Impact of elobixibat on liver tumors, microbiome, and bile acid levels in a mouse model of nonalcoholic steatohepatitis

Elevated bile acid levels have been associated with liver tumors in fatty liver. Ileal bile acid transporter inhibitors may inhibit bile acid absorption in the distal ileum and increase bile acid levels in the colon, potentially decreasing the serum and hepatic bile acid levels. This study aimed to investigate the impact of these factors on liver tumor. C57BL/6J mice received a one-time intraperitoneal injection of 25-mg/kg diethylnitrosamine. They were fed a choline-deficient high-fat diet for 20 weeks starting from 8 weeks of age, with or without elobixibat (EA Pharma, Tokyo, Japan). Both groups showed liver fat accumulation and fibrosis, with no significant differences between the two groups. However, mice with elobixibat showed fewer liver tumors. The total serum bile acid levels, including free, tauro-conjugated, glyco-conjugated, and tauro-α/β-muricholic acids in the liver, were noticeably reduced following elobixibat treatment. The proportion of gram-positive bacteria in feces was significantly lower in the group treated with elobixibat (5.4%) than in the group without elobixibat (33.7%). Elobixibat suppressed tumor growth by inhibiting bile acid reabsorption, and decreasing total bile acid and primary bile acid levels in the serum and liver. Additionally, the presence of bile acids in the colon may have led to a significant reduction in the proportion of gram-positive bacteria, potentially resulting in decreased secondary bile acid synthesis.


Introduction
Nonalcoholic steatohepatitis (NASH) and its associated liver cancer are increasing worldwide.Recent research has revealed that the causes of NASH and its resulting tumors include a multifactorial condition influenced by aberrant lipid metabolism, gut microbiome, bile acids (BAs), and oxidative stress.
BAs play a role in developing metabolic disorders, such as obesity, type 2 diabetes, and nonalcoholic fatty liver disease (NAFLD) [1].Primary BAs are synthesized in the liver and secreted into the duodenum, where they are metabolized into secondary BAs through gut microbiome metabolism, including conjugation and dehydroxylation.In the terminal ileum, 95% of BAs are reabsorbed into the liver through the portal vein via the ileal BA transporter (IBAT) [2].This process is known as enterohepatic circulation.Previous studies have reported that mice fed a high-fat diet showed significantly increased levels of secondary BA [3].High levels of BAs have been reported to contribute to liver tumors [4].
BAs also affect the gut microbiome.BA salts can regulate the growth of gram-positive bacteria by adjusting their concentration.The connection between BA concentration and the microbiome may play a crucial role in the gut-liver axis [5].This could in turn affect conditions, such as NASH and liver tumors [6].Several studies have reported decreased tumor incidence in NASH tumor models by reducing grampositive bacteria through vancomycin treatment.
In 2018, elobixibat (A3309), a selective inhibitor of the IBAT, was approved for treating functional chronic constipation.Elobixibat reduces BA reabsorption in the terminal ileum, resulting in increased BA excretion in stool and higher BA concentration in the colon, which enhances the secretion of water and electrolytes into the colon, improves intestinal motility, and eases colonic transit [7].
The choline-deficient high-fat (CDHF) diet + diethylnitrosamine (DEN) carcinogenesis model was used to evaluate the impact of elobixibat on serum BAs, liver tumors, and the microbiome [8].This study analyzed the effects of elobixibat administration on serum BA levels, liver tumors, and the microbiome.

Animals and study design
Three-week-old male C57BL/6J mice were randomly divided into two groups (Fig. 1a): (1) CDHF diet + DEN (control group) and (2) CDHF diet + DEN + elobixibat (elobixibat group) groups.The mice received a single intraperitoneal injection of 25-mg/kg DEN at 3 weeks of age.Then, they were fed a standard diet until they reached 8 weeks of age.For the next 20 weeks, mice in the control group were fed a CDHF diet (60 kcal% fat, Oriental Yeast, Tokyo, Japan), while those in the elobixibat group were fed a CDHF diet mixed with elobixibat (EA Pharma., Tokyo, Japan).The animals were housed in a controlled environment (temperature 23 ± 1 °C, humidity 50 ± 10%, 12-h light/dark cycle) at the animal facility with unlimited access to food and water.

