HucMSC-derived exosomes delivered BECN1 induces ferroptosis of hepatic stellate cells via regulating the xCT/GPX4 axis

Activated hepatic stellate cells (HSCs) are significant in liver fibrosis. Our past investigations have shown that human umbilical cord mesenchymal stem cells (hucMSCs) and their secreted exosomes (MSC-ex) could alleviate liver fibrosis via restraining HSCs activation. However, the mechanisms underlying the efficacy were not clear. Ferroptosis is a regulatory cell death caused by excessive lipid peroxidation, and it plays a vital role in the occurrence and development of liver fibrosis. In the present study, we aimed to study the proferroptosis effect and mechanism of MSC-ex in HSCs. MSC-ex were collected and purified from human umbilical cord MSCs. Proferroptosis effect of MSC-ex was examined in HSCs line LX-2 and CCl4 induced liver fibrosis in mice. Gene knockdown or overexpression approaches were used to investigate the biofactors in MSC-ex-mediated ferroptosis regulation. Results: MSC-ex could trigger HSCs ferroptosis by promoting ferroptosis-like cell death, ROS formation, mitochondrial dysfunction, Fe2+ release, and lipid peroxidation in human HSCs line LX-2. Glutathione peroxidase 4 (GPX4) is a crucial regulator of ferroptosis. We found that intravenous injection of MSC-ex significantly decreased glutathione peroxidase 4 (GPX4) expression in activated HSCs and collagen deposition in experimental mouse fibrotic livers. Mechanistically, MSC-ex derived BECN1 promoted HSCs ferroptosis by suppressing xCT-driven GPX4 expression. In addition, ferritinophagy and necroptosis might also play a role in MSC-ex-promoted LX-2 cell death. Knockdown of BECN1 in MSC diminished proferroptosis and anti-fibrosis effects of MSC-ex in LX-2 and fibrotic livers. MSC-ex may promote xCT/GPX4 mediated HSCs ferroptosis through the delivery of BECN1 and highlights BECN1 as a potential biofactor for alleviating liver fibrosis.


INTRODUCTION
Liver fibrosis is a pathophysiological process caused by chronic viral hepatitis and metabolic disorders. It is characterized by the deposition of many collagen fibers and the proliferation of hepatic stellate cells (HSCs) [1]. Currently, about 500 million patients are suffering from fibrotic liver diseases worldwide. Hepatic fibrosis could cause liver cirrhosis or liver cancers, and there is no effective treatment [2]. Increased activation of HSCs is the main factor leading to the progression of liver fibrosis. Therefore, HSCs are critical targets for ameliorating liver fibrosis [3]. Targeted induction of HSCs inactivation will help establish a more effective antifibrosis treatment method.
Mesenchymal stem cells (MSCs) are pluripotent stem cells with self-renewal and multidirectional differentiation capabilities. MSCs have been increasingly applied to clinical treatment, including acute and chronic liver injury [4,5]. Many clinical trials exploring MSC-based therapy have suggested good advantages in liver disease. However, the potential risks involved in cell therapy are still uncertain [6]. MSC-derived exosomes are membrane vesicles encapsulated with MSC-derived proteins, lipids, mRNAs, and noncoding RNAs [5,7,8]. MSC-ex has been proven to play a therapeutic role in eye diseases [9], wound regeneration [10], Parkinson's disease (PD) [11], and breast cancer [12]. Human bone marrow MSCs (BM-MSC) derived exosomes can inhibit the progression of liver fibrosis by regulating the Wnt/β-catenin signaling pathway [13]. Chorionic plate MSCs (CP-MSC) derived exosomes transferred microRNA-125b inhibits the activation of Hh signaling and promotes liver regeneration [14]. Moreover, umbilical cord MSCs derived exosomes can attenuate inflammation and liver damage in the acute inflammatory liver injury model [15]. In addition, our previous studies have demonstrated that MSC-ex can ameliorate CCl 4 -induced liver fibrosis [16,17]. These findings suggest that MSC-ex-based therapy can effectively alleviate liver fibrosis, but these curative effect mechanisms are unclear.
Ferroptosis is a new type of programmed cell death that is irondependent and different from apoptosis, cell necrosis, and autophagy [18]. Ferroptosis is related to the occurrence and development of many diseases, including Parkinson's syndrome [19], pancreatic cancer [20], and liver damage [21]. Several studies have demonstrated that ferroptosis induction could ameliorate organizational fibrosis, suggesting a new anti-fibrotic therapy [22].
It has been reported that HSCs ferroptosis could be mediated by chrysophanol [23], magnesium isoglycyrrhizinate [24], or artesunate [25] and thereby alleviated liver fibrosis. However, the role and mechanisms of MSC-ex in regulating HSCs ferroptosis and repairing liver injury is unknown. BECN1 is a crucial regulator of ferroptosis [26], but whether MSC-ex can transport BECN1 to induce HSCs ferroptosis has not been studied.
In the present study, we aimed to investigate the function of MSC-ex in HSCs ferroptosis and the mechanism of action. We observed that MSCs-ex promoted ferroptosis of HSCs. In addition, MSC-ex-derived BECN1 decreased xCT/GPX4 expression and inhibited HSCs activation, providing a novel insight into the proferroptosis effect of MSC-ex in liver fibrosis.

