Celastrol alleviates esophageal stricture in rats by inhibiting NLR family pyrin domain containing 3 activation

Abstract

BACKGROUND

The role of NLR family pyrin domain containing 3 (NLRP3) in post-endoscopic submucosal dissection (ESD) esophageal stricture remains incompletely understood. The effect of celastrol (CEL) on the prevention of esophageal strictures has not yet been investigated.

AIM

To explore the effect of CEL on the prevention of esophageal stricture in rats.

METHODS

NLRP3, interleukin (IL)-1β, and IL-18 mRNA levels were measured in patients’ tissues after esophageal ESD. NLRP3 expression in esophageal fibroblasts was determined using immunohistochemistry and immunofluorescence staining. Lentiviral transfection was used to induce NLRP3 overexpression and thioredoxin reductase 1 (TXNRD1) silencing. The CCK8 assay was used to determine the optimal CEL concentration. Reactive oxygen species (ROS) generation was detected via fluorescence and flow cytometry. Masson’s trichrome staining and barium esophagography were performed to assess collagen deposition and esophageal stenosis.

RESULTS

The mRNA levels of NLRP3 and IL-1β were higher in human tissues from the ESD resection bed than in normal esophageal mucosa. NLRP3 overexpression in primary rat esophageal fibroblasts led to high collagen 1 expression. Thus, NLRP3 participated in esophageal inflammation and tissue repair after ESD. Comparable to prednisolone, CEL significantly inhibited NLRP3 activation in vitro and in vivo, and esophageal strictures were markedly alleviated. Mechanistically, CEL upregulated TXNRD1 expression and reduced ROS production, thereby inhibiting NLRP3 expression. This effect was reversed by TXNRD1 silencing. Furthermore, TXNRD1 interacted with NLRP3 and promoted its ubiquitination.

CONCLUSION

CEL is a promising alternative therapeutic agent for the prevention of post-ESD esophageal strictures.

Key Words: Esophageal stricture after endoscopic submucosal dissection; Celastrol; NLR family pyrin domain containing 3 inflammasome; Early-stage esophageal neoplasm; Natural medicine

Core Tip: NLR family pyrin domain containing 3 (NLRP3) and interleukin-1β are increased in the endoscopic submucosal dissection resection bed of patients with esophageal neoplasms. Celastrol (CEL) prevented esophageal stricture in rats, and the effects were comparable to the positive control, prednisolone. Thioredoxin reductase 1 (TXNRD1) was increased by CEL in lipopolysaccharide plus adenosine triphosphate stimulated primary rat esophageal fibroblasts. Independent of its effect on reactive oxygen species, TXNRD1 interacts with NLRP3 and promotes its ubiquitination.

INTRODUCTION

Esophageal strictures negatively affect patients' quality of life and may lead to severe complications[1]. With the development of endoscopic technology, endoscopic submucosal dissection (ESD) has become the standard therapy for superficial esophageal neoplasms in clinical practice[2]. Consequently, the incidence of esophageal strictures after endoscopic treatment for early esophageal cancer has increased, posing a major clinical challenge.

Following endoscopic therapy, the esophageal wall is prone to severe inflammation and ulceration due to irritation from food passage and exposure to digestive fluids, such as saliva, and gastric acid reflux[3]. Inflammation potently induces transforming growth factor-β signaling, activates fibroblasts, and induces tissue fibrosis[4]. Excessive fibrosis and scar formation during the healing process are the most commonly recognized causes of esophageal strictures[5-7]. To prevent esophageal strictures, inflammation at mucosal defects should be minimized to limit damage to the muscularis propria and prevent excessive fibrosis[8].

Many inflammation-associated molecules are involved in the process of injury repair. When tissue injury occurs, innate immune cells express damage-associated molecular patterns (DAMPs); DAMPs are recognized by inflammasomes and activate subsequent cascade pathways, triggering the inflammatory response[9]. NLR family pyrin domain containing 3 (NLRP3) is the most widely recognized inflammasome[10]. It has been reported that NLRP3 promotes esophageal stricture formation following ulcer healing in rats[11]. However, no studies have confirmed the expression of NLRP3 in human esophageal tissues after ESD.

Celastrol (CEL), a pentacyclic triterpenoid derived from Tripterygium wilfordii Hook F, has shown anti-inflammatory effects in several inflammatory diseases[12]. However, CEL has not been applied in esophageal stricture prevention. Here, we examined NLRP3, interleukin (IL)-1β, and IL-18 mRNA in human tissues after esophageal ESD procedure. Then, the effect of CEL was investigated in vitro and in vivo.

MATERIALS AND METHODS

ESD procedures and data collection

The protocols were approved by the Ethics Committee of Tongji Medical College of Huazhong University of Science and Technology (No. S167). From November 2023 to May 2024, 21 consecutive patients were included. Inclusion criteria was as follows: (1) Adult patients aged 18 to 60 years; (2) Willingness to comply with all study procedures; (3) Provision of written informed consent; (4) Initial endoscopic diagnosis suggesting early esophageal cancer requiring ESD treatment; and (5) Discontinuation of antiplatelet and anticoagulant medications for at least 1 week, absence of cardiovascular diseases, diabetes, hepatic or renal insufficiency, and good tolerance to anesthesia.

The ESD procedure has been described previously[13]. Following a previously published method[14], after completing the resection, three mucosal biopsies were collected from the resection bed, and from healthy mucosa which is 2 cm proximal to the resection bed, respectively, with the patients’ written informed consent, under the precondition of no obvious bleeding in the resection bed. During the first month following the resection, all patients took a double-dose proton pump inhibitors. Clinical characteristics of the patients are shown in Table 1. All esophageal biopsies were immediately fixed in 10% formalin for histological analysis or stored at -80 °C for further RNA extraction.

