SAE1 knockdown attenuates hepatic fibrosis by promoting ubiquitination and proteasomal degradation of TGF-β receptors

Abstract

Objective

Liver fibrosis is characterised by excessive accumulation of extracellular matrix and can ultimately progress to cirrhosis and malignant transformation. Activation of hepatic stellate cells (HSCs), largely driven by transforming growth factor-beta (TGF-β) signalling, is a central event in fibrogenesis. SAE1 is a subunit of the small ubiquitin-like modifier-activating enzyme E1; its role in liver fibrosis and HSC activation remains unclear. This study aimed to investigate the regulatory function and underlying mechanism of SAE1 in hepatic fibrogenesis.

Methods

A carbon tetrachloride (CCl₄)-induced mouse model of liver fibrosis was established to assess SAE1 expression in vivo. SAE1 was silenced using an adeno-associated virus serotype 6-mediated approach. In vitro, primary HSCs and TGF-β-treated LX-2 cells were used to evaluate the effects of SAE1 knockdown on HSC activation and profibrotic gene expression. The activity of the TGF-β/SMAD signalling pathway and the stability of TGF-β receptors were also examined.

Results

SAE1 expression was significantly increased in fibrotic livers. SAE1 knockdown attenuated HSC activation and downregulated profibrotic genes. Mechanistically, SAE1 silencing promoted ubiquitin-proteasome-dependent degradation of TGF-β receptor I/II, thereby reducing SMAD2/3 phosphorylation. In vivo, Sae1 knockdown alleviated CCl₄-induced liver fibrosis.

Conclusion

SAE1 contributes to liver fibrosis by maintaining TGF-β receptor stability and enhancing TGF-β/SMAD-driven HSC activation. These results highlight SAE1 as a potential target for antifibrotic therapy.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• SAE1 is a core subunit of the small ubiquitin-like modifier (SUMO)-activating enzyme E1 complex and is involved in multiple cellular processes through SUMOylation. Hepatic stellate cell (HSC) activation is a central event in liver fibrogenesis and is largely driven by transforming growth factor-beta (TGF-β)/SMAD signalling, but the role of SAE1 in this process remains poorly understood. Although SUMOylation has been implicated in the regulation of fibrotic signalling, whether SAE1 contributes to hepatic fibrosis and how it may regulate TGF-β signalling have not been clearly established. Therefore, this study was undertaken to define the role of SAE1 in HSC activation and liver fibrosis and to investigate its underlying molecular mechanism.

WHAT THIS STUDY ADDS

• This study identifies SAE1 as a positive regulator of HSC activation and hepatic fibrosis. SAE1 depletion reduces TGF-β/SMAD signalling by promoting ubiquitin-dependent, proteasome-mediated degradation of TβRI/II, whereas SAE1 overexpression enhances TGF-β signalling and profibrotic responses. C-immunoprecipitation assays further demonstrate an association between SAE1 and TGF-β receptor I/II (TβRI/II), supporting a receptor-associated mechanism underlying the regulation of receptor stability. Importantly, the distinct effects of SAE1 and UBA2 depletion suggest that the profibrotic function of SAE1 is not dependent on the canonical SAE1/UBA2 SUMO E1 complex.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• These findings identify SAE1 as a potential molecular target for modulating TGF-β/SMAD signalling and HSC activation in liver fibrosis. The study provides a basis for further investigation of SAE1-targeted strategies as potential antifibrotic interventions. The association between SAE1 and TβRI/II also provides a new direction for studying the regulation of TGF-β receptor stability during fibrogenesis. Further studies, particularly in human cohorts and disease models, are needed to determine the therapeutic relevance and safety of targeting SAE1.

Introduction

Liver fibrosis is a pathological consequence of chronic liver injury, marked by excessive extracellular matrix (ECM) deposition, persistent inflammation, and progressive parenchymal damage.^{1 2} Without intervention, this fibrotic process can advance to cirrhosis and ultimately hepatocellular carcinoma (HCC).^{3 4} During chronic liver injury, hepatic stellate cells (HSCs) undergo activation and transdifferentiate into contractile,^{5 6} myofibroblast-like cells with enhanced ECM-secreting properties, serving as the principal driver of liver fibrogenesis.⁷ The resulting aberrant ECM deposition, particularly collagen and fibronectin, leads to architectural distortion of the liver, functional impairment and the progression of fibrotic disease.^8–11

