1. Introduction
Metabolic dysfunction-associated steatohepatitis (MASH) is a progressive phenotype of metabolic dysfunction-associated steatotic liver disease (MASLD), characterized by the cooccurrence of inflammation, hepatocellular damage, and fibrotic restructuring. MASH is a major risk factor of hepatocellular carcinoma and the consequential demand for liver transplantation [
1], [
2]. Despite extensive research, the underlying pathophysiology of MASH still lacks comprehensive understanding [
3].
Altered expression of multiple lipid-related genes is a crucial factor in MASH development [
4]. In particular, carboxyl ester lipase gene (
Cel) is vital for regulating lipid metabolism by breaking down cholesteryl esters and triglycerides (TGs), leading to the formation of free cholesterol and fatty acids [
5], [
6]. Emerging evidence indicates that carboxyl ester lipase (CEL) contributes to the development of metabolic disorders such as dysglycemia [
7] and atherosclerosis [
8], [
9]. Another study also reported that CEL regulates selective hepatic uptake and metabolism of cholesteryl esters from high-density lipoprotein through interactions with the scavenger receptor pathway [
10]. Our previous studies revealed that
Cel is recurrently mutated in obese mice, and knockdown or mutation of
Cel increases lipid levels and endoplasmic reticulum stress in hepatocytes [
11], both of which are well-documented contributors to MASH development [
12], [
13]. Nevertheless, the precise function of CEL in MASH remains unclear.
Herein, our results show that hepatocyte-specific Cel knockout (CelΔHEP) mice exhibited exacerbated steatohepatitis upon exposure to choline-deficient high-fat diet (CD-HFD) or methionine- and choline-deficient diet (MCD). Mechanistically, CEL bound directly to fatty acid synthase (FASN), leading to reduced SUMOylation and consequently enhanced proteasomal degradation of FASN. Hepatocyte-specific overexpression of CEL significantly attenuated steatohepatitis induced by CD-HFD and MCD in mouse models. These findings collectively highlight the protective effect of CEL against MASH by modulating de novo lipogenesis mediated by FASN.
2. Materials and methods
2.1. Mouse models and treatment
Transgenic mouse model: Male mice with hepatocyte-specific knockout of Cel (CelΔHEP) were bred by Biocytogen Biocytogen Pharmaceuticals (Beijing) Co., Ltd. (China). To induce MASH, male CelΔHEP mice or wildtype (WT) mice were subjected to 16 weeks of CD-HFD feeding, with normal chow (NC) used as control. Alternatively, mice were subjected to three weeks MCD feeding, and MCD was used as control. CD-HFD and MCD was purchased from Research Diets, Inc. (USA). Before sacrifice, mice were fasted and their body weights were measured. Liver tissues and serum were collected upon sacrifice.
Adeno-associated viruses (AAVs)-mediated hepatocyte-specific CEL overexpression mouse model: Adeno-associated viral, serotype 8 (AAV8) carrying the thyroxine-binding globulin promoter with/without Cel exon sequence (AAV8-Cel and AAV8-null) were obtained from OBiO Technology (Shanghai) Corp., Ltd. (China). Mice received intravenous injections of AAV8-Cel or AAV8-null at a concentration of 1 × 1011 genome copies per mouse. Two weeks after injection, MASH was induced following the method described in the preceding paragraph.
To determine the effect of FASN inhibition on Cel knockout in vivo, we utilized AAV8 vectors carrying Cel single guide RNA (sgRNA) and clustered regularly interspaced short palindromic repeats (CRISPR)-associated protein 9 (AAV8-sgRNA (Cel), and AAV8-sgRNA (NC); OBiO Technology (Shanghai) Corp., Ltd.). Mice received intravenous injections of AAV8-sgRNA (Cel) or AAV8-sgRNA (NC) at a concentration of 1 × 1011 genome copies per mouse. Two weeks after injection, MASH was induced by MCD and simultaneously orally gavaged with FASN inhibitor TVB-3664 at a dosage of 3 mg∙kg−1 per day.
The animal experiments conducted were approved by the Animal Experimentation Ethics Committee of the Chinese University of Hong Kong.
