MFGE8 regulates the EndoMT of HLMECs through the BMP signaling pathway and fibrosis in acute lung injury

Abstract

Background

To investigate the effects and mechanisms of MFGE8 on LPS-induced endothelial-to-mesenchymal transition (EndoMT) and pulmonary fibrosis in human lung microvascular endothelial cells (HLMECs) and a mouse model of acute lung injury.

Methods

Serum MFGE8 levels were compared between ARDS patients and controls. In vitro, HLMECs were treated with LPS, siRNA targeting MFGE8, and recombinant human MFGE8 (rhMFGE8).HLMEC morphology, invasion, migration, and EndoMT markers (CD31, ɑ-SMA) were evaluated. BMP/Smad1/5-Smad4 signaling and Snail expression were assessed via immunofluorescence, western blotting, and qRT-PCR. In vivo, rhMFGE8 effects on pulmonary fibrosis and EndoMT were analyzed in a mouse model of acute lung injury.

Results

MFGE8 levels were significantly reduced in ARDS patients, with higher levels correlating to better survival. In vitro, rhMFGE8 improved HLMEC morphology, reduced invasion and migration, and attenuated LPS-induced EndoMT by increasing CD31 and decreasing α-SMA. MFGE8 knockdown increased BMP/Smad1/5-Smad4 signaling and Snail expression, while rhMFGE8 inhibited these effects. In vivo, rhMFGE8 ameliorated pulmonary fibrosis and EndoMT in mice.

Conclusions

MFGE8 regulates LPS-induced EndoMT in HLMECs via the BMP/Smad1/5-Smad4 pathway and protects against pulmonary fibrosis in acute lung injury, suggesting it as a therapeutic target for ALI and ARDS.

Introduction

Acute lung injury (ALI) and its severe form of acute respiratory distress syndrome (ARDS) are caused by pulmonary aetiology (pulmonary infection, aspiration) and extrapulmonary aetiology (sepsis, pancreatitis, trauma), this diseases is diagnosed by doctors in the intensive care unit [1,2,3], and their pathological features are acute diffuse inflammatory injury accompanied by a lack of surfactant and acute respiratory distress. Traditionally, the pathological process of ALI is a continuous process that includes exudation, proliferation and fibrosis [3,4,5]. However, the process of fibrosis in ALI was recently shown to be initiated during the early stage. Ameliorating pulmonary fibrosis is beneficial for improving refractory hypoxemia, decreasing the use of ventilators and improving patient survival [3, 6]. The current research aim is to alleviate pulmonary fibrosis [7, 8].

Many studies have shown that myofibroblasts from endothelial cells (ECs) and can generate amounts of extracellular matrix (ECM) and deposited in the pulmonary interstitium, which can cause pulmonary fibrosis [9, 10]. Endothelial cells induced by inflammation, hypoxia, radiation, hyperglycemia, and TGF-β and then acquire a mesenchymal phenotype, this process was called endothelial-to-mesenchymal transition (EndoMT) [11]. Under specific physiological and pathological conditions, the polarity of EndoMT ECs is weakened, the tight junctions between ECs are lost, and the cells acquire the movement and contraction characteristics of mesenchymal cells. After these changes, the cells separate itself from the endothelial layer, and their migration and invasion are significantly enhanced after the cells migrate to the interstitium and transform into myofibroblasts. During this transformation, the cells change from cobblestone-like to elongated spindle-shaped cells, lose distinctive molecular markers of ECs, including TIE1, TIE2, von Willebrand factor (vWF), CD31 and VE-cadherin, and acquire specific molecular markers of mesenchymal cells, containing α-SMA, SFP1, vimentin, collagen of type I and N-cadherin [12].

Several studies have demonstrated that the generation of myofibroblasts depends on the transforming growth factor beta (TGF-β) signalling pathway [13,14,15]. Bone morphogenetic proteins (BMPs) was part of the TGF-β family and can bind on the endothelial cell membrane serine-threonine kinase receptors, including II and I type receptors [10]. Studies have shown that the receptor of BMP type I can promote Smad 1, Smad 5, and Smad 8 binding and undergo phosphorylation and these receptors can also associate with phosphorylated Smad4, form a complex in the nucleus and cooperate with other transcription factors to dominate the transcriptional responses of aim genes, which results in regulation of the EndoMT process [16,17,18].

Milk fat globule epidermal growth factor 8 (MFGE8) is an important molecule that mediates phagocytosis and is expressed in the alveolar epithelium, vascular ECs and macrophages [19]. It has been reported that this protein has crucial physiological and pathophysiological functions in various diseases, including liver, pulmonary and cardiovascular diseases. Laura W. Hansen found that MFGE8 deficiency leads to increased inflammation, tissue damage, neutrophil infiltration, and apoptosis, which causes sepsis in neonatal mice and results in elevated morbidity and mortality [20]. The Monowar Aziz study clearly indicated the importance of MFGE8 in improving neutrophil infiltration and demonstrated the potential of MFGE8 as a new therapeutic approach for ALI [21]. Furthermore, in recent years, a fair number of research have highlighted the potential function of MFGE8 in multiple of organs fibrosis, as revealed by the findings that MFGE8 slows the degree of skin and renal fibrosis in mice and reduces tissue fibrosis in mouse models of liver and cardiac fibrosis [19, 20, 22]. However, the molecular mechanisms of MFGE8 and EndoMT remain unclear and need to be examined. It was hypothesized that rhMFGE8 exerts a protective effect against EndoMT induced by lipopolysaccharide (LPS) via a BMP/Smad1/5/Smad4 signaling pathway.

