Abstract

   Conjugated lithocholic acid (LCA) plays a critical role in the
   development of metabolic dysfunction‐associated steatotic liver disease
   (MASLD). In this process, hepatocyte inflammation‐caused upregulation
   of its receptor, Takeda G protein‐coupled receptor 5 (TGR5) is a
   crucial factor. Serum bile acid profiling shows an increase in
   conjugated LCA, which correlates with disease severity. Depletion of
   Gpbar1 in hepatocytes significantly protects against the progression
   from MASLD to metabolic dysfunction‐associated steatohepatitis (MASH)
   that is related to conjugated LCA. In vivo and in vitro experiments
   indicate that TGR5 activation in hepatocytes promotes
   lipotoxicity‐induced cell death and inflammation by suppressing de novo
   carnitine biosynthesis. Mechanistically, TGR5 binding to CD36
   facilitates E3 ubiquitin ligase TRIM21 recruitment, leading to the
   degradation of BBOX1, a crucial enzyme in de novo carnitine
   biosynthesis. Targeting TGR5 therapeutically can restore carnitine
   biosynthesis, which may offer a potent strategy to prevent or reverse
   the transition from MASLD to MASH.

   Keywords: carnitine, gamma‐butyrobetaine hydroxylase 1, Gpbar1, MASH,
   secondary bile acids, TRIM21
     __________________________________________________________________

   Takeda G protein‐coupled receptor 5 (TGR5) serves as a central
   regulator in the metabolic dysfunction‐associated steatotic liver
   disease (MASLD)/metabolic dysfunction‐associated steatohepatitis (MASH)
   transition via conjugated lithocholic acid (LCA). Mechanistically, TGR5
   orchestrates lipotoxicity‐induced cell death through its interaction
   with CD36, which facilitates TRIM21 recruitment and BBOX1 degradation,
   ultimately disrupting de novo carnitine biosynthesis. Importantly,
   Targeting hepatic TGR5 prevents transition of MASLD/MASH.

   graphic file with name ADVS-12-2410602-g002.jpg

1. Introduction

   Metabolic dysfunction‐associated steatotic liver disease (MASLD)
   encompasses a spectrum ranging from simple steatotic liver disease
   (SLD) to metabolic dysfunction‐associated steatohepatitis (MASH). It
   may involve fatty liver disease at various stages, potentially
   accompanied by hepatic fibrosis, cirrhosis, liver failure, and advanced
   hepatocellular carcinoma (HCC).^[ [44]^1 , [45]^2 , [46]^3 ^] The key
   distinction among these stages primarily depends on the presence of
   significant liver inflammation. In the initial stage, impaired lipid
   processing capability of the liver is the dominant pathological
   feature, which leads to substantial lipid accumulation in hepatocytes.
   Under this condition, hepatocytes may resist lipotoxic effects, and
   cell death remains relatively rare. However, as the disease progresses,
   various factors (including secondary insults) may overwhelm the ability
   of hepatocytes to manage lipid toxicity, thereby increasing
   immune‐related cell mortality and triggering inflammatory responses.^[
   [47]^4 , [48]^5 ^] Inflammation may lead to progressive liver fibrosis
   and even cirrhosis in extreme cases, ultimately resulting in
   irreversible conditions (such as hypohepatia and liver cancer).^[
   [49]^6 ^] Therefore, prevention of the transition from steatosis to
   MASH has become a primary focus in MASLD research.

   Gastrointestinal (GI) metabolites play a crucial role in the transition
   from MASLD to MASH.^[ [50]^7 ^] These metabolites are predominantly
   produced by the gut microbiota and can affect hepatic metabolic
   pathways and immune responses, thus exacerbating liver inflammation and
   promoting the progression to more severe liver conditions.^[ [51]^8 ,
   [52]^9 ^] An urgent but unsolved question is why GI metabolites that do
   not induce inflammation in healthy livers can promote the transition
   from MASLD to MASH. Our preliminary research has discovered that CD36,
   a free fatty acid (FFA) receptor on hepatocytes, can interact with the
   upregulated GI metabolite receptor TGR5 in an inflammatory environment.
   This interaction highlighted the critical role of receptor dynamics in
   the liver's response to GI metabolic signals.

   Bile acids mediate their physiological and pathological effects through
   major cellular receptors, primarily the farnesoid X receptor (FXR) and
   TGR5.^[ [53]^10 ^] Compared with FXR, TGR5 exhibits a stronger response
   to GI metabolites, especially secondary bile acids (SBAs). TGR5
   displays a more pronounced response to SBAs, particularly lithocholic
   acid (LCA) and its derivatives.^[ [54]^11 ^] There is a certain
   concentration of SBAs in healthy livers, but TGR5 is poorly expressed
   in healthy hepatocytes.^[ [55]^12 ^] The limited expression of TGR5 may
   help maintain the low reactivity of a healthy liver to GI metabolites.
   This suggests a protective mechanism against excessive inflammatory and
   metabolic responses under normal physiological conditions.

2. Results

2.1. TGR5 Reduces CD36‐Mediated Lipid Metabolism in Hepatocytes in an
Inflammatory Microenvironment

   The impact of inflammation on hepatocyte fatty acid metabolism was
   elucidated using a 3D culture model. Hepatocytes were isolated,
   co‐cultured with macrophages at a 1:1 ratio, and exposed to oleic acid
   (OA) and palmitic acid (PA) (Figure [56]1A). As indicated by the
   results, hepatocytes co‐cultured with macrophages showed increased
   lipid droplet accumulation under exposure to OA+PA (Figure [57]1B),
   which suggested that macrophages inhibited hepatocyte lipid metabolism
   in an inflammatory environment. This may illustrate the
   immune‐metabolic interplay in hepatic tissues.

Figure 1.

   Figure 1
   [58]Open in a new tab

   TGR5 reduces CD36‐mediated lipid metabolism in hepatocytes within an
   inflammatory environment. A) Schematic representation of the
   experimental setup for co‐culturing hepatocytes with macrophages and
   treating with oleic acid and palmitic acid (OA+PA). The pull‐down assay
   was performed to isolate CD36 binding proteins, followed by LC/MS
   analysis to identify the protein interactions. B) Comparison of the
   effects of different treatment conditions on lipid accumulation in
   hepatocytes using Oil Red O staining. Statistical significance was
   determined using on one‐way ANOVA test with Dunnett adjustment, each
   group n = 6. C) LC/MS quantitative analysis of proteins (in peptides).
   D) Tandem mass spectrometry of TGR5. E) Co‐immunoprecipitation (Co‐IP)
   assay confirming the interaction between CD36 and TGR5 in hepatocytes.
   F) Measurements of different serum bile acids in MASLD patients with
   different liver fibrosis stages. G) Comparing the serum concentration
   of different bile acids in patients with different MASLD stages (HC:
   healthy control, MASLD: metabolic associated steatohepatitis liver
   disease, MASH: metabolic associated steatohepatitis). Statistical
   significance was determined using one‐way ANOVA test with Dunnett
   adjustment. Data are presented as mean ± SD. All experiments were
   performed in triplicate. **p < 0.01, ***p < 0.0002, ****p < 0.0001, ns:
   p > 0.05.

   Given the critical role of CD36 in hepatocyte FFA metabolism,^[ [59]^13
   ^] this research investigated how inflammatory conditions might alter
   protein interactions with CD36. Specifically, CD36‐specific antibodies
   were used for immunoprecipitation to isolate CD36‐binding proteins
   under various conditions (Figure [60]1A). According to liquid
   chromatography‐mass spectrometry (LC/MS) results, TGR5 was identified
   as the most significantly upregulated protein in these complexes
   (Figure [61]1C,D). Co‐immunoprecipitation (Co‐IP) results confirmed the
   interaction between TGR5 and CD36. Notably, TGR5 was absent in
   hepatocytes under non‐inflammatory conditions. However, TGR5 was
   remarkably upregulated and bound to CD36 after the introduction of OA
   and PA in hepatocyte‐macrophage co‐cultures (Figure [62]1E). The above
   findings suggested that TGR5 played a responsive role under lipid
   stress, indicating a dynamic regulatory mechanism in hepatic lipid
   metabolism and inflammatory response.

   TGR5 is encoded by the Gpbar1 gene and functions as a receptor for bile
   acids, especially GI metabolites SBAs.^[ [63]^14 ^] Consequently, the
   serum bile acids profiles, specifically LCA and its derivatives in
   various stages of liver fibrosis (F0–F4) in MASLD patients (n = 550)
   were investigated. It was found that as liver fibrosis or MASLD
   pathology progresses, the levels of LCA derivatives were notably
   increased, rather than LCA itself (Figure [64]1F,G).

2.2. TGR5 Expression Increases during the Transition from MASLD to MASH in an
NF‐κB Signaling‐Dependent Manner

   TGR5 expression dynamics in MASLD were investigated. In short, human
   liver samples were obtained and MASLD mouse models were constructed
   using high‐fat and high‐fructose (HFHfr) and Western diets (WD). Based
   on morphological assessments and serum hepatic enzyme levels, these
   samples were categorized into normal, MASLD, and MASH groups.
   Quantitative real‐time PCR (qRT‐PCR) and immunostaining were used to
   detect TGR5 expression in liver tissues at various MASLD stages. As
   revealed by the results, Gpbar1 gene expression and its encoded protein
   TGR5 were significantly upregulated at the MASH stage (Figure
   [65]S1A–C, Supporting Information). This suggested that increased TGR5
   expression may be associated with the progression of MASLD to more
   severe stages.

