ABSTRACT
-
Background/Aims
Endoplasmic reticulum (ER) stress in hepatocytes plays a causative role in alcohol-associated liver disease (ALD). The incomplete inhibition of ER stress by targeting canonical ER stress sensor proteins suggests the existence of noncanonical ER stress pathways in ALD pathology. This study aimed to delineate the role of RAB25 in ALD and its regulatory mechanism in noncanonical ER stress pathways.
-
Methods
RAB25 activation was examined in liver samples from ALD patients and ethanol-fed mice. The interaction between RAB25 and GCN1 was confirmed through mass spectrometry and co-immunoprecipitation (Co-IP) assays in vitro. The role of RAB25/GCN1 in promoting noncanonical ER stress in ALD was assessed both in vitro and in vivo.
-
Results
RAB25 expression was upregulated and specifically accumulated on the ER in ALD. Mass spectrometry and Co-IP assays confirmed that RAB25 interacts with GCN1, thereby activating a noncanonical ER stress pathway that facilitates ALD progression. Further analysis revealed that RAB25 interaction with GCN1 inhibits K33-ubiquitination-mediated degradation of GCN1, promotes GCN2 phosphorylation, and subsequently activates ATF4-mediated ER stress. This activation modulates lipid metabolism, mitochondrial function, and inflammation, thereby facilitating ALD progression. Knockdown of RAB25 in hepatocytes inhibited ER stress activation and mitigated associated mitochondrial dysfunction, excessive lipid synthesis, and the exaggerated inflammatory response in an ALD model.
-
Conclusions
Our findings demonstrate a causal role for RAB25-GCN1 signaling in activating the ER stress pathway, which contributes to ALD progression. This pathway may provide a proof-of-concept target for treating ALD and associated metabolic disorders.
-
Keywords: Alcohol-associated liver disease; Endoplasmic reticulum stress; Ras-related protein 25; Ubiquitination
Study Highlights
• Using transcriptomic profiling and mechanistic approaches, our study demonstrates that RAB25 is significantly upregulated in ALD and is associated with disease progression. Specially, we found that RAB25 inhibits K33-mediated GCN1 degradation and subsequently promotes GCN2 phosphorylation, activating ATF4-mediated ER stress, driving mitochondrial dysfunction, lipid metabolism disorder and inflammation in vitro and in vivo.
• This study indicated the specific upregulation of RAB25 in ALD contributes to disease progression by activating ER stress responses. Suppression of RAB25 may alleviate ER stress and consequently mitigate ALD progression.
Graphical Abstract
INTRODUCTION
Alcohol-associated liver disease (ALD) is one of the most common causes of liver-related morbidity and mortality. It includes a spectrum of disease states ranging from simple liver steatosis and steatohepatitis to fibrosis and/or cirrhosis [
1]. Approximately 2.2 million people in the US were affected by alcohol-associated cirrhosis in 2017, and 90-day mortality rates ranged between 20–50% [
2]. Currently, pharmacological treatment for alcohol-associated hepatitis relies primarily on corticosteroids. However, the efficacy of these treatments is limited, and patients cannot obtain long-term survival benefits [
3]. Therefore, further understanding the pathogenesis and molecular mechanisms of ALD could provide profound benefits.
Alcohol consumption can directly induce the accumulation of unfolded or misfolded proteins in the endoplasmic reticulum (ER) lumen, which activates the unfolded protein response (UPR), a state referred to as ER stress [
4]. Classic ER stress sensor proteins include protein kinase (PKR)-like endoplasmic reticulum kinase (PERK), inositol-requiring enzyme 1α (IRE1α), and activating transcription factor 6 (ATF6). However, simple inhibition of ER stress sensor proteins cannot completely block ATF4 activation and associated ER stress damage in ALD, which implies that alternative pathways might exist for the transmission of ER stress signals in ALD pathology [
5]. On the other hand, noncanonical ER stress responses intersect with components of the UPR to drive function in the face of stress. Researchers have demonstrated that AGGF1 blocks ER stress-induced apoptosis in a heart failure model through a noncanonical pathway [
6]. However, the role of noncanonical ER stress in ALD remains unclear.
Rab proteins, the largest family of small GTPases, play essential roles in regulating cellular activities [
7]. RAB8a is a mitochondrial receptor that regulates lipid droplets (LDs) in skeletal muscle [
8]. Considering the involvement of the Rab GTPase family in lipid regulation and the crucial role of the ER in lipid synthesis, we investigated the potential contribution of RAB25 to ER stress in ALD.
Here, we employed single-cell database analysis to determine that RAB25 is an important stress protein that is upregulated in ALD. Mechanistically, we discovered that RAB25 interacts with GCN1, thereby inhibiting K33-mediated degradation of GCN1 and subsequently promoting GCN2 phosphorylation, activating ATF4-mediated ER stress, and facilitating the progression of ALD through the modulation of lipid metabolism, mitochondrial function, and inflammation. Overall, we demonstrated a causal role of RAB25-GCN1 signaling in activating the ER stress pathway to facilitate the progression of ALD, which may reveal a potential target for ALD treatment.
MATERIAL AND METHODS
RESULTS
RAB25 is upregulated in ALD patients and associated with disease progression
We employed single-cell transcriptomes (integrated from the GEO accessions GSE115469 and GSE136103) to clarify the genes possibly involved in ALD progression. By using uniform manifold approximation and projection (UMAP) visualization, we identified 11 clusters of cells (
Fig. 1A,
Supplementary Fig. 1A–
1F). On the basis of the aforementioned cell type categorization, hepatocytes were meticulously isolated for comprehensive analysis, culminating in the classification of hepatocytes into ten clusters (
Supplementary Fig. 1G–
1I). The odds ratio (OR) analysis revealed that ALD liver samples were strongly distributed in the 1-hepatocyte and 5-hepatocyte clusters (
Fig. 1B). Single-cell transcriptomic analysis revealed RAB25 predominantly expressed in hepatocyte clusters 1 and 5 (
Fig. 1C,
Supplementary Fig. 1J), which featured established ALD-associated genes like KRT7 and CCL4 (Supplement Data 1, 2).
We then analyzed the gene expression of ALD patients using GEO datasets, which ranged from healthy control, simple steatosis to fibrosis, and found that abnormally high RAB25 expression was closely associated with ALD progression (
Fig. 1D). Meanwhile, RAB25 protein levels were significantly higher in our collected ALD patient samples, which was related to ALD stage progression (
Fig. 1E,
1F,
Supplementary Table 1). Taken together, these findings revealed that RAB25 expression might be correlated with the progression of human ALD.
High RAB25 expression was strongly associated with the ER stress signaling pathway (
Fig. 2A,
Supplementary Fig. 2A). We found notable co-staining of RAB25 and ATF4 in hepatocytes from ALD patients compared with those from healthy controls. Immunohistochemical analysis also revealed that the increased RAB25 and ATF4 expression was associated with ALD stage progression (from mild steatosis to cirrhosis), suggesting that RAB25 may involve in ER stress in ALD patients (
Fig. 2A,
Supplementary Fig. 2B,
2C).
