Estrogen Receptor β Activation Mitigates Colitis-associated Intestinal Fibrosis via Inhibition of TGF-β/Smad and TLR4/MyD88/NF-κB Signaling Pathways

Abstract

Background

Intestinal fibrosis, a complex complication of colitis, is characterized by excessive extracellular matrix (ECM) deposition. Estrogen receptor (ER) β may play a role in regulating this process.

Methods

Intestinal tissue samples from stenotic and nonstenotic regions were collected from Crohn’s disease (CD) patients. RNA sequencing was conducted on a mouse model to identify differentially expressed mRNAs. Histological, immunohistochemical, and semiquantitative Western blotting analyses were employed to assess ECM deposition and fibrosis. The roles of relevant pathways in fibroblast transdifferentiation, activity, and migration were examined.

Results

Estrogen receptor β expression was found to be downregulated in the stenotic intestinal tissue of CD patients. Histological fibrosis score, collagen deposition, and profibrotic molecules in the colon of an intestinal fibrosis mouse model were significantly decreased after activation of ERβ. In vitro, ERβ activation alleviated transforming growth factor (TGF)-β-induced fibroblast activation and migration, as evidenced by the inhibition of col1α1, fibronectin, α-smooth muscle actin (α-SMA), collagen I, and N-cadherin expression. RNA sequencing showed that ERβ activation affected the expression of genes involved in ECM homeostasis and tissue remodeling. Enrichment analysis of differentially expressed genes highlighted that the downregulated genes were enriched in ECM-receptor interaction, TGF-β signaling, and Toll-like receptor (TLR) signaling. Western blotting confirmed the involvement of TGF-β/Smad and TLR4/MyD88/NF-κB signaling pathways in modulating fibrosis both in vivo and in vitro. The promoter activity of TGF-β1 and TLR4 could be suppressed by ERβ transcription factor.

Conclusion

Estrogen receptor β may regulate intestinal fibrosis through modulation of the TGF-β/Smad and TLR4/MyD88/NF-κB signaling pathways. Targeting ERβ activation could be a promising therapeutic strategy for treating intestinal fibrosis.

Graphical Abstract

Summary of the underlying mechanisms by which estrogen receptor (ER) β inhibits intestinal fibrosis through TGF-β/Smad and TLR4/Myd88/NF-κB pathway.

Open in new tabDownload slide

Introduction

Inflammatory bowel disease (IBD) poses a significant global health challenge, with its recurrent nature and escalating severity causing substantial distress for affected individuals and society.¹ A prevalent complication of Crohn’s disease (CD), intestinal fibrosis, often results in intestinal narrowing and blockage.² Research suggests that at least half of those with CD are likely to experience an intestinal perforation or stricture.³ Despite advancements in anti-inflammatory treatments, surgery remains the primary approach for managing fibrostenotic CD due to the absence of effective anti-fibrotic therapies. This highlights the urgent need for innovative strategies to prevent or lessen fibrogenesis in CD.

The estrogen receptor (ER)-β, a conventional ER subtype, is a member of the nuclear receptor family known for its moderating role in osteoporosis, obesity, and gastrointestinal diseases.^4,5 Predominantly present in the colon, ERβ is integral for anti-inflammation and colorectal cancer prevention.^6,7 Given its potential as a crucial mediator against organ fibrosis, including in the skin, prostate, heart, and liver, ERβ has attracted increasing research interest.^8–11 Animal studies have shown that treating with an ERβ isoform alleviated interstitial heart fibrosis, reducing maladaptive remodeling.⁹ However, ERβ’s impact on intestinal fibrosis remains underexplored. Therefore, further investigation into the interplay between CD-related fibrosis and ERβ is paramount.

Fibrosis is generally the result of prolonged exposure to chronic irritation, leading to significant extracellular matrix (ECM) deposition.^2,12 In this process, fibroblasts in the intestine are drawn to the site of irritation, activated, and transformed into ECM-producing myofibroblasts, which are central to fibrosis development.¹³ Soluble molecules such as N-cadherin¹⁴ and transforming growth factor (TGF)-β^15,16 are thought to drive this transformation. Recent studies suggest that myofibroblasts may also react to pathogen-associated molecular patterns recognized by Toll-like receptors (TLRs), further promoting collagen synthesis.¹⁷ These findings present potential targets for therapies aimed at reducing ECM deposition and CD-associated fibrosis. Deciphering the role and mechanism of ERβ in intestinal fibrosis could greatly contribute to the development of preventative and therapeutic strategies for intestinal fibrosis.

In this study, we used intestinal tissue samples from CD patients to ascertain the ERβ expression pattern in stenotic and nonstenotic regions. We employed a dextran sulfate sodium (DSS)-induced mouse model of colitis-associated intestinal fibrosis to investigate the potential antifibrotic role of ERβ in vivo, focusing on its effect on ECM deposition. Additionally, we evaluated the influence of ERβ agonists on the migration and activation of human colonic fibroblasts in vitro and sought to understand the underlying molecular mechanisms using RNA sequencing data from our in vivo models.

Materials and Methods

Human Tissue Acquisition and Processing

Intestinal specimens were collected from Crohn’s disease (CD) patients (n = 8) undergoing surgical intervention. Participant inclusion in the study necessitated evidence of intestinal stenosis, verified either through endoscopic methods (colonoscopy or enteroscopy) or relevant imaging techniques. Conditions that could mimic stenosis were excluded. These conditions included intestinal tuberculosis, intestinal malignancy, ischemic bowel disease, infectious bowel disease, amyloidosis, and lymphoma. Both stenotic and nonstenotic regions of each tissue sample were investigated. Detailed clinical data of the patients are presented in Supplementary Table 1. After collection, the specimens were dissected into fragments. A fragment was preserved in paraformaldehyde, followed by paraffin embedding for subsequent histological and immunofluorescence staining. The remaining specimens were snap-frozen in liquid nitrogen, stored at −80°C, and later utilized for Western blot analyses. All patients furnished informed consent for participation. This study received approval from the Independent Ethics Committee of Wuhan Union Hospital (No. 2018-S425).