   Subsequently, the factors contributing to the transcriptional
   upregulation of the Gpbar1 gene in hepatocytes were elucidated. The
   in‐vitro experiments were conducted to simulate three conditions that
   may influence hepatocytes in MASH, including supernatant from
   lipopolysaccharide (LPS)‐stimulated macrophages (SUP.), high‐fat
   (OA+PA), and high glucose (H.G.). The outcomes indicated that SUP.
   treatment notably elevated the transcription and protein levels of
   Gpbar1 (Figure [66]S1D,E, Supporting Information). In addition, Gpbar1
   expression in a transwell co‐culture system with macrophages and
   hepatocytes was determined. It was found that LPS‐stimulated macrophage
   secretion significantly enhanced Gpbar1 gene expression, particularly
   at higher macrophage‐to‐hepatocyte ratios, suggesting a dose‐dependent
   relationship (Figure [67]S1F, Supporting Information). Next,
   hepatocytes were exposed to TNFα, interleukin (IL)‐6, and IL‐1β to
   identify specific responsible macrophage cytokines. According to the
   results, only TNFα remarkably increased Gpbar1 expression (Figure
   [68]S1G, Supporting Information). Notably, TGR5 expression in
   hepatocytes was progressively increased when TNFα concentration
   exceeded 5 ng/mL (Figure [69]S1H, Supporting Information). Furthermore,
   it was demonstrated that TNFα upregulated Gpbar1 expression in
   hepatocytes via the NF‐κB signaling pathway (Figure [70]S1I,J,
   Supporting Information). Taken together, the above results proposed a
   direct mechanistic link between TNFα signaling and the modulation of
   Gpbar1 gene expression in hepatocytes.

   Subsequently, the relationship between TGR5 and TNFα was analyzed in
   vivo in both human and mouse samples. Multiplex immunohistochemistry
   (mIHC) staining was performed to assess the presence of TGR5, TNFα, and
   CD68‐positive macrophages in MASLD tissues. The results revealed a
   significantly positive correlation between TGR5 and TNFα levels. In
   addition, there was a positive correlation between TGR5 expression and
   macrophage numbers in liver tissues (Figure [71]S2A,B, Supporting
   Information). Further investigations revealed that higher TGR5
   expression in hepatocytes was markedly associated with elevated levels
   of liver function markers (alanine aminotransferase [ALT], aspartate
   aminotransferase [AST], and alkaline phosphatase [ALP]) in MASLD
   patients (Table [72]1 ). These findings suggested that TGR5 was a
   marker of inflammatory and metabolic disturbances in liver tissues and
   correlated with liver dysfunction severity in MASLD.

Table 1.

   Baseline characteristics of the clinical samples, according to the
   expression of TGR5[73] ^a ^)
Variable                     TGR5^high ) (N = 44) TGR5^low ) (N = 45)   Total
Demographic characteristics
Age (years)                         56 ± 6              54 ± 8          56 ± 8
Sex‐no. (%)
Male                               20 (45)              23 (51)        43 (48)
Female                             24 (55)              22 (49)        46 (52)
Body mass index[74] ^b ^)
Distribution‐no. (%)
<25                                 2 (5)                1 (2)          3 (3)
25 to <30                           7 (16)              15 (33)        22 (25)
30 to <35                          13 (29)              17 (38)        30 (34)
≥35                                22 (50)              12 (27)        34 (38)
Histologic features
NAFLD activity score[75] ^c)      5.2 ± 1.7            4.0 ± 1.5      4.2 ± 1.6
Steatosis score                   1.6 ± 0.8            1.2 ± 0.6      1.5 ± 0.9
Lobular inflammation score        1.9 ± 0.7            1.4 ± 0.4      1.6 ± 0.8
Ballooning score                  1.7 ± 0.6            0.8 ± 0.5      1.2 ± 0.6
Fibrosis stage[76] ^d ^)          4.0 ± 0.0            0.9 ± 0.7      1.5 ± 1.2
MASH diagnosis‐no. (%)
Definite or borderline MASH        41 (93)              27 (60)        68 (76)
MASLD, not MASH                     3 (7)               18 (40)        21 (24)
Liver tests
Median ALT(IQR) (U L^−1)          40 (26–57)          48 (30–74)      52 (31–80)
Median AST(IQR) (U L^−1)          45 (32–64)          33 (25–52)      36 (29–59)
Median ALP(IQR) (U L^−1)         89 (73–114)          70 (56–87)      74 (62–92)
   [77]Open in a new tab
   ^^a)

   ± values are means ± SD. Percentages may not total 100 because of
   rounding. IQR interquartile range, MASLD metabolic
   dysfunction‐associated steatotic liver disease, and MASH metabolic
   dysfunction‐associated steatohepatitis
   ^^b)

   Body mass index is the weight in kilograms divided by the square of the
   height in meters. A body mass index of less than 25 indicates normal
   weight, of 25 to less than 30 overweight, 30 to less than 35 obese, and
   35 or above morbid obesity.
   ^^c)

   NAFLD activity score was assessed on a scale of 0–8, gauges disease
   severity through three criteria: steatosis (0–3 scale), lobular
   inflammation (0–3), and hepatocellular ballooning (0–2).
   ^^d)

   Fibrosis stage was assessed on a scale of 0–4, with higher scores
   indicating more severe fibrosis.

2.3. Hepatocyte TGR5 Exacerbates Conjugated LCA‐Related MASLD Model

   Next, the impact of hepatic TGR5 expression in hepatocytes on MASLD was
   comprehensively investigated. Specifically, Gpbar1 ^fl/fl mice were
   generated and crossed with Alb‐Cre mice for selective depletion of TGR5
   expression in hepatocytes. Clinical bile acid profiling data have
   indicated that the natural TGR5 ligand glycine/taurine conjugated
   lithocholic acid (G/TLCA) significantly increases as MASLD progresses.
   Therefore, it is hypothesized that TGR5 may exacerbate hepatocyte
   damage. To verify this hypothesis, a series of experiments were
   conducted using two HFHfr mouse MASLD models: HFHfr + Clostridium
   difficile (C. diff) fecal microbiota transplantation (FMT) [treated
   with C. diff transplantation] and HFHfr + LCA (treated with a high LCA
   diet) (Figure [78]2A). According to the results, both treatments
   remarkably exacerbated the MASH phenotype, with high LCA diet treatment
   showing a more pronounced effect. Interestingly, mice with
   hepatocyte‐specific depletion of Gpbar1 did not show significant MASH
   progression, even with treatments of C. diff transplantation and a high
   LCA diet (Figure [79]2B,C and Figure [80]S3A, Supporting Information).
   This suggested that TGR5 played a crucial role in mediating the liver's
   response to conjugated LCA, which can affect MASH development.

Figure 2.

   Figure 2
   [81]Open in a new tab

   Hepatic TGR5 exacerbates conjugated LCA induced hepatic inflammation,
   fibrosis in in vivo mice models. A) A schematic representation of the
   construction of the two mouse MASH models. B,C) Comparison of H&E, Oil
   Red O, Masson trichrome, F4/80 immunofluorescence staining, and NAS
   scores in Gpbar1 ^fl/fl and Gpbar1 ^fl/fl×Alb‐Cre mice at 16 weeks of
   the two conjugated LCA induced MASH model, each group n = 12. D)
   Comparison of H&E, Oil Red O, Masson trichrome, F4/80
   immunofluorescence staining, and NAS scores in Gpbar1 ^fl/fl and Gpbar1
   ^fl/fl×Alb‐Cre mice at 20 weeks of HFHfr and WD diet MASH model, each
   group n = 18. E) Comparison of H&E, Oil Red O, Masson trichrome, F4/80
   immunofluorescence staining, and NAS scores in Rosa26‐LSL‐Cas9 mice,
   Gpbar1 ^fl/fl×Lyz‐Cre cross with Rosa26‐LSL‐Cas9 mice with different
   treatment indicated in the figure until 16 weeks of HFHfr and
   HFHfr+FMT:C. diff mouse MASH model, each group n = 18. Data are
   presented as mean ± SD. All experiments were performed in triplicate.
   **p < 0.01, ***p < 0.0002, ****p < 0.0001, ns: p > 0.05.

   Next, the HFHfr and WD models at 8, 16, and 20 weeks were established
   to compare MASLD phenotypes at different time points. Similar to FMT
   and LCA diet mouse MASLD models, Gpbar1 knockdown in hepatocytes
   noticeably inhibited MASLD progression in both models, as manifested as
   reduced hepatocyte vacuolization by H&E staining, decreased Oil Red
   O‐stained cells, attenuated hepatic fibrosis by Masson's trichrome
   staining, and alleviated hepatic inflammation by F4/80 staining,
   compared to Gpbar1 ^fl/fl mice (Figure [82]2D and Figure [83]S3B,
   Supporting Information). These differences were particularly prominent
   at 20 weeks (Figure [84]2D). Despite significant liver inflammation and
   fibrosis in Gpbar1 ^fl/fl mice, hepatocyte‐specific Gpbar1 knockout
   mice did not show these severe symptoms. In addition,
   hepatocyte‐specific Gpbar1 knockout not only reduced mouse body weight
   but also improved serum markers. As indicated by serological analysis
   results, Gpbar1 knockout mice exhibited markedly decreased ALT, TNFα,
   MCP1, TG, and TC levels compared to the control group (Figure
   [85]S3C,D, Supporting Information).

   Previous studies have shown that TGR5 is a protective factor in MASLD
   development in the liver, as TGR5‐knockout mice exhibit a more
   pronounced MASLD phenotype. Further research has indicated that TGR5
   exerts an anti‐inflammatory effect in hepatic macrophages.^[ [86]^15 ,
   [87]^16 ^] Hence, we then shifted to comparing the roles of macrophage
   and hepatic TGR5 in the progression of MASLD to MASH. In short, Gpbar1
   ^fl/fl mice were specifically knocked out of the Gpbar1 gene in myeloid
   cells using Lyz‐Cre mice, with Gpbar1 ^fl/fl mice as controls. The
   results in HFHfr and WD models demonstrated that TGR5 in myeloid cells
   can protect against MASLD onset and progression (Figure [88]S3E,
   Supporting Information).

   Furthermore, Lyz‐Cre‐Gpbar1fl/fl mice were crossed with Rosa26‐LSL‐Cas9
   mice. At the 12th week of the HFHfr and C. diff FMT models, AAV8
   carrying Alb‐Cre‐Gpbar1 sgRNA was injected via the spleen for TGR5
   knockout in hepatocytes. At the 16th week, it was observed that mice
   with hepatocyte TGR5 knockout showed significantly milder MASH
   phenotype compared to those with TGR5 knocked out in myeloid cells
   (Figure [89]2E). These results suggested that hepatocyte TGR5 played a
   dominant role in MASH progression.