Materials and methods

Clinical study and data collection

In the present study, 44 patients with ARDS and healthy volunteers from the Second Affiliated Hospital of Chongqing Medical University, between April and August 2021, were enrolled. A total of 11 patients died during their hospital stay, whereas 33 patients achieved unassisted respiration after discharge. The inclusion criteria were in accordance with the 2012 Berlin definition of ARDS [1]. The exclusion criteria included congestive heart failure, pulmonary disease induced by medicine, interstitial pulmonary disease, and connective-tissue diseases such as polymyositis, diffuse alveolar hemorrhage, carcinoma of cancer, and tuberculosis. When the patients were diagnosed with ARDS, we collected the patient’s blood and tested the expression level of MFGE8. Blood serum samples were obtained from the patients and volunteers and cryopreserved at -80 ˚C for subsequent experiments. The purpose of the study was explained to the healthy volunteers and the patients and their families, and all participants provided signed informed consent.

Cell culture

Human lung microvascular endothelial cells (HLMECs) were provided by company of ScienCell, locating in USA, (lot number: 3000) and stored at the College of Life Science. Cells were cultured in DMEM/F-12 medium obtained from ScienCell (lot number:09421), 1% penicillin–streptomycin were obtained by company of ScienCell (lot number: 0513), and 10% standard FBS was purchased from company of HyClone, locating in Cytiva (lot number: SH30084.03) in incubator with concentration of 5% CO2 and the temperature at 37 ℃.

Mouse model of ALI

Mice (C57BL/6, aged 7–8 weeks, male) were provided by the Laboratory Animal Center, which belong to Chongqing Medical University. In this study, all animal experimental procedures were performed using school-approved animal equipment at 24 ℃, with a 12 h light/dark cycle, and supplied sterile mouse chow and adequate water. The experimental procedures were implemented in a pathogen-free manner. Mice were healthy and checked daily by staff of Laboratory Animal Center. A total of 45 mice were separated into three groups according to a completely randomized design, 5 mice per cage. The mice in the control group were injected with physiological saline into the peritoneal cavity on the lower right side. Those in the LPS group were injected with 5 mg/kg LPS dissolved in physiological saline into the peritoneal cavity on the lower right side, and those in the LPS + rhMFGE8 group, on the first day, the mice were injected with 20 µg/kg rhMFGE8 into the peritoneal cavity on the lower right side, and then on the second day, the mice were continuously injected with 5 mg/kg LPS and 20 µg/kg rhMFGE8 into the peritoneal cavity on the lower right side for 5 days. At the endpoint of the experiment, to obtain lung tissue, the experimental mice were anaesthetized with sodium pentobarbital (60 mg/kg) and euthanized with potassium chloride (100 mg/kg) on days 7, 14 and 28.

Antibodies and reagents

LPS (from Escherichia coli 055: B5, 5 mg/ml, cat. no. L8880) was purchased from Solarbio Limited Company. LDN-2,128,549 (BMP receptor inhibitor, cat. no.1432597-26-6) was procured from APEXBIO Company. rh-MFGE8 (cat. no. 2767-MF) was purchased from R&D Systems, a biotech company. The MFGE8 ELISA kit (cat. no. SEB286Hu) was obtained from Cloud-Clone Limited Company. siRNA-MFGE8-1 (lot number: 430019983), siRNA-MFGE8-2 (lot number: 430084101), and siRNA-MFGE8-3 (lot number: 430048100) were obtained from Qiagen Limited Company. AB Antibody Technology provided primary antibodies, such as CD31 (1:1,000, cat. no. ab9498) and BMP4 (1:1,000, cat. no. ab39973) antibodies, and Cell Signaling Technology (Danvers, CT, USA) provided other antibodies, including α-SMA (1:1,000, cat. no. 19245), Snail (1:1,000, cat. no. 3879), Smad1 (1:1,000, cat. no. 6499), Smad4 (1:1,000, cat. no.46535), and Smad5 (1:1,000, cat. no.12534), p-Smad1/5 (1:1,000, cat. no.9576s) antibodies. Fluorochrome-conjugated goat anti-rabbit IgG antibody (lot number: bs-0295G-AF488) was purchased from Bioss Antibodies Limited Company. The Beyotime Institute of Biotechnology, locating in Shanghai municipality belong to China, provided secondary antibodies, including goat anti-rabbit IgG (HRP, lot number: A0208) and anti-mouse IgG (HRP, lot number: A0192).

