Interleukin 33 Promotes Liver Sinusoidal Endothelial Cell Dysfunction and Hepatic Fibrosis in Diabetic Mice
Article information
Abstract
Background
Interleukin 33 (IL33) drives liver fibrosis, and individuals with type 2 diabetes mellitus are more likely advanced to liver fibrosis. However, the role of IL33 in diabetic liver fibrosis remains unclear, prompting our investigation.
Methods
We developed a diabetes model in wild-type, IL33−/−, and suppression of tumorigenicity 2 (ST2−/−, IL33 receptor) mice. Furthermore, wild-type diabetic mice were injected with IL33 neutralizing antibody (αIL33). We also co-cultured human liver endothelial cells and human hepatic stellate cells to identify the role of IL33 in high palmitic acid and high glucose conditions.
Results
Hepatic collagen deposition was increased in diabetic mice, which was alleviated by IL33 knockout, ST2 knockout, or administration of αIL33. Also, αIL33 treatment blunted liver sinusoidal endothelial cell (LSEC) dysfunction and inflammation during diabetic liver fibrosis progression. Recombinant IL33 (rIL33) treatment aggravated autophagy disruption in the presence of palm acid and high glucose in LSECs, which was blunted by autophagy agonist rapamycin administration and ERK/MAPK inhibitor PD98059 treatment. Hepatic stellate cell line LX-2 co-cultured with rIL33-pretreated LSECs displayed augmented activation, which was also attenuated by rapamycin or PD98059 pretreated.
Conclusion
IL33 drives LSEC dysfunction and promotes diabetic hepatic fibrosis, thus a potential therapeutic target for diabetic liver fibrosis.
Highlights
• IL33 drives hepatic fibrosis in diabetic mice.
• IL33 promotes liver sinusoidal endothelial cell dysfunction and inflammation.
• Recombinant IL33 treatment aggravates endothelial-mesenchymal transition in LESCs.
• IL33 is an attractive therapeutic target for diabetic liver fibrosis.
INTRODUCTION
Non-alcoholic fatty liver disease (NAFLD) is a kind of most common metabolic disease affecting 1.2 billion individuals globally [1]. This chronic liver disease represents a disease spectrum, ranging from steatosis, non-alcoholic steatohepatitis, liver cirrhosis, and hepatocellular carcinoma. Strikingly, the fibrosis stage is a key prognostic marker for liver-related mortality in NAFLD [2]. In a meta-analysis comprising 4,428 individuals with NAFLD from 13 studies, individuals with any stage of fibrosis increased liver-related mortality (relative risk, 11.13; 95% confidence interval [CI], 4.15 to 29.84) [3]. Type 2 diabetes mellitus (T2DM) is the vital risk factor for hepatic fibrosis in NAFLD [4]. An analysis involving 3,841 participants from National Health and Nutrition Examination Survey 2015 to 2016 reported that T2DM doubled the risk of hepatic fibrosis in individuals with obesity and NAFLD (odds ratio, 2.06; 95% CI, 1.30 to 3.25) [5]. Moreover, a Mendelian randomization study indicated that the genetic liability of T2DM was associated with the risk of liver cirrhosis (odds ratio, 1.12; 95% CI, 1.03 to 1.21), independently of the genetically predicted body mass index [6]. Notably, preventative efforts and research have largely focused on traditional vascular risk factors, the causes of death from T2DM have shifted. The vascular death rates decreasing and liver disease death rates increasing, therefore, these efforts toward excessive risk should be expanded to consider the wide causes of death, especially in patients with liver disease [7]. Worryingly, despite the strong association between T2DM and liver disease especially fibrosis has been shown [8], the mechanisms are still not fully elucidated.
T2DM is normally characterized by systemic low-grade inflammation which probably contributes to the deleterious impact of T2DM on NAFLD progression [9]. Recent research underscores the significance of cytokines in liver fibrosis, with interleukin 33 (IL33), belonging to the interleukin 1 family, emerging as a key player in various organ fibrosis conditions like pulmonary, pancreatic, and renal fibrosis [10]. This cytokine is known as a damage-associated molecular pattern derived from necrotic or impaired cells, and it binds and signals through its receptor suppression of tumorigenicity 2 (ST2) (also known as IL-1 receptor-like 1, IL1RL1) [11]. In fibrotic liver, IL33 and ST2 are overexpressed [12], and their expression is significantly increased with the progression of fibrosis [13]. Notably, despite IL33 is reported to be highly expressed in both circulating and tissue samples from diabetic individuals [14], the role of IL33 in diabetic liver fibrosis remains unknown.
