Comparison of the effects of different mesenchymal stem cells on attenuating phosgene-induced lung injury in rats

Phosgene-induced lung injury is an important type of acute lung injury (ALI), which mainly leads to acute pulmonary edema. Currently, no effective clinical treatment has been developed yet. In the present study, the effects of lung-resident mesenchymal stem cells (LRMSCs), bone marrow mesenchymal stem cells (BMSCs) and human chorionic villi-derived MSCs (hCMSCs) were compared in phosgene-induced lung injury of male SD rats.


Introduction
Phosgene, as a highly toxic gas, is widely used as an intermediate in the chemical synthesis. Lung injury caused by inhalation of phosgene mainly leads to acute toxic pulmonary edema. However, there is no clear consensus on the pathogenesis of poisoning caused by phosgene and there is no targeted treatment for phosgene induced lung injury. Currently, phosgene-induced lung injury is often treated with clinical support, and new treatment methods are urgently necessary. In recent years, MSCs have become an important research hotspot in lung injury, bringing hope to the treatment of phosgene inhalation lung injury. BMSCs, LRMSCs, and hCMSCs are intensively studied. To date, BMSCs are the most extensively studied stem cells, due to their potential of self-renewal and multidirectional differentiation. BMSCs have a speci c role in the treatment of multiple systemic diseases such as acute pancreatitis, ischemic stroke, brotic scar and so on [1][2][3]. Previous studies have con rmed that BMSCs had decreases the severity of lung injury caused by phosgene [4]. LRMSCs are a class of MSCs derived from the lungs. Studies have shown that LRMSCs co-cultured with type II epithelial cells (ATII) can differentiate into ATII cells [5].
LRMSCs are reported to treat hyperoxia-induced lung injury in mice [6]. Studies have also con rmed that LRMSCs can signi cantly improve in ammatory response and reduce pulmonary edema in endotoxininduced lung injury [7]. HCMSCs also play a repair role in lung injury [8]. However, the speci c effects of Recent studies in naphthalene-damaged mice have shown that BMSCs are distributed in the lungs following naphthalene exposure to help regenerate club cells. Besides, the migration of stem cells which is mediated by chemokines is an important differentiation point between BMSCs and hCMSCs [9]. This suggests that different migration ability of MSCs may be the key to the differences in the roles of different MSCs in repairing lung injury. HGF is a cytokine with functions such as mitogen, cytogen, morphogen and angiogenesis. In normal and tumor tissue, HGF can be produced by broblasts and other stromal cells and aggregated in the extracellular matrix and basement membrane to bind to glycosaminoglycan sulfate [10]. Studies have reported that HGF plays a major role in promoting migration in many tumors [11,12]. Other studies have con rmed that HGF promotes the directional migration of MSCs [13,14]. Roggia et al. con rmed HGF enhanced the differentiation of embryonic stem cells [15]. HGF was proved to be an effective differentiating factor in the comparison of embryonic stem cells, cardiac stem cells and adipose tissue stem cells. Therefore, HGF is associated with the differential mechanism of the anti-lung injury action of stem cells by affecting the activity and function of MSCs. HGF in uences the role of stem cells and this may be an important factor affecting the differences in protective effects caused by the three MSCs in lung injury.
β-catenin plays an indispensable role as an adhesion adapter and transcriptional synergistic regulator. It also regulates and determines the fate of cells during embryogenesis and also involved in the induction of embryonic mesoderm [16,17]. Activated β-catenin can improve the migration of MSCs in mice and humans [18,19]. The migration of human LRMSCs has been closely related to the β-catenin signaling pathway [20]. Wnt, as the most famous activator, promotes β-catenin accumulation in the cytoplasm, which is transferred into the nucleus and binds to transcription factors of the LEF-1/TCF family to change the expression of the β-catenin gene [21]. Meanwhile, other ligands also affect the β-cateninn pathway [22]. GSK-3β is an important component involved in this signal pathway. GSK-3β phosphorylation affects β-catenin expression. Numerous studies have con rmed that GSK-3β plays an important role in the protection of the viscera. Several pathways in the body converge on GSK-3β making it a key protein. Here, it is suggested that HGF triggers a signaling pathway in MSCs. This pathway may be that HGF induces phosphorylation of GSK-3β and futher leads to the accumulation and nuclear translocation of β-catenin in the cytoplasm. To verify this biological effect, it is hypothesized that HGF induces cell migration by activating β-catenin pathway. In this study, effects of BMSCs, LRMSCs, and hCMSCs in phosgene inhalation lung injury was compared and the mechanism of difference in the function of different MSCs was studied.