MATERIALS AND METHODS Cell culture
The human umbilical cord was gained from informed, healthy parturients, and MSCs were isolated as previously reported [27]. Isolated MSCs were maintained in α-MEM (Minimum Essential Medium-α) with 10% FBS (fetal bovine serum). Human hepatic stellate cell line LX-2 (#CC4023) was purchased from Saiku Culture Bank (Guangzhou, China) and cultured in H-DMEM (High glucose Dulbecco's Modified Eagle Medium) with 10% FBS. Human hepatic cell line L-02 cells was purchased from the Chinese Academy of Science and cultured in RPMI 1640 containing 10% FBS at 37°C with 5% CO2. Human fetal lung fibroblasts (HFL-1) were purchased from the Chinese Academy of Sciences and cultured with α-MEM containing 15% FBS at 37°C with 5% CO2. All the cell lines were authenticated by STR profiling and tested clean for mycoplasma contamination.

Isolation and characterization of MSC-ex
After reaching 80-90% confluence, the media of MSCs was changed to an FBS-free medium (Bioind) for 48 h and centrifuged at 10,000 × g for 30 min to remove cells and cell debris. MSC-ex was isolated and purified as reported previously [28]. The supernatant collected after that is filtered and sterilized through a 0.22 µm pore filter (Millipore). MSC-ex were then collected and washed with PBS by centrifugation at 10,000 × g for 3 h. Protein concentration of MSC-ex was examined by BCA assay (CWBIO, Beijing, China) and subjected to western blot analysis for CD9, CD63, Calnexin, TSG101, and β-actin. We observed the size and structure of MSCex with a transmission electron microscope (TEM) (FEI Tecnai 12; Amsterdam), Atomic Force Microscope (AFM) (Bruker, Germany), and nanoparticle tracking analysis (NTA; NanoSight, UK).

qRT-PCR
RNA from LX-2 or liver tissues was collected utilizing Trizol (GIBCO) following the manufacturer's instructions. cDNA was synthesized from the RNA using SuperScript Reverse Transcriptase Kit (Vazyme, Nanjing, China). Then qRT-PCR assays were performed using SYBR green (CWBIO). Relative gene expression normalized to β-actin was calculated using the 2 −ΔΔCt method. All primers were synthesized by Invitrogen and listed in Table 1. There were three replicates per group.
Immunofluorescence LX-2 was planted on coverslips in a 24-well plate and treated with PBS or MSC-ex (400 or 800 μg/ml). After MSC-ex treatment, LX-2 was fixed in 4% (w/v) paraformaldehyde (Sigma-Aldrich) for 15 min. LX-2 was then permeabilized with 0.1% Triton X-100 for 10 min. 5% BSA was utilized to block the cells at room temperature for 1 h. After that, LX-2 was hybridized with primary antibodies against α-SMA, GPX4, and xCT (Abcam) at 4°C overnight and visualized with fluorescence-labeled secondary antibodies (Abclonal) for 45 min at 37°C. Hoechst 3342 was used to stain nuclei at room temperature. Images were captured using a confocal microscope (Nikon, Tokyo).