IL33 was reported to trigger thioacetamide or carbon tetrachloride-induced hepatic fibrosis by activating liver resident innate lymphoid cells (ILC2s) and hepatic stellate cells (HSCs), which is ST2-dependent [13,15]. Liver fibrosis was mainly perisinusoidal either in NAFLD or in liver fibrosis induced by carbon tetrachloride administration [16]. Strikingly, researchers performed single-cell transcriptomics analysis of liver endothelial cells from normal and cirrhotic livers collected from mice or humans, they determined that perisinusoidal liver sinusoidal endothelial cells (LSECs) are most susceptible to damage characterized by capillarization-associated genes upregulated [17]. Healthy LSECs function as gatekeepers of liver by maintaining HSC quiescence and hepatic immune homeostasis [18]. Several lines of evidence demonstrated that LSEC dysfunction (characterized by reduced bioavailability of nitric oxide [NO] and proinflammatory phenotype) always occurs early and promotes NAFLD progress to fibrosis [19,20]. LSECs are highly specialized endothelial cells characterized by a discontinuous basement membrane and open fenestrae. These structures enable themselves to sense the metabolic change in circulation and exchange metabolites between hepatic sinusoids and liver cells [21]. The metabolic alterations and systemic low-grade inflammation always drive endothelial dysfunction in individuals with severe T2DM [22]. Dysfunction of LSEC disrupts liver immune homeostasis and HSC activation, both of which are mechanisms by which IL33 promotes fibrosis. Furthermore, IL33 shows a proinflammatory effect on endothelial cells in an autocrine/paracrine manner [23,24]. It is reported that IL33 induced remodeling and capillary-like network formation of endothelial cells [24,25]. Taken together, IL33-mediated changes in LSECs may be the key to preventing diabetic liver fibrosis. Thus, there is a pressing need to dissect the relationship between IL33 and LSEC dysfunction in diabetic liver fibrosis. Taken together, we intended to investigate the role of IL33 in diabetic liver fibrosis and the potential mechanisms.
METHODS
Animal model
Eight-week years old male C57BL/6J mice were purchased from Vital River Laboratory Animal Technology Co. Ltd. (Beijing, China), and IL33−/−, and ST2−/− mice (C57BL/6J) were kindly gifted by the Department of Immunology, Tongji Medical College. All of them were housed in a temperature-controlled room (22°C±2°C) with a 12-hour–12-hour light: dark cycle. All mice were fed with either a normal diet (ND) (Xietong, Guangzhou, China) or a high-fat diet (HFD) (Xietong, 60% calories from fat) for 12 weeks, and mice fed with HFD were given intraperitoneal streptozotocin (STZ) injection (50 mg/kg) or vehicle solution daily for 5 consecutive days. Seven days later, mice with fasting blood glucose levels ≥11.1 mmol/L were considered to be diabetic [26] and were fed with an HFD or ND for continuous 12 weeks. Wild-type diabetic mice were given an intraperitoneal injection of anti-IL33 neutralizing antibody (αIL33, 37 μg/kg body weight, R&D Systems, Minneapolis, MN, USA) or recombinant IL33 (rIL33, 12.5 μg/kg body weight, R&D Systems) twice a week for last 12 weeks. On the last day, mice were intraperitoneal injected with sodium pentobarbital (50 mg/kg) and then euthanized by cervical dislocation.
Ethic approval statement
All the procedures were approved by the Institutional Animal Care and Use Committee at Huazhong University of Science and Technology (approval reference number, S820). All mice received human care according to the criteria outlined in the Guide for the Care and Use of Laboratory Animals.
Cell culture and treatments
The human LSEC cell line SK-Hep1 and human HSC line LX-2 were kindly gifted by Dr. Kong Qinghong from Zunyi Medical University and Prof. Zhou Ping from Huazhong University of Science and Technology [27]. Human LSECs were purchased from Zhao Rui Biotech (iCell-0019a, Shanghai, China), and cultured with endothelial cell medium (ScienCell, Carlsbad, CA, USA) containing 5% fetal bovine serum, 1% penicillin/streptomycin solution, and 1% endothelial cell growth supplement. SK-Hep1 cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM) (Gibco, Grand Island, NY, USA) at 37°C in a 5% CO2 incubator, supplemented with 10% fetal bovine serum (Gibco) and 1% penicillin/streptomycin solution (Solarbio, Beijing, China). Human monocyte leukemia cell line THP-1 was purchased from the Cell Bank of the Type Culture Collection of the Chinese Academy of Sciences. THP-1 was cultured in Roswell Park Memorial Institute (RPMI) 1640 medium containing 10% fetal bovine serum and 1% penicillin/streptomycin solution. LX-2 cells were cultured in DMEM supplemented with 10% fetal bovine serum (Gibco) and 1% penicillin/streptomycin solution at 37°C with 5% CO2. Palmitic acid (PA) was dissolved in 0.1 M NaOH by heating 90°C for 10 minutes, and then the stock solution was prepared to contain 20 mM PA with 5% (w/v) bovine serum albumin (Solarbio). Cells were treated with normal glucose (5.5 mM) and PA plus high glucose (HG, 30 mM) (PAHG) for 24 hours.
Co-cultures
SK-Hep1 cells or human LSECs and LX-2 cells were co-cultured using Transwell (BIOFIL, Guangzhou, China). SK-Hep1 cells or human LSECs were seeded in inserts, and LX-2 cells were plated in the bottom and cultured with DMEM. They were cultured with DMEM containing 10% fetal bovine serum and 1% penicillin/streptomycin. SK-Hep1 cells or human LSECs were treated with rIL33 or rapamycin in the presence of PAHG, after 24 hours, the culture medium was removed and replaced with a fresh culture medium. Then, SK-Hep1 cells or human LSECs and LX-2 cells were co-cultured for the following 48 hours.
Monocyte adhesion assay
THP-1 cells were labeled with a fluorescent probe, Calcein AM (C2012, Beyotime, Haimen, China) according to the manufacturer’s instructions. Briefly, Calcein-labeled THP-1 cells were washed and then co-cultured with SK-Hep1 cells or human LSECs at 37°C for 1 hour. Removing nonadherent THP-1 cells, and the number of adherent cells was examined by a fluorescent microscope (Olympus, Tokyo, Japan) and analyzed by Image Pro Plus 8.0 (Media Cybernetics, Rockville, MD, USA).