Transforming growth factor-β (TGF-β) is a well-established profibrotic cytokine in the liver and a vital mediator of HSC activation.^{12 13} Upon liver injury, TGF-β expression is markedly increased, resulting in activation of SMAD2 and SMAD3 through the TGF-β receptor I (TGFBR1; TβRI) and TGF-β receptor II (TGFBR2; TβRII).^14–16 Phosphorylated SMAD2 and SMAD3 subsequently associate with SMAD4 to form heteromeric complexes that translocate to the nucleus, where they promote profibrotic gene expression.¹⁷ Several inhibitors of the TGF-β/SMAD pathway have been explored as potential treatments for hepatic fibrosis.^{18 19} TGF-β drives fibrosis by increasing collagen and other extracellular matrix production in fibroblasts while limiting matrix breakdown, which distorts tissue architecture and impairs liver function.^{12 20} Given its central role, therapeutic inhibition of the TGF-β signalling pathway is a major focus in the development of antifibrotic treatments.^{21 22}

SAE1 was originally identified as a subunit of the E1 activating enzyme (SAE1/UBA2) involved in the small ubiquitin-like modifier (SUMO) modification (SUMOylation), which regulates diverse cellular processes, including transcriptional control, stress response and tumourigenesis.^{23 24} Notably, SUMOylation modulates TGF-β signalling through modification of key components such as SMAD4, thereby influencing tumour epithelial-mesenchymal transition and cell migration.^{25 26} Accumulating evidence suggests that inhibition of SUMOylation may represent a promising strategy for fibrosis intervention.²⁷ However, the role of SAE1 in liver fibrosis remains largely unexplored. In this study, we established a carbon tetrachloride (CCl₄)-induced mouse model of liver fibrosis and found that SAE1 expression increased in fibrotic liver tissue. Using TGF-β-activated HSCs or a spontaneous activation model of primary mouse HSCs in vitro, the effects of SAE1 knockdown on TGF-β signalling cascades and HSC activation were investigated. Furthermore, adeno-associated virus serotype 6 (AAV6)-mediated delivery of short hairpin RNA (shRNA) targeting Sae1 was used to achieve Sae1 silencing in fibrotic mice,²⁸ and profibrotic gene expression and key components of the TGF-β signalling axis were evaluated. The results presented here showed that depletion of SAE1, but not UBA2, significantly reduced TGF-β receptor I/II (TβRI/II) protein levels, leading to attenuation of TGF-β/SMAD signalling and suppression of profibrotic gene expression. Collectively, these findings demonstrated that SAE1 served as a pivotal regulator of HSC activation and liver fibrogenesis, functioning independently of the canonical SAE1/UBA2 SUMO E1 complex.

Methods

Animal experiments

Male C57BL/6 mice (6–8 weeks old, weighing 20–25 g) were obtained from GemPharmatech (licence no. SCXK (Su) 2023–0009). All animal experiments were approved by the Animal Ethics Committee of Anhui Medical University (approval no. LLSC20242520) and conducted in accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals.

For the AAV6-mediated Sae1 knockdown experiment, 24 mice were randomly assigned to the experimental groups (n=6 per group). Each legend shows the corresponding number of samples. Group sizes were determined based on preliminary experiments and previous studies using comparable liver fibrosis models. Chronic liver fibrosis was induced by intraperitoneal administration of CCl₄ (1 µL/g body weight) once every 3 days for 6 weeks. To knock down Sae1, mice received AAV6-mediated shRNA via tail-vein injection 2 weeks before the initiation of CCl₄ treatment. At the end of the experiment, mice were euthanised by carbon dioxide (CO₂) inhalation in accordance with institutional animal welfare guidelines, and death was confirmed before tissue collection. All procedures were performed in accordance with the institutional guidelines. Liver tissues were then harvested for subsequent analyses. No unexpected adverse events were observed during the study.

All animals were randomly assigned to the experimental groups before intervention. The investigator responsible for drug administration and tissue collection was blinded to group allocation. Outcome assessments, including histological scoring and biochemical analyses, were performed by a separate investigator who was also blinded to group identity.

Isolation of primary mouse HSCs

C57BL/6 mice were anaesthetised, and the liver was perfused through the portal vein. Perfusion was performed in two sequential steps: initial perfusion with prewarmed ethylene glycol tetraacetic acid buffer to flush out blood, followed by type IV collagenase to digest the tissue. The cell suspension was centrifuged at 50×g for 5 min at 4 °C and washed several times by low-speed centrifugation to remove hepatic parenchymal cells and debris. The supernatant containing non-parenchymal cells was then collected and centrifuged at 200×g at 4°C to pellet the cells. The pellet was resuspended in 20% OptiPrep and separated using a discontinuous density gradient prepared in a 15 mL tube (bottom to top: 20% OptiPrep cell suspension, 11.5% OptiPrep and Hank's Balanced Salt Solution (HBSS)). The gradient was centrifuged at 1,800×g for 17 min at 4°C. HSCs were collected from the interface between the 11.5% OptiPrep layer and HBSS using a Pasteur pipette.⁸ Cells were plated on collagen-coated dishes and cultured in high-glucose Dulbecco's modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS) at 37°C in a 5% CO₂ incubator.