Table 1 Clinical characteristics of 21 patients.

Label	Gender	Age, years	Lesion number	Longitudinal diameter (cm)	Circumferential diameter (cm)	Histopathology
1	Female	53	3	11	4	Carcinoma in situ
2	Female	46	4	3	2	Carcinoma in situ
3	Male	54	2	9	2	HGIN
4	Female	53	1	2	2	HGIN
5	Female	47	5	9	5	Squamous carcinoma (M2)
6	Male	55	1	6	4	Squamous carcinoma (M2)
7	Male	56	1	6	5	Squamous carcinoma (M2)
8	Male	45	1	3	4	carcinoma in situ
9	Male	57	1	3	3	carcinoma in situ
10	Male	56	1	3	3	HGIN
11	Female	55	1	3	3	Squamous carcinoma (M3)
12	Male	53	3	3	2	Squamous carcinoma (M2)
13	Female	57	1	3	2.5	Squamous carcinoma (M2)
14	Male	65	1	2	2.5	HGIN
15	Male	67	1	3	2	HGIN
16	Male	62	2	4	3	Squamous carcinoma (M3)
17	Male	71	1	3	2	Squamous carcinoma (M2)
18	Male	56	1	3	3	carcinoma in situ
19	Male	62	1	5	3	Squamous carcinoma (M2)
20	Male	61	1	5	2	Squamous carcinoma (M3)
21	Male	73	1	3	2	Squamous carcinoma (M2)

M2: Epithelium; M3: Lamina propria mucosa; HGIN: High-grade intraepithelial neoplasia.

Immunohistochemistry

Paraffin-embedded human and rat tissue sections complete antigen retrieval and blocking. Then, sections were incubated with primary antibodies (NLRP3, SAB5700723, Sigma-Aldrich) overnight and secondary antibody for 1 hour. Subsequently, the slides were stained and dehydrated. Quantitative analysis was performed using Image J software by at least two experimenters who were blinded to the subgroup details.

Immunofluorescence staining

Paraffin-embedded human and rat tissue sections were dewaxed and hydrated. After antigen retrieval, the sections were further incubated overnight with primary antibodies against vimentin (SAB5700070, Sigma-Aldrich) and NLRP3 (SAB5700723, Sigma-Aldrich), and then stained with Alexa Fluor 647 and Alexa Fluor 488-conjugated secondary antibodies. DAPI was used for nuclear counterstaining.

Differences in colocalization between the control and model groups were analyzed by comparing the Pearson correlation coefficient for each two-channel comparison[15]. It is calculated using Image J software.

Colocalization of thioredoxin reductase 1 (TXNRD1) (sc-28321, Santa Cruz Biotechnology) and NLRP3 (SAB5700723, Sigma-Aldrich) in primary rat esophageal fibroblasts (REFs) was also detected by immunofluorescence (IF) staining.

Hematoxylin-eosin and Masson trichrome staining

Hematoxylin-eosin (HE) staining and Masson trichrome staining were used to detect the presence of esophageal fibroblasts and the degree of fibrosis. Rat esophageal tissues on day 9 were embedded in paraffin and sectioned at a thickness of 3 μm for HE and Masson trichrome staining. For HE staining, the fibroblast cell count was determined using Image J software. For Masson’s staining, the degree of fibrosis in the stained sections was quantitatively evaluated using Image-Pro Plus 6.0. All quantitative analyses were performed by at least two experimenters who were blinded to the subgroup details.

Animals and reagents

The sample size was identified by resource equation. Due to our study design, 4-6 rats were needed in each group. Rats were numbered and divided into 5 groups according to a randomized blocked design (assigned using R package “randomizr”).

Specific-pathogen-free male Sprague-Dawley rats (200-230 g, 6 weeks old) (Shulaibao, Wuhan, China) were fasted for 12 hours and subjected to laparotomy under inhalation anesthesia with isoflurane (4.0 L/minutes for induction, and 2.0 L/minutes for maintenance). All experimental procedures were approved by the Institutional Animal Care and Use Committee of Huazhong University of Science and Technology (No. 4016).

Esophageal ulcers were induced as previously described[11]. Rats that underwent laparotomy but not induction of esophageal ulcers were used as negative controls for analysis. The rats were divided into five groups and given the following reagents via gastric lavage: (1) Sham group: 0.5% sodium carboxymethyl cellulose; (2) Vehicle group: 0.5% sodium carboxymethyl cellulose; (3) CEL group: CEL (1.5 mg/kg/day; S1290, Selleck) suspended in vehicle; (4) Prednisolone (PRED) group (as the positive control): PRED (3 mg/kg/day; HY-17463, MedChemExpress) suspended in vehicle; and (5) Combination therapy group: CEL (1.5 mg/kg/day) plus PRED (3 mg/kg/day) suspended in vehicle. All reagents were administered once daily starting on day 3. The rats were euthanized on day 6 and day 9, to analyze inflammation and fibrosis, respectively.

Body weight was recorded on days 1, 3, 5, 7, and 9. To assess esophageal stricture, barium esophagography was performed on day 9 under intraperitoneal anesthesia with 2% tribromoethanol by a radiologist who was blinded to the subgroup details.

Animal care and use statement

Every effort was made to minimize their suffering. Rats were acclimatized to laboratory conditions for 2 weeks prior to experimentation. Food and water were sufficient. Intragastric gavage administration was carried out with No. 16 straight gavage needles. All rats were euthanized by 2% tribromoethanol overdose for tissue collection.

Primary REF isolation and cell culture

Primary REF cultures were established from Sprague-Dawley rats (3-5 days old), as previously described[16]. Immunohistochemistry (IHC) was performed to evaluate the expression of vimentin (SAB5700070, Sigma-Aldrich) and cytokeratin 19 (ab220193, Abcam). Passages 3-5 of primary esophageal fibroblasts were used in subsequent experiments.

