Enhancement of liver mitochondrial complex I and energy metabolism induced by enteritis: The key role of gut microbiota derived endotoxins

Inflammation is an energy-intensive process and the liver is a key organ in energy regulation. Since the intestine and liver exchange nutrients and metabolites, enteritis can affect the liver. To investigate the correlation between enteritis and liver metabolism, we developed an intestinal inflammation model with concentration-dependent 2,4,6-trinitrobenzene sulfonic acid (TNBS) in gibel carp (Carassius gibelio). The results showed the dysregulation of intestinal tight junction, increased permeability of the gut barrier, and apoptosis of epithelial cells during the development of enteritis. The liver metabolome was analyzed by LC-MS and the live respiration was determined using Oxygraph-2k. The results showed that glycolysis, the TCA cycle and pyrimidine metabolism were affected by intestinal inflammation. In particular, the activity of hepatic mitochondrial respiratory chain complex I was significantly increased. Structure and abundance changes of gut microbiota were analyzed by 16S rRNA sequencing analysis. Pathogenic bacteria in the intestine, as well as plasma LPS, increased significantly. Using a liver cell line, we verified that the dysfunctional metabolism of the liver is related to the dislocation of LPS. All results imply the existence of a connection between enteritis and liver metabolism in gibel carp, and the gut microbiome plays a critical role in this process.


Introduction
The gastrointestinal tract is continuously exposed to numerous bacteria as well as food-derived and environmental toxins (1), potentially leading to inflammatory bowel diseases (IBD) in humans (2, 3) and colitis in animals (4)(5)(6). IBD presents with abdominal pain, fever, and clinical signs of bowel obstruction or diarrhea with passage of blood or mucus, or both. Enteric diseases in commercial animals contribute to losses in productivity and increased mortality. The dysregulation of intestinal permeability, tight junction, immune function, and apoptosis has shown to be involved in the development of enteritis (7). As an important part of the intestinal immune barrier, the gut microbiota is crucial for maintaining the host's health (8). The development of intestinal inflammation is accompanied by dysbiosis of intestinal microbiota. The leakiness and dysregulated permeability of the gut not only enables bacteria to invade the host, but also results in the infiltration of macrophages and inflammation. In response, the body produces cytokines and other mediators, effectively launching a systemic inflammatory response (9). Thus, enteritis is a systemic disease that not only affects the gastrointestinal tract, but also impacts the extraintestinal organs. The effects of intestinal microbiota and microbial metabolites on other organs during enteritis is therefore worthy of future study.
Dysregulated energy homeostasis has been observed to be implicated in the pathogenesis of enteritis (10). Mitochondrial oxidative phosphorylation is the main source of cellular adenosine 5'-triphosphate (ATP). A study in ulcerative colitis patients found that complex II and III were decreased by around 50-60% in the colonic mucosa (11). Other studies in mice showed that increasing mitochondrial ATP synthesis in intestinal epithelial cells could relieve colitis (10). Therefore, the regulation of energy metabolism is anticipated as a promising strategy for counteracting enteritis. As a main organ of metabolism, the liver is exposed to intestinal-derived substances including nutrients and bacteria. Some studies have found that some IBD patients or animals with enteritis exhibited liver diseases, such as non-alcoholic fatty liver disease (NAFLD) and metabolic disturbances (12, 13). However, the mechanisms which alter mitochondrial activity in the liver during intestinal inflammation are still unknown.
In the present study, we speculated that intestinal inflammation may impact the gut microbiota, thereby affecting the gut-liver axis and causing a change in liver mitochondrial metabolism in response to inflammation. To test this hypothesis, we developed an intestinal inflammation model using gibel carp (Carassius gibelio). The structure and abundance of changes in the gut microbiota were analyzed by 16S rRNA sequencing analysis. Liver metabolome was analyzed by LC-MS, and we quantified mitochondrial activity in the liver by Oxygraph-2k. We also treated a liver cell line with LPS in order to verify that the dysfunction of energy metabolism during intestinal inflammation is related to the gut microbiota.

