Combined Inhibition of the TGF-β1/Smad Pathway by Prevotella copri and Lactobacillus murinus to Reduce Inflammation and Fibrosis in Primary Sclerosing Cholangitis

Primary sclerosing cholangitis (PSC) is a chronic cholestatic disease characterized by inflammation and fibrosis of the bile ducts. Cholestasis may lead to hepatic inflammation and fibrosis, and amelioration of cholestasis may allow recovery from inflammatory and fibrotic pathological damage. Prevotella copri (P. copri) interventions have been reported to significantly improve cholestasis and liver fibrosis in 3,5-diethoxycarbonyl-1,4-dihydrocollidine (DDC)-induced PSC mouse models. Even though P. copri treatment alone cannot bring about recovery from DDC-induced inflammation, it increases the abundance of Lactobacillus murinus (L. murinus) compared with DDC treatment, which has been reported to have anti-inflammatory effects. The abundance of L. murinus still not recovering to a normal level may underlie hepatic inflammation in P. copri + DDC mice. Separate or combined interventions of P. copri and L. murinus were used to investigate the molecular mechanism underlying the improvement in PSC inflammation and fibrosis. P. copri and L. murinus significantly reduced the hepatic inflammatory cell aggregation and inflammatory factor expression as well as the hepatic collagen content and fibrin factor expression in the PSC mice. Further analysis of phosphorylation and dephosphorylation levels revealed that treating the PSC mice with the P. copri and L. murinus combined intervention inhibited the activity of the DDC-activated TGF-β1/Smad pathway, thereby reducing liver inflammation and fibrosis. The combination of P. copri and L. murinus inhibits the TGF-β1/Smad pathway and reduces inflammation and fibrosis in PSC.


Introduction
Primary sclerosing cholangitis (PSC) is a chronic cholestatic liver disease of unclear etiology [1][2][3]. With long-term cholestasis and inflammatory and fibrotic lesions, most patients eventually develop cirrhosis and liver failure with poor prognosis [4,5]. There are no effective drugs available, which stimulates the need for research to elucidate interventions and treatments for PSC. Studies have confirmed that the intestinal microbiota of the PSC population is disturbed and P. copri is significantly deficient in PSC patients [6][7][8]. In our previous study, we constructed a PSC mouse model exhibiting bile duct stricture, liver inflammation, and fibrosis using a 0.1% (w/w) 3,5-diethoxycarbonyl-1,4-dihydrocollidine (DDC)-supplemented diet [9]. Subsequently, P. copri was administered to the PSC model animals to further explore the effect of the microbial intervention. Our previous study found that the symptoms of cholestasis of the PSC mice significantly improved after the P. copri treatment, but P. copri did not reduce liver inflammation [9]. 16S rRNA sequencing analysis determined that P. copri could regulate the microenvironment of intestinal microbiota by changing the abundance of specific genera, among which the abundance of Lactobacillus Our previous study confirmed that the Lactobacillus genus was decreased in DDCinduced PSC mice and monocolonized Gram-negative P. copri could recover its abundance, and further analysis found that Lactobacillus murinus (L. murinus) was the most abundant Gram-positive bacterium in the DDC group [9] (Figure S1A,B). The qPCR analysis revealed a noticeable decrease in the abundance of P. copri and L. murinus within the DDC group when compared with the control group, confirming our previous research findings ( Figure 1A). After that, an experimental schematic diagram was designed, as shown in Figure 1B; in the experiment, the control and PSC model mice were fed a normal diet or the 0.1% (w/w) DDC diet for one week, respectively. Subsequently, the mice were orally administered bacterial suspensions of P. copri and L. murinus, either alone or in combination for an additional week. The serum levels of ALT, ALP, and TBIL were significantly elevated in the DDC group compared with the control mice, indicating that DDC induced liver damage, bile duct obstruction, and cholestasis ( Figure 1C). Compared with the DDC group, administering P. copri or L. murinus alone significantly reduced the levels of ALT, ALP, and TBIL in the DDC-induced groups ( Figure 1C), while administering P. copri and L. murinus together significantly reduced the level of ALT in the DDC-induced groups ( Figure 1C). There was no significant difference between the groups receiving P. copri or L. murinus treatment alone and the combined treatment group. Notably, the intervention of P. copri and L. murinus in the PSC mice resulted in a reduction in hepatic TBAs to a level comparable to that of the control mice ( Figure 1D). Similarly, the P. copri or P. copri + L. murinus treatment blunted the reducing effect on intestinal TBAs in the DDC-induced mice, except for the L. murinus treatment ( Figure 1D). Furthermore, the high expression of the bile duct epithelial marker cytokeratin 19 (CK19) after being fed the DDC diet was attenuated by the P. copri and/or L. murinus treatment, which further confirmed that P. copri and/or L. murinus intervention could alleviate the obstruction of the bile duct induced by the DDC diet in liver ( Figure 1E). induced mice, except for the L. murinus treatment ( Figure 1D). Furthermore, the high expression of the bile duct epithelial marker cytokeratin 19 (CK19) after being fed the DDC diet was a enuated by the P. copri and/or L. murinus treatment, which further confirmed that P. copri and/or L. murinus intervention could alleviate the obstruction of the bile duct induced by the DDC diet in liver ( Figure 1E). Immunohistochemistry staining of CK19 in liver tissues, and the box shows the obstruction of bile components in the bile duct. (Scale bar = 100 µm; original magnification 20×; * indicates statistical difference compared to the Con + PBS (Con is short for control) group, # indicates statistical difference compared to the DDC + PBS group, and & indicates statistical difference compared to the DDC+ P. copri + L. murinus group; * p < 0.05, ** p < 0.01, *** p < 0.001; # p < 0.05, ## p < 0.01; and & p < 0.05).

