Ferrostatin-1 improves prognosis and regulates gut microbiota of steatotic liver transplantation recipients in rats

Liver transplantation (LT) remains the only therapeutic option for acute liver failure, end-stage liver disease and malignant liver cancer [1]. However, increasing imbalance between donor liver shortage and liver transplant candidates forces transplant centers to utilize extended-criteria donors, including donation after circulatory death and steatotic livers [2]. With the rising prevalence of obesity, metabolic syndrome and type 2 diabetes, the incidence rate of nonalcoholic fatty liver disease (NAFLD) is increasing, affecting approximately 25% of the population. The increasing prevalence of NAFLD has multiple impacts on liver transplant donors, with early fatty liver disease accounting for approximately 25% of liver transplant donors. The United Network for Organ Sharing database confirms that NAFLD has reduced the liver donor pool, and overall utilization of liver grafts is predicted to decrease from 78 to 44% by 2030 [3]. Although liver transplantation technology is maturing, the abandonment rate due to severe fatty liver disease remains high [4]. Therefore, optimizing the utilization rate of severe steatotic liver grafts in liver transplantation is crucial.

Steatotic liver transplantation ischemia reperfusion (IR) can damage recipient liver and extrahepatic organs, especially the heart, lungs and intestines. These organ injuries are also the main factors leading to poor prognosis following liver transplantation, especially intestinal injuries, which can increase perioperative complications of liver transplantation, leading to increased perioperative mortality [3]. Although ischemic preconditioning, preservation solutions and machine perfusion had encouraging preclinical data, they still failed to make a significant difference in clinical practice [5].

Protecting intestinal mucosal barrier function is an important element of perioperative management of liver transplantation [6,7]. Intestinal symbiotic bacteria and their products can reach the liver through the portal vein, thereby regulating liver innate and adaptive immunity, causing liver inflammation and promoting liver fibrosis [8–10]. Some specific bacterial populations are associated with the occurrence of bloodstream infections before and after transplantation [11]. LT significantly affects the gut microbiota [12,13]. Moreover, ecological imbalance following LT seems to be related to the colonization of multidrug-resistant bacteria [14].

Recent research indicates that progression from simple fat to fatty hepatitis to liver fibrosis is closely related to increased lipid peroxidation, and is caused by accumulation of fatty acids and large amounts of iron in liver cells [15–17]. Iron overload is another risk factor for liver ischemia-reperfusion injury (IRI) [18]. Our recent research showed that regulating ferroptosis in steatotic liver grafts can improve IRI in steatotic liver transplantations [19,20].

Therefore, the present study aimed to establish a rat model of liver transplantation with severe steatotic liver grafts. Using 16S rRNA amplicon sequencing and bioinformatics algorithms to identify key subgroups of the gut microbiota, we explored the effects of ferroptosis inhibitor Ferrostatin-1 (Fer-1) on improving the prognosis of LT recipients with severe steatotic liver grafts in rats. We also assessed changes in the gut microbiota, providing research directions for expanding the donor liver pool and improving the prognosis of recipients.

Materials & methods

Establishment of animal & severe steatotic liver graft models

Specific pathogen-free grade rats (China Food and Drug Administration, Beijing, China) were used. We obtained severe steatosis liver grafts following a high-fat diet for 12 weeks (final body weight 450–500 g). Hematoxylin and eosin staining showed that the area of mixed macrovesicular steatosis was more than 60% under the microscope, indicating that a model of severe steatotic liver was established successfully [13]; liver transplantation recipients were male Sprague–Dawley rats fed on a normal basal diet until 22–24 weeks of age (final body weight 450–500 g). All animal experiments followed guidance for the care and use of experimental animals, and the study was reviewed and approved by the Animal Ethics Committee of Nankai University (ethics no. 2021-SYDWLL-000331).

Establishment of a rat liver transplantation model

Nonarterialized orthotopic liver transplantation was performed in rats using the double-cuff technique [21]. Recipients were divided into three groups: sham operation (sham group), steatotic liver transplantation (LT group) and Fer-1 group. The Fer-1 group received intraperitoneal injection (10 mg/kg/day) 24 h before surgery (MCE, Shanghai, China) [19]. A total of 24 recipients (six per group) were used for survival analysis; 35 rats were used for biochemical analysis (blood, liver and intestinal tissue: five in the sham operation group and five in the other experimental groups at postoperative day 1 (POD 1) and postoperative day 7 (POD 7).

