 at the mid-scale city of China should be paid more attention.

Conclusion
This work for the rst time estimated the lung injury posed by PM 2.5 collected from Baoji, a representative midscale city of China. We also elucidated the underlying mechanisms through integratedly analyzing the pro les of mRNA-seq and microRNA-sEq. Obvious lung injuries including pulmonary dysfunction, in ammatory response, pulmonary brosis were observed in the lung tissues of mice from exposure group. As revealed by KEGG annotations, the main pathway induced by PM 2.5 was immune system associated pathway, especially B cell receptor signaling pathway and cell adhesion molecules (CAMs). Moreover, the expression levels of the key genes and microRNAs involved into the pathway, as well as the associated protein expression were veri ed.
The results from this work may provide deeper insight into the mechanisms for the pulmonary toxicity posed by PM 2.5 in the developing area. The more adverse effects on lung tissue posed by PM 2.5 from mid-scale city of China than it from the mega city suggested the potential health risk of PM 2.5 in developing area should be more concerned.

PM 2.5 particle collection and analysis
Ambient PM 2.5 were collected on quartz lter (2 µm pore size) by a High-Volume Ultra ne Particle Sampler (Ju Kang Technology, China) at a 5 L/min ow rate from Baoji, shaanxi, China, maintained for 24 h. PM 2.5 was extracted described previously [59]. Brie y, the lters were chopped into small fragments, and then shaken for 20 min after treated by ultrasonic in ice for 30 min in a 50-mL centrifuge tube. The process was repeated for three times. The mixture was centrifuged at 8000 rpm for 10 min, and then the precipitate was transplanted into a new tube and dried by lyophilization. The powder was stored at -80 ℃ until experiment. The concentrations of seventeen EPA priority polycyclic aromatic hydrocarbons (PAHs) bonding on the PM 2.5 were determined by gas chromatography coupled with a mass spectrometer (GC-MS, Agilent GC 6890, MS 5973, USA) following the previous publication [60]. The details were shown in supporting information. Instrumental and method detection limits were showed in Table S1.

Animals and whole-body inhalation
Male Balb/C mice at 8 weeks of age with body mass of 20 ± 2 g were purchased from Chengdu Dashuo Company (Chengdu, Sichuan, China) and acclimated for a week before exposure in SPF level animal house of West China Hospital. Subsequently, twenty mice were divided randomly into two groups (n = 10 for each group) exposed to either ltered air (control) or concentrated PM 2.5 air (exposure) in a "real-world" exposure system for  Figure.S2. According to the criteria provided previously, the daily ventilation volume of adult male Balb/C mice is 0.0864 m 3 [61]. Previous investigation has revealed that 75% of PM 2.5 particles would enter and touch alveoli through inhalation [62]. Hence, the daily exposure dose of PM 2.5 for one mouse was: 298.52 µg/m 3  Lung tissues preparation and histopathology assay After exposure, the mice were anesthetized by pentobarbital sodium. Blood samples were collected from the cardiac vein. Serum was obtained by centrifuged at 3000 rpm for 15 min with no anticoagulant added to detect the routine blood indexes. The lung tissue of each mice was immediately separated. After washed by precooled phosphate buffered saline (PBS) for 3 times, the left lung tissues were divided into two parts. One part was used for ow cytometry analysis, and the other was xed with 10% formaldehyde solution immediately and then embedded in para n. Para n-embedded tissues (60 ℃) were sectioned consecutively with a thickness of 5 µm. The slices were then depara nized and stained with hematoxylin and eosin (H&E) or Masson trichrome. More than ve replicates per lung tissue each mouse were observed with an optical microscope to determine the histopathology from exposure group. The right lung tissues were kept at -80℃ for further analysis.

Flow cytometry analysis
The ow cytometry analysis was performed as previously described. Brie y, one part of the left lung tissues were minced on ice, and then digested by 0.1% Type I and Type Collagenase in 4 ml of the PBS system, in 37 ℃ in the incubator, shaking incubation for 1 h. After digestion completely, cell suspension was ltered through The paired-end RNA reads were aligned to the reference genome using Hisat2 v2.0.5 software. The gene expression level was quanti ed by feature Counts v1.5.0-p3. The microRNA reads were quality controlled using miRBase20.0 as reference, modi ed by software miredeep2. The potential microRNA secondary structures were obtained using srna-tools-cli. MicroRNA expression levels were estimated by TPM (transcript per million).
Differential expressions of RNA and microRNA between exposure and control group were analyzed using the

Quantitative real-time PCR analysis
To verify the data from the transcriptomics, 27 genes and 29 microRNA were chosen as candidates for RT-qPCR analysis. Primer sequences were listed in Table S2. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and β-actin genes were used as the endogenous control to monitor the quality of the target genes. MicroRNA U6 was used as the internal standard to monitor the quality of the target microRNA. The variations in the endogenous control expression were below 10% for all groups. Ampli cation was performed using SYBR Green PCR master mix (Applied Biosystems) according to the manufacturer's instructions. Gene and microRNA expressions were quanti ed using the 2 −△△ Ct method suggested by Applied Biosystems (Foster City, CA, USA).
The fold change between the exposure group and the control were calculated by the geometric mean of the relative expression normalized by the two housekeeping genes (β-Actin and GAPDH) [63].

Statistical analysis
The statistical program SPSS 18.0 (Chicago, IL, USA) was used to analyze all the collected data. Five or more replicates of each parameter were determined to eliminate the variability of the results. All the data were expressed as mean ± standard deviations (S.D). Two-tail student's T-test (95% con dence interval) was used to examine the signi cance of differences between control and exposure groups.

Consent for publication
Not applicable.
Not applicable.Availability of data and materials Not applicable.