Lithocholic Acid May Improve Type 2 Diabetes by Balancing Gut Bacteria and Bile Acids

Data for this analysis were derived from our group’s previously published study in type 2 diabetes [13]. We reclassified participants by fasting plasma glucose (FPG) in accordance with ADA Standards of Care (T2DM: FPG ≥ 7.0 mmol/L, n = 80; controls: FPG < 7.0 mmol/L, n = 40) and compared serum bile acid profiles between groups.

LCA (Sigma-Aldrich, St. Louis, MO, USA), metformin hydrochloride tablets (0.5 g per tablet; Merck Ltd., Darmstadt, Germany; batch ACD0048), and streptozotocin (MP Biomedicals, Solon, OH, USA; cat. no. 100557) were used in this study. Phosphate-buffered saline (PBS; Adamas Reagent Co., Shanghai, China; cat. no. C8020), Humulin insulin injection (Eli Lilly and Company, Indianapolis, IN, USA; batch D386919A), and citrate–sodium citrate buffer (SenBeiJia Biotechnology, Nanjing, China; cat. no. BL-S031) were used for sample preparation and physiological support.

The Concept 400 anaerobic workstation was sourced from Panmax Technology Co. (Garden City, NY, USA). The H1850 high-speed centrifuge was obtained from Xiang Yi Company (Jiaxing, China), and the Eppendorf 5424R refrigerated microcentrifuge was sourced from Eppendorf (Hamburg, Germany). The RTS-1C Automatic Microbial Growth Curve Analyzer was obtained from Guangzhou Shengrui Biotechnology Co. (Guangzhou, China).

Male C57BL/6 mice (7–8 weeks, 23–25 g) were acclimated for one week before random assignment into five groups: normal control (NC, n = 3), high-fat diet (HFD, n = 4), type 2 diabetes model (TM, n = 4), metformin-treated (MET, n = 4), and lithocholic acid-treated (LCA, n = 4). All animals were specific pathogen-free, wild-type mice and had not undergone any experimental procedures prior to study initiation. Animals were housed under a 12 h li12 h12-h dark cycle at room temperature (22–24 °C), with ad libitum access to food and water. No additional environmental enrichment was provided.

The experimental unit was a single animal. Animals were randomly allocated to experimental groups using a random number table. Potential confounders were minimized by maintaining identical conditions across groups, including temperature, humidity, light–dark cycle, diet, and handling procedures. No inclusion or exclusion criteria were defined a priori. All animals that entered the experiment were included in the analysis, and no animals or data points were excluded. These group sizes reflect the numbers of animals analyzed in each group. Group allocation was known to the investigator responsible for randomization, whereas outcome assessment and data analysis were performed blinded to group allocation.

Apart from NC, other groups were provided with a high-fat diet (D12492; Research Diets, lot 22090102), while NC received standard chow. MET and LCA groups received daily oral gavage of metformin (200 mg/kg) or lithocholic acid (50 mg/kg) for eight weeks; the remaining groups received saline. These procedures were designed to establish a T2DM model and to evaluate the metabolic effects of LCA. All animal procedures were approved by the Institutional Animal Care and Use Committee of the Shanghai Model Organisms Center (Approval Code: 2022-0044; Approval Date: 26 October 2022). A detailed animal use protocol, including the study objectives, experimental design, and planned procedures, was developed prior to study initiation and approved by the same IACUC.

Body weight was recorded weekly, and fasting glucose monitored biweekly. In week 6, TM, MET, and LCA mice were administered a single intraperitoneal dose of streptozotocin (40 mg/kg), whereas NC and HFD mice were given citrate buffer. Ten days after streptozotocin administration, metabolic evaluations were carried out, and body weight, fasting blood glucose, serum insulin concentrations, HOMA-IR, OGTT, and insulin sensitivity were selected as outcome variables.

Animals were routinely inspected throughout the study for overall health and well-being, with particular attention to body weight, food consumption, posture, spontaneous activity, grooming behavior, and coat appearance. All procedures were carried out in compliance with institutional animal welfare regulations, with measures taken to limit discomfort and distress. No adverse events, whether anticipated or unanticipated, occurred during the experimental period. Humane endpoints were not triggered, and no animals required premature removal from the study. This study was exploratory in design and did not define a single primary outcome for sample size determination.