We then employed a murine model developed by the National Institute on Alcohol Abuse and Alcoholism (NIAAA, EtOH-fed) [
9] and found that the body weights of EtOH-fed mice were similar to those of pair-fed mice, but their liver-to-body weight ratios were increased. The serum alanine transaminase (ALT), aspartate transaminase (AST), serum triglyceride (TG) and total cholesterol (T-CHO) levels were greater in ethanol-fed mice (
Fig. 2B). Histopathological staining showed that ethanol administration induced hepatic steatosis, including ballooning degeneration of hepatocytes and mild-to-moderate lobular inflammation (
Fig. 2C,
Supplementary Fig. 2D,
2E). In summary, the EtOH-fed mouse model exhibited ALD-associated pathological features. Furthermore, compared with those in pair-fed mice, the protein and mRNA levels of RAB25 were significantly greater in EtOH-fed mice (
Fig. 2D,
Supplementary Fig. 2F). Transmission electron microscopy (TEM) showed notable dilation and membrane damage in the ER, along with swelling and rupture of mitochondrial cristae, indicating the occurrence of ER stress and mitochondrial damage in ethanol-fed mice (
Fig. 2E). Immunohistochemical staining and western blot showed markedly greater expression of both RAB25 and ER stress markers in ethanol-fed mice than in pair-fed mice (
Fig. 2F,
Supplementary Fig. 2G). We reexamined the data of multiple human liver diseases (non-alcoholic steatohepatitis, fibrosis, drug-induced liver injury, and primary sclerosing cholangitis) and found that RAB25 was specifically elevated in ALD patients and was associated with ER stress (
Supplementary Fig. 3A). We constructed mouse liver disease models comprising high-fat, high-calorie diet (HFCD), methionine-choline deficient (MCD), fibrosis, and bile duct ligation (BDL), metabolic dysfunction-associated ALD (MetALD) [
10] and found that RAB25 was elevated only in the ALD mouse model, suggesting that RAB25 might be a unique and important mediator of ER stress in ALD (
Supplementary Fig. 3B–
3M,
Supplementary Fig. 4). Given the alcohol-specific induction of RAB25, we investigated upstream regulators. Furthermore, single-cell analysis and experimental validation identified HIF1A as a key transcription factor to upregulated RAB25 and ER stress to drive disease progression (
Supplementary Fig. 5).
After excluding RAB25’s effect on alcohol-metabolizing enzymes expression，we focus on its role in modulating ER stress (
Supplementary Fig. 6). We treated primary human hepatocytes (PHHs) and HepG2 cells (which stably overexpressed human ADH and CYP2E1 for efficient alcohol metabolism)11 with ethanol at concentrations ranging from 0 to 200 mM (
Supplementary Fig. 7A,
7B). We observed dose-dependent increases in RAB25 and ER markers protein levels in PHHs and HepG2 cells following ethanol exposure (
Fig. 3A), and we selected 100 mM for subsequent experiments and observed a same trend in PHHs and primary mouse hepatocytes (PMHs) (
Supplementary Fig. 7C,
7D). Confocal microscopy revealed that RAB25 protein levels were increased and RAB25 accumulated on the ER in ethanol-stimulated PHHs, HepG2 cells and ALD patients, which indicated that RAB25 might involve in the process of ER stress (
Fig. 3A,
Supplementary Fig. 7E). To identify the key downstream targets of RAB25 in ER stress, we performed coimmunoprecipitation by mass spectrometry (MS). Coomassie brilliant blue staining of the gel revealed that cells treated with ethanol presented increased levels of RAB25-interacting proteins (
Fig. 3B,
Supplementary Fig. 7F). Among these, GCN1 is the most closely related to ER stress, and was subsequently chosen for further investigation. Subsequent analysis confirmed the RAB25-GCN1 interaction (
Fig. 3C) and their co-localization in ethanol-treated cells (
Fig. 3D,
Supplementary Fig. 7G,
7H).
To elucidate RAB25’s regulation of GCN1 and ER stress activation, we knocked down RAB25 in HepG2/AML12 cells using lentiviral shRNA. RAB25 knockdown attenuated ethanol-induced GCN1 upregulation. Given GCN1’s role in triggering GCN2-eIF2α phosphorylation, our findings revealed that RAB25 knockdown also decreased the levels of phosphorylated GCN2 (p-GCN2), p-eIF2α, and ATF4 in ethanol-stimulated cells (
Fig. 3E,
Supplementary Fig. 7I). We conducted rescue experiments and observed that adenovirus-mediated GCN1 overexpression restored the reduction in the GCN2-ATF4 signaling in HepG2 cells with stable RAB25 knockdown (
Fig. 3E). These results confirmed that RAB25 is involved in the regulation of GCN1, thereby influencing the subsequent GCN2-ATF4 signaling pathway. Furthermore, we evaluated the protein levels of PERK, IRE1α, and ATF6, and found that they were upregulated in response to ethanol. Surprisingly, knocking down RAB25 did not influence the upregulation of their protein expression, which indicated that RAB25 may activate ER stress independent of previous reports (PERK, IRE1α, and ATF6) (
Fig. 3E). We observed that silencing PERK could not completely abolish the alcohol-induced increase in p-Eif2α, ATF4 or CHOP expression (
Fig. 3F). Moreover, we rechecked the MS data and did not find any additional ER stress-related proteins that interact with RAB25 (
Supplementary Table 2). We further performed an immunoprecipitation assay and detected no interaction between RAB25 and PERK or eIF2α (
Fig. 3F), indicating that RAB25 may mediate ER stress in a GCN1/GCN2/ATF4 manner rather than in a PERK manner. We subsequently used tunicamycin, a well-established and mechanistically distinct chemical ER stress inducer that acts by inhibiting N-linked glycosylation, to induce ER stress. Surprisingly, we found that the suppression of RAB25 did not influence the ER stress caused by tunicamycin, indicating that RAB25 did not influence the stimulation of ER stress by PERK, IRE1α or ATF6 (
Fig. 3F,
Supplementary Fig. 7J). These findings suggest that RAB25 was specifically upregulated and localized in the ER and mediated ER stress through the GCN1/GCN2/ATF4 pathway in response to ethanol stimulation.
Next, we investigated the potential mechanism by which RAB25 regulates GCN1 protein expression. The protein expression levels of RAB25 and GCN1 were significantly greater in ALD patients and mice (
Fig. 4A,
Supplementary Fig. 8A). Interestingly, the mRNA level of GCN1 remained unchanged in ALD patients and mice (
Supplementary Fig. 8B–
8L). RAB25 knockdown attenuated GCN1 protein upregulation in ethanol-treated HepG2 cells without affecting the expression of the corresponding mRNAs (
Fig. 4A,
Supplementary Fig. 8M). The inconsistency between the protein and mRNA levels of GCN1 led us to hypothesize that RAB25 may employ protein posttranslational modifications to regulate GCN1 protein expression. Since ubiquitination and deubiquitination are crucial posttranslational modifications that regulate protein degradation, we performed a cycloheximide (CHX) and MG132 chase assays. RAB25 knockdown significantly accelerated GCN1 protein degradation (
Fig. 4B). The proteasome inhibitor MG132 reversed the decrease in GCN1 levels induced by RAB25 knockdown, indicating that RAB25 regulates GCN1 stability in a proteasome-dependent manner (
Fig. 4B). Moreover, the suppression of RAB25 increased the ubiquitination of GCN1. Overexpression of RAB25 reduced the K33-linked ubiquitination of GCN1 (
Fig. 4C). Taken together, these results suggest that RAB25 stabilizes GCN1 in a ubiquitin-dependent manner. We performed protein‒protein docking and found that RAB25 and GCN1 formed hydrogen bonds at specific amino acid residue sites, establishing a robust and stable protein docking model. To determine which part of RAB25 is necessary for interaction with GCN1, we employed several truncated forms of RAB25. Full-length RAB25 interacted well with GCN1, and truncation analysis revealed that a specific domain (amino acids 120-190) is required for the interaction between RAB25 and GCN1 (
Fig. 4D). The role of RAB25 in blocking GCN1 ubiquitination was abolished when the binding sites of RAB25 to GCN1 were blocked. Additionally, overexpression of RAB25 may partially reverse the decrease of GCN1 in HepG2 cells with stable knockdown of RAB25, whereas truncation of a specific domain (amino acids 120-190) may not have the same effect (
Fig. 4E). Taken together, these findings suggest that RAB25 interacts with GCN1 through a specific domain (amino acids 120-190), thereby further influencing ER stress.
Considering that RAB25 cannot directly affect protein ubiquitination, we speculated that a ubiquitin enzyme that mediates RAB25-regulated GCN1 expression exists. From the IP-MS database, we identified USP5 as a potential candidate. Notably, USP5 interacted with both RAB25 and GCN1, impeding the K33-linked ubiquitination and degradation of GCN1. RAB25 enhanced the interaction between USP5 and GCN1 (
Fig. 4F). Taken together, these findings suggest that RAB25 may promote GCN1 deubiquitination by recruiting USP5.