Phosgene inhalation lung injury
Clean grade Sprague-Dawley rats (Male, 180-220 g, 5-6 weeks old) were provided by the Medicine School of Fudan University and maintained in a clean animal house at 22 ~ 24℃ (dark/light cycle, 12 h). We obey the requirements of the Ethics Committee of Jinshan Hospital a liated to Fudan University and complied with international guidelines for animal experimentation. The method of establishing the lung injury model by phosgene was developed as previously described [23]. Brie y, rats were poisoned with phosgene (8.33 mg/L) in a 24L sealed chamber for 5 minutes. Animals were randomly divided into several groups (n = 6 at each time point).
Isolation of LRMSCs and cell culture LR-MSCs were isolated from BALF of rats treated by KGF-2 as previously described [24]. The lungs of each rat were added into rhKGF-2 (5 mg/kg, 200µL). After 72 h, LRMSCs were extracted from BALF. The cells precipitation was left and plant in T25 bottles. Cells were cultured in a incubator with 5% CO 2 at 37℃ for 14 days. Flow cytometry analysis demonstrated that CD73, CD105, CD90 and CD44 were positively expressed for cell surface markers while CD45 was negatively expressed for hematopoietic markers. Multi-lineage differentiation potential of LR-MSCs was con rmed by induction into other mesenchymal lineages such as osteocytes, adipocytes, and chondrocytes as previously demonstrated [24]. BMSCs and hCMSCs were provided by ScienCell Research Laboratory (Carlsbad, CA, USA) with primary cells (processing institutional identi cation certi cate). Cells were cultured in high-glucose DMEM (Invitrogen, Carlsbad, CA) which was supplemented with 10% FBS, Penicillin/Streptomycin (100 U/ml) and fungizone (0.5%) in a humidi ed incubator with 5% CO 2 at 37℃. MSCs at P3 ~ P6 were used in the experiments.

CCK8 assay
When the numbers of MSCs reached 5 × 10 4 after being cultured in DMEM, MSCs were digested with 0.25% trypsin. The detached cells were transferred into the 96-well plates (10 3 cells/well). The medium (100µL) was added to each well and 8 repeated holes were set up in each group at each time. CCK8 (85 mg/mL, 10µL) was added. The cells were put back into the cabinet and incubated at 37℃ for 120 min. The OD450 values were determined every 24 hours until the ending of the experiment at the 96th hour.

Treatment of MSCs
The treatment groups were given with MSCs (1 × 10 6 ) in 100µL of PBS suspension by tracheal infusion immediately after being exposed in phosgene. The normal group and phosgene exposed group were given the same volume (100µL) of PBS as the control group through the trachea by using the same method. The rats were sacri ced in 6 h, 24 h and 48 h after being poisoned, respectively. Their lungs, BALF and serum were collected for subsequent analyses.

Histological study
The collected lungs were xed in 10% formaldehyde. After that, lungs were embedded in para n and wad made to 5 µm sections. Then they were stained with hematoxylin and eosin (H&E Lung wet to dry ratio A small piece of tin foil paper (2cm × 2 cm) was weighed as W1. The right lower lung lobe was taken from every rat and wrapped in foil after its residual blood was wiped off the surface. The weight was measured as W2. Then it was placed in a 60 ℃ dryer, dried for 48 h and weighed as W3 then. According to the formula, W/D = (W2-W1)/(W3-W1), lung wet to dry ratio was recorded for evaluating the severity of pulmonary congestion and edema.

Bronchoalveolar lavage uid (BALF) collection
The animal's neck skin was cut and opened to nd the trachea at 6 h, 24 h and 48 h after phosgene inhalation. The trachea was cut 1/3 under the cricoid cartilage. The scalp needle connection tube with a length of about 3 cm was prepared and connected with a 5 mL needle tube. The connection was inserted from the incision along the direction of the rat trachea. Then, PBS (1 mL) was injected and lavage was performed for 3 times. The obtained BALF was transferred and centrifuged (1000 rpm × 10 min, at 4℃) to remove the precipitation and the supernatant was immediately detected.

Protein concentration and cell count
Protein concentration in BALF was assayed by bicinchoninic acid (BCA) assay kit (Pierce, Rockford, IL, USA) following the instructions. The cell pellet was obtained from BALF after centrifuge (1500 rpm × 10 min, at 4 °C). The total cell count was recorded by a hemocytometer.