Dose setting for elobixibat
Elobixibat was calculated to be 0.27 mg/kg/day.This study used animals (mean body weight of 23 g) based on the previously published data [9].The 50% inhibitory concentration of human IBAT is 0.53 nmol/L, and that of mouse IBAT is 0.13 nmol/L.Therefore, the inhibitory activity is four times higher in mice than in humans.The concentration of elobixibat in mice at 70% effective dose is 2.7 mg/kg; while at 50% effective dose, it is 0.27 (70% × [0.023/60]) 0.33 = 2.23 mg/ kg; this would be 110 mg/day for a 50-kg human, 11 times the amount normally used.Therefore, we set our effective capacity at 50%: 0.27 (50%) × [0.023/60]) 0.33 = 0.223 mg/kg.A CDHF + elobixibat diet containing 3 mg elobixibat per kg of CDHF diet was created and used based on the mean expected body weight and expected food intake.

Laboratory and staining methods
The details of the experimental method were given previously [9].Blood was obtained from the heart at the time of killing after 12-h fasting, and blood analysis was performed using a device by SRL Inc. (Tokyo, Japan).Total BAs were measured by Oriental Yeast Co., Ltd.(Tokyo, Japan).They measured the serum BA levels using the Aqua-auto Kainos total BA test kit (Kainos Laboratories, Inc., Tokyo, Japan).The serum BA levels were determined by combining 2.2-μL sample with 150-μL β-thionicotinamide adenine dinucleotide (oxidized) reagent.The mixture was allowed to react for Fig. 1 a Protocol: At 3 weeks of age, mice in both experimental groups received diethylnitrosamine (DEN) and were fed a normal diet (CE-2) for 8 weeks.After 8 weeks, the mice were randomly divided into two groups: the control group received only the choline-deficient high-fat (CDHF) diet, while the elobixibat group received CDHF and Elobixibat diet.a Liver Macro and hematoxylin and eosin (H&E) staining: The control group displayed a noticeable mass on the liver surface, while the elobixibat group showed a smaller mass.The tumor showed a distinct, densely packed cell mass on H&E staining.In the background, the liver showed fatty and fibrotic areas.b There were no significant differences in fibrosis and fatty areas.c Polymerase chain reaction analysis of liver tissue showed no significant differences between the two groups in αSMA and TGFβ levels.The liver tissue in the elobixibat group showed increased p21 expression.d The total number of liver surface tumors and the number of tumors > 2 mm were lower in the elobixibat group than in the control group.e The serum of the elobixibat group had a significantly lower total bile acid (TBA) level than that of the control group ◂ 5 min.Then, 50-μL mixture containing beta-nicotinamide adenine dinucleotide (reduced) and 3α-hydroxysteroid dehydrogenase reagent was added to the reaction mixture.The reaction rate was evaluated by performing a rate assay at 405/660 nm.
The liver was divided into subsamples for pathological analysis and mRNA purification.For staining, the extracted livers were fixed in 4% paraformaldehyde in phosphate-buffered saline for 16 h, dehydrated, and embedded in paraffin.The paraffin blocks were subsequently sliced into 3-μm-thick sections for hematoxylin and eosin (H&E) staining and Sirius red staining.Liver tissue sections were imaged at a magnification of 20 × using a BZ-9000 microscope (KEY-ENCE, Osaka, Japan).We calculated the ratio of fatty liver, inflammation score, Sirius red-positive area, and fibrosis score.The samples were randomized and the evaluators were unaware of the sample grouping.More than 10 images per mouse were evaluated independently by three experts including pathology expert (A.E).
The number of tumors on the liver surface was counted, and tumors were evaluated by two or more hepatologists following the methodology reported earlier [9] wherein, the level of c-Jun was increased, while that of p21 decreased.Tumor evaluation included assessing the size and abnormalities of the nuclei and hepatocyte arrangement primarily through H&E staining.In cases of ambiguity, Ki67 staining was performed to reach a definite conclusion.The diagnoses were made with the advice and cooperation of A.E.
Immunohistochemistry for Ki67 was performed using an automatic Ventana Discovery ULTRA Stainer (Ventana Medical Systems, Tucson, AZ, USA).The primary antibody (Rb Ki67/NB600-1252 antibody, Novus Biologicals, Abingdon, UK) was prediluted (1:100) and applied.The slides were incubated for 32 min at 37 °C.The slides were developed for 16 min with OmniMap anti-Rb HRP (Roche Diagnostics, Rotkreuz, Switzerland) and were detected using ChromoMap DAB (Ventana).The slides were counterstained with Ventana Hematoxylin II reagent for 4 min to visualize the nuclei.It was followed by a bluing reagent for 4 min, dehydrated, and mounted as in routine processing.