2.4. TGR5 Facilitates Hepatocyte Death by Suppressing Hepatocyte Carnitine
Synthesis

   The above findings have pinpointed TGR5 as a pivotal factor
   exacerbating lipotoxicity and hepatocyte death within an in‐vivo MASLD
   model. To delve deeper into the underlying action mechanisms, we
   extended our research involving experiments using primary hepatocytes.

   The impact of various factors on hepatocyte FFA processing was assessed
   using OA+PA treatment and Oil Red O staining. As revealed by the
   results, TNFα did not notably promote lipid accumulation, while OA+PA
   remarkably increased Oil Red O‐positive areas in wild‐type hepatocytes,
   not in Gpbar1 knockout hepatocytes. TGR5 agonists (G/TLCA and INT‐777)
   further intensified FFA accumulation in AML12 cells (Figure [90]S4A,
   Supporting Information). TNFα combined with TGR5 agonists elevated TG,
   TC, Il6, and Il8 mRNA levels, particularly in wild‐type hepatocytes
   (Figure [91]S4B, Supporting Information). As indicated by flow
   cytometry results, TNFα and TGR5 agonists significantly increased cell
   death in wild‐type hepatocytes, but not in Gpbar1 knockout hepatocytes
   (Figure [92]S4C, Supporting Information).

   RNA‐seq and metabolomics analyses on cells and mouse liver tissues were
   conducted to elucidate the potential mechanisms of hepatocytes in
   promoting MASLD progression. According to RNA‐seq analysis results of
   MASLD tissues from cells and LCA‐fed mice, differentially expressed
   genes were enriched in KEGG pathways related to cell death, fatty acid,
   and lipid metabolism (Figure [93]S4D–F, Supporting Information).
   Notably, metabolomics assays demonstrated that wild‐type hepatocytes or
   Gpabr1fl/fl mouse liver tissues showed significantly decreased levels
   of carnitine metabolites compared to Gpbar1 knockout hepatocytes
   treated with TNFα+OA+PA+INT‐777 or Gpbar1 hepatocyte‐specific knock out
   MASLD mice treated with LCA diet (Figure [94]3A,B, Supporting
   Information). We confirmed L‐carnitine concentration in cell lines,
   mouse MASH tissues, and human MASH samples (Figure [95]3C and Figure
   [96]S4G, Supporting Information).

Figure 3.

   Figure 3
   [97]Open in a new tab

   Hepatic TGR5 exacerbates conjugated LCA induced MASH by decreasing de
   novo carnitine synthesis. A) Heatmap of metabolomic profiling data
   showing differences in metabolite levels between Gpbar1 ^fl/fl and
   Gpbar1 ^fl/fl×Alb‐Cre mice in LCA+HFHfr model at 16 weeks. B) Volcano
   plot comparing metabolites levels in hepatocyte isolated from Gpbar1
   ^fl/fl versus Gpbar1 ^fl/fl×Alb‐Cre mice and treated with 50 × 10^−6 m
   G/TLCA. C) l‐carnitine concentration in hepatocytes treated
   differently, serum l‐carnitine levels in human liver samples (Normal n
   = 23, MASLD n = 39, MASH n = 24) and mouse (Gpbar1fl/fl vs Gpbar1
   ^fl/fl×Alb‐Cre, each group n = 6) liver samples in two MASH model.
   Statistical significance was determined using one‐way ANOVA test with
   Dunnett adjustment. D) Representative figures and qualification of Oil
   Red O staining for mouse primary hepatocyte treated differently.
   Statistical significance was determined using two‐way ANOVA test with
   Dunnett adjustment, each group n = 6. E,F) Comparisons of TG, TC
   levels, and the transcription levels of the inflammatory cytokines Il6
   and Il8 in primary mouse hepatocytes after treatment as shown in the
   figure. Statistical significance was determined using two‐way ANOVA
   test with Dunnett adjustment, each group n = 6. G,H) Comparison of Oil
   Red O, Masson's trichrome, and F4/80 immunofluorescence staining and
   NAS score in liver tissues obtained from two mouse MASH models treated
   as shown in the figure. Statistical significance of (H) and (I) was
   determined using one‐way ANOVA test with Šidák adjustment, each group n
   = 6. Data are presented as mean ± SD. All experiments were performed in
   triplicate. *p < 0.05, **p < 0.01, ***p < 0.0002, ****p < 0.0001, ns: p
   > 0.05.

   Subsequently, the role of reduced carnitine levels in the death of
   wild‐type hepatocytes was investigated following TNFα‐bound TGR5
   agonist treatment. Specifically, 50 × 10^−6 m l‐carnitine was added to
   wild‐type hepatocytes in vitro. This treatment reduced Oil Red O
   staining areas, cell death, intracellular TG and TC levels, and Il6 and
   Il8 mRNA levels (Figure [98]3D–F). In vivo, 0.28% l‐carnitine
   supplementation in wild‐type mouse from week 12 to 20 in HFHfr+C. diff
   FMT and HFHfr+LCA models effectively reduced nonalcoholic fatty liver
   disease activity score (NAS), Oil‐Red O staining areas, hepatic
   fibrosis (indicated by Masson's trichrome staining), cell death rate,
   and F4/80‐positive cells (Figure [99]3G,H).

2.5. TGR5 Decreases Hepatocyte De Novo Carnitine Biosynthesis by Degrading
Gamma‐Butyrobetaine Hydroxylase 1 (BBOX1) Proteins

   Research has indicated that TGR5 facilitates the transition from MASLD
   to MASH primarily by downregulating carnitine biosynthesis. This
   process involves the capacity of hepatocytes to synthesize carnitine
   from amino acids and uptake exogenous carnitine via the OCTN2
   transporter. To investigate this mechanism, we evaluated the expression
   of relevant metabolic enzymes and transporters in cells and animals. As
   revealed by qRT‐PCR analysis results, the transcript levels of genes
   involved in carnitine synthesis (Bbox1, Tmlhe, Aldh9a1, and Shmt2) and
   carnitine transport (Slc22a5) remained largely unchanged with the
   combined action of TNFα, OA+PA, and the TGR5 agonist, except for Bbox1
   (Figures [100]4A and [101]S5A, Supporting Information). According to
   Western blot (WB) analysis results, OCTN2 (protein name of Slc22a5) and
   SHMT2 protein levels did not significantly change; however, BBOX1 level
   was almost completely degraded 24 h after treatment with TNFα, OA+PA,
   and the TGR5 agonist (Figure [102]4B). Furthermore, mIHC staining was
   used to evaluate the expression levels of BBOX1 and OCTN2. Notably,
   BBOX1 underwent significant degradation in models involving FMT with C.
   difficile and a high‐LCA diet, whereas OCTN2 remained unaffected.
   However, there were no significant changes in mice with
   hepatocyte‐specific depletion of Gpbar1 (Figure [103]4C). Meanwhile, a
   significant decrease in BBOX1 protein level was observed at 20 weeks in
   both models, while OCTN2 expression level remained relatively stable
   over time (Figure [104]S5B,C, Supporting Information). Moreover,
   similar results were observed in human MASLD tissues, where BBOX1
   protein level was remarkably reduced in MASH patients’ tissues, while
   OCTN2 expression level remained relatively stable (Figure [105]4D). As
   revealed by the correlation analysis results, TGR5 expression was
   markedly associated with BBOX1 protein levels in human MASLD tissues
   (Figure [106]4E).

Figure 4.

   Figure 4
   [107]Open in a new tab

   Activation of hepatic TGR5 decrease de novo carnitine biosynthesis by
   degrading BBOX1 proteins. A) Relative mRNA expression of Bbox1 in
   hepatocytes with different treatment. Statistical significance was
   determined using one‐way ANOVA test with Dunnett adjustment, each group
   n = 6. B) WB monitors the protein expression of CD36, OCTN2, TGRR5,
   BBOX1, and SHMT2 in hepatocytes after the treatment shown in the
   figure, with β‐actin as the internal control. C) mIHC staining and mean
   fluorescence intensity (MFI) qualification of TGR5, BBOX, and OCTN2
   expression in liver sections from Gpbar1 ^fl/fl and Gpbar1
   ^fl/fl×Alb‐Cre mice treated with two MSAH models. Statistical
   significance was determined using two‐way ANOVA test with Šidák
   adjustment, each group n = 6. D) mIHC staining and MFI qualification of
   TGR5, BBOX1, OCTN2 expression in human samples with normal liver n =
   23, MASLD n = 39, and MASH n = 24. E) Correlation analysis of TGR5 with
   OCTN2 or BBOX1 expression in human liver samples. F) The l‐carnitine
   concentration in AML12 cells under different treatments, each group n =
   6. G–I) Comparison of cell death rate, Oil Red O staining positivity,
   and intracellular TG and TC levels in AML12 cells under different
   transfections and treatments. Statistical significance was determined
   using two‐way ANOVA test with Dunnett adjustment, each group n = 6.
   Data are presented as mean ± SD. All experiments were performed in
   triplicate. **p < 0.01, ***p < 0.0002, ****p < 0.0001, ns: p > 0.5.

   The above results from both in vivo and in vitro studies indicated that
   the BBOX1 protein level was remarkably decreased during MASLD
   progression, despite stable mRNA expression. To confirm this, a
   doxycycline (DOX)‐inducible system with the p‐TRE‐Tight‐BI plasmid was
   used, enabling controlled Bbox1 overexpression, which can be suppressed
   by DOX. After DOX treatment, l‐carnitine levels were measured at 6 and
   24 h. It was observed that l‐carnitine level returned to baseline after
   24 h, similar to that in empty vector control cells in the presence of
   TNFα, OA+PA, and G/TLCA (Figure [108]4F). In addition, hepatocytes
   overexpressing BBOX1 showed similar outcomes to empty vector controls
   in terms of cell death rate, Oil Red O staining positivity,
   intracellular TG and TC levels, and Il6 and Il8 transcription after 24
   h of DOX induction (Figure [109]4H,I and Figure [110]S5D, Supporting
   Information). Notably, hepatocytes treated only with OA+PA and
   overexpressing BBOX1 showed notably less recovery compared to groups
   where TGR5 was activated by G/TLCA (Figure [111]S5E, Supporting
   Information).