Supplementation with P. copri Accelerates the Improvement That L. murinus Has on Inflammation in PSC Mice
The H&E staining showed that the DDC treatment induced the aggregation of a large number of inflammatory cells in the liver to activate inflammatory response, while the addition of P. copri and L. murinus intervention resulted in a noticeable reduction in both the extent and number of inflammatory cell infiltrates (Figure 2A). IHC staining, along with quantitative analysis, was employed to assess the expression of the macrophage marker F4/80 in the liver tissues of the mice ( Figure 2B,C). According to the quantitative results ( Figure 2C), compared with the intervention with P. copri or L. murinus alone, the combined intervention can significantly reduce the expression of F4/80, indicating that the combined intervention of the two strains has a be er effect on reducing liver inflammation (E) Immunohistochemistry staining of CK19 in liver tissues, and the box shows the obstruction of bile components in the bile duct. (Scale bar = 100 µm; original magnification 20×; * indicates statistical difference compared to the Con + PBS (Con is short for control) group, # indicates statistical difference compared to the DDC + PBS group, and & indicates statistical difference compared to the DDC+ P. copri + L. murinus group; * p < 0.05, ** p < 0.01, *** p < 0.001; # p < 0.05, ## p < 0.01; and & p < 0.05).

Supplementation with P. copri Accelerates the Improvement That L. murinus Has on Inflammation in PSC Mice
The H&E staining showed that the DDC treatment induced the aggregation of a large number of inflammatory cells in the liver to activate inflammatory response, while the addition of P. copri and L. murinus intervention resulted in a noticeable reduction in both the extent and number of inflammatory cell infiltrates (Figure 2A). IHC staining, along with quantitative analysis, was employed to assess the expression of the macrophage marker F4/80 in the liver tissues of the mice ( Figure 2B,C). According to the quantitative results ( Figure 2C), compared with the intervention with P. copri or L. murinus alone, the combined intervention can significantly reduce the expression of F4/80, indicating that the combined intervention of the two strains has a better effect on reducing liver inflammation caused by the DDC diet than the single bacterial intervention. The above finding was corroborated by reevaluating the mRNA levels of the macrophage marker Mpeg1 ( Figure 2D). Furthermore, the suppression of monocyte chemoattractant protein-1 (Mcp-1) and vascular cell adhesion factor (Vcam-1) is responsible for the effect of the combined intervention of L. murinus and P. copri, which indicates that the combined intervention with these two bacteria might reduce the inflammation induced by the DDC diet through reducing the migration and infiltration of macrophages ( Figure 2D). However, the intervention with P. copri or L. murinus alone did not decrease the expression of Mcp-1 and Vcam-1, further confirming that the P. copri treatment facilitates the improvement that L. murinus has on inflammation in PSC mice. caused by the DDC diet than the single bacterial intervention. The above finding was corroborated by reevaluating the mRNA levels of the macrophage marker Mpeg1 ( Figure 2D). Furthermore, the suppression of monocyte chemoa ractant protein-1 (Mcp-1) and vascular cell adhesion factor (Vcam-1) is responsible for the effect of the combined intervention of L. murinus and P. copri, which indicates that the combined intervention with these two bacteria might reduce the inflammation induced by the DDC diet through reducing the migration and infiltration of macrophages ( Figure 2D). However, the intervention with P. copri or L. murinus alone did not decrease the expression of Mcp-1 and Vcam-1, further confirming that the P. copri treatment facilitates the improvement that L. murinus has on inflammation in PSC mice.  and & indicates statistical difference compared to the DDC + P. copri + L. murinus group; * p < 0.05, ** p < 0.01, *** p < 0.001; # p < 0.05, ### p < 0.001; & p < 0.05, && p < 0.01, and &&& p < 0.001).