Treatment of primary REFs

REFs were divided into five groups: (1) Control group: PBS was used as negative control; (2) NLRP3 activation group: The NLRP3 inflammasome primer, lipopolysaccharide (LPS) (1 μg/mL, L2630, Sigma-Aldrich) for 6 hours and followed by the activator, adenosine triphosphate (ATP) (5 mmol/L, HY-B2176, MedChemExpress) for 30 min to ensure activation of the NLRP3 inflammasome; (3) CEL group: CEL (200 nM, S1290, Selleck) for 1 hour followed by the LPS plus ATP stimulation; (4) PRED group: PRED (1 μM, HY-17463, Medchemexpress) for 1 hour followed by the LPS plus ATP stimulation; and (5) Combination therapy group: CEL (200 nM, S1290, Selleck) and PRED (1 μM, HY-17463, MedChemExpress) for 1 hour followed by the LPS plus ATP stimulation.

CCK8 assay

To assess the effect of CEL on primary REFs, CCK8 (HY-K0301, MedChemExpress) assay was performed. Cells were cultured with Dulbecco’s modified Eagle’s medium (DMEM) containing PBS, dimethyl sulfoxide, or CEL (200 nM, 1 μM, 5 μM, and 10 μM) for 24, 48, and 72 hours. Subsequently, the CCK8 solution was added to each well for 4 hours, and absorbance at 450 nm was measured.

Lentiviral vector preparation and infection

NLRP3 overexpression was achieved via lentivirus transfection (pLent-EF1a-FH-CMV-GFP-P2A-Puro, purchased from Weizhen Biology, Wuhan). TXNRD1 knockdown (HBLV-r-TXNRD1-shRNA1-ZsGreen-Puro) and overexpression (HBAD-Adeasy-r-TXNRD1-3xflag-EGFP) were achieved via lentivirus transfection (purchased from Hanheng Biology, Wuhan). Primary REFs were seeded and incubated with either lentiviral particles or control lentiviral particles along with polybrene in the DMEM medium (multiplicity of infection = 150) when the cells reached 30% confluence. Transfection efficiency was verified after 72 hours by fluorescence. In the primary REFs, the aforementioned lentivirus showed a transfection efficiency of over 90%.

Reactive oxygen species assay

Reactive oxygen species (ROS) analysis kit (HY-D0940, MedChemExpress) was used to measure the ROS levels in each group after different treatments. Primary REFs were divided into three groups: (1) Control group: PBS as negative control; (2) LPS group: LPS (1 μg/mL) for 6 hours and followed by ATP (5 mmol/L) for 30 minutes; and (3) CEL + LPS group: CEL (200 nM) for 1 hour followed by the LPS plus ATP stimulation. Cells were incubated with DCFH-DA and observed under a fluorescence microscope. The fluorescence intensity of ROS in each group was quantified. For flow cytometry analysis, after incubation with DCFH-DA, cells were digested and resuspended in PBS.

TXNRD activity assay

TXNRD activity was measured using the thioredoxin reductase assay kit (BC1155, Solarbo) following the manufacturer’s protocol. Primary REFs were divided into three groups: (1) Control group: PBS as negative control; (2) LPS group: LPS (1 μg/mL) for 6 hours and followed by ATP (5 mmol/L) for 30 minutes; and (3) CEL + LPS group: CEL (200 nM) for 1 hour followed by the LPS plus ATP stimulation. TXNRD activity was measured through the absorbance at 412 nm.

Western blot

Protein samples were transferred onto a PVDF membrane. The membrane was blocked for 1 hour. After washing with Tris-buffered saline containing Tween-20, the membrane was incubated overnight at 4 °C with the following primary antibodies: NLRP3 (ab263899, Abcam), IL-1β (ab283818, Abcam), Caspase1 (2225, Cell Signaling Technology), nuclear factor erythroid 2-related 2 (Nrf2) (A21176, ABclonal), TXNRD1 (11117-1-AP, Proteintech), GAPDH (ET1601-4, Huabio) and Collagen 1 (COL1) (14695-1-AP, Proteintech).

Reverse transcription and quantitative real-time PCR

Total RNA was extracted from tissues and cultured cells using TRIzol reagent, followed by reverse transcription (RR037A, TaKaRa). PCR was conducted in a reaction volume of 20 μL using PCR mix (RR820A, TaKaRa), and PCR amplification was performed using a PCR cycler (Thermofisher). Primer sequences are listed in Table 2.

Table 2 Primers for qPCR.

	Gene name	Forward primer	Reverse primer
Rat	ACTB	TTCGCCATGGATGACGATATC	TAGGAGTCCTTCTGACCCATAC
	NLRP3	GACTGCCGTCTATGTCTTTTTC	GAAGCTGTAAAATCTCTCGCAG
	IL1β	AACTGTGAAATAGCAGCTTTCG	CTGTGAGATTTGAAGCTGGATG
	COL1	AAAGATGGACTCAACGGTCTC	CAGGAAGCTGAAGTCATAACCA
Human	ACTB	CCAGAGACAGTTATGCGAATTG	TTCTGGGAATTCGGGAACATAA
	NLRP3	GGCAAATTCGAAAAGGGGTATT	CTGATTTGCTGAGAGATCTTGC
	IL1β	GCCAGTGAAATGATGGCTTATT	AGGAGCACTTCATCTGTTTAGG
	IL18	GCTGAAGATGATGAAAACCTGG	CAAATAGAGGCCGATTTCCTTG

ACTB: Β-actin; NLRP3: NLR family pyrin domain containing 3; IL-1β: Interleukin-1β; COL1: Collagen 1.