Mouse Model and Experimental Design

We used C57BL/6 male mice, aged 8 weeks and weighing 23-25 g; we acquired them from HFK Bioscience (Beijing, China) and kept them under specific pathogen-free conditions. They were randomly assigned into 3 groups (10 per group): control, DSS, and DSS + ERB041. The chronic colonic fibrosis model was established as previously described by our department.¹⁸ In brief, mice were provided with 3% DSS (MW, 36 000–50 000 Da; MP Biomedicals, Santa Ana, CA, USA) in their drinking water over a course of 4 cycles. Each cycle was structured to include 1 week of DSS administration, immediately followed by 1 week of regular drinking water. On day 14, ERB041 (Tocris Biotech, Minneapolis, MN, USA), a selective ERβ agonist, was administered subcutaneously at a concentration of 1 mM in dimethyl sulfoxide (5 mg/kg/d) to the DSS + ERB041 group. The mice in the control and DSS groups received concurrent injections of the vehicle alone. All subjects were killed on the fourth day after the final DSS exposure, and their colonic tissues were collected for analysis. The whole structure of the colon was carefully extracted to record its length and weight. A small segment of the colon (1 cm proximal from the distal rectum) was preserved in 10% formalin for a 24-hour period, then dehydrated and embedded in paraffin. All animal procedures complied with institutional animal ethics guidelines and received approval from the Institutional Animal Care and Use Committee of Tongji Medical College, Huazhong University of Science and Technology (protocol No. 2021-2758), Wuhan, China.

Histological Assessment

Tissue samples were sectioned at 4-μm intervals for subsequent pathological evaluation. The depth of invasion, degree of inflammation, destruction of crypts and epithelia, and extent of intestinal lesions were assessed by microscopically observing hematoxylin and eosin (HE)-stained sections.¹⁹ Collagen deposition, structural morphology, and mucosal/submucosal boundary definition were histologically evaluated using Masson trichrome staining.²⁰ High-resolution digital images of the tissue specimens were captured using a Pannoramic MIDI microscope (3DHistech, Budapest, Hungary; numerical aperture, 0.8). The colonic tissue at the distal rectum was harvested for further western blotting and quantitative real-time polymerase chain reaction (qRT-PCR) assays.

RNA Sequencing

For the purpose of RNA sequencing, colon samples located 1 cm from the distal rectum were randomly selected from the control (n = 3), DSS (n = 3), and DSS + ERB041 (n = 3) groups. These samples were processed with TRIzol reagent (Invitrogen, Waltham, MA, USA), per the guidelines provided by the manufacturer, to extract total RNA. The quality of the extracted RNA was meticulously evaluated utilizing a Nanodrop 2000 instrument (Thermo Scientific, Waltham, MA, USA), while an Agilent 2100 device (Agilent Technologies, Santa Clara, CA, USA) was employed to determine RNA concentrations.

The RNA sequencing was carried out seamlessly on the Illumina NovaSeq 6000 platform (Biomarker Technologies, Beijing, China). Initial raw reads were exported and subjected to filtration via SOAPnuke software (V2.1.0) to eliminate adapters, low-quality, and undetected tags, thus generating clean reads. The expression levels of genes and transcripts were quantified using the RSME software, assisted by Bowtie2, and presented as fragments per kilobase of transcript per million mapped reads. Differentially expressed genes (DEGs) among the groups were analyzed using DESeq2. Genes showing enrichment, characterized by a false discovery rate <0.05 and log2FC >1 or < −1, were mapped to terms in the Gene Ontology (GO) database for GO analysis and to those in the Kyoto Encyclopedia of Genes and Genomes (KEGG) database for KEGG pathway analysis. The top 20 functional pathway clusters that included the DEGs were annotated with bubble diagrams. The distribution differences in the DEGs were visualized by principal component analysis and the R package. The raw sequencing data have been responsibly submitted to the Sequence Read Archive under the BioProject accession number PRJNA858675.

Immunofluorescence

Human intestinal specimens were prepared for immunofluorescence staining by deparaffinizing the slices and repairing the antigens with a Tris-ethylenediamine tetraacetic acid (EDTA) solution. Concurrently, the cells were stabilized by fixing them in 4% paraformaldehyde for a 30-minute duration. In the subsequent step, the slides were overlaid with 0.3% Triton X-100 and donkey serum, each for 30 minutes at room temperature. The samples were then exposed to primary antibodies, specifically rabbit anti-α-smooth muscle actin (anti-α-SMA; Proteintech, Rosemont, IL, USA, 1:200), rabbit anti-vimentin (ABclonal, Wuhan, China, 1:200), and mouse anti-ERβ antibodies (Genetex, Irvine, CA, USA, 1:200). Following the application of primary antibodies, the samples were kept in incubation at 4°C overnight. The antibodies were then linked with Alexa Flour 594 secondary antibody (1:200) for a 60-minute period. For counterstaining, the slices were treated with DAPI (Antengene, Shanghai, China). The stained specimens were examined under a fluorescence microscope (BX53, Olympus; numerical aperture, 0.4–0.95), and the digital images were captured for further analysis.

Immunohistochemistry

The entire procedure was conducted in strict compliance with the instructions provided in the reagent kit (SA1028, Boster Bio, Pleasanton, CA, USA). Initially, the colonic slices underwent a deparaffinization process, followed by immersion in 3% hydrogen peroxide. Antigens were retrieved using an EDTA buffer solution, and the slices were subsequently blocked using 5% bovine serum albumin. The blocked slices were then incubated overnight with primary antibodies, specifically α-SMA (Proteintech, 1:200), fibronectin (Proteintech, 1:200), and collagen I (Cell Signaling Technology, Danvers, MA, USA, 1:200), at a temperature of 4°C. This was followed by a 30-minute incubation period with a secondary anti-rabbit IgG antibody and a streptavidin-biotin complex. The final phase of the procedure involved staining with a diaminobenzidine chromogenic agent and counterstaining with hematoxylin. Digital images of the treated slices were captured using a photomicroscope (BX53, Olympus; numerical aperture, 0.75). Quantitative analysis of the images was performed using Image-Pro Plus (V6.0) software.