Cell culture and treatment

LX-2 and HEK293T cells were obtained from the Chinese Academy of Sciences (Shanghai, China). LX-2 cells were cultured in DMEM (HyClone, Utah, USA) with 2% FBS (Gibco, California, USA),²⁹ and HEK293T cells were cultured in DMEM with 10% FBS. Cells were maintained at 37°C in a 5% CO₂ incubator. For experiments, LX-2 cells were plated in six-well plates at 5×10³ cells/mL. After 12 hours, cells were transfected with small interfering RNA (siRNA) using Lipofectamine 3000 (Thermo Fisher Scientific, L3000015) to reduce expression of the target genes. 24 hours later, cells were treated with TGF-β1 (10 ng/mL) or left untreated.

Plasmid construction

The shRNA sequences in the pLKO.1-shRNA constructs were obtained from the Broad Institute and inserted into the lentiviral vector at the AgeI/EcoRI sites. The Sae1 single-guide RNA (sgRNA) sequences were designed using the online tool Benchling and inserted into the lenti-guideV2 vector using the restriction enzymes BfuAI and BsmBI. For AAV vector construction, the U6 promoter-driven shRNA expression cassette from the pLKO.1-shCtrl-puro and pLKO.1-shSae1-puro plasmids was inserted into the MfeI site of the pAAV-CAG-GFP vector. The sequences of sgRNA and shRNA are listed in online supplemental table S1.

AAV6 vector production

HEK293T cells received simultaneous transfection of AAV vectors and helper plasmids (pHelper, pREPCAP6) in an equimolar ratio (1:1:1). Following a 72-hour transfection period, adenoviral particles were successfully generated.^{28 30}

Histological staining

Mouse liver tissue was fixed with 4% paraformaldehyde and embedded in paraffin. Then the tissue was cut into 4 µm sections and stained with H&E, Masson’s trichrome and Sirius Red. After staining, the slides were sealed with neutral resin. The liver tissue injury was observed and analysed under a light microscope.

Hydroxyproline content

Liver samples were homogenised in 0.9% saline, and the liver hydroxyproline content was measured using a standard commercial kit according to the manufacturer’s instructions (Nanjing Jiancheng Bioengineering Institute, Nanjing, China).

Biochemical analysis

Blood samples were incubated at room temperature for 30 min to allow for complete clotting and then centrifuged at 3000 rpm for 15 min to obtain the serum, which was subsequently stored at −80°C until analysis. Serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) activities were determined using commercial kits from the Nanjing Jiancheng Bioengineering Institute (ALT, Cat. No. C009-2-1; AST, Cat. No. C010-2-1), according to the manufacturer’s instructions, which are based on the Reitman-Frankel colourimetric method.

Western blot analysis

Cells were lysed in lysis buffer on ice for 30 min, and tissues were homogenised in the same buffer. Lysates were centrifuged at 12,000×g for 15 min at 4°C, and the supernatants were collected for protein analysis. Protein concentration was measured using an enhanced bicinchoninic acid assay kit (Beyotime, Cat. No. P0010) by reading absorbance at 562 nm. Equal amounts of protein (30–50 μg per lane) were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to polyvinylidene fluoride membranes. Membranes were blocked with 5% skim milk at room temperature for 1 hour, washed three times and proteins were detected using an enhanced chemiluminescence kit.

RNA extraction and quantitative reverse transcription-PCR

Total RNA was extracted from liver tissue and cell lines using TRIzol reagent (Invitrogen, #15596018) according to the manufacturer’s instructions. Reverse transcription was performed using 1 µg of total RNA and the HiScript III RT SuperMix for quantitative PCR (qPCR) (genomic DNA wiper) kit (Vazyme, R323-01). qPCR was carried out in a 20 µL reaction volume using 2× HyperMB Universal SYBR Green qPCR Master Mix (BBI, #B690016). The reaction system contained 10 µL of 2×Master Mix, 0.4 μM forward and reverse primers and 1 μL of complementary DNA template. The reactions were run under the recommended conditions on a real-time PCR instrument. Glyceraldehyde 3-phosphate dehydrogenase was used as the internal control, and the relative expression of the target gene was calculated using the 2-ΔΔCt method. Primer sequences are listed in online supplemental table S2.