Co-immunoprecipitation assay

Cells in the control and TXNRD1-overexpression groups were collected and lysed with lysis buffer. The supernatant fractions were collected and incubated with control immunoglobulin G or primary antibody overnight, followed by incubation with resuspended Protein A/G beads (HY-K0202, MedChemExpress) to pull down bound proteins. Immunoprecipitated proteins were separated by SDS-PAGE after western blotting. The following primary antibodies were used: Anti-ubiquitin (3936S, Cell Signaling Technology), anti-NLRP3 (ab263899, Abcam), anti-TXNRD1 (sc-28321, Santa Cruz Biotechnology), anti-GFP (AE078, Abclonal) and anti-GAPDH (ET1601-4, Huabio).

Statistical analysis

Statistical significance was determined using the Student’s t-test (two groups) or One-way ANOVA with Dunnett’s multiple comparisons as a post hoc test (three or more groups) for data with a normal distribution. For data without normal distribution or where there was a violation of the assumption of homogeneity of variance, the Mann-Whitney U test (unpaired samples) or the Wilcoxon signed-rank test (paired samples) was performed; for three or more groups, the Kruskal-Wallis nonparametric ANOVA was used. Differences with P values < 0.05 were considered significant. The statistical methods of this study were reviewed by a biomedical statistician from the School of Public Health, Tongji Medical College, Huazhong University of Science and Technology. Statistical analysis was performed in GraphPad Prism 10.0.2 software.

RESULTS

NLRP3 and IL-1β mRNA were increased in tissues from the ESD resection bed

The study design was shown in Figure 1. Between November 2023 and May 2024, 21 patients who underwent planned esophageal endoscopic resection for superficial esophageal neoplasms agreed to participate in the study. Tissues were acquired as Figure 2A. The clinical and histological characteristics of the patients are presented in Table 1. The mRNA levels of NLRP3 and its downstream target IL-1β and IL-18 were measured. The results demonstrated that the mRNA levels of NLRP3 and IL-1β were higher in the ESD resection bed than in the healthy mucosa, but IL-18 was not elevated in the ESD resection bed (Figure 2B).

Figure 1 Schematic diagram of the study design. NLRP3: NLR family pyrin domain containing 3; IL: Interleukin; RT-qPCR: Reverse transcription and quantitative real-time PCR; IHC: Immunohistochemistry; IF: Immunofluorescence; ROS: Reactive oxygen species; CEL: Celastrol; PRED: Prednisolone; LPS: Lipopolysaccharide; ATP: Adenosine triphosphate; HE: Hematoxylin-eosin.

Figure 2 NLR family pyrin domain containing 3 associated inflammatory mediators in the endoscopic submucosal dissection resection bed. A: Tissue acquisition site in the post-endoscopic submucosal dissection (ESD) esophagus. Arrowhead: The ESD resection bed; Arrow: The normal esophageal mucosa; B: mRNA levels of NLR family pyrin domain containing 3 (NLRP3), interleukin (IL)-1β, and IL-18 (presented as the mean ± SD). Control: Tissues from normal esophageal mucosa; ESD: Tissues from the resection bed. Statistical analysis was performed using the Wilcoxon signed-rank test, ^aP = 0.009; ^bP = 0.02; C: Immunohistochemistry (IHC) of NLRP3 in the post-ESD resection bed: Vascular endothelial cells, myocytes, and fibroblasts (arrows). Scale bar = 50 μm; D: IHC of vimentin in human tissues from post-ESD esophageal stricture. Scale bar = 100 μm; E: Dual labeling for NLRP3 and vimentin was performed in human tissues from the resection bed. Scale bar = 20 μm. ESD: Endoscopic submucosal dissection; IL: Interleukin; NLRP3: NLR family pyrin domain containing 3.

Owing to the mucosal loss status of the ESD resection bed, NLRP3 was detected in fibroblasts, myocytes, and vascular endothelial cells (Figure 2C). Activated fibroblasts play an important role in esophageal strictures[3,5,6]. Consistent with this finding, fibroblast (vimentin-positive spindle-like cell) is the main component in the submucosa of a patient with post-ESD esophageal stricture (Figure 2D). The expression of NLRP3 in fibroblasts was further confirmed by the colocalization of NLRP3 and vimentin in human tissues (Figure 2E).

NLRP3 expression in fibroblasts participated in esophageal stricture formation

To determine the role of NLRP3 in esophageal stricture formation, an esophageal stricture rat model was established as described in Method. The Pearson correlation coefficient revealed significantly increased colocalization of NLRP3 and vimentin in esophageal scar tissue after ulcer healing compared to the normal esophagus in rats (Figure 3A and B). Then, we isolated primary REFs and achieved NLRP3 overexpression by lentivirus transfection. IHC identified the cells as fibroblasts, not epithelial cells[16] (Figure 3C). NLRP3 overexpression significantly increased the mRNA and protein levels of COL1 in primary REFs (Figure 3D), indicating that NLRP3 activation in esophageal fibroblasts promotes fibrosis. Recently, the NLRP3 inflammasome was shown to connect a persistent proinflammatory cytokine milieu with the progression of structural remodeling by activated fibroblasts[17]. These results suggest that NLRP3 could be a potential target for post-ESD esophageal stricture prevention.