Western Blotting

Total protein and nucleoprotein were extracted from the intestinal tissues and cells using a standard extraction buffer, and the concentrations were measured using the bicinchoninic acid assay kit (Aspen, Wuhan, China). Equivalent quantities of proteins underwent separation through 10% to 12.5% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (Epizyme, Shanghai, China) and were subsequently electroblotted onto 0.45-μm polyvinylidene difluoride membranes (Millipore, Billerica, MA, USA). The membranes were then prepared for antibody binding by blocking with 5% nonfat milk in Tris-buffered saline with Tween20 (TBST) for an hour. This was followed by an overnight incubation with primary antibodies at a 4°C temperature. The primary antibodies employed in this study included those against ERβ (Genetex), collagen I (Cell Signaling Technology), α-SMA (Cell Signaling Technology), E-cadherin (Proteintech), N-cadherin (Proteintech), vimentin (ABclonal), TGF-β (Cell Signaling Technology), Smad2 (Cell Signaling Technology), p-Smad2 (Cell Signaling Technology), Smad2/3 (Cell Signaling Technology), p-Smad2/3 (Cell Signaling Technology), TGF-β-RII (Cell Signaling Technology), TLR4 (Arigo, Shanghai, China), MyD88 (Proteintech), NF-κB p65 (Proteintech), and phosphorylated NF-κB p65 (NF-κB p-p65; Abcam, Eugene, OR, USA). After incubation with primary antibodies, the membranes were washed with TBST 3 times and subsequently incubated with secondary antibodies (Antengene, 1:2000). Bands were brought into visual focus with the aid of electrochemical luminescence reagents (Vazyme, Nanjing, China). Images of the blots were captured using a chemiluminescence imaging system (CLiNX, Shanghai, China).

Quantitative Real-Time PCR

Total RNA extraction was carried out using TRIzol reagent (Invitrogen), which was then reverse transcribed into cDNA via the PrimeScript RT reagent kit (Takara, Shiga, Japan). The qRT-PCR assays were executed with SYBR Green PCR Master Mix (Takara) and run on a Light Cycler 480 System instrument (Roche, Basel, Switzerland) in accordance with the instructions provided by the manufacturers. All data were normalized against glyceraldehyde-3-phosphate dehydrogenase (GAPDH), with relative transcript expression levels calculated through the 2^-△△Ct method. The primer sequences utilized for qRT-PCR are provided in Supplementary Table 2.

In Vitro cell Culture

The CCD-18Co cell line was propagated in Dulbecco’s modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin, in a humidified 5% CO₂ atmosphere at 37 °C. To investigate the interactions between ERβ activation and fibroblasts, cells were initially treated with 10 ng/mL of TGF-β (Peprotech, Cranbury, NJ, USA) for 2 hours, followed by treatment with or without 100 nM of ERB041 for 46 hours. The fibroblasts underwent transfection with TLR4-overexpression plasmids or negative control plasmids (ov-NC) using LipofectamineTM 3000 (Invitrogen) as per the manufacturer’s guidelines. After transfection, the cells were once again treated with TGF-β and ERB041 before collection. All experiments were replicated twice independently.

CCK-8 Assay for Cell Viability

Cell viability was evaluated using the Cell Counting Kit-8 (CCK-8) assay (Beyotime, Haimen, Jiangsu, China). Briefly, cells were harvested and seeded into 96-well plates to attach overnight. The medium was then replaced with DMEM (FBS-free) supplemented with varying concentrations of ERB041 or TGF-β. After 48 hours of exposure to different treatments, 10 μL of the working reagent was added, and the cells were incubated for an additional hour at 37°C. The optical density of each well was then measured to determine cell viability.

Transwell Migration Assay

The migratory behavior of the cells was assessed using an 8 μm pore-sized Transwell chamber (Corning, Glendale, AZ, USA). Cells were seeded into the upper chamber, while the lower chamber was loaded with 600 μL of medium enriched with 10% FBS and varying compounds. Following a 48-hour incubation, the upper chamber was detached and the cells that had migrated and adhered to the lower membrane surface were rinsed with PBS. Subsequently, the cells were fixed using 4% paraformaldehyde, stained with crystal violet for 30 minutes, and then examined and quantified under an IX73 microscope (Olympus, Tokyo, Japan; numerical aperture, 0.45).

Dual Luciferase Reporter Assay

The promoter sequences for TGF-β1 and TLR4, termed pGL4.10-TGFB1-promotor and pGL4.10-TLR4-promotor, were respectively cloned into a pGL4.10-basic luciferase reporter plasmid (Obio Technology, Shanghai, China). An ERβ transcription factor overexpression plasmid (1593 bp) and a negative control (NC) plasmid were constructed using the pcDNA3.1 vector. For the luciferase assay, 293T cells were cotransfected with ERβ-carrying plasmids and the wild-type TGF-β1 and TLR4 promoter reporter plasmids, using the LipofectamineTM 2000 reagent (Invitrogen). The dual luciferase reporter gene assay system (Promega, Madison, WI, USA) was employed to measure the activity of firefly and Renilla luciferases in 96-well plates.

Statistical Analysis

Each experimental group comprised more than 3 animals, with the term “n” representing the count of independent values. Data comparisons between 2 groups utilized unpaired Student’s t tests, while 1-way analysis of variance was applied for multiple group comparisons. Data are presented as mean values ± standard deviation and analyzed using the GraphPad Prism (V7.0) software package (GraphPad Software, San Diego, CA, USA). The threshold for statistical significance was set at P < .05 (2-tailed).