Constructing an injury model using IAR20 cells & steatotic IAR20 cells

Lipopolysaccharide (LPS; 10 μg/ml) was administered to rat hepatocyte IAR20 cells for 24 h to establish an IAR20 cell injury model. Cells were divided into 1) NC (normal IAR20 cells); 2) NL (10 μg/ml LPS to stimulate normal IAR20 cells for 24 h); and 3) NF (10 μg/ml LPS plus 2 μM Fer-1) groups.

Palmitic acid (Sigma-Aldrich, MO, USA; 100 μM) and 200 μM oleic acid (Sigma) were used to stimulate IAR20 cells for 24 h to obtain adipose IAR20 cells. These were divided into 1) FC (adipose degeneration IAR20 cells); 2) FL (10 μg/ml LPS to stimulate adipose degeneration IAR20 cells for 24 h); and 3) FF (10 μg/ml LPS plus 2 μM Fer-1) groups. The activity of IAR20 cells and levels of reactive oxygen species (ROS) and ferroptosis-related proteins were measured.

Constructing an injury model of intestinal epithelial cells

LPS (10 μg/ml) was applied to growing rat intestinal epithelial IEC-6 cells and human colon adenocarcinoma Caco-2 cells for 24 h to establish injury models of human and rat intestinal epithelial cells. They were divided into 1) NC (normal intestinal epithelial cells); 2) NL (10 μg/ml LPS the stimulate normal intestinal epithelial cells for 24 h); and 3) NF (10 μg/ml LPS plus 2 μM Fer-1) groups.

Liver function test

Collected sera were analyzed by a fully Cobas 800 automated biochemical analyzer (Roche, Basel, Switzerland) to assess liver function by measuring alanine aminotransferase (ALT) and aspartate aminotransferase (AST) and alkaline phosphatase (ALP) enzyme activities, and total bilirubin (TBil) and total cholesterol levels. Very-low-density lipoprotein (VLDL), low-density lipoprotein (LDL) and high-density lipoprotein (HDL) contents were also measured. All operations were performed according to the manufacturer's instructions.

Histopathology

Liver and intestinal tissues were fixed in 4% formalin then embedded in paraffin. Slices 4 μm thick were then stained with hematoxylin and eosin.

Immunohistochemistry

After sectioning tissue specimens, paraffin was dewaxed with xylene and samples were dehydrated using a graded alcohol series. Antigen repair was performed by heating in ethylene diamine tetraacetic acid (pH 9.0) in a microwave oven for 15 min. Tissue slides were then incubated in hydrogen peroxide for 30 min to block endogenous peroxidase activity, and further blocked with 5% normal goat serum. Slides were incubated overnight at 4 °C with antibody-recognizing glutathione peroxidase 4 (GPX4; 1:500), cyclooxygenase-2 (COX2; 1:500), myeloperoxidase (MPO; 1:300) or CD68 (1:500). Slides and biotinylated secondary antibody were then incubated at room temperature for 60 min, and streptavidin-conjugated horseradish peroxidase was added and incubated at room temperature for 30 min. Finally, slides were stained with diaminobenzidine and restained with hematoxylin.

Transmission electron microscopy

Fresh liver and ileocecal intestinal tissue samples were cut into 1 × 1 × 2 mm³ pieces and observed using an HT7800 transmission electron microscope (Hitachi, Tokyo, Japan).

ELISA

Serum levels of diamine oxidase (DAO), d-lactic acid (D-LA) and LPS were measured using ELISA kits (Multisciences Biotechnology Co., Hangzhou, China) according to the manufacturer's instructions.

Western blotting

Radioimmunoprecipitation assay and phenylmethylsulphonyl fluoride were used to extract proteins from tissues and cells. For detailed experimental methods, please refer to our previous study [12]. ImageJ 7.0 software (NIH, MD, USA) was used to analyze grayscale values and calculate relative protein expression levels.

Flow cytometry

Collected cells were subjected to extracellular staining (CD3, CD4; BioLegend, CA, USA). A FACS Canto II instrument (BD Biosciences, CA, USA) was used for cell identification and data were analyzed using FlowJo software (ThreeStar, OR, USA).