Feces were collected 24 h prior to euthanasia and stored at −80 °C. At study endpoint, following a 12 h fast, mice were induced with 4% isoflurane anesthesia, and blood was drawn through cardiac puncture. After clotting at room temperature for 30 min, samples were centrifuged at 4 °C, and serum stored at −80 °C. Liver, spleen, and cecal contents were also harvested and frozen at −80 °C for subsequent analyses.

Fasting blood glucose was measured using an Accu Chek Performa meter (Roche Diagnostics, Basel, Switzerland) with compatible test strips. Fasting insulin (FINS) was measured using an insulin ELISA kit (Crystal Chem, Elk Grove Village, IL, USA, Cat. No. 90080). Assays for alanine aminotransferase (ALT), aspartate aminotransferase (AST), total cholesterol (TC), triglycerides (TG) were performed using kits from Nanjing Jiancheng Bioengineering Institute, Nanjing, China (Cat. Nos. C009-2-1, C010-2-1, A111-1-1, A110-1-1). The levels of lipopolysaccharide-binding protein (LBP) in mice were quantified using the Mouse LBP ELISA Kit (LPS Binding Protein) from Abcam, Cambridge, UK (Catalog No. ab269542).

During the oral glucose tolerance test (OGTT), mice fasted for 6 h (ad libitum water). Baseline tail-tip blood was sampled at 0 min for FBG. A 50% glucose solution (2 g/kg) was administered by oral gavage, and additional tail-tip glucose measurements were taken at 15, 30-, 60-, 90-, and 120 min. Glucose tolerance was quantified by AUC analysis, and glucose responses were also analyzed after normalization to baseline.

During the insulin tolerance test (ITT), Following a 6 h fast, baseline blood glucose was recorded at 0 min, then insulin (0.5 U/kg) was given intraperitoneally. Blood glucose readings were collected at 15, 30-, 60-, 90-, and 120 min post-injection. Similarly, AUC values were derived from both raw glucose measurements and baseline-normalized responses in the ITT.

Fecal samples from each mouse group were collected, and genomic DNA was extracted using the OMEGA Soil DNA Kit (Omega Bio-Tek, Norcross, GA, USA). DNA quality was evaluated by Nanodrop spectrophotometry and electrophoresis on an agarose gel. The V3-V4 region of the 16S rRNA gene was amplified with universal primers, purified using Vazyme VAHTS™ DNA Clean Beads (Nanjing, China), and quantified with the PicoGreen assay. Amplicon sequencing was conducted on the Illumina NovaSeq 6000 platform, San Diego, CA, USA (2 × 250 bp). Raw sequences were demultiplexed with Demux, and primers were removed with Cutadapt. Quality control, denoising, and chimera removal were performed using DADA2 to create an amplicon sequence variant (ASV) table. Alpha and beta diversity were analyzed with QIIME2, and principal coordinate analysis (PCoA) was conducted. Taxa with differential abundance were identified using LEfSe (LDA score > 4).

Bile acids in liver samples were analyzed using a Waters (Milford, MA, USA) ACQUITY UPLC HSS C18 column (1.8 μm, 100 mm × 2.1 mm). The mobile phase consisted of 0.01% acetic acid with 5 mmol/L ammonium acetate in ultrapure water (A) and 0.01% acetic acid in acetonitrile (B). Mass spectrometry was conducted in negative electrospray ionization (ESI) mode. Quantification was performed using multiple reaction monitoring (MRM) mode. Data acquisition and analysis were carried out with SCIEX Analyst 1.6.3 and MultiQuant 3.0.3 software. Standard curves were established using known standards, and integrated peak area ratios were applied to ensure accurate quantification.

The A. muciniphila strain (ATCC BAA-835) used in this study was purchased from Beijing Baio Bio Biotechnology Co., Ltd. Prior to experimentation, the strain was activated in Brain Heart Infusion (BHI) medium (Hopebio, Qingdao, China) under anaerobic conditions at 37 °C for 24 h. After activation, the bacteria were exposed to different concentrations of LCA for 24 h. The range of LCA concentrations was selected based on previous studies that reported effective dose levels [14,15]. The optical density (OD) at 600 nm was measured every six hours to monitor bacterial growth.