ELISA
The protein expressions of IL-6, IL-1β, TNF-α, IL-10 and HGF in BALF and serum were detected with the corresponding kits (KANGLANG Technology Co., Ltd., Shanghai, China) according to the manufacturer's instruction. The OD450 value of each well was measured.

RT-PCR
The total RNA was isolated from each sample using Column Animal RNAout (Shanghai Zeye Biotechnology Co., Ltd., Shanghai, China) according to the instruction. An Easyscript rst-strand cDNA synthesis super mix kit (Saiweier, Shanghai, China) was used to synthesize the rst strand of cDNA by RT- Wound-healing assay MSCs (5 × 10 4 /L) suspension (2 mL) was dropped to each hole of the test plate. The plate was shaken well and then was put into the 37℃ cabinet. When MSCs grew to 80% of each hole, the end of the pipette was used to draw a vertical line in the center of the hole. Then the plate was put into the 37℃ cabinet again for 48 h. MSCs were taken photos under the mirror.

Transwell assay
MSCs migration was evaluated using the transwell chamber (8 µm speci cation, provided by BD) was used. MSCs were seeded into the upper chamber (1 × 10 5 /200µL) which were put into 24-well plates loaded with 600µL DMEM. In the rst migration experiment, six visual elds counts were randomly selected after 24 h. 10%FBS, 30 ng/mL CXCL12 and 25 ng/mL HGF were added to the lower chamber in turn in the second migration experiment. After 4 hours, MSCs migrated through the chamber and were xed and stained using Hema 3 staining kit (Fisher Scienti c) was used and then the chamber was rinsed with water. The remaining cells in the upper chamber were swabbed and left to dry. HGF-induced migration of LRMSCs and hCMSCs were tested by transwell migration assays and stained with trypan blue. The cell counts were recorded from 5 random elds of per transwell chamber. The results were expressed as the mean number of cells.

Statistical analysis
All statistical analyses were performed using SPSS v19.0 software. Signi cance was determined by Nonparametric tests followed by Bonferroni correction. The results are presented as the median (quartile). P < 0.05 was considered statistically signi cant.

Results
Effects of BMSCs, hCMSCs and LRMSCs on histopathological changes and PaCO 2 in ALI It could be seen from the pathological sections (H&E staining) that no in ammatory changes were detected in the normal rats (Fig. 1a) while the model group showed swollen lung interstitium and in ltration of many blood cells (Fig. 1a). Treatment with three types of MSCs restored damaged alveolar structure and signi cantly alleviated pathological damage (Fig. 1a). Compared to those of the normal group, lung injury scores, total protein content and cell count in BALF in the model group were all increased after 24h. However, while those in treatment groups were signi cantly decreased (Fig. 1b-e). PaCO 2 was signi cantly decreased at 6h, 24h and 48h after inhalation of phosgene ( Fig. 1f-h). MSCs treatment resulted in the recovery of PaCO 2 (Fig. 2f-h). These changes were more prominent in the BMSCs group compared to the other groups.

Effects of BMSCs, hCMSCs and LRMSCs on reducing in ammatory response and immune regulation in ALI
The expressions of TNF-1α, IL-6 and IL-1β in BALF and serum were increased after phosgene inhalation while their expressions were decreased in three MSCs groups (Fig. 2a-c, e-g). Moreover, the changes of BMSCs were more signi cant than those of other MSCs groups (Fig. 2a-c, e-g). These results demonstrated that BMSCs could better resist in ammatory response both in the lungs and the entire body. IL-10 possess immunomodulatory properties [25]. After the rats were poisoned, the levels of IL-10 in both BALF and serum were increased and the increase was even higher after MSCs treatment (Fig. 2d, h). The effects of BMSCs were stronger than those of LRMSCs and hCMSCs, indicating that BMSCs had higher immune regulation ability in ALI than LRMSCs and hCMSCs.

Changes in HGF after the treatment of BMSCs, hCMSCs and LRMSCs
The expression of HGF in BALF and serum of each group were detected at 6h, 24h and 48h, respectively ( Fig. 3a-f). Compared with healthy rats, the levels of HGF in BALF and serum in the phosgene group were increased. HGF expression was increased in all MSCs treatment groups. BMSCs showed the most signi cant effects followed by hCMSCs and LRMSCs.