16S rRNA sequencing

Sequencing of fecal 16S rRNA was performed by Metware Biotechnology Co. Ltd. (Wuhan, China). Fresh fecal samples (∼0.2 g) from rats in each group were place in a sealed, sterile, frozen storage tube and stored in liquid nitrogen. According to the characteristics of the V3–V4 region of the 16S rDNA gene of bacteria, a small fragment library was constructed and paired-end sequenced using an Illumina NovaSeq sequencing platform. First, original data were spliced and filtered to obtain valid data (clean data). Operational taxonomic unit (OTU) clustering was then employed and based on the OTU clustering results, representative sequences of each OTU were annotated to obtain corresponding species information and abundance distribution based on species. Furthermore, OTUs were analyzed based on abundance, alpha-diversity, Venn plots and petal plots to obtain information on species richness and evenness within samples, and to identify common and unique OTUs among different samples. In addition, multiple sequence alignment was performed on OTUs and phylogenetic trees were constructed. Through dimensionality reduction analyses such as Shannon, Simpson, chao1, abundance-based coverage estimator (ACE), principal co-ordinates analysis (PCoA), principal component analysis (PCA) and non-metric multi-dimensional scaling (NMDS) sample cluster trees were generated and differences in community structure among different samples or groups were investigated. In order to further explore differences in community structure among grouped samples, statistical analysis methods such as T-test, similarity percentage (Simper) and line discriminant analysis effect size (LEfSe) were used to assess the significance of differences in species composition and community structure of grouped samples. Annotation results for amplified genes were assessed against GreenGenes (16S), FAPROTAX and KEGG functional databases to perform functional prediction analysis of microbial communities in samples.

Statistical analysis

SPSS 13.0 (SPSS GmbH, Munich, Germany) and GraphPad 8.0 (GraphPad Software Company, CA, USA) were used for statistical analysis. Results are expressed as mean ± standard deviation. One-way analysis of variance (ANOVA) was employed to test statistical significance. Results are expressed as a percentage (%), and a Chi-squared test was used to assess statistical significance. Kaplan–Meier survival curves and log bank tests (Mantel–Cox) were used for survival analysis (p < 0.05 was considered statistically significant).

Results

Fer-1 inhibits ferroptosis in transplanted steatotic liver & recipient intestinal tissues, & prolongs recipient survival rate

Morphological changes of severe steatotic grafts were typically yellow with a greasy texture, blunt edges, and tight and smooth capsules (Figure 1A), indicating that the donor liver model with severe steatosis was successfully established. Using normal SD rats as recipients, we successfully performed liver transplantation on transplanted livers with severe steatosis (Figure 1B). Survival curves show that the median survival (30 days) of the Fer-1 group was significantly higher than that of the LT group (16 days; p < 0.05). Therefore, inhibiting ferroptosis could prolong the survival of liver transplant recipients with severe steatotic liver grafts (Figure 1C).

Immunohistochemical results of liver tissue analyses showed that at POD 1, the level of GPX4 was significantly decreased in the LT group. Application of ferroptosis inhibitor Fer-1 increased GPX4 levels (Figure 1D). Electron microscopy results showed that mitochondria of steatotic liver grafts were swollen with prominent mitochondrial cristae, and no rupture of the outer membrane was observed. In the LT group, the intracellular structure was disordered, the mitochondrial volume decreased and cristae were significantly fewer. At POD 7 there was significant mitochondrial injury (indicated by a red arrow). However, in the Fer-1 group more normal mitochondria remained visible (Figure 1D). This indicates significant ferroptosis in liver tissue after liver transplantation from severe steatotic liver grafts, which was more severe at POD 7. Application of ferroptosis inhibitor Fer-1 could significantly reduce ferroptosis in liver cells after transplantation.

Immunohistochemical results for intestinal tissues showed that at POD 1, GPX4 levels in the LT group were significantly decreased, while in the Fer-1 group they were increased (Figure 1E). Electron microscopy observations showed that in the sham group, microvilli were neatly arranged. In the LT group, intestinal microvilli became shorter and atrophied, with obvious mitochondrial swelling and rupture of the adventitia (indicated by a red arrow). In the Fer-1 group, microvilli of the intestinal wall were arranged more neatly, with more normal mitochondria visible (Figure 1E). At POD 7, significant mitochondrial outer membrane rupture remained evident in the LT group. In the Fer-1 group, GPX4 levels were increased, with more normal mitochondria visible. This indicates ferroptosis in both liver and intestinal epithelial cells after severe steatotic liver transplantation, and there were no significant differences in intestinal epithelial cells between POD 1 and POD 7. After application of Fer-1, ferroptosis in liver and intestinal epithelial cells was decreased.