2.2. Cell culture and treatment
Male conventional C57BL/6 mice were used for mouse primary hepatocytes (MPHs) isolation as previously described [
14]. Briefly, mouse livers were perfused with ethylene glycol tetraacetic acid (EGTA) followed by collagenase D (11088858001; Sigma-Aldrich, USA) to obtain MPH through mechanical and density-based separation. The isolated MPHs were quantified and plated onto plates coated with collagen for 3 h and cultured in Dulbecco’s modified Eagle’s medium and Ham’s F-12 medium (DMEM/F12) with 10% fetal bovine serum (FBS). Murine immortalized hepatocyte cell line AML12 was cultured in DMEM/F12 with 10% FBS and insulin-transferrin-selenium.
In vitro MASH model was established by culturing cells with palmitic acid and oleic acid dissolved in 10% fatty acid-free bovine serum albumin (BSA). AML12 cells or MPH were fasted in FBS-free medium for 8 h and then cultured with palmitic acid and oleic acid (0.33/0.66 mmol∙L
−1) in DMEM/F12 or DMEM without FBS for 24 h. Another
in vitro MASH model was established by culturing MPH exposed to the control medium or DMEM/F12 medium deficient in methionine and choline (MCD medium, Shanghai Anwei Biotechnology Co., Ltd., China) for 48 h.
To knockdown Cel or Fasn, specific siRNAs targeting Cel (siCel) or Fasn (siFasn) and a non-targeting control siRNA (siNC) were obtained from OBiO Technology (Shanghai) Corp., Ltd. siRNA transfection was performed using LipofectamineTM RNAiMAX (Thermo Fisher Scientific, USA) according to the manufacturer’s instructions.
2.3. Histological examination
Liver tissues were fixed in 10% formalin overnight at room temperature. Following fixation, the tissues were dehydrated and subsequently embedded in paraffin. Hematoxylin and eosin (H&E) staining was performed for histological analysis. Two independent investigators evaluated and scored the sections for inflammation, steatosis, and ballooning according to the Nursing Activities Score (NAS) scoring system. The NAS were determined by summing steatosis, inflammation, and ballooning scores. To assess steatosis, frozen liver sections were subjected to Oil Red O staining. Sections (4 μm) were cut using a cryostat and fixed in 4% (v/v) paraformaldehyde for 10 min before staining with Oil Red O working solution (Sigma-Aldrich).
2.4. Western blot (WB) assay
Protein samples were lysed with radio immunoprecipitation assay (RIPA) buffer and quantified using a bicinchoninic acid (BCA) assay kit (Thermo Fisher Scientific). WB was performed as previously described. The primary antibodies used are detailed in Table S1 in Appendix A.
2.5. Pull-down assay
Mouse liver tissues were lysed in RIPA buffer for co-immunoprecipitation (co-IP) analyses. Protein lysates (1 mg) were incubated overnight at 4 °C with either 2 μg of anti-CEL or control mouse immunoglobulin G (IgG). The resulting lysate-antibody mixture was then subjected to a 4 h incubation with 20 μL Protein A/G Plus Agarose Beads (sc-2003; Santa Cruz Biotechnology, Inc., USA). Beads were washed using phosphate-buffered saline (PBS) containing 0.05% Tween-20 and then heated with the loading buffer. The separated proteins were examined via silver staining or WB and characterized with liquid chromatography-mass spectrometry.
For the pull-down assay, 1 μg recombinant mouse CEL protein (50583-M08B; Sino Biological, China) and 1 μg of recombinant mouse FASN protein (ABIN3009348; Antibodies-online GmbH, Germany) were incubated overnight at 4 °C. Subsequently, the mixture was exposed to incubation with either 1 μg of anti-CEL antibody or 1 μg of anti-FASN antibody overnight at 4 °C, then incubated with 50 μL of Protein A/G Plus Agarose Beads or Protein A/G Magnetic Beads. The beads were then washed with PBS containing 0.05% Tween-20 and boiled to elute the proteins. The proteins were then analyzed by WB.
2.6. Liver total cholesterol (TC), TG, and thiobarbituric acid reactive substances (TBARS) assays
Liver tissue samples weighing between 20 and 30 mg were homogenized. The levels of TC were detected using Cholesterol Quantification Kit (ab65359; Abcam, UK) according to manufacturer’s instructions. TG levels were detected using the Wako E-test TG kit (Wako Pure Chemical Industries, Ltd., Japan). Lipid peroxidation was evaluated as malondialdehyde level through TBARS assay (MAK085; Sigma-Aldrich). The experiments were conducted in triplicates.