Western blot
Liver tissue or cells were lysed in radioimmunoprecipitation assay (RIPA) buffer (Beyotime) supplemented with phosphatase inhibitors (Beyotime). Protein concentration was detected by a bicinchoninic acid kit (Beyotime). Western blot analysis was performed as previously described [28]. Primary antibodies targeting IL33 (Proteintech, Rosemont, IL, USA; 1:1,000; 12372-1-AP), ST2 (Proteintech, 1:1,000; 11920-1-AP), P62 (Cell Signaling Technology [CST], Danvers, MA, USA; 1:1,000; 5114T), microtubule-associated protein 1 light chain 3 B (CST, 1:1,000; 3868S), phosphor-extracellular signal-regulated kinase (p-ERK) (CST, 1:1,000; 4370T), ERK (CST, 1:2,000; 9108S), and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (CST, 1:2,000; 2118S). Western blot data in the figures were all representative of more than three independent experiments.
Statistical analysis
Graph plots and P values were generated using GraphPad Prism 8 software (GraphPad Software Inc., San Diego, CA, USA), and all data are shown as the mean±standard error of the mean from replicate experiments. For comparisons between groups for continuous variables, unpaired Student t-tests or one-way analysis of variance (ANOVA) test was performed. P value <0.05 was considered statistically significant. All experiments were repeated at least three times. More details on methods were provided in the Supplementary Methods.
RESULTS
Increased IL33 expression and liver injury in diabetic mice
The T2DM mice model with advanced liver disorders was established by STZ injection and HFD diet (Fig. 1A), and then measured diabetic liver injury. Compared with non-diabetic mice, diabetic mice showed a pale-yellow surface in the liver (Fig. 1B), and higher serum alanine transferase (ALT) and aspartate transaminase (AST) were detected in diabetic mice compared to controls (Fig. 1C). The diabetic mice also exhibited increased hepatic collagen deposition as shown by Sirius Red staining and hepatic hydroxyproline content (Fig. 1D-F). In addition, the data showed higher immunohistochemical staining score of alpha smooth muscle actin (αSMA) in diabetic mice compared with control mice (Fig. 1G and H). To determine the role of IL33 in diabetic liver injury, we first tested the level of circulating and hepatic IL33 levels by using enzyme-linked immunosorbent assay (ELISA) and Western blot, respectively. Both serum IL33 concentrations and hepatic IL33 expression were significantly elevated in diabetic mice (Fig. 2B and C). Together, these findings indicate that diabetic mice show liver injury and exhibit higher IL33 expression.
Diabetic mice exhibit severe liver injury and advanced fibrosis. (A) Schematic diagram of the experimental procedure, mice were fed with normal diet (ND) or high-fat diet (HFD) for a total of 26 weeks. After 12 weeks, HFD-fed mice were consecutively injected with streptozotocin (STZ, 50 mg/kg) for 5 days. Diabetic phenotype was validated by fasting blood glucose ≥11.1 mmol/L at 7 days later (14 weeks). (B) Macroscopic liver phenotype. (C) Serum levels of alanine transferase (ALT) and aspartate transaminase (AST) were measured in mice after 26 weeks of ND feeding or HFD challenge (n=8). (D) Representative images of H&E and Sirius Red staining of liver sections from control (Ctrl) mice and diabetic (DM) mice. Scale bar, 100 μm. (E) A positive area of Sirius Red staining (middle) is used to quantify liver fibrosis (n=6). (F) Hepatic hydroxyproline levels (right) in mice (n=6). (G, H) Representative images and quantify of immunohistochemical (IHC) staining of alpha smooth muscle actin (αSMA) in liver sections (n=4). Scale bar, 100 μm. Data was shown as mean±standard error of the mean. i.p., intraperitoneal. aP<0.01.
Diabetic liver fibrosis is alleviated in interleukin 33 (IL33)-deficient mice. (A) Schematic diagram of the experimental procedure, mice were fed with normal diet (ND) or high-fat diet (HFD) for a total of 26 weeks. After 12 weeks, HFD-fed mice were consecutively injected with streptozotocin (STZ, 50 mg/kg) for 5 days. Diabetic phenotype was validated by fasting blood glucose ≥11.1 mmol/L at 7 days later (14 weeks). HFD-fed mice were consecutively injected with αIL33, immunoglobulin G (IgG), recombinant IL33 (rIL33), or vehicle for 12 weeks. (B) Serum IL33 level in control (Ctrl) mice and diabetic (DM) mice assessed by enzyme-linked immunosorbent assay (ELISA) (n=12). (C) Western blot and measurement for IL33 expression of whole liver (n=6). (D) Representative images of H&E and Sirius Red staining of liver sections from DM mice and IL33 knockout (KO) diabetic mice. Scale bar, 100 μm. (E) A positive area of Sirius Red staining is used to quantify liver fibrosis (right) (n=4). (F) Hepatic hydroxyproline levels in mice (n=4). (G, H) Representative images and quantify of immunohistochemical (IHC) staining of alpha smooth muscle actin (αSMA) in liver sections (n=4). Scale bar, 100 μm. Data was shown as mean±standard error of the mean. ST2, suppression of tumorigenicity 2; i.p., intraperitoneal; PBS, phosphate buffered saline; GAPDH, glyceraldehyde-3-phosphate dehydrogenase. aP<0.01.