Effects of BMSCs, hCMSCs and LRMSCs on repairing alveolar epithelia and vascular endothelia in ALI
Many red cells were shown to leak into the alveoli in the injured lungs (Fig. 4a). After MSCs treatment, fewer red cells were present in the alveoli indicating that vascular endothelial permeability had been signi cantly recovered (Fig. 4a). The expression of VE-CAD, a vascular endothelial cell marker, was decreased after lung injury but was increased in rats treated with MSCs (Fig. 4e, f). BMSCs, hCMSCs and LRMSCs were found to have protective effects in repairing vascular endothelia. BMSCs were shown to have the best protective effects followed by hCMSCs and LRMSCs. SP-C expression re ects the secretory function of normal alveolar epithelia [26]. SP-C was signi cantly reduced due to alveolar epithelial damage caused by phosgene (Fig. 4b-d). The expression of SP-C mRNA in lung tissues in each group was detected by RT-PCR (Fig. 4b-d). SP-C mRNA expression was signi cantly decreased in the phosgene group. Meanwhile, MSCs treatment signi cantly increased SP-C mRNA expression, and the BMSCs group showed the highest increase.
Comparison of the proliferation potential of three types of MSCs by CCK8 assay The proliferation of the three types of MSCs were found not to be signi cantly different (Fig. 5a). However, the proliferation of BMSCs was slightly higher compared to that of hCMSCs and LRMSCs at 72h. All MSCs entered the logarithmic growth from the second to the fourth day.

Comparison of the migration potential of BMSCs, hCMSCs and LRMSCs
The horizontal migration of MSCs was different based on the wound-healing test (Fig. 5b). The BMSCs group exhibited signi cantly higher cell migration with obvious cell aggregation at 48h, followed by the hCMSCs group and LRMSCs group. The vertical migration of MSCs was determined using the transwell migration assay (Fig. 5c-d). At 24h, the number of hCMSCs, BMSCs and LRMSCs that migrated to the lower compartment was 181.27±62, 708.75±20 and 106.31±17, respectively. The three types of MSCs exhibited different migration rates and BMSCs showed the highest migration. This nding was consistent with the change in HGF in each treatment group in vivo. It is speculated that the high migration potential of BMSCs may be related to HGF, which may also be important in de ning the roles of different MSCs.
According to the experimental result (Fig. 5e), the optimal concentration of HGF was 25ng/mL which was added to the MSCs medium for 48h. MSCs with HGF showed a higher migration ability compared to the group without HGF (Fig. 5e). The experimental result indicated that HGF promoted the migration of MSCs in a dose-dependent manner in a certain range. Previous studies have shown that CXCL12 is the most important factor affecting the homing ability of hBMSCs compared with hCMSCs which was determined by PCR array in all the chemokines and receptors [9]. The effect of HGF on BMSCs migration was compared with that of CXCL12. The HGF group was compared with the CXCL12 group (30ng/mL) [20] and the FBS (10%) group (Fig. 5f). HGF was shown to signi cantly induce migration compared to CXCL12. HGF increased MSCs' migration by about 2 fold when compared with 10%FBS. These results indicated that HGF was a strong inducer of BMSCs' migration. Different concentrations of HGF were cocultured with hCMSCs and LRMSCs for 48 hours (Fig. 6a). The migration of LRMSCs and hCMSCs were increased with different doses of HGF (Fig. 6b, c). The migration ability in HGF (25ng/mL) treatment group was highest, which veri ed that the optimal concentration of HGF was 25ng/mL. The migration potential of the three types of MSCs after HGF treatment was different (Fig. 6d). BMSCs showed signi cantly increased migration potential when compared to LRMSCs and hCMSCs. This indicated that BMSCs had the highest HGF sensitivity among the three types of MSCs.
The expressions of proteins in MSCs were tested by Western blot (Fig. 7a-d) to explore the mechanism of the effect of HGF on MSCs. Compared with the control group, the expressions of p-GSK-3β and β-catenin in MSCs after HGF treatment were signi cantly increased in a dose-dependent manner that indicated that HGF activated this signaling pathway. Studies have con rmed the MMP-2 gene is an important target gene for the β-catenin signal pathway and its expression is associated with the degree of cell migration [27]. Therefore, the changes in MMP-2 mRNA after treatment were detected to investigate the effect of HGF on the target gene. Compared with the control group, MMP-2 expression after HGF treatment increased signi cantly (Fig. 7e). MMP-2 transcription was inhibited by the addition of FH535 (20μM), a potent inhibitor of the β-catenin signal pathway [28]. This showed that HGF promoted the transcription of MMP-2 by activating the β-catenin signaling pathway which promoted the migration of MSCs.