2.7. Glucose and insulin tolerance tests
Following overnight fasting, mice were intraperitoneally injected of 1 g glucose per 1 kg body weight for the intraperitoneal glucose tolerance test (IPGTT) or 0.75 U insulin per 1 kg body weight for the intraperitoneal insulin tolerance test (IPITT). Blood glucose levels were monitored at designated time intervals.
2.8. Serum biochemical assay
Serum levels of TC, TG, alanine aminotransferase (ALT), aspartate aminotransferase (AST), and blood urea nitrogen (BUN) were determined using Catalyst One Chemistry Analyzer (IDEXX Laboratories, Inc., USA) following the manufacturer’s protocol.
2.9. RNA extraction and mouse genotyping
TRIzol Reagent (15596018; Invitrogen, USA) was used to extract total RNA, and subsequent complementary DNA (cDNA) synthesis was performed with the PrimeScript RT Reagent Kit (RR037B; Takara, Japan) using 1000 ng of RNA. The primers employed for genotyping purposes are listed in Table S2 in Appendix A.
2.10. Fatty liver polymerase chain reaction (PCR) array
Mouse liver cDNA was used for the Mouse Fatty Liver PCR Array (PAMM-157ZF; Qiagen, Germany). The relative expression levels for the candidate genes were evaluated by ΔΔCt methodology. Genes with fold changes > 2 were considered statistically significantly different.
2.11. Immunohistochemistry (IHC)
IHC staining was performed on mouse liver, pancreas, and intestine samples. After paraffin embedding, the samples were sectioned, deparaffinized, and rehydrated. Antigen retrieval was carried out using sodium citrate buffer. Subsequently, the sections were treated with 3% H2O2 and blocked with 5% goat serum. The CEL primary antibody (Table S1) was applied overnight at 4 °C, followed by incubation with goat anti-rabbit IgG-horseradish peroxidase (HRP) conjugate (Bio-Rad, USA). Counterstaining was performed using hematoxylin, followed by the capture of microscopic images.
2.12. Real-time PCR
Real-time PCR was conducted using TB Green Advantage Quantitative Polymerase Chain Reaction Premix (639676; Takara Bio INC, Japan) in the QuantStudio 7 Flex Real-Time PCR System (Thermo Fisher Scientific). Each sample was subjected to triplicate testing. Primers used for real-time PCR are listed in Table S2.
2.13. Statistical analysis
Results are presented as means ± standard deviation (SD). Student’s t-test or Mann-Whitney U test was used for comparison between two groups. Differences in numerical variables among multiple groups were evaluated using the analysis of variance (ANOVA) method. Glucose tolerance and insulin release test differences were compared using the two-way ANOVA method. Statistical analyses were performed using the SPSS (Version 26.0; International Business Machines Corporation, USA) or GraphPad Prism (Version 8.4.3; USA).
3. Results
3.1. Hepatocyte-specific Cel knockout accelerates CD-HFD-induced MASH in mice
We first detected the hepatic CEL protein level of mice fed with CD-HFD or MCD. The results demonstrated the downregulation of CEL protein in these MASH mouse models (
Fig. 1(a)). To explore the role of CEL in MASH, we generated hepatocyte-specific
Cel knockout (
CelΔHEP) mice (
Fig. 1(b)
, Figs. S1(a) and S1(b) in Appendix A) and fed them with CD-HFD for 16 weeks (
Fig. 1(c)). Hepatocyte-specific
Cel knockout was validated by quantitative PCR (qPCR) to detect the messenger RNA (mRNA) expression of
Cel in the liver, pancreas, and intestine of
CelΔHEP mice and WT littermates. The results demonstrated the specific loss of
Cel expression in the liver of
CelΔHEP mice (Fig. S1(c) in Appendix A). The absence of hepatic CEL in
CelΔHEP mice was further validated by WB (Fig. S1(d) in Appendix A) and IHC staining (Fig. S2 in Appendix A), while CEL was expressed in the liver of WT mice. In contrast, we observed high expression of CEL in the pancreas and weak expression in the intestines of both
CelΔHEP and WT mice (Figs. S1(d) and S2).