2.6. Hepatocyte TGR5‐CD36 Interaction Facilitates TRIM21 Recruitment and
Downregulation of BBOX1

   Subsequently, the mechanism of BBOX1 protein degradation in response to
   TGR5 activation was investigated. Specifically, hepatocytes were
   treated with autophagy inhibitors (3‐methyladenine [3‐MA] and
   chloroquine [CQ]) and proteasome inhibitors (MG‐132 and Carfilzomib).
   The results indicated that only the proteasome inhibitors effectively
   blocked BBOX1 protein degradation (Figure [112]5A). According to
   ubiquitination level detection results, treatment with
   TNFα+OA+PA+G/TLCA notably increased BBOX1 protein ubiquitination,
   especially with the TGR5 agonist G/TLCA (Figure [113]5B). Next, the E3
   ubiquitin ligase responsible for BBOX1 degradation was detected.
   Pull‐down studies were conducted using a BBOX1 antibody under 3
   conditions: control, TNFα+OA+PA, and TNFα+OA+PA+G/TLCA. According to
   LC/MS and GO enrichment analysis results, the identified proteins were
   enriched in ubiquitination pathways (Figure [114]5C,D and Figure
   [115]S6A, Supporting Information). Our analysis results revealed that
   the binding of the E3 ubiquitin ligase TRIM21 increased progressively
   with TNFα+OA+PA and G/TLCA treatment.

Figure 5.

   Figure 5
   [116]Open in a new tab

   TGR5‐CD36 interaction facilitates E3 ubiquitin ligase recruitment and
   downregulation of gamma‐butyrobetaine hydroxylase 1. A) Western blot
   analysis reveals the expression of BBOX1 in wild‐type hepatocytes at
   different time points upon treatment with autophagy inhibitors (3‐MA
   and CQ) and proteasome inhibitors (MG‐132 and Carfilzomib). B)
   Co‐immunoprecipitation (Co‐IP) and Western blot analysis of BBOX1 with
   ubiquitination, CD36, TGR5, and BBOX1 under treatment, with IgG as a
   negative control. C) The summarized table lists the top five proteins
   with the most significant differences identified in pull‐down LC/MS
   experiments using a BBOX1 antibody under different conditions. D)
   Tandem mass spectrometry of TRIM21. E) Co‐IP and WB verification show
   the interactions between TRIM21, CD36, TGR5, and BBOX1 in AML12 and
   AML12‐Gpbar1 shRNA hepatocytes under different treatment conditions.
   IgG serves as a negative control. F) Co‐localization of TGR5 with
   TRIM21 and CD36 under different treatments by using laser confocal
   microscopy. G) Co‐IP and WB analyses reveal BBOX1 and its mutants
   (D282E, K283N) overexpression, BBOX1 ubiquitination, and the
   interaction between BBOX1 and TRIM21 under different treatment
   conditions, with IgG serves as a negative control. H) mIHC is used to
   localize TRIM21 and BBOX1 proteins within liver sections from MASLD and
   MASH patients. Correlation analysis indicates the relationship between
   OD values of TRIM21 and BBOX1 in liver sections. Statistical
   significance of the correlations was determined using the Pearson
   correlation coefficient. Data are presented as mean ± SD. All
   experiments were performed in triplicate.

   Next, the interaction between BBOX1 protein and TRIM21 was validated
   through Co‐IP and mIHC staining experiments. As revealed by these
   experiment results, the interaction between BBOX1 and TRIM21 was
   enhanced with increasing TGR5 expression and activation
   (Figure [117]5E,F). Notably, in Gpbar1 or Cd36 knockdown hepatocytes,
   there was no binding between the BBOX1 protein and TRIM21
   (Figure [118]5E and Figure [119]S6B, Supporting Information). This
   indicated the crucial role of TGR5‐CD36 interaction for the binding of
   BBOX1 and TRIM21.

   Subsequently, the mechanism underlying TRIM21‐mediated ubiquitination
   and degradation of BBOX1 was elucidated. The interaction sites between
   the proteins were predicted using ROSETTA software. Our analysis
   results indicated that TRIM21 bound to BBOX1 at an aspartic acid
   residue at position 282 (Figure [120]S6C, Supporting Information). A
   potential E3 ubiquitination site at lysine 283^[ [121]^17 ^] was
   identified, which was adjacent to the aspartic acid at position 282. We
   hypothesized that BBOX1's interaction with TRIM21 at these residues may
   facilitate ubiquitination. For verification of this hypothesis, BBOX1
   mutants D282E and K283N were generated and Co‐IP experiments were
   conducted. The results showed that both mutants reduced BBOX1
   ubiquitination and degradation (Figure [122]5G,H, Supporting
   Information). AML12 overexpressing these mutants maintained lipid
   metabolism even with TNFα+OA+PA+G/TLCA treatment (Figure [123]S6D,
   Supporting Information) and exhibited reduced cell death compared to
   wild‐type cells (Figure [124]S6E, Supporting Information). In addition,
   TG and TC levels were markedly decreased (Figure [125]S6F, Supporting
   Information). mIHC analysis of human MASLD samples demonstrated a
   linear negative correlation between TRIM21 and BBOX1 expression at the
   MASH stage (Figure [126]5H). From the above compelling findings, it can
   concluded that TGR5, when activated in hepatocytes as a bridge docking
   CD36, promoted the E3 ubiquitination ligase TRIM21 to catalyze the
   ubiquitination of BBOX1, which in turn degraded BBOX1.

2.7. Inhibiting TGR5 Activity In Vitro Prevents and Reverses the Progression
of Conjugated LCA‐Induced Steatosis to MASH

   For investigating whether targeting hepatocyte TGR5 and its downstream
   pathways can halt the progression from MASLD to MASH, an in‐vitro model
   was established using 3D culture techniques, including two systems:
   primary hepatocytes alone and a co‐culture with macrophages. Two
   treatment scenarios were tested: inducing MASH with G/TLCA and using
   SBI‐115 (a TGR5 receptor inhibitor), either as a “prevention model”
   (administered simultaneously with G/TLCA) or a “therapeutic model”
   (introduced 48 hours after G/TLCA treatment) (Figure [127]6A).

Figure 6.

   Figure 6
   [128]Open in a new tab

   Inhibiting TGR5 activity has the potential to prevent and reverse the
   progression of MASLD to MASH in vitro. A) A schematic diagram
   illustrating how 3D culture system models conditions. B–E) Comparison
   of the Oil Red O staining positive area, TG and TC levels,
   transcription of Il6 and Il8 and cell death rate under different
   treatments in the in vitro MASH model, each group n = 6. F)
   Representative images showing the expression of TGR5, TRIM21, BBOX1,
   and RIP3 under different treatments in the in vitro MASH model using
   mIHC staining. G) Comparison of the MFI of TGR5, TRIM21, BBOX1, and
   RIP3 in the in vitro MASH model under different treatments. Statistical
   significance was determined using on one‐way ANOVA test with Šidák
   adjustment, each group n = 6. H) Heatmap showing metabolites levels
   under different treatments. Data are presented as mean ± SD. All
   experiments were performed in triplicate. *p < 0.05, **p < 0.01, ***p <
   0.0002, ****p < 0.0001, ns: p > 0.05.

   Initially, SBI‐115 did not enhance lipid metabolism in the isolated
   primary hepatocytes, as indicated by Oil Red O staining findings, as
   well as detection results of cell viability, and TG and TC levels.
   However, in hepatocytes co‐cultured with macrophages, SBI‐115
   significantly reduced Oil Red O staining positive area, improved cell
   viability, and decreased TG and TC levels in the prevention model. It
   also reduced the production of inflammatory factors (such as Il8 and
   Il6) within the 3D culture system. Remarkably, even with the occurrence
   of the transition from steatosis to MASH, SBI‐115 treatment can reverse
   this transition, exhibiting therapeutic effects similar to the
   prevention model (Figure [129]6B–E and Figure [130]S7A,B, Supporting
   Information). These findings suggested that targeting TGR5 and its
   downstream pathways with SBI‐115 could effectively prevent and treat
   the progression from simple fatty liver to steatohepatitis, even after
   the disease has developed.

   The therapeutic impact of SBI‐115 was further elucidated using mIHC
   staining and metabolomics. In the steatosis model, SBI‐115 treatment
   did not significantly alter TGR5 expression, BBOX1 protein level, or
   hepatocyte cell death rate (as indicated by RIP3 positivity). However,
   in both the therapeutic and prevention models, SBI‐115 treatment
   notably preserved BBOX1 protein level and reduced cell death rates.
   (Figure [131]6F,G and Figure [132]S7B,C, Supporting Information).
   Metabolomics analysis further confirmed these findings, showing that
   the inhibition of downstream TGR5 activity by SBI‐115 effectively
   restored carnitine metabolite levels in hepatocytes (Figure [133]6H).

2.8. Targeting TGR5 and Its Downstream Pathways Is Effective for Reversing
Conjugated LCA‐Induced MASLD to MASH Progression In Vivo

   Considering the promising results of in vitro inhibition of TGR5
   activity, its efficacy in vivo was evaluated. Currently, the specific
   inhibitors for hepatocyte‐targeted TGR5 have not been developed. To
   address this, Rosa26‐LSL‐Cas9 knock‐in mice were used with two mouse
   models of conjugated LCA‐induced MASLD and AAV8 virus‐mediated gene
   therapy was administered via the tail vein (Figure [134]7A). The roles
   of manipulating hepatocyte TGR5 and TRIM21, as well as overexpressing
   BBOX1 and other therapeutic targets, in the progression from steatosis
   to MASH were explored in this study, considering both “therapeutic” and
   “prevention” approaches (Figure [135]7A).

Figure 7.