A fecal sample was collected from an older individual who was clinically healthy and had not experienced any gastrointestinal conditions or been exposed to antibiotics or prebiotics in the preceding three months. The samples were suspended in a 0.1 mol/L phosphate-buffered saline solution (pH 7.2) at a mass-to-volume ratio of 10:1 and vortexed for two minutes to achieve homogenization. The mixture was centrifuged at 2000 rpm for 10 min. The supernatant and pellet were then collected separately. The supernatant was subjected to bile acid profiling, whereas the pellet served for 16S rRNA gene assessment and quantification of total bacterial load.

Quantification of total bacterial abundance in sediment from in vitro fecal fermentation systems was achieved through real-time PCR. A plasmid harboring the full-length 16S rRNA gene derived from Lactobacillus served as the quantification reference. Serial dilutions of the plasmid, spanning from 10³ to 10⁹ copies per microliter, were prepared to construct standard curves. Amplification targeted the conserved bacterial 16S rRNA gene using a universal primer pair (Uni331F: 5′-TCCTACGGGAGGCAGCAGT-3′; Uni797R:5′-GGACTACCAGGGTATCTAATCCTGTT-3′), generating an amplicon of approximately 467 base pairs. Each PCR reaction was carried out in a final volume of 10 μL. Ct values obtained from the serial standards were plotted against log₁₀-transformed copy numbers to generate the calibration curve. Absolute bacteria copy numbers in test samples were determined by substituting the corresponding Ct values into the regression model.

Fecal samples (50 mg per mouse) were suspended in 0.1 M phosphate-buffered saline (PBS, pH 6.8), vortexed for 3 min, and homogenized. After centrifugation at 9000× g for 30 min at 4 °C, the supernatant was used to measure β-glucuronidase (β-GUS) activity. The assay mixture contained 20 μL supernatant, 175 μL PBS, and 5 μL of 4-methylumbelliferyl β-D-glucuronide (4-MUG), incubated at 37 °C for 30 min [16]. The reaction was stopped with acetonitrile, and fluorescence was measured at 340 nm excitation and 460 nm emission.

Fecal extracts were prepared as described for the GUS assay, and the clarified supernatant was used as the enzyme source. BSH activity was quantified with a far-red fluorogenic bile-acid substrate reported [17]. Briefly, the supernatant and probe were combined in PBS (pH 6.8) and incubated at 37 °C; fluorescence was recorded kinetically on a microplate reader. Heat-inactivated samples served as negative controls.

Total RNA was carried out using the FastPure^® RNA Isolation Kit V2 (Vazyme, RC112). Reverse transcription was conducted with HiScript^® III RT SuperMix for qPCR (+gDNA wiper) (Vazyme, R323). Quantitative PCR (qPCR) was conducted using ChamQ SYBR qPCR Mix (Vazyme, Q711). Primers were available by Sangon Biotech (Shanghai) Co., Ltd. (Shanghai, China).

In this study, statistical analyses were performed using GraphPad Prism 10.0 (GraphPad Software, San Diego, CA, USA). Comparisons were performed between the HFD and TM groups versus the NC group, and between the MET and LCA groups versus the TM group. For comparisons between two groups, unpaired or paired data was conducted using two-tailed Student’s t-test or the Mann–Whitney U test, depending on data distribution. For comparisons involving three or more groups, one-way analysis of variance (ANOVA) was employed, followed by Tukey’s multiple comparisons test. Nonparametric data were analyzed using the Kruskal–Walli’s test, with Dunn’s post hoc test for pairwise comparisons. Categorical data were assessed using Fisher’s exact test. In the figures, the “#” symbol indicates comparisons between the HFD or TM groups and the NC group, whereas the “*” symbol indicates comparisons between the Met or LCA groups and the TM group. All data are expressed as mean ± standard deviation (SD), and a p-value of less than 0.05 was considered statistically significant.

Building on our previous cohort [13], we compared the profile of serum bile acids in 40 healthy individuals and 80 patients with T2DM (Figure 1A,B). A supervised OPLS-DA model showed clear separation between the two groups based on the bile-acid matrix (Figure 1C). In the volcano plot, the T2DM group showed one bile acid significantly increased and eight significantly decreased relative to controls (Figure 1D). Applying p < 0.05 with VIP ≥ 1 retained six differential bile acids (Figure 1E–K); among them, 3-keto-LCA carried the top VIP (VIP value = 2.064), indicating the strongest driver of group separation.