CelΔHEP mice had significantly higher body weight and liver weight (
P < 0.01) compared to their WT littermates (
Fig. 1(d)). Hepatic levels of TG, TC, and TBARS were markedly increased in
CelΔHEP mice compared to WT mice (all
P < 0.01,
Fig. 1(e)), accompanied by reduced insulin sensitivity, as demonstrated by the IPGTT/IPITT assay (all
P < 0.01,
Fig. 1(f)). Serum ALT and AST, as well as TG and TC levels, were also significantly elevated in
CelΔHEP mice (all
P < 0.01,
Fig. 1(g)). Histological analysis revealed that
CelΔHEP mice developed more severe hepatic steatosis, inflammation, and ballooning in the livers (
Fig. 1(h)). The mRNA expressions of genes related to cholesterol metabolism (sterol regulatory element-binding protein 2 gene (
Srebp2), β-hydroxy β-methylglutaryl-CoA reductase gene (
Hmgcr)
, squalene epoxidase gene (
Sqle)
, low density lipoprotein receptor gene (
Ldlr)), lipogenesis (acetyl-CoA carboxylase alpha gene (
Acaca), fatty acid synthase gene (
Fasn)), and lipid oxidation (acyl-CoA oxidase (
Acox)
, protein kinase B (AKT) serine/threonine kinase 1 gene (
Akt1)
, mechanistic target of rapamycin kinase gene (
Mtor)) were significantly elevated in
CelΔHEP mice compared to WT mice (
Fig. 1(i)). Moreover, activation of the pro-inflammatory nuclear factor kappa-light-chain-enhancer of activated B cell (NF-κB) pathway was observed in
CelΔHEP mice, as evidenced by reduced levels of inhibitor of NF-κB alpha (IκBα) and increased phosphorylation of NF-κB p65 in liver tissues from
CelΔHEP mice (
Fig. 1(j)). More severe macrophage infiltration was observed in the liver of
CelΔHEP mice, as indicated by IHC staining of the macrophage marker F4/80 (
P < 0.01,
Fig. 1(k)). The mRNA expressions of inflammatory markers interleukin 6 (
Il6) and tumor necrosis factor alpha (
Tnfα) were also elevated in
CelΔHEP mice (
P < 0.01,
Fig. 1(i)). Additionally,
CelΔHEP mice had more pronounced liver fibrosis by Sirius Red staining (
P < 0.01,
Fig. 1(l)), accompanied by elevated mRNA expressions of fibrosis markers α-smooth muscle actin (
αSma) and collagen type I alpha 1 chain (
Col1a1) (
P < 0.01,
Fig. 1(i)). Thus, our findings illustrate that the hepatocyte-specific knockout of
Cel exacerbates the manifestation of MASH induced by CD-HFD in mice.
3.2. Hepatocyte-specific Cel knockout accelerates MCD-induced MASH in mice
To validate our findings, we employed an MCD-induced MASH mouse model using
CelΔHEP mice (
Fig. 2(a)). Our results indicated increased hepatic TG, TC, and TBARS levels in
CelΔHEP mice compared to WT mice (all
P < 0.01,
Fig. 2(b)). Serum ALT (
P < 0.05) and AST (
P < 0.01) were also significantly higher in
CelΔHEP mice than in WT mice (
Fig. 2(c)). Histological analysis revealed that
CelΔHEP mice developed severe hepatic steatosis, inflammation, and ballooning in the liver (
Fig. 2(d)). Pro-inflammatory NF-κB pathway activity (
Fig. 2(e)) and mRNA expressions of inflammation-related genes (
Il6 and
Tnfα) (
Fig. 2(f)) were elevated in
CelΔHEP mice. Furthermore, increased macrophage infiltration was observed in hepatocyte-specific knockout of
Cel by F4/80 staining (
Fig. 2(g)). Our consistent observations from the two mouse models confirm that hepatocyte-specific
Cel knockout exacerbates the development of diet-induced MASH.