   Figure 7
   [136]Open in a new tab

   In vivo therapies targeting hepatocyte TGR5 and downstream pathways
   effectively reverse the progression from LCA‐induced MASLD to MASH. A)
   Experimental design illustrating the therapeutic and prevention models
   for the progression from conjugated LCA induced steatosis to MASH. B)
   WB analysis of BBOX1, TGR5, and TRIM21, with β‐actin expression in
   hepatocytes of mice treated indicated in the figure. C) Representative
   images of H&E, Oil Red O, Masson trichrome, and F4/80
   immunofluorescence staining reflect the effects of prevention and
   therapeutic treatment on mice in different groups in the HFHfr+C. diff
   FMT mouse MASH model. D) Quantitative analysis of H&E, Oil Red O,
   Masson trichrome, and F4/80 immunofluorescence staining in the liver of
   mice in the HFHfr+C. diff FMT mouse MASH model with prevention and
   therapeutic treatment. Statistical significance was determined using on
   two‐way ANOVA test with Dunnett adjustment, each group n = 18. E)
   Representative images of H&E, Oil Red O, Masson trichrome, and F4/80
   immunofluorescence staining reflect the effects of prevention and
   therapeutic treatment on mice belong to different groups in the
   LCA+HFHfr mouse MASH model. F) Quantitative analysis of H&E, Oil Red O,
   Masson trichrome, and F4/80 immunofluorescence staining in the liver of
   mice in the LCA+HFHfr mouse MASH model with prevention and therapeutic
   treatment. Statistical significance was determined using two‐way ANOVA
   test with Dunnett adjustment, each group n = 18. G) GSEA analysis based
   on the KEGG dataset reflects the impact of treatment and preventive
   treatment on genes related to lipid metabolism, inflammation, and
   fibrosis in mice. H) The heatmap shows the levels of four groups of
   indicated metabolites in the therapeutic model. Data are presented as
   mean ± SD. All experiments were performed in triplicate. *p < 0.05, **p
   < 0.01, ***p < 0.0002, ****p < 0.0001.

   Initially, the accurate expression of target proteins in isolated
   hepatocytes in vitro was confirmed using WB (Figure [137]7B). Next, the
   effects of the “therapeutic” and “prevention” models in mouse MASH
   models were assessed. In both models, the three treatments were
   effective when administered either early (at 8 weeks) or starting from
   the 12th week. Compared to the control group, overexpression of BBOX1
   in hepatocytes or specific knockdown of Gpbar1 and Trim21 genes reduced
   liver vacuolation and the proportion of Oil Red O positive staining
   areas. Similarly, promising results were also observed in assessments
   of liver fibrosis by Masson's trichrome staining and liver inflammation
   by F4/80 staining (Figure [138]7C–F). Notably, simultaneous targeting
   of these three factors (particularly knockdown of Gpbar1 in
   hepatocytes) significantly reduced mouse body weight. Serological tests
   confirmed the effectiveness of these treatments, as evidenced by
   protected liver function and reduced serum levels of TG, TC, and
   inflammatory factors (MCP1 and TNFα) (Figure [139]S8A,B, Supporting
   Information).

   Moreover, RNA‐seq analysis was used to assess the therapeutic effects.
   The results revealed that hepatic lipid metabolism capacity was
   substantially enhanced, accompanied by reduced inflammation and
   fibrosis levels, and post‐targeted prevention or treatment reflected by
   GSEA analysis (Figure [140]7G). Metabolomics analysis results further
   supportably indicated that hepatic carnitine metabolite levels were
   markedly improved (Figure [141]7H).

3. Discussion

   It has been shown that persistent inflammation plays a pivotal role in
   the progression of steatosis to steatohepatitis.^[ [142]^18 ^]
   Inflammation arises from hepatocyte death, triggered by excessive
   accumulation of lipids, ultimately resulting in lipotoxicity and cell
   death.^[ [143]^19 ^] A previous study has revealed that hepatocyte
   lipid metabolism is crucial for counteracting lipotoxicity‐induced cell
   death, and CD36 acts as a vital transporter in hepatocyte lipid
   metabolism.^[ [144]^13 ^] Hence, in this study, an in‐vitro 3D culture
   experiment was conducted to investigate whether inflammation can affect
   the functionality of CD36. Consequently, our research proved consistent
   results with previous studies, revealing that inflammation increased
   lipid accumulation in hepatocytes; LC‐MS analysis identified TGR5 as
   the most significantly increased protein in its binding to CD36,
   compared to the non‐inflammatory environment.

   This discovery is of significant importance as CD36 combined to TGR5 is
   associated with a substantial increase in hepatocyte mortality,
   primarily due to severe impairment of CD36 and its normal function.
   CD36, also known as fatty acid translocase (FAT), is a cell surface
   protein that plays a pivotal role in the metabolism of FFAs in various
   tissues.^[ [145]^20 , [146]^21 , [147]^22 ^] It has been shown that
   CD36 can actively participate in hepatocyte lipid metabolism through
   diverse functions such as FFA uptake, fatty acid storage, energy
   production, insulin sensitivity, etc.^[ [148]^23 , [149]^24 , [150]^25
   ^] The function of CD36 depends on a fundamental cellular process
   called endocytosis, which is a vital means for cells to transport
   extracellular biomacromolecules.^[ [151]^26 ^] Given the binding
   between TGR5 and CD36, both membrane receptors are present within
   entotic vesicles. Our discovery indicated that TGR5 can serve as a
   bridge to facilitate the degradation of BBOX1, a pivotal protein
   responsible for FFAs entering mitochondria through the ubiquitination
   proteasome pathway. Consequently, this process hinders the
   beta‐oxidation of FFAs before vesicles enter the mitochondria.

   TGR5, short for G protein‐coupled bile acid receptor 1 (Gpbar1), is a
   cell membrane receptor that can significantly regulate various
   physiological processes.^[ [152]^14 , [153]^27 , [154]^28 ^] TGR5
   primarily responds to bile acids and participates in controlling
   metabolism, inflammation, and energy balance.^[ [155]^12 , [156]^14 ^]
   It is present in many tissues throughout the body, including the liver,
   intestine, and immune cells.^[ [157]^14 ^] A previous study has pointed
   out that TGR5 exhibits a higher sensitivity to SBAs than to primary
   bile acids.^[ [158]^29 ^] Notably, the existing research has indicated
   that TGR5 expression is absent in healthy hepatocytes,^[ [159]^30 ,
   [160]^31 ^] which serves as a protective mechanism against the hepatic
   effects of bile acid originating from intestinal bacteria.^[ [161]^32
   ^] Our investigation revealed that, in the presence of inflammatory
   stimulation, the Gpbar1 gene was transcriptionally active under the
   regulation of NF‐κB. This elevated expression enhanced the
   responsiveness of both murine and human liver cells to SBAs.
   Furthermore, our data elucidated that a notable elevation in serum
   concentrations of SBAs (particularly LCA and its derivatives) was
   correlated with MASH fibrosis progression. These data strongly
   supported that TGR5 may serve as a promising target for MASH prevention
   or treatment.

   We found that TGR5 in hepatocytes emerges in an inflammatory
   environment and can promote MASH development by inducing hepatocyte
   death. Nonetheless, TGR5 is a classic metabolism‐related receptor that
   plays a relatively well‐understood role in macrophages, especially
   hepatic macrophages, where it exerts an anti‐inflammatory effect by
   inhibiting the NF‐κB pathway activation.^[ [162]^15 , [163]^16 ^]
   Therefore, in the MASLD model of macrophage TGR5 knockout mice, TGR5
   plays a protective role, which is opposite to its role in hepatocytes.
   To address this issue, we used a dual knockout model and found that
   when TGR5 in hepatocytes was further knocked out at 16 weeks in the
   MASH model, the MASH phenotype in the dual knockout mice was noticeably
   alleviated, whether compared to macrophage TGR5 knockout mice or
   wild‐type (WT) mice. These results indicated that TGR5 on hepatocytes,
   rather than on macrophages, plays a more important role in the
   progression of MASH. The reason for this phenomenon may be that TGR5
   inhibits the activation of the NF‐κB pathway. In hepatocytes, the NF‐κB
   pathway activation is generally considered protective, especially
   during inflammatory injury, as it promotes the expression of
   anti‐apoptotic genes and helps maintain liver function. However, in
   macrophages, the NF‐κB pathway activation is primarily associated with
   the cell's ability to promote inflammation.^[ [164]^33 ^] Under
   non‐inflammatory conditions, hepatocytes do not express TGR5. However,
   once TGR5 expression is upregulated, the de novo synthesis of
   carnitine‐related substances is reduced, resulting in a significant
   decrease in beta‐oxidation in hepatocytes, accompanied by a marked
   increase in hepatocyte death. This phenomenon may also be attributed to
   the inactivation of the NF‐κB pathway caused by elevated TGR5 levels in
   hepatocytes. Therefore, although the activation of TGR5 and its
   downstream effects in macrophages are associated with reduced
   pro‐inflammatory capacity, its activation in hepatocytes primarily
   leads to cell death. Both phenomena may be linked to the NF‐κB
   signaling pathway, highlighting the cell type‐specific roles of TGR5.

   TGR5 inhibitors and CRISPR/Cas9 techniques were used for both in vivo
   and in vitro targeted interventions, yielding favorable outcomes, as
   shown by improved liver fibrosis, mitigated liver inflammation, and
   augmented hepatic lipid metabolism. It is worth noting that inhibition
   of TGR5 receptor activity has emerged as an efficacious therapeutic
   strategy. The primary mechanism underlying this intervention involves
   the disruption of TRIM21 recruitment around the BBOX1 protein when TGR5
   activity is inhibited or knocked out. Consequently, this intervention
   could impede the efficient degradation of BBOX1 protein, thereby
   safeguarding the lipid metabolic capacity of hepatocytes, alleviating
   the burden of lipid‐induced cytotoxicity, and ultimately leading to
   reduced hepatocyte mortality.