Immunofluorescence

LX-2 cells (5×10⁴) were plated on sterile coverslips in 24-well plates. After siRNA transfection and TGF-β1 treatment, cells were fixed in 4% paraformaldehyde and permeabilised with 0.1% Triton X-100 in phosphate-buffered saline. Samples were blocked with 10% goat serum for 30 min and incubated with primary antibodies at 4°C overnight. Staining was performed using a panoramic immunofluorescence kit (Panovue, 10522100050) following the manufacturer’s instructions. For tissue sections, slides were deparaffinised, rehydrated through graded ethanol and heated in 1% sodium citrate buffer (pH 6.0) for antigen retrieval. Sections were then blocked with 10% goat serum for 30 min and incubated with primary antibodies at 4°C overnight. Staining was performed with the same kit, and images were captured using a fluorescence microscope.

Cell-surface protein biotinylation

To assess TGFBR2-HA internalisation in LX-2 cells, we generated a stable line expressing HA-tagged TGFBR2 by lentiviral transduction (pCDH-TGFBR2-HA). Before the assay, cells were treated with TGF-β (10 ng/mL) for 24 hours. Cell-surface proteins were labelled at 4°C for 15 min using the cleavable, membrane-impermeable biotin reagent Sulfo-NHS-SS-Biotin (Cell Surface Biotinylation and Isolation Kit; Beyotime, Cat# P0660S), which restricts labelling to the plasma membrane. Cells were assigned to three groups: (1) background control (kept at 4°C to prevent internalisation), (2) endocytosis control and (3) SAE1 knockdown (endocytosis). For internalisation, cells in groups (2) and (3) were incubated at 37°C in an incubator for 30 min. After incubation, the remaining surface biotin was removed by treating intact cells with Mesna (50 mM) at 4°C for 20 min, followed by iodoacetamide (5 mg/mL) at 4°C for 5 min to block reformation of disulphide bonds. Cells were then lysed, and biotin-labelled proteins protected from Mesna cleavage (internalised fraction) were enriched using streptavidin magnetic beads.³¹

Statistical analysis

All experimental results are presented as mean±SD. Statistical comparisons were performed using two-sample t-tests or one-way analysis of variance, followed by Sidak’s post hoc test. A p-value <0.05 was considered statistically significant.

Results

SAE1 expression is upregulated during the progression of liver fibrosis

To evaluate the role of SAE1 in liver fibrosis, its expression was evaluated in a CCl₄-induced mouse liver fibrosis model. Compared with control mice, CCl₄-treated mice exhibited marked collagen deposition, as demonstrated by Sirius Red and Masson’s trichrome staining (online supplemental figure S1A). Consistently, hepatic hydroxyproline content was significantly elevated, accompanied by increased serum ALT and AST levels (online supplemental figure S1B-D), confirming successful induction of liver fibrosis and associated liver injury. Notably, SAE1 expression was significantly upregulated in fibrotic liver tissues at both the transcript and protein levels, as evidenced by quantitative reverse transcription-PCR and western blot assays (online supplemental figure S1E, H). Furthermore, SAE1 expression was positively correlated with established fibrotic markers, including Timp1, Acta2, Col1a1, Col1a2 and Col3a1 (online supplemental figure S1F-G). Together, these results suggest that SAE1 is associated with liver fibrogenesis.

SAE1 knockdown suppressed HSC activation in vitro

To clarify the role of SAE1 in HSC activation, LX-2 cells and primary mouse HSCs were treated with TGF-β. Knockdown of SAE1 was achieved in LX-2 cells and primary mouse HSCs using siRNA and lentiviral-mediated shRNA, respectively. In both types of cells, SAE1 depletion attenuated the effect of TGF-β on the elevation of profibrotic markers, including α-SMA, collagen type I and collagen type III (figure 1A–D). During culture-induced activation of primary mouse HSCs, SAE1 knockdown further reduced the expression of key fibrosis-related genes (figure 1E,F). In contrast, under the same experimental conditions, knockdown of UBA2 did not result in comparable changes in the aforementioned liver fibrosis-related indicators (online supplemental figure S2A,B). This divergence suggests that SAE1 may regulate HSC activation and liver fibrosis through a mechanism that is independent of the canonical SUMOylation pathway.