Animal ethics
All experiments were conducted under the Guidance of the Care and Use of Laboratory Animals in China. This study was approved by the ethics committee of the Institute of Hydrobiology, Chinese Academy of Sciences.
The administration of TNBS (Sigma-Aldrich) was performed by gavage using a lavage tube (OD 1.00 mm). After a 24h period of fasting, the fishes were anesthetized with MS-222 (50 mg/L, Sigma, Saint Louis, Missouri, USA) and a TNBSethanol solution (50% v/v) was administered at a dose of 75 mL or 125 mL/kg body weight (TNBS 75 mL/kg; TNBS 125 mL/kg). The control groups were treated with the same volume of 0.68% stroke-physiological saline solution (SS 75 mL/kg; SS 125 mL/kg).
Fish were anesthetized for sampling after 24h and blood samples were taken from the caudal vein, centrifuged at 3,000 g (5 min, 4°C), and stored at -80°C until further analyses. Tissue samples were collected and stored at -80°C for further analyses including biochemical assays, gene expression assessment, and metabolomic analysis. A small section of each liver and mid-gut was fixed in 4% paraformaldehyde solution for histological analysis. Another portion of the mid-gut was stored in 2.5% glutaraldehyde solution for ultrastructural observation. In addition, another 7 fish in the TNBS 125 mL/kg and SS 125 mL/kg groups were sampled for testing intestinal epithelial resistance by the Ussing Chamber. The liver was collected for analysis of respiration by Oxygraph-2k immediately after sampling.

Histological analyses
Gross pictures of the whole intestine from all groups of fish were taken immediately following dissection (DC-GF10K, Panasonic). Hematoxylin-eosin (H&E) staining of the intestine and liver samples was performed by Powerful Biology Co. (Wuhan, China), and photographed using light microscopy (Leica DM2500, Leica, Solms, Germany). Transmission electron microscopy (TEM) observation of the intestine was conducted as described in Liu et al. (14).
A terminal deoxynucleotidyl transferase-mediated dUTPbiotin nick end labeling (TUNEL) technique was applied to verify apoptosis in the liver. The TUNEL and DAPI (4', 6diamidino-2-phenylindole) double staining of the liver was performed by Powerful Biology Co. (Wuhan, China) and slides were observed under a Tecnai transmission electron microscope.

Ussing chamber experiments
The mucosal biopsies were immediately immersed in icecold Krebs-Ringer solution (117 mmol/L NaCl, 25 mmol/L NaHCO 3 , 1.2 mmol/L NaH 2 PO 4 , 2.5 mmol/L CaCl 2 , 1.2 mmol/L MgCl 2 , 4.7 mmol/L KCl 2 and 11 mmol/L glucose), at pH 7.4 and gassed with 95% O 2 and 5% CO 2 at 25°C. The biopsied intestine was mounted in Ussing chambers (Beijing Kingtech Technology Co.) with an insert with a square area of 0.04 cm 2 , and each chamber was filled with 5 mL Krebs-Ringer solution which was maintained at 25°C and continuously oxygenated with 95% O 2 and 5% CO 2 , and stirred by gas flow in the chambers as previously described (15). Through a short voltage pulse (5V), the epithelial resistance (Rep) is calculated by applied voltage with the current deflection.

qPCR analysis
Total RNA was extracted and the quality was checked using 1.0% agarose electrophoresis. Subsequently, the mRNA quantitative reverse transcription-polymerase chain reaction (RT-PCR) was performed as previously described (16). The specific primers used for RT-PCR are referenced in previous studies (16) and the additional primers are listed in Table 1. The expression levels in the intestine were calculated as fold-change in expression relative to the housekeeping gene (b-actin and ef-1a).