   Carnitine is a natural compound in various foods, especially in meat
   and dairy products, which plays a crucial role in the metabolism of
   fatty acids. The primary function of carnitine is to transport
   long‐chain fatty acids to the mitochondria, the energy‐producing
   organelles within cells. Carnitine facilitates the movement of these
   fatty acids across the mitochondrial membranes, where they undergo
   beta‐oxidation, a process in which fatty acids are broken down to
   produce ATP.^[ [165]^34 ^] In addition, carnitine can be synthesized de
   novo within the human body from amino acids, primarily in the liver and
   kidneys, using specific amino acids and cofactors.^[ [166]^35 ^] Our in
   vivo investigations demonstrated that the administration of exogenous
   carnitine could effectively prevent MASLD development. Nonetheless, it
   is noteworthy that exogenous supplementation alone cannot rectify the
   diminished capacity of hepatic cells to biosynthesize carnitine. This
   limitation is associated with the interaction between TGR5 and the E3
   ubiquitination ligase TRIM21, which leads to the degradation of BBOX1,
   a pivotal enzyme in the de novo carnitine biosynthesis pathway,
   specifically via the K283 site.

   TRIM21 is a member of the TRIM family of proteins and has gained much
   attention for its multifaceted roles in immune regulation,
   intracellular signaling, and protein degradation.^[ [167]^36 ^] In the
   context of MASH, the involvement of TRIM21 is particularly intriguing.
   Recent research has shed light on the potential effect of TRIM21 on
   MASH pathogenesis. It is believed that TRIM21 can modulate the immune
   response and inflammation within the liver and influence the severity
   of liver injury and fibrosis. In addition, TRIM21 may play a role in
   the regulation of cellular responses to oxidative stress, which is a
   key contributor to MASH development and progression.^[ [168]^37 ,
   [169]^38 ^] Furthermore, it has been shown that the capacity of TRIM21
   to target specific proteins for ubiquitination and subsequent
   degradation is crucial in MASH.^[ [170]^39 ^] Taken together, TRIM21
   may affect the degradation of proteins involved in lipid metabolism,
   inflammation, and fibrosis, thereby affecting the overall course of the
   disease.

   In summary, our investigation initially focused on the influence of
   inflammation on CD36‐mediated lipid metabolism. Remarkably, we
   uncovered a novel phenomenon that TGR5, a CD36‐bound SBA receptor, was
   expressed on hepatocytes in response to inflammatory cues. This
   unregulated TGR5 expression, when overexpressed, disrupted the crucial
   supply route of FFAs into mitochondria. This disruption was achieved
   through recruiting E3 ubiquitination enzyme TRIM21, leading to the
   degradation of BBOX1, an enzyme pivotal in the rate‐limiting step of
   the de novo carnitine biosynthesis pathway. Consequently, this
   disruption amplified lipid accumulation within hepatocytes and
   intensified lipotoxicity, ultimately triggering cell death,
   inflammation, and eventually liver fibrosis. In this study, critical
   factors (including TGR5, carnitine metabolism, and TRIM21) were
   pinpointed as potential targets for preventing and treating the
   transition from simple fatty liver to MASH. These findings not only
   offer novel therapeutic approaches for this condition but also
   significantly enhance our understanding of the underlying mechanisms.

4. Experimental Section

Human Study

   Participants for this study were enrolled between January 2021 and June
   2022 at the liver clinic associated with the First Affiliated Hospital
   of Anhui Medical University, Gulou Hospital, and People's Hospital of
   Taizhou. Inclusion criteria were adults between the ages of 35 and 65
   years, and all participants were categorized as Healthy control, MASLD,
   and MASH, as evidenced by histopathological analysis of liver biopsy
   specimens. Liver histopathology for all study groups was independently
   verified by a pair of medical society‐certified pathologists to ensure
   the validity of inclusion. Individuals with one of the following
   conditions were excluded from participation in this study: documented
   history of chronic viral hepatitis, excessive alcohol consumption, or
   other diagnoses of liver disease. Liver biopsy samples were collected
   using an ultrasound‐guided technique. A venous blood sample was also
   collected from each participant for further analysis. All collected
   samples were rapidly frozen and preserved in liquid nitrogen and then
   stored at −80 °C until analysis. The study was approved by the
   Institutional Ethics Committee of the First Affiliated Hospital of
   Anhui Medical University to ensure compliance with ethical standards
   and participant safety.

   Cohort from Wenzhou is a well‐characterized Prospective Epidemic
   Research Specifically of MASH (PERSONS) cohort. All MASLD patients in
   this study were consecutively recruited from 2016 to 2019 at the First
   Affiliated Hospital of Wenzhou Medical University in Wenzhou (China).
   The inclusion and exclusion criteria and liver histology assessment
   have been described extensively elsewhere.^[ [171]^1 ^] Written
   informed consent was obtained from each subject before study
   participation. The research protocol was approved by the ethics
   committee of the First Affiliated Hospital of Wenzhou Medical
   University (2016‐246, December 01, 2016) and registered in the Chinese
   Clinical Trial Registry (ChiCTR‐EOC‐17013562). A total of 550 Chinese
   adults with biopsy‐proven MASLD and 15 heathy volunteers as controls
   were included in the present study. Serum samples were collected for
   targeted metabolomics analysis of bile acids that had been reported in
   a previous work.^[ [172]^2 ^]

Mouse Models

   C57BL/6 mice, Gpbar1 ^fl/fl mice, Alb‐Cre mice, and Lyz‐Cre mice were
   purchased from GemPharmatech, while Rosa26‐LSL‐Cas9 knockin mice were
   obtained from The Jackson Laboratory. To generate hepatocyte‐specific
   deletions of Gpbar1, Gpbar1^fl/fl mice were crossed with Alb‐Cre mice.
   To generate myeloid‐specific deletions of Gpbar1, Gpbar1^fl/fl mice
   were crossed with Lyz‐Cre mice. To generate myeloid‐specific Gpbar1
   knockout mice with Cas9, we crossed myeloid‐specific Gpbar1 knockout
   mice (Gpbar1 ^fl/fl × Lyz‐Cre) with Rosa26‐LSL‐Cas9 knockin mice. All
   procedures involving animals were performed in compliance with the
   ethical standards of the Institutional Animal Care and Use Committee
   (IACUC) at the First Affiliated Hospital of Anhui Medical University.
   The mice were housed in a specific pathogen‐free environment at the
   same institution. The housing conditions included standard
   polypropylene cages with a 12‐h light/dark cycle, and were maintained
   at a stable temperature of 22 ± 2 °C and humidity of 55 ± 10%. Food and
   water were provided ad libitum. The mice were fed either a chow diet
   (Cat: D12450B, Research Diets), a high‐fat, high‐fructose (HFHfr) diet
   consisting of 40 kcal% fat (primarily palm oil), 20 kcal% fructose, and
   2% cholesterol (Cat: D09100310, Research Diets), or a WD containing 40%
   of total calories from fat, 1.25% cholesterol, and 0.5% cholic acid
   (Cat: D12109C, Research Diets).

Methods Details–Animal Experiments–Animal Experiment 1: HFHfr+C. diff FMT

   The FMT process, which involved the utilization of the C. diff strain
   Bio‐18725, was carefully executed following established protocols. The
   C. diff strain was sourced from Beijing Biobw Company. For the
   cultivation of bacteria, an anaerobic environment was provided by the
   AnaeroPack system (Model MGC‐041, Mitsubishi Gas Chemical Company). The
   growth medium used was brain‐heart infusion (BHI) medium, procured from
   HiMedia Laboratories GmbH in Einhausen, Germany. This medium was
   further enriched with 5 g L^−1 of yeast extract (sourced from Biolab
   Zrt., Budapest, Hungary) and 0.1% l‐cysteine (supplied by Acros
   Organics, Thermo Fisher Scientific, Waltham, MA). Bacterial cultures
   were incubated at optimal growth conditions for 24 h to reach an
   optical density of 1.5 in BHI medium. Subsequently, the cultures were
   aliquoted into 1 mL volumes to prepare for gavage administration into
   mice. Following a 12‐week regimen of a HFHfr diet (Research Diets,
   D12040701), the mice underwent biweekly oral gavage C. diff and
   continued on the high‐fat diet for a total of 16 weeks, as detailed in
   the protocol cited.^[ [173]^3 ^]

Animal Experiment 2: LCA+HFHfr Model

   Following a 12‐week regimen of a HFHfr diet, mice were orally
   administered LCA a dose of 50 mg kg^−1 dissolved in saline once a day
   along with HFHfr diet until 16 weeks.

Animal Experiment 3: HFHfr Model

   Mice (6–8 weeks old C57BL/6J, male) were randomly divided and treated
   with HFHfr diets (Research Diets, D12040701). Six mice from each group
   were euthanized after 0, 8, and 16 weeks. The liver specimen was
   collected for further histological evaluation and multi‐omics analysis.

Animal Experiment 4: WD Model

   This experiment followed a similar protocol to Animal experiment 3, but
   with the inclusion of a WD (Research Diets, D12079B).

Animal Experiment 5: Therapeutic Model

   To model MASH therapeutically, Rosa26‐LSL‐Cas9 knockin mice, aged 8
   weeks, were treated with HFHfr diet for 8 weeks. Mice were processed
   with the initiation of either HFHfr+C. diff FMT or LCA+HFHfr model for
   four weeks. Then, the mice were randomized into three treatment groups:
   (1) Bbox1 OE group (n = 8), injected with liver‐targeted
   Adeno‐Associated Virus 8 (AAV8) carrying the Alb‐Bbox1 gene; (2) Gpbar1
   KO group (n = 8), receiving AAV8 with Alb‐Cre‐U6‐Gpbar1 sgRNA; (3)
   Trim21 KO group (n = 8), injected with AAV8 carrying the
   Alb‐Cre‐U6‐Trim21 sgRNA. Each injection had a dose of 1 × 10^10 viral
   genomes (vg) per mouse. Each mouse received 200 µL of the virus
   solution. Each receiving weekly intravenous tail vein injections, along
   with either HFHfr+C. diff FMT or LCA+HFHfr model, with continued
   proceeding until the 16th week.

Animal Experiment 6: Prevention Model

   In the prevention model for MASH, Rosa26‐LSL‐Cas9 knockin mice at 8
   weeks of age were treated with HFHfr diet for 8 weeks. Then, the mice
   were randomized into three treatment groups, similar to the therapeutic
   model. Each group received tail vein injections weekly along with the
   treatment of the HFHfr+C. diff FMT or LCA+HFHfr model, with continued
   proceeding until the 16th week.

