Danggui-Baishao herb pair protects against dextran sulfate sodium-induced colitis by modulating the Wnt/β-catenin pathway

Abstract

BACKGROUND

The Danggui-Baishao herb pair is the foundation of a traditional Chinese medicine formula known as Shaoyao decoction, which is widely used in the treatment of colitis.

AIM

To uncover the mechanisms underlying the anti-colitis effects of the Danggui-Baishao herb pair.

METHODS

The chemical composition of the herb pair was characterized by high performance liquid chromatography-quadrupole/time of flight mass spectrometry analysis. A mouse model of colitis was induced by administering 2.5% dextran sulfate sodium. The therapeutic effects of the herb pair were evaluated based on body weight changes, colon length, histopathological, intestinal inflammation, and barrier function. To investigate the underlying mechanisms, RNA sequencing, metabolomics, 16S rRNA sequencing, metagenomics, and the β-catenin inhibitor ICG-001 were utilized. Furthermore, molecular docking and dextran sulfate sodium-treated HCT 116 cells were conducted to explore the protective mechanisms of benzoylpaeoniflorin.

RESULTS

The herb pair improved body weight, colon length, intestinal inflammation, and barrier function. Additionally, the herb pair upregulated the expression of intestinal stem cells marker leucine-rich repeat-containing G-protein coupled receptor 5 and proliferation-related proteins. RNA sequencing analysis showed that the herb pair activated the Wnt/β-catenin signaling pathway. Metabolomic analysis revealed changes in bile acids composition. Through 16S rRNA and metagenomic sequencing, it was observed that the herb pair modulated the gut microbiota, with an enrichment of probiotics and a depletion of pathogenic bacteria. Following intraperitoneal injection of antagonist ICG-001, the therapeutic efficacy was diminished. Molecular docking showed that benzoylpaeoniflorin can bind to β-catenin. Furthermore, benzoylpaeoniflorin can activated the Wnt/β-catenin signaling pathway and the therapeutic efficacy was also diminished by the ICG-001 in vitro.

CONCLUSION

The herb pair effectively reduces colonic inflammation and maintains the integrity of the intestinal barrier. Moreover, the anti-colitis efficacy of the herb pair is closely associated with activation of the Wnt/β-catenin pathway.

Key Words: Danggui-Baishao herb pair; Benzoylpaeoniflorin; Ulcerative colitis; Wnt/β-catenin pathway; Intestinal barrier; Gut microbiota; Bile acids

Core Tip: The Danggui-Baishao herb pair has shown definite anti-colitis effects in traditional Chinese medicine. By detecting the active ingredients of the herb pair and using transcriptome sequencing for mechanism prediction, our results show that the herb pair protects against colitis by activating the Wnt/β-catenin pathway in vivo, yet this effect is diminished by the β-catenin antagonist ICG-001. Through metabolomics, 16S rRNA, and metagenomic further revealed that the herb pair regulates bile acids metabolism and gut microbiota composition. Finally, molecular docking revealed that benzoylpaeoniflorin has a structural basis for binding to β-catenin. Benzoylpaeoniflorin can activating the Wnt/β-catenin pathway in vitro and diminished by the ICG-001. This study provides experimental evidence for research on anti-colitis drugs.

INTRODUCTION

Ulcerative colitis (UC) is one of the primary clinical subtypes of inflammatory bowel disease (IBD), and is characterized by intestinal damage, loss of intestinal barrier integrity, and an expanded inflammatory response[1]. Approximately 5 million people worldwide are affected by UC, and the incidence rate is on the rise[2]. This has thus imposed a significant burden on healthcare systems. The etiopathogenesis of UC remains complex and enigmatic, involving a multifaceted interplay of factors that disrupt intestinal barrier function, impair immune regulation, cause dysbiosis of the gut microbiome, and induce genetic alterations. Despite the existence of multiple therapeutic approaches that have yet to achieve clinical efficacy, the majority of currently available treatments for IBD primarily focus on suppressing inflammation. Patients may experience residual symptoms during disease remission due to defective wound healing processes[3], and the use of immunosuppressants can also increase the risk of infection and various cancers. Consequently, in-depth research on the pathogenesis of UC and the pursuit of new intervention drugs are of vital importance.

