C57BL/6J adult mice were purchased from Animal Experimental Center of Xi’an Jiaotong University Health Science Center, and Mfge8-knockout (Mfge8-KO) mice were purchased from Nanfang Biotech Technology Co., Ltd (Shanghai, China). Mfge8-KO mice were obtained by knocking out 2–6 exons of mfge8 gene using CRISPR/Cas9 gene editing technology. All experimental animals are housed in a temperature-controlled room on a 12-h light/dark cycle with ad libitum access to food and water. The mice were fasted for 12 h before collecting samples. The study protocol was approved by the Institutional Animal Care and Use Committee of the Ethics Committee of Xi’an Jiaotong University Health Science Center.

Chronic pancreatitis was induced in male adult mice by six intraperitoneal injections of cerulein (50 μg/kg/body weight, C6660, Solarbio, Beijing, China) twice a week for 10 weeks, as described by Sendler M. et al. (Sendler et al., 2015) and used by us recently (Ren et al., 2020). During the last 5 weeks, 1 hour after cerulein injection, normal saline (vehicle) or 20 μg/kg recombinant murine MFG-E8 (RD System, Inc. Minnesota, United States) was administered through intraperitoneal injection. The animals were sacrificed 2 days after the last injection of cerulein. The doses of MFG-E8 used in this study were chosen on the basis of our previous publications in acute pancreatitis (Ren et al., 2021). Blood and tissue samples were collected.

Human pancreatic stellate cells (PSCs) were purchased from FengHui Biotechnology (FHHUM-CELL-0124, China) and cultured in Ham’s F-12K medium (PM150910C, Procell, Wuhan, China) with 20% fetal bovine serum (164210-100, Procell, Wuhan, China) in a humidified incubator at 37°C with 5% CO₂. HPSCs (1X10⁶/well) were planted into 6-well plates for 24 h. The cells appeared to be in quiescent state. Then, the cells were treated with 5 ng/mL TGF-β1 (5154LC, Cell Signaling Technology, United States) with 10 ng/ml or 20 ng/mL recombinant human MFG-E8 or equal volume of PBS for 24 h. In additional groups of cells, QX77 (5 ng/ml, S6797, SELLEK, United States), a LAMP2A-specific activator (Zhang et al., 2017), was added simultaneously with 20 ng/mL MFG-E8 in TGF-β1-treated HPSCs. 24 h later, the cells were collected for various measurements.

Pancreatic tissue sections were stained with H&E. The pathological staining scoring system introduced by Schmidt et al. (Schmidt et al., 1992) was used to evaluate the pancreatic tissue damage.

Immunohistochemical staining was performed as we described before (Bi et al., 2020). Paraffin sections of mouse pancreatic tissue were prepared, and Sirius red and Masson-Goldner staining were used to indicate the degree of tissue fibrosis. α-smooth muscle actin (α-SMA) staining (A5228, mouse monoclonal clone, Sigma-Aldrich, United States) and Collagen I (ab34710, rabbit polyclonal, Abcam, United States) staining were used to mark the deposition of extracellular matrix. MPO (ab45977, rabbit polyclonal, Abcam, United States) was used to mark the infiltration of neutrophils. Gr1 (LY6G) (ab25377, antibody, Abcam, United States), CD11b (ab216445, rabbit polyclonal, Abcam, United States) and F4/80 (ab240946, rabbit polyclonal, Abcam, United States) were used to demonstrate macrophage infiltration, LC3B (ab48394, rabbit polyclonal, Abcam, United States) were used to demonstrate autophagy levels. The immunohistochemical staining was photographed with a light microscope. Three fields were randomly selected for each image, and the images were quantitatively and statistically analyzed with ImageJ Pro Plus 6.0 software.

The expression of Collagen I, α-SMA and DHE were assessed by immunofluorescence staining as described by us previous (Ren et al., 2020). Quantitative determination of fluorescence intensity was perfumed by Image Pro Plus 6.0 software.

The mouse IL-6 ELISA kit (SEA079Mu, Cloud-Clone Corp USCN Life Science, Wuhan, CN), tumor necrosis factor-α (TNF-α) ELISA kit (SEA133Mu, Cloud-Clone Corp USCN Life Science, Wuhan, CN), IL-10 ELISA kit (SEA056Mu, Cloud-Clone Corp USCN Life Science, Wuhan, CN) and MFG-E8 ELISA kit (SEB286Mu, Cloud-Clone Corp USCN Life Science, Wuhan, CN) were used for the detection of IL-6, TNF-α, IL-10 and MFG-E8 according to the manufacturer’s instructions.