Cell morphology and imaging

HLMECs were seeded in culture bottles and used in the logarithmic growth phase. The control cells were treated with physiological saline, the cells in the LPS group were treated with 20 µg/ml LPS dissolved in physiological saline [7], and the cells in the LPS + rhMFGE8 group were treated with administered 20 µg/ml LPS and 80 µg/ml rhMFGE8. The cells in the LPS + siRNA group were treated with 20 µg/ml LPS and 100 nM siRNA, and the cells in the LPS + siRNA + rhMFGE8 group were treated with 20 µg/ml LPS, 80 µg/ml rhMFGE8, and 100 nM siRNA. All treatments were performed for 96 h after which the morphology of cells was observed applying an inverted system microscope provided by Olympus Company and photographed.

Transwell migration assay

HLMECs were treated with LPS (20 µg/ml), siRNA, and/or rhMFGE8 as described above. Trypsin was used to digest adherent-growing cells, which was stopped by the addition of media. Subsequently, cells were centrifuged. We washed twice with PBS and add an appropriate amount of serum-free medium to resuspend the cells with concentration 1 × 10⁵ cells/ml, which were drop to upper chamber (8 μm pore size, Corning, 3413) in a 24-well plate. We dropped 600 µL medium containing DMEM-F12 into the lower chamber and the cells were incubated for 24 h. The chambers were washed and gently erased the upper unmigrated cells with a cotton buds. Dropped 4% paraformaldehyde to the migrated cells locating in the lower surface for fixing, 30 min and then added crystal violet to stain for 30 min. The cell numbers distributed in the five distinct regions were counted to evaluate ability of cell migration.

Wound-healing assay

Using a marker pen, five horizontal and vertical lines were drawn on a clean 6-well culture plate. The interval between the lines is 0.5 cm and clearly visible at the hole plate. Added cells with concentration of 5.0 × 10⁵ cells/ml to 6-well plates. When the confluence of the cells approximately covered 90%, putted sterile ruler on the 6-well plate along the straight line behind the hole plate and scratch the cell with sterile pipette tip, and washed the shedding of cells with PBS. Images were taken using an inverted system microscope provided by Olympus Company, and the migration distance was detected applying Image-Pro Plus 6.0.

Cellular immunofluorescence

The cells at a concentration of 2 × 10⁶ cells/ml. To detect surface CD31 expression, cells were incubated with 10 µL anti-CD31 for 30 min at 4 ℃, gently shaking, using normal IgG as isotype, and then cells were then incubated with sufficient amount of diluted primary antibody (CD31).To examine intracellular α-SMA expression 0.5%Triton X-100 was added with a pipete gun to disrupt the cell wall permeability at room temperature for 15 min and then incubated with 10 µL anti-α-SMA for 30 min at 4 ℃, gently shaking, using normal IgG as isotype, and then cells were then incubated with sufficient amount of diluted primary antibody (α-SMA). And then the treated cells were send into the department of Flow Cytometry in graduated college of Chongqing medical university and tested by technicians. The cell fluorescence intensity was tested by a FACSCanto II instrument and quantified applying Flow Jo software.

RNA interference

siRNAs against MFGE8 were obtained from Qiagen GmbH. Cells were transiently transfected using Lipofectamine^® 8000 (Beyotime Institute of Biotechnology) and 100 nM siRNA. Non-targeting siRNA (mock control) was also transfected in parallel to account for any effects caused by the transfection reagents. Western blotting and quantitative (q)PCR were used to confirm the efficiency of knockdown. The sequences of the siRNAs were siRNA-MFGE8-1, AGTGGAGAACACGAAT; siRNA-MFGE8-2, TCACTCCGGATAAAT; and siRNA-MFGE8-3, TGGCTGTCAGGAATTG.

Reverse transcription (RT)-qPCR

The RNA from the HLMECs was extracted applying TRIzol reagent (Takara, Code No. 9108) and transcribed into complementary DNA (cDNA) adopting the PrimeScript RT kit (Takara, Code No. RR047A). Quantitative real-time polymerase chain reaction (qRT‒PCR) was executed in a 50 µL reaction volume using SYBR Green (Takara, Code No. RR820A). β-Actin was applied as the reference gene, and the change in mRNA was quantified using the 2-ΔΔCq method. The primer pairs are exhibited in Table 1.

Table 1 Sequences of the primers

Western blotting

The protein was obtained from the HLMECs applying cell lysate, and BCA reagent (Solarbio cat. no. PC0020) was added incubated for 15–30 min at 37 ℃and then the protein concentration was analyzed by the standard curve. Added 40 µg of the sample protein to the hole above the PAGE gel, setting the voltage to 50 V. When the Marker protein is electrophoresis to the upper edge of the lower gel, adjust the voltage to 100 V to continue electrophoresis, and end the electrophoresis when the Marker protein is close to the bottom of the lower gel. Set the condition of transfer membranes and then the target proteins were transferred to PVDF (0.22–0.45 μm) membranes. The PVDF membranes were put into 5% nonfat milk to blocked for 1 h at room temperature, and then dilution target proteins were added to PVDF membranes for overnight at 4 °C, and TBST was dropped to remove the excess antibodies, and then sufficient amount of diluted secondary antibody was dropped to PVDF membranes for incubating. ECL reagents were used to visualize the signals. Image Lab software was used for densitometry analysis of the immunoblots.