Role of IL33 in diabetic liver fibrogenesis
To investigate whether IL33 deficiency affects diabetic hepatic fibrosis progression in vivo, IL33−/− mice were also subjected to STZ injection and an HFD diet. After 26 weeks of feeding, IL33−/− mice displayed a significant reduction in collagen deposition as evidenced by Sirius Red staining and hepatic hydroxyproline content (Fig. 2D-F). Also, IL33−/− mice was lower in immunohistochemical staining score of αSMA (Fig. 2G and H). Additionally, we investigated the therapeutic effects of exogenous blocked IL33 by intraperitoneal injection of IL33 neutralizing antibody αIL33 (twice per week at a dose of 37 μg/kg body weight) (Fig. 2A, Supplementary Fig. 1A). Consistently, mice treated with αIL33 exhibited lower ALT and AST (Supplementary Fig. 1B), a marked reduction in extracellular collagen accumulation (Supplementary Fig. 1C-E), and lower immunohistochemical staining score of αSMA (Supplementary Fig. 1F and G). Our data suggest that IL33 deficiency alleviates hepatic fibrosis in diabetic mice.
IL33 contributes to LSEC dysfunction during liver fibrosis progression in diabetic mice
T2DM individuals are normally complicated with endothelial dysfunction, whether LSEC (highly specialized endothelial cells localized in liver sinusoid) injury occurs in diabetic mice remains unknown. We thus assessed hepatic NO concentration (a signal molecular produced by health LSEC) and lymphatic vessel endothelial hyaluronan receptor 1 (LYVE-1) expression (a marker of differentiated LSECs). Consistently, scanning electron microscopy images showed a loss of hepatic sinusoidal fenestrae in diabetic mice compared to non-diabetic mice (Fig. 3A), and diabetic mice exhibited lower hepatic NO levels, decreased LYVE-1 fluorescence intensity, and increased fluorescence intensity of CD31 (Fig. 3B-D). Also, mRNA expression of LSEC capillarization markers (endothelin 1 [Edn-1] and nitric oxide synthase 2 [Nos2]) and inflammation markers (vascular cell adhesion molecule 1 [Vcam1] and intercellular adhesion molecule 1 [Icam1]) were up-regulated, and differentiated phenotype marker (nitric oxide synthase 3 [Nos3] and KLF transcription factor 2 [Klf2]) were down-regulated in the diabetic liver (Fig. 3E). We further evaluated the role of IL33 in LSEC dysfunction during diabetic liver fibrosis progression. IL33 neutralizing antibody injection recovered sinusoidal fenestrae, LYVE-1 expression, and hepatic NO level in diabetic mice. Furthermore, diabetic mice subjected to IL33 neutralizing antibody exhibited blunted LSEC capillarization and inflammation based on the expression of hallmark genes (Fig. 3E). However, rIL33 injection did not aggravate LSEC dysfunction and inflammation in diabetic liver fibrosis (Fig. 3E). These findings indicate that diabetic liver is characterized by LSEC dysfunction, capillarization, and endothelial inflammation, and these changes are alleviated by endogenous IL33 neutralization.
Interleukin 33 (IL33) promotes liver sinusoidal endothelial cell (LSEC) dysfunction in diabetic mice. (A) Representative scanning electron micrographs of livers of mice. Scale bar, 5 μm. (B) Hepatic nitric oxide (NO) levels (n=8). (C, D) Representative immunofluorescent images and analysis of lymphatic vessel endothelial hyaluronan receptor 1 (LYVE-1) and CD31 of liver sections (n=4) (scale bar, 100 μm). (E) Relative mRNA abundance of vascular cell adhesion molecule 1 (Vcam1), intercellular adhesion molecule 1 (Icam1), nitric oxide synthase 2 (Nos2), endothelin 1 (Edn-1), nitric oxide synthase 3 (Nos3), and KLF transcription factor 2 (Klf2) of liver tissue (n=8). Data was shown as mean±standard error of the mean. Ctrl, control; DM, diabetes mellitus; IgG, immunoglobulin G; PBS, phosphate buffered saline; rIL33, recombinant IL33; NS, no significant. aP<0.05, bP<0.01.
Extracellular IL33 always binds with its receptor ST2 and activates downstream progress. Thus, we explored whether IL33/ST2 signaling has a role in the development of liver fibrosis and LSEC dysfunction in diabetic mice, we established diabetes in ST2 knockout mice. In our research, hepatic ST2 expression was significantly higher in diabetic mice (Supplementary Fig. 2A and B). We demonstrated that LSEC dysfunction and liver fibrosis were alleviated in ST2−/− diabetic mice as compared with wild-type diabetic mice (Supplementary Fig. 2C-G), indicating that diabetic liver fibrosis and LSEC dysfunction are diminished in the absence of ST2.