Coordination between P. copri and L. murinus Downregulates DDC-Induced Liver Fibrosis
The formation of extracellular matrix causes the aggregation of collagen and its winding spiral arrangement, which drives fibrosis in the liver [20,21]. To further investigate the effect of P. copri and/or L. murinus on liver fibrosis, SR staining was employed and revealed that the DDC diet could induce onion-like fibrosis by recruiting collagen; either P. copri and L. murinus alone or in combination induced an apparent alleviation of collagen deposit ( Figure 3A). In addition, the qPCR detected that the expression of Collagen 1a1 in the DDC-induced PSC mice decreased after treatment with P. copri ( Figure 3A) or L. murinus alone, or in combination, which suggests that these interventions may retard type I collagen accumulation in PSC mice. Moreover, we detected the expression of metalloproteinase 9 (Mmp9), adhesion molecule E-cadherin/Cah1, and tissue inhibitor of metalloproteinase 1 (Timp-1) in the liver ( Figure 3B). Compared with the control group, the levels of Mmp9 and Timp-1 in the DDC group significantly increased, whereas the levels markedly reduced in the DDC + P. copri and DDC + P. copri + L. murinus groups, suggesting that both interventions significantly slow the accumulation of extracellular matrix and attenuate fibrosis under the background of DDC diet. Nevertheless, L. murinus administration could not offset the increase in Mmp9, Cdh1, and Timp-1 in the PSC mice, suggesting that the alleviating effect of L. murinus treatment alone is not as effective as its co-intervention with P. copri in PSC mice. Compared to the DDC + P. copri and DDC + L. murinus groups, the DDC + P. copri + L. murinus group could significantly reduce the expression of the pro-fibrogenic factor Timp-1 in the PSC mice, indicating that the combined treatment with these two bacteria could better alleviate liver fibrosis caused by the DDC diet. As a marker of HSC activation, α-smooth muscle actin (α-SMA) plays an important role in liver fibrosis. As shown in Figure 3C,D, the IHC analysis demonstrated that the DDC group had the highest α-SMA expression, and both P. copri and L. murinus, either alone or in combination, significantly decreased the expression of α-SMA in the PSC mice. Consistent with this result, the Western blotting further confirmed that α-SMA protein was also reduced in the PSC mice after being given the treatment with L. murinus alone or the combined treatment with the two bacteria ( Figure 3E,F). Collectively, these results indicate that administering either P. copri or L. murinus alone, or both together, could alleviate DDC-induced liver fibrosis by downregulating the expression of Collagen 1a1, Mmp9, and α-SMA.

Treatments with P. copri and L. murinus Alone or in Combination Alleviate DDC-Induced Liver Fibrosis by Inhibiting TGF-β1/Smad Signaling
To elucidate the underlying mechanism through which P. copri or L. murinus inhibits fibrosis in the PSC mouse model, we detected the TGF-β1/Smad pathway. First, we found that the DDC diet increased the mRNA expression of Tgf-β1 and Tgf-βr II compared with the control group ( Figure 4A). Treatment with P. copri alone or combined treatment of P. copri and L. murinus could offset the increase in Tgf-β1 observed in the PSC mice. The Western blotting (Scale bar = 100 µm; original magnification 20×; * indicates statistical difference compared to the Con + PBS group, # indicates statistical difference compared to the DDC + PBS group, and & indicates statistical difference compared to the DDC + P. copri + L. murinus group; * p < 0.05, ** p < 0.01, *** p < 0.001; # p < 0.05, ## p < 0.01, ### p < 0.001; & p < 0.05, && p < 0.01, and &&& p < 0.001).