Hematoxylin and eosin staining

Lung tissue blocks were embedded in dissolved paraffin and cut into suitable specimens which was put into xylene for deparaffinization. Ethanol was then used for dehydration. The specimens were soaked in haematoxylin for staining for dyeing and hydrochloric acid in ethanol to differentiate and then put into distilled water for 15 min, transferred into eosin solution for 2 min and dehydrated with ethanol. The sections were finally sealed using neutral resin.

Masson staining

The degree of lung fibrosis was determined using Masson staining. The specimens were dyed with Weigert ferrohematoxylin staining solution and then placed in hydrochloric acid in ethanol. Subsequently, the specimens were dyed with Ponceau acid fuchsin staining buffer and soaked into 5% phosphomolybdic acid. Next, samples were stained with 2% aniline blue for 3 min. The specimens were routinely dehydrated, incubated with xylene, and then observed using an optical microscope.

Sircol assay

Lung tissue was harvested and processed for Sircol-assay.According to the manufacturer’s instructions, soluble collagen secretion (collagens type I to V) in the conditioned media was determined using a Sircol Soluble Collagen Assay (Biocolor, lot number: S1000).

Statistical analysis

Statistical analysis was performed using SPSS version 22.0 (IBM, Corp.) statistical software and GraphPad Prism version 6.0 (GraphPad Software, Inc.). Data were compared using a two-tailed Student’s t-test or a one-way ANOVA. P < 0.05 was considered to indicate a statistically significant difference.

Results

MFGE8 is a protective factor in ARDS

The baseline characteristics of all the patients with ARDS and healthy volunteers did not differ significantly regarding age or sex (Tables 2 and  3). The expression of MFGE8 protein in serum samples collected from patients with ARDS and healthy volunteers was assessed using ELISA, and the results suggested that the expression of MFGE8 in the ARDS group (61.11 ± 22.77 pg/ml) was inferior to that in the control group (79.36 ± 14.9 pg/ml) (Table 2). The MFGE8 levels were markedly higher in patients who were successfully discharged from the hospital (survivors, 65.6 ± 18.94 pg/ml) than in those who died in the hospital (nonsurvivors, 45.09 ± 10.79 pg/ml) (Table 3). These data indicated that the high expression of MFGE8 was a protective factor associated with hospital survival.

Table 2 Clinicopathological characteristics

Table 3 Clinicopathological characteristics of the patients with ARDS based on survival

HLMEC EndoMT is induced by LPS in vitro

We cultured HLMECs with LPS (20 µg/mL) for 96 h and observed their morphology by microscopy, which revealed a clear changed from the cobblestone shape of ECs to spindle-like fibroblasts (Fig. 1A). In addition, we performed a wound-healing assay to detect endothelial cell migration and a Transwell assay to analyse invasion. The results indicated that endothelial cell migration and invasion were strengthened in the LPS group compared with the control group (Fig. 1B and C). The changes in the expression of CD31 and α-SMA protein were assessed by fluorescence microscopy (Fig. 1D), and LPS greatly inhibited the protein expression of CD31 and raised the protein expression of α-SMA. Western blotting showed the same changes at the protein level (Fig. 1E). The mRNA levels of CD31 and a-SMA in the LPS group were 0.67-fold and 2.3-fold of those found in the control group (Fig. 1F). These data showed that LPS could induce EndoMT in HLMECs.

Fig. 1

LPS treatment of HLMECs induces EndoMT. (A) HLMECs were induced with 20 µg/ml LPS for 96 h, and subsequently the morphology of cells was observed under a microscope. Magnification, × 100; scale bar, 50 μm. (B) HLMECs were induced with 20 µg/ml LPS for 48 h, and a Transwell migration assay was used to examine cell invasion. Magnification, × 100; scale bar, 50 μm. (C) HLMECs were induced with 20 µg/ml LPS for 24 h, and a wound-healing assay was used to assess cell migration. Magnification, × 100; scale bar, 50 μm. (D) HLMECs were induced with 20 µg/ml LPS for 96 h, and fluorescence microscopy was used to visualize CD31 and α-SMA expression. Magnification, × 200; scale bar, 20 μm. (E) HLMECs were induced with 20 µg/ml LPS for 96 h, and the protein expression of CD31 and α-SMA was examined using western blotting. β-Actin was used as the loading control. (F) HLMECs were induced with 20 µg/ml LPS for 96 h, and the mRNA expression levels of the CD31 and α-SMA genes were examined. Data are presented as the mean ± SD of three repeats. ^*P < 0.05 vs. control. HLMEC, Human lung microvascular endothelial cell; EndoMT, endothelial-mesenchymal transition

HLMEC EndoMT is increased following siRNA-mediated knockdown of MFGE8

The expression of MFGE8 at the protein level was decreased by ~ 40% in HLMECs treated with LPS (Fig. 2A). The mRNA levels of MFGE8 in the LPS group were 0.67-fold those of the control group (Fig. 2B). Subsequently, whether siRNA-mediated knockdown of MFGE8 could promote EndoMT in HLMECs was assessed. Specific siRNAs (siRNA-MFGE8-1, siRNA-MFGE8-2, and siRNA-MFGE8-3) were transfected into HLMECs to knockdown MFGE8 expression, and the efficiency of knockdown was assessed using western blotting (Fig. 2C) and RT-qPCR (Fig. 2D). The results illustrated that siRNA-MFGE8-2 decreased the expression of MFGE8 at the protein level by ~ 90% compared with that found in the negative control group (Fig. 2C). The mRNA levels of MFGE8 in the siRNA-MFGE8-1, siRNA-MFGE8-2 and siRNA-MFGE8-3 group were 0.36-fold, 0.12-fold, and 0.75-fold those in the negative control group, respectively (Fig. 2D).