Analysis of fecal microbiome
DNA isolated from fecal samples was amplified using universal primers targeting the V3-4 regions of bacterial 16S rRNA.The PCR products were pooled to construct a sequencing library, sequenced using an Illumina MiSeq sequencer to generate pair-end reads using the MiSeq Reagent Kit v3 with 2 × 300 reads and 600 cycles (Illumina, San Diego, CA, USA).

Analysis of BA in serum and liver tissue
BAs in the serum and liver were analyzed per the method described by Asano et al. [13].Specifically, 530-μL methanol (MeOH)/acetonitrile (ACN) (1:1, v / v ) and 10-μL internal standard (IS) solution (cholic acid [CA]-d 4 , chenodeoxycholic acid [CDCA]-d4, deoxycholic acid [DCA]-d4, and ursodeoxycholic acid [UDCA]-d4 solutions, 0.05 mg/mL each) were added to 60-μL serum sample.After vortexing, the sample was centrifuged at 13,000 g for 10 min, and 500-μL supernatant was concentrated using a centrifugal concentrator at room temperature.The residue was reconstituted in 50 μL of 10% MeOH and filtered through a 0.22-mm filter.For processing liver samples, approximately 20-mg liver sample was homogenized in a fivefold volume of the MeOH/ ACN (1:1, v/v) solution.The homogenate was centrifuged at 13,000 g for 10 min.Next, 10-µL IS mixture solution (250 ng/mL each) was added to 50 µL supernatant and the sample was concentrated using a centrifugal concentrator at room temperature.The residue was reconstituted in 50 µL of 10% MeOH and filtered through a 0.22-µm filter.

Statistical analysis
Continuous variables were expressed as medians (interquartile range) and analyzed using the Mann-Whitney U-test.Categorical variables were analyzed using the chisquare test or Fisher's exact test.Statistical significance was set at p < 0.05.All statistical analyses were performed using GraphPad Prism version 9.0 (GRAPH PAD Software Inc, San Diego, CA, USA).
Microbiome comparisons, heat map, and diversity were visualized and statistically analyzed using the online microbiome data analysis platform (Microbiome Analyst [14]: https:// www.micro biome analy st.ca/ Micro biome Analy st).Its settings were as follows: low-count filter (minimum count: 0, percentage to remove: 0%) and data rarefying (data rarefying: Rarefy to the minimum library size).Alpha diversity was calculated with the Shannon index using the Mann-Whitney U-test.Beta diversity was estimated using the unweighted UniFranc distance and visualized using the principal coordinate analysis (PCoA) method.The significance of the PCoA plot was analyzed using permutational multivariate analysis of variance (PERMANOVA), which uses distance metrics by the Bray-Curtis index to confirm the strength and statistical significance of the sample groupings.Differential taxonomy between the two groups was compared using linear discriminant analysis effect size (LEfSe [15]: http:// hutte nhower.sph.harva rd.edu/ galaxy/) with linear discriminant analysis (LDA) score > 2 and p-value < 0.05.