3.3. CEL modulates hepatocyte steatosis and inflammation in vitro
We further investigated the role of
Cel in MASH by treating MPHs and mouse normal hepatocytes (AML12) with fatty acids (oleic acid and palmitic acid (OA/PA)). The results showed that treatment with fatty acids led to more pronounced lipid accumulation in MPH and AML12 cells (all
P < 0.01,
Fig. 3(a)). Elevated cellular TG levels (all
P < 0.01,
Fig. 3(b)) were also observed in cells after knockdown of
Cel by siRNA, together with NF-κB pathway activation as shown by decreased IκBα and elevated NF-κB p65 phosphorylation (
Fig. 3(c)). In contrast, CEL overexpression ameliorated lipid accumulation in both MPH (
P < 0.01) and AML12 cells (
P < 0.05,
Fig. 3(d)), along with decreased cellular TG level (all
P < 0.01,
Fig. 3(e)) and expressions of pro-inflammatory proteins (
Fig. 3(f)).
We established another in vitro MASH model by culturing MPH and AML12 cells with MCD medium to induce severe inflammation and steatosis. The results revealed that the utilization of MCD medium resulted in heightened lipid accumulation, a phenomenon exacerbated by Cel knockdown (all P < 0.01, Figs. S3(a) and (b) in Appendix A). Additionally, cellular inflammation levels were assessed, revealing increased mRNA expressions of Il6 and Tnfα in both MPH and AML12 cells after Cel knockdown (all P < 0.01, Fig. S3(c) in Appendix A), along with activation of the NF-κB pathway, as indicated by decreased IκBα and elevated NF-κB p65 phosphorylation (Fig. S3(d) in Appendix A). Together, our in vitro and in vivo findings suggest that CEL regulates steatosis and inflammation in MASH.
3.4. CEL directly targets FASN to protect against MASH
To elucidate the underlying mechanism, co-IP and mass spectrometry were conducted on the mouse liver tissues to identify the binding targets of CEL. Our results suggest that FASN could be a target of CEL (
Fig. 4(a)). For validation, we conducted anti-CEL/anti-FASN co-IP on mouse liver tissues, to pull down FASN and CEL, respectively (
Fig. 4(b)). The direct interaction between FASN and CEL was consistently confirmed by performing co-IP (
Fig. 4(c)) and far-WB (
Fig. 4(d)) using recombinant proteins.
Previous studies have suggested that SUMOylation, particularly small ubiquitin-related modifier 2 (SUMO2)-mediated SUMOylation, protects FASN from proteasomal degradation [
15]. We, therefore, examined SUMOylation of FASN after overexpressing CEL in hepatocytes. Our results demonstrated that the level of SUMOylated FASN decreased in cells after FASN was pulled down, while the SUMOylation level in total cell lysates remained unchanged. This suggested that CEL may specifically target and reduce SUMOylation of FASN (
Fig. 4(e)). We also observed that CEL overexpression resulted in the downregulation of FASN protein expression in mouse hepatocytes (
Fig. 4(f)). To explore whether such a decrease was due to proteasomal degradation, we treated CEL-overexpressing cells with a proteasome inhibitor, MG132. Upon treatment, we observed an upregulation of FASN protein, compared to cells without MG132 treatment (
Fig. 4(f)). FASN protein expression was also markedly increased in
CelΔHEP mice, compared to WT mice fed with CD-HFD or MCD (
Fig. 4(g)). Moreover, immunofluorescence co-staining revealed the co-localization of CEL and FASN in mouse hepatocytes (
Fig. 4(h)). Taken together, these results suggest that CEL overexpression promotes FASN degradation via the proteasomal pathway by modulating the SUMOylation level of FASN.
3.5. Targeting FASN abolishes MASH induced by Cel knockdown
We investigated whether modulating FASN could affect the function of CEL in MASH by using siRNA to knock down
Cel and/or
Fasn in mouse hepatocytes. The results showed that the knockdown of
Cel led to significantly increased lipid accumulation (
P < 0.01,
Fig. 5(a)), cellular TG levels (
P < 0.01,
Fig. 5(b)), and NF-κB pathway activity (
Fig. 5(c)) in cells treated with fatty acids. These effects from
Cel knockdown were reversed by
Fasn siRNA treatment. For validation, we treated the cells with TVB-3664, a chemical inhibitor of FASN [
16]. In line with the siRNA results, treatment with TVB-3664 markedly reduced lipid accumulation (
P < 0.01,
Fig. 5(d)) and cellular TG levels (
P < 0.01,
Fig. 5(e)), along with increased protein expressions of IκBα and reduced ratio of phosphorylated p65/total p65 (
Fig. 5(f)) in AML12 cells treated with fatty acids.