IL33 amplifies PA and high glucose-induced LSEC dysfunction in vitro
To further confirm the function of IL33 in LSEC dysfunction, human LSEC cell line SK-Hep1 cells were cultured with PAHG (0.2 mM PA and 30 mM glucose) for 24 hours. In the presence of PAHG condition, SK-Hep1 cells were treated by rIL33 at a dose of 1, 5, 10, 50, 100 ng/mL. Then, the NO production was quantified and compared. PAHG treatment down-regulated NO production in vitro, and rIL33 further aggravated NO decrease (Fig. 4C). To characterize IL33-induced LSEC dysfunction, we performed a monocyte adhesion assay by co-culturing Calcein AM-labeled THP-1 and SK-Hep1 cells for 1 hour. Fluorescence microscope analysis revealed that PAHG enhanced monocyte-endothelial adhesion in vitro, and the adhesion was augmented by rIL33 treatment (Fig. 4A and B). Besides, PAHG induced reactive oxygen species (ROS) production in SKHep1 cells and enhanced by rIL33 treatment (Fig. 4A and B). Furthermore, all these changes induced by PAHG plus rIL33 treatment were partially abrogated by small interfering RNA targeting ST2 (siST2) (siST2-1) transfection (Fig. 4G-I). Similar results were detected in human LSECs (Supplementary Fig. 3A-D). Taken together, IL33 amplifies LSEC dysfunction in the presence of PAHG.
Interleukin 33 (IL33) amplifies palmitic acid and high glucose-induced liver sinusoidal endothelial cell (LSEC) dysfunction in vitro. (A) Representative fluorescent images of adhered monocytes (upper), and microscopy images of reactive oxygen species (ROS) in LSEC (lower). (B) Analysis of adherent monocytes (upper) and average fluorescent intensity of ROS (lower) (n=4) (scale bar, 25 μm). (C) Nitric oxide (NO) levels of culture medium from LSEC administrated with recombinant IL33 in vitro (0, 1, 5, 10, 50, or 100 ng/mL) for 24 and 48 hours in the presence of palmitic acid (PA) plus high glucose (PAHG; 0.2 mM PA and 30 mM glucose) (n=4). (D) Schematic representation of Transwell coculture system. (E, F) Representative fluorescent images and analysis of alpha smooth muscle actin (αSMA) in LX-2 co-cultured with LSEC (n=4) (scale bar, 25 μm). (G) Western blot images and quantification of suppression of tumorigenicity 2 (ST2) from L02 cells transfected with small interfering RNA targeting ST2 (siST2) or scramble RNA. (H) Representative fluorescent images of adhered monocytes (upper), and microscopy images of ROS in LSEC (lower) after transfected with siST2 or scramble RNA (n=3) (scale bar, 25 μm). (I) Representative fluorescent images and analysis of αSMA in LSEC transfected with siST2 or scramble RNA (n=3) (scale bar, 25 μm). (J) Representative fluorescent images and analysis of αSMA in LX-2 cultured with culture medium collected from LSECs transfected with siST2 or scramble RNA (n=3) (scale bar, 25 μm). Data was shown as mean±standard error of the mean. NG, normal glucose; NS, no significant; HSC, hepatic stellate cell; DAPI, 4’,6-diamidino-2-phenylindole; wt, wild-type; GAPDH, glyceraldehyde-3-phosphate dehydrogenase. aP<0.05, bP<0.01.
IL33-induced LSEC dysfunction promotes HSC activation in vitro
To investigate whether IL33-induced LSEC dysfunction affects HSC activation in vitro, we co-cultured a human HSC cell line, LX-2, with LSECs pretreated with rIL33 and PAHG for 48 hours (Fig. 4D). LX-2 co-cultured with rIL33-pretreated LSECs displayed augmented activation as demonstrated by immunoblot staining of αSMA (Fig. 4E and F), and similar results were detected in human LSECs (Supplementary Fig. 3E and F). Additionally, siST2 transfection combated the increasing of fluorescence intensity of αSMA in LX-2 (Fig. 4J). These data indicates that IL33-induced LSEC dysfunction promotes HSC activation in vitro.
IL33 elicits LSEC dysfunction by blunting autophagy
To mechanistically dissect the role of IL33 in LSEC dysfunction, we monitored autophagy levels in LSECs in vitro. PAHG resulted in an increase in LC3II and p62, which was aggravated by rIL33 treatment (Fig. 5A). We next sought to determine whether autophagy inhibition is associated with IL33-driven LSEC dysfunction and the following HSC activation. Autophagy agonist rapamycin attenuated autophagy blockade induced by rIL33 combined PAHG treatment in LSECs (Fig. 5A). Additionally, rapamycin treatment alleviated IL33-induced NO reduction (Fig. 5B) and partially decreased monocyte-endothelial adhesion (Fig. 5C and D) in LSECs in the presence of PAHG. We further monitored the oxidative stress in LSECs, and rapamycin also attenuated IL33-driven ROS production in LSECs in the presence of PAHG (Fig. 5C and D). Then we assessed the effect of IL33-induced autophagy inhibition in LSECs on HSC activation. Rapamycin administration in LSECs exhibited a decreased average intensity of αSMA, which is an indication of HSC activation inhibition (Fig. 5E and F). Most results were verified in human LSECs (Supplementary Fig. 4). Collectively, IL33 drives LSEC dysfunction and the following HSC activation via inhibiting autophagy.
Interleukin 33 (IL33) drives liver sinusoidal endothelial cell (LSEC) dysfunction by blocking autophagy in vitro. (A) Western blot and measurement for light chain 3 (LC3) and p62 of LSEC (n=4). (B) Nitric oxide (NO) levels of culture medium from LSEC cultured with different administration (n=4). (C) Representative fluorescent images of adhered monocytes (upper), and microscopy images of reactive oxygen species (ROS) in LSEC (lower) (scale bar, 25 μm). (D) Analysis of adherent monocytes number (left) (n=4) and average fluorescent intensity of ROS in LSEC (right) (n=4). (E, F) Representative fluorescent images and analysis of alpha smooth muscle actin (αSMA) in LX-2 (scale bar, 25 μm). Data was shown as mean±standard error of the mean. PAHG, palmitic acid plus high glucose; Rapa, rapamycin; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; NG, normal glucose; NS, no significant; DAPI, 4’,6-diamidino-2-phenylindole. aP<0.05, bP<0.01.
