Comparative proteomic analysis of silica-induced pulmonary fibrosis in rats based on tandem mass tag (TMT) quantitation technology

Silicosis is a systemic disease characterized by chronic persistent inflammation and incurable pulmonary fibrosis with the underlying molecular mechanisms to be fully elucidated. In this study, we employed tandem mass tag (TMT) based on quantitative proteomics technology to detect differentially expressed proteins (DEPs) in lung tissues of silica-exposed rats. A total of 285 DEPs (145 upregulated and 140 downregulated) were identified. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses were performed to predict the biological pathway and functional classification of the proteins. Results showed that these DEPs were mainly enriched in the phagosome, lysosome function, complement and the coagulation cascade, glutathione metabolism, focal adhesion and ECM-receptor interactions. To validate the proteomics data, we selected and analyzed the expression trends of six proteins including CD14, PSAP, GM2A, COL1A1, ITGA8 and CLDN5 using parallel reaction monitoring (PRM). The consistent result between PRM and TMT indicated the reliability of our proteomic data. These findings will help to reveal the pathogenesis of silicosis and provide potential therapeutic targets. Data are available via ProteomeXchange with identifier PXD020625.


Introduction
Silicosis represents a pulmonary fibrosis disease caused by long-term inhalation of free silica dust, primarily produced in the mining and construction industries [1]. It is the most serious occupational lung disease especially in developing countries [2,3], affecting the quality of life of individuals. The main characteristic histopathological feature of silicosis is the build-up and accumulation of fibrosing nodular lesions with progressive massive fibrosis and gradual loss of respiratory functions [4,5]. The disease is characterized by fatal, irreversible, incurable signs (n = 10) intratracheally instilled with 1.0 mL normal saline. All rats were sacrificed on day 28 after treatment.

Histopathologic examination
The right lower lungs from all rats were isolated and fixed with 4% paraformaldehyde for 24 hours, dehydrated in a series of graded ethanol solutions, then embedded in paraffin. Serial paraffin sections were cut in 4 μm thick. Subsequently, sections were stained with hematoxylin-eosin (HE) and masson trichrome to evaluate the histopathological changes in lung. The score of alveolitis and pulmonary fibrosis was determined as previously described [17]. The left lungs were preserved immediately snap-frozen in liquid nitrogen and stored at −80˚C.

Extraction of total protein from lung tissue
Based on the results of the histological examination, three lungs were selected from each group for quantitative proteomic analysis. The procedures for protein preparations were according to previous papers [18]. Briefly, 300 μL lysis buffer SDS (P0013G, Beyotime Biotechnology, China) and 1 mM protease inhibitor Phenylmethanesulfonyl fluoride (PMSF, PB0425-5G, Amresco, USA) were added to the frozen sample, followed by ultrasonic treatment (1 s/1 s intervals, 80 W) on ice for 3 min and centrifugation (12,000 g, 4˚C) for 10 min. The supernatant was collected, packaged and frozen at −80˚C. The protein concentrations were assayed by the method of BCA (23227, Thermo, USA) according to the manufacturer's instructions. Next, 12% SDS-PAGE (17-1313-01, Sinopharm, China) was applied to separate 10 μg protein from each sample. The corresponding protein bands were observed by Coomassie Blue R-250 staining to conform the quality of proteins.

Protein digestion and TMT labeling
Protein digestion was carried out as previously described [19]. After protein quantification, 100 μg of protein was incubated with 120 μL reduction buffer 10 mM DL-Dithiothreitol (DTT,  A620058-

Mass spectrometry data and bioinformatic analysis
Proteome Discoverer (v.2.2, Thermo, America) was used to search all of the Q Exactive raw data thoroughly against the UniProt database (https://www.uniprot.org/). Various search parameters were set: a peptide mass tolerance of ±10 ppm, variable modifcations of oxidation (M), a fragment mass tolerance of 0.02 Da, decoy as the database pattern and a peptide false discovery rate (FDR) of �0.01. For protein quantization, the protein was required to contain at least one unique peptide. The quantitative protein ratios were weighted and normalized by the median ratio in Mascot [20,21]. For three biological replicates, the ratio of mean expression between model and control was defined as fold changes (FC). Those proteins with significant differences between control and model groups are considered DEPs. The DEPs were performed using the Database for Annotation Visualization and Integrated Discovery (DAVID). The biological process (BP), cellular component (CC) and molecular function (MF) were annotated by the Gene Ontology (GO) database. The signaling pathways of proteins were elucidated by searching against the Kyoto Encyclopedia of Genes and Genomes (KEGG) database. The protein-protein interaction (PPI) of selected proteins was analyzed by Search Tool for the Retrieval of interacting Genes/proteins (STRING) Software.