Cell Culture and Reagents

   The mouse liver cell line AML12, procured from the China Model Culture
   Center, Wuhan, and the Raw264.7 cell line and THP‐1 cell line, sourced
   from the Cell Bank of the Chinese Academy of Sciences, Shanghai, and
   HepRG cell line, sourced from Thermo Fisher Scientific were all
   authenticated as free from mycoplasma contamination. AML12, a mouse
   hepatocyte cell line, was cultured following standard protocols. The
   cells were maintained in Dulbecco's Modified Eagle Medium/Nutrient
   Mixture F‐12 Ham (DMEM/F12) supplemented with 10% fetal bovine serum
   (FBS) (Catalog No. 085‐150, WISENT), 1% insulin‐transferrin‐selenium
   (ITS), and 40 ng mL^−1 dexamethasone. The culture medium was also
   supplemented with 100 U mL^−1 penicillin and 100 µg mL^−1 streptomycin
   to prevent bacterial contamination. Raw264.7 cells were cultured in
   DMEM, enriched with 1% penicillin‐streptomycin (Catalog No. BL505A,
   Biosharp Life Sciences) and 10% high‐quality FBS (Catalog No. 085‐150,
   WISENT). THP‐1 cells were cultured in RPMI 1640 medium (Catalog No.
   350‐000‐CL, WISENT) supplemented with 10% fetal bovine serum, 1%
   penicillin‐streptomycin, and 0.05 × 10^−3 m β‐mercaptoethanol (Catalog
   No. 444203, Sigma‐Aldrich). To induce differentiation into macrophages,
   THP‐1 cells were seeded at a density of 5 × 10^5 cells mL^−1 and
   treated with 100 ng mL^−1 PMA (Catalog No. P8139, Sigma‐Aldrich) for 48
   h. The medium was then replaced with fresh PMA‐free medium, and cells
   were allowed to rest for 24 h to complete differentiation. For
   polarization into M1 macrophages, differentiated macrophages were
   stimulated with 20 ng mL^−1 IFN‐γ (Catalog No. UA040066, UA BIOSCIENCE)
   and 100 ng mL^−1 LPS (Catalog No. L8880, Solarbio) for 24 h. HepRG
   cells were cultured according to the manufacturer's protocol (Catalog
   No. HPRGC10, Thermo Fisher Scientific). The cells were maintained in
   two different media sequentially: HepaRG Thaw, Plate & General Purpose
   Medium for initial thawing and plating, followed by HepaRG
   Maintenance/Metabolism Medium for long‐term maintenance and metabolic
   studies. All cell lines were maintained in a Thermo direct heat
   incubator at 37 °C and 5% CO[2].

   To study the effects of specific treatments, cells were exposed to
   INT777 (MCE, HY‐15677) at a concentration of 6 × 10^−6 m, G/TLCA (GLCA,
   Sigma‐Aldrich, Catalog No. 474‐74‐8; TLCA, Sigma‐Aldrich, Catalog No.
   516‐90‐5) at 50 × 10^−6 m, and doxycycline hyclate (Sigma‐Aldrich,
   D5207) at 1 µg mL^−1. These reagents were added to the culture medium
   for specified durations to investigate their impact on cell behavior
   and gene expression.

3D Cell Culture

   For the 3D co‐culture of hepatocytes and macrophages (primary mouse
   hepatocyte, peritoneal macrophages [PMs] and HepRG, THP‐1), agarose
   microwells were used. Following manufacturer instructions, 330 µL of
   sterile saline containing 2% agarose (Catalog No. R0491, Thermo Fisher
   Scientific) was dispensed into sterile 3D Petri dishes (Catalog No.
   Z764051, Sigma‐Aldrich) to form the agarose microwells. After
   solidification, the microwells were transferred to a 24‐well plate. The
   microwells were equilibrated by adding 1 mL of culture medium to each
   well, followed by an incubation of over 15 min. Then, a mixture of 1.5
   × 10^5 liver cells and 1.5 × 10^5 macrophages in 75 µL of culture
   medium was seeded into the agarose microwells. Seven days post‐seeding,
   cells were harvested for the preparation of O.C.T. (Catalog No.
   G6059‐110ML, Servicebio) or paraffin‐embedded, or preparing a
   single‐cell suspension, then removing macrophages by using flow
   cytometry with an F4/80 antibody.

Biochemical Analysis

   Serum cytokine concentrations in mice were quantified utilizing
   enzyme‐linked immunosorbent assay (ELISA) kits. These kits specifically
   targeted tumor necrosis factor (TNFα) (Cloud‐Clone, Catalog No.
   SEA133Mu) and monocyte chemoattractant protein‐1 (MCP1) (Cloud‐Clone,
   Catalog No. SEA087Mu). Serum levels of ALT were determined using an
   ADVIA 2400 Chemistry System analyzer (Siemens, Tarrytown, NY). All
   assays were conducted meticulously following the manufacturer's
   instructions. For the quantitative assessment of triglycerides (TG) and
   total cholesterol (TC) in liver samples, commercial assay kits (Wako,
   Catalog No. 290‐63701 for TG, and 294‐65801 for TC) were used, with
   strict adherence to the provided protocols.

Histopathological Examination

   Histological analyses were performed using multiple staining
   approaches. Hematoxylin and eosin (H&E) staining was applied to
   paraffin‐embedded sections to examine the general tissue architecture.
   Lipid droplets in frozen liver tissues, embedded in Tissue‐Tek O.C.T.
   Compound (Servicebio, G6059‐110ML), were visualized using Oil Red O
   staining. Masson's Trichrome staining (Solarbio, Catalog No. G1346) was
   used to evaluate liver fibrosis. All images were captured using an
   Olympus light microscope and were subjected to quantitative analysis
   utilizing Image‐Pro Plus 6.0 (Media Cybernetics) software.

Multiple‐Colored Immunohistochemistry (mIHC)

   Tissue samples, meticulously evaluated by two experienced pathologists,
   were sectioned into 5‐µm slices. Immunohistochemical staining was
   performed using specific antibodies against TGR5 (1:200, ab72608,
   Abcam), BBOX1 (1:200, 16099‐1‐AP, Proteintech), CD36 (1:200,
   18836‐1‐AP, Proteintech), TRIM21 (1:200, 12108‐1‐AP, Proteintech), RIP3
   (1:200, 17563‐1‐AP, Proteintech), CD68 (1:200, Catalog No. 26042, CST),
   F4/80 (1:200, Catalog No. 70076, CST), TNFα (1:200, Catalog No.
   17590‐1‐AP, Proteintech), and OCTN2 (1:200, Catalog No. 16331‐1‐AP,
   Proteintech), followed by application of HRP‐conjugated secondary
   antibodies and various PPD dyes (Catalog No. 10234100050, PANOVUE). The
   identification of positive cells and subsequent analyses were automated
   using the Vectra 3 quantitative pathology imaging system and inForm
   software. inForm software was used, which uses an advanced
   machine‐learning‐based cell segmentation algorithm to classify and
   distinguish different cell types according to cell shape, size, and
   nuclear characteristics.

Targeted Identification of LCA and Its Derivatives Concentration in MASLD
Cohort

   Targeted metabolomics analysis of bile acids had been reported in a
   previous work.^[ [174]^2 ^] A total of 550 Chinese adults with
   biopsy‐proven MASLD and 15 heathy volunteers as controls were included
   in the present study. All MASLD patients were consecutively recruited
   from 2016 to 2019 at the First Affiliated Hospital of Wenzhou Medical
   University in Wenzhou (China). This is a well‐characterized Prospective
   Epidemic Research Specifically of MASH (PERSONS) cohort. The inclusion
   and exclusion criteria and liver histology assessment were described
   elsewhere.^[ [175]^1 ^] Written informed consent was obtained from each
   subject before study participation. The research protocol was approved
   by the ethics committee of the First Affiliated Hospital of Wenzhou
   Medical University (2016‐246, December 01, 2016) and registered in the
   Chinese Clinical Trial Registry (ChiCTR‐EOC‐17013562).

RNA‐Sequencing

   Total RNA extraction, library construction, and sequencing for this
   study were outsourced to Gene De novo Biotechnology Co. (Guangzhou,
   China). In brief, RNA was extracted using Trizol reagent and underwent
   quality assessment. Eukaryotic mRNA was enriched using Oligo (dT)
   beads, followed by fragmentation using fragmentation buffer. Reverse
   transcription to cDNA was carried out using the NEBNext Ultra RNA
   Library Prep Kit for Illumina (Catalog No. 7530, New England Biolabs,
   Ipswich, MA). The resulting double‐stranded cDNA fragments were
   processed through end repair, A‐tailing, and ligation to Illumina
   sequencing adapters. After purification with AMPure XP Beads (1.0×) and
   size selection via agarose gel electrophoresis, PCR amplification was
   performed. The final cDNA library was sequenced on an Illumina NovaSeq
   6000 platform.

KEGG Pathway Enrichment Analysis

   For KEGG pathway enrichment analysis, Fisher's exact test was utilized
   through a custom R script. KEGG pathway annotations for the selected
   genome's entire gene set were sourced from the KEGG database. Pathways
   were considered significantly enriched if they met a statistical
   significance threshold of a P‐value below 0.05.

Gene Set Enrichment and Gene Set Variation Analysis

   Gene set enrichment analysis (GSEA) and gene set variation analysis
   (GSVA) were conducted to assess pathway activity. Specifically, GSEA
   was executed using the Java GSEA platform version 3.0, utilizing gene
   sets defined for each known biological pathway sourced from the KEGG
   database. The “Signal2Noise” metric was used for ranking, and
   significance was determined based on nominal P‐values less than 0.05
   and false discovery rate (FDR) values less than 0.25. Concurrently,
   GSVA was carried out using the GSVA R package version 1.32.0 to
   evaluate sample‐wise fluctuations in KEGG pathway activity.