Figure 3 NLR family pyrin domain containing 3 expression in rat esophageal fibroblasts in the narrowed esophagus. A: Representative Immunofluorescence image of NLR family pyrin domain containing 3 (NLRP3) and vimentin colocalization in the healthy esophagus (control) and fibrotic esophagus (model) in rats. Scale bar = 50 μm; B: Pearson correlation coefficient of NLRP3 and vimentin in the rat esophagus (presented as mean ± SE). Statistical analysis was performed using the Student’s t test, ^aP = 0.009; C: Immunohistochemistry of vimentin and cytokeratin 19 in primary rat esophageal fibroblasts (REFs). Scale bar = 500 μm; D: Western blot analysis and mRNA levels of collagen 1 in primary REFs transfected with control or NLRP3 overexpression lentiviral particles for 72 hours (presented as mean ± SE). Statistical analysis was performed using the Student’s t test. ^bP < 0.001; ^cP = 0.003; NLRP3: NLR family pyrin domain containing 3; NC: Negative control.

CEL had a better inhibitory effect on NLRP3/IL-1β than PRED in primary REFs

LPS and ATP were used to initiate NLRP3 inflammasome activation. PRED is identified as a positive control as the standard prophylaxis for esophageal strictures. First, the results of the CCK-8 assay showed that 200 nM CEL did not induce any obvious cytotoxicity until 72 horus (Figure 4A). The optimal PRED dose was determined in a preliminary experiment. The two-step challenge (LPS followed by ATP) significantly increased the protein levels of NLRP3, caspase-1 p20, and cleaved IL-1β. Pretreatment with CEL remarkably inhibited the increased protein levels of NLRP3 and cleaved IL-1β. As a positive control, PRED exhibited limited inhibitory effects on NLRP3 and cleaved IL-1β, but its efficacy was enhanced in the combination treatment (Figure 4B). The mRNA levels of NLRP3 and IL-1β were also reduced by CEL and combination treatment (Figure 4C). Thus, CEL has shown an excellent effect on NLRP3 in vitro.

Figure 4 The effect of celastrol on NLR family pyrin domain containing 3 activation in primary rat esophageal fibroblasts. A: CCK8 assay of celastrol (CEL) in primary rat esophageal fibroblasts (REFs) (presented as mean ± SE). Statistical analysis was performed using Dunnett’s multiple comparison test. Compared with control, ^aP = 0.001; ^bP < 0.001; ^cP = 0.003; ^dP < 0.001; ^eP < 0.001; ^fP = 0.005; ^gP < 0.001; ^hP < 0.001; B: Western blot analysis of NLR family pyrin domain containing 3 (NLRP3), proCaspase1, Caspase1, pro-interleukin (IL)-1β, and IL-1β in primary REFs treated as described (presented as mean ± SE). Statistical analysis was performed using Dunnett’s multiple comparison test, with the “lipopolysaccharide (LPS)” group as the control; ⁱP = 0.04; ^jP < 0.001; ^kP = 0.03; ^LP = 0.02; ^mP = 0.02; ⁿP = 0.03; ^oP = 0.04; ^pP = 0.03; C: mRNA levels of NLRP3 and IL-1β in primary REFs treated as described (presented as mean ± SE). Statistical analysis was performed using Dunnett’s multiple comparison test, with the “LPS” group as the control; ^qP = 0.002; ^rP < 0.001; ^sP = 0.03; ^tP < 0.001; ^uP = 0.009; ^vP = 0.009; ^wP = 0.009; NLRP3: NLR family pyrin domain containing 3; CEL: Celastrol; PRED: Prednisolone; LPS: Lipopolysaccharide.

CEL also alleviated esophageal stricture formation and body weight restriction comparable to PRED

Targeting NLRP3, the effect of CEL on esophageal stricture prevention was explored in vivo. Based on the results of our pretreatment experiments, we used an animal experimental procedure similar to that described previously[11] (Figure 5A). Barium esophagography was performed to evaluate the extent of esophageal stricture on day 9. It revealed esophageal stenosis with a bird’s beak sign in the vehicle group compared to the sham group, with significant alleviation in the drug-treatment group (representative image shown in Figure 5B).

Figure 5 The effect of celastrol in preventing esophageal stricture in vivo. A: Rats were divided into the following five groups: Sham group, vehicle group, celastrol (CEL) group, prednisolone (PRED) group, and CEL plus PRED group. Inflammation was assessed on day 6, and fibrosis was evaluated on day 9; B: Representative barium esophagography image of esophageal stricture on day 9 in rats; C: Body weight (g) of rats (n = 5 in each group) (presented as mean ± SE). Statistical analysis was performed using Kruskal-Wallis ANOVA test for Day 1 and Day 3 and one-way ANOVA test for Day 5, Day 7, and Day 9, with the “vehicle” group as the control. Only statistically significant comparisons are marked in corresponding colors. ^aP < 0.001; ^bP = 0.005; ^cP = 0.01; ^dP = 0.01; ^eP = 0.01; ^fP = 0.004; ^gP = 0.02; ^hP = 0.01; ⁱP = 0.005; ^jP = 0.002; ^kP = 0.004; ^LP = 0.001; NLRP3: NLR family pyrin domain containing 3; CEL: Celastrol; PRED: Prednisolone; LPS: Lipopolysaccharide.

The body weight of the rat was recorded until day 9, correlating with the extent of esophageal stricture. The analysis showed that body weight in the vehicle group decreased significantly compared to those of the sham group from day 5 to day 9. CEL, PRED, and CEL + PRED treatment significantly mitigated body weight loss associated with esophageal strictures during this period, suggesting that drug treatment improves food intake restriction in rats with esophageal strictures. On day 9, the average body weight of the CEL group was the largest of the three drug-treatment groups (Figure 5C).

CEL showed greater inhibitory effects on collagen deposition and esophageal thickness than PRED and CEL+PRED

Masson’s trichrome staining showed that in the sham group, the submucosal layer was thin with a soft wall and a linear distribution of collagen fibers. However, in the vehicle group, the esophageal wall was thickened and stiff with numerous collagen fibers in the submucosal and muscularis layers. Stiffness and collagen deposition were improved in the drug-treatment group (CEL, PRED, and CEL + PRED) (Figure 6A).