Mouse model of liver fibrosis and MSC-ex injection
Mice (BALB/c, female, 4-5 weeks) were from Cavens animal laboratory Co., Ltd from Changzhou, China. Following the protocols approved by the Jiangsu University ethics committee (UJS-IACUC-AP-2020033127), a mouse model of liver fibrosis was established randomly with an intraperitoneal injection of 10% CCl4 (dissolved in mineral oil, 0.1 ml/100 g body weight) for eight weeks, twice a week. To analyze the effect of MSC-ex on GPX4, BECN1, and xCT expression and liver fibrosis, mice were injected with MSCex twice a week at a dose of 20 mg/kg body weight (n = 6) or PBS (n = 6). Liver samples were collected for further analysis at two weeks post-MSC-ex injection.

Sirius red staining
According to the manufacturer's instruction, the manufacturer's instruction measured collagen deposition in liver sections in Sirius Red (Chondrex, USA). To analyze hepatic collagen distribution, fibrotic septa randomly selected from the right and left liver lobes of 6 individual mice/groups were assessed. Images were acquired using a digital slide scanner (3DHISTECH, Hungary). Collagen extent was expressed as a percentage of stained area in each liver section.

Immunohistochemistry
The liver sections were deparaffinized and rehydrated for immunohistochemical staining for BECN1 and GPX4; the endogenous peroxidase was inactivated by 3% H2O2 for 30 min. Next, the liver slides were immersed into preheated antigen retrieval solution (0.01 M, pH 6.0, citrate buffer) for 30 min. After blocking with 5% BSA, the liver slides were probed with primary antibody against BECN1 (Proteintech) and GPX4 (Abcam) overnight at 4°C and incubated with biotin-conjugated anti-rabbit IgG and streptavidin-biotin. Finally, liver tissue sections were visualized with DAB Horseradish Peroxidase Color Development Kit (Boster, Wuhan, China) and counterstained with hematoxylin. Images were acquired using a pathological section scanner (3DHISTECH, Hungary).
Mitochondrial membrane potential assay JC-1 staining was performed to analyze the mitochondrial membrane potential in LX-2 or L-02 (Beyotime, Shanghai, China). JC-1 probe was diluted in fresh medium at 1:1000. Cells were incubated with a diluted JC-1 probe in the dark at 37°C for 20 min. After fixing with 4% paraformaldehyde, LX-2 or L-02 were counterstained with Hoechst 3342. Images were captured using a confocal microscope (Nikon, Tokyo, Japan). Mean fluorescent intensity was calculated using Image-Pro Plus 6.0 software. There were three replicates per group.

Iron concentration assay
MSC-ex (400 or 800 μg/ml) or BECN1 cDNA treated LX-2 or L-02 were lysed utilizing RIPA (Pierce, Rockford, USA), and iron concentrations were determined using an iron assay kit following the manufacturer's instructions (Abcam). There were three replicates per group.

GSH and MDA quantification analysis
Levels of GSH and MDA in MSC-ex (400 or 800 μg/ml) treated LX-2 or L-02 were detected with GSH assay kit (Beyotime, Shanghai, China) and MDA assay kit (Beyotime, Shanghai, China) following the manufacturer's instructions, respectively. There were three replicates per group.

Lentiviral knockdown of BECN1 in MSCs
Recombinant lentiviruses were produced by transient transfection in HEK293T cells. BECN1 shRNA sequences were synthesized and cloned into pPLK/GFP-Puro lentiviral plasmid to generate pPLK/GFP-Puro-BECN1 shRNA/ control shRNA. Sequences of BECN1 shRNA are designed using the following: GTGGACACGAGTTT CAAGA. Then HEK293T cells were cotransfected with three lentiviral plasmids of psPAX2 (packaging vector), pMD2G (envelope vector), and pPLK/GFP-Puro-BECN1 shRNA (transfer vector). GFP expression of HEK293T cells was evaluated, and supernatants were collected at 72 h post-co-transfection. MSCs were transfected with lenti-BECN1-shRNA/lenti-control-shRNA lentivirus. BECN1 expression in MSCs was examined with qRT-PCR and western blot. Exosomes of lenti-BECN1 shRNA transfected MSCs (Ex shBECN1 ) or lenti-control shRNA transfected MSCs (Ex sh-ctr ) were collected and purified. To investigate the regulation of Ex sh-ctr and Ex shBECN1 on GPX4, mice were injected with PBS (n = 6), Ex sh-ctr (n = 6), and Ex shBECN1 (n = 6) at a dose of 20 mg/kg body weight twice a week. At two weeks post Ex shBECN1 injection, liver samples were collected for further analysis.