Elobixibat did not affect body weight, serum lipid concentrations, and fatty and fibrotic areas
No significant differences were noted in body weight, food intake, serum lipid levels, or alanine transaminase (ALT) between the elobixibat and control groups (Table 1).However, the elobixibat group had higher alkaline phosphatase and cholinesterase levels than the control group.Both groups showed increased aspartate aminotransferase and ALT levels as revealed by histological examination of inflammatory cell infiltration.Both the groups had comparable extent of fatty and fibrotic areas (Fig. 1b, c).The levels of alpha-smooth muscle actin and transforming growth factor beta were also comparable between the two groups (Fig. 1d), suggesting that elobixibat does not significantly reduce body weight or improve fatty liver or fibrosis.
About 94.6% of the fecal microbiomes in the elobixibat group mainly comprised gram-negative bacteria Alpha diversity is shown in Fig. 2a.The samples from the pre-elobixibat group were taken at 8 weeks of age from mice fed a normal diet before elobixibat administration.Alpha diversity was lower in the CDHF diet + DEN group than in the pre-elobixibat group and even lower in the elobixibat group.The PCoA plot (Fig. 2b) also showed a clear difference in microbiome compositions, with the pre-elobixibat and elobixibat groups being significantly different despite being obtained from the same mouse.The proportions at the phylum level of Bacteroidota, Firmicutes, and Proteopacetia were 52.1%, 41.1%, and 2.78%, respectively, in the preelobixibat group (Fig. 2c).These bacteria have been reported to be dominant in many previous reports and are believed to have a typical distribution.In contrast, in the CDHF diet + DEN control group, Firmicutes and Bacteroidota proportion was 32.9% and 44.3%, respectively, lower than those in the pre-elobixibat group.The relative abundances of Verrucomicrobiota, Desulfobacterota, and Proteobacteria were higher in the elobixibat group than in all other groups, with Bacteroidota at 56.4%.The total relative abundance of gram-positive bacteria Actinobacteria (0.081%), Firmicutes (5.32%), and Patescibacteria (0%), was 5.4% in the elobixibat group, significantly lower than that in the pre-elobixibat (41.7%) and control groups (33.7%).The proportions of genus-level bacteria are shown in Fig. 2d.In the Control group, there included Bacteroides (22.0%),Akkermansia (11.5%), and Alloprevotella (11.4%).In contrast, the elobixibat group was dominated by gramnegative bacteria, with Bacteroides (27.2%),Akkermansia (22.1%), and Alloprevotella (8.8%) being the three most abundant bacteria.Gram-positive bacteria had a lower proportion: Lactobacillus (Firmicutes): 1.04%, Faecalibaculum: 0.5%, Lachnospiraceae_FCS020_group: 0.4%, and Clostrid-ium_sensu_stricto_1: 0.3%.The heat map in Supplementary Fig. 2 demonstrates the clear differences in microbiome distribution among the three groups.

Comparison of microbiome between the two groups and pre/post changes in the elobixibat group
Figure 3 shows the bacteria with significantly different relative abundances between the two groups.Figure 3a compares the CDHF diet + DEN and CDHF diet + DEN + elobixibat groups using different mice of the same age and diet.Elobixibat significantly reduced the number of grampositive bacteria from the Firmicutes, Deferribacterota, and Actinobacteria phyla and increased the number of Proteobacteria compared to the control group.Similar changes were observed in same mice before and after elobixibat treatment (Fig. 3b), but one group was 8-week-old and received a normal diet (Pre-elobixibat) and the other group was 28-week-old and received CDHF diet + elobixibat (Post-elobixibat).These results indicate that the elobixibat-induced changes in the microbiome do not revert to the condition before CDHF diet + elobixibat administration but rather result in an even lower proportion of gram-positive bacteria than before CDHF diet + elobixibat administration.A detailed list of the bacteria with changed relative abundances is provided in Supplementary Table 1, which also highlights the increased relative abundance of Bifidobacterium and Clostridium bile salt hydrolase-rich bacteria in the elobixibat group compared to the control group.