3-keto-LCA, also known as 3-oxo-LCA, arises from LCA through C3-position oxidation as a downstream product [14]. Given the limited murine gavage literature on 3-keto-LCA and its unclear target profile, we administered its precursor, LCA, orally in mice to evaluate bile-acid-based intervention effects relevant to T2DM.

To evaluate the metabolic benefits of LCA in T2DM, we employed a diet-and streptozotocin-induced mouse model of T2DM and administered LCA (50 mg/kg/day) for 4 weeks (Figure 2A,B). Food and water intake showed only minor fluctuations over the experimental period, with no marked differences among the HFD, TM, MET, and LCA groups, while the NC group consistently exhibited lower food intake and slightly higher water consumption (Figure 2C,D). Compared to TM group, LCA-treated mice significantly attenuated body weight gain throughout the intervention period (Figure 2E). LCA treatment significantly reduced both fasting blood glucose and insulin concentrations (Figure 2F,G), leading to a marked reduction in HOMA-IR values and indicating improved insulin sensitivity (Figure 2H). OGTT (Figure 2I,J) and ITT (Figure 2K,L) results not significantly different between the LCA and TM groups, although a smaller trend was observed.

In addition to regulating glucose, LCA has beneficial effects on the lipid profile and liver enzyme levels. LCA can significantly reduce TC, while TG levels show a downward trend, though this is not statistically significant (Figure 2K,L). Moreover, LCA attenuated liver injury, as indicated by reduced ALT activity and a downward trend in AST levels (Figure 2M,N).

Taken together, these data suggest that LCA provides T2DM mice with a range of metabolic benefits, including improved glucose and lipid metabolism, insulin sensitivity, and liver function.

This study employed 16S rRNA gene sequencing to investigate the impact of LCA on the gut microbiota. Alterations in the gut microbiota composition were observed in the T2DM mouse. Contrary to typical findings of reduced α-diversity in T2DM mice, the diabetic model in this study exhibited a significant increase in α-diversity. Following LCA treatment, Chao1 and Shannon index decreased significantly, approaching levels observed in the normal control group (Figure 3A,B). Further analysis of β-diversity based on PCoA and PCA revealed that the microbiota structure of LCA-treated mice was like that of the control group (Figure 3C,D).

Taxonomic profiling at phylum and genus levels revealed that the LCA significantly altered the relative abundance of major microbial taxa (Figure 3E,F). LEfSe analysis indicated that in TM group, dominant taxa included members of the phyla Firmicutes (e.g., Clostridiales, Ruminococcaceae) and Bacteroidetes (such as Bacteroides, Bacteroidaceae) (Figure 3G). In contrast, LCA treatment significantly enriched Verrucomicrobia, especially the genus Akkermansia, along with select Clostridium and Lachnospiraceae taxa relative to the TM group (Figure 3H). Interestingly, the taxon Akkermansia, along with its constituent species A. muciniphila, was also identified as characteristic of the NC group microbiota.

Moreover, targeted quantification of SCFAs revealed that total SCFAs, acetic acid, propionic acid and Valeric acid were significantly reduced in diabetic mice, whereas LCA supplementation markedly increased the levels of total SCFAs and acetate acid (Figure 3I–M). Although butyrate did not reach statistical significance, a rising trend was observed (Figure 3L). These results suggest that LCA mitigates T2DM-associated dysbiosis and restores beneficial microbial metabolites, which may contribute to improved host metabolic outcomes.

To assess the effect of LCA on bile acid metabolism in diabetic mice, hepatic bile acid profiling was performed. OPLS-DA plots showed a clear separation between NC, TM, and LCA groups, while stacking plots showed changes in multiple bile acid levels after LCA treatment (Figure 4A,B). Eight bile acids were elevated, and seven other bile acids were decreased in LCA-treated mice compared with the TM group. Notably, multiple bile acids exhibited significant alterations following exposure to LCA. Elevated species included murine deoxycholic acid (MDCA), LCA and glycyrlurea deoxycholic acid (GUDCA). In contrast, significant reductions in 12-oxo chenodeoxycholic acid (12 oxo CDCA) and ωMouse cholic acid (ωMCA), as well as several other bile acid derivatives, were observed (Figure 4C,D). Spearman’s analysis of blood glucose indices and differential bile acids revealed a negative correlation between elevated bile acids, including LCA, MDCA and GUDCA and blood glucose indices such as FBG and HOMA-IR. This suggests that these bile acids may play a role in maintaining blood glucose homeostasis (Figure 4E).