Statistical analysis
Data are presented as means ± SD (standard deviation). The student's t-test used GraphPad Prism version 5.0 software to analyze statistical significance between two groups. One-way analysis of variance (ANOVA) followed with Dunnett was used for studies involving more than two groups. A two-sided P < 0.05 was considered statistically significant.
ferroptosis inhibitors NAC addition (Supplemental Fig. 1A-C). In addition, transmission electron microscopy images of mitochondrial structure demonstrated induced LX-2 ferroptosis in the MSC-ex group. MSC-ex treated LX-2 showed crumpled mitochondria with increased membrane density than the PBS group (Fig. 1H).
To elucidate the specificity of MSC-ex on LX-2 cell death, the effects of MSC-ex on human hepatic cell line L-02 and human fetal lung fibroblast (HFL-1) ferroptosis were also investigated with FDA assay and ROS assay. The FDA and ROS assay results showed that L-02 cell death and ROS production were decreased by MSC-ex ( Fig. 2A, B). MSC-ex exerted an inhibitory effect of L-02 ferroptosis. Expression of ferroptosis marker glutathione peroxidase 4 (GPX4) was also not changed in MSC-ex treated L-02 (Fig. 2C, D,  Supplemental Fig. 9B). MSC-ex treatment did not induce cell death and ROS production in HFL-1 (Supplemental Fig. 2). In conclusion, these data showed that MSC-ex caused HSCs ferroptosis in vitro.
MSC-ex-derived BECN1 downregulates System xc-/GPX4 expression in LX-2 xCT and SLC7A11 are components of the xc system amino acid antiporter, which import cysteine and export glutamate into cells [18]. BECN1 induces ferroptosis through binding to xCT and blocking xc-activity [29]. Thus, we hypothesized that BECN1 could initiate the proferroptosis effect of MSCs-Ex. We found that BECN1 was enriched both in MSCs and MSC-ex, whereas not observed in MSC-ex-free CM (Fig. 3A, Supplemental Fig. 9C). Downregulated xCT/GPX4 signaling contributes to ferroptosis, which is necessary to activate HSCs. Then, expression of BECN1, xCT, GPX4, and HSCs activation marker a-SMA were detected in MSC-ex treated LX-2. BECN1 mRNA level was not changed (Fig. 3B), whereas BECN1 protein was significantly increased (Fig. 3C, F, Supplemental Fig. 3, Supplemental Fig. 9D). Furthermore, expression of xCT and α-SMA mRNA in MSC-ex treated LX-2 decreased compared with PBS control (Fig. 3B). Similarly, immunofluorescence and western blot analysis confirmed that MSC-ex could downregulate xCT, GPX4, and a-SMA protein expression (Fig. 3D-F, Supplemental Figs. 3, 9D). These results show that MSC-ex may induce HSCs ferroptosis and inactivation through exosomes mediated BECN1 protein transfer but not BECN1 mRNA regulation.

MSC-ex induces BECN1 mediated ferroptosis in mouse fibrotic liver
Our previous report showed that MSC-ex could significantly decrease liver fibrosis by downregulating HSCs activation [16]. To investigate the mechanisms of the MSC-ex mediated proferroptosis effect, we speculated that MSC-ex delivered BECN1 induced ferroptosis of activated HSCs. Then the expression of ferroptosisand HSCs activation-associated biomarkers such as BECN1, xCT, GPX4, and a-SMA were examined in liver tissues from MSC-extreated mice. Consistent with the expression of BECN1 mRNA in MSC-ex treated LX-2 (Fig. 3B), qRT-PCR results showed that MSC-ex also did not affect BECN1 mRNA expression in fibrotic livers (Fig. 4A). xCT, GPX4, and a-SMA mRNA levels were downregulated in MSC-ex (10 or 20 mg/kg) treated mouse fibrotic livers (Fig. 4A). Immunohistochemistry staining assay and western blot analysis revealed that BECN1 level was increased in MSC-ex (10 or 20 mg/ kg) treated mouse fibrotic liver compared to which in CCl 4 group (Fig. 4B, C, Supplemental Figs. 4, 9E). xCT, GPX4, and a-SMA were cumulated in fibrotic livers and were decreased by MSC-ex ( Fig. 4B, C, Supplemental Fig. 4). Furthermore, Sirius red staining revealed that MSC-ex reduced collagen deposition in fibrotic livers (Fig. 4D). Thus, MSC-ex containing BECN1 may induce BECN1/ System xc-/GPX4 mediated ferroptosis and reduce HSCs activation in mouse fibrotic livers.