Figure 6 The effect of celastrol in preventing esophageal stricture in vivo. Data presented as mean ± SE. Statistical analysis was performed using Dunnett’s multiple comparison test, with the “vehicle” group as the control. A: Masson staining image of the esophageal stenosis area. Scale bar = 200 μm; B: Quantitative analysis of Masson staining, ^aP < 0.001; ^bP = 0.01; C: Relative thickness of the scar area, ^cP < 0.001; ^dP = 0.004, ^eP = 0.02; ^fP = 0.02; D: Representative Hematoxylin-eosin (HE) staining image of the esophageal stenosis area in rats on day 9. Scale bar = 80 μm, 20 μm; E: Cell counts in HE staining, ^gP < 0.001; ^hP = 0.001; ⁱP = 0.005; ^jP = 0.002; F: Western blot analysis of Collagen 1 (COL1) in esophageal stenosis tissues on day 9, ^kP = 0.04; ^LP < 0.001; ^mP < 0.001; ⁿP < 0.001; G: mRNA levels of COL1 in esophageal stenosis tissues on day 9, ^oP < 0.001; ^pP < 0.001; ^qP < 0.001; ^rP < 0.001; NLRP3: NLR family pyrin domain containing 3; CEL: Celastrol; PRED: Prednisolone; LPS: Lipopolysaccharide.

Quantitative analysis of collagen deposition showed significant increase in the vehicle group than the sham group, consistent with the worse stenosis on the X-ray. There was marked reduction in the CEL group compared with the vehicle group. Although PRED and CEL + PRED also exhibited inhibitory effects, these were not statistically significant (Figure 6B). CEL also reduced the esophageal wall thickness more significantly than the PRED and CEL + PRED groups (Figure 6C).

HE staining showed that fibroblasts in the drug treatment groups were distributed more sparsely than those in the vehicle group (Figure 6D and E). Besides, the mRNA and protein levels of COL1 were significantly reduced by the drug treatments (Figure 6F and G).

CEL inhibited NLRP3/IL-1β more significantly than PRED and CEL + PRED in vivo

The suppression of inflammation contributed to fibrosis alleviation. To assess the inhibitory effect of CEL on NLRP3 activation in vivo, the protein and mRNA levels of NLRP3 and its downstream effector IL-1β were examined in esophageal ulcer tissues on day 6. As shown in Figure 7A and B, mRNA and protein levels of NLRP3 and IL-1β were significantly higher in the vehicle group than in the sham group, whereas the treatment of CEL significantly reduced the mRNA and protein levels. PRED and CEL + PRED exhibited weaker inhibitory effects than CEL. In addition, CEL and PRED treatment significantly reduced the NLRP3-positive area compared to the vehicle group (Figure 7C and D).

Figure 7 The effect of celastrol on NLR family pyrin domain containing 3 activation in vivo. Data presented as mean ± SE. Statistical analysis was performed using Dunnett’s multiple comparison test, with the “vehicle” group as the control. A: Western blot analysis of NLR family pyrin domain containing 3 (NLRP3) and interleukin (IL)-1β in the esophageal ulceration area of rat tissues on day 6. ^aP < 0.001; ^bP < 0.001; ^cP = 0.004; ^dP = 0.01; ^eP < 0.001; ^fP = 0.01; B: mRNA levels of NLRP3 and IL-1β in rat esophagus on day 6. ^gP = 0.009; ^hP = 0.001; ⁱP = 0.01; ^jP = 0.02; ^kP < 0.001; ^LP < 0.001; ^mP = 0.002; ⁿP = 0.009; C: Representative immunohistochemistry (IHC) image of NLRP3 in the esophageal ulceration area of rats. Scale bar = 80 μm, 20 μm; D: Quantitative analysis of the NLRP3-positive area (%) for IHC, ^oP = 0.004; ^pP = 0.02; ^qP = 0.01; NLRP3: NLR family pyrin domain containing 3; CEL: Celastrol; PRED: Prednisolone; LPS: Lipopolysaccharide.

CEL reduced ROS production and NLRP3 expression by upregulating TXNRD1

Then we investigated the mechanism of CEL on NLRP3 activation. ROS, byproducts of oxidative stress, act as secondary signals that promote NLRP3 activation[18]. ROS fluorescence was performed and fluorescence intensity was remarkably lower in the CEL-treated group than in the LPS + ATP-induced group (Figure 8A and B). These findings were further supported by flow cytometry analysis (Figure 8C).