Fer-1 can improve structural & functional changes in transplanted livers & the recipient intestinal tract

Liver histology

In the sham group mixed steatosis was observed. At POD 1, steatosis remained evident in the LT group, accompanied by disorder of the hepatic cord, severe congestion of hepatic sinuses, and obvious cell necrosis and infiltration of inflammatory cells. In the Fer-1 group there was less infiltration of inflammatory cells than in the LT group. At POD 7 the LT group showed significant liver tissue necrosis; in the Fer-1 group mild steatosis of hepatocytes was observed, and some hepatic cords (near the portal vein) were arranged neatly (Figure 2A).

Liver function

At POD 7 the LT group had the highest ALT, AST, TBiL, ALP and TC values (Figure 2B). In the LT group, VLDL was significantly increased at POD 1 but decreased at POD 7 to a level similar to that of the Fer-1 group. There were no differences between groups for LDL and HDL at POD 1. However, at POD 7 LDL was significantly increased in LT, followed by Fer-1 (p < 0.05). These results indicate that inhibiting ferroptosis can improve donor liver histology, liver function and lipid metabolism in steatosis.

Intestinal histology

Intestinal villi of rats in the sham group were arranged regularly. In the LT group, at POD 1 intestinal villi were significantly damaged, with fracture and epithelial loss visible; at POD 7 significant intestinal villi fracture remained evident. In the Fer-1 group, severe villi epithelial loss was observed at POD 1; at POD 7 intestinal villi were regularly arranged with slight villi epithelial loss (Figure 2C).

Intestinal function

At POD 1 LT had the highest plasma levels of D-LA, DAO and LPS, followed by Fer-1 (p < 0.05). At POD 7 only D-LA in the LT group continued to increase. In the Fer-1 group, D-LA, DAO and LPS levels were lower than at POD 1 (p < 0.05; Figure 2D). These results suggest that inhibiting ferroptosis could improve recipient intestinal histology and permeability.

Effects of Fer-1 on injured hepatocytes & intestinal epithelial cells in vitro

In the IAR20 cell injury model, the same concentration of LPS stimulated normal and steatosis IAR20 cells, and cell activity in the FL group was lowest (p < 0.05) (Figure 3A). Intracellular lipid ROS content in the FL group was highest, followed by the FF group (Figure 3B). COX2 protein content was highest, while GPX4 protein content was lowest in FL and NL groups (Figure 3C). These results suggest that LPS can aggravate ferroptosis in steatosis IAR20 cells, while Fer-1 can alleviate ferroptosis in IAR20 cells induced by LPS to some extent.

In the intestinal epithelial cell injury model, cell activity in the NL group was lowest (Figure 3D). Intracellular lipid ROS content in the NL group was highest (Figure 3E–F). In addition, COX2 protein levels in the NL group were significantly increased, while GPX4 protein levels were decreased (p < 0.05; Supplementary Figure 1). These results suggest that LPS can cause ferroptosis in intestinal epithelial cells, while Fer-1 can alleviate ferroptosis in intestinal epithelial cells caused by LPS to a certain extent.

Fer-1 reduces inflammatory responses of fatty liver transplantation & recipients

Neutrophil (marker = MPO) and Kupffer cells (marker = CD68) play a key role in the occurrence and development of nonalcohol fatty liver disease and liver IR [19]. Changes in T-cell subsets are closely related to transplant rejection and the survival of transplant recipients. Therefore, we explored changes in liver MPO and CD68 levels, blood and liver T-cell subsets in each group.

Immunohistochemical results for liver tissues revealed a small number of MPO- and CD68-positive cells in the sham group, indicating inflammatory damage in the steatotic liver graft itself (Figure 4A & B). Levels of MPO and CD68 in the LT group were highest, especially at POD 7 (Figure 4A & B).

Flow cytometry of blood T-cell subsets showed that the proportion of CD8⁺ T cells in the LT group was increased significantly, especially at POD 7. Fer-1 could reduce the proportion of CD8⁺ T cells and increase the proportion of CD4⁺ T cells, and its effect at POD 7 was significantly higher than at POD 1 (Figure 4C). In liver grafts, the proportion of CD8⁺ T cells in the LT group was increased significantly. Compared with the LT group, the proportion of CD8⁺ T cells in the Fer-1 group was decreased, while the proportion of CD4⁺ T cells was significantly increased (Figure 4D).