DISCUSSION
This study found MSC-ex-derived BECN1 induced HSCs ferroptosis and ameliorated liver fibrosis. In vitro, MSC-ex increased BECN1 expression in LX-2 through exosomal BECN1 transfer and subsequently inhibited the expression of xCT and GPX4 in HSCs. Then the word of α-SMA in LX-2 was decreased following ferroptosis. Our results showed that overexpression of BECN1 in LX-2 enhanced ferroptosis, whereas knockdown of BECN1 in LX-2 abolished ferroptosis, further suggesting that BECN1 was crucial in HSCs ferroptosis induction. In vivo, the ability of MSC-ex to induce ferroptosis and ameliorate liver fibrosis was weakened following BECN1 knockdown. Briefly, MSC-ex could enhance ferroptosis in HSCs and subsequently reduce CCl4-induced liver injury through exosomes/ BECN1/xCT/GPX4 pathway.
xCT mRNA levels were log-fold lower in primary mouse hepatocytes than in primary mouse HSCs. Inhibiting xCT in LX-2 cells by Erastin induced massive ferroptosis in HSCs did not affect hepatocytes [36]. Thus, we reasoned that MSC-ex-mediated xCT inhibition might preferentially promote LX-2 ferroptosis but not L-02 hepatocytes. L-02 cells were resistant to the treatment with MSC-ex, whereas the LX-2 cells were sensible.
Exosomes contain complex RNAs and proteins and have aroused widespread interest in the regeneration and repair of various tissues [5,8,37]. Evidence suggests that MSC has therapeutic potential in liver fibrosis. However, whether MSC-ex can induce ferroptosis remains unclear. We aimed to investigate the critical ferroptosis components present in MSC-ex. BECN1 is an important ferroptosis that can regulate various cellular processes [38]. Our findings confirmed that BECN1 was enriched in MSC-ex. Overexpression and knockdown of BECN1 in LX-2 suggested that BECN1 played crucial roles in ferroptosis and inactivation of HSCs. In vivo, we knocked down BECN1 in MSC-ex and confirmed the part of MSC-ex-derived BECN1 in ferroptosis and anti-fibrotic effects. These results indicated that MSC-ex exosomal BECN1 might be crucial for MSC-ex-mediated HSCs ferroptosis and liver fibrosis treatment. We compared BECN1 expression in MSC-ex treated LX-2 cells, and recombinant BECN1 cDNA transfected LX-2. As shown in Figs. 3F and 5B, BECN1 presentation in the MSC-ex 800 μg/ml group increased by 2.86 times, while the 1 μg BECN1 cDNA group increased by 3.01 times. Furthermore, xCT mRNA and protein expression in both MSC-ex 800 μg/ml group (Fig. 3A, F) and BECN1 cDNA 1 μg group (Fig. 5A, B) was significantly suppressed. MSC-exosomal BECN1 may suppress xCT activity through transcriptional inhibition. Thus MSC-ex treatment can correspond to the amounts of BECN1 levels after overexpression.
We further explored the underlying mechanisms by which MSCex derived BECN1 can induce ferroptosis and present hepatoprotective effects. BECN1, xCT, and GPX4 inhibition are all reported to trigger ferroptosis [39]. BECN1 can inhibit system xc-activity and induce lipid peroxidation through binding to xCT [38]. We then A Western blot analysis of BECN1 in exosomes from MSCs transfected with control shRNA (Ex sh-ctr ) or BECN1-shRNA (Ex shBECN1 ). B mRNA expression of GPX4 and α-SMA in Ex sh-ctr or Ex shBECN1 treated LX-2 (n = 3; *P < 0.05 compared with Ex sh-ctr group, **P < 0.01 compared with Ex sh-ctr group). C Western blot analysis of GPX4 and α-SMA expression in Ex sh-ctr or Ex shBECN1 treated LX-2. D ROS production in Ex sh-ctr or