Liver respiration
The liver was excised from sacrificed fishes and precooled respiration medium was added to the tissues (0.05mM EGTA, 110 mM sucrose, 1g/L BSA, 20 mM HEPES, 10 mM KH 2 PO 4 , 20 Mm taurine, 60 mM K-lactobionic acid, and 3 mM MgCl 2 , pH7.1). A two-channel titration injection respirometer (Oxygraph-2k, Oroboros Instruments) was used to analyze liver respiratory function after homogenization (temperature was 28°C). The oxygen flux per tissue mass (pmol/(s*mg)) and oxygen concentration (mM) were recorded in real-time using Oroboros Datlab software (Oroboros Instruments). Routine respiration was measured when respiration was stabilized without any additives. After titration with malate (M, 2 mM) and pyruvate (P, 5 mM), the respiratory leak of the complex I was tested. The oxidative phosphorylation capacity of complex I (CI OXPHOS) was examined after adding 5 mM ADP. Succinate (10 mM) was titrated to measure OXPHOS of CI and complex II (CI&II OXPHOS). Then, stepwise addition of carbonylcyanidem-chlorophenylhydrazone (CCCP, 2 mM) was carried out in order to obtain the maximal uncoupled respiratory capacity of the electron transfer system (CI&II ETS). Subsequently, the CII ETS was determined after adding rotenone (Rot, 0.5 mM). Finally, antimycin A (Ama, 2.5 mM) was added to evaluate the residual oxygen consumption (ROX). All substrates were dissolved in H 2 O and the inhibitors were dissolved in ethanol.

Metabolome analysis of liver using LC-MS
The 50-milligram liver sample was placed into the 2-mL EP tube and 500 mL of pre-cooled extractant was added (70% methanol aqueous solution). After homogenizing, the sample was centrifuged at 1500 rpm for 5 min and then allowed to stand for 15 min on ice before being centrifuged again for 10 min at 12000 rpm. Subsequently, 200 mL of the supernatant was collected for analysis using an LC-ESI-MS/MS system (UPLC, Shim-peck UFLC SHIMADZU CBM30A system; MS, QTRAP ® System) which was equipped with electrospray ionization (ESI) Turbo Ion-Spray interface, operating in positive and negative ion mode and controlled by Analyst 1.6.3 software (Sciex). A specific set of multiple reaction monitoring (MRM) transitions was monitored for each period according to the metabolites eluted within this period.
All data were processed and normalized using MetaboAnalyst 5.0 (https://www.metaboanalyst.ca/). The samples were grouped using principal component analysis (PCA) and significantly upregulated and down-regulated metabolites between groups were determined by variable importance in the projection (VIP) ≥ 1 and absolute Log 2 FC (fold change) ≥ 1.

Metabolome analysis of cells using GC-MS
The 1×10 7 cells were homogenized with 0.6 mL of cold 80% methanol and ground with a tissue-grinding machine. The samples were centrifuged at 13000 rpm for 20 min at 4°C after standing for 30 min at -20°C. Ribitol (15 mL) was added to the supernatant as an internal standard before drying with

Statistical analyses
Statistical comparisons were made among treatments using the statistical software SPSS 20.0 (SPSS Inc., Chicago, IL, USA). Student's t-tests were conducted (unpaired, two-tailed). All values were expressed as the mean ± standard error (S.E.), and P<0.05 was considered significant.