Results

ERβ Expression Is Reduced in Stenotic Intestinal Tissues of CD Patients

Our investigation encompassed both stenotic and nonstenotic intestinal specimens derived from CD patients. In the stenotic regions, we observed notable histological alterations, including shortened villi, absence or disruption of crypt structures, widened spacing between some crypts, and a conspicuous infiltration of inflammatory cells, predominantly lymphocytes and plasma cells, around the crypts and at the mucosal base. Additionally, localized hyperplasia of the mucosal muscularis layer was a consistent finding (Figure 1A). In contrast, the nonstenotic regions presented a more regular histological architecture. The crypts were uniformly arranged with their length approximately one-third the height of the villi. A mild degree of inflammatory cell infiltration was noted, but without the extent or pattern observed in stenotic areas. Notably, our detailed microscopic examination did not reveal any features suggestive of heterogeneous proliferative cells or nuclear schizophrenia, which are often indicative of neoplastic transformations. Masson’s trichrome staining indicated an expanded blue-stained submucosal layer within the stenotic region, an indication of increased collagen deposition, a hallmark of fibrosis (P = .002; Figure 1A). Immunofluorescence staining displayed considerably reduced ERβ expression in the stenotic tissues, contrasting with the abundant ERβ distribution within the mucosal layers of adjacent nonstenotic intestinal tissues (Figure 1B). The variations in ERβ expression were further substantiated at the protein level through western blot analysis, which confirmed the lower ERβ expression in the stenotic tissues (P = .0005). In addition, this analysis demonstrated elevated levels of the fibrotic markers, collagen I (P = .0024), and α-SMA (P < .0001), within the stenotic intestines (Figure 1C). Taken together, our findings paint a compelling picture of diminished ERβ expression concurrent with ECM deposition in the stenotic intestines of CD patients.

In addressing concerns about gender representation, we executed a subanalysis focused exclusively on the male sample. This procedure aimed to determine whether the inclusion of 1 female participant could significantly skew the observed data trends. Our analysis confirmed that the trends in the male-only sample were consistent with those in the full sample, indicating that the inclusion of a single female subject did not materially impact the overall results. Specifically, we noted a significantly higher percentage of blue-stained area in the stenotic tissues of male subjects (P = .0067) compared with their nonstenotic counterparts (data not shown). This finding mirrors the general trend observed across the entire study cohort. Additionally, in the male stenosis group, there was a significant reduction in ERβ expression (P = .0032) and a notable increase in α-SMA (P < .0001) and collagen I (P = .0018) expressions (data not shown). These patterns of expression in the male-only group were in alignment with those observed in the full sample, further validating the reliability of our data and indicating that gender distribution did not substantially affect the study’s key findings.

ERB041 Treatment Ameliorates DSS-induced Chronic Colitis and Fibrosis in Mice

We examined the role of the ERβ agonist ERB041 in DSS-induced chronic colitis in mice. The experimental protocol is depicted in Figure 2A. Our findings revealed that ERB041 treatment contributed to an increase in body weight over the course of the experiment (Figure 2B). On the last day of modeling, the mice were humanely killed for further evaluation. Morphological observation revealed significant shortening and increased weight of colon in the DSS-treated group compared with the control group (Figure 2C and D). Notably, ERB041 intervention led to recuperation towards normal colon morphology and decreased the colon weight/length ratio, suggestive of fibrogenesis attenuation (Figure 2E). Pathological examination of the DSS and DSS + ERB041 groups disclosed architectural anomalies, loss of crypt structure, and reduced goblet cell numbers. The DSS group exhibited a substantial influx of inflammatory cells in the mucosa, lamina propria, and submucosa (Figure 2F). Daily ERB041 administration ameliorated the histological indications of colonic fibrosis in DSS-treated mice, as demonstrated by the reduced or faded blue-stained areas in the mucosa and submucosa (Figure 2F and G). The DSS group scored higher on the fibrosis index compared with the control, while ERB041 intervention significantly lowered this score (Figure 2H).

Gene sequencing was performed to identify the functional mRNAs involved in the development of fibrosis. Colon samples from the control, DSS, and DSS + ERB041 groups were randomly selected for sequencing. A total of 31 283 mRNAs were identified. Principal component analysis (PCA) indicated discernible differences in patterns across the treatment groups (Figure 2I). Heatmap analysis disclosed differences in the DEG profiles among the groups (Figure 2J). M-versus-A (MA) plots (Figure 2K and L) were used to represent the genes expression across the various groups. A total of 701 genes were upregulated, and 827 genes downregulated in the DSS group compared with the control group (Figure 2K); the DSS + ERB041 group demonstrated 340 upregulated and 199 downregulated genes relative to the DSS group (Figure 2L). Supplementary Table 3 presents a comprehensive list of the differentially expressed genes as indicated in the MA plots.

ERB041 Modulates Fibrosis-related Genes and Protein Expression Patterns and Reduces Intestinal Myofibroblast Count

A selection of fibrogenesis-related genes were showcased in the heatmap, providing a clear depiction of how DSS enhances fibrosis in the animal model and the mitigating role ERB041 can play. The heatmap analysis identified genes associated with ECM deposition and the regulatory balance between Matrix Metalloproteinase (MMP)s and Matrix Metalloproteinase Inhibitor 1 (TIMPs; Figure 3A). To verify the RNA sequencing analysis, we carried out qRT-PCR on frozen colonic specimens. The genes Col1α1, Timp1, Tgf-β, and Il-6 were found to be significantly upregulated in the DSS group relative to the control group. Of note, Tgf-β, a key activator of fibroblasts, experienced a striking 15-fold increase in expression in the DSS group compared with the control. The ERB041 administration notably downregulated Tgf-β mRNA expression levels in the DSS + ERB041 group (Figure 3B), suggesting ERB041’s potential to inhibit the release of profibrotic molecules in the colon.