Figure 8 The effect of celastrol on the thioredoxin reductase 1/reactive oxygen species axis. A: Representative fluorescence image and of intracellular reactive oxygen species (ROS) production detected by DCFH-DA. Scale bar = 500 μm; B: Quantitative analysis of ROS fluorescence intensity (presented as mean ± SE). Statistical analysis was performed using Dunnett’s multiple comparison test, with the “lipopolysaccharide (LPS)” group as the control, ^aP < 0.001; ^bP < 0.001; C: Flow cytometry of ROS levels in cells. The mean value of the BL1-A channel represents the ROS production level (presented as mean ± SE). Statistical analysis was performed using Dunnett’s multiple comparison test, with the “LPS” group as the control, ^cP < 0.001; ^dP < 0.001; D: Western blot analysis of nuclear factor erythroid 2-related 2 (Nrf2) in the nucleus and cytoplasm indicates nuclear translocation (presented as mean ± SE). Statistical analysis was performed using Dunnett’s multiple comparison test, with the “LPS” group as the control, ^eP = 0.005; ^fP = 0.03; ^fP = 0.03; E and F: mRNA, protein expression levels, and activity of thioredoxin reductase 1 (TXNRD1) (presented as mean ± SE). Statistical analysis was performed using Dunnett’s multiple comparison test, with the “LPS” group as the control, ^hP = 0.01; ⁱP = 0.03; ^jP = 0.02; ^kP < 0.001; G: Flow cytometry of ROS levels in cells transfected with shNC or TXNRD1 shRNA. The mean value of the BL2-A channel represents the ROS production level (presented as mean ± SE). Statistical analysis was performed using the Student’s t test, ^LP = 0.002 vs shNC-LPS; ^mP < 0.001 vs shTXNRD1-LPS; H: Western blot analysis of NLR family pyrin domain containing 3 protein expression level in cells with shNC or shRNA of TXNRD1 (presented as mean ± SE). Statistical analysis was performed using Dunnett’s multiple comparison test, with the “shTXNRD1-LPS” or “shNC-LPS” as the control. ⁿP = 0.04 vs shTXNRD1-LPS; ^oP = 0.007 vs shNC-LPS; ^pP = 0.04 vs shNC-LPS. NLRP3: NLR family pyrin domain containing 3; CEL: Celastrol; PRED: Prednisolone; LPS: Lipopolysaccharide; TXNRD1: Thioredoxin reductase 1.

Subsequently, we examined whether CEL counteracted ROS generation by activating the Nrf2 antioxidant pathway in primary REFs. We found that LPS plus ATP induced a two-fold increase in Nrf2 nuclear translocation, whereas CEL pretreatment (200 nM for 1 hour) did not further enhance Nrf2 nuclear translocation (Figure 8D).

Considering the top 20 differentially expressed genes in CEL reported previously[19] and negative feedback effect, we investigated the mRNA and protein levels of one of the downstream targets of Nrf2, TXNRD1. The results showed that CEL significantly increased TXNRD1 mRNA and protein expression compared with those of the LPS plus ATP group. Besides, the TXNRD1 activity was higher in the CEL group than in others (Figure 8E and F).

To determine whether TXNRD1 plays a key role in CEL-mediated ROS alleviation and NLRP3 inhibition, we performed rescue experiments using TXNRD1 shRNA. TXNRD1 inhibition abrogated the CEL-induced reduction in ROS levels (Figure 8G) and NLRP3 protein expression (Figure 8H).

TXNRD1 overexpression promoted NLRP3 ubiquitination in primary REFs

In addition to regulating ROS production, TXNRD1 altered protein stability by catalyzing the modification of protein substrates[20]. Thus, the interaction between TXNRD1 and NLRP3 in primary REFs was investigated.

TXNRD1 overexpression was successfully induced in primary REFs using lentiviral particles (Figure 9A). We first examined NLRP3 mRNA and protein levels in control and TXNRD1-overexpressing cells. The results showed that while TXNRD1 overexpression did not affect NLRP3 mRNA levels, it significantly reduced NLRP3 protein levels (Figure 9B). Following cycloheximide treatment for 12 hours and 24 hours, NLRP3 protein levels in TXNRD1-overexpressing cells were significantly lower than those in the control group (Figure 9C), indicating that TXNRD1 overexpression promoted NLRP3 degradation.

Figure 9 Thioredoxin reductase 1 and NLR family pyrin domain containing 3 interaction in primary rat esophageal fibroblasts. Data presented as mean ± SE). Statistical analysis was performed using the Student’s t test. A: Western blot analysis of thioredoxin reductase 1 (TXNRD1) and NLR family pyrin domain containing 3 (NLRP3) in the control and TXNRD1-overexpression cells, ^aP < 0.001; ^bP = 0.001; B: mRNA and protein levels of NLRP3 in the control and TXNRD1-overexpression groups, ^cP = 0.006; C: Western blot analysis of NLRP3 in the control and TXNRD1-overexpression groups after incubation with CHX (10 μM) for 0, 2, 4, 8, 12 and 24 hours, ^dP = 0.02; ^eP = 0.02; D: Protein levels of NLRP3 ubiquitination in primary rat esophageal fibroblasts (REFs) from the control and TXNRD1-overexpression groups following MG132 (5 μM) treatment for 6 hours (n = 3 in each group); E: Reciprocal co-immunoprecipitation assays showing the TXNRD1-NLRP3 interaction with endogenous proteins in primary REFs (n = 3 in each group); F: Colocalization of TXNRD1 and NLRP3 observed under a fluorescence microscope. Scale bar = 20 μm. NLRP3: NLR family pyrin domain containing 3; CEL: Celastrol; PRED: Prednisolone; LPS: Lipopolysaccharide; TXNRD1: Thioredoxin reductase 1; NC: Negative control.

Ubiquitination facilitates NLRP3 degradation and inhibits NLRP3 activation[21]. We further explored whether CEL-induced TXNRD1 upregulation could suppress NLRP3 levels via the ubiquitin-proteasome system. We then investigated NLRP3 ubiquitination following MG132 treatment and found that TXNRD1 overexpression accelerated NLRP3 ubiquitination (Figure 9D). To identify the interaction between TXNRD1 and NLRP3, co-immunoprecipitation was performed and the results demonstrated that TXNRD1 co-immunoprecipitated with NLRP3 (Figure 9E). IF staining confirmed that TXNRD1 was colocalized with NLRP3 (Figure 9F). These results indicate an interaction between TXNRD1 and NLRP3.

Consequently, TXNRD1 upregulation not only inhibits ROS production but also enhances NLRP3 ubiquitination, thereby suppressing NLRP3 activation.