This suggests that Fer-1 could reduce the infiltration of neutrophils and Kupffer cells in severe steatotic liver grafts. In addition, it could improve inflammatory responses of recipients after liver transplantation by regulating the ratio of CD4⁺ T cells to CD8⁺ T cells. As time progressed the effects of Fer-1 became more obvious.

Fer-1 alters the gut microbiota composition of transplant recipients

Gut microbiota sequencing depth

In vitro experiments showed that Fer-1 could only improve damaged hepatocytes and intestinal epithelial cells to a certain extent, but could not completely restore the activity of damaged cells, reducing the content of lipid ROS and COX2 proteins in cells. In addition, changes in the liver and intestinal functions of recipients can have a significant impact on the composition and function of the gut microbiota. Therefore, we collected and sequenced fecal samples. Species accumulation boxplots, rarefaction and rank abundance curves and goods coverage index reflected observed species, amount of sequencing data and sequencing depth were enough to cover all species in the sample (Figure 5A–D).

Dominant bacteria in the intestines of transplant recipient rats in each group

At the family level Peptostreptococcaceae, Bacteroidaceae, Clostridiaceae, Muribaculaceae, Enterobacteriaceae and Lactobacillaceae were the most abundant microbiota (Figure 5E). At the phylum level Firmicutes, Bacteroidota, Proteobacteria, Verrucomicrobiota, Fusobacteriota and unidentified bacteria had the highest abundance in each group. At POD 7 in the LT group the Firmicutes/Bacteroidota ratio was significantly decreased (Figure 5F).

Alpha-diversity

Alpha-diversity clustering analysis showed that steatotic liver transplantation significantly disrupted the number and diversity of gut microbiota species. ACE, Shannon, Simpson and Chao1 indices all showed that species diversity within the LT group was greatest (Supplementary Figure 2A–D), and the number of OTUs in the LT group was largest, especially at POD 7 (Supplementary Figure 2E).

Beta-diversity

To explore microbial diversity between groups, unweighted UniFrac and weighted UniFrac algorithms were used for beta-diversity analysis (Supplementary Figure 3A & B). In non-NMDS, PCA and PCoA analyses there were significant differences in spatial indices between any two groups, and the LT group had the largest difference from the sham group at POD 7, while the Fer-1 group had the smallest difference from the sham group at POD 7 (Supplementary Figure 3C–F).

Diversity of bacterial populations in each group

Simper (similarity percentage) was used to decompose the Bray-Curtis difference index to quantify the contribution of each species to differences between two groups. The results revealed the top ten species and their abundance at the family level that contribute to differences between two groups.

In the LT group at POD 1, compared with the sham group, the abundances of Enterobacteriaceae and Staphylococcaceae were higher; the abundances of Akkermansiaceae, Muribaculaceae, Lactobacillaceae and Lachnospiraceae were decreased significantly (p < 0.05; Supplementary Figures 4 & 5A). At POD 7 the abundances of Bacteroidaceae and Fusobacteriaceae were increased significantly while the abundances of Peptostreptococcaceae, Enterobacteriaceae, Staphylococcaceae and Lactobacillaceae were decreased significantly (Supplementary Figure 4 & 5B).

In the Fer-1 group at POD 1, compared with the sham group, the abundances of Clostridiaceae and Oscillospiraceae increased were increased while the abundances of Peptostreptococcaceae and Akkermansiaceae were decreased (Supplementary Figures 4 & 5C). At POD 7 the abundance of Peptostreptococcaceae was increased while the abundance of Clostridiaceae was decreased (Supplementary Figures 4 & 5D). Compared with the LT group, the abundances of Clostridiaceae and Muribaculaceae were increased in the Fer-1 group at POD 1, while the abundances of Peptostreptococcaceae, Enterobacteriaceae, Bacteroidaceae and Staphylococcaceae were lower in the Fer-1 group, and differences were significant (p < 0.05; Supplementary Figures 4 & 5E). At POD 7 the abundances of Peptostreptococcaceae, Muribaculaceae and Prevotellaceae were increased while the abundances of Bacteroidaceae, Fusobacteriaceae and Enterobacteriaceae were decreased (Supplementary Figures 4 & 5F).