In vivo validation was performed using the MCD mouse model. Hepatic
Cel knockdown was achieved via tail vein injection of AAV8-sgRNA (
Cel), with or without intragastric administration of FASN inhibitor TVB-3664 (
Fig. 5(g)). The results showed that TVB-3664 effectively reversed the effects of
Cel knockdown in MCD-fed mice, with decreased levels of hepatic TG, TC, TBARS, and serum ALT (
Fig. 5(h)). The use of TVB-3664 also decreased NAS and mitigated hepatic lipid accumulation induced by
Cel knockdown in MCD-fed mice (
Fig. 5(i)). Moreover, the decreased protein expressions of IκBα after TVB-3664 administration further confirmed that FASN blockade reversed the heightened inflammation induced by
Cel knockdown in MCD-fed mice (
Fig. 5(j)). These results suggest that targeting FASN could abolish the development of MASH induced by CEL knockdown.
3.6. CD-HFD-induced MASH was ameliorated by overexpression of CEL in mice
We utilized AAV8-
Cel to selectively overexpress CEL in the mouse liver to evaluate the therapeutic potential of CEL in CD-HFD-induced MASH (
Fig. 6(a)). CEL was confirmed to be overexpressed only in mouse liver (
Fig. 6(a)) but not in other organs such as the pancreas (Fig. S4 in Appendix A). In mice fed with CD-HFD, we observed that AAV8-
Cel-treated mice exhibited reduced liver weight (
P < 0.01) and body weight (
P < 0.01) compared to AAV8-null-treated mice (
Fig. 6(b)). In mice overexpressing CEL and fed with CD-HFD, hepatic TC (
P < 0.01), TG (
P < 0.01), and TBARS (
P < 0.05) levels were reduced (
Fig. 6(c)), along with improved insulin sensitivity (
Fig. 6(d)). These mice also exhibited significantly lower serum ALT (
P < 0.05), AST (
P < 0.05), TC (
P < 0.01) and TG (
P < 0.01) compared to control mice (
Fig. 6(e)). Histological analysis revealed that CEL overexpression ameliorated hepatic steatosis, inflammation, and ballooning induced by CD-HFD (
Fig. 6(f)). Furthermore, the mRNA expressions of lipogenesis-related genes including
Acaca and
Fasn, and lipid oxidation-related genes such as
Akt1, Mtor, and
Acox were significantly reduced in CEL-overexpressing mice (
Fig. 6(g)). Downregulated NF-κB activation was also observed, as evidenced by decreased phosphorylation of NF-κB p65 and increased IκBα (
Fig. 6(h)), along with a reduction in F4/80 positive cells (
Fig. 6(i)) and decreased mRNA expressions of inflammatory markers
Il6 and
Tnfα (
Fig. 6(j)) in the liver of CEL-overexpressing mice. Moreover, CEL-overexpressing mice showed less severe fibrosis with reduced collagen composition (
Fig. 6(k)) and decreased mRNA expressions of
αSma and
Col1a1 (
Fig. 6(l)), as well as downregulated hepatic FASN protein expression (
Fig. 6(h)).
3.7. MCD-induced MASH was ameliorated by overexpression of CEL in mice
To validate our findings, we employed another MASH mouse model induced by MCD, together with AAV8-
Cel or AAV8-null treatment (
Fig. 7(a)). Treatment with AAV8-
Cel significantly decreased hepatic TG, TC, and TBARS, compared to the control group treated with AAV8-null (all
P < 0.01,
Fig. 7(b)). Mice treated with AAV8-
Cel also exhibited reduced liver injury, as demonstrated by reduced serum ALT and AST (
Fig. 7(c)). Histological examination revealed that AAV8-
Cel-treated mice showed alleviated hepatic steatosis, inflammation, and ballooning (
Fig. 7(d)). The activity of NF-κB pathway was markedly reduced in CEL
-overexpressing mice (
Fig. 7(e)), accompanied by lowered macrophage infiltration (
Fig. 7(f)) and downregulated mRNA expressions of inflammatory genes (
Il6 and
Tnfα) (
Fig. 7(g)). Moreover, the protein expression of FASN was decreased in CEL-overexpressing mice fed with the MCD (
Fig. 7(e)). Collectively, our consistent findings from the two diet-induced MASH mouse models indicate that CEL-overexpression could represent a potential therapeutic approach for alleviating MASH development.