Figure 1. SAE1 knockdown reduces HSC activation. (A–B) Transcriptional and translational profiling via qRT-PCR and immunoblotting of SAE1 and fibrotic mediators in LX-2 cells following 24-hour TGF-β stimulation (10 ng/mL). (C–D) Primary mouse HSCs transduced with lentiviral vectors encoding shCtrl or shSae1. Cellular protein/RNA was collected after 48 hours of TGF-β stimulation for qRT-PCR and immunoblot evaluation of SAE1 and fibrosis-associated genes. (E–F) Primary mouse HSCs were cultured for the specified time intervals (first, fourth and eighth days). mRNA and protein expression were measured separately. The values presented are the mean±SD of three independent experiments. (*p<0.05, **p<0.01, ***p<0.001, ****p<0.0001). α-SMA, alpha-smooth muscle actin; HSC, hepatic stellate cell; mRNA, messenger RNA; qRT-PCR, quantitative reverse transcription PCR; SAE1, subunit of the E1 activating enzyme; TGF-β, transforming growth factor-beta.

SAE1 knockdown impairs the activation of the TGF-β signalling pathway

To explore the mechanism by which SAE1 regulates HSC activation, we performed RNA sequencing in LX-2 cells following SAE1 knockdown. Transcriptome analysis showed that genes downregulated upon SAE1 knockout under TGF-β stimulation were enriched in extracellular matrix organisation, collagen-containing matrix, integrin-mediated cell adhesion, leucocyte migration and angiogenesis. Several TGF-β-related gene sets, including TGF-β1 signalling, also showed significant negative enrichment, indicating an overall weakening of the TGF-β transcriptional response in the SAE1 knockdown group (figure 2A,B). In line with these results, SAE1 knockdown in TGF-β-treated LX-2 cells and primary mouse HSCs reduced the protein levels of TβRI/II, accompanied by decreased phosphorylation of SMAD2 and SMAD3 (figure 2C,D), with p-SMAD2/3 levels quantified relative to total SMAD2/3 levels in online supplemental figure S3 A-D). Immunofluorescence staining confirmed diminished expression of TβRI/II in SAE1-deficient cells (figure 2E). In contrast, qPCR showed no change in TGFBR1 or TGFBR2 mRNA (figure 2F). Together, these data suggest that SAE1 maintains TGF-β receptor protein levels through a post-transcriptional mechanism, potentially through modulation of protein stability or turnover rather than transcriptional regulation.

Figure 2. SAE1 depletion attenuates TGF-β/SMAD signalling. (A–B) The gene ontology and gene set enrichment analysis of differentially expressed mRNAs in the control and SAE1 knockout groups. (C–D) LX-2 and primary mouse HSCs were treated with 10 ng/mL TGF-β for 24 hours and 48 hours or phosphate-buffered saline and examined for SAE1, SMAD2/3, p-SMAD2/3 and TβRI/II using immunoblotting. (E) Fluorescence micrographs depicting immunofluorescence labelling of TβRI/II (green) in TGF-β-treated LX-2 cells (10 ng/mL, 24 hours). Scale bar: 50 μm. (F) TGFBR1/2 expression in TGF-β-stimulated LX-2 cells was examined using qRT-PCR. The values presented are the mean±SD of three independent experiments (*p<0.05, **p<0.01, ****p<0.0001). DAPI, 4′,6-diamidino-2-phenylindole; HSC, hepatic stellate cell; mRNA, messenger RNA; n.s., non-significant; qRT-PCR, quantitative reverse transcription PCR; SAE1, subunit of the E1 activating enzyme; TGF-β, transforming growth factor-beta; TβRI/II, TGF-β receptor I/II.

SAE1 overexpression exacerbates HSC activation

To investigate the role of SAE1 in HSC activation, we generated LX-2 cells stably overexpressing SAE1 using a pCDH-SAE1 lentiviral vector, detecting overexpression of SAE1 (online supplemental figure S6A). Under TGF-β stimulation, SAE1 overexpression significantly upregulated the protein and mRNA levels of α-SMA, COL1A1, COL1A2 and COL3A1 compared with the empty vector control (figure 3A,C). Mechanistically, SAE1 overexpression increased TβRI/II protein levels and subsequently enhanced SMAD2 and SMAD3 phosphorylation (figure 3C). These results indicate that SAE1 potentiates TGF-β signalling to promote a profibrotic activation state in HSCs.