Treatments with P. copri and L. murinus Alone or in Combination Alleviate DDC-Induced Liver Fibrosis by Inhibiting TGF-β1/Smad Signaling
To elucidate the underlying mechanism through which P. copri or L. murinus inhibits fibrosis in the PSC mouse model, we detected the TGF-β1/Smad pathway. First, we found that the DDC diet increased the mRNA expression of Tgf-β1 and Tgf-βr II compared with the control group ( Figure 4A). Treatment with P. copri alone or combined treatment of P. copri and L. murinus could offset the increase in Tgf-β1 observed in the PSC mice. The 2.4. Treatments with P. copri and L. murinus Alone or in Combination Alleviate DDC-Induced Liver Fibrosis by Inhibiting TGF-β1/Smad Signaling To elucidate the underlying mechanism through which P. copri or L. murinus inhibits fibrosis in the PSC mouse model, we detected the TGF-β1/Smad pathway. First, we found that the DDC diet increased the mRNA expression of Tgf-β1 and Tgf-βr II compared with the control group ( Figure 4A). Treatment with P. copri alone or combined treatment of P. copri and L. murinus could offset the increase in Tgf-β1 observed in the PSC mice. The Western blotting further confirmed that the DDC diet stimulates the release of TGF-β1 to activate the phosphorylation of SMAD2/3, which leads to P-SMAD2/3 translocation from the cytoplasm to the nucleus and induces HSC activation and liver fibrosis ( Figure 4B). Compared with the DDC group, P. copri treatment and L. murinus treatment by themselves notably reduced P-SMAD2, and administration of P. copri combined with L. murinus led to a remarkable reduction in the levels of SMAD2 and P-SMAD2 ( Figure 4C). In addition, as shown in Figure 4D, SMAD3 protein expression was significantly activated by the DDC diet. When P. copri and L. murinus were used as interventions in the PSC mice alone, the SMAD3 protein level could not be significantly restored; in contrast, when both bacteria were administered together, significant reduction in SMAD3 expression was observed. P-SMAD3 protein was significantly inhibited by the bacterial interventions and could be restored to normal level after P. copri and L. murinus treatments alone or in combination. Collectively, these results suggest that P. copri and L. murinus treatments might alleviate DDC-induced liver fibrosis by inhibiting TGF-β1/Smad signaling, and the effect of combining P. copri with L. murinus is more obvious. cantly restored; in contrast, when both bacteria were administered together, significant reduc tion in SMAD3 expression was observed. P-SMAD3 protein was significantly inhibited by th bacterial interventions and could be restored to normal level after P. copri and L. murinus treat ments alone or in combination. Collectively, these results suggest that P. copri and L. murinu treatments might alleviate DDC-induced liver fibrosis by inhibiting TGF-β1/Smad signaling and the effect of combining P. copri with L. murinus is more obvious. The expression of TGF-β1 and its receptor, Smad2 in liver tissues was determined using qPCR. (B) Immunoblo ing bands of total and phosphorylated proteins of SMAD2 and 3. The gray values of the total and phosphorylated protein bands of SMAD (C) and SMAD3 (D) in the liver tissues of mice in each group were statistically analyzed using Im ageJ software. (* indicates statistical difference compared to the Con + PBS group, # indicates statis tical difference compared to the DDC + PBS group, and & indicates statistical difference compared to the DDC + P. copri + L. murinus group; * p < 0.05, ** p < 0.01, *** p < 0.001; # p < 0.05, ## p < 0.01; and & p < 0.05).

Changes in the Microbiome of Mice in Different Treatment Groups Related to P. copri and L murinus Interventions and Correlations between Different Bacterial Genera and Indicators
The PCoA analysis demonstrated that each group was distinguishable from others a the OTU level ( Figure 5A) and, compared to the intestinal microbiota diversity of the mice in the DDC group, the microbiota diversity was significantly higher in the DDC + P. copri DDC + L. murinus, and DDC + P. copri + L. murinus groups ( Figure 5B). The PLS-DA anal ysis further showed that each of the four control groups could be distinguished from each other, while there was a partial overlap in the background of the DDC treatment group ( Figure 5C), indicating that there are some common genera in the gut microbiota of PSC mice after treatment with P. copri and/or L. murinus. A Venn diagram was used to analyze