4. Discussion
In this study, we provide mechanistic insights into the protective role and function of CEL against MASH in mice. MASH is widely recognized as a metabolic disorder, and accumulating evidence suggests that lipid accumulation is central to its pathogenesis [
17]. However, despite its increasing global incidence, our understanding of MASH development, especially concerning the role of host genetics in MASH, remains largely unclear. While previous investigations have identified correlations between MASH and genetic alterations or single nucleotide polymorphisms [
18], the exact molecular interactions between these genetic factors and metabolic dysregulation remain elusive. In this study, we identified a notable decrease in CEL protein levels in the liver of in diet-induced MASH mice (
Fig. 1), indicating a potential protective role of CEL against MASH.
Although CEL is primarily synthesized in the pancreas, it is also expressed in the liver [
10], with its presence mainly observed in hepatocytes [
19], [
20], [
21]. A previous study indicated the presence of CEL in specific endosomal compartments within hepatocytes, suggesting that hepatic-derived CEL could be directly secreted by liver cells [
22]. The presence of CEL on the surface and within endosomal compartments of liver cells suggests its potential involvement in the hepatic secretion-capture pathway, which plays a role in lipoprotein metabolism [
23]. Indeed, hepatic CEL is crucial for hepatic lipoprotein metabolism by facilitating the selective uptake of high-density lipoprotein (HDL)-derived cholesterol esters [
24]. Our findings demonstrate that hepatic CEL directly binds to FASN, leading to a reduction in FASN SUMOylation. This, in turn, promotes FASN degradation through the proteasome pathway, thereby impeding the progression of MASH. These suggest that hepatic CEL could exert anti-MASH effects by modulating hepatic lipid metabolism.
In contrast, while our results suggest the presence of intestinal CEL, it is apparent that CEL expression in the intestine is not intrinsic but likely originates from the pancreas [
21]. Specifically, enterocytes may uptake CEL from pancreatic juice in conjugation with the chaperone protein GRP94 through an endocytic pathway, enabling CEL adherence to the duodenal epithelium and leading to its expression in the intestinal lumen [
25]. Consistent with our results, a previous study reported a positive signal of CEL protein in the intestine despite the absence of CEL mRNA expression [
21]. In this study, our primary focus was on CEL in the liver, where we examined its role and mechanism in MASH.
Specifically, we established two mouse models of diet-induced MASH with hepatocyte-specific
Cel knockout, confirming that CEL deficiency markedly accelerated MASH development (
Fig. 2). These findings are consistent with our previous research, where we identified recurrent mutations of
Cel in obese mice, and observed that knockdown of
Cel resulted in a significant accumulation of cholesterol ester in hepatocytes
in vitro [
11]. Additionally, other studies have reported insignificant impacts on insulin sensitivity and serum cholesterol or lipids levels in mice with non-specific
Cel knockout [
26], [
27]. However, our consistent results from two different MASH mouse models provide solid evidence of the notable acceleration of disease progression upon hepatocyte-specific CEL deficiency. Taken together, we propose that the anti-steatosis effect of CEL may be tissue-specific, with its primary manifestation observed in the liver during MASH development.