Levels of certain primary and secondary BAs reduced in liver tissue and serum in the elobixibat group
Out of 39 targeted BAs, 25 were detected.The primary BAs, including CA, glycocholic acid (GCA), and taurocholic acid (TCA), were significantly reduced in the serum samples of the elobixibat group (Fig. 4a).In rodents, CDCA is metabolized to alpha-muricholic acid (α-MCA) and beta-muricholic acid (β-MCA).Therefore, MCAs, and not CDCA, are the specific BAs in mice.BAs related to MCA, including tauro-α-MCA (TαMCA), β-MCA, and tauro-β-muricholic acid (TβMCA), were also significantly decreased in the elobixibat group.All types of primary BAs were increased in the elobixibat group.Secondary BAs, including 3-keto-cholic acid (3-ketoCA), 7-ketodeoxycholic acid (7-ketoDCA), ursocholic acid (UCA), and norcholic acid (NorCA), were significantly lower in the elobixibat group.Besides, dehydrocholic acid level was significantly higher, and that of DCA was not significantly different between the two groups (Fig. 4b).To show the relationship between these bile acids and the relative abundance of bacteria at the genus level, we computed correlation coefficients that were corrected with False Discovery Rate (FDR) and represented them in a heatmap (supplement figure 2).This analysis suggested a correlation between certain bacteria and bile acids.However, as these results were derived from comparisons between two groups, they indicate correlations between bile acids and microbial communities that are affected by elobixibat.Therefore, additional investigations are necessary to directly elucidate the relationship between bile acids and bacteria.