IL33 drives endothelial-mesenchymal transition in LSEC by activating ERK/MAPK pathways
Previous researches illustrated that mitogen-activated protein kinase (MAPK) pathway is involved in the process downstream of IL33 stimulation and ST2 activation [15]. In our work, Western blotting analysis indicated that ERK was activated by IL33 stimulation in the presence of PAHG (Fig. 6A and B). To examine whether this pathway mediates IL33-induced autophagy inhibition and LSEC dysfunction, ERK inhibitor PD98059 was used to incubate LSECs in the presence of IL33 and PAHG. Pharmacological inhibition of the ERK/MAPK pathway reversed the increase in p62 expression in LSECs after IL33 and PAHG incubation (Fig. 6A). In addition, ERK/MAPK pathway inhibition alleviated IL33-induced NO reduction, monocyte adhesion, and ROS production in LSECs (Fig. 6C and D). Previous studies demonstrated that the crosstalk between LSECs and HSCs is mediated by the paracrine signals from them [20,29]. Thus, we cultured LX-2 cells with conditioned medium (CM) collected from different groups of LSECs. As expected, ERK/MAPK pathway inhibition reversed HSC activation induced by CM from LSECs treated with IL33 and PAHG (Fig. 6E). Furthermore, similar data were exhibited in human LSECs (Supplementary Fig. 5A-F).
Interleukin 33 (IL33) elicits liver sinusoidal endothelial cell (LSEC) dysfunction by activating the extracellular signal-regulated kinase 1 (ERK1)/mitogen-activated protein kinase (MAPK) pathway. (A, B) Western blot and measurement for p-ERK, ERK, light chain 3 (LC3), and p62 of LSEC (n=3). (C) Nitric oxide (NO) levels of culture medium from LSEC cultured with different administration (n=4). (D) Representative fluorescent images and analysis of adhered monocytes (upper), and microscopy images of reactive oxygen species (ROS) in LSEC (lower) (scale bar, 25 μm). (E) Representative fluorescent images and analysis of alpha smooth muscle actin (αSMA) in LX-2 cultured with culture medium collected from different groups of LSECs (n=4) (scale bar, 25 μm). (F) Representative fluorescent images and analysis of αSMA in LSEC cultured with different administration (n=4) (scale bar, 25 μm). Data was shown as mean±standard error of the mean. GAPDH, glyceraldehyde-3-phosphate dehydrogenase; NG, normal glucose; NS, no significant; PAHG, palmitic acid plus high glucose; DAPI, 4’,6-diamidino-2-phenylindole. aP<0.05, bP<0.01.
Intriguingly, it is determined that partial LSECs undergo an endothelial-mesenchymal transition (EndoMT) to drive extracellular matrix deposition in chronic liver disease [30]. Thus, we examined whether IL33 promotes LSECs acquiring a myofibroblast-like phenotype. Immunoblot staining of αSMA exhibited that IL33 aggravated PAHG-induced αSMA expression in LSECs, which was combated by PD98059 treatment (Fig. 6F, Supplementary Fig. 5G and H). This indicates that IL33 promotes EndoMT by activating ERK/MAPK pathway in LSECs.
DISCUSSION
In our work, we identified the role of IL33 on LSEC dysfunction and its implications in the development of diabetic liver fibrosis. IL33 is elevated in the circulation and the liver in animal models with late-stage of T2DM. Moreover, in vivo studies indicated that IL33 knockout or αIL33 administration alleviates diabetic liver fibrosis and LSEC inflammation, and rIL33 treatment favors LSEC dysfunction and HSC activation in vitro. Further, IL33 elicits LSEC dysfunction by inhibiting autophagy via activating ERK/MAPK pathway. These novel findings expand our knowledge about the molecular mechanisms involved in diabetic liver fibrosis.
T2DM is an established risk factor for the progression of NAFLD to advanced hepatic fibrosis [4]. We identified severe hepatic fibrosis in diabetic mice. NAFLD individuals with metabolic dysfunction are normally complicated with chronic and low-grade inflammation [31]. As an inflammatory factor, IL33 has been implicated in T2DM pathogenesis [14,32]. A progressive significant increase in IL33 levels has been examined in diabetic mice [14,32], obese individuals [33], and NAFLD individuals [34]. Similarly, we detected increased serum IL33 and hepatic IL33 expression in diabetic mice. Despite IL33 playing an undetermined role in obesity and T2DM [14,35], it is determined that IL33 plays a role in organ fibrosis (the lung, the heart, the liver, etc.) since this kind of cytokine has long been known to promote wound healing and tissue repair [10]. In our research, we determined that IL33 promotes diabetic liver fibrosis. IL33 knockout mice were challenged to the HFD-fed plus STZ injection, besides, αIL33 was injected into diabetic mice. As expected, both endogenous IL33 deficiency and αIL33 injection alleviated hepatic fibrosis in diabetic mice. Similarly, 11 weeks subcutaneous injection of anti- IL33 receptor antibody decreased αSMA-positive myofibroblasts in diabetic mice modeled by 16 weeks HFD feeding and STZ injection [36], indicating that IL33 drives hepatic fibrosis in diabetic mice. IL33 being reported to trigger experimental hepatic fibrosis by activating liver resident ILC2 [13] and directly activating HSCs [15], while the role of IL33 in diabetic liver fibrosis is unknown. Our data suggest that IL33 promotes hepatic fibrosis in diabetic mice.