We then conducted
in vitro experiments to elucidate the mechanism of CEL in MASH development. Our results confirmed CEL could regulate steatosis and inflammation in mouse hepatocytes (
Fig. 3). Through mass spectrometry, we identified a novel interaction between CEL and FASN, a crucial enzyme involved in
de novo lipogenesis often upregulated in MASLD/MASH [
28]. Specifically, CEL directly bound to FASN, leading to a reduction in the SUMOylation of FASN (
Fig. 4). Consequently, this promoted the proteasomal degradation of FASN, leading to a decrease in FASN-related lipogenesis. Furthermore, we expanded our study by performing
in vitro functional experiments, demonstrating that both FASN siRNA and chemical inhibitors could counteract the MASH-promoting effects induced by
Cel knockdown (
Fig. 5). Studies have shown that SUMOylation regulates MASH pathogenesis. For example, SUMOylation of SREBP1c represses its transcriptional activity and inhibits lipid production [
29], while impaired SUMOylation of the nuclear receptor liver receptor homolog 1 promotes MASH by inducing the expression of oxysterol binding protein-like 3 [
30]. Acetylation of farnesoid X receptor has also been reported to block its interaction with the SUMO ligase protein inhibitor of activated signal transducer and activator of transcription (STAT) Y (PIASy), inhibiting its SUMO2 modification, and activating inflammatory genes in obese mice [
31]. Collectively, our results provide additional evidence supporting the involvement of SUMOylation in the development of MASH.
MASH lacks approved pharmacological treatments, posing a significant global concern due to its increasing prevalence and potential progression to liver malignancy [
32]. We therefore investigated the therapeutic potential of targeting CEL against MASH development by generating AAV8 vectors that could overexpress
Cel specifically in mouse livers. Our results showed that hepatocyte-specific CEL overexpression effectively mitigated disease phenotypes in mice with CD-HFD-induced MASH (
Fig. 6). These therapeutic effects were verified in another mouse model of MCD-induced MASH (
Fig. 7). Several clinical trials have been conducted using recombinant AAV vectors against liver disease, predominantly focusing on Phase I/II primarily targeting inherited metabolic liver disorders [
33]. The application of recombinant AAV particles carrying specific DNA sequences for therapeutic purposes has emerged as a highly promising and safe approach for gene therapy.
In the context of MASLD, AAV-based gene therapy has predominantly been explored in animal experiments, despite demonstrating therapeutic potential [
34], [
35]. Amidst the scarcity of approved treatment options and the growing body of evidence linking gene mutations to MASLD, the utilization of gene therapy with AAV vectors presents a promising solution for MASLD treatment. However, a thorough assessment of the efficacy of ongoing AAV studies is essential. While preliminary findings are encouraging, it is crucial to evaluate the translational impact of these trials from animal models to human clinical settings to establish the clinical promise of AAV-based interventions for MASLD. Moreover, the unique hepatocyte-specific anti-steatosis effect of CEL highlights its potential as a robust therapeutic intervention. Our exploration of AAV-mediated CEL overexpression in MASH treatment could introduce an innovative avenue to gene therapy targeting MASH. Aligned with numerous previous studies on genetic mutations implicated in MASLD, our findings contribute positively to the outlook for AAV investigations in MASLD and emphasize the need for meticulous clinical validation.
5. Conclusions
In summary, our study identified that hepatocyte-specific CEL deficiency is associated with the acceleration of MASH development, confirming the anti-steatosis role of CEL against MASH. Our mechanistic investigation revealed that CEL protects against MASH by directly targeting the SUMOlyation of FASN. Additionally, we highlighted the therapeutic potential of hepatocyte-specific CEL overexpression as a novel and robust strategy for MASH treatment. Overall, this study presents a promising direction for the development of targeted therapies against MASH, a prevalent liver disease.
Acknowledgments
This work was supported by the National Natural Science Foundation of China (82222901, 82103355, and 82272619), the Innovation and Technology Fund—Guangdong-Hong Kong Technology Cooperation Funding Scheme (GHP/086/21GD), the Research Grants Council (RGC) Theme-based Research Scheme (T12-703/19-R), the Research Grants Council-General Research Fund (14117422 and 14117123), the Health and Medical Research Fund, Hong Kong (08191336 and 07210097), the CUHK Research Startup Fund (FPU/2023/149), and the Natural Science Foundation of Fujian Province (2020J01122587).
We thank Dr. Olabisi Oluwabukola Coker for language editing.
Compliance with ethics guidelines
Yang Song, Wei Zhong, Harry Cheuk-Hay Lau, Yating Zhang, Huayu Guan, Mingxu Xie, Suki Ha, Diwen Shou, Yongjian Zhou, Hongzhi Xu, Jun Yu, and Xiang Zhang declare that they have no conflict of interest or financial conflicts to disclose.
PDF
(9719KB)