Figure 3. Knockdown of TGFBR1/2 restores fibrotic marker levels increased by SAE1 overexpression. (A) LX-2 cells transduced with lentiviral vectors encoding pCDH-EV or pCDH-SAE1-Flag. Cellular RNA was collected after 24 hours of stimulation with TGF-β (10 ng/mL) for qRT-PCR evaluation of fibrosis-associated genes. (B) LX-2 cells were transduced with an SAE1-Flag overexpression lentivirus and transfected with small interfering RNAs targeting TGFBR1/2, followed by TGF-β stimulation. The mRNA levels of COL1A1, COL1A2 and COL3A1 were measured by qRT-PCR. (C) Western blot analysis of fibrosis-associated proteins in SAE1-overexpressing LX-2 cells under TGF-β stimulation. (D) Protein levels of α-SMA, COL1A1 and COL3A1 were analysed by immunoblotting. The values presented are the mean±SD of three independent experiments (*p<0.05, **p<0.01, ***p<0.001, ****p<0.0001). α-SMA, alpha-smooth muscle actin; EV, empty vector; HSC, hepatic stellate cell; mRNA, messenger RNA; qRT-PCR, quantitative reverse transcription PCR; SAE1, subunit of the E1 activating enzyme; TGF-β, transforming growth factor-beta; TβRI/II, TGF-β receptor I/II; OE, overexpression.

To determine whether SAE1-mediated fibrogenesis is dependent on TGF-β receptors, we silenced TGFBR1 and TGFBR2 in SAE1-overexpressing LX-2 cells treated with TGF-β (online supplemental figure S6B,C). Knockdown of either receptor effectively attenuated the SAE1-driven increase in fibrosis markers (including α-SMA and COL1A1), restoring their levels to those comparable to those observed in TGF-β-treated empty-vector control cells (figure 3B,D). These findings indicate that SAE1 promotes fibrotic signalling in HSCs primarily by enhancing the TGF-β receptor-dependent pathway, identifying TGF-β receptor I/II as essential mediators of the profibrotic effects of SAE1.

SAE1 knockdown promotes proteasomal degradation of TβRI/II

To elucidate the mechanism by which SAE1 regulates TGF-β receptor protein levels to modulate TGF-β signalling, we assessed the stability of TGF-β receptors using a cycloheximide chase assay. Upon TGF-β stimulation, TβRI and TβRII exhibited significantly accelerated protein turnover in SAE1-depleted cells compared with controls (figure 4A), indicating that SAE1 deficiency promoted TGF-β receptor degradation. To identify the degradation pathway, cells were treated with the proteasome inhibitor MG132 or the lysosome inhibitor chloroquine. MG132, but not chloroquine, effectively blocked TGF-β receptor degradation induced by SAE1 knockdown (figure 4B), demonstrating that SAE1 prevents proteasome-dependent degradation of these receptors. Consistently, immunoprecipitated TβRI and TβRII showed increased polyubiquitination upon SAE1 depletion (figure 4C,D). Furthermore, receptor trafficking of TβRII was evaluated using surface biotin labelling followed by a streptavidin pull-down of internalised proteins. SAE1 knockdown markedly enhanced the internalisation of TβRII (figure 4E). Collectively, these data suggested that SAE1 sustained TGF-β signalling by limiting receptor internalisation and inhibiting ubiquitin-dependent proteasomal degradation of TβRI and TβRII. To further determine whether SAE1 interacts with TGF-β receptors, we performed co-immunoprecipitation (Co-IP) assays in HEK293T cells. SAE1 was detected in TβRI- and TβRII-immunoprecipitated complexes, and reciprocal Co-IP assays further confirmed the association of SAE1 with both receptors (online supplemental figure S4A-D). These findings provide evidence that SAE1 interacts with TGF-β receptors and support a receptor-associated mechanism by which SAE1 regulates TβRI/II protein stability.

Figure 4. SAE1 deficiency accelerates the protein degradation of TβRI/II. (A) Immunoblotting analysis for the protein levels of TβRI/II after treatment with CHX (150 µM) at the indicated time points in TGF-stimulated LX-2 cells. (B) TβRI/II expression profiles in control versus SAE1-depleted LX-2 cells following pharmacological treatment with CHX (150 µM) combined with MG132 (10 µM) or CQ (20 µM). (C–D) HEK293T cells transfected with plasmids of HA-Ub, Flag-TGFBR1 or Flag-TGFBR2 were subjected to MG132 treatment (10 µM, 10 hours), followed by co-immunoprecipitation and immunoblot analysis. (E) Perform biotinylation detection of cell surface TβRII on HA-TGFBR2-expressing LX-2 cells in the control group and the SAE1 knockdown group. The values presented are the mean±SD of three independent experiments (*p<0.05, **p<0.01). CHX, cycloheximide; CQ, chloroquine; DMSO, dimethyl sulfoxide; HSC, hepatic stellate cell; SAE1, subunit of the E1 activating enzyme; ns, non-significant; TGF-β, transforming growth factor-beta; TβRI/II, TGF-β receptor I/II; HA, HA-tag; IP, immunoprecipitation; IB, immunoblotting.

In vivo AAV6-mediated Sae1 silencing attenuates CCl₄-induced liver fibrosis in mice

To investigate whether Sae1 knockdown in HSCs affects fibrosis development, an AAV6-Enhanced Green Fluorescent Protein (EGFP)-shSae1 vector, which preferentially transduces HSCs, was injected via the tail vein, followed by 6 weeks of CCl₄ administration. Efficient and HSC-preferential delivery of the AAV6 vector was confirmed by the colocation of GFP and the HSC marker desmin in liver sections (online supplemental figure S5). Consistent with our in vitro findings, Sae1 knockdown significantly attenuated SMAD signalling in the liver of mice treated with CCl₄, as evidenced by reduced phosphorylation of SMAD2 and SMAD3 (figure 5A). Indeed, the expression levels of TβRI and TβRII were significantly decreased upon Sae1 depletion. qPCR confirmed a significant reduction in Sae1 mRNA levels in liver tissues, which correlated with a marked downregulation of the profibrotic genes, including Timp1, Acta2, Col1a1, Col1a2 and Col3a1 (figure 5B). Biochemical measurements revealed a significant reduction in serum ALT, AST and hepatic hydroxyproline levels in the AAV6-shSae1 group following CCl₄ treatment, demonstrating that Sae1 silencing mitigated liver injury and fibrosis (figure 5C). Furthermore, H&E, Sirius Red and Masson’s trichrome staining revealed that Sae1 knockdown markedly diminished CCl₄-induced liver injury and collagen deposition (figure 5D). Similarly, multiplex fluorescence imaging in the CCl₄ model of the Sae1 knockdown group, fibrosis and TβRI/II were reduced (figure 5E,F). Together, these results demonstrated that HSC-preferential Sae1 silencing alleviates CCl₄-induced liver fibrosis by suppressing SMAD activation and the associated fibrogenetic programme.

Figure 5. AAV6-mediated Sae1 silencing alleviates liver fibrosis in the CCl₄ mouse model. (A) Mice received a tail-vein injection of AAV6-EGFP-shCtrl or AAV6-EGFP-shSae1 for Sae1 knockdown. 2 weeks later, hepatic fibrosis was induced by CCl₄ administration for 6 weeks; the expression of Sae1, Acta2, Col1a1, Col1a2, Col3a1 and Timp1 using qRT-PCR and (B) the protein levels of SAE1, α-SMA, COL1A1, COL3A1, SMAD2/3, p-SMAD2/3 and TβRI/II were examined in fibrotic and control liver using immunoblotting (n=3 mice per group). (C) Histopathological evaluation via H&E, Masson’s trichrome and Sirius Red staining in CCl₄-induced fibrotic livers. Scale bar: 100 µm. (D) Serum ALT/AST concentrations and liver tissue hydroxyproline were assayed across experimental cohorts. (E and F) Multiplex fluorescence imaging of liver sections from experimental groups. Scale bar: 20 µm. Data represent mean±SD from n=6 mice (*p<0.05, **p<0.01, ***p<0.001, ****p<0.0001). AAV6, adeno-associated virus serotype 6; ALT, alanine aminotransferase; AST, aspartate aminotransferase; α-SMA, alpha-smooth muscle actin; CCl₄, carbon tetrachloride; DAPI, 4′,6-diamidino-2-phenylindole; mRNA, messenger RNA; qRT-PCR, quantitative reverse transcription PCR; SAE1, subunit of the E1 activating enzyme; TGF-β, transforming growth factor-beta; TβRI/II, TGF-β receptor I/II; EGFP, Enhanced Green Fluorescent Protein.

Discussion

Prior studies show that the expression level of SAE1 tends to be altered in a variety of tumours, suggesting its possible involvement in tumourigenesis and progression.³² For example, several studies have shown that the expression level of SAE1 is increased in some cancer types, such as HCC, breast cancer, gastric cancer, pancreatic cancer and lung cancer.^32–37 In breast cancer, higher SAE1 levels are linked to greater metastatic potential and worse survival.³⁴ In glioma models, SAE1 knockdown suppresses cell proliferation and reduces xenograft tumour growth.³⁸ In particular, in HCC, SAE1 expression is closely correlated with the tumour stage, metastatic phenotype and survival rate.^{32 33} However, the function and mechanism of SAE1 in the development of liver fibrosis remain unclear.