The intestinal barrier, recognized as the body’s internal defense, can effectively prevent the entry of detrimental substances and pathogens. It is widely recognized that persistent injury to the intestinal barrier is the main pathological mechanism underlying the early onset of IBD[4]. The intestinal epithelial cells constitute the fundamental component of mucosal barrier function. It has been reported that all the cell types derived from intestinal stem cells (ISCs) reside at the crypt base and are shielded from the influence of soluble metabolites in the intestinal lumen, including intestinal microbiota-derived metabolites that have an inhibitory effect on the proliferation potential of stem cells[5]. Studies have shown that in a colitis model, the leucine-rich repeat-containing G-protein coupled receptor 5 (Lgr5) gene in ISCs is almost completely depleted by the third day[6]. Moreover, scientists have focused on mucosal barrier healing as a therapeutic approach to treating UC, with the modulation of induced ISCs emerging as the most promising strategy[7].

The precise mechanisms governing the regulation of ISCs expression or stability remain a subject of ongoing research and debate. It is well-established that microbiota-related mucosal barrier injury plays a pivotal role in colitis and that the subsequent mucosal barrier injury facilitates the dysregulation of microbiota-induced inflammatory cascades[8,9]. A recent study demonstrated that microbiota-derived tyramine from Enterococcus suppresses the proliferation of ISCs, thereby impairing epithelial regeneration and exacerbating dextran sulfate sodium (DSS)-induced colitis through the activation of adrenergic receptor alpha 2[10]. Lactobacillus-derived lactate activates Wnt/β-catenin signaling through Gi-protein-coupled receptor 81, thereby promoting the proliferation of ISCs epithelium[11]. Researchers also identified that the subsequent activation of farnesoid X receptor (FXR) by microbiota-derived deoxycholic acid triggered Wnt signal-dependent ISCs maintenance[12]. Wnt/β-catenin is a well-documented regulator of ISCs proliferation and fate control. Disorders in Wnt signaling can reduce the number of intestinal secretory cells, further impairing the mucus barrier and ultimately resulting in integrity damage[13]. Glycogen synthase kinase-3β (GSK-3β) degrades β-catenin through phosphorylation. Unphosphorylated β-catenin transfers to the nucleus, thereby activating the expression of ISCs-related target genes[14]. Consequently, the function of ISCs and the regulation of ISCs homeostasis are precisely controlled by both microbiota and host signals.

Traditional Chinese medicine (TCM) has a long history of treating UC and plays a crucial role in its management. Herb pair (referred to as “yao dui” or “dui yao” in TCM) represent the simplest and fundamental form of multiple herbal preparations. Compared with individual plants, they exhibit significantly enhanced pharmacological efficacy, a broader therapeutic range, and reduced toxicity. Danggui and Baishao are derived from the dried root of Angelica sinensis (Oliv.) Diels and Paeonia lactiflora Pall, respectively. The Danggui-Baishao herb pair (hereafter referred to as “herb pair”) is the central component of the classic anti-colitis formula, Shaoyao decoction, and is extensively used in clinical practice for colitis treatment[15] and was recently confirmed to restore the mucus layer in colitis mice[16]. Yin and blood deficiency are important pathological mechanisms of UC, and the herb pair has the function of regulating yin and blood[17]. In the Song dynasty, Dou Cai initially used the herb pair for treating diarrhea accompanied by blood or pus and abdominal pain, which are the primary symptoms of UC. Experimental research has also demonstrated that Angelica sinensis polysaccharide exhibits substantial efficacy in improving body weight loss, colon shortening, and the disease activity index score[18]. Paeonia lactiflora Pall. extract showed positive effects on colon length and cytokine levels in a colitis model[19]. Given the favorable outcome of the herb pair, its underlying mechanism requires further investigation.

In the present study, we initially explored the restorative effects of the Danggui-Baishao herb pair on intestinal mucosal barrier integrity and ISCs proliferation. Then, we investigated its potential in influencing the microbiota, bile acids (BAs) metabolism, and the colonic Wnt/β-catenin signaling pathway. Furthermore, we elucidated the potential mechanism of action of the herbs in the treatment of colitis with the aid of the β-catenin inhibitor ICG-001. Finally, molecular docking and in vitro model were conducted to explore the protective mechanisms of benzoylpaeoniflorin.

MATERIALS AND METHODS

Preparation of herb pair and high performance liquid chromatography-quadrupole/time of flight mass spectrometry analysis

According to the ancient Chinese book Bianquexinshu by Dou Cai during the Song dynasty, the herb pