Mesenchymal stem cells ameliorate silica‐induced pulmonary fibrosis by inhibition of inflammation and epithelial‐mesenchymal transition

Abstract Silicosis is a devastating occupational disease caused by long‐term inhalation of silica particles, inducing irreversible lung damage and affecting lung function, without effective treatment. Mesenchymal stem cells (MSCs) are a heterogeneous subset of adult stem cells that exhibit excellent self‐renewal capacity, multi‐lineage differentiation potential and immunomodulatory properties. The aim of this study was to explore the effect of bone marrow‐derived mesenchymal stem cells (BMSCs) in a silica‐induced rat model of pulmonary fibrosis. The rats were treated with BMSCs on days 14, 28 and 42 after perfusion with silica. Histological examination and hydroxyproline assays showed that BMSCs alleviated silica‐induced pulmonary fibrosis in rats. Results from ELISA and qRT‐PCR indicated that BMSCs inhibited the expression of inflammatory cytokines TNF‐α, IL‐1β and IL‐6 in lung tissues and bronchoalveolar lavage fluid of rats exposed to silica particles. We also performed qRT‐PCR, Western blot and immunohistochemistry to examine epithelial‐mesenchymal transition (EMT)–related indicators and demonstrated that BMSCs up‐regulate E‐cadherin and down‐regulate vimentin and extracellular matrix (ECM) components such as fibronectin and collagen Ⅰ. Additionally, BMSCs inhibited the silica‐induced increase in TGF‐β1, p‐Smad2 and p‐Smad3 and decrease in Smad7. These results suggested that BMSCs can inhibit inflammation and reverse EMT through the inhibition of the TGF‐β/Smad signalling pathway to exhibit an anti‐fibrotic effect in the rat silicosis model. Our study provides a new and meaningful perspective for silicosis treatment strategies.


| INTRODUC TI ON
Silicosis is a devastating occupational disease associated with longterm silica dust particle inhalation and deposition, and it is characterized by chronic inflammatory changes, granuloma formation and diffuse pulmonary interstitial fibrosis. 1 It is currently the most common occupational disease with the highest incidence and prevalence rates in China. Rapid industrialization has increased occupational exposure to respirable silica particles in various industries and work environments, such as coal and metal mines; building material, glass product and ceramic manufacturing; and sandblasting, as well as in iron, steel and non-ferrous foundry work. As artificial stone has recently become widely popular in interior designing, a silicosis epidemic is emerging in the artificial stone industry. 2 Owing to insufficient protective measures, workers form a high-risk group for silicosis development. Silicosis may develop after many years, even with exposure cessation. Despite prevention efforts for many decades, silicosis remains a serious problem worldwide. 3 Although the incidence of silicosis is still very high in the developing countries, many people believe, owing to rapid economic development, that silicosis is a historical and ancient disease and therefore rarely attracts clinical attention. Unlike those employed for common lung diseases such as asthma, lung cancer and other fibrotic lung diseases, modern medical research techniques are rarely used for the research and treatment of silicosis. Clinically, there are no effective treatments available for silicosis except lung transplantation. 4,5 Therefore, new therapies are desired to prevent, improve or inhibit silicosis.
Mesenchymal stem cells (MSCs) are adult multipotent progenitor stem cells with self-renewal capacity and multi-lineage differentiation potential. 6 MSCs have low immunogenicity, multidifferentiation potential, immunomodulatory properties, paracrine function, homing effect and abundant sources and are easily isolated and amenable to culture expansion in vitro. Thus, MSCs are considered promising candidates for clinical applications, such as neurodegenerative diseases, 7 cardiovascular diseases 8 and lung diseases. 9,10 Till date, multiple studies have focused on MSCs.
Silicosis initiation and development are a complex process involving multiple cells regulated by complex molecular networks.
Silicosis begins with inflammation, alveolar epithelial cell damage and apoptosis, followed by tissue repair, epithelial-mesenchymal transition (EMT) stimulation, myofibroblast activation, fibroblast and myofibroblast foci formation, and extracellular matrix (ECM) deposition, ultimately leading to pulmonary fibrosis. Intratracheal Au nanoparticle-labelled or unlabelled bone marrow-derived mesenchymal stem cells (BMSCs) transplantation into mice with silica-induced pulmonary fibrosis has shown that BMSCs inhibit silica-induced macrophage activation and have an anti-inflammatory and anti-fibrotic role. 11 During EMT, epithelial cells transdifferentiate into mesenchymal cells, which is essential for embryogenesis and tissue regeneration. EMT also contributes to fibrotic disease 12 and cancer progression. 13 During this process, loss of cell-cell adherens junctions and apical-basal polarity triggers cytoskeletal composition shift, and cell-cell and cell-ECM interactions are remodelled. 14 Epithelial cells lose their polarity and acquire mesenchymal phenotypes such as high migration and invasion, anti-apoptosis and ECMdegrading ability. EMT involves the functional cooperation of several signalling pathways, such as transforming growth factor β (TGFβ), Notch, Wnt/β-catenin and Hedgehog. Among these, TGFβ family signalling plays a predominant role. 15 The Smad protein family is the downstream substrate of the TGFβ receptor kinase present in the cytoplasm. It directly transduces the TGFβ signals from the cell membrane to the nucleus and is an important signalling molecule in the TGFβ signalling pathway. A study has reported that BMSCs and their exosomes promote endometrial repair and reverse EMT via the TGF-β1/Smad signalling pathway. 16 Another study has shown that adipose-derived stem cell (ADSC) transplantation attenuates renal interstitial fibrosis through inhibition of EMT via the TGF-β1 signalling pathway. 17 Therefore, we investigated the effects of BMSCs on silicainduced pulmonary fibrosis in rats. The underlying mechanism was also discussed. This study provides a theoretical basis for silicosis treatment.

| Culture and identification of BMSCs
Sprague-Dawley (SD) rat BMSCs were purchased from Cyagen Biosciences Inc (Suzhou, China). The cells were recovered and seeded into T25 culture flasks with sufficient complete medium (Cyagen) in a saturated humidified atmosphere with 5% CO 2 at 37°C. The medium was exchanged every 3 days, and when the cells reached 80%-90% confluence, they were passaged. BMSC morphology was observed under a light microscope. The multi-lineage differentiation of BMSCs was determined using adipogenic and osteogenic differentiation medium (Cyagen). The passage 2 cells (2 × 10 4 cells/

| Animals
All animals received humane care, and all methods were carried out

| Rat model of silicosis and BMSC treatment
After one week of adaptive breeding, these animals were randomly divided into three groups: control group, silica group and BMSCs group (n = 20 rats per group, total 60 rats). Crystalline silica particles (0.5-10 μm, approximately 80% between 1 and 5 μm; Sigma-Aldrich, USA) were autoclaved and suspended in sterile saline at a concentration of 100 mg/mL. Silica suspensions were sonicated for 10 minutes before use. After the animals were anaesthetized by isoflurane inhalation, the silica and BMSCs group rats received non-exposed intratracheal instillation of 500 μL silica suspension (100 mg/mL/rat). The rats in the control group were treated with 500 μL sterile saline. Then, 1 mL BMSCs (2 × 10 6 cells/mL) were injected into rats of the BMSCs group by tail vein, whereas an equal volume of sterile saline was injected in the control and silica group rats on days 14, 28 and 42 after silica suspension administration. A total of 10 rats in each group were killed on days 28 and day 56 after intratracheal instillation. After intraperitoneal anaesthesia with pentobarbital sodium (2%, 50 mg/kg), the rats were weighed and fixed on an anatomic device. After collecting the bronchoalveolar lavage fluid (BALF), the complete lung tissue was removed, the trachea and fascia were obtuse removed, washed with PBS and dried on a filter paper. The lung was accurately weighed, and the lung coefficient was calculated as the ratio of lung weight (g) to bodyweight (g) × 100. Lung tissue and BALF were stored for subsequent analyses.
The study design is shown in Figure S1.

| Histological examination
The left lungs were fixed with 4% paraformaldehyde, embedded in paraffin and then cut into 6μm-thick paraffin sections. The sections were, respectively, stained with haematoxylin and eosin (H&E) to observe lung morphology and inflammatory infiltrates and Masson staining to estimate fibrosis severity. A semi-quantitative score of pulmonary inflammation and fibrosis was performed according to the inflammation score 18 and modified Ashcroft scale. 19 For immunohistochemical (IHC) staining, sections were incubated with rabbit anti-rat primary antibodies against fibronectin (1:600), collagen I (1:1000), E-cadherin (1:1000) and vimentin (1:5000) (both from ProteinTech, Rosemont, IL, USA), followed by a horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG secondary antibody (1:5000, ProteinTech). The whole section was then scanned, and the images were obtained using CaseViewer.

| Hydroxyproline assay
To analyse pulmonary fibrosis severity, pulmonary hydroxyproline (HYP) content was measured using hydroxyproline detection kit according to the manufacturer's instructions (Nanjing Jian Cheng Institute, Nanjing, China). The HYP levels in lung tissue were detected at an absorbance of 550 nm. The results were expressed as μg HYP/mg of lung weight.

| Enzyme-linked immunosorbent assay (ELISA)
The lung tissues were homogenized using saline. The homogenates and collected BALF were centrifuged for ELISA. TNFα, IL-1β, IL-6 and TGF-β1 levels in the lung tissues and BALF were measured using ELISA detection kits according to the manufacturer's instructions (Elabscience, Houston, TX, USA).

| Quantitative reverse transcription polymerase chain reaction (qRT-PCR)
The lung tissues were homogenized using saline. Total RNA was extracted from the homogenates using RNAiso Plus Reagent (Takara, Kyoto, Japan) and quantified using a NanoDrop 2000 Spectrophotometer (Thermo Fisher Scientific). Total RNA was reverse-transcribed into cDNA using the PrimeScript™ RT Reagent Kit with gDNA Eraser (Takara). qRT-PCR was performed with SYBR Green PCR Kit (Takara) and an ABI 7500 Fast Real-Time PCR System (Applied Biosystems). All samples were assayed in triplicates. All results were calculated using the 2 −ΔΔCt method and normalized to that of GAPDH. The primer sequences are listed in Table S1.

| Western blot
The lung tissues were lysed using radioimmunoprecipitation assay Protein bands were detected using an enhanced chemiluminescence (ECL) kit (Absin, Shanghai, China) and imaged on Amersham Imager 600 (Amersham, Little Chalfont, UK). The intensities were analysed and quantified using ImageJ software to calculate relative expression by normalizing to GAPDH.

| Statistical analysis
All data obtained from the experiments were presented as mean ±standard deviation (SD). Comparisons between two groups were analysed using the Student's t test. Comparisons among multiple groups were performed with one-way analysis of variance (ANOVA).
Statistical analysis was performed with SPSS 21.0. Values of P < .05 were considered statistically significant.

| Identification of BMSCs
We confirmed the biological characteristics and purity of BMSCs by observing the cell morphology under a light microscope, determining the differentiation ability from adipocyte and osteogenic induction, and detecting the cell surface markers using flow cytometry. The cells were spindle-shaped or fibroblast-like, converged and arranged in a swirl ( Figure 1A). Lipid droplets and red calcium nodules of various sizes were visualized by Oil Red O ( Figure 1B) and Alizarin Red ( Figure 1C) staining, respectively, which demonstrated that the cells exhibited the capacity to differentiate into adipocytes and osteoblasts under the induction of adipogenic and osteogenic differentiation medium, respectively. In addition, flow cytometry analysis showed that CD90 and CD29 expression was 98.6% and 99.0%, respectively, whereas CD45 and CD11b expression was <1% ( Figure 1D).

| BMSCs improved lung tissue morphology of and lung coefficient
The altered lung morphology of rats in each group was observed by the naked eye. The lungs of the control group rats were red, uniformly coloured, soft and elastic, and the surface of the pulmonary lobes was smooth without spots or nodules. On the contrary, the silica group rats showed visibly injured lungs; their lungs were grey, hard and lacked elasticity, with different-sized spots and nodules in the pulmonary lobes. Future, the lung volume of the silica group rats increased compared with that of the control group rats. Although scattered spots and nodules were present in the lungs of the BMSCs group rats, lung injury was lesser than that in the silica group rats ( Figure S2). Lung coefficient is another indicator of lung injury. As shown in Table 1, compared with that of the control group, the lung coefficient of the silica group rats was significantly increased both on days 28 and 56, BMSC transplantation reduced the increase in lung coefficient caused by silica-induced pulmonary fibrosis on day 56, but no statistical difference was observed on day 28.

| BMSCs ameliorated the pulmonary pathological changes and pulmonary fibrosis induced by silica
To detect the pathological changes, the lung tissues were stained with H&E and Masson to observe inflammatory infiltrates and  Figure S3). HYP content reflects collagen metabolism in tissues and is an index of disease fibrosis.
HYP content in lung tissue of the silica group was significantly increased compared with that in the control group, whereas compared with the silica group, HYP content in lung tissue of the BMSCs group was significantly decreased on day 56, but no statistical difference was observed on day 28 ( Figure 2C). The mRNA expression levels of fibronectin and collagen I detected by qRT-PCR showed the same trend ( Figure 2D-H).

| BMSCs alleviated silica-induced inflammatory cytokine production
The expression levels of inflammatory cytokines including TNFα, IL-1β and IL-6 in each group were examined by qRT-PCR and ELISA.

| BMSCs reserved silica-induced EMT
In addition to the anti-inflammatory effects, we also explored another biological role of BMSCs in reducing silicosis. We performed qRT-PCR, Western blot and immunohistochemistry to measure the expression levels of several EMT indicators and ECM components such as fibronectin and collagen I. As shown in Figure 4A-C, F-G, the expression of epithelial marker E-cadherin decreased significantly, and the mesenchymal marker vimentin increased in the silica group rats. However, after three BMSCs treatments, E-cadherin expression increased and vimentin expression decreased. At the same time, we also tested the expression of ECM components fibronectin and collagen I. Under the stimulation of silica for 56 days, ECM deposition was significant; however, BMSCs reduced the ECM deposition ( Figure 2E-F, Figure 4C-E, H-I).

| BMSCs blocked TGFβ/Smad pathway activation
We further explored the mechanism by which BMSCs could alleviate EMT. qRT-PCR, Western blot and ELISA were performed to measure the mRNA and protein expression levels of TGF-β1. These results showed the TGF-β1 was significantly up-regulated in the silica group rats, whereas the mRNA and protein expression levels of TGF-β1 were significantly decreased in the BMSCs group rats compared with those in the silica group rats ( Figure 5A-E). Western blot analysis showed p-Smad2/Smad2 and p-Smad3/Smad3 protein levels increased and Smad7 protein levels decreased in the silica group rats compared with those in the control group rats, demonstrating that silica activates the TGFβ/Smad pathway. However, after BMSCs treatment, expressions of p-Smad2/Smad2 and p-Smad3/Smad3 were down-regulated and that of Smad7 was up-regulated compared with that in the silica group rats (P < .05) ( Figure 5D, F-H). transplanted MSCs immediately after constructing the silicosis model using silica. However, in real life, most people exposed to dust and may suffer from silicosis are workers with poor economic conditions. It is unrealistic to perform stem cell therapy before or shortly after dust exposure. They often receive treatment after symptoms or even a diagnosis of silicosis. The pathogenesis of silicosis generally includes acute inflammation, chronic inflammation and fibrosis; however, the boundaries between the stages are not clear. Our previous studies have shown that inflammation is predominant before 14 days, and fibrosis is predominant afterwards in a silica-induced mouse pulmonary fibrosis model. After 28 days, mature cellular silicon nodules are formed, accompanied by a large amount of collagen deposition. 24,25 A recent study has reported the collagen fibres and the number of nodules formed on day 15 after silica instillation. 26 In summary, we transplanted BMSCs on 14 days after perfusion of the silica suspension. It has been reported that BMSCs begin to decline gradually and disappear within 30 days in the body. 27 To maintain a high level of BMSCs in the silica-exposed rats, we repeated the intervention three times every 14 days, and rats were killed after 28

F I G U R E 2
BMSCs improved pulmonary pathological changes and pulmonary fibrosis. (A) H&E staining and Masson's Trichrome Staining were performed to detect the pathological changes in lung tissues after rats were exposed to silica for 28 and 56 days (200×). (B-C) Severity of pulmonary inflammation and fibrosis evaluated by inflammation score and modified Ashcroft score. n = 7 rats per group. (D) Pulmonary HYP content of each group after rats was exposed to silica for 28 and 56 days. (E-H) The mRNA expression of fibronectin and collagen I of each group after rats were exposed to silica for 28 and 56 days. n = 3 rats per group. mRNA expression values were normalized to GAPDH. Data were presented as mean ±SD. * P <.05, ** P <.01, *** P <.001, **** P <.0001 and 56 days after a single and three interventions, respectively. We also calculated the lung coefficients and detected their lung HYP content and fibrosis-associated genes on days 28 and 56. These results showed that BMSCs reduced lung injury and improved pulmonary fibrosis on day 56, but the effect was not significant on day 28, suggesting that continuous BMSCs transplantation ameliorated silicosis in rats.
After proving that three BMSCs transplantations may be used to treat silica-induced pulmonary fibrosis in rats, we investigated the mechanism through which BMSCs protect against silicosis in rats. In addition, we explored other processes involved in BMSCsinduced silica-induced pulmonary fibrosis amelioration. EMT describes the process by which epithelial cells lose their functionality and characteristics and adopt mesenchymal characteristics that confer motility. There is increasing evidence that sustained EMT is the key mechanism of multi-organ fibrosis pathology. 12  I and fibronectin. 32 We found that compared with the control group,

| CON CLUS ION
In this study, the results showed that BMSCs could alleviate silica-induced pulmonary fibrosis in rats by regulating inflammatory response and EMT process and reducing ECM deposition.
Furthermore, silica stimulation activates the TGFβ/Smad pathway, whereas BMSCs could inhibit this activation. Ultimately, these findings suggest that BMSCs may be a promising therapeutic strategy for treating silicosis.

ACK N OWLED G EM ENTS
The work was supported by the National Natural Science Foundation of China. (Grant No. U1904209 to WY).

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.