Establishment of intestinal inflammation in TNBS-treated gibel carp
Macroscopically, visible damage in TNBS-treated fishes was observed compared to the control group. As demonstrated by abnormal gross morphology and H&E staining ( Figures 1A, B), TNBS-treatment induced dose-dependence gut congestion and erosion, and intestinal epithelial villi detachment and destruction.
TNBS significantly decreased intestinal barrier function. The expression of intestinal tight junction mRNA including jam, zo1, occludin, and claudin4 is suggested to be representative of paracellular barrier integrity. Epithelial resistance (Rep) is determined by the permeability of tissue. As shown in Figure 1C, we found that the expression of occludin and zo1 were significantly decreased in the 75 mL/kg TNBS treated group compared to the control, while occludin was significantly decreased in the 125 mL/kg TNBS treated group compared to the 125mL/kg SS group. Similar to the results of gene expression, the ultrastructural changes of the intestine ( Figure 1D that tight junctions were significantly destructed in the 125 mL/ kg TNBS group. The Rep also decreased significantly at this concentration of TNBS ( Figure 1E), further confirming TNBSinduced increased permeability. Additionally, the expression of claudin4 significantly increased in TNBS treated groups. Proinflammatory cytokines and apoptosis were also upregulated. Compared to the SS groups, there was higher expression of il-1b and tgfb in TNBS-treated groups (Figure 2A). DAPI and TUNEL staining of the intestine indicated a greater degree of apoptosis in the TNBS-treated groups compared to the control groups ( Figure 2C). In addition, 125 mL/kg TNBS significantly induced the transcription levels of bcl2, while capsase3 was significantly downregulated in the 75 mL/kg TNBS group (Figures 2B, D). However, no significant difference in the expression of ero1a was observed ( Figure 2E).

Dysbiosis of intestinal microbiota
Alpha diversity indices of the bacterial community showed that 125 mL/kg TNBS significantly decreased the community diversity and abundance. As shown in Table 2, the Shannon and Simpson diversity, and Chao and Ace richness were significantly affected in the 125 mL/kg TNBS group, but not in the 75 mL/kg TNBS group. Based on the distribution of samples on the plot, principal co-ordinates analysis (PCoA) at the genus level indicated differences between the four groups ( Figure 3A).
Proteobacteria, Firmicutes, Actinobacteriota and Fusobacteriota were dominant phyla in the control groups ( Figure 3B). However, 75 mL/kg TNBS significantly increased the relative abundances of Proteobacteria and Firmicutes phylum, and decreased the Fusobacteriota. In addition, 125 mL/kg TNBS increased Firmicutes while decreasing the Actinobacteriota.
At the genus level, Mycobacterium and Cetobacterium were the abundant genera in the control groups. A TNBS concentration of 75 mL/kg significantly decreased the abundance of Cetobacterium while increasing the abundance of Gemmobacter ( Figure 3C). In addition, 125 mL/kg TNBS significantly increased the abundance of gram-negative bacteria, including Acinetobacter, Plesiomonas, and Aeromonas ( Figure 3C). In particular, the abundance of Lactococcus significantly increased in TNBS-treated groups. We found that TNBS specifically increased the pathogenic bacteria Lactococcus garvieae, Plesiomonas shigelloides, Acinetobacter junii and Aeromonas veronii ( Figure 3D).

Effect of TNBS-induced enteritis on host injury parameters
In order to evaluate whether acute enteritis induced by TNBS causes host injury, various biochemical indexes of plasma were analyzed. As shown in Figure 4A, IL-1b was significantly increased in the 125 mL/kg TNBS group, while the other cytokines (IL-6 and IL-10) were significantly decreased in the TNBS groups compared to the controls. The antioxidant biomarker SOD was significantly decreased in the TNBS groups, while GSH was significantly increased in the 75 mL/kg TNBS group.
Plasma glucose was significantly higher, while insulin was significantly decreased in the 125 mL/kg TNBS group compared to the control group ( Figure 4B). In addition, both doses of TNBS significantly decreased the content of plasma lipids, including NEFA, cholesterol and triglycerides.
The increase in gut permeability led to the paracellular transport of luminal endotoxins into circulation. Compared to the SS groups, the level of LPS was significantly elevated in the 125 mL/kg TNBS group ( Figure 4C). In agreement with the previous intestinal permeability results, the level of DAO in plasma was significantly higher in the 125 mL/kg TNBS group ( Figure 4D).