We administered an adeno-associated virus (AAV)-specific shRNA to mice via the tail vein to establish hepatocyte-specific RAB25-knockdown mice (
Fig. 5A,
Supplementary Fig. 9A,
9B). Body weight did not change in AAV-shRAB25 mice, but these mice presented lower liver-to-body weight ratios than did control mice in response to ethanol feeding. RAB25 knockdown partially reversed the increase in the serum ALT, AST, TG and T-CHO levels in ethanol-fed mice (
Fig. 5A). Moreover, H&E and Oil Red O staining revealed that LD accumulation and injury were lower in the livers of AAV-shRAB25 mice in the ethanol-fed groups (
Fig. 5B). TEM revealed that the expansion and membrane damage of the ER, as well as the degree of mitochondrial crista swelling and dissolution, were alleviated in the AAV-shRAB25 mice (
Fig. 5B). Additionally, RAB25 knockdown significantly decreased the upregulation of GCN1, p-GCN2, and downstream ER stress signaling pathway proteins (
Fig. 5C,
5D). However, RAB25 knockdown did not reverse liver injury or ER stress in other liver diseases (
Supplementary Fig. 9C–
9J). To further confirm the regulatory effects of RAB25 on GCN1, we utilized AAV serotype 8 (AAV8) to achieve hepatocyte-specific GCN1 overexpression (
Supplementary Fig. 9K). We observed that GCN1 overexpression significantly increased the liver-to-body weight ratios, serum ALT and AST levels, serum TG and T-CHO levels, and lipid accumulation compared to RAB25 knockdown in response to ethanol feeding (
Fig. 5E,
Supplementary Fig. 9L–
9O). Furthermore, we observed that GCN1 overexpression restored the reduction in the GCN2-ATF4 signaling in RAB25 knockdown mice (
Fig. 5F). Additionally, GCN20 may act as a co-activator cooperating with GCN1 to promote GCN2 phosphorylation under ethanol stress (
Supplementary Fig. 10). Moreover, further experiments demonstrated that alcohol metabolism remains unaffected by RAB25 knockdown in ethanol-fed mice (
Supplementary Fig. 11). Taken together, these findings suggest that RAB25 contributes to the regulation of ER stress through interaction with GCN1 especially in ALD.
ER stress can modulate the expression of mitochondrial regulatory components, thereby regulating mitochondrial function and dynamics [
12]. However, the impact of ER stress on mitochondrial function in ALD remains unclear. We found that ethanol significantly reduced the oxygen consumption rate (OCR), and RAB25 suppression partially reversed this effect. The ethanol-induced reductions in basal respiration and ATP production were notably ameliorated in RAB25-knockdown cells (
Fig. 6A). Moreover, we observed that ethanol reduced the protein levels of electron transport chain components I, II, III, IV, and V both in vitro and in vivo, and this reduction was partially reversed by RAB25 knockdown (
Fig. 6B). TEM revealed that ethanol increased mitochondria-associated membranes (MAMs) formation and elongated the MAM length, changes partially reversed by RAB25 suppression (
Fig. 6C). We further observed that ethanol enhanced ER-mitochondria colocalization, increased the MAM-resident protein MFN2, and elevated mitochondrial Ca
2+ concentrations. Conversely, RAB25 knockdown suppressed these ethanol-induced effects (
Supplementary Fig. 12). Collectively, our findings suggest that RAB25 suppression restores MAM structural integrity, attenuates pathological calcium flux, and alleviate mitochondrial dysfunction in ALD.
Since ER stress can change global lipid metabolism by upregulating lipogenic genes, we found significant increases in the expression levels of peroxisome proliferator-activated receptor gamma (PPARγ), sterol regulatory element binding protein 1c (SREBP-1c), and acetyl-CoA carboxylase 1 (ACC1) in ethanol-fed mice. RAB25 knockdown partially reversed these changes (
Fig. 6D). Since mTORC1 plays an important role in ALD-associated lipogenesis, we utilized
Tsc1 LKO mice with ethanol feeding. RAB25 knockdown failed to alleviate ethanol-induced liver injury, reduce SREBP1 activation, or suppress mTORC1 signaling (
Supplementary Fig. 13), These results demonstrate that the RAB25-PPARγ axis represents an mTORC1-independent mechanism governing de novo lipogenesis during ethanol exposure.
Similarly, lipidomic analysis of the tissue revealed the significant impact of RAB25 on hepatic lipid metabolism. Ethanol administration led to alterations in the levels of numerous lipid types, as indicated by the normalized heatmaps. Among the most abundant lipid types, TG, phosphatidylcholine (PC), ceramide (CER), and lysophosphatidylcholine (LPC) were significantly increased. These changes were alleviated in the RAB25-knockdown groups (
Fig. 6E).
Furthermore, increased hepatic ROS accumulation exacerbates ALD progression [
13]. We confirmed RAB25 knockdown significantly reduced ROS accumulation (
Supplementary Fig. 14). As 4-hydroxynonenal (4-HNE) is the ultimate byproduct of ROS and lipid peroxidation in ALD [
14], we found significantly lower levels of 4-HNE in the AAV-shRAB25 mice following ethanol administration (
Fig. 6F). Taken together, these findings indicate that RAB25 knockdown can reshape lipid metabolism, leading to a reduction in lipid peroxidation toxicity, thereby inhibiting the progression of ALD.
Alcohol-exposed hepatocytes activate inflammation by producing various mediators including cytokines, chemokines, and damage-associated molecular patterns (DAMPs) [
15]. Liver tissues from ALD patients exhibited elevated inflammation with increased neutrophils, macrophages, and T cells infiltration (
Fig. 7A). Single cell data proved the link between RAB25 expression to cytokine pathway activation (
Supplementary Fig. 15A,
15B). Cell-PhoneDB analysis revealed intensive communication between hepatocytes and immune cells, such as CD4, CD8, DC and NK cells, identifying hepatocytes as key producers of chemokines including IL-1, CCL2, CXCL10, and CXCL12 in ALD (
Fig. 7B). In ethanol-fed mice, RAB25 knockdown partially attenuated macrophage, neutrophil, CD4+/CD8+ T cell infiltration with decreased multiple inflammatory mediators such as IL-1α, IL-1ra, CXCL9, CXCL12 and TNFα (
Fig. 7C–
7E,
Supplementary Fig. 15C–
15E).
Mechanistically, RAB25 suppression decreased NLRP3 protein levels—a known downstream effector of the ATF4-CHOP axis [
16]—and reduced caspase-1 activation (
Supplementary Fig. 15F–
15I). By integrating cell-cell communication analysis, we identified CXCL12 as a central mediator of inflammation in ALD, with elevated levels confirmed in ALD patients (
Fig. 7E,
Supplementary Fig. 16A). RAB25 knockdown significantly reduced ethanol-induced CXCL12-positive cells (
Fig. 7F). Western blot analysis demonstrated that RAB25 knockdown reduced CXCL12 protein levels in liver tissues and HepG2 cells by downregulating the ATF4-CHOP-NLRP3 cascade (
Supplementary Fig. 16B,
16C). These results establish that RAB25 promotes inflammation through coordinated ATF4-CHOP-NLRP3 signaling, with CXCL12 serving as a principal effector chemokine that mediates immune cell recruitment in ALD.
DISCUSSION
ER stress contributes to the progression of ALD; currently, the majority of related research has focused predominantly on classic ER stress sensors. All three canonical pathways of the ER stress have been shown to contribute to fatty liver development [
17]. Whereas activation of the PERK-ATF4 pathway was found to contribute to TAG accumulation by increasing de novo lipogenesis [
18], ATF4 deficiency ameliorated both age-related and diet-induced fatty liver development [
11]. These studies involved manipulating ATF4 to regulate ER stress. However, the knockdown of PERK or IRE1α alone cannot completely block ER stress in ALD, suggesting that alternative factors may activate ER stress in ALD [
19]. In addition to PERK-mediated eIF2α phosphorylation, double-stranded RNA-activated protein kinase (PKR) and heme-regulated inhibitor kinase (HRI), can phosphorylate Ser51 of eIF2α during viral infection and heme deprivation, respectively [
19]. However, the involvement of secondary ER stress factors in ALD is poorly understood, and the specific mechanisms by which these cells participate remain unknown. In our study, we observed an abnormal increase in RAB25 levels and RAB25 translocation to the ER, where it interacts with GCN1. We discovered that the binding of RAB25 to GCN1 activates ER stress in ALD. Importantly, there was a significant reduction in ER stress and an improvement in liver injury when RAB25 was inhibited.