Fig. 2

Analysis of the knockdown efficiency of MFGE8. (A) Western blotting was used to examine the expression of MFGE8 in LPS-induced HLMECs. β-Actin was used as the loading control. (B) mRNA expression levels of MFGE8 in LPS-induced HLMECs. (C and D) For knockdown three specific siRNAs and a non-targeting siRNA control were transfected into HLMECs to knock down MFGE8 expression, and the efficiency of knockdown was assessed using western blotting (C) and RT-qPCR (D). Data are presented as the mean ± SD of three repeats. ^*P < 0.05 vs. control. HLMEC, Human lung microvascular endothelial cell; MFGE8, Milk Fat Globule Epidermal Growth Factor 8; siRNA, small interfering

Therefore, siRNA-MFGE8-2 was used in the following experiments and is henceforth referred to as si-MFGE8. Microscopy was used to observe the cell morphology and it was found that the si-MFGE8 induced a change from the cobblestone morphology of ECs to a spindle-like morphology associated with fibroblasts and that this change was more prominent in the LPS + si-MFGE8 group compared with the si-MFGE8 group (Fig. 3A). The results from the wound-healing and Transwell assays illustrated that the siRNA-mediated knockdown of MFGE8 group increased the effect of LPS on endothelial cell migration and invasion (Fig. 3B and C). The changes in the protein expression levels of CD31 and α-SMA were assessed by fluorescence microscopy, and the results illustrated that the knockdown of MFGE8 markedly decreased the CD31 protein and increased α-SMA protein. These changes in expression were significant when both LPS and si-MFGE8 were administered, comparing with control group. (Fig. 3D). MFGE8 konckdown markedly enhanced the effect of LPS on EndoMT induction by decreasing CD31 and increasing α-SMA protein and gene expression (Fig. 3E and F). Taken together, these results indicated that the knockdown of MFGE8 facilitated LPS-induced EndoMT in HLMECs.

Fig. 3

Knockdown of MFGE8 exacerbated LPS-induced EndoMT in HLMECs. (A) Microscopy was used to observe the morphology of HLMECs. Magnification, x100; scale bar, 50 μm. (B) A Transwell migration assay was used to examine cell invasion. Magnification, × 100; scale bar, 50 μm. (C) A wound-healing assay was used to assess cell migration. Magnification, x100; scale bar, 50 μm. (D) Fluorescence microscopy was used to examine the protein expression of CD31 and α-SMA. Magnification, × 200; scale bar, 20 μm. (E) Protein and (F) mRNA expression levels of CD31 and α-SMA. Data are presented as the mean ± SD of three repeats. ^*P < 0.05. siRNA, small interfering RNA; HLMEC, Human lung microvascular endothelial cell; MFGE8, Milk Fat Globule Epidermal Growth Factor 8; LPS, lipopolysaccharide

rhMFGE8 attenuates LPS-induced EndoMT in HLMECs

To further assess whether exogenous MFGE8 could exert a protective effect on EndoMT, HLMECs were treated with rhMFGE8. As shown in Fig. 4A, B, and C, rhMFGE8 ameliorated the changes in cell morphology and markedly decreased the invasion and migration of ECs induced by LPS or LPS + si-MFGE8. In addition, immunofluorescence staining indicated that the expression of CD31 in the LPS + rhMFGE8 group was greater than that in the LPS group. The expression of CD31 in the LPS + si-MFGE8 + rhMFGE8 group was greater than that in the LPS + si-MFGE8 group, and the increase in CD31 expression in the LPS + si-MFGE8 + rhMFGE8 group was inferior to that in the LPS + rhMFGE8 group. The expression of α-SMA in the LPS + rhMFGE8 group was inferior to that in the LPS group, the expression of α-SMA in the LPS group was inferior to that in the LPS + si-MFGE8 group, the expression of α-SMA in the LPS + si-MFGE8 + rhMFGE8 group was inferior to that in the LPS + si-MFGE8 group, and the decrease in α-SMA expression in the LPS + si-MFGE8 + rhMFGE8 group was inferior to that in the LPS + rhMFGE8 group (Fig. 4D). The results of western blotting indicated that rhMFGE8 markedly attenuated the effect of LPS or LPS + si-MFGE8 on EndoMT induction by increasing the CD31 protein levels and decreasing the α-SMA protein levels (Fig. 4E). The changes in the mRNA expression levels were consistent with that of the western blot results (Fig. 4F). These data showed that the administration of rhMFGE8 reduced LPS- or LPS + si-MFGE8-induced EndoMT in vitro.