Discussion
BAs, microbiome, and lipid metabolism are believed to be associated with NASH and tumorigenesis.However, this is the first study to show that elobixibat reduces tumor incidence without any fibrotic and lipid changes; it lowers the serum BA and primary BA levels in the liver by inhibiting BA absorption in the terminal ileum.The study also found that BAs in the colon significantly reduced proportion of gram-positive bacteria.The decrease in Clostridium clusters, a gram-positive bacteria, is related to the 7α-dehydroxylation carried out by enteric bacteria during the production of secondary BAs, which lowers serum BA levels, suggesting that BAs are directly related to liver tumors and that elobixibat does not worsen liver fibrosis.
Patients with NAFLD reportedly have higher total plasma BAs (with a disproportionate increase in primary BAs) than healthy individuals and high levels of GCA and TCA are associated with liver inflammation.Furthermore, 7-ketoDCA is related to ballooning and fibrosis [16].Moreover, in Japanese population, the fecal levels of BAs reportedly change relatively early in patients with chronic hepatitis C [17].In the early stage of cirrhosis, the concentrations of total BAs and primary conjugated BAs significantly increase in patients with hepatocellular carcinoma (HCC) and cirrhosis compared to those with cirrhosis only [18].Luo et al., [19] in their large-scale, multicenter study, suggested that phenylalanyl-tryptophan, metabolized by intestinal bacteria, and GCA could be used as biomarkers for early HCC detection.BAs and the microbiome are gaining attention for their potential association with HCC as they may act as tumor promoters or initiators.Fu et al. [20] suggested that BAs may promote cell proliferation, inflammation, and oxidative stress, possibly causing DNA damage and tumor growth.Furthermore, Sun et al. [21] showed that Sirt5 deficiency synergizes with oncogenes to increase BA biosynthesis in hepatocyte peroxisomes.This abnormal accumulation of BAs distorts microphage polarization via activating farnesoid X receptor (FXR) (Nr1h4), thereby creating an immunosuppressive tumor microenvironment.Furthermore, cholestyramine almost abolishes the effect of Sirt5 deficiency on the increased liver tumor volume.This indicates the importance of suppressing the BA levels in the liver.Elobixibat is an inhibitor of the IBAT, primarily located in the terminal ileum, and reabsorbs approximately 95% of secondary BAs into the portal vein.Graffner et al. [22] reported that A4250, another IBAT inhibitor, decreased the plasma levels of total BAs in phase-1 clinical trials compared to controls.Similarly, elobixibat suppresses reabsorption, which may reduce primary BAs in the liver.Nakajima et al. [23] showed that when patients with chronic constipation were administered elobixibat, their fecal BA excretion increased and the serum BA levels decreased.The improvement in constipation following elobixibat treatment was associated with an increase in total BAs, particularly primary BAs, consistent with the results of our study.Using a mouse model of methionine-and choline-deficient diet-induced NASH, Yamauchi et al. [24] found that elobixibat administration significantly reduced liver fibrosis and inflammation.They also noted that although the total serum BA increased, elobixibat decreased total serum BA to the same level as in mice fed on a standard diet.We also found a marked decrease in serum BA levels, but no change in fibrosis or lipidosis in response to choline deficiency alone.As the CDHF diet has a higher fat component than methionine-and choline-deficient diets, the difference was not significant.
Yoshimoto et al. [3] reported that high levels of BAs are associated with increased liver tumors.They also found that an increase in DCA, a type of secondary BA, is involved in carcinogenesis when a high-fat diet increases.
Ma et al. [4] showed a relationship between serum BAs and liver tumors, independent of fatty liver and fibrosis.They reported that intestinal bacteria use BAs as messengers to regulate anti-tumor immunity against primary and metastatic liver tumors by regulating hepatic NKT cell accumulation.However, the TβMCA results were different from those obtained in their study.They demonstrated an increase in hepatic TβMCA in a mouse model treated with antibiotics, and showed that TβMCA induced Cxcl16 mRNA in vitro.This result is related to tumor suppression and differs from our findings, wherein TβMCA level was significantly decreased.As aforementioned, liver tumors are associated with high levels of BAs.The major BAs in rodents are CA, CDCA, UDCA, αMCA, and βMCA (with βMCA being the major component).TαMCA and TβMCA exert their effects primarily by exhibiting antagonistic activities against FXR [25,26].This means that they may act antagonistically against tumor formation or growth owing to the high BA levels.In their model, levels of primary BAs, such as βMCA and TCDCA, were high.In contrast, elobixibat inhibits the reabsorption of primary BAs, leading to decreased bile stasis in the liver.Hence, this mechanism may be an extension of the mouse model subjected to antibiotic treatment, but with reduced hepatic BA levels.This facilitates anti-tumor effects even in the presence of lower levels of hepatic TβMCA.Rather, it may be a model where tumor suppression is achieved using cholestyramine to inhibit BA absorption.We hypothesize that modulating microbiome facilitates formation of a liver internal environment where tumors are more suppressed.
Sun et al. [27] showed that BAs can promote liver cancer via increased inflammatory signaling, such as regulating Nuclear factor kappa B, tumor necrosis factor alpha, and interleukin-1β.In our study, we observed a decrease in the primary BA, CA, which may have inhibited tumorigenesis.Although increased secondary BAs have been reported in liver tumors and serum, we found little change in secondary BAs in the liver tissue.Some minor changes were observed, including a reduction in NorCA, which has been linked to tumorigenesis via negative regulation of FXR [28].We believe that the reduction of primary BAs was an important factor in the reduction of tumor incidence in liver tissue with respect to BAs.
Large amounts of BAs inhibit the growth of gram-positive bacteria, and may induce shift from gram-positive to gram-negative bacteria in the colon.Several types of bacteria synthesize secondary BAs.Primary and some secondary BAs may decrease in the intestine, when bacteria rich in bile salt hydrolase, such as Clostridium spp., are in a lower proportion.In T5KO mice, vancomycin treatment decreased the proportion of gram-positive Lachnospiraceae and Ruminococcaceae, belonging to the phylum Firmicutes; Bifidobacteria of the phylum Actinobacteria; and Clostridium cluster XXII, which produces secondary BAs [29].Selective depletion of intestinal bacterial communities, such as Clostridium cluster XIVa, which produces secondary BAs, was observed.The absence of hepatocarcinogenesis in vancomycin-treated mice strongly correlated with a significant decrease in circulating secondary BAs [28].Notably, we found that decrease in bacteria involved to BA synthesis inhibited carcinogenesis.In our study, gram-positive bacteria may have been significantly depleted, leading to a similar pathogenesis.
The gram-negative bacteria of microbiome contain pathogen-associated molecular patterns, such as lipopolysaccharides, in their cell walls [30], which can enter the portal vein and cause inflammation and fibrosis via activating astrocytes in the liver, thus increasing the risk of carcinogenesis.However, next-generation sequencing shows the relative abundance of bacterial flora and not the absolute abundance and we were unable to measure this in our study.However, the relative number of gram-negative bacteria may have been lower than the overall bacterial abundance owing to the selective reduction of gram-positive bacteria.
This study had some limitations.First, it was performed only in a single model; further examination using various models is necessary.Second, our sample size was relatively small.Kishida et al. [8] reported 100% tumor rate in a mouse model at 24 weeks of age, with an average tumor diameter of 2.9 ± 2.8 mm.Our 28-week model has also demonstrated a high incidence of tumors consistently.Here, we repeated our experiment in two separate batches, with three mice in each batch, to ensure reproducibility of our results.Third, although the relative abundance of the microbiome was measured, it is important to compare the actual number of gram-positive bacteria to confirm if there was an increase.Finally, further molecular biological studies are required to understand the relationships between fibrosis, inflammation, and fat.The relationship between the microbiome and BAs, in particular, will require well-designed animal experiments and data analysis.A complete elucidation of the mechanisms will likely require network analysis and validation using other models in the future.
In conclusion, elobixibat reduced the levels of BAs in both the serum and liver by inhibiting BA absorption in the terminal ileum.This may inhibit the tumor development.Gram-positive bacteria play a significant role in synthesizing secondary BAs, but are negatively impacted by BAs.This study showed that elobixibat effectively suppressed the incidence of tumors and the presence of gram-positive bacteria in stool samples (Fig. 5).