To date, the majority of studies on the Rab family of small GTPases have focused on their involvement in membrane trafficking processes [
20]. Previous reports have indicated that RAB27 plays a role in controlling the biogenesis and release of exosomes, whereas various RAB regulators have been shown to interact with essential components of the autophagic machinery [
21]. Exosomes are small vesicles that are released extracellularly and originate from the ER. Most related studies have concentrated on GTPase functions in the process of exosome trafficking, whether GTPases can change ER function has not been determined. We propose that RAB25 may play a potential role in the development of ER stress. In our study, we systematically demonstrated that upon exposure to ethanol, RAB25 translocates to the ER and acts as a mediator of GCN1 stability through its ubiquitination effects. Our study revealed that RAB25 directly binds to GCN1 and acts as a platform for K33-linked deubiquitination-mediated stabilization of GCN1. Furthermore, GCN20 may cooperate with GCN1 as a co-activator to induce GCN2 phosphorylation under ethanol stress.
ER stress has been reported to disrupt mitochondrial function, activate inflammation, and contribute to lipid accumulation in ALD patients [
11,
22]. The ER and mitochondria are two crucial organelles in eukaryotic cells that exhibit dynamic interactions and are physically connected through MAMs. In our study, alcohol disrupted mitochondrial homeostasis and induced mitochondrial dysfunction via MAMs. Interestingly, we found that suppressing RAB25 could reverse these effects on ALD by modulating MAMs. RAB25 knockdown decreased alcohol-induced lipogenesis by reducing the protein levels of lipogenic genes. It also mitigated lipid peroxidation toxicity. These findings provide an understanding of how RAB25 regulates ER stress and prevents excessive lipogenesis in ALD. Mitochondrial dysfunction can lead to the generation of high levels of ROS, and lipid accumulation can induce lipotoxicity, both of which can exacerbate hepatic steatosis and inflammation [
23]. Our study revealed that excessive alcohol intake increased the infiltration of inflammatory cells into the liver, while suppressing RAB25 reduced the presence of neutrophils and macrophages. It has been reported that ER stress can mediate inflammation through multiple pathways [
24]. However, the role of ER stress-mediated inflammatory factors in the progression of ALD has not been determined. We discovered that RAB25 knockdown specifically reduces CXCL12 levels to improve the immune microenvironment. Hence, RAB25 may regulate inflammation in ALD patients by targeting CXCL12. The involvement of noncanonical ER stress and whether it mediates ALD damage in contrast to classic ER stress remain uncertain. Therefore, we investigated the major pathogenic factors involved in ALD, including mitochondrial damage, lipid dysregulation, and inflammation. The results revealed that secondary ER stress has pathogenic effects similar to those of classic ER stress, and they synergistically contribute to the disease progression. Thus, targeting secondary ER stress has potential for ALD treatment.
In summary, this study revealed that the specific upregulation of RAB25 contributes to ALD progression by activating ER stress responses. RAB25 plays a crucial role in regulating the protein level of GCN1 by modulating its ubiquitination, thereby driving the development of ALD by promoting chronic ER stress responses. Suppressing RAB25 can attenuate ALD by preventing mitochondrial dysfunction, reducing lipogenesis, and suppressing inflammation. These findings improve the understanding of the involvement of RAB25 in ALD pathogenesis and suggest that RAB25 is a novel therapeutic target for the prevention and clinical treatment of ALD.
FOOTNOTES
-
Authors’ contributions
J.W., Y.G., and F.Z. designed and supervised the research. X.L. and J.W. wrote the paper. Z.Z. and Z.L. extracted and analyzed the data. X.L., Z.Z., Y.Z., X.Q., X.L., X.L. K.B., and X.X. performed the experiments. All authors contributed to the article and approved the submitted version.
-
Acknowledgements
This work was supported by the National Natural Science Foundation of China (82200703, 92068206), Guangdong Basic and Applied Basic Research Foundation (2021A1515110228).
-
Conflicts of Interest
The authors have no conflicts to disclose.
SUPPLEMENTARY MATERIAL
Supplementary material is available at Clinical and Molecular Hepatology website (
http://www.e-cmh.org).
Supplementary Figure 1.
RAB25 is upregulated in ALD patients. (A) UMAP showing the annotation and color codes for cell types in the ALD ecosystem. (B, C) UMAP plots showing the distribution of healthy control and ALD samples across different cell clusters, with cluster identities and group origins indicated by distinct color. (D, E) Feature plots displaying the expression patterns of selected marker genes across cell clusters. (F) Percent contribution of health and ALD human cells from each cluster. (G) The expression of marker genes in the ten different hepatocytes. (H) UMAP plot showing the distribution of healthy control and ALD samples within ten hepatocyte subclusters. (I) The cell ratio of health and ALD cells from each hepatocyte cluster. (J) The mRNA expression level of RAB25 in each cell types. ALD, alcohol-associated liver disease; UMAP, uniform manifold approximation and projection.
Supplementary Figure 2.
RAB25 is closely related to ER stress in ALD. (A) Feature plots of ER stress relate genes from ten hepatocyte (The red circle represents high expression of RAB25). (B) The co-staining percentage between ATF4 and RAB25 in livers obtained from ALD patients compare to health. (C) Quantification of the percentage of the positive cells of ATF4 in health and ALD patients. (D) Representative images of CK19 immunohistochemistry (IHC) in liver tissues obtained from pair-fed and EtOH-fed mice. Scale bars, 100 μm. (E) Quantification of the percentage of the positive cells of CK19. (F) The mRNA expression levels of RAB25 in pair-fed and EtOH-fed mouse liver. (G) Extrahepatic tissue sections of ethanol-fed mice were stained with H&E and analyzed by IHC for RAB25 levels. Notably, no significant elevation of RAB25 was observed in other tissues. Scale bars, 100 μm. The data are presented as mean±SEM. ER, endoplasmic reticulum; ALD, alcohol-associated liver disease. **P<0.01 and ***P<0.001 by Student’s t-test or two-way ANOVA.
Supplementary Figure 3.
Validation of RAB25 expression across non-ALD liver injury models and corresponding histological features. (A) Gene expression analysis of RAB25 in liver tissues from patients with NASH, cirrhosis, drug-induced liver injury (DILI), and primary sclerosing cholangitis (PSC), using publicly available GEO datasets from the NCBI database. (B) Top panel: Schematic representation of the 16-week high-fat, high-cholesterol diet (HFCD) mouse model. Bottom panel: Representative H&E and immunohistochemical staining for RAB25 and BIP in liver sections (scale bars, 100 μm). (C) Serum ALT and AST levels in normal diet (ND) and HFCD mice. (D) Hepatic mRNA expression levels of RAB25 in ND and HFCD mice. (E) Top panel: Schematic of the 4-week methionine-choline-deficient (MCD) diet model. Bottom panel: Representative H&E and immunohistochemical staining for RAB25 and BIP (scale bars, 100 μm). (F) Serum ALT and AST levels in methionine-choline-supplemented (MCS) and MCD mice. (G) Hepatic mRNA expression of RAB25 in MCS and MCD mice. (H) Top panel: Experimental timeline of the 8-week CCl₄-induced liver fibrosis model. Bottom panel: Representative H&E and immunohistochemical staining of RAB25 and BIP in liver sections (scale bars, 100 μm). (I) Serum ALT and AST levels in control and CCl-treated mice. (J) The mRNA expression levels of RAB25 in control and CCl₄₄-treated mice. (K) Top panel: Experimental timeline of the 28-day bile duct ligation (BDL) model. Bottom panel: Representative H&E and immunohistochemical staining for RAB25 and BIP (scale bars, 100 μm). (L) Serum ALT and AST levels in control and BDL mice. (M) Hepatic mRNA expression of RAB25 in control and BDL mice. The data are presented as mean±SEM. ALD, alcohol-associated liver disease; NASH, non-alcoholic steatohepatitis; GEO, Gene Expression Omnibus; ALT, alanine transaminase; AST, aspartate transaminase. ***P<0.001; n.s., not significant by Student’s t-test or two-way ANOVA.
Supplementary Figure 4.