Quantitative Real‐Time PCR

   Total RNA from cells and liver tissue was extracted using the Total RNA
   Isolation Kit (Catalog No. RC101‐01, Vazyme Biotech Co., Ltd., China),
   according to the manufacturer's guidelines. RNA concentration was
   quantified using a NanoDrop 2000C spectrophotometer (Thermo Fisher
   Scientific, Waltham, MA). For reverse transcription, 500 ng of purified
   RNA per sample was used, employing the Reverse Transcription Kit
   (Catalog No. R223, Vazyme Biotech Co., Ltd., China). Quantitative
   real‐time PCR was conducted on a QuantStudio 6 Flex Real‐Time PCR
   System (Applied Biosystems, CA, USA) using SYBR Green (Catalog No.
   Q221, Vazyme Co., Ltd., China). Primer sequences for qPCR, synthesized
   by GENEray Biotechnology (Shanghai, China), are listed in Table [176]S1
   (Supporting Information).

Western Blot Analysis

   Cellular and mouse liver tissue proteins were extracted using RIPA
   lysis buffer (Catalog No. P0013B, Beyotime) with added protease and
   phosphatase inhibitors (Catalog Nos. HY‐K0010, HY‐K0021, MCE). After
   centrifugation at 13000g for 20 min, protein concentration was
   determined using a BCA Protein Assay Kit (Catalog No. P0011, Beyotime).
   Protein samples underwent 10% SDS‐PAGE and were transferred to PVDF
   membranes (Catalog No. 03010040001, Roche). Membranes were blocked with
   5% non‐fat milk (Catalog No. P0216‐1500g, Beyotime) and incubated
   overnight with primary antibodies. Following TBST washes, membranes
   were incubated with appropriate HRP‐conjugated secondary antibodies.
   Antibodys were performed using specific antibodies against CD36
   (1:1000, Catalog No. 18836‐1‐AP, Proteintech), TRIM21 (1:1000, Catalog
   No. 12108‐1‐AP, Proteintech), TGR5 (1:1000, Catalog No. ab72608,
   Abcam), SHMT2 (1:1000, Catalog No. 11099‐1‐AP, Proteintech), BBOX1
   (1:1000, Catalog No. 16099‐1‐AP, Proteintech), β‐actin (1:20000,
   Catalog No. 66009‐1‐Ig, Proteintech), and OCTN2 (1:1000, Catalog No.
   16331‐1‐AP, Proteintech), followed by application of HRP‐conjugated
   secondary antibodies.

Primary Mouse Hepatocyte Isolation and Cultivation

   Primary hepatocytes were isolated from 8‐week‐old male Gpbar1
   ^fl/fl×Alb‐Cre and Gpbar1 ^fl/fl mice using a two‐step collagenase
   perfusion method, as described by Li et al.^[ [177]^4 ^] Mice were
   anesthetized with 3% pentobarbital sodium (90 mg kg^−1, Catalog No.
   P3761, Sigma‐Aldrich), and the liver was perfused via the portal vein
   using specific media. Isolated hepatocytes were cultured under
   specified conditions and subjected to various treatments as outlined in
   the text.

Isolation of Peritoneal Macrophages

   PMs were induced with an intraperitoneal injection of 3% Brewer's yeast
   polysaccharide into mice, administered the day prior to harvest to
   boost cell number and activity, under sterile conditions. Mice were
   subsequently euthanized via CO[2] asphyxiation and the abdomen
   sterilized with 75% ethanol. 5–10 mL of cold PBS or DMEM was injected
   into the peritoneum, the volume tailored to the size of the mouse.
   After gently massaging the abdomen to dislodge cells, the peritoneal
   fluid was collected into centrifuge tubes. The cells were pelleted by
   centrifugation at 300g for 5 min, followed by counting with a
   hemocytometer for further experimental use.

Genotyping

   Lyz Cre genotyping was performed by PCR using DNA from myeloid cells.
   The primers used were: F:AGTGCTGAAGTCCATAGATCGG and R:
   CTGATTCTCCTCATCACCAGG to detect the targeted (Cre) allele (543 bp). PCR
   was conducted with a 2×Taq Master Mix, starting with denaturation at
   95 °C for 5 min, followed by 20 cycles of 98 °C for 30 s, 65 °C for 30
   s (decreasing 0.5 °C cycle^−1), and 72 °C for 45 s, with a final
   extension at 72 °C for 5 min. Alb Cre genotyping was performed by PCR
   using DNA from hepatocytes. The primers used were: F:
   ATTTGCCTGCATTACCGGTC and R: ATCAACGTTTTCTTTTCGG to detect the targeted
   (Cre) allele (350 bp). PCR was conducted with a 2×Taq Master Mix,
   starting with denaturation at 94 °C for 3 min, followed by 20 cycles of
   94 °C for 20 s, 64 °C for 30 s (decreasing 0.5 °C cycle^−1), and 72 °C
   for 35 s, with a final extension at 72 °C for 2 min. PCR products were
   analyzed by gel electrophoresis.

Plasmid

   The sgRNA sequences of mouse Gpbar1 (F:TGGCTAGGGCTCTCACCTGG,
   R:CCAGGGTTGAGGGTACATCG) and Trim21 (F:CAAAGGATCGGAGACAAGTG,
   R:GATGATGAGCCATGAGTTGG) sequences were cloned into the lentiCRISPR
   v2‐purinomycin vector using BfuAI and BsmBI to construct CRISPR
   plasmids. For overexpression of target genes, mouse Bbox1 genes were
   cloned into the pCDH‐CMV‐MCS‐EF1‐Puro vector. Bbox1 variants with
   specific mutations (D282E or K283N) were synthesized and similarly
   cloned by Corues Biotechnology (Nanjing, China). Plasmids were
   transfected using ExFect Transfection Reagent (Vazyme, T101‐01) and
   verified by Western blotting.

   AAV constructs (AAV8‐Alb‐Bbox1, AAV8‐Alb‐Cre‐U6‐Gpbar1 sgRNAs, and
   AAV8‐Alb‐Cre‐U6‐Trim21 sgRNAs) were created using the pAAV‐TBG‐CBh‐
   3xFLAG vector. CDS of mouse Bbox1, sgRNA of Trim21, and sgRNA of Gpbar1
   were inserted and cloned in DH5α. Post‐sequence verification, 293 cells
   were transfected with these plasmids, pAAV‐RC, and pHelper plasmid. AAV
   particles were purified (Catalog No. V1469, Biomiga) and quantified
   using PCR. For in vivo administration, the AAV8 vectors were delivered
   via tail vein injections at a dose of 1 × 10^10 vg per mouse. Each
   mouse received 200 µL of the virus solution. Injections were performed
   once per week for four consecutive weeks to ensure sustained expression
   of the overexpressed genes or knockout constructs in hepatocytes. Virus
   titers were verified using a quantitative PCR‐based assay.

   For the Tet‐off system in AML12 cells, Bbox1 was cloned into the
   pTRE‐Tight Tet‐off vector. Stable cell lines were established using
   G418 antibiotic selection, ensuring conditional BBOX1 expression
   controlled by Doxycycline in Tet‐off conditions. This system's efficacy
   was validated via qRT‐PCR and WB.

Protein‐Docking

   CD36 and TGR5 structures were predicted using Alphafold2, while BBOX1
   and TRIM21 structures were obtained from RCSB PDB (PDB ID: 4BG1, 7BBD).
   Post‐preprocessing, initial conformations were adjusted through rigid
   docking on ClusPro 2.0, followed by flexible docking using RosettaDock.
   Interprotein interactions were analyzed using Ligplot, and final
   protein conformations were visualized with Pymol software.

Co‐Immunoprecipitation and LC/Mass Spectrometry

   For immunoprecipitation assays, cells were lysed using Co‐IP lysis
   buffer (Catalog No. P0013, Beyotime) and centrifuged at 12000×g for 20
   min at 4 °C to pellet the lysate. Protein concentration was
   standardized to 1 mg mL^−1 based on BCA assay results. Protein AG beads
   (Beyotime, P2108) were pre‐equilibrated with 1 mL of wash buffer
   (Beyotime, ST661) and incubated overnight at 4 °C with 1 mg of lysate
   along with specific antibodies (5 µg of BBOX1(16099‐1‐AP, Proteintech),
   CD36(18836‐1‐AP, Proteintech), or TRIM21(12108‐1‐AP, Proteintech)) or a
   rabbit IgG isotype control. The samples were then subjected to
   electrophoresis, followed by staining with Coomassie Blue Fast Staining
   Solution (Beyotime, P0017). Mass spectrometry analysis was performed
   (Shanghai Applied Protein Technology) and the immunoprecipitated
   proteins were further assessed through immunoblotting using appropriate
   antibodies.

Statistical Analysis

   In this study, data are presented as the mean ± standard Deviation
   (SD). The generation of all bar plots was accomplished using GraphPad
   Prism 9.0 (GraphPad Software, La Jolla, CA). To determine the
   distribution of the samples, normality tests were performed. For the
   statistical comparison of normally distributed variables, one‐way ANOVA
   was used to analyze more than two groups. Two‐way ANOVA was used to
   analyze data involving two independent factors. These analyses were
   conducted using SPSS 26.0 (IBM SPSS, Chicago, IL). In addition,
   Spearman's rank correlation coefficient was used to assess the
   relationship between variables, with graphical visualizations of these
   correlations produced in R studio (RStudio, Boston, MA).

Conflict of Interest

   The authors declare no conflict of interest.

Author Contributions

   S.L., M.L., L.J., and B.Z. contributed equally to this work. R.J.,
   G.X., Y.N., Y.Y., and S.L. designed the research. S.L., M.L., J.L.,
   B.Z., Y.F., X.W., and Z.H. performed the experiments and analyzed the
   data. M.L., J.L., X.W., and Y.F. participated in mouse feeding and
   experiments. S.L., M.L., and B.Z. were responsible for data statistics.
   S.L. and Y.F. validated the manuscript and data. R.J., G.X., Y.N.,
   Y.Y., and S.L. drafted and revised the manuscript. R.J., G.X., Y.N.,
   Y.Y., and S.L. designed, supervised, wrote, and revised the paper.

Supporting information

   Supporting Information
   [178]ADVS-12-2410602-s001.docx^ (11.5MB, docx)

Acknowledgements