Interestingly, our data suggest that the profibrotic role of SAE1 may not be fully explained by its canonical function within the SAE1/UBA2 SUMO E1 complex. While SAE1 knockdown markedly suppressed HSC activation and TGF-β/SMAD signalling, knockdown of UBA2 did not produce comparable effects under the same experimental conditions. Given that SAE1 and UBA2 function as obligate partners in the SUMO E1-activating enzyme, this discrepancy raises the possibility that SAE1 may exert additional regulatory functions beyond global SUMOylation activity.³⁹ In this study, we systematically clarified the pathogenic role of SAE1 in liver fibrosis and linked its effects to TGF-β signalling.⁴⁰ SAE1 was increased in CCl₄-induced fibrotic mouse livers and in activated HSCs, and higher SAE1 levels were associated with more severe fibrosis. In TGF-β-stimulated HSC models, SAE1 knockdown reduced extracellular matrix production and lowered key fibrosis markers, including α-SMA, type I collagen and type III collagen, consistent with reduced HSC activation. RNA sequencing further supported an anti-fibrotic effect after SAE1 depletion and pointed to changes in the TGF-β pathway. Mechanistically, SAE1 knockdown decreased TβRI/II protein levels and reduced SMAD2/3 activation, indicating weaker TGF-β signalling. This finding suggests that SAE1 acts upstream by maintaining TGF-β receptor abundance and thereby promoting fibrotic progression. In TGF-β-stimulated LX-2 cells overexpressing SAE1, silencing TGFBR1 or TGFBR2 abolished the SAE1-driven increase in fibrosis markers, indicating that SAE1 acts through TGF-β receptors rather than downstream targets. We also found that SAE1 deficiency accelerated TβRI/II protein turnover in a proteasome-dependent manner,^{41 42} which likely explains the reduced receptor levels and dampened SMAD2/3 activation.⁴² We observed the same pattern in the CCl₄ fibrosis model, supporting an in vivo relevance. Together, these findings point to TGF-β receptor I/II as key mediators of SAE1 action and highlight receptor stability as a key point of regulation.

This study has limitations and leaves questions unanswered. Specifically, the way SAE1 controls TβRI/II remains to be defined at a molecular level. Clarifying whether SAE1 binds the receptors directly and which regions and residues are required will require targeted mutagenesis and structural studies. Although AAV6-mediated Sae1 silencing effectively reduced hepatic fibrosis, future studies using HSC-specific SAE1 conditional knockout models are required to further define the cell-specific role of SAE1 in fibrogenesis. Further work along these lines is needed to fully explain how SAE1 regulates TβRI/II stability and TGF-β signalling. While SAE1 expression was increased in our experimental models of liver fibrosis and HSC activation and in immunohistochemical staining of human fibrotic liver specimens (online supplemental figure S7), analysis of human activated HSC populations did not show a significant increase in SAE1 expression. This discrepancy likely reflects post-transcriptional regulation, species-specific regulation, variations in disease aetiology and fibrosis stage or differences in the cellular microenvironment between experimental models and human liver disease. Further studies involving larger human cohorts and single-cell approaches will be required to determine the precise clinical relevance of SAE1 in human HSC activation.

In summary, our data indicate that SAE1 supports TGF-β-SMAD signalling by preserving TβRI/II protein stability, which sustains stellate cell activation and extracellular matrix buildup. Reducing SAE1 increases TβRI/II ubiquitination and proteasome-dependent degradation,⁴³ weakens SMAD signalling and limits fibrotic changes. These results identify SAE1 as an upstream regulator of TGF-β receptor abundance and a potential target for anti-fibrotic therapy.

Conclusion

SAE1 loss reduced TGF-β signalling by lowering TβRI/II protein levels through ubiquitin-dependent, proteasome-mediated degradation. In cultured HSCs, SAE1 silencing weakened TGF-β-induced activation and decreased fibrosis markers. In mice, Sae1 knockdown alleviated CCl₄-induced fibrosis and liver injury. Mechanistically, SAE1 supported fibrotic signalling by maintaining TGF-β receptor protein stability in a receptor-dependent manner; when SAE1 was reduced, TβRI/II ubiquitination increased, receptor protein levels fell, and SMAD2/3 phosphorylation decreased. Together, these data identify SAE1 as an important driver of TGF-β-SMAD signalling and liver fibrosis and suggest that targeting SAE1 may have anti-fibrotic potential.

Data availability statement

Data are available upon reasonable request.