Fig. 4

rhMFGE8 attenuated LPS-induced EndoMT in HLMECs. (A) Microscopy was used to observe the morphology of HLMECs. Magnification, x100; scale bar, 50 μm. (B) Transwell migration assays were used to examine cell invasion. Magnification, × 100; scale bar, 50 μm. (C) A wound-healing assay was used to assess cell migration. Magnification, x100; scale bar, 50 μm. (D) Fluorescence microscopy was used to analyze the protein expression of CD31 and α-SMA. Magnification, x200; scale bar, 20 μm. (E) Protein and (F) mRNA expression levels of CD31 and α-SMA. Data are presented as the mean ± SD of three repeats. ^*P < 0.05. siRNA, small interfering RNA; HLMEC, Human lung microvascular endothelial cell; MFGE8, Milk Fat Globule Epidermal Growth Factor 8; LPS, lipopolysaccharide; rh, recombinant human; EndoMT, endothelial-mesenchymal transition

MFGE8 regulates EndoMT induced by LPS via the BMP/Smad1/5/Smad4 signaling pathway and alters the expression of EndoMT-associated transcription factors

Previous studies have illustrated that the Snail gene is associated with EndoMT [23]. In the present study, western blotting and RT-qPCR analyses showed that the protein and mRNA expression levels of Snail were greater in the LPS and LPS + si-MFGE8 groups compared with the control group. The expression of Snail in the LPS + rhMFGE8 group was inferior to that in the LPS group. The expression of Snail in the LPS + si-MFGE8 + rhMFGE8 group was inferior to that in the LPS + si-MFGE8 group, suggesting rhMFGE8 ameliorated the increased expression of Snail in the LPS and LPS + si-MFGE8 groups (Fig. 5A and C). Compared with the control group, the expression of BMP was upregulated in the cells treated with LPS and LPS + si-MFGE8. In contrast, rhMFGE8 blocked the upregulation of BMP expression in the LPS and LPS + si-MFGE8 groups, and the expression of Smad1/5 and Smad4 were increased (Fig. 5A and C). These results illustrated that this signaling pathway may be activated by LPS and ameliorated by rhMFGE8. The BMP inhibitor LDN-212,854 was used to verify the role of BMP, and the results indicated that LDN-212,854 reduced the activity of Smad1/5 and Smad4 in the LPS and LPS + si-MFGE8 groups (Fig. 5B). Additionally, phospho-SMAD1/5 levels were examined to confirm pathway activation and showed an increase in response to LPS, which was attenuated by rhMFGE8 treatment (Fig. 5A). Overall, these results showed that rhMFGE8 regulated EndoMT in ECs via the BMP/Smad1/5/Smad4 signaling pathway and altered the expression of associated transcription factors.

Fig. 5

LPS activated the BMP/Smad1/5/Smad1/4 signaling pathway. (A and C) Protein and mRNA expression levels of BMP, Smad1, Smad5, Smad4, Snail, and pSmad1/5. (B) The BMP inhibitor LDN-212,854 was used to confirm the relationship between LPS and the BMP signaling pathway. Western blotting was used to detect the protein expression levels of Smad1, Smad5, and Smad4. Data are presented as the mean ± SD of three repeats. ^*P < 0.05. HLMEC, Human lung microvascular endothelial cell; MFGE8, Milk Fat Globule Epidermal Growth Factor 8; LPS, lipopolysaccharide; rh, recombinant human; siRNA, small interfering RNA

rhMFGE8 reverses ECM deposition and plays a protective role during the early stage of EndoMT

We induced mice with LPS and successfully constructed a mouse injury model. The expression of MFGE8 protein in serum samples collected from mouse with ALI and control group was assessed using ELISA, and the results suggested that the expression of MFGE8 in the ALI group (346.34 ± 24.68 pg/ml) was inferior to that in the control group (862.89 ± 15.8 pg/ml). C57 mice were used to determine whether exogenous MFGE8 exerted a similar protective effect on EndoMT induced by LPS in vivo. The mice were divided into three groups: Control group, LPS group, and LPS + rhMFGE8 group. Lung sections were embedded in paraffin, and hematoxylin and eosin staining was adopted to test the mice lung tissues. The LPS group exhibited a thickened alveolar septum and the thickness increased over time. In contrast, rhMFGE8 reversed the increase in thickness induced by LPS (Fig. 6A). Masson staining results illustrated that LPS-induced ECM deposition during the early stage, and this LPS-induced ECM deposition increased over time, and the amount of deposited ECM was higher than that in the control group. The levels of ECM deposited in the LPS + rhMFGE8 group were significantly inferior to that in the LPS group (Fig. 6B). Sircol assay results showed that LPS-induced collagen increased depending times, the amount of collagen was higher than that in the control group.The levels of collagen in the LPS + rhMFGE8 group were obviously inferior to that in the LPS group (Fig. 6G-I). These results indicated that fibrosis in ALI was initiated during the early stage and that rhMFGE8 could reverse ECM deposition and collagen induced by LPS. In addition, the results of immunofluorescence staining indicated that the expression α-SMA (blue fluorescence) was increased by LPS compared with that in the control group but decreased by rhMFGE8 (Fig. 6C). In vivo, the western blot results indicated that LPS-treated group exhibited lower protein expression levels of CD31 and higher protein expression levels of α-SMA compared with the control group; the LPS + rhMFGE8 group showed higher protein expression levels of CD31 and lower protein expression levels of α-SMA compared with the LPS group (Fig. 6D-F). Taken together, these results illustrated that rhMFGE8 exerted a protective effect during the early stage of EndoMT in a mouse model of ALI induced by LPS.