Compared with the control group, the DSS group exhibited a pronounced decline in ERβ expression, which ERB041 administration could stimulate (Figure 3C). Immunohistochemistry and western blotting disclosed a significant increase in the myofibroblast activation marker α-SMA and ECM components, collagen I, and fibronectin in the DSS group compared with the control group (Figure 3D and E). These components were predominantly distributed in the mucosa and submucosa. Remarkably, ERB041 treatment demonstrated potential in slowing intestinal fibrosis progression. Immunohistochemistry revealed a significant reduction in α-SMA immunopositivity in the mucosa and submucosa of ERB041-treated mice compared with mice in the DSS group, indicating a reduction in the quantity of intestinal myofibroblasts (Figure 3D). Activation of ERβ also appeared to reverse the expression patterns of vimentin and E-cadherin observed in the DSS group (Figure 3E). Collectively, these findings suggest a crucial role for ERβ in matrix remodeling in the context of DSS-induced intestinal fibrosis.

ERβ Agonist Inhibits Fibroblast Migration and Activation In Vitro

Given the observed reduction in α-SMA immunopositive myofibroblast infiltration in the colonic mucosa and submucosa of mice treated with the ERβ agonist, we explored the impact of ERB041 on fibroblast activation in vitro. The cytoactive effects of increasing concentrations of TGF-β and ERB041 were evaluated by CCK8 assays. Similar to the TGF-β-treated cells, those exposed to ERB041 at concentrations of 50 nM, 100 nM, and 200 nM for 48 hours maintained growth comparable to untreated cells (Figure 4A). Transwell assays were employed to examine cell migration in response to TGF-β with or without ERB041. We observed a decrease in the number of migrating fibroblasts under the influence of ERB041 (Figure 4B). Taken together with the CCK8 results, these findings suggest that ERB041 inhibits fibroblast migration through a mechanism not dependent on proliferation inhibition.

We further examined whether TGF-β or ERB041 stimulation could influence the fibrotic phenotype of fibroblasts. Cells were treated with 100 nM of ERB041 for 46 hours following a 2-hour stimulation with TGF-β. Subsequent qRT-PCR analysis of the harvested cells revealed that the expression of FN, α-SMA, COL1α1, TIMP1, and SERPINE1 significantly increased in the presence of TGF-β and subsequently significantly decreased upon ERβ activation (Figure 4C). Immunofluorescent staining of α-SMA and vimentin showed lower mean gray values in fibroblasts treated with ERB041 compared with the TGF-β-treated fibroblasts (Figure 5A and B). Consistent with prior findings, α-SMA, vimentin, collagen I, and N-cadherin levels were significantly elevated in TGF-β-induced fibroblasts, while ERB041 treatment suppressed the expression of these fibrosis markers (Figure 5C). These findings suggest that fibroblasts adopt a myofibroblast-like phenotype following incubation with TGF-β, a transformation that ERB041 treatment appears to inhibit.

Functional Enrichment Analyses Confirm that ERβ Activation Targets TGF-β/Smad and TLR Signaling In Vivo

Based on our prior observations, we undertook comprehensive functional enrichment analyses to gain a deeper understanding of ERβ’s role. By utilizing GO enrichment analyses of DEGs derived from RNA-Seq, we were able to depict the dynamic changes in gene expression. Post-DSS treatment, downregulated genes were associated with 651 GO terms, and upregulated genes linked with 1747 GO terms (Supplementary Table 4). The top 20 GO terms encompassed responses to stimuli, responses to bacteria, and proteinaceous extracellular matrix (Figure 6A). In comparison, in the DSS + ERB041 group relative to the DSS group, downregulated genes were associated with 449 GO terms, and upregulated genes linked with 243 GO terms (Supplementary Table 5). The top 20 GO terms in this comparison included the extracellular region, inflammatory cell migration, responses to lipids, and flavonoid metabolism (Figure 6B). Notably, 2 biological process terms—cell migration and wound healing—were enriched from downregulated genes post-ERB041 injection.

We correlated the DEGs with the KEGG database and analyzed the top 20 enriched pathways to identify key signaling pathways. Interestingly, pathways of upregulated gene enrichment in the DSS group relative to the control group overlapped with those of downregulated gene enrichment following ERB041 intervention. These shared pathways included the TGF-β signaling pathway, TLR signaling pathway, Jak-STAT signaling pathway, hypoxia inducible factor (HIF)-1 signaling pathway, among others (Figure 6C and D). Our research uncovers a novel mechanism in which ERB041 targets key signaling pathways, thereby exerting a protective effect against DSS-induced intestinal fibrosis.

ERB041 Blocks Fibrogenesis In Vivo by Abrogating TGF-β/Smad and TLR4/MyD88/NF-κB Signaling

A key pathway significantly tied to the development of tissue fibrosis is the TGF-β/Smad signaling pathway. Utilizing insights from transcriptomic outcomes, we individually evaluated each protein of interest. Our western blot analysis suggested that chronically injured colon tissue exhibited enhanced expression levels of TGF-β, p-Smad2, and p-Smad2/3 when compared with the control group (Figure 6E). Moreover, the administration of ERB041 notably mitigated these elevated expressions in DSS-exposed mice. These observations substantiate that the traditional TGF-β/Smad signaling pathway is activated in mice subjected to DSS-induced chronic colitis and that ERB041 exhibits its antifibrotic impact by targeting this pathway.

To further understand the modulatory mechanism of ERB041, we examined the crosstalk of a classical signaling pathway with the innate immune system, namely, TLR4-NF-κB signaling. Enrichment analyses indicated that ERB041 intervention involved the following biological processes: TLR signaling pathway, and bacterial stimulation and response (Figure 6B, D). Colonic specimens were examined using western blotting for the expression of TLR4 and its adaptor molecule MyD88 and downstream transcription factor NF-κB. The results showed that the expressions of both TLR4 and MyD88 were increased in the DSS group compared with the control group, with concomitant accumulation of NF-κB p-p65. These elevated protein expressions were significantly decreased under the action of ERB041 (Figure 6F). Consistent with this, p-p65 protein expression was drastically increased in the nucleus after TGF-β stimulation, while the translocation of p65 into the nucleus was reversed by ERB041 (Figure S1A).