Ex shBECN1 treated LX-2 by DCF probe staining (n = 3; *P < 0.05 compared with Ex sh-ctr group). Scale bar, 20 μm. E mRNA expression of GPX4 and α-SMA in livers from Ex sh-ctr (20 mg/kg) or Ex shBECN1 (20 mg/ kg) treated mice (n = 3; *P < 0.05 compared with Ex sh-ctr group, **P < 0.01 compared with Ex sh-ctr group). F Western blot analysis of BECN1, GPX4, and α-SMA expression in livers from Ex sh-ctr (20 mg/kg) or Ex shBECN1 (20 mg/kg) treated mice. G Sirius Red staining of collagen deposition in livers from Ex sh-ctr (20 mg/kg) or Ex shBECN1 (20 mg/kg) treated mice (n = 6; **P < 0.01 compared with Ex sh-ctr group). ROS Reactive oxygen species. DCF, 2' ,7' Dichlorofluorescein. α-SMA Alpha smooth muscle actin. examined BECN1, xCT, and GPX4 expression in MSC-ex treated LX-2 and mouse fibrotic livers. BECN1 was significantly increased, whereas xCT and GPX4 expression were decreased in vitro and in vivo. Alteration of BECN1 affects xCT and GPX4 expression and ferroptosis. Our results suggested that MSC-exosomal BECN1 may promote ferroptosis by reducing the xCT expression. BECN1 is an autophagy initiation factor and has a central role in autophagy [26]. We found that BECN1 overexpression and MSC-ex upregulated autophagy marker LC3B expression promoted Fe2+ release in LX-2 cells. These results indicated that MSC-exosomal BECN1 might promote ferroptosis by inducing ferritinophagy. Furthermore, MSC-ex mediated LX-2 cell death, and ROS production was also attenuated by necroptosis inhibitor NEC-1. However, the exact role and mechanisms of MSC-ex in ferritinophagy and necroptosis regulation in HSCs need to be further elucidated.
Finally, NAC has significant protective activity against liver fibrosis through attenuation of oxidative stress. Our previous study also showed that MSC-ex could act as an antioxidant to reduce oxidative stress and rescue liver failure in CCl4 induced liver injury mouse model [8,17]. Our results showed that MSC-ex was a ferroptosis inducer to inactivate HSCs in vitro and vivo. Thus, we reasoned that combining the antioxidant effect of NAC and the proferroptotic effect of MSC-ex may produce a better efficacy in liver disease associated with fibrosis. However, NAC also interferes with the proferroptotic development of MSC-ex. The specific dose and effectiveness need to be further explored.

CONCLUSION
In summary, we demonstrated that MSC-ex targeted hepatic stellate cells activation in vitro and in vivo. This may be mediated by the delivery of BECN1 to induce ferroptosis via the downregulation xCT/GPX4 pathway (Fig. 7). These findings will provide different mechanisms for understanding the protective effect of MSC-ex against CCl 4 -induced liver fibrosis, and HSCs ferroptosis may be a new therapeutic approach for liver fibrosis treatment in the future.

DATA AVAILABILITY
All data generated or used during the study appear in the submitted article. Fig. 7 Schematic representation of hypothetic mechanism for BECN1 enriched MSC-ex mediating LX-2 ferroptosis through upregulating xCT/GPX4 signaling. MSC-ex can deliver BECN1 protein into HSCs and inhibits SLC7A11/xCT transcription in the nucleus. Decreased SLC7A11/ xCT results in cysteine deficiency and GSH reduction in HSCs, leading to GPX4 reduction and ferroptosis. MSC-ex Mesenchymal stem cell derived exosome, xCT cystine/glutamate exchange transporter. GSH glutathione, r-glutamyl cysteingl glycine, GPX4 Glutathione peroxidase 4.