Fig. 6

rhMFGE8 attenuated ECM deposition and exerted a protective role in the early stages of EndoMT. A total of 45 mice were divided into three groups: Control group, LPS group, and LPS + rhMFGE8 group. Mice were sacrificed after 7, 14, or 28 days. (A) Hematoxylin and eosin staining was used to observe the LPS-induced changes in cell morphology and the alveolar septum in the lung. Magnification, x100; scale bar, 100 μm. (B) Masson staining was used to examine ECM deposition in the lung tissues. Magnification, × 100; scale bar, 100 μm. (C) Fluorescence microscopy was used to analyze the protein expression levels of CD31 and α-SMA in lung tissue. CD31 is shown by red fluorescence, and α-SMA is shown by green fluorescence. Magnification, × 100; scale bar, 20 μm). (D-F) Western blotting was used to examine the protein expression levels of CD31 and α-SMA after 7,14, and 28 days. Data are presented as the mean ± SD of five repeats. (G-I) Collagen were measured by Sircol assay.*P < 0.05. HLMEC, Human lung microvascular endothelial cell; MFGE8, Milk Fat Globule Epidermal Growth Factor 8; LPS, lipopolysaccharide; rh, recombinant human; siRNA, small interfering RNA; ECM, extracellular matrix; EndoMT, endothelial-mesenchymal transition

Discussion

In this study, we first illustrated that LPS could induce EndoMT in HLMECs. Our previous studies showed that LPS, which is an important inflammatory inducer, could be used in cell and animal models to investigate the mechanism and feasible treatments of ALI [24, 25]. Toshio Suzuki et al. found that EndoMT induced by LPS is rely on the ROS signalling pathway in the cells and transient EndoMT was detected in the lung tissue of mice with septic, which could be rehabilitated [26]. Several environmental factors, including inflammation, hypoxia, radiation, hyperglycaemia and TGF-β, have been shown to trigger EndoMT [27,28,29]. During initiation of EndoMT, the molecular and structural properties of ECs are altered by various stimulating factors. Cells lose their endothelial properties (reduced EC markers including CD31 and vWF) and cell adhesion abilities and acquire high migratory potential and fibroblast-like phenotypes [11]. Our results are in agreement with previous research. This study demonstrated that LPS greatly inhibited the level of CD31 protein and increased the level of α-SMA protein, which showed that LPS could induce EndoMT.

MFGE8 was found in macrophages and acted as a vital role in reducing inflammation and sustaining tissue homeostasis [30, 31]. The present study demonstrates that nonsurvivors had significantly lower MFGE8 levels compared to survivors, suggesting a potential link between MFGE8 and survival outcomes in ARDS. This observation aligns with MFGE8’s known role in mitigating inflammation and promoting tissue repair. Future studies should further validate this association and explore MFGE8 as a potential prognostic marker. Our results show that endogenous MFGE8 levels decrease in response to LPS in both HLMECs and mouse lung tissue, which exacerbates the inflammatory response and fibrosis. This finding underscores the protective nature of MFGE8 in regulating lung injury. Xiao Wang et al. found that LPS suppresses MFGE8 gene expression in macrophages [32]. Atabai et al. discovered that the expression of MFGE8 was down-regulated, which could exacerbate the severity of lung injury [33]. We confirmed these reports and also showed that MFGE8 was decreased by approximately 40% in HLMECs after LPS stimulation and in ARDS patients. Furthermore, recent studies indicate that rhMFGE8 demonstrates biologic activity in mice [34], paralleling its effects on HLMECs. Evidence from the Atabai group also substantiates that MFGE8 exhibits activity in murine models [35], where recombinant mouse MFGE8 effectively influences endothelial-mesenchymal transitions and fibrosis, validating the cross-species functionality of rhMFGE8 [36]. In contrast to previous findings, Wang et al. illustrated that the protein expression of MFGE8 was elevated in the serum of patients with chronic pulmonary hypertension [19]. The distinction in the results of these studies are probably attributable to the discrimination in pathophysiological processes between ALI and pulmonary hypertension. Aziz et al. showed that exogenous supplementation with MFGE8 could reduce neutrophil and cellular infiltration induced by LPS by upregulating GRK2 expression and downregulating CXCR2 expression, which reduces lung injury [21]. A previous study revealed that MFGE8 is generated and separated from mesenchymal stem cells and acts as an autocrine factor that suppresses TGF-β signalling by combining to αvβ3 integrin and decreases ECM deposition and liver fibrosis in mice [37]. Our results that endogenous supplementation with rhMFGE8 could attenuate LPS-induced EndoMT in vitro are in accordance with Wang’s study, who showed that MFGE8 could exert a protective effect on EndoMT induced by TGF-β and could be a novel treatment strategy for fibrosis of heart [38].