RAB25 upregulation is conserved across chronic and metabolic dysfunction-associated ALD (MetALD) murine models of alcohol-associated liver disease. (A) Schematic overview of murine models: mice were fed a Lieber-DeCarli ethanol diet (5%) for 2 weeks, 4 weeks, or subjected to a MetALD model (n=6 mice per group). (B) Representative H&E, Oil Red O, and immunohistochemical staining of liver sections for RAB25 and BIP (Scale bars, 100 μm). (C) Serum levels of ALT, AST, and TG from pair-fed and ethanol-fed groups in each model (n=6 mice per group). (D) Quantification of Oil Red O-positive areas in liver sections. Data are presented as mean±SEM. ALT, alanine transaminase; AST, aspartate transaminase; TG, triglyceride; HFCD, high-fat, high-cholesterol diet. **P<0.01 and ***P<0.001 by Student’s t-test or two-way ANOVA.
Supplementary Figure 5.
HIF1A mediates alcohol-induced upregulation of RAB25 in ALD. (A) Venn diagram showing the overlap of differentially expressed genes between Hep-1 and Hep-5 hepatocyte subpopulations. (B) Heatmap of transcription factors positively correlated with RAB25 expression. (C) Gene expression levels of HIF1A and RAB25 in healthy controls and ALD samples. (D–F) HIF1A expression analysis in liver tissues from ALD patients using three independent GEO datasets (GSE155907, GSE142530, and GSE28619). (G–I) Correlation analysis between HIF1A and RAB25 expression in the corresponding GEO datasets. (J) Immunohistochemical staining of HIF1A and RAB25 in liver tissues from healthy controls and ALD patients (scale bars, 100 μm). (K) Western blot analysis of HIF1A and RAB25 protein levels in human liver tissues. (L, M) RAB25 protein (L) and mRNA (M) expression levels in liver tissues of pair-fed and ethanol-fed mice. (N) Confocal immunofluorescence staining for HIF1A (red), RAB25 (green) and nuclei (blue) in health and ALD human livers. Scale bars, 50 µm. (O, P) Protein expression in HepG2 cells following shRNA-mediated knockdown of HIF1A (O) or RAB25 (P) under alcohol exposure. The data are presented as mean±SEM. ALD, alcohol-associated liver disease; GEO, Gene Expression Omnibus. *P<0.05, **P<0.01; ***P<0.001 by Student’s t-test or two-way ANOVA.
Supplementary Figure 6.
Temporal and spatial expression patterns of RAB25 in murine ALD models. (A) Schematic illustration of mouse models fed a Lieber-DeCarli ethanol diet (5%) for 2 weeks or 4 weeks. (B) Serum levels of ALT, AST, and TG in pair-fed and ethanol-fed groups (n=6 mice per group). (C, D) Representative H&E, Oil Red O, and immunohistochemical (IHC) staining of liver sections from the indicated groups (scale bars, 100 μm). (E) Protein expression levels of RAB25, CYP2E1, ADH, and ALDH in liver tissues from 2-week and 4-week ethanol-fed mice, assessed by western blot. (F) IHC staining of CYP2E1, ALDH1A1, and RAB25 in liver sections from pair-fed and NIAAA model mice (scale bars, 100 μm). The data are presented as mean±SEM. ALD, alcohol-associated liver disease; ALT, alanine transaminase; AST, aspartate transaminase; TG, triglyceride; NIAAA, National Institute on Alcohol Abuse and Alcoholism. **P<0.01 and ***P<0.001; n.s., not significant by Student’s t-test or two-way ANOVA.
Supplementary Figure 7.
RAB25 interacts with GCN1 to mediate ER stress. (A) Validation of ADH and CYP2E1 protein overexpression by Western blot in transfected HepG2 cells. (B) qPCR analysis of ADH and CYP2E1 mRNA expression. (C, D) Immunoblot examination showing protein levels of RAB25, p-eIF2α, eIF2α, ATF4, CHOP and BIP in PHH (C) and PMH (D) following ethanol exposure for 48 hours. (E) Immunofluorescence analysis of RAB25 (green) and ER using calnexin (red) antibodies in HepG2 cells. Colocalized portion of RAB25 and ER is shown in yellow. Scale bars, 20 µm. (F) Using Coomassie staining to reveal proteins that may bind to RAB25. (G, H) Confocal immunofluorescence staining for GCN1 (red), RAB25 (green) and nuclei (blue) in AML12 cells treated with or without ethanol for 48 hours. Scale bars, 50 µm. (I) Immunoblot examination showing protein levels of GCN1, p-GCN2, GCN2, p-eIF2α, eIF2α, ATF4, CHOP and BIP in AML12 cells with stable knockdown of RAB25 after 48 hours of ethanol exposure. (J) Western blot analysis showing protein levels of ATF4, CHOP and BIP in AML12 cells following tunicamycin exposure. The data are presented as mean±SEM. ER, endoplasmic reticulum; PHH, primary human hepatocyte; PMH, primary mouse hepatocyte. *P<0.05, **P<0.01 and ***P<0.001 by Student’s t-test or two-way ANOVA.
Supplementary Figure 8.
The expression levels of GCN1 and BIP in liver diseases. (A) Quantification of the percentage of the positive cells of RAB25 and GCN1 in health and ALD patients. (B) Real-time PCR analysis of mRNA levels of GCN1 in liver samples from health and ALD. (C–E) Gene expression analysis in liver tissues from patients with ALD using GEO datasets from NCBI database. (F–I) Gene expression analysis in liver tissues from patients with NASH, cirrhosis, DILI and PSC using GEO datasets from NCBI database. (J) Quantification of the percentage of the positive cells of RAB25 and GCN1 in pair-fed and EtOH-fed mice. (K) Real-time PCR analysis of mRNA levels of GCN1 in liver samples from pair-fed and EtOH-fed mice. (L) Real-time PCR analysis of mRNA levels of GCN1 in HepG2 cells treated with ethanol. (M) Real-time PCR analysis of mRNA levels of GCN1 in HepG2 cells with stable knockdown of RAB25 treated with or without ethanol. The data are presented as mean±SEM. ALD, alcohol-associated liver disease; GEO, Gene Expression Omnibus; NASH, non-alcoholic steatohepatitis; DILI, drug-induced liver injury; PSC, primary sclerosing cholangitis. *P<0.05 and ***P<0.001; n.s., not significant by Student’s t-test or two-way ANOVA.
Supplementary Figure 9.
RAB25 knockdown did not reverse liver injury or ER stress in other liver diseases. (A, B) Confocal immunofluorescence confirming liver-specific expression of AAV-GFP (green) and AAV-shRAB25 (green) (scale bars, 50 μm). (C) Top panel: Schematic of the 16-week HFCD model with AAV injection 2 weeks prior to feeding. Bottom panel: H&E and immunohistochemical staining for ATF4 (scale bars, 100 μm). (D) Serum ALT and AST levels in AAV-GFP and AAV-shRAB25 mice in the HFCD model. (E) Top panel: Schematic of the 4-week MCD model with prior AAV injection. Bottom panel: H&E and immunohistochemical staining for ATF4 (scale bars, 100 μm). (F) Serum ALT and AST levels in MCD-fed mice. (G) Top panel: Schematic of the 8-week CCl₄-induced fibrosis model with prior AAV injection. Bottom panel: H&E, Masson and immunohistochemical staining for ATF4 (scale bars, 100 μm; n=6). (H) Serum ALT and AST levels in CCl₄-treated mice. (I) Top panel: Schematic of the 28-day bile duct ligation (BDL) model with prior AAV injection. Bottom panel: H&E and immunohistochemical staining for ATF4 (scale bars, 100 μm). (J) Serum ALT and AST levels in BDL mice. (K) Schematic depiction of the in vivo rescue experiment design. (L, M) Body weight changes (L) and liver-to-body weight ratios (M) in the indicated groups. (N, O) Serum biochemical measurements, including ALT, AST, triglycerides (TG), and total cholesterol (T-CHO), in the indicated groups. The data are presented as mean±SEM. ER, endoplasmic reticulum; AAV, adeno-associated virus; HFCD, high-fat, high-cholesterol diet; ALT, alanine transaminase; AST, aspartate transaminase; MCD, methionine-choline deficient. *P<0.05, **P<0.01; n.s., not significant by Student’s t-test or two-way ANOVA. NES>1 and P<0.05 were considered significant for enrichment analysis.