Fer-1 causes functional changes in the gut microbiota of transplant recipients

Bugbase functional prediction analysis

See Supplementary Figure 6A–I. Based on the GreenGenes (16S) database, microbial communities are classified according to seven phenotypes: Gram-positive, Gram-negative, biofilm-forming, pathogenic potential, mobile element-containing, oxygen-utilizing (aerobic, anaerobic and facultatively anaerobic) and oxidative stress-tolerant. Gram-positive were most abundant in the sham group. At POD 7 the number of Gram-positive bacteria in the LT group was lowest, while the number of Gram-negative bacteria was significantly increased. In addition, in the LT group at POD 1, the abundance of bacteria with pathogenic potential was highest, and it was decreased at POD 7. Facultative anaerobic bacteria were most abundant in the LT group at POD1 and lowest in the Fer-1 group at POD 7. This suggests that the number of harmful stimuli received by the transplanted liver from the intestinal tract was increased, and Fer-1 could decrease pathogenic bacteria and obesity-causing bacteria.

FAPRITAX function prediction

See Figure 6A & Supplementary Figure 7. In the LT group the chemoheterotrophy function was significantly reduced, especially for aerobic chemoheterotrophy. In addition, human pathogens including those causing diarrhea were increased significantly in the LT group but decreased significantly in the Fer-1 group at POD 7. Dark iron oxidation iron respiration of gut microbiota functions were highest in the sham group and the Fer-1 group at POD 1, while at POD 7 they were significantly increased in the LT group (Figure 6B & Supplementary Figure 7). In addition, gut microbiota function in the LT group at POD 1 and the sham group displayed the greatest difference (Figure 6C). This indicates that liver transplantation with fatty liver could significantly affect the functions of the gut microbiota.

PICRUSt functional prediction

Gut microbiota Brite Hierarchies and Cellular Processes were highest in the sham group, while human diseases and environmental information-processing pathways were significantly enhanced in the LT group at POD 1 (Figure 6D). As with FAPRITAX functional prediction, functional differences in the gut microbiota were greatest between the LT group at POD 1 and the sham group (Figure 6E).

Gut microbiota may influence the recovery of recipients after liver transplantation

We also conducted T-tests between groups based on quantitative results from PICRUSt functional annotation. In the LT group at POD 1, compared with the sham group, pathways related to bacterial growth such as cell growth were decreased, while bacterial invasion such as Yersinia infection, bacterial invasion of epiphytic cells and LPS biosynthesis proteins were increased (Supplementary Figure 8A). Pathways related to energy metabolism such as carbon hydrate and fat were significantly decreased, and functions related to iron metabolism (biosynthesis of siderophores and nonribosomal peptides) were increased. At POD 7 bacterial invasiveness of the LT group was further enhanced (two-component systems, ABC transporters and the bacterial secretion system; Supplementary Figure 8B).

In the Fer-1 group at POD 1, compared with the sham and LT groups, pathways related to bacterial growth and reproduction ability such as cell growth were enhanced, bacterial invasion ability was decreased, and carbon fixation and fat metabolism ability were significantly enhanced. Compared with the sham group, the ferroptosis function in the Fer-1 group was enhanced (Supplementary Figures 9 & 10A). At POD 7 the metabolism of carbohydrates and fats in the Fer-1 group was significantly enhanced, and metabolic activity of the gut microbiota associated with ferroptosis (terpenoid backbone biosynthesis) was also enhanced (Supplementary Figure 10B). Thus, Fer-1 could decrease the invasiveness of the gut microbiota, enhance the metabolism of energy-related substances and stimulate ferroptosis-related pathways.

Discussion

Steatotic livers are often used in liver transplantations. Such organs are more prone to IRI, which affects liver regeneration, leading to poor prognosis after liver transplantation [19,20]. Therefore, the use of severe steatotic liver grafts is currently avoided in clinical practice [3]. Therefore, it is very important to develop methods for improving severe steatotic liver graft function after transplantation. The present work investigated the role of ferroptosis inhibitor Fer-1 in liver transplantation and postoperative intestinal injury in recipients, and revealed the role of inflammation and changes in gut microbiota under the influence of Fer-1.