Pancreatic tissue and HPSCs homogenate was obtained and superoxide dismutase (SOD), total antioxidant capacity assay kit with FRAP method (FRAP value), glutathione (GSH) and malonaldehyde (MDA) were measured as we described before (Ren et al., 2019a; Ren et al., 2019b; Ren et al., 2020).

Pancreatic tissues were lysed in cold RIPA (P0013B, Beyotime, Beijing, China). The protein concentration was evaluated with the BCA Protein Assay Kit (P0012S, Beyotime, Beijing, China). After gel electrophoresis, the protein was transferred to PVDF membrane and incubation in blocking solution (3% BSA or 5% skimmed milk) at room temperature. Then the membranes were incubated overnight at 4°C with the primary antibodies (Supplementary Table S1). Primary antibodies were diluted in Primary Antibody Dilution Buffer for Western Blot (P0256, Beyotime, Beijing, China). Membranes were washed and then incubated with specific HRP‐conjugated secondary antibodies (Supplementary Table S1) for 1.5 h at room temperature. Bands were developed using Digital gel image analysis system (Bio‐Rad, California, United States) and quantitative of protein level were calculated by ImageJ2x software. Information about the antibodies used in this study are listed in Supplementary Table S1.

All measurement data are expressed as the mean ± standard error (SEM). The t-test or one-way ANOVA with the Tukey-Kramer test was used to analyze the differences between groups. All analyses were conducted with data statistics software GraphPad Prism version 8.0 (GraphPad Software, Inc., San Diego, CA, United States). p < 0.05 represented a significant difference.

To evaluate the pathophysiological role of MFG-E8 in CP, we induced CP in Mfge8-KO mice (The efficacy of MFG-E8 knockout was confirmed by Western blotting, Supplementary Figure S1) by repeated intraperitoneal injection of cerulein. H&E staining showed that MFG-E8 deficiency did not result in any significant changes in pancreatic histomorphology in sham mice (Figure 1A). However, repeated cerulein injection caused much more severe pancreatic damage in Mfge8-KO mice than their WT littermates (Figures 1A,B). Mfge8-KO mice also had larger area of necrosis than WT mice after repeated cerulein injection (Figures 1A,C). Masson and Sirius red staining showed more severe fibrosis in the pancreatic tissue of Mfge8-KO mice than that of WT mice (Figures 1A,D). Consistently, α-SMA and Collagen I staining showed that repeated cerulein injection caused extracellular matrix deposition in the interstitial space of Mfge8-KO mice than that of WT mice (Figures 1A,D). These findings were confirmed by western blot analysis of α-SMA and Collagen I protein expression in the pancreatic tissues (Figures 1E,F). MFG-E8 Deficiency also potentiated inflammatory responses in CP. As shown in Figure 1G, serum proinflammatory cytokines TNF-α and IL-6 were further increased, while anti-inflammatory cytokine IL-10 was further decreased in Mfge8-KO mice than WT mice after repeated cerulein injection. Inflammatory cell infiltration in the pancreas was measured by F4/80, CD11b, and Gr1 staining. As shown in Figures 1H,I, repeated cerulein injection also resulted in more inflammatory cell infiltration in the pancreas of Mfge8-KO mice than that of WT mice.

To further investigate the role of MFG-E8 in CP, we measured MFG-E8 protein expression in the pancreas after repeated cerulein injection by western blot analysis. As shown in Figures 2A,B, repeated intraperitoneal injection of cerulein significantly decreased MFG-E8 protein levels in the pancreas of WT mice. Because MFG-E8 is a secreted protein, we also measured the serum MFG-E8 levels in mice. As shown in Figure 2C, serum MFG-E8 levels in cerulein-treated mice were significantly lower than those in the control group, while repeated intraperitoneal injection of exogenous MFG-E8 effectively maintained serum MFG-E8 levels (p < 0.05). Administration of recombinant MFG-E8 significantly reduced pancreatic injury (Figures 2D,E) and necrosis (Figure 2F) in cerulein-CP mice. Pancreatic fibrosis was evaluated by Sirius red, Masson, α-SMA and Collagen I staining. As shown in Figures 2D,G, exogenous MFG-E8 treatment reduced the positive staining of the above fibrosis-related indicators in the pancreatic tissue of cerulein-CP mice by 55.2, 59.5, 68.7, and 54.6%, respectively (p < 0.05). Western blot analysis also confirmed that α-SMA and Collagen I protein expression in pancreatic tissue of cerulein-CP mice was downregulated by MFG-E8 treatment (Figures 2H,I). Similar beneficial effects of exogenous MFG-E8 treatment was observed in repeated L-arginine injection-induced CP in mice (Supplementary Figures 2A–F). MFG-E8 also inhibited inflammatory responses in cerulein-CP mice. As shown in Figures 2J,K, administration of exogenous MFG-E8 reduced the numbers of F4/80, CD11b and MPO positive cells in the pancreas of cerulein-CP mice by 82.6, 72.5 and 47.9%, respectively (p < 0.05). The abnormal serum levels of inflammatory mediators (IL-6 and IL-10) in cerulein-CP mice were restored to almost sham levels by exogenous MFG-E8 treatment (Figure 2L).