Supplementary Figure 10.
GCN20 is essential for GCN1-mediated activation of GCN2 under ethanol exposure. (A, B) GCN20 mRNA expression in liver tissues of patients with ALD and healthy controls, analyzed using publicly available GEO datasets (GSE155907 and GSE142530). (C) Co-immunoprecipitation (Co-IP) analysis reveals no detectable interaction between RAB25 and GCN20 in HepG2 cells. (D) Western blot analysis of RAB25, GCN1, and p-GCN2 in alcohol-treated HepG2 cells transfected with control siRNA or siGCN20. (E) Western blot analysis of RAB25, GCN1 and GCN20 in liver tissues of EtOH-fed mice with or without RAB25 knockdown or GCN1 overexpression. The data are presented as mean±SEM. ALD, alcohol-associated liver disease; GEO, Gene Expression Omnibus. *P<0.05, **P<0.01 and ***P<0.001 by Student’s t-test or two-way ANOVA.
Supplementary Figure 11.
RAB25 knockdown does not affect alcohol metabolism in EtOH-fed mice. (A) Quantification of serum ethanol concentrations in the pair-fed and EtOH-fed groups with RAB25 knockdown. (B) Western blot analysis of ADH1, ADLH2, and CYP2E1 proteins in mouse livers from EtOH-fed mice with or without RAB25 knockdown. (C) Immunohistochemical staining of ADH1, ALDH2 and CYP2E1 in liver sections from EtOH-fed mice with or without RAB25 knockdown (Scale bars, 100 μm). The data are presented as mean±SEM. AAV, adeno-associated virus. ***P<0.001; n.s., not significant by Student’s t-test or two-way ANOVA.
Supplementary Figure 12.
RAB25 knockdown restores MAM integrity and alleviates mitochondrial calcium overload in ALD. (A) Representative confocal images of HepG2 cells stained with Rhod-2 AM (red) and Mito-Tracker Green (green) following RAB25 knockdown. (B) Quantification of mitochondrial calcium levels ([Ca2+]ₘ) using Manders’ colocalization coefficient (M₂) for Rhod-2 AM and Mito-Tracker signals. (C) Immunofluorescence analysis of ER and mitochondria using PDI (dark green) and TOMM20 (red) antibodies in AAV-GFP and AAV-shRAB25 mice liver. Colocalized portion of ER and mitochondria is shown in white. Scale bars, 50 µm. (D) Quantification of the percentage of the co-localization of PDI and TOMM20. The data are presented as mean±SEM. MAM, mitochondria-associated membrane; ALD, alcohol-associated liver disease; ER, endoplasmic reticulum; AAV, adeno-associated virus. *P<0.05, ***P<0.001 by Student’s t-test or two-way ANOVA.
Supplementary Figure 13.
RAB25 knockdown fails to suppress mTORC1-driven lipogenesis in TSC1-deficient mice. (A) PCR genotyping confirming successful TSC1 knockout (red arrow). (B) Schematic overview of the in vivo experimental design (n=6 mice per group). (C, D) Serum ALT (C) and triglyceride (TG) (D) levels in pair-fed, ethanol-fed, and TSC1 LKO mice with or without RAB25 knockdown. (E, F) Representative H&E and Oil Red O staining (E) and quantification of hepatic lipid accumulation (F) in liver sections (Scale bars, 100 μm). (G, H) Western blot analysis of hepatic protein levels of SREBP1 and mTORC1 pathway effectors in the indicated groups. The data are presented as mean±SEM. ALT, alanine transaminase. *P<0.05, **P<0.01 and ***P<0.001; n.s., not significant by Student’s t-test or two-way ANOVA.
Supplementary Figure 14.
RAB25 knockdown attenuates ethanol-induced ROS and hepatocyte apoptosis. (A) Intracellular ROS levels measured by DCFH-DA staining in HepG2 cells. (B) Western blot analysis of PERK and ATF4 expression in HepG2 cells transfected with control siRNA or PERK-targeting siRNA. (C) 4-hydroxynonenal (4-HNE) protein adduct levels in liver tissues from healthy controls and ALD patients. (D) Confocal immunofluorescence of 4-HNE (red) and nuclei (blue) in AAV-GFP and AAV-shRAB25 mice (Scale bars, 50 µm). (E) Quantification of 4-HNE-positive areas in liver tissues. (F) Western blot analysis of 4-HNE protein levels in ethanol-treated HepG2 cells with RAB25 or PERK knockdown. (G, H) TUNEL staining and quantification of apoptotic cells in liver sections from the indicated groups (Scale bars, 100 μm). The data are presented as mean±SEM. ROS, reactive oxygen species; PERK, PKR-like endoplasmic reticulum kinase; ALD, alcohol-associated liver disease; AAV, adeno-associated virus. *P<0.05, **P<0.01, ***P<0.001 by Student’s t-test or two-way ANOVA.
Supplementary Figure 15.
RAB25 knockdown protects liver cells from ER stress-induced inflammatory activation. (A) Gene expression analysis in liver tissues from patients with ALD using GEO datasets from NCBI database. (B) GO analysis showed that high expression of RAB25 significantly upregulated TNF, cytokine-cytokine receptor interaction signaling pathways. (C) Representative images of TNFα immunohistochemistry in liver tissues obtained from AAV-GFP and AAV-shRAB25 mice after ethanol feeding. Scale bars, 100 μm. (D) Representative images of CD206 immunohistochemistry in liver tissues obtained from AAV-GFP and AAV-shRAB25 mice after ethanol feeding. Scale bars, 100 μm. (E) Quantification the protein levels of related genes in WT, AAV-GFP and AAV-shRAB25 mice after ethanol feeding. (F) Western blot analysis was conducted using liver tissue lysates from AAV-GFP and AAV-shRAB25 mice. (G) Quantitative real-time PCR analysis of hepatic mRNA expression of NLRP3, caspase-1, and IL-1β in the same groups. (H, I) Western blot analysis of ER stress and inflammatory proteins in HepG2 cells following RAB25 knockdown or overexpression. The data are presented as mean±SEM. ER, endoplasmic reticulum; ALD, alcohol-associated liver disease; GEO, Gene Expression Omnibus; GO, Gene Ontology; AAV, adeno-associated virus. *P<0.05, **P<0.01 and ***P<0.001 by Student’s t-test or two-way ANOVA.
Supplementary Figure 16.
RAB25 promotes CXCL12 expression via the ER stress–NLRP3 inflammatory axis. (A) Feature plots showing CCL and CXCL chemokine expression across ten hepatocyte clusters in single-cell RNA-sequencing data (red circles indicate clusters with high RAB25 expression). (B, C) Western blot analysis of CXCL12 and related inflammatory markers in HepG2 cells (B) and liver tissues (C) following RAB25 knockdown. The data are presented as mean±SEM. ER, endoplasmic reticulum; UMAP, uniform manifold approximation and projection; AAV, adeno-associated virus. *P<0.05, **P<0.01 and ***P<0.001 by Student’s t-test or two-way ANOVA.
Figure 1.RAB25 is upregulated in ALD patients and associated with disease progression. (A) UMAP showing the annotation and color codes for cell types in the ALD ecosystem (left). Heatmap showing the expression of marker genes in the indicated cell types (middle). Clustering of hepatocytes population (right). (B) The ORs of metaclusters occurring in each hepatocyte cluster (left). The top five genes between health and ALD (right). (C) The expression of RAB25 between health and ALD (left). Feature plots of specific marker genes from ten hepatocyte clusters (right). (D) Gene expression analysis in liver tissues from patients with ALD using GEO datasets. (E) The protein levels of RAB25 and CK18 in liver tissues from health and ALD patients. (F) Liver sections of health and ALD patients were stained with H&E and analyzed by IHC for RAB25 levels. Scale bars, 100 μm. The data are presented as mean±SEM. ALD, alcohol-associated liver disease; UMAP, uniform manifold approximation and projection; OR, odds ratio; IHC, immunohistochemistry; GEO, Gene Expression Omnibus. *P<0.05, **P<0.01, ***P<0.001 by Student’s t-test or two-way ANOVA.