ERB041 Inhibits Fibroblast Activation In Vitro via Mitigating TGF-β/Smad and TLR4/MyD88/NF-κB Signaling

In fibroblasts, western blotting revealed that TGF-β treatment significantly increased the expression of TGF-β and the phosphorylation of Smad2 and Smad2/3 (Figure 7A). Moreover, intervention with ERB041 dramatically lowered the expressions of TGF-β, p-Smad2, p-Smad2/3, and TGF-β receptor II (TGF-β RII). These results suggested that ERB041 could partially abrogate TGF-β responsiveness by targeting TGF-β/Smad signaling in fibroblasts, possibly via altered TGF-β RII expression.

Given the alterations in the TLR4/MyD88/NF-κB axis observed in the murine model, we further conducted western blotting to evaluate the corresponding effects in vitro. As shown in Figure 7B, ERB041 treatment moderately attenuated the activity of TLR4/MyD88/NF-κB signaling and inhibited the nuclear translocation of p65 (Figure S1B). These outcomes suggest that ERβ activation can influence the differentiation of intestinal fibroblasts via the TLR4/MyD88/NF-κB pathway, thereby inhibiting fibrogenesis both in vitro and in vivo.

We subsequently assessed the influence of TLR4, a critical molecule in the TLR4/NF-κB pathway, on fibroblast activation in the presence of ERB041. Our aim was to ascertain whether TLR4 overexpression counteracts ERB041’s antifibrotic effects. Compared with the control and ov-NC groups, TLR4 overexpression through plasmid transfection significantly escalated TLR4 mRNA expression, corroborated by qRT-PCR results (Figure S2). Overexpression of TLR4 countered ERB041’s inhibitory impact on TLR4 signaling pathway proteins (Figure 7C). Additionally, TLR4 overexpression significantly diminished ERB041’s suppression of TGF-β1 stimulation-induced fibrosis markers such as α-SMA and collagen I in intestinal fibroblasts (Figure 7D). These findings suggest that the ERβ agonist counteracts fibrosis by regulating fibroblast activation through TLR4.

Transcription Factor ERβ Curbs the Promoter Activities of TGF-β and TLR4

Our findings suggest that the inhibitory effect of ERβ on fibroblast activation might be counteracted by the TGF-β/Smad and TLR4/Myd88/NF-κB pathways; the exact mechanism, however, remains unclear. Consequently, we utilized a dual luciferase reporter gene assay to examine whether ERβ directly moderates the promoter activities of TGF-β and TLR4. The 293T cells were cotransfected with an ERβ overexpression vector and the wild-type reporter plasmid, along with the negative control (NC). The results from the luciferase reporter assay revealed a significant decrease in luciferase expression due to the transcription factor ERβ (Figure 7E and F), suggesting that ERβ diminishes the promoter activities of both TGF-β and TLR4. Through the dual luciferase assay, we discerned the molecular regulatory mechanism of ERβ, which targets TGF-β or TLR4, thereby outlining potential interactions between ERβ-TGF-β and ERβ-TLR4.

Discussion

In this study, we discovered a distinct elevation of ECM deposition in tandem with diminished ERβ expression within the stenotic intestinal tissues of CD patients compared with their nonstenotic counterparts. We harnessed transcriptomic technologies to delve deeply into the ERβ’s antifibrotic influence, and the potential underlying mechanisms, within a murine model. Activating ERβ notably restrained overexpression of profibrotic genes and modulated the mRNA levels of MMPs and TIMP1 in vivo. In vitro experiments demonstrated ERB041’s ability to curb fibroblast activation, thereby reducing fibrogenesis. These mitigative effects are potentially mediated through the inhibition of TGF-β/Smad and TLR4/MyD88/NF-κB signaling pathways.

The relationship between ERβ and stenotic intestinal tissue in CD has been largely unexplored up until now. In this study, we pioneer the discovery of ERβ’s diminished expression in stenotic intestinal tissue of CD patients. Our sample consisted of 8 CD patients undergoing surgical intervention, of which only 1 was female. Therefore, keeping this concern in mind, we subanalyzed the data for the male sample. The results imply that the inclusion of 1 female did not significantly alter the data’s reliability. Additionally, we primarily investigated the role of ERβ, a key receptor through which estrogen exerts its biological effects, particularly in gastrointestinal processes.⁶ Estrogen receptor β is the predominant ER subtype in intestinal tissues and contributes prominently to intestinal regulatory T cell loss of function.^6,21,22 Goodman et al’s work with ERβ knockout mice, which demonstrated increased susceptibility to intestinal inflammation, guided our focus on ERβ rather than on estrogen broadly.²³ Previous research has shown no significant difference in ERβ expression between male and female subjects in both UC patients and healthy intestinal tissues, as well as in male and female mice models.^7,23–25 This finding informed our approach to not prioritize gender as a criterion in our participant selection. However, it is important to note that the specific expression patterns of estrogen receptors in the intestinal tissues of CD patients across different genders remain unclear. Recognizing this gap, we plan to expand our future studies to include a more diverse and balanced gender representation.

In our study, we have explored the role of ERβ in the colon, where its expression is notably high. Previous studies have established the antifibrotic potential of ERβ in various organs such as the skin, myocardium, and liver.^9–11 For instance, ERβ-selective agonists inhibit TGF-β signaling in benign prostatic hyperplasia disease and prevent fibrosis from progressing.⁸ Similarly, ERβ activation has been associated with reduced myocardial fibrosis and amelioration of cardiac hypertrophy,⁹ as well as inhibiting hepatic stellate cells from secreting extracellular matrix (ECM) components such as collagen IV and hyaluronic acid.¹¹ Interestingly, while abundance of ERβ in colonic tissues is well-documented, its functional role in this context has been less explored deeply. Using ERB041, a specific ERβ agonist, we analyzed the resultant changes in ERβ expression in colonic tissues of mice. Our findings, as depicted in Figure 3C, demonstrated a significant increase in ERβ expression post-ERB041 administration, suggesting its activation and potential influence on fibrotic processes. We unveiled that ERβ activation mitigates colitis-related fibrosis in vivo, as exemplified by the reduced levels of α-SMA, fibronectin, collagen I, and vimentin, along with the transcriptional downregulation of fibrogenic molecules such as Spp1, Ptpn13, and Thbs1. Typically, a disturbance in the balance between MMP and TIMP can accelerate the dysfunctional secretion and deposition of the ECM.²⁶ Our data also suggest that ERB041 administration reverses the transcription levels of Ecm1, Mmp2, Mmp3, Mmp8, Mmp9, and Timp1 compared with the DSS group. These observations collectively support the assertion that ERB041 acts as an antagonist to intestinal fibrosis.