Apoptosis, necrosis, ferroptosis and other kinds of cell deaths exist in steatotic liver transplantation [22]. These cell deaths have considerable overlap and crosstalk. However, since apoptosis requires energy, the chronic ATP depletion in steatosis liver my lead to failure to induce apoptosis [22]. Necrosis in the liver has been long considered an unregulated process [23]. Indeed, iron overload in steatosis livers suggests an important role of ferroptosis [19,20].

We observed ferroptosis in both transplanted liver and intestine after liver transplantation with severe steatotic liver grafts. Following liver transplantation, ferroptosis of transplanted livers was most obvious at POD 7, with more mitochondria swelling, cristae disappearance and outer membrane rupture than at POD 1. There were no significant differences in the degree of ferroptosis between POD 1 and POD 7 in intestinal epithelial cells of recipients. However, Fer-1 could improve ferroptosis in the transplanted liver and intestinal epithelium of recipients, and the effect became more evident over time.

In addition to directly inhibiting ferroptosis, we found that Fer-1 had a repair effect on the structure and function of the transplanted liver and the recipient intestinal tract. After liver transplantation, significant hepatocyte steatosis, congestion and inflammatory cell infiltration could still be seen at POD 7. However, Fer-1 brought about transplanted liver recovery at POD 7, and lipid droplet and inflammatory cell infiltration were significantly decreased. Liver function tests showed that Fer-1 could significantly improve liver function and lipid metabolism in rats following steatotic liver transplantation. Due to clamping of the portal vein and inferior vena cava during liver transplantation with severe steatotic grafts, mesenteric arteries become congested, resulting in structural changes in the intestinal wall and pathophysiological changes including vascular congestion, edema and decreased or relaxed tight connections [13,24]. We found that plasma D-LA, DAO and LPS levels were significantly increased after liver transplantation, indicating an increase in intestinal permeability. However, under the action of Fer-1, levels were decreased significantly. This indicates that Fer-1 could improve the function of steatotic liver transplantation while also reducing intestinal mucosal damage and intestinal permeability.

However, in vitro experiments showed that Fer-1 did not significantly improve the activity of rat hepatocyte IAR20 and intestinal epithelial cells (IEC-6 and Caco-2), intracellular lipid ROS content, or ferroptosis-related protein levels induced by LPS. This is significantly different from the apparent role of Fer-1 in our previous in vivo experiments, hence we continued to explore whether other factors were involved. To this end we investigated the inflammatory responses of transplant recipients.

The proportion of CD8⁺ T and CD4⁺ T cells in transplant recipients is closely related to the clinical outcome of organ transplantation [25]. Neutrophils and Kupffer cells play important roles in the development of IRI during steatotic liver transplantation [26]. Therefore, we tested inflammatory cells associated with IRI and T-cell subsets associated with recipient prognosis. We found that after liver transplantation, there was a significant inflammatory response in the recipient. The number of neutrophils, Kupffer cells and CD8⁺ T cells was significantly increased, while Fer-1 significantly reduced the number of proinflammatory cells and the CD8⁺ T-cell ratio, especially at POD 7, thereby reducing IRI injury and inflammatory responses in steatotic liver transplantation. This indicates that Fer-1 could also inhibit inflammatory reactions and improve the prognosis of recipients.

Considering changes in portal vein pressure, severe damage to recipient liver and intestinal mucosa, and changes in immune function, we suspected that interactions between the liver and the gut microbiota had also altered. Based on the study that gut microbiota metabolites enhance GPX4 expression and suppress ferroptosis in intestinal I/R injury [27] and the therapeutic role of gut microbiota in NAFLD [28], we hypothesize that the gut microbiota may also play a role in the occurrence of ferroptosis in steatotic liver grafts.

Therefore, we explored changes in the gut microbiota after transplantation and detected significant changes in Firmicutes/Bacteroides, suggesting changes in the composition of the gut microbiota caused by liver damage and Fer-1. As the time after transplantation increased, the increase in Peptostreptococcaceae and Enterobacteriaceae became more significant. Members of the Peptostreptococcaceae family can secrete over 20 known toxins and enzymes that may affect homeostasis of the gut microbiota and induce pathogenesis of the host gut, leading to the development of colitis and tumors [29]. In addition, pathogenic bacteria Enterococcus faecalis, Streptococcus bovis, Bacteroides fragilis, Escherichia coli and Fusobacterium spp. may be involved in liver and intestinal mucosal damage after transplantation. LPS, a typical product of Gram-negative bacteria, is a marker of systemic inflammation and prognosis that is negatively correlated with survival [30,31]. After liver transplantation the number of Gram-negative bacteria significantly increased, while under the effect of Fer-1 the abundance of Gram-negative bacteria decreased. Erysipelotrichaceae and facultative anaerobic bacteria are considered ‘bad’ bacteria that cause obesity and promote diabetes [32], and they were more abundant after liver transplantation with steatosis. Thus, after steatotic liver transplantation there was a significant increase in pathogenic bacteria and obesity-causing bacteria, with increased bacterial invasiveness and an increase in harmful stimuli received by the transplanted liver from the gut.