Verification of protein expression levels by PRM
The candidate proteins were verified by PRM on a Q Exactive mass spectrometer (Thermo Scientific) combined with Easy-nLC system1200 (Thermo Scientific). The lung tissues used for RPM validation were same to the TMT analysis and the peptides were prepared according to TMT. Tryptic peptides of each sample were spiked with the equivalent heavy isotope AQUA peptide (an internal standard) [22] and loaded onto a C18 column (75 μm × 15 cm, C18, 3 μm, 120 A, hromXP Eksigent, America). The full MS scan from 350 to 1650 m/z was acquired with an orbitrap resolution of 30000 (at m/z 200), AGC value was set to 3e6 and 200 ms maximum ion injection time (IT). Ion activation and dissociation was performed at normalized a collision energy of 27 in HCD collision cells. Following this step, 20 MS2 scans (target precursor ions) were performed and orbitrap resolution was set to 30000 (at m/z 200), isolation window was set to 1.6Th. ACG target value was set to 3e6 and maximum IT was set to 120 ms. The raw data were analyzed using Skyline (MacCoss Lab, University of Washington) software (V.4.2).

Data sharing
The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the iProX partner repository [23] with the dataset identifier PXD020625.

Statistical analysis
Statistical analysis was performed with Statistical Program for Social Sciences (SPSS) (SPSS Inc., version 20.0, United States). The quantitative data were reported as the means ± Standard Deviation (SD), and the significant difference was analyzed with t-test between two groups, P values <0.05 was considered statistically significant. In TMT proteins with p values <0.05 and fold changes �±1.8 were considered as DEPs. A multiple testing correction was performed using Benjamini and Hochberg procedure to control the False Discovery Rate (FDR), using P value (<5%) [24]. GO and KEGG analyses were carried out using Fisher's exact test, using the entire quantifed protein annotations as the background dataset. Only categories and pathways with p values <0.05 were considered statistically significant.

Histopathological evaluation of lung tissue
Lung architecture was normal in sections of lungs from the control group (Fig 1A and 1C). HE staining showed thickened alveolar walls, damaged alveolar structures, more infiltrating inflammatory cells and silicotic nodules (Fig 1B), and Masson staining showed damaged alveolar septa, diffuse silicon nodules, more collagen fibers and inflammatory cell infiltration in model group (Fig 1D). Lungs of model group showed a marked increase in scores of alveolitis and pulmonary fibrosis compared with the control group (P<0.05) ( Table 1).

Protein identification and differential expression
A total of 3099 proteins were identified (S1 and S2 Tables), of which 285 DEPs (145 upregulated and 140 downregulated) were identified between the control and model groups respectively (Fig 2A, S3 and S4 Tables). Heat maps were generated using these 285 DEPs (Fig 2B). The relative expression levels are shown by the intensity of the color. Red, green, or black colors indicate relative increase, decrease, or no quantitative information regarding protein content for a particular protein.

PRM validation of the protein expression
Six significantly changed proteins, including CD14 (UniProt identifier Q63691), PSAP (Uni-Prot identifier P10960), GM2A (UniProt identifier Q6IN37), COL1A1 (UniProt identifier P02454), CLDN5 (UniProt identifier Q9JKD6) and ITAG8 (UniProt identifier B5DEG1) were examined by PRM. These proteins have larger FC value and potentially important biological functions related to inflammation or fibrosis. The results showed that the expression levels of CD14、PSAP、COL1A1 and GM2A were all increased, and ITAG8 and CLDN5 were decreased in the model group than those in the control group. This was exactly the same trend as that observed when the protein levels were quantified by TMT (Fig 3).