Of note, chronic vascular complications are major clinical problems in T2DM [37,38]. Endothelial dysfunction occurs early and plays a major role in the pathophysiology of T2DM-associated macrovascular and microvascular complications [39]. In our work, diabetic animals developed significant endothelial dysfunction and inflammation. Similarly, PA and HG (mimic the diabetic microenvironment) treatment triggered LSEC dysfunction in vitro. The pathogenesis of endothelial dysfunction in T2DM is intricate and multifactorial [37-39]. Our data exhibited that both serum and hepatic IL33 levels were elevated in diabetic mice. Furthermore, both endogenous IL33 deficiency, IL33 neutralization, and ST2 deficiency alleviated hepatic endothelial inflammation in diabetic animals, and rIL33 treatment exacerbated LSEC dysfunction in vitro. As highly specialized endothelial cells localized at the liver sinusoid, LSECs direct contact with blood and liver cells and sense changes in circulation and intrahepatic microenvironment at the same time [40]. Similarly, IL33 is previously shown to induce angiogenesis and inflammatory activation of endothelial cells [25,41]. Our results are the first to implicate that IL33 triggers LSEC inflammation and dysfunction through the ST2-dependent pathway in the diabetic liver.
Recently, it has been determined that LSEC dysfunction is an early event that precedes and contributes to fibrosis in NAFLD [18,19]. Considering that LSECs play a pivotal role in chronic liver disease initiation and progression [20,21]. We detected whether IL33-induced endothelial dysfunction contributes to liver fibrosis. Our data clarified that IL33 induces LSEC dysfunction and drives diabetic liver fibrosis. In addition, coculture with IL33-treated LSECs facilitated HSC activation in vitro, supporting the concept that IL33 triggers diabetic liver fibrosis by promoting LSEC dysfunction. LSECs display permeable properties since numerous fenestrae are located on their surface [21]. The loss of fenestrae, also called defenestration or capillarization, is an early and remarkable characteristic of LSEC dysfunction. In our work, we identified that the LSEC capillarization occurred in the diabetic liver was mediated by the IL33/ST2 axis. Except for IL33, other macromolecules such as gut microbiota-derived products may also contribute to LSEC capillarization in NAFLD [42,43]. These triggers probably drive oxidative injury in endothelial cells [38]. Autophagy is a constitutive process to maintain cellular homeostasis by eliminating damaged cellular material and controlling ROS production. It has been identified to promote LSEC defenestration and thus prevent HSC activation [16,44,45]. Notably, LSEC autophagy deficiency promotes fibrosis in NAFLD individuals [16]. In our study, autophagy was blocked in PA and HG-treated endothelial cells in vitro, which was exacerbated by IL33 administration. Additionally, autophagy activator rapamycin treatment abrogated ROS production and autophagy disruption in LSECs, and the following dysfunction induced by IL33 administration in vitro. Although the precise mechanisms of how IL33 contributes to autophagy in LSECs are still incompletely understood, we can speculate that the following pathways play a role. The results of the Gene Set Enrichment Analysis demonstrated that IL33 was closely related to MAPK pathway in hepatocellular carcinoma [46]. Furthermore, MAPK signaling was dose-dependently activated by rIL33 stimulation [15]. Interestingly, the MAPK pathway not only functions as a negative modulator of autophagy [47] but is also crucial for paracrine signal production in LSECs [29,42]. In our work, IL33 activated the ERK/MAPK signaling pathway and thus induced LSEC dysfunction by inhibiting autophagy. Moreover, pharmacological inhibition of the ERK/MAPK pathway blunted IL33-induced EndoMT in LSECs and associated HSC activation. Nonetheless, several studies have implied other mechanisms account for how IL33 regulates autophagy such as B cell lymphoma-2 [48,49] and ST2/phosphoinositide 3-kinase (PI3K)/mammalian target of rapamycin (mTOR) pathway [50]. Collectively, despite the mechanisms of IL33-induced LSECs autophagy block and dysfunction in the diabetic liver are not fully identified and need further investigation, one can certainly be confirmed that ERK/MAPK pathway contributes to IL33-induced LSEC dysfunction in vitro.
In conclusion, our work underscores the importance of IL33/ST2 signaling in the development of diabetic liver fibrosis. In T2DM-associated liver disorder, IL33 contributes to LSEC dysfunction and promotes liver fibrosis, which is related to the ERK/MAPK pathway-mediated autophagy regulation. Our findings demonstrate that IL33 is an attractive therapeutic target for diabetic liver fibrosis.
SUPPLEMENTARY MATERIALS
Supplementary materials related to this article can be found online at https://doi.org/10.4093/dmj.2024.0532.