Metabolomic analysis of the liver
TNBS significantly influenced energy metabolism in gibel carp. In line with the changes in plasma biomarkers, histological analyses with H&E staining clearly revealed liver damage induced by TNBS, characterized by obvious macrovesicular steatosis and irregular arrangement of hepatocytes that was not seen in the control group ( Figure 5A). To further explore the metabolic changes in the liver, the energy metabolism of liver samples was analyzed. The PCA result is shown in Figure 5B. Significant differences between the 125 mL/kg TNBS-treated groups and the SS groups were observed. Based on the top 25 metabolites, we also found that the heatmap plot clearly distinguished 125 mL/ kg TNBS groups from the control groups ( Figure 5C). Figure 5D shows the glycolysis metabolism, purine metabolism, pyrimidine metabolism as well as TCA and PPP pathways in the liver. Compared to the control group, changes in the metabolites of the group treated with 125 mL/kg TNBS were clearly visualized using boxplots. Compared to the control groups, the glyceraldehyde 3-phosphate, dihydroxyacetone phosphate, thiamine pyrophosphate (TPP) and aspartate were significantly decreased while the citrate and dUMP were significantly increased in 125 mL/kg TNBS groups.

Liver respiration was affected in TNBSinduced enteritis of gibel carp
The substrate-uncoupler-inhibitor-titration (SUIT) approach together with high-resolution respirometry is able to assess the function of electron-transferring CI and CII, and ETS capacity (Figures 6A, B). Oxygen fluxes of the CI OXPHOS, as well as CI and CII OXPHOS (CI & CII OXPHOS) states were significantly higher than the control for the TNBS groups. A significant increase was also observed when the uncoupler (CCCP) was applied for the assessment of the maximal ETS (CI and CII ETS) capacity in 125 mL/kg TNBS groups. However, no differences in the oxygen fluxes of CII ETS states were observed after subsequent addition of CI inhibitor rotenone in HTNBS groups. Thus, 125 mL/kg TNBS significantly enhanced the CI OXPHOS, not CI OXPHOS capacity. At the end of the experiments, the resulting ROX in the 125 mL/kg TNBS groups was significantly lower than the control, demonstrating that the oxygen consumption was mainly of mitochondrial origin.

Metabolomic analysis of the LPS treated NCTC1469 cells
The PCA result was shown in Figure 7A and it demonstrated significant differences between the LPS-treated group and the control (Con) group. An enriched chart of the metabolites provided a graphical display of the top 25 most enriched pathways ( Figure 7B). The results showed that the most affected pathways were primarily related to amino acid and carbohydrate metabolism. The heatmap showed that most metabolites were downregulated in LPS groups while the itaconic acid was significantly upregulated ( Figure 7C).