Bioinformatic analysis
Using DAVID software to initially explore the potential functions of those DEPs in silicosis. The colors of the bar charts represent the top ten terms of the three different categories (Fig 4). For BP, response to external stimulus, immune system process and cellular response to chemical stimulus were the top three significantly enriched terms (blue in Fig 4). For CC, extracellular region part, extracellular region and extracellular vesicle were the top three significantly enriched terms (red in Fig 4). For MF, protein binding, cell adhesion molecule binding and lipid binding were the top three significantly enriched terms (yellow in Fig 4).
Data from KEGG pathway analysis indicated that DEPs between the groups were involved in 34 pathways, Fig 5 shows the top twenty terms. The results showed that the DEPs were enriched in phagosome (14 proteins), lysosome (8 proteins), leukocyte transendothelial migration (8 proteins), cell adhesion molecules (CAMs) (8 proteins) and focal adhesion (7 proteins), collecting duct acid secretion (5 proteins), complement and coagulation cascades (5 proteins) and glutathione metabolism (4 proteins). Additional, ECM-receptor interactions (4 proteins) and antigen processing and presentation (4 proteins) were also observed to be significantly enriched. It was observed that the DEPs were involved in physical or functional interactions to constitute a network through STRING database analysis (Fig 6). The PPI network analysis found that some DEPs interact with each other, such as CD14-CD68-Atp6v0c-Paps-Gm2a-Gns, Cldn5-Esam-CD34-Col1a1-Itga8-Col4a3 and Rac2-Cyba-Ncf2-Ncf4. These key focus hubs have important biological functions in biological regulation, oxidative stress, enzyme activity, cell migration and motility, lysosome, biological adhesion, exponse to stimulus, receptor binding, etc.

Discussion
TMT technique is used for quantitative proteomics with high throughput and high reproducibility. PRM analysis was used for validating the accuracy and reproducibility of the proteomic  data. Our results showed that pulmonary fibrosis was induced by a single exposure to silica particles by intratracheally instillation in rats. Two hundred and eighty-five DEPs were identified between the control and silicosis model groups (Fig 1). We selected six proteins for verification using PRM and the results showed a similar expression trend with TMT (Fig 3), suggesting the reliability of our TMT analysis. The DEPs mainly enriched in the pathways of phagosome, lysosome, oxidative stress and ECM-receptor interactions.

Silica exposure influences phagosome
Inhaled silica dust is predominantly phagocytosed by alveolar macrophage (AM) when it enters the pulmonary alveolus [25]. In this study, phagosome proteins such as SFTPA1, RT1-Ba, CD14, SCARB1 and SEC61B were upregulated in the lung of silica-exposed rats. Macrophages are innate immune cells with various types of receptors such as Fc receptors, scavenger receptors, and Toll-like receptors (TLRs), of which TLRs are crucial to macrophage phagocytosis. Silica particle, recognized as pathogen-associated molecular patterns by the innate immune system, binds the membrane-bound TLRs to active two different signaling cascades: the "myeloid differentiation primary response gene 88 (MyD88)-dependent" and the "Toll/ interleukin 1 receptor domain-containing adaptor-inducing interferon-β (TRIF)dependent" cascades. These signaling cascades induce the activation of nuclear factor κB (NF-κB) and interferon regulatory factor 3 (IRF3). Finally, inflammatory cytokines such as tumor necrosis factor (TNF)-α and interleukin (IL)-1β are produced to promote fibrosis [26]. CD14 has been shown to be required for TLR4 endocytosis to activate downstream signaling [27]. The core fucosylation deficiency in CD14 suppressed the endocytosis of TLR4 and impaired TLR4 signaling in mouse embryonic fibroblasts [28]. Alveolar macrophage may phagocytose silica particles through TLR4-mediated recognition. Additionally, Scavenger Receptor Class B Member 1 (SCARB1) is a silica receptor associated with canonical inflammasome activation [29]. Phagocytosed silica particles cannot be digested in phago-lysosomes, which induces lysosomal stress and activates NLRP3 inflammasomes, followed by progressive lung fibrosis [30].

Silica exposure influences lysosome
Inhaled silica dust disrupted lysosomes, which released lysosomal cathepsins (Cats) and other hydrolases into the cytosol [31]. In this study, we confirmed Cat D, S, H, L and PSAP were significantly upregulated in the lung of silica-exposed rats. Upregulated Cat S, L, B and K were also observed in the lung of silica-exposed mice [32]. Cat B contributes to lung fibroblast differentiation into myofibroblasts by triggering TGF-β1-driven canonical SMAD pathway. Inhibition of Cat B diminished α-SMA expression and delayed lung fibroblast differentiation [33], and also reduced hepatic inflammation, collagen deposition and fibrogenesis [34]. Conversely, Cat K could inactivate TGF-β1 and restrict excessive ECM deposition to control lung fibrosis [35], and Cat S may proteolytically inactivate Cat K and thus would control its collagenolytic or elastinolytic activity [36]. Additionally, multiple Cathepsins such as Cat B, L, C, S and X promote pro-IL-1β synthesis and NLRP3-mediated IL-1β activation in murine macrophages [37]. Prosaposin (PSAP) is a precursor for four sphingolipid activator proteins known as saposins A-D, which serve as activators for lysosomal hydrolases [30]. PSAP can reverse the inhibitory effects of Cystatin C (CST3) on Cathepsins by forming a complex that changes the conformational properties [38]. In prostate cancer cell, downregulation of PSAP decreased b1A-integrin expression, its cell-surface clustering, and adhesion to basement membrane proteins. Cat D expression and proteolytic activity, migration, and invasion were also significantly decreased in PSAP knock-down cells [39]. Downregulation of PSAP might be contribute to silicosis therapy.