Changes in the Microbiome of Mice in Different Treatment Groups Related to P. copri and L. murinus Interventions and Correlations between Different Bacterial Genera and Indicators
The PCoA analysis demonstrated that each group was distinguishable from others at the OTU level ( Figure 5A) and, compared to the intestinal microbiota diversity of the mice in the DDC group, the microbiota diversity was significantly higher in the DDC + P. copri, DDC + L. murinus, and DDC + P. copri + L. murinus groups ( Figure 5B). The PLS-DA analysis further showed that each of the four control groups could be distinguished from each other, while there was a partial overlap in the background of the DDC treatment groups ( Figure 5C), indicating that there are some common genera in the gut microbiota of PSC mice after treatment with P. copri and/or L. murinus. A Venn diagram was used to analyze the number of common and unique OTUs in the intestinal microbiota of each group of mice, and it was found that the number of OTUs decreased after the DDC diet, with a total of 166 identical out sequences ( Figure 5D). Two-matrix correlation analysis was performed between the values of g_Prevotella, g_Lactobacillus, and g_Staphylococcus, which had the most abundant OTU values, and the detected values of ALP, ALT, TBIL, F4/80, and αSMA ( Figure 5E). The results showed a significant positive correlation between g_Staphylococcus and changes in these indicators (r = 0.44, p = 0.002; r = 0.46, p = 0.0009; r = 0.47, p = 0.0007; r = 0.50, p = 0.0003; and r = 0.44, p = 0.002, respectively), while g_Lactobacillus showed a significant negative correlation (r = −0.56, p < 0.001; r = −0.43, p = 0.002; r = −0.53, p = 0.0001; r = −0.65, p < 0.001; and r = −0.58, p < 0.001, respectively) and g_Prevotella showed only a statistically significant negative correlation with ALT (r = −0.31, p = 0.03). The above results illustrated that the increase in Lactobacillus abundance was signifi-cantly correlated with the recovery of serum biochemical indicators ALP, ALT, and TBIL ( Figure 1C), the decrease in macrophage number ( Figure 2C), and the decrease in αSMApositive area indicators ( Figure 3D). The addition of Prevotella significantly modulated the reduction in ALT indicators in Figure 1C, suggesting the addition of bacteria led to a significant improvement in DDC-induced liver injury, inflammation, and fibrosis. and changes in these indicators (r = 0.44, p = 0.002; r = 0.46, p = 0.0009; r = 0.47, p = 0.0007; r = 0.50, p = 0.0003; and r = 0.44, p = 0.002, respectively), while g_Lactobacillus showed a significant negative correlation (r = −0.56, p < 0.001; r = −0.43, p = 0.002; r = −0.53, p = 0.0001; r = −0.65, p < 0.001; and r = −0.58, p < 0.001, respectively) and g_Prevotella showed only a statistically significant negative correlation with ALT (r = −0.31, p = 0.03). The above results illustrated that the increase in Lactobacillus abundance was significantly correlated with the recovery of serum biochemical indicators ALP, ALT, and TBIL ( Figure 1C), the decrease in macrophage number ( Figure 2C), and the decrease in αSMA-positive area indicators ( Figure 3D). The addition of Prevotella significantly modulated the reduction in ALT indicators in Figure 1C, suggesting the addition of bacteria led to a significant improvement in DDC-induced liver injury, inflammation, and fibrosis. Two matrices were used to analyze the correlation between the abundance values of g_Staphylococcus ("g_" is short for "genus of"), g_Lactobacillus, and g_Prevotella and the quantitative values of ALP, ALT, TBIL, F4/80, and αSMA, * p < 0.05, ** p < 0.01, *** p < 0.001. Two matrices were used to analyze the correlation between the abundance values of g_Staphylococcus ("g_" is short for "genus of"), g_Lactobacillus, and g_Prevotella and the quantitative values of ALP, ALT, TBIL, F4/80, and αSMA, * p < 0.05, ** p < 0.01, *** p < 0.001.

Inner Communication and Fluorescence Localization of P. copri and L. Murinus in the Intestines of Mice
According to the above results, we suspected that P. copri and L. murinus might be co-located in the intestine and regulate hepatic fibrosis and inflammation-related signaling pathways through the liver-intestine axis to improve their synergistic effects. We constructed fluorescence-labeled strains of P. copri and L. murinus named Cy5-P. copri and TAMRA-L. murinus, in which two FDAA probes, the D-type amino acids DDap (Cy5)-NH 2 and DDap (5TAMRA)-NH 2 , were added into the media to separately cultivate these strains, as shown in Figure 6A. The fluorescence-labeled bacteria could be clearly ob-served by changing the excitation wavelength under a laser confocal microscope after in vitro culture ( Figure 6C,D). The in vivo imaging observation showed that both greenfluorescence-labeled Cy5-P. copri and red-fluorescence-labeled TAMRA-L. murinus localized in the gastrointestinal tract in the gavaged mice ( Figure 6E,F). As expected, observation of the frozen sections of the SI and colon using confocal microscopy further confirmed that the two types of fluorescence were visible in the same location in the intestine when the two strains were administered together ( Figure 6B). These results indicated that P. copri and L. murinus were stably present in the intestine and could coexist within the same location, providing a theoretical foundation for the combined intervention of P. copri and L. murinus in PSC mice to improve their liver injury, inflammation, and fibrosis.