Discussion
In the present study, we aimed to discover metabolic changes in the liver during enteritis. The results indicated that hepatic metabolism, in particular the upregulation of mitochondrial complex I, might be a compensatory or acute phase response against changes in intestinal inflammation. The intestinal microbiota and their endotoxins mediate this process.
Several intestinal inflammation models have been created in fish to clarify the mechanisms of inflammatory reactions and evaluate novel prevention methods, including soybean meal (16), pathogenic bacteria Aeromonas hydrophila (17), and chemical reagent TNBS (18). In the present study, the histological examination revealed that TNBS caused typical intestinal pathology, including villus atrophy, inflammatory cells infiltrating the lamina propria and disruption of the epithelial barrier as reported previously (16,19).
The formation and permeability of tight junctions determine the resistance and integrity of the intestine (20). Dysregulation of A B FIGURE 6 The TNBS enhanced the respiratory of mitochondria in liver.  tight junction proteins contributes to the pathogenesis of intestinal diseases (21). Meanwhile, intestinal inflammation affects the expression and cellular localization of TJ proteins, including the claudin family proteins, JAM and ZO. Some ultrastructural studies have demonstrated a dilation of TJ and increased intercellular space in mucosal biopsies of enteritis patients (22). Gene expression and ultrastructural examination in the current study revealed that tight junctions were disrupted after being treated with TNBS thereby confirming the ability of our model to simulate intestinal barrier disruption. Alterations of intestinal epithelial homeostasis during inflammation may increase permeability of the gut barrier. Intestinal permeability was evaluated on the basis of the transepithelial electrical measurements with the Ussing chamber. The present study showed decreased TER revealing decreased physical intestinal barrier function which is in line with previous studies in Atlantic salmon (23). In addition, DAO is barely present in the circulation and its basal plasma levels are correlated with the integrity of the intestinal mucosa (24). In agreement with the TER results, the level of DAO in plasma was significantly elevated in TNBS groups compared to SS groups. In mammals, the cytokines (eg. IL-1b, IL-10, TNFa and TGFb) mediate tight junction function (25). Thus, the barrier disruption in gibel carp may involve up-regulation of il-1b and tgfb as observed in the current study.
As TNBS induced intestinal damage, high dosages may cause apoptosis of epithelial cells (26). The present study found that TUNEL staining was enhanced and the expression of apoptosisrelated genes (blc2 and caspase3) was significantly affected. Redox and oxidant-mediated regulation of apoptosis (27), and Nrf2 stimulates cytoprotective genes in response to oxidative stress both in fish and mammals and leads to activation of the downstream genes in Nrf2-Keap1 signaling pathway (28). We found that the gene expression of nrf2, bach1 and cat was significantly down-regulated while the expression of hsp70 was elevated after treatment with 125 mL/kg TNBS indicating that TNBS induced oxidative stress.
As a second factor associated with barrier function, intestinal microbiota was one of the factors responsible for enhancement of intestinal permeability (29). Several reports have shown that the changes in gut microbiota composition are associated with intestinal barrier function. In this study, we have clearly shown that the barrier disruption was associated with a significantly reduced gut microbiota diversity. In addition, numerous studies have found that an increase in the abundance of Proteobacteria reflects dysbiosis in gut microbiota (30) and plays a causative role in intestinal inflammation (31). Fusobacteriota metabolizes carbohydrates into butyrate which provides energy for gastrointestinal cells and has antiinflammatory propoerties (31). Thus, in this study, the increased relative abundance of Proteobacteria and decreased relative abundance of Fusobacteriota had a detrimental effect on the maintenance of host intestinal energy supply and normal immuno homeostasis. Meanwhile, the genus level was significantly affected as well after treatment with TNBS. Lactococcus garvieae, Plesiomonas shigelloides, Acinetobacter junii and Aeromonas veronii are all pathogenic bacteria in aquaculture (32), and may cause gastroenteritis and septicemia (33,34). The relative abundance of these bacteria increased significantly in the TNBS groups in a dose-dependent manner. Accordingly, TNBS-induced enteritis in gibel carp can be used as a model of intestinal inflammation.
With the development of enteritis, the intestinal barrier will be broken and bacteria and bacterial endotoxins will enter circulation. In response to enteritis, the intestinal epithelium changes structurally, allowing LPS to enter the bloodstream, resulting in an increase in the plasma levels of LPS (35). LPS activates Toll-like receptor-4 (TLR4) leading to the production of pro-inflammatory cytokines and systemic inflammation (35). Our results showed that intestinal inflammation increases plasma LPS, and leads to liver damage. The same result was found in the SBM-induced intestinal inflammation model (16).