Further classification of the bile acid profile showed that treatment with LCA increased the proportion of conjugated species, especially secondary conjugated bile acids (Figure 4F–H). Conjugated secondary bile acid levels were significantly increased in the LCA group (Figure 4I), indicating changes in bile acid metabolism and conjugation patterns. Taken together, these findings suggest that LCA’s improvement in glycemic control may be related to changes in the composition and structure of bile acids.

To investigate the relationship between gut microbiota and bile acid metabolism in T2DM mice, we first performed LEfSe analysis to identify differentially abundant taxa at the genus and species levels between the TM and LCA groups (Figure 3H). Subsequently, based on the differential bile acid profiles between the two groups (Figure 4C), Spearman’s rank correlation analysis was conducted to assess the associations between these differentially abundant taxa and bile acid metabolites (Figure 5A). Significant positive or negative correlations were observed between bacterial genera such as Bacteroides, Akkermansia and Clostridium, and bile acid metabolites such as DCA, MDCA and LCA. Notably, Akkermansia exhibited a strong positive correlation with several bile acids, particularly DCA, 3β-HDCA, 23-norDCA, and GUDCA (Spearman’s ρ > 0.6, p < 0.05).

The regulatory influence of LCA on bile acid receptors was investigated by measuring the levels of Tgr5 and Fxr mRNA in the ileal tissue of mice. Compared to the TM group, Tgr5 expression increased significantly, while Fxr expression decreased notably in the LCA-treated group (Figure 5B). Moreover, LCA could significantly increase the level of occludin at the gene level (Figure 5C). Lipopolysaccharide binding protein (LBP) levels were significantly decreased in TM group and decreased after LCA treatment, which was close to the control level (Figure 5D). Furthermore, the activities of GUS and BSH enzymes were significantly lower in the feces of LCA group mice than in those of the model group (Figure 5E,F).

Overall, LCA’s improvement of metabolic homeostasis may be related to modulation of the intestinal microbe-bile acid network, as well as Tgr5 activation and Fxr inhibition.

To assess the direct growth-promoting effect of LCA on A. muciniphila, gradient concentrations of LCA were co-cultured with Akk, and OD600 was measured every 6 h. The results indicated that LCA treatment caused an initial increase in Akk growth followed by stabilization, with the peak OD600 reached at 12 h for most concentrations (Figure 6A). Therefore, 12 h was selected as the optimal time point for analysis. At this time point, 0.5 μM LCA was determined to be the most effective concentration for promoting Akk growth (Figure 6B).

To investigate the influence of LCA on A. muciniphila within a complex microbial environment, an in vitro fermentation model was constructed using fecal samples from a healthy elderly donor. PCoA analysis revealed that the 0.5 μM LCA treatment group and the untreated group were partially distinct on PCoA1, which accounted for 95.39% of the variation. This suggests that this concentration of LCA could alter the composition of microbes (Figure 6C). Treatment with 0.5 μM LCA significantly altered the abundance of specific bacteria, increasing the abundance of A. muciniphila (p = 0.042) and S. vestibularis (p = 0.021). Conversely, Limosilactobacillus mucosae (p = 0.026) and Veillonella atypica (p = 0.002) decreased in abundance. These results show that LCA can selectively promote the growth of certain species, such as A. muciniphila, thereby reshaping the complex gut microbiota (Figure 6D). Analysis of bile acid targeting showed that exposure to 0.5 μM LCA significantly altered the composition of bile acids, resulting in a notable decrease in oxygen-derived bile acids, including 3α-Hydroxy-6-oxo-5β-cholan-24-oic acid (6-ketoLCA), 3-oxochenodeoxycholic acid (3-oxoDCA) and Chenodeoxycholic acid 3-sulfate disodium salt (CDCA-3S) (Figure 6E,F). These results suggest that LCA increases the abundance of A. muciniphila in both pure cultures and complex communities, while also significantly altering the bile acid profile in vitro.