Further, we investigated the impact of ERβ activation on human intestinal-derived fibroblasts. The results were compelling: ERβ activation led to a notable decrease in the mRNA expression levels of fibrosis-associated genes such as COL1α1, FN, α-SMA, TIMP1, and SERPINE1 and also reduced the expression of fibrotic markers like collagen I, α-SMA, and N-cadherin. Moreover, ERβ activation inhibited the migratory capabilities of these fibroblasts (Figures 4-5). These observations underscore the potential of ERβ activation as an antifibrotic intervention in intestinal tissues. Previous findings indicate that ERB041 significantly curtails the migration of colon cancer cells.²⁷ In this current study, we discovered that while ERB041 does not impact fibroblast proliferation, it can modulate their migratory ability triggered by TGF-β. Existing research by Burke et al also supports that TGF-β enhances fibroblast migration and causes a notable increase in the expression of protein N-cadherin and gene S1004a in fibroblasts extracted from patients with fibrostenosing CD.^14,28 N-cadherin, an adhesive protein, is pivotal for cell-cell adhesion, cell migration, cell invasion, and wound healing.^29,30 Aligning with these observations, our investigation reveals that N-cadherin was significantly downregulated following ERB041 treatment. Moreover, the metastatic potential of several cancers, including breast, pancreatic, and gastric, has been linked with S100a4 expression.²⁸ In our study, the analysis of differentially expressed genes suggested a decrease in the mRNA level of S100a4 after ERB041 intervention. These collective data propose that ERB041 holds the potential to modulate the migration of intestinal fibroblasts.

To identify the key profibrotic signaling pathways involved in the ERB041-induced amelioration of intestinal fibrosis, we conducted a KEGG enrichment analysis, focusing on the TGF-β/Smad signaling pathway. The TGF-β/Smad pathway is recognized as a strong promoter of fibroblast-myofibroblast transition and plays a crucial role in regulating both the production and degradation of ECM components.³¹ The initiation of this signaling pathway is triggered by TGF-β binding to its receptors, TGF-β-RI and TGF-β-RII, leading to the formation of complexes that stimulate the phosphorylation of proteins in the Smad family.³² Following this, p-Smad2 and p-Smad3 bind with Smad4 (co-Smad), and the resultant complex moves into the nucleus, acting as a transcription factor that modifies the expression of target genes. Our current research demonstrated through western blotting analysis that activation of ERβ significantly suppresses the expressions of TGF-β, along with p-Smad2 and p-Smad2/3 in the murine model.

Our KEGG analysis spotlighted another critical profibrotic pathway significantly enriched—the TLR signaling pathway, a research hot spot in recent years due to its suspected role in excessive tissue remodeling.^33–35 Among the various TLRs, TLR4 has been identified as a significant contributor to fibroblast activation.³⁶ Research by Jun et al indicated milder symptoms of intestinal edema and fibrosis in TLR4-deficient mice.¹⁷ There is also mounting evidence of interaction between TLR4 and TGF-β signaling pathways contributing to fibrosis in other organs.^36,37 This is primarily driven by TGF-β’s role in activating the TLR4/NF-κB signaling pathway, resulting in a fibrotic phenotype. Generally, TLR4 signaling unfolds via a Myeloid Differentiation Factor 88 (MyD88)-dependent or an MYD88-independent pathway, leading to NF-κB’s nuclear translocation. The TLR4/NF-κB signaling pathway has been successfully targeted to mitigate fibrosis in various organs, such as the liver, kidneys, and heart.^38–40 Our research shows that ERB041, in therapeutic applications, led to inhibition of the TLR4/NF-κB signaling pathway. Recently, research has indicated that the elimination of intestinal ERβ could modify the variety and abundance of gut microbiota in mice with colitis.⁴¹ This phenomenon might be explained by the potential interaction between ERβ and TLR4, given the known strong link between the gut microbiota and TLR4. This intriguing scientific question merits additional exploration.

We subsequently assessed the impact of ERB041 on TGF-β/Smad signaling in the context of human intestinal fibroblast activation. Through western blotting, we were able to confirm that the inhibitory effect of ERB041 on fibroblast activation is likely a result of the suppression of the TGF-β/Smad pathway. Remarkably, our dual luciferase reporter gene assay revealed that ERβ could impede TGF-β signaling by interacting with the promoter of its target gene, TGF-β, a finding hitherto unreported. Notably, our study observed a significant increase in TLR4 expression in fibroblasts upon TGF-β stimulation. While it remains unclear whether this is due to a direct interaction between TLR4 and TGF-β, or a result of ECM secretion induced by fibroblast response to TGF-β stimulation, the finding is noteworthy. A broad range of ligands, including heat shock proteins, fibronectin, hyaluronic acid, and fibrinogen are all known to be recognized by the TLRs.⁴² In summary, we demonstrated that ERB041 reverses the fibrotic phenotype via a MyD88-dependent pathway both in vivo and in vitro. Our dual luciferase reporter gene experiments also revealed a potential interaction mechanism between ERβ and TLR4, highlighting a significant avenue for further exploration.