Discussion
Phosgene-induced lung injury remains a common critical disease with high morbidity and mortality, however, its speci c mechanism is not fully identi ed. In recent years, MSCs have become promising interventions for various diseases and are shown to be effective in both experimental and clinical practice. The main functions of MSCs are homing differentiation (limitation), immune regulation and secretion of cytokines. BMSCs, LRMSCs and hCMSCs have been widely studied in recent years. They are derived from the bone marrow, placental chorionic membrane and the lung respectively. Previous studies have shown that BMSCs, LRMSCs and hCMSCs play protective roles in lung injury caused by mechanical ventilation injury, endotoxin and plateau factors [6,8]. Other studies have reported that BMSCs can improve the lung injury caused by phosgene [4]. In this study, the roles of LRMSCs and hCMSCs in phosgene poisoning were con rmed and the effects of BMSCs, LRMSCs and hCMSCs in the same disease were compared. The results revealed that the three types of MSCs improved PaO 2 , lung injury score, leukocytes in ltration, and the expressions of various pro-in ammatory and anti-in ammatory factors in BALF and serum in rats. Besides, they promoted the recovery ability of functional cells to reduce the phosgene inhalation lung injury. BMSCs showed better activity and effect compared to LRMSCs and hCMSCs. Previously, BMSCs were reported to have better therapeutic potential compared with hCMSCs in naphthalene-injured mice [9]. These ndings suggested that BMSCs may be the most suitable MSC in the clinical treatment of phosgene inhalation lung injury.
This study revealed that different MSCs played different roles in the same disease model. However, this mechanism needs to be studied further. There was no signi cant difference seen in the proliferation of MSCs. However, BMSCs showed the highest migration ability compared to LRMSCs and hCMSCs. These ndings suggested that the differences in the role of the MSCs were partly related to their migration ability. In vivo, the expressions of HGF in each MSCs group differed and this was consistent with the trend of the effects of stem cells. Previous studies have reported that HGF can promote the directional migration of a variety of cells [13,14]. Moreover, HGF is closely related to the vascular endothelium function [29]. Therefore, it was speculated that HGF was related to the mechanism of the differential effects of the MSCs. Our results found that HGF affected the migration ability and the role of MSCs, and this was one of the important factors affecting the differences in the three types of MSCs. Besides, HGF was found to promote the migration ability of MSCs by activating the p-GSK-3β/β-catenin signal pathway. Previous studies have con rmed that activated β-catenin also plays an important role in the Wnt signaling pathway involved in phosgene inhalation lung injury. The activation of the β-catenin signaling pathway improves MSCs migration [18,19]. GSK-3β phosphorylation is reported to affect the βcatenin signaling pathway. HGF activated GSK-3β/β-catenin signaling pathway has been found to promote cell migration and repair activity. Recent oncology studies have found that the activation and expression of the c-Met receptor of HGF are indispensable, and the HGF/ c-Met pathway is currently one of the research focus of tumor migration [30][31][32]. However, c-Met changes in MSCs needs to be explored further.
According to the collected literatures and previous studies, the therapeutic effects of BMSCs from rats and BMSCs from human sources in phosgene induced lung injury were compared. However, no signi cant differences in the therapeutic effects have been reported when the two types of MSCs are used in the same disease model. Other studies have shown that there is no signi cant difference in the repairing effect of dermal MSCs from mouse and human sources on skin broblast tissue. Since the mice are genetically similar to humans, it can be concluded that the species origin was not the key reason for the different effects of MSCs in this experiment. In addition to the in uencing factors such as the source, differentiation state, passages and dosage, the effects of stem cells in diseases more depend on the key factors which include sensitivity, reactivity and genetic inheritance. The optimal treatment in animal models will need to undergo rigorous clinical trials in the future. This study payed attention to the migration of MSCs in the basic experiment.

Conclusions
In conclusion, this study suggested that BMSCs have the best protective effects in phosgene-induced lung injury, which providing a theoretical basis for the next clinical trial.

Availability of data and materials
The data used in supporting study are available from the corresponding author on reasonable request.

Ethics approval and consent to participate
The animal protocol in this work was in accordance with guidelines for the care and use of laboratory animals sanctioned by the Ministry of Science and Technology of the People's Republic of China and approved by Jinshan Hospital of Fudan University.

Consent for publication
Not applicable

Competing interests
The authors declare that they have no competing interests.