Previous studies have demonstrated that the TGF-β and BMP families of growth factors act as mediators that promote EndoMT, including downstream Smad-dependent and Smad-independent proteins. After their phosphorylation, Smad1 and Smad5 recruit Smad4 and then create a complex that plays a crucial role in enhancing signal cascade conduction [13]. The increased levels of phospho-SMAD1/5 in the LPS and LPS + si-MFGE8 groups confirm SMAD pathway activation, which was ameliorated by rhMFGE8, suggesting a regulatory effect on SMAD1/5 phosphorylation during EndoMT. After shifting from the cytoplasm to the nucleus, the complexes interact with target transcription factors, including Slug, Snail, Twist and Zeb1. We found that the expression of BMP, Smad1/5 and Smad4 was elevated in response to stimulation with LPS, and LDN-212,854, a BMP/Smad1/5/4 signalling pathway inhibitor, was used to verify the involvement of this signalling pathway. The results illustrated that knockdown of MFGE8 up-regulated the activity of the BMP signalling pathway, whereas the administration of rhMFGE8 reversed this effect.

In this study, we discovered that LPS induced the deposition of blue-stained ECM in the alveolar interstitial space and around mouse pulmonary arteries at the early stage and that the level of deposition increased over time. Kun Xiao found that proinflammatory cytokines are the most effective inducers of EndoMT, which leads to inflammatory conditions that transform ECs into myofibroblasts and thereby facilitates fibrotic diseases [37]. In the lungs, fibrosis may occur as a result of abnormal remodeling following ALI and inflammatory diseases, idiopathic and autoimmune diseases. Pulmonary tissue architecture is determined by the balance among collagen production, deposition and removal, which is a dynamic regulation process [8]. When sustained collagen production exceeds collagen removal, massive of collagen is deposited in the pulmonary interstitium, resulting in tissue fibrosis. Atabai et al. demonstrated that mice with deficiency in the MFGE8 gene exhibit collagen accumulation in pulmonary tissues after bleomycin treatment and that rhMFGE8 is effective in binding and internalizing collagen and accelerating the removal of collagen. We also found that rhMFGE8 ameliorated ECM deposition and collagen induced by LPS. Chaung et al. reported that rhMFGE8 sharply attenuates sepsis in mouse caused by acute alcohol exposure and would be prove to be a safe and effective therapy for sepsis in alcohol abusers [39]. Hansen et al. used male and female newborn mice and showed that a lack of MFGE8 exacerbates lung injury and elevates ratio of mortality in sepsis of mice and that rhMFGE8 has a protective effect in sepsis of mice [20]. In addition to its impact on EndoMT, the protective effects of MFGE8 in this model may also arise from its broader anti-inflammatory and anti-fibrotic actions, which target multiple cellular populations, including macrophages and epithelial and mesenchymal cells. This multi-target effect aligns with MFGE8’s reported functions in other studies, where its role extends beyond EndoMT regulation to modulating inflammation and fibrosis across various tissue types. Therefore, the therapeutic benefits observed in our study might not be limited to EndoMT suppression but may also involve these additional mechanisms. Wu et al. study showed high MFGE 8 levels were significantly associated to reducing risk of mortality in sepsis using Cox hazards model. Furthermore, they found that septic mice demonstrated hepatocyte and kidney injury compared with sham-operated mice, and MFGE8 may be a protective factor in animal experiments [40].

The protective effects of rhMFGE8 on fibrosis in this study might partially result from its ability to attenuate acute lung injury. In vivo, rhMFGE8 reduced the thickening of alveolar septa, decreased ECM deposition, and improved lung morphology, all of which are indicative of ameliorated ALI. These findings align with previous reports highlighting MFGE8’s role in mitigating inflammation and promoting tissue repair [20, 21]. Thus, while our data indicate a significant reduction in EndoMT markers in rhMFGE8-treated mice, the observed anti-fibrotic effects may also arise from the mitigation of initial lung injury and inflammation. Further studies using lineage-tracing approaches and models distinguishing ALI from fibrosis would help delineate the relative contributions of these mechanisms.

The present study has several limitations. Patient lung tissues were not obtained, thus it was not possible to compare these tissues and see if similarities were observed with the mouse lung tissues. While our study observed decreased CD31 and increased α-SMA protein expression following LPS treatment, these findings alone may not definitively confirm EndoMT, as similar changes could reflect endothelial injury or responses from stromal cell populations that express α-SMA upon injury. Future studies should incorporate lineage-traced endothelial cells and a broader panel of myofibroblast markers to confirm EndoMT. This additional approach would allow for a more accurate distinction between endothelial-derived mesenchymal cells and other cellular sources in this model. The serum expression levels of MFGE8 in patients only when they were diagnosed with ARDS, but the serum expression levels of MFGE8 after the patients had recovered were not assessed; thus, these comparisons will be performed in future studies. Additionally, rhMFGE8 was only administered to mice for only 6 days; longer administration periods should be assessed to better determine the efficacy and safety of rhMFGE8 administration. Further studies distinguishing cases with fibroproliferative progression from those primarily affected by acute alveolar damage could enhance our understanding of MFGE8’s role across ARDS progression stages.

In conclusion, these results indicate that MFGE8 regulates LPS-induced EndoMT of HLMECs via the BMP/Smad1/5/Smad4 signaling pathway and plays a protective role in pulmonary fibrosis induced by acute lung injury.