Figure 2.RAB25 is closely related to ER stress in human and murine ALD. (A) GO analysis showed that high expression of RAB25 significantly upregulated ER stress (left). Confocal immunofluorescence staining for ATF4 (red), RAB25 (green) and nuclei (blue) and immunohistochemistry (IHC) staining in health and ALD human livers. Scale bars, 50 µm (right). (B) Schematic image of a murine model with chronic-plus-binge ethanol feeding. The effect of EtOH feeding on body weight changes and the ratios of liver-to-body weight. Serum ALT, AST, TG and T-CHO levels (n=6 mice per group). (C) Liver sections were stained with H&E, Oil Red O, and analyzed by IHC (Scale bars, 100 μm). (D) The protein levels of RAB25 in liver tissues from the pair-fed and EtOH-fed mice. (E) TEM showed damaged ER and mitochondrial structure in EtOH-fed mice liver (Scale bars, 5 µm). The red arrow represents ER. Blue arrows represent fat vacuoles. M, mitochondria. (F) Representative images of RAB25, BIP and ATF4 IHC in liver tissues. Scale bars, 100 μm (left). The protein levels in liver tissues from the pair-fed and EtOH-fed mice (right). The data are presented as mean±SEM. ER, endoplasmic reticulum; ALD, alcohol-associated liver disease; GO, Gene Ontology; ALT, alanine transaminase; AST, aspartate transaminase; TG, triglyceride; T-CHO, total cholesterol; TEM, transmission electron microscopy; MPO, Myeloperoxidase. *P<0.05, **P<0.01 and ***P<0.001; n.s., not significant by Student’s t-test or two-way ANOVA
Figure 3.RAB25 interacts with GCN1 to mediate ER stress. (A) The protein levels of RAB25, p-eIF2α, eIF2α, ATF4 and BIP in PHH and HepG2 cells following ethanol exposure with varying concentrations of ethanol for 48 hours. Immunofluorescence analysis of RAB25 (green) and ER using calnexin (red) antibodies in PHH cells and human livers (right). Scale bars, 20 µm. (B) Using Coomassie staining to reveal proteins that may bind to RAB25 (left). The top five identified candidate interacting proteins of RAB25. (C) The protein levels of GCN1 in HepG2 cells following ethanol exposure with varying concentrations of ethanol for 48 hours. Co-IP assays were conducted to investigate the interaction between RAB25 and GCN1 in HepG2 cells and AML12 cells following ethanol exposure for 48 hours (right). (D) Confocal immunofluorescence staining for GCN1 (red), RAB25 (green) in PHH, PMH, and HepG2 cells treated with or without ethanol for 48 hours. Scale bars, 50 µm. (E) The protein levels in HepG2 cells with stable knockdown of RAB25 after 48 hours of ethanol exposure (left). The protein levels in HepG2 cells with stable knockdown of RAB25 and adenovirus-mediated overexpression of GCN1 after 48 hours ethanol exposure (middle). The protein levels in HepG2 cells with stable knockdown of RAB25 after 48 hours of ethanol exposure (right). (F) The protein levels in HepG2 cells with control siRNA or siRNA against Perk. Co-IP analysis shows RAB25 cannot co-immunoprecipitate with PERK or eIF2α. The protein levels in HepG2 cells following tunicamycin exposure. Schematic diagram of the possible role of RAB25 in ALD. The data are presented as mean±SEM. ER, endoplasmic reticulum; PHH, primary human hepatocyte; PMH, primary mouse hepatocyte; Co-IP, co-immunoprecipitation; PERK, PKR-like endoplasmic reticulum kinase; ALD, alcohol-associated liver disease. *P<0.05, **P<0.01 and ***P<0.001 by Student’s t-test or two-way ANOVA.
Figure 4.RAB25 stabilizes the GCN1 protein by preventing K33-dependent ubiquitination and degradation. (A) Representative staining of liver sections from health and ALD patients, pair-fed and EtOH-fed mice for RAB25 and GCN1. Scale bars, 100 µm (left). The protein levels of RAB25 and GCN1 in the indicated groups. (B) The protein levels of GCN1 in HepG2 cells with stable knockdown of RAB25 stimulated with ethanol and together treated with cycloheximide (CHX) for the indicated time periods. The protein levels of GCN1 in HepG2 cells with stable knockdown of RAB25 stimulated with ethanol and together treated with MG132 (10 μM) for 4 hours (right). (C) Immunoblot analysis of lysates of HepG2 cells with stable knockdown of RAB25 stimulated with ethanol and together treated with MG132 (10 μM) for 4 hours, and followed by IP with GCN1 antibody (left). Ubiquitination screening of GCN1 by RAB25 with indicated types of ubiquitin and treated with MG132 (10 μM) for 4 hours in HEK293T cells (middle). The ubiquitination of GCN1 in response to RAB25 overexpression was investigated in HEK293T cells transfected with mutated Myc-Ub plasmids (K33O Ub with the intact Lys33 residue alone; K33R Ub only Lys33 residue was mutated) (right). (D) Surface diagram of the docking model and their interfacing residues between RAB25 and GCN1 protein. Co-immunoprecipitation analysis shows the interaction of GCN1 with different truncations of RAB25. (E) The ubiquitination levels of GCN1 in HEK293T cells transfected with the K33O vector in different truncations of RAB25. The protein levels in response to overexpression of RAB25 were investigated in HepG2 cells. (F) Candidate RAB25-interacted ubiquitin enzymes identified by IP–MS analysis. The interaction of RAB25 and USP5, GCN1 and USP5 in HEK293T cells was analyzed using IP. K33-linked ubiquitination of GCN1 in HEK293T cells transfected with USP. The protein levels of GCN1 in HEK293T cells transfected with USP5. The interaction of GCN1 and USP5 in HepG2 cells. The data are presented as mean±SEM. ALD, alcohol-associated liver disease. **P<0.01 and ***P<0.001 by Student’s t-test or two-way ANOVA.
Figure 5.RAB25 knockdown attenuates ethanol-induced ER stress via GCN1 in murine ALD. (A) Schematic image of a murine model with chronic-plus-binge ethanol feeding. The mice were injected with adeno-associated virus (AAV-shRAB25 and AAV-GFP) through tail vein injection two weeks before the modeling process (n=6 mice per group). The protein levels in mouse livers after injection of AAV-GFP or AAV-shRAB25 (middle). The effect of EtOH feeding on body weight changes, the ratios of liver-to-body weight, serum ALT, AST, TG and T-CHO levels of AAV-GFP and AAV-shRAB25 mice (n=6 mice per group). (B) H&E, Oil Red O in livers sections obtained from AAV-GFP and AAV-shRAB25 mice. Scale bars, 100 μm. TEM analysis revealed that the damaged ER and mitochondrial structure in the liver of AAV-shRAB25 mice was improved. Scale bars, 5 µm (right). (C) Western blot analysis was conducted using liver tissue lysates from AAV-GFP and AAV-shRAB25 mice. (D) Representative images of GCN1, BIP and ATF4 in liver tissues obtained from AAV-GFP and AAV-shRAB25 mice. Scale bars, 100 μm. (E) H&E and Oil Red O staining and corresponding quantification of liver sections of the mice from the indicated groups (n=6). Scale bars, 100 μm. (F) The protein level in the livers of the mice from the indicated groups. The data are presented as mean±SEM. ER, endoplasmic reticulum; ALD, alcohol-associated liver disease; ALT, alanine transaminase; AST, aspartate transaminase; TG, triglyceride; T-CHO, total cholesterol; TEM, transmission electron microscopy; ns, not significant. *P<0.05, **P<0.01 and ***P<0.001 by Student’s t-test or two-way ANOVA.