The interplay between inflammation and fibrosis in the intestinal context is indeed complex.^12,43 The activation of ERβ has been shown to reduce intestinal inflammation, as evidenced by our own research and others.^6,44 It is important to note that anti-inflammatory treatments alone are not entirely effective in preventing or treating intestinal fibrosis.¹² This suggests that the anti-inflammatory effects of ERβ activation do not fully account for its protective role against fibrosis. Our research provides further insights into the specific mechanisms by which ERβ exerts its anti-fibrotic effects. We demonstrated that the administration of ERB041 significantly reduced TGF-β-induced fibroblast activation and migration. This was accompanied by a marked decrease in the expression of TGF-β and its downstream effectors, phosphorylated Smad2 and Smad3, in colonic tissues and fibroblasts. Using an in vitro model, we elucidated how ERβ activation downregulates the TLR4 pathway, thereby contributing to its antifibrotic action. Additionally, ERB041 effectively inhibited the migratory and ECM-producing capabilities of activated fibroblasts. Based on these findings, we propose that the antifibrotic effects of ERβ are predominantly mediated through its modulation of the TGF-β/Smad and TLR4/NF-kB pathways, rather than solely through its anti-inflammatory action.

While this study provides valuable insights, there are several limitations worth noting. In general, intestinal fibrosis models are categorized into 7 groups: chemically induced, spontaneous, gene-targeted, immune-mediated, bacteria-induced, and radiation-induced, as well as postsurgical fibrosis.^45–47 It is widely recognized that no single animal model can capture all the pathogenic and clinical features of human intestinal fibrosis due to the disease’s multifactorial nature.⁴⁵ Immune-mediated and bacteria-induced models are valuable for studying immune responses and microbial interactions in fibrosis.⁴⁶ The evidence supporting the applicability of gene-targeted models to studies of ERβ is currently insufficient. The SAMP1/Yit model, with its segmental lesions primarily in the terminal ileum, closely mimics human Crohn’s disease.^45,48 However, the difficulty in breeding these mice limits their widespread use. The chemically induced model, particularly DSS and trinitrobenzene sulfonic acid (TNBS), remains the most extensively used due to its reproducibility, control over inflammation severity, and its well-characterized pathogenesis, which includes distinct phases of injury and repair.^47,48 This model has proven especially useful in studying the regulatory role of ERβ on mucosal immunity.⁶

Dextran sulfate sodium, particularly in a regime involving prolonged exposure to low doses or cyclic administration of higher doses, is recognized for replicating a chronic colitis model, eventually leading to fibrosis.⁴⁵ This approach is considered to be the easiest and most reproducible protocol to induce colonic inflammation with associated fibrosis, making it relevant for our investigation into chronic intestinal fibrosis.^45,46,49 In contrast, considerable variation exists in TNBS-induced colitis, depending on the source and amount of TNBS.⁴⁵ Trinitrobenzene sulfonic acid leads to a shift in the immune response from Th1 to Th2/Th17 dominance, followed by collagen deposition and fibrosis.^45,47 There is no evidence that haptens play a role in human IBD.^45,46 Studies have demonstrated that mice can develop colonic fibrosis following single or multiple DSS cycles, depending on their genetic background.^18,50 Increased levels of TGF-β1, MMP-2, MMP-9, and collagen deposition have been documented in mice post-DSS treatment, aligning with the pathological features observed in human IBD.⁴⁸ These considerations, alongside the specific objectives of our study, guided our decision to use the DSS model to simulate intestinal fibrosis.

Another potential limitation lies in our inability to extract primary colonic fibroblasts from human tissues. Consequently, the precise mechanisms of ERβ action on fibrostenotic CD remain elusive and necessitate further investigation. Recognizing this limitation, we strategically utilized both stenotic and nonstenotic intestinal tissues from CD patients. This approach allowed us to investigate the effects of ERβ activation in a more clinically relevant context. To further elucidate the role of ERβ in intestinal fibrosis, we employed an in vitro model using human intestinal-derived fibroblasts. Subsequent treatment with the ERβ agonist ERB041 provided critical insights into how ERβ activation influences fibroblast behavior and fibrosis development. Future studies focusing on the isolation and characterization of primary colonic fibroblasts from human CD tissues would be invaluable in delineating the specific pathways through which ERβ exerts its effects. Finally, it is essential to highlight that while our study demonstrated that ERB041 mitigated intestinal fibrosis by inhibiting fibroblast activation, the cellular landscape involved in intestinal fibrosis is diverse and extends beyond fibroblasts. It includes epithelial cells, endothelial cells, and muscularis propria muscle cells, all of which may play significant roles in the fibrotic process. Further research is warranted to elucidate the precise roles these various cell types play in the development and progression of intestinal fibrosis.

Conclusions

Our research strongly indicates that activating ERβ could serve as an effective approach for addressing intestinal fibrosis, notably through impeding both TGF-β/Smad and TLR4/MyD88/NF-κB signaling pathways. These findings not only substantiate the promise held by ERβ as a therapeutic target for intestinal fibrotic strictures but also pave the way for future in-depth investigations into the therapeutic potential of ERβ in the management of intestinal fibrosis. Therefore, we invite further research to expand on these findings and explore the full breadth of ERβ’s applicability in treating intestinal fibrotic strictures.

Supplementary Data

Supplementary data is available at Inflammatory Bowel Diseases online.

Author Contributions

F.L. performed the experiments and wrote the article. Y.C. conducted the bioinformatics analysis. F.L., Y.C., and M.X. were responsible for the mouse feeding and model establishment. L.T., J.L., H.W., and S.L. were involved in data curation and analysis. L.Z. conceptualized the study and revised the manuscript. G.S. supervised the study and edited the manuscript. The schematic in the article was created with BioRender.com.

Funding

This work is supported by National Key R&D Program of China (2023YFC2507300), the National Natural Science Foundation of China (82170547 and 81873558), the Major Projects of the Ministry of Science and Technology of China (2018YFC0114600), Guangxi Natural Science Foundation (2024GXNSFBA010039), the Natural Science Foundation of Hubei Province (2021CFB450), Wuhan Knowledge Innovation Special Basic Research Project (2022020801010462), Science Research Foundation of Union Hospital (2022xhyn009), and Teaching Reform Project of the First Clinical College (202120).

Conflicts of Interest

The authors declare that there is no conflict of interest.

Data Availability

The data presented in this study are all contained within the main body and the Supplementary Materials of this article.

References