Fig. 2 aFig. 2
Fig.2a Stools collected from 28-week-old mice in the Control group were labeled "post control," while stools collected at 8 weeks from the elobixibat group were labeled "pre-elobixibat."The stools collected at 28 weeks from the Elobixibat group were designated as "post-elobixibat.""Pre-elobixibat" and "post-elobixibat" were collected from the same mouse.Stool analysis showed that "pre-elobixibat" had the highest alpha diversity, as measured by the Chao1, Shannon, and all observed species indices.In contrast, "post-elobixibat" had the lowest alpha diversity.Furthermore, "pre-elobixibat" had the highest alpha diversity, as measured by the Chao1, Shannon, and all

Fig. 3 a 3 ◂Fig. 4 a
Fig. 3 a The comparison of the control group and the elobixibat group by LEfSe.b the comparison of the microbiome before (8-week-old) and after elobixibat administration.The names of bacteria with red bars indicate a significantly higher relative abundance in the control group (a) or the post-elobixibat group (b) compared to the other group.The upper (central) circular graph presents a plot cladogram with information on the higher level.The linear discriminant analysis (LDA) size effect of the LEfSe and the name of the bacteria in the LEfSe results are shown in the Supplementary Fig.3◂

Table 1
Weight and blood testsThe Mann-Whitney U-test showed no significant difference (p < 0.05) between the two groups CDHF choline-deficiency high-fat, DEN diethylnitrosamine, ALT alanine aminotransferase, AST aspartate aminotransferase