Impaired autophagy and oxidative stress activate PSCs and promote their release of large amounts of extracellular matrix (ECM), which, along with collagen deposition, initiates and accelerates the progression of pancreatic fibrosis (Bhardwaj and Yadav, 2013; Ryu et al., 2013; Diakopoulos et al., 2015; Li et al., 2018a). To investigate the mechanism responsible for MFG-E8’s beneficial effects in CP, we measured indicators of autophagy and oxidative stress. We have confirmed that deleting the exons 4 to 6 of the MFG-E8 gene (mfge8-knockout) has no significant effect on the antioxidant capacity (Ren et al., 2021) and autophagy (Supplementary Figure S3) in the mouse pancreatic tissue. In this study, we further explored the effects of exogenous MFG-E8 on the levels of oxidative stress and autophagy in the pancreas of cerulein-treated CP mice. As shown in Figures 3A,B, pancreatic levels of ATG7, ATG5 and LC3 II/LC3 I increased, while P62 decreased significantly after repeated cerulein injection, indicating activated autophagy process in CP. Exogenous MFG-E8 treatment reversed these changes in cerulein-CP mice, suggesting MFG-E8 suppresses autophagy in CP. The enhancement of autophagy could induce the disorder of oxygen free radical regulation, resulting in oxidative stress (Li et al., 2021). Our results also indicated that repeated cerulein injection induced oxidative stress in the pancreas. As shown in Figures 3C,D, DHE staining in the pancreas increased dramatically after repeated cerulein injection. Consistently, pancreatic tissues levels of MDA (Figure 3E) were also significantly elevated in cerulein-CP. In the meantime, anti-oxidative indicators including FRAP (Figure 3F), GSH (Figure 3G) and SOD (Figure 3H) decreased after repeated cerulein injection. Exogenous MFG-E8 treatment decreased DHE, MDA, and increased FRAP, GSH, and SOD in the pancreas of cerulein-CP mice, suggesting MFG-E8 reduces oxidative stress in CP.

PSCs activation plays a fundamental role in the development of pancreatic fibrosis. Activated PSCs have upregulated α-SMA expression and release a large amount of extracellular matrix proteins such as collagen I. To determine the effects of MFG-E8 on PSCs activation in vitro, we treated human PSCs with TGF-β1 in the presence of various concentrations of MFG-E8. As shown in Figures 4A–D, MFG-E8 dose-dependently suppressed TGF-β1-induced collagen I and α-SMA production in cultured human PSCs. MFG-E8 also blocked TGF-β1-induced autophagy (Figures 4E,F) and oxidative stress (Figures 4G–K) in cultured human PSCs.

TGF-β1 treatment increased ER stress-related protein GRP78 expression and PERK phosphorylation in human PSCs (Figures 5A,B), suggesting activated ER stress. MFG-E8 decreased TGF-β1-induced GRP78 expression and PERK phosphorylation in human PSCs. ER stress can lead to CMA (Li et al., 2018b). LAMP2A is the rate-limiting receptor for CMA substrate flux. And increased CMA activity leads to MEF2D degradation (Li et al., 2017). As shown in Figures 5A,B, TGF-β1 also increased LAMP2A expression and decreased MEF2D expression in human PSCs, suggesting increased CMA activity. MFG-E8 decreased TGF-β1-induced LAMP2A express, while increased MEF2D expression in the meantime. Exogenous MFG-E8 also reduced ER stress and CMA in cerulein-treated CP mice (p < 0.05, Supplementary Figures S4A,B). QX77 is a specific CMA activator. It can upregulate LAMP2A expression (Zhang et al., 2017). As shown in Figures 5C,D, QX77 reversed MFG-E8’s effects on LAMP2A and MEF2D expression. To explore the role of CMA in MFG-E8’s effects on PSC activation, the expression of collagen I and α-SMA was measured. As shown in Figures 5E–H, QX77 eliminated MFG-E8’s effects on collagen I and α-SMA expression. Similarly, the suppressive effect of MFG-E8 on oxidative stress in activated PSCs was also mitigated by the addition of QX77 (Figures 5I–M).