Silica exposure influences oxidative stress
Silicosis is a disease associated with oxidative stress. In AM, ROS was mainly generated by NADPH oxidase (NOX) from alveolar macrophages. In this study, we conformed the subunits of NADPH oxidase complex such as NOX 2 (gp91 phox ), p22 phox , p47 phox , p40 phox and p67 phox were upregulated in the lung of silica-exposed rats. By activating p38 MAPK signaling pathway ROS disrupted lung endothelial integrity and increased vascular hyperpermeability [40], which created a pro-fibrotic intra-alveolar environment to promote several profibrotic responses, such as intra-alveolar coagulation and provisional matrix establishment [41]. CLDN5, a marker for endothelium tight junctions and permeability [42], predominantly expressed in the cell-cell junctions of alveolar endothelial cells and played critical roles in the pulmonary endothelial barrier. Downregulation of CLDN5 was associated with disrupted endothelium tight junctions in bleomycin-induced pulmonary fibrosis, and which may be involved in epithelial-mesenchymal transition (EMT). TGF-β also disrupted the alveolar epithelial and endothelial tight junctions by downregulating CLDN5 expression [43]. In cardiac fibroblasts and endothelial cells TGF-β also induced COL1A1 expression by downregulating CLDN5 expression, which also promoted macrophage infiltration and pro-fibrotic responses [44]. We Confirmed CLDN5 was downregulated in the lung of silica-exposed rats in this study. Endothelial hyperpermeability induced by oxidative damage may contribute to silica -induced pulmonary fibrosis. Therefore, a therapeutic approach of limiting the extent of vascular leak may be an effective strategy for treating silicosis.

Silica exposure influences Extracellular Matrix (ECM)
The primary pathological characteristic of silicosis is the imbalance of extracellular matrix anabolism and catabolism. MMPs degrade all ECM components as well as divers nonmatrix proteins including cytokines, chemokines, and receptors, but the catalytic activity of MMPs can be compromised by the tissue inhibitor of metalloproteinases (TIMP) family. In this study, upregulated MMP-8 and downregulated TIMP-3 were identified in the lung of silica-exposed rats. MMP-8 can degrade basement membrane and extracellular proteins, causing airway disruption and remodeling in chronic obstructive pulmonary disease (COPD) [45]. TIMP-3 has been recognized as a key regulator in lung homeostasis, which plays a versatile part in the development of inflammation as well as fibrosis, rather than merely acting through the restriction of ECM degradation. More severe fibrosis occurs in bleomycin-injured TIMP3-deficient mice [46].
Integrin α8 (ITGA8), an important component of ECM-receptor interaction pathway, played important roles in the expression of extracellular matrix components [47]. ITGA8 may participate in the degradation of extracellular matrix, including collagen type XI alpha 1, aggrecan, collagen type VI alpha 1 [48]. Additionally, ITGA8 attenuated tubulointerstitial fibrosis by regulation of TGF-β /Smad2/3 signaling, fibroblast activation and immune cell infiltration [49]. Deficiency of ITGA8 worsened tubulointersititial fibrosis [50] and delayed healing in a model of glomerulonephritis [51]. In the lung, ITGA8 expression was restricted to interstitial stromal cells, and that was increased in bleomycin-induced fibrosis. ITGA8 deletion increased COL1A1 production during lung fibrosis in vitro, but did not affect pulmonary fibrosis in the bleomycin animal model [52]. We confirmed ITGA8 was decreased in silicainduced pulmonary fibrosis, and the role of ITGA8 in silica-induced needs further research.

Conclusion
In summary, we found some proteins which are closely relevant to the occurrence and development of silicosis using TMT coupled with PRM technology. Most proteins were enriched in immune system processes, oxygen transporter activity, phagosome, lysosome and ECM-receptor interactions. These findings will further provide useful clues to elucidate pathogenesis of silicosis and reveal more potential therapeutic targets.