DISCUSSION

Esophageal strictures associated with submucosal fibrosis often develop during the healing of post-ESD ulcers, particularly when the resection area extends to approximately 75% of the circumference[2]. Owing to its dual anti-inflammatory effects, oral or local steroid injection has been recommended to prevent esophageal stenosis following ESD. However, severe complications, such as perforation, immunosuppression, delayed ulcer healing, and osteoporosis, often occur[22], and the incidence of stenosis remains between 10% and 45% despite prophylactic measures[2]. Therefore, alternative treatments should be explored.

Inflammation at mucosal defects should be minimized to prevent excessive fibrosis and stricture formation[8]. To clarify the inflammatory response in the esophagus at an early stage after endoscopic treatment, we examined the activation of inflammatory pathways in the resection bed immediately after ESD for superficial esophageal neoplasms. Increased NLRP3 and IL-1β mRNA levels were detected in the submucosal tissues. In contrast, IL-18 mRNA level was not significantly different compared to normal mucosa. Differences in the expression trends of IL-18 and IL-1β have been reported previously[23,24], but the mechanisms have not been elucidated.

IL-18 has been reported to present dual functions in chronic gut inflammation[25,26]. In a mouse model of acute colitis, IL-18 had an anti-inflammatory role, promoting epithelial barrier repair and maintaining intestinal homeostasis[27]. Butyrate-producing gut bacteria may aid in mucosal healing in inflammatory bowel disease by increasing IL-18 expression[28]. A previous study reported that IL-18 signaling activity is lower in patients with Barrett’s esophagus and esophageal adenocarcinoma[29]. However, the role of IL-18 in esophageal diseases has not been clarified due to the scarcity of relevant studies. Given the protective effect of IL-18 against epithelial damage, we hypothesized that IL-18 may affect more than the single synergistic pro-inflammatory effect with IL-1β in esophageal tissues after ESD. Further exploration is warranted on this topic.

Upon exposure to digestive juices and thermal injury, we observed positive NLRP3 expression in multiple cell types in the submucosal space. Fibroblasts are central to scarring and fibrosis in post-ESD esophageal strictures[3,5,7]. However, their potential role in sensing tissue damage by pattern recognition via the inflammasome in the ESD resection bed remains unexplored. NLRP3 activation in fibroblasts has been reported in the pathology of several fibrotic diseases[30,31]. In the context of esophageal fibrosis, our study demonstrated that NLRP3 overexpression in primary REFs increased collagen deposition in vitro and that NLRP3 colocalization with fibroblasts was elevated in the rat model of esophageal stricture. These findings establish a link between NLRP3 activation and esophageal stricture. Thus, targeting NLRP3 activation in esophageal fibroblasts may offer a potential strategy for preventing stricture formation.

CEL alleviates fibrosis through NLRP3 inhibition in several fibrotic diseases[32,33]. Based on this, we evaluated the suppressive effects of CEL on NLRP3 activation both in vitro and in vivo. Our results showed that CEL significantly inhibited NLRP3 activation and IL-1β secretion in primary REFs. Moreover, CEL significantly alleviated NLRP3 activation following collagen deposition and esophageal stenosis in the rat model. The inhibitory effect of CEL was comparable to that of PRED and, in some aspects, superior. To our knowledge, we are the first to demonstrate the potential of CEL in preventing post-ESD esophageal strictures.

However, despite its great efficacy, the clinical application of CEL is limited due to its poor oral utilization and water solubility[33]. It also has a rather narrow dosage window, with severe toxicity reported in a rat arthritis model at doses higher than 2.5-5 mg/kg/day[34]. In our study pre-test, we found that CEL had significant toxicity when administered intraperitoneally at 1 mg/kg/day, whereas gastric gavage markedly attenuated this toxicity in rats, similar to a previous report[35]. As the duration of CEL treatment was only 10 days in this study, it is unclear whether a longer duration would induce significant cumulative toxicity in vivo. Therefore, although CEL has great potential as an alternative treatment for patients who cannot tolerate steroid treatment, the preventive effect of CEL must be validated in large animal models (e.g. porcine or canine models of esophageal ESD). Whether CEL toxicity can be ameliorated by topical application or combination therapy requires further investigation.

Inhibition of ROS production partially inhibits the ROS-dependent activation of NLRP3[36]. Among the numerous ROS-associated signaling pathways, the Nrf2 signaling pathway has garnered most attention[37]. We found that LPS induced an increase in Nrf2 nuclear translocation, but CEL pretreatment did not enhance this effect. Interestingly, TXNRD1 expression and activity were increased by CEL.

TXNRD1 is an endogenous antioxidant enzyme that regulates redox homeostasis and protects DNA from oxidative stress-associated damage[38]. TXNRD1 knockdown using shRNA blocked CEL-induced ROS reduction and NLRP3 inhibition, confirming that TXNRD1 is essential for CEL-mediated ROS suppression and NLRP3 inhibition. ROS could switch on and off diverse signaling pathways including the Nrf2/Keap1, PI3K/AKT, and MAPK pathways[39]. Thus, the observed reduction in Nrf2 nuclear translocation under CEL treatment may be a secondary effect resulting from ROS suppression.

Independent of its enzymatic activity, we identified a novel regulatory function of TXNRD1 in NLRP3 ubiquitination and, for the first time, demonstrated ubiquitin-proteasome degradation of NLRP3 in primary REFs. We hypothesized that TXNRD1 alters the protein stability of NLRP3 by their interaction, but the specific mechanism requires further exploration.

CONCLUSION

This study demonstrates that CEL has significant potential for preventing post-ESD esophageal strictures by reducing collagen deposition and esophageal fibrosis in rats. It may serve as an alternative treatment for patients in whom steroid prophylaxis is ineffective. In addition, our study provides a more nuanced understanding of the interplay between TXNRD1 and ubiquitination.

ACKNOWLEDGEMENTS

We thank the Experimental Medicine Center, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan 430030, People’s Republic of China for their support.