Figure 6.RAB25 knockdown alleviated mitochondrial dysfunction and reshapes lipid metabolism in ALD. (A) OCR in HepG2 cells was measured by XF-analyzer. Rot, Rotenone; AA, Antimycin A. Basal respiration, ATP production and maximal respiration were calculated (right). (B) Western blot analysis was performed from HepG2 cells with stable knockdown of RAB25 and AAV-GFP and AAV-shRAB25 mice to determine the protein levels. Schematic diagram of RAB25 regulation of MAMs (right). (C) MAM formation (red arrows) in the AAV-GFP and AAV-shRAB25 mice liver sections were analyzed. Scale bars, 5 µm. (D) Representative images of in liver tissues obtained from AAV-GFP and AAV-shRAB25 mice. Scale bars, 100 μm. Western blot analysis was conducted using liver tissue lysates from AAV-GFP and AAV-shRAB25 mice. (E) Heatmap analysis of selected lipid molecules from the indicated groups. (F) Western blot analysis was conducted using liver tissue lysates from AAV-GFP and AAV-shRAB25 mice. The data are presented as mean±SEM. ALD, alcohol-associated liver disease; OCR, oxygen consumption rate; AAV, adeno-associated virus; MAM, mitochondria-associated membrane; ER, endoplasmic reticulum. *P<0.05, **P<0.01 and ***P<0.001 by Student’s t-test or two-way ANOVA.
Figure 7.RAB25 knockdown protects liver cells from ER stress-induced inflammatory activation. (A) Representative staining of liver sections from health and ALD patients. Scale bars, 100 µm. (B) Cell-cell communications between 1-hepatocyte and other cells (left). Overview of selected ligand–receptor interactions of 1-hepatocyte and other cells (right). (C) Representative images of F4/80 and MPO in liver tissues obtained from AAV-GFP and AAV-shRAB25 mice. Scale bars, 100 μm. (D) Representative images of CD4 and CD8 in liver tissues obtained from AAV-GFP and AAV-shRAB25 mice after ethanol feeding. Scale bars, 100 μm. (E) Mouse liver lysates were analyzed using the mouse cytokine antibody array kit (Cat. ARY006) for the detection of 40 different cytokines and chemokine (n=4). (F) CXCL12 expression analysis in liver tissues from patients with ALD using GEO datasets. Representative images of CXCL12 in liver tissues obtained from health and ALD patients and mice after ethanol feeding (below). Scale bars, 100 μm. The data are presented as mean±SEM. ER, endoplasmic reticulum; ALD, alcohol-associated liver disease; MPO, MPO, Myeloperoxidase; AAV, adeno-associated virus; GEO, Gene Expression Omnibus; ns, not significant. *P<0.05, **P<0.01 and ***P<0.001 by Student’s t-test or two-way ANOVA.
Abbreviations
alcohol-associated liver disease
activating transcription factor 6
Dulbecco’s Modified Eagle Medium
inositol-requiring enzyme 1α
mitochondria-associated membrane
methionine-choline deficient
PKR-like endoplasmic reticulum kinase
peroxisome proliferator-activated receptor gamma
reverse transcription-quantitative PCR
transmission electron microscopy
uniform manifold approximation and projection
unfolded protein response
REFERENCES
- 1. Louvet A, Mathurin P. Alcoholic liver disease: mechanisms of injury and targeted treatment. Nat Rev Gastroenterol Hepatol 2015;12:231-242.
- 2. GBD 2017 Cirrhosis Collaborators. The global, regional, and national burden of cirrhosis by cause in 195 countries and territories, 1990-2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet Gastroenterol Hepatol 2020;5:245-266.
- 3. Singh S, Murad MH, Chandar AK, Bongiorno CM, Singal AK, Atkinson SR, et al. Comparative Effectiveness of pharmacological interventions for severe alcoholic hepatitis: a systematic review and network meta-analysis. Gastroenterology 2015;149:958-970.e12.
- 4. Walter P, Ron D. The unfolded protein response: from stress pathway to homeostatic regulation. Science 2011;334:1081-1086.
- 5. Wu X, Fan X, Miyata T, Kim A, Cajigas-Du Ross CK, Ray S, et al. Recent advances in understanding of pathogenesis of alcohol-associated liver disease. Annu Rev Pathol 2023;18:411-438.
- 6. Yao Y, Lu Q, Hu Z, Yu Y, Chen Q, Wang QK. A non-canonical pathway regulates ER stress signaling and blocks ER stress-induced apoptosis and heart failure. Nat Commun 2017;8:133.
- 7. Yin G, Huang J, Petela J, Jiang H, Zhang Y, Gong S, et al. Targeting small GTPases: emerging grasps on previously untamable targets, pioneered by KRAS. Signal Transduct Target Ther 2023;8:212.
- 8. Ouyang Q, Chen Q, Ke S, Ding L, Yang X, Rong P, et al. Rab8a as a mitochondrial receptor for lipid droplets in skeletal muscle. Dev Cell 2023;58:289-305.e6.
- 9. Bertola A, Mathews S, Ki SH, Wang H, Gao B. Mouse model of chronic and binge ethanol feeding (the NIAAA model). Nat Protoc 2013;8:627-637.
- 10. Kim GA, Moon JH, Kim W. Critical appraisal of metabolic dysfunction-associated steatotic liver disease: Implication of Janus-faced modernity. Clin Mol Hepatol 2023;29:831-843.
- 11. Hao L, Zhong W, Dong H, Guo W, Sun X, Zhang W, et al. ATF4 activation promotes hepatic mitochondrial dysfunction by repressing NRF1-TFAM signalling in alcoholic steatohepatitis. Gut 2021;70:1933-1945.
- 12. Abdallah MA, Singal AK. Mitochondrial dysfunction and alcohol-associated liver disease: a novel pathway and therapeutic target. Signal Transduct Target Ther 2020;5:26.
- 13. Diehl AM, Day C. Cause, pathogenesis, and treatment of nonalcoholic steatohepatitis. N Engl J Med 2017;377:2063-2072.
- 14. Dey A, Cederbaum AI. Alcohol and oxidative liver injury. Hepatology 2006;43:S63-S74.
- 15. Cai Y, Xu MJ, Koritzinsky EH, Zhou Z, Wang W, Cao H, et al. Mitochondrial DNA-enriched microparticles promote acute-onchronic alcoholic neutrophilia and hepatotoxicity. JCI Insight 2017;2:e92634.
- 16. Lebeaupin C, Proics E, de Bieville CH, Rousseau D, Bonnafous S, Patouraux S, et al. ER stress induces NLRP3 inflammasome activation and hepatocyte death. Cell Death Dis 2015;6:e1879.
- 17. DeZwaan-McCabe D, Sheldon RD, Gorecki MC, Guo DF, Gansemer ER, Kaufman RJ, et al. ER stress inhibits liver fatty acid oxidation while unmitigated stress leads to anorexia-induced lipolysis and both liver and kidney steatosis. Cell Rep 2017;19:1794-1806.
- 18. Lauressergues E, Bert E, Duriez P, Hum D, Majd Z, Staels B, et al. Does endoplasmic reticulum stress participate in APD-induced hepatic metabolic dysregulation? Neuropharmacology 2012;62:784-796.
- 19. Bhardwaj M, Leli NM, Koumenis C, Amaravadi RK. Regulation of autophagy by canonical and non-canonical ER stress responses. Semin Cancer Biol 2020;66:116-128.
- 20. Homma Y, Hiragi S, Fukuda M. Rab family of small GTPases: an updated view on their regulation and functions. FEBS J 2021;288:36-55.
- 21. Ostrowski M, Carmo NB, Krumeich S, Fanget I, Raposo G, Savina A, et al. Rab27a and Rab27b control different steps of the exosome secretion pathway. Nat Cell Biol 2010;12:19-30 sup pp 1-13.
- 22. Ratna A, Lim A, Li Z, Argemi J, Bataller R, Chiosis G, et al. Myeloid endoplasmic reticulum resident chaperone GP96 facilitates inflammation and steatosis in alcohol-associated liver disease. Hepatol Commun 2021;5:1165-1182.
- 23. Choi SE, Hwang Y, Lee SJ, Jung H, Shin TH, Son Y, et al. Mitochondrial protease ClpP supplementation ameliorates diet-induced NASH in mice. J Hepatol 2022;77:735-747.
- 24. Ajoolabady A, Lebeaupin C, Wu NN, Kaufman RJ, Ren J. ER stress and inflammation crosstalk in obesity. Med Res Rev 2023;43:5-30.
, Yi Gao1,2