In agreement with many other studies (10,(36)(37)(38), our study indicated that dysfunction of energy metabolism is involved in intestinal inflammation. The increased plasma glucose and decreased NEFA, cholesterol, and triglyceride in the present study suggest glycolipid metabolism may be altered in response to intestinal inflammation. Zebeli et al. found that an oral LPS challenge significantly increased plasma glucose and decreased the plasma concentration of NEFA in cows (39). This suggests that the LPS translocation may contribute to the dysfunction of metabolism.
The energy metabolic analysis provides comprehensive information on central carbon metabolism-related changes. In the present study, we analyzed the influence of high dose TNBS with a more obvious phenotype on metabolism. The results showed that changes in glycolysis, the TCA cycle and pyrimidine metabolism occurred in the liver in response to severe intestinal inflammation after treatment with 125 mL/kg TNBS. Similarly, Kim et al. found that the levels of liver pyrimidine metabolismrelated metabolites (uracil, UMP and UDP) increased significantly in the state of DSS-induced gut inflammation. These metabolites are associated with the synthesis of nucleotides and are important in the anti-inflammatory response and in preventing DNA damage (13). Although altered metabolism can confer several advantages with respect to inflammation, cells may be burdened with toxic byproducts as a consequence of a deranged or simply overactive metabolism (40). The level of glycolysis-related metabolites (DHAP and GAP) decreased significantly, whereas G6P, PGA and lactate increased in TNBS groups, suggesting enhanced glycolysis to lactate production. The elevated levels of citric acid in the TCA cycle revealed that, in addition to glycolysis, lipolysis may also be enhanced in the TNBS groups. In addition, succinate has been identified as an inflammatory signal and glutamine-dependent anaplerosis is the principal source of succinate during inflammation (41). Thus, the significantly higher levels of succinate and glutamate indicated that the liver actively responded to inflammation.
Some intermediates of central carbon metabolism are involved in the OXPHOS. Inflammation is an energy-intensive process and the mitochondria play crucial roles in supporting these energy-requiring responses. The colonic mucosa in enteritis is in a state of energy deficiency characterized by low energy charge and low adenosine triphosphate (ATP) levels. Herein, we found mitochondrial abnormalities in the liver as well. However, different from the gut, the activity of Complex I (NADH: ubiquinone oxidoreductase) was significantly promoted in liver. We observed that enteritis increased the respiratory function in liver at the CI & CII OXPHOS, and CI & CII ETC states. Our results demonstrate that liver increased ATP synthase activity and enhanced mitochondria oxidative phosphorylation during enteritis. The site for ROS production in mitochondria is normally ascribed to the activity of Complex I (42) which accepts electrons from nicotinamide adenine dinucleotide (NAD+) generated in the TCA. Thus, active mitochondria in the liver produce high levels of ATP as a compensatory response to the acute challenge of TNBSinduced inflammation. However, in a high metabolic state, the liver also produces more ROS which would damage the mitochondria and impede the production of high ATP levels. Some researchers have found that mitochondrial respiratory chain complex I inhibitors (rotenone and metformin) can alleviate the LPS-induced fulminant liver injury (43,44). Thus, mitochondrial respiratory chain complex I may potentially be involved in the control of TNBS-induced inflammation.
We further confirmed that LPS induces disturbances in liver metabolism. Most studies focused on the metabolic changes in macrophages after LPS treatment. The in vitro research has demonstrated that upon LPS stimulation, the hepatic macrophages activate to release the inflammatory cytokines (45). The activated macrophages shift from producing ATP by oxidative phosphorylation to aerobic glycolysis. In the present study, we analyzed the metabolome of NCTC1469 cells after LPS treatment and demonstrated that LPS significantly affects the metabolism of hepatocytes, not only the macrophages. Thus, we speculated that intestinal inflammation directly affected liver metabolism.

Conclusion
In conclusion, we developed an intestinal inflammation model in gibel carp with TNBS. We found that gut microbiota-derived endotoxins mediated changes in liver metabolism, especially the activity of Complex I in response to inflammation. Our results further indicate that intestinal inflammation results in dysbiosis of microbiota and leads to the production of a large number of endotoxic LPS which enter the circulation. LPS induced the activation of mitochondria to enhance energy production. Thus, targeting the energy biosynthetic pathways in the liver might be a novel strategy for the therapy of systemic inflammation during enteritis.

Data availability statement
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: NCBI: PRJNA856173, MetaboLights: MTBLS5270.