In the Fer-1 group, some bacterial populations involved in receptor energy metabolism, especially fat metabolism, were increased significantly. Oscillospiraceae members are related to regulating body lipid metabolism and promoting lipid lowering and weight loss [33]. Oscillospiraceae members may use host polysaccharides as a growth substrate to reduce chronic inflammation caused by obesity, hence a decrease in the abundance of Oscillospiraceae may lead to mild inflammation and hyperglycemia [33]. Muribaculaceae members have beneficial effects on intestinal ecological disorders through immune regulation and controlling intestinal homeostasis [34,35]. Prevotellaceae is an important bacterial group associated with degradation of polysaccharides and the formation of short-chain fatty acids [36,37]. These bacterial groups related to fat metabolism were increased to varying degrees in the Fer-1 group. Thus, changes in the gut microbiota caused by Fer-1 were mainly related to recipient lipid metabolism.

Lactobacillus, a major class of bacteria in the intestinal tract of humans and animals, is widely used as a probiotic. Lactobacilli have immunomodulatory effects and promote intestinal host defences through interactions with the immune system [37]. After application of Fer-1, the abundance of Lachnospiraceae and Clostridiaceae increased significantly. These are important butyrate-producing bacteria [38] that can significantly improve carbohydrate and lipid metabolism in recipients. Butyrate has been shown to protect the intestinal epithelial barrier, regulate energy metabolism and inflammation, and influence tumor growth and development [39]. In addition, butyrate can also improve obesity, inhibit insulin resistance and reduce cholesterol synthesis [39–41]. This indicates that Fer-1 increases butyrate-producing bacteria and indirectly participates in recipient lipid metabolism.

Lactobacillus members also detect intestinal iron levels and reduce host iron absorption [42]. Lactobacillus was increased significantly in the LT group at POD 1, but decreased significantly in the LT group at POD 7. However, after application of Fer-1, it increased significantly over time. Pasteurellaceae showed the strongest negative correlation with serum ferritin levels, consistent with their dependence on the use of iron in host transferrin for growth and survival [42]. Pasteurellaceae was significantly increased in the Fer-1 group at POD 1. This indicates that Fer-1 promoted the utilization of iron in gut microbiota and reduced iron absorption of recipients. This also explains why Fer-1 enhanced the ferroptosis-related pathways of gut microbiota.

Our previous study demonstrated that that bone marrow mesenchymal stem cells exhibited significant protective effects on steatotic liver grafts [19,20], as well as the ability to modulate the gut microbiota of recipients [13]. It was evident from our findings that the effectiveness of Fer-1 in promoting recovery of steatotic liver grafts was comparatively weaker than that of bone marrow mesenchymal stem cells. Furthermore, the functional changes observed in the gut microbiota with the presence of Fer-1 primarily involve potential pathogenic bacteria and iron absorption. This indicates that the recovery benefits of Fer-1 for steatotic liver grafts are limited in their scope. Moreover, our study focused solely on the changes in gut microbiota and did not investigate alterations in the metabolites of the gut microbiota or conduct experiments to abolish the gut microbiota and validate our results. In future studies, we will focus on the molecular effects of Fer-1 in steatotic liver transplantation and gut microbiota metabolism changes under the effect of Fer-1.

Conclusion

Fer-1 directly inhibits ferroptosis, improves inflammatory responses of recipients after liver transplantation, and promotes recovery of the histological structure and function of transplanted liver and recipient intestinal tracts. Fer-1 reduces the abundance of intestinal pathogenic bacteria and obesity-causing bacteria in recipients and reduces the pathogenicity of the gut microbiota. It also protects transplanted liver with steatosis by improving lipid metabolism and reducing iron absorption of recipients.