Interleukin 33 (IL33) neutralization attenuates diabetic liver fibrosis. (A) Serum IL33 level in diabetic (DM) mice and DM mice treated with IL33 neutralizing antibody (αIL33) assessed by enzyme-linked immunosorbent assay (ELISA) (n=7). (B) Serum levels of alanine transferase (ALT) and aspartate transaminase (AST) were measured in DM mice and mice treated with αIL33 (n=7). (C) Hepatic hydroxyproline levels in mice (n=6). (D) Representative images of H&E and Sirius Red staining of liver sections from DM mice and DM mice treated with αIL33 (scale bar, 100 μm). (E) A positive area of Sirius Red staining is used to quantify liver fibrosis (n=6). (F, G) Representative images and quantify of immunohistochemical (IHC) staining of alpha smooth muscle actin (αSMA) in liver sections (n=4) (scale bar, 100 μm). Data was shown as mean±standard error of the mean. IgG, immunoglobulin G. aP<0.05, bP<0.01.
Liver sinusoidal endothelial cell (LSEC) dysfunction and liver fibrosis are reduced in diabetic mice in the absence of suppression of tumorigenicity 2 (ST2). (A, B) Western blot and measurement for ST2 of the whole liver (n=6). (C) Representative images of H&E and Sirius Red staining of liver sections from wild-type diabetic (DM) mice and ST2 knockout (KO) DM mice (left). Positive area of Sirius Red staining is used to quantified liver fibrosis (right) (n=4) (scale bar, 100 μm). (D) Hepatic hydroxyproline levels in mice (n=4). (E) Representative images and quantify of immunohistochemical (IHC) staining of alpha smooth muscle actin (αSMA) in liver sections (n=4) (scale bar, 100 μm). (F) Representative immunofluorescent images (left) and analysis (right) of vessel endothelial hyaluronan receptor 1 (LYVE-1) of liver sections (n=4) (scale bar, 100 μm). (G) Hepatic nitric oxide (NO) levels (n=6). Data was shown as mean±standard error of the mean. Ctrl, control; GAPDH, glyceraldehyde-3-phosphate dehydrogenase. aP<0.05, bP<0.01.
Interleukin 33 (IL33) amplifies palmitic acid (PA) and high glucose-induced human liver sinusoidal endothelial cell (LSEC) dysfunction in vitro. (A, B) Representative fluorescent images and analysis of adhered monocytes (upper), and microscopy images of reactive oxygen species (ROS) (C) in human LSECs (lower) (n=3) (scale bar, 25 μm). (D) Nitric oxide (NO) levels of culture medium from human LSECs administrated with recombinant 50 ng/mL IL33 in vitro for 24 hours in the presence of palmitic acid plus high glucose (PAHG) (0.2 mM PA and 30 mM glucose) (n=4). (D) Schematic representation of Transwell coculture system. (E, F) Representative fluorescent images and analysis of alpha smooth muscle actin (αSMA) in LX-2 co-cultured with human LSECs (n=3) (scale bar, 25 μm). Data was shown as mean±standard error of the mean. NG, normal glucose; DAPI, 4’,6-diamidino-2-phenylindole. aP<0.05, bP<0.01.
Interleukin 33 (IL33) drives human liver sinusoidal endothelial cell (LSEC) dysfunction by blocking autophagy in vitro. (A, B, C) Representative fluorescent images and analysis of adhered monocytes (upper), and microscopy images of reactive oxygen species (ROS) in human LSECs (lower) (scale bar, 25 μm; n=3). (D) Nitric oxide (NO) levels of culture medium from human LSECs cultured with different administration (n=3). (E, F) Representative fluorescent images and analysis of alpha smooth muscle actin (αSMA) in LX-2 (scale bar, 25 μm; n=3). Data was shown as mean±standard error of the mean. NG, normal glucose; PAHG, palmitic acid plus high glucose; Rapa, rapamycin; NS, no significant; DAPI, 4’,6-diamidino-2-phenylindole. aP<0.05, bP<0.01.
Interleukin 33 (IL33) elicits human liver sinusoidal endothelial cell (LSEC) dysfunction by activating the extracellular signal-regulated kinase 1 (ERK1)/mitogen-activated protein kinase (MAPK) pathway. (A, B, C) Representative fluorescent images and analysis of adhered monocytes (upper), and microscopy images of reactive oxygen species (ROS) in human LSECs (lower) (scale bar, 25 μm). (D) Nitric oxide (NO) levels of culture medium from human LSECs cultured with different administration (n=3). (E, F) Representative fluorescent images and analysis of alpha smooth muscle actin (αSMA) in LX-2 cultured with culture medium collected from different groups of human LSECs (n=3) (scale bar, 25 μm). (G, H) Representative fluorescent images and analysis of αSMA in human LSECs cultured with different administration (n=3) (scale bar, 25 μm). Data was shown as mean±standard error of the mean. NG, normal glucose; PAHG, palmitic acid plus high glucose; DAPI, 4’,6-diamidino-2-phenylindole. aP<0.05, bP<0.01.
Notes
CONFLICTS OF INTEREST
No potential conflict of interest relevant to this article was reported.
AUTHOR CONTRIBUTIONS
Conception or design: H.C., Y.T.
Acquisition, analysis, or interpretation of data: L.M., X.Y., L.C., B.W., C.L., B.X.
Drafting the work or revising: H.C., C.C., Y.Z., D.L., P.Y., Y.T.
Final approval of the manuscript: all authors.
FUNDING
This work was supported by Applied Basic Frontier Special Project of Wuhan Science and Technology Bureau (2020020601012246) and National Natural Science Foundation of China (NO.81602858).
ACKNOWLEDGMENTS
We sincerely appreciate the Department of Immunology, Tongji Medical College for providing IL33−/− and ST2−/− mice. We also thank Prof. Zhou Ping from Huazhong University of Science and Technology for the kindly gifted cell lines.