Discussion
It has been reported that a change in gut microbiome composition is not only associated with some diseases but also modulates host metabolism [22][23][24][25]. Our previous study found that P. copri intervention significantly improved cholestasis and BA metabolism in PSC mice [9], but P. copri failed to reduce liver inflammation. Therefore, we speculated that the combination of beneficial bacteria might improve disease status in PSC mice, including reducing inflammation. Previous research studies reported that L. murinus was able to mediate anti-inflammatory effects in calorie-restricted mice [11], and we found that Lactobacillus showed a significant decrease in abundance after the DDC diet and an increase in abundance after the DDC + P. copri treatment in our earlier studies [9], suggesting that the addition of L. murinus might help P. copri to exert anti-inflammatory effects in PSC mice. In line with our hypothesis, the combined intervention of P. copri and L. murinus effectively improved liver injury and inflammatory and fibrotic processes in our DDC-induced PSC mouse model ( Figure 2). Moreover, the combined intervention of the two bacteria significantly reduced the expression of inflammatory factors such as Mpeg1 and Mcp-1 ( Figure 2D). Mpeg1 is a gene that encodes macrophage-specific expression, which represents changes in the number of macrophages [26,27]. Mcp-1 is a chemokine that specifically acts on monocytes to accelerate the aggregation of monocytes [28,29]. It has been found that activation of monocytes and macrophages is stimulated by the microenvironment in vivo, including infections by pathogens such as Mycobacterium tuberculosis infection, delayed hypersensitivity reactions and other diseases [30,31]. This leads to some qualitative and quantitative changes in the macromolecular components of the macrophage membrane and promotes inflammation [32,33]. In our previous study, we discovered that only when the intestinal mucosa was damaged by treatment with the DDC diet did P. copri breach the barrier into the internal environment to induce inflammation; in contrast, when the integrity of the gut barrier was preserved, this pro-inflammatory effect of P. copri did not manifest [9]. Therefore, the DDC diet induced severe liver inflammation and fibrosis in the PSC mice by recruiting a large number of macrophages and monocytes. The combined intervention of P. copri and L. murinus significantly reduced the response of macrophages and monocytes in the PSC mice, and the improvement in inflammatory immune response in the PSC mice was characterized by the expression of Mpeg1 and Vacm-1.
Previous research has reported that Staphylococcus is the most common purulent cocci, and organisms are susceptible to its infection when their skin mucosa is traumatized or when the immunity of the organisms is reduced [34,35]. In our study, the 16S rRNA intestinal microbiota sequencing analysis indicated that g_Staphylococcus was the most abundant genus and showed a significant positive correlation with the expression of ALP, ALT, TBIL, F4/80, and αSMA, suggesting that its increased abundance contributed to the exacerbation of disease indications in the DDC-induced PSC mice ( Figure 5E). Unlike Staphylococcus, Prevotella and Lactobacillus were negatively correlated with the expression of ALP, ALT, TBIL, F4/80, and αSMA, suggesting that an increase in P. copri and L. murinus abundance after the addition of the bacterial intervention contributed to reducing disease indicators in the PSC mice. This difference might be due to the enterotoxin produced by Staphylococcus during proliferation, which harms the metabolism of the gut microbiota and leads to disease symptoms in vivo [36]. Both P. copri and L. murinus have been found to regulate metabolism in vivo; for example, P. copri can help decompose proteins and carbohydrates [37,38], and L. murinus can produce lactic acid through fermentation of carbohydrates, promote animal growth, and regulate normal gastrointestinal microbiota [10,39]. Other studies have shown that Lactobacillus can significantly activate the phagocytosis of macrophages, stimulate peritoneal macrophages, and induce the production of interferon, which have special advantages in improving inflammation and fibrosis [40][41][42][43]. This is consistent with our findings, as shown in Figure 2, that L. murinus significantly reduced liver inflammation in the PSC mice after intervention. However, variations in the abundance of Staphylococcus, Prevotella, and Lactobacillus genera found in this study did not have a direct effect on liver inflammation and fibrosis; rather, the effect was exerted through an intermediary, which warrants further investigation. TGF-β1 and the SMAD protein family are pivotal players in the development and progression of liver fibrosis [16,[44][45][46][47]. In hepatic stellate cells, TGF-β1 produced by paracrine and autocrine can strongly stimulate the synthesis of collagen type I and other mechanistic components [48,49]. It has been confirmed that TGF-β1 can encode the expression of major extracellular matrix component genes and regulates the progression of liver fibrosis by activating metalloproteinase inhibitors to inhibit metalloproteinase activity [50,51]. SMAD-family proteins play a key role in the transmission of TGF-β1 signals from cell surface receptors to the nucleus, and different SMADs mediate the signal transduction of different TGF family members [52][53][54]. According to their functions, SMAD proteins can be divided into three subfamilies: receptor-activated SMAD (R-SMAD), common pathway SMAD (Co-SMAD), and suppressed SMAD (I-SMAD) [44,55]. R-SMAD can be activated by TGF-β1 receptors and form complexes with them, including SMAD2 and SMAD3. It has been pointed out that with the aggravation of liver fibrosis, the expression of SMAD2 and SMAD3 increases [56]. In a previous study, SMAD3 knockout mice had a lower degree of liver fibrosis induction than wild-type mice [57]. Co-SMAD, including SMAD4, is a medium commonly required in the signaling process of the TGF family. Studies have found that increasing the content of SMAD4 can reduce the synthesis of hepatic ECM, thus slowing down the process of liver fibrosis, which has potential clinical application value [58]. I-SMAD includes SMAD6 and SMAD7, which bind to the TGF complex and inhibit or regulate TGF-β1 signaling to improve liver inflammation and fibrosis [59]. The DDC-induced PSC mouse model used in this study presented severe liver inflammation and fibrosis. By detecting the expression levels of TGF-β1 and SMAD proteins in vivo (Figure 4), it was found that DDC activated the TGF-β1/Smad signaling pathway and significantly increased the expression of TGF-β1 and SMAD 2/3 proteins to induce liver inflammation and fibrosis. After being treated with the P. copri and L. murinus combined intervention, the reception and binding of TGF-β1 signal decreased, and the relative level of SMAD 2/3 also significantly decreased, suggesting that the TGF-β1/Smad signaling pathway might be inhibited by the bacterial intervention, thus leading to significant improvement in liver inflammation and fibrosis in the PSC mice. Furthermore, PSC is more common in men, although there is also a high number of women with PSC [2,60]. Female sex and estrogens are important regulators of bile acid production through critical hepatic feedback mechanisms, affecting enzyme activity and bile acid pool composition [61]. Moreover, we have confirmed that P. copri treatment regulates the metabolism of bile acids; thus, the interference caused by estrogen in the experiment was substantially attenuated and we chose male mice in this study [9]. We believe that our pioneering work has validity and will further explore the gender factor in future studies.

Culture and Quantification of Bacterial Strains
P. copri (DSM18205), blood agar base, and sterile defibrinated sheep whole blood were acquired from Mingzhou Biotechnology Co., Ltd. (Ningbo, China). Both L. murinus (BNCC194688) and MRS medium (lactic acid bacterial culture medium) were purchased from BeNa Culture Collection (Beijing, China). The above bacteria were cultured in an anaerobic culture tank (volume: 2.5 L, Mitsubishi, Tokyo, Japan) equipped with anaerobic gas-producing bags and indicators to provide an anaerobic environment. Colonies were cultured using Gram stain to determine the bacterial species. Bacterial DNA was extracted using the TIANamp Bacteria DNA Kit (Tiangen Biotechnology Co., Ltd., Beijing, China). qPCR was performed using a QuantStudio ® 5 Real-Time System (Applied Biosystems, Carlsbad, CA, USA), and the cyclic conditions were as follows: 95 • C for 5 min, 40 cycles at 95 • C for 10 s, 56 • C for 20 s, and 72 • C for 30 s. Eub was used as an internal reference to calculate the relative expressions using the 2 −∆∆Ct calculation method. The primer sequences used are shown in Table 1. Table 1. List of primer sequences used in this study.

Animal Experiments
Male 7-week-old C57BL/6J specific pathogen-free (SPF) mice were purchased from GemPharmatech Co., Ltd. (Nanjing, China) and housed in a 12 h light/dark cycle, with no restrictions on their food or water under conditions of controlled humidity (50 ± 5%) and temperature (22 ± 2 • C). All experiments involving mice were approved by the Nanjing Medical University Institutional Animal Care and Use Committee (Approval No. IACUC-2005013). A total of 80 mice were randomly divided into 8 groups of 10 mice each for the experimental treatment, and there was no significant difference in mouse body weight between the 8 groups. All experiments involving mice were conducted in parallel. There was a three-day adaptation period after the mice entered the facility. All mice in the facility drank purified water after disinfection. The mice were randomly divided into eight groups using random numbers based on the weight of each mouse, and statistical analysis of the body weight of the mice in each group after grouping showed no significant differences. The eight groups of mice were divided into four groups each of Con (short for control) and DDC groups and fed a normal or a customized diet for one week, respectively. The customized diet was 0.1% (w/w) DDC (CAS: 632-93-9; Sigma-Aldrich Co., Ltd., Saint Louis, MO, USA) supplemented diet synthesized by Jiangsu Xietong Pharmaceutical Bioengineering Co., Ltd. (Nanjing, China). The mice in the diet model group were raised separately from the other groups to ensure they maintained their SPF status. After that, the eight groups of mice were treated with gavage for one week with continuous feed supply, with four groups each of control and DDC groups, and each group was gavaged with the same volume of PBS (0.01 M, pH 7.2-7.4), P. copri, L. murinus, and a mixture of the two bacterial suspensions. The bacterial suspensions were washed, centrifuged, and re-suspended in PBS. A microplate reader read that OD 600nm was equal to 1, which corresponds approximately to a strain concentration of 1 × 10 9 CFU/mL. All mice were sacrificed after 12 h of starvation on the last day, and fresh small intestinal contents were collected. The collected blood was stored overnight at 4 • C, and the serum was collected via centrifugation. Liver tissues were immediately excised, sectioned, and fixed for further histological and immunohistochemical analysis, with the remainder stored at −80 • C until use.

Determination of Total Bile Acids (TBAs)
The content of TBAs in the liver and intestine tissues were determined using TBA kits (No. E003-2-1; Jiancheng Bioengineering Institute, Nanjing, China).

RNA Extraction and Quantitative Real-Time PCR
Total RNA was extracted from the liver tissues using the FastPure Cell/Tissue Total RNA Isolation Kit (Cat: RC101-01; Vazyme Biotech Co., Ltd., Nanjing, China) and reverse transcribed to cDNA using the HiScript II Q RT SuperMix (Cat: R223-01; Vazyme Biotech Co., Ltd.). Total RNA was measured for concentration and purity using a NanoDrop2000 spectrophotometer (NanoDrop Technologies; Thermo Fisher Scientific, Inc., Waltham, MA, USA). Then, real-time PCR was used to quantitatively detect the gene expression. Enriched cDNA, forward primer and reverse primer (0.2 µM, respectively, see Table 1 of oligonucleotides sequences; synthesized by Genscript Biotech Corporation, Nanjing, China), and Hieff ® qPCR SYBR ® Green Master Mix (5 µL, Cat: 11202ES08, Yeasen Biotech Co., Ltd., Shanghai, China) were mixed to a volume of 10 µL. The mixture was subjected to qPCR on AB RT-PCR (Q5). Gapdh was used as an internal control for target gene expression levels, and the relative gene expression levels were assessed using the 2 −∆∆Ct method.

16S rRNA Sequencing
Microbial DNA was extracted from the small intestine samples using the E.Z.N.A. ® soil DNA Kit (Omega Bio-tek, Norcross, GA, USA) according to the manufacturer's protocols. The final DNA concentration and purification were determined using a NanoDrop 2000 UV-vis spectrophotometer (Thermo Scientific, Wilmington, NC, USA), and DNA quality was checked using 1% agarose gel electrophoresis. The V3-V4 hypervariable regions of the bacterial 16S rRNA gene were amplified with primers 338F and 806R using a thermocycler PCR system (GeneAmp 9700, ABI, Carlsbad, CA, USA). Sequencing was pooled on an Illumina MiSeq platform. The classification of each 16S rRNA gene sequence was analyzed using the RDP Classifier algorithm (http://rdp.cme.msu.edu/, accessed on 15 July 2021) against the Silva 138/16S rRNA database (https://www.arb-silva.de/, accessed on 15 July 2021) with confidence threshold of 70%. The data were stored and analyzed using the I-Sanger Bio-cloud platform by Shanghai Majorbio Bio-pharm Technology Co., Ltd. (Shanghai, China).

Bacterial Labeling with Fluorescent Probes
Two fluorescent D-amino acid (FDAA) probes were purchased from Chinese Peptide Company (Hangzhou, China): (a) DDap (Cy5)-NH 2 , with 95% purity, excitation light wavelength (EX) = 649 nm, and emitted light wavelength (EM) = 670 nm, and (b) DDap (5TAMRA)-NH 2 , with 95% purity, EX = 555 nm, and EM = 580 nm. The structure and dosage were used as described in [62]. P. copri and L. murinus were cultured to an early stationary phase. FDAA was then added to the culture media to a final concentration of 300 µM. The DDap (Cy5)-NH 2 marker for P. copri and DDap (5TAMRA)-NH 2 marker for L. murinus were used, respectively. Two hours after the bacteria were labeled, they were washed twice with PBS and re-suspended in PBS to achieve an appropriate concentration for analysis using the Small Animal Live Visible 3D Imaging System (IVIS Spectrum, PerkinElmer, Waltham, MA, USA). Cy5-P. copri and 5TAMRA-L. murinus were administered by gavage to different groups of mice in a volume of 200 µL each, and another group of mice was gavaged with an equal amount of PBS as a reference. Immediately after the in vivo visible light 3D imaging of the mice, each intestinal segment was removed, frozen sections were created, the nucleus was stained with DAPI, and the location of bacteria in each intestinal segment was observed under a confocal microscope (Zeiss LSM700, Oberkochen, Germany).

Statistical Analysis
The data in this study were expressed as mean ± standard deviation (SD). Statistical comparisons were analyzed using SPSS (Version 20). For data that conform to the normality, homogeneity of variance was tested using Levene's test, and then a t-test was carried out to calculate significant differences between groups. A Kruskal-Wallis test was used when data did not follow a normal distribution. Bonferroni correction was used to adjust the overall significance level. The data were plotted using GraphPad Prism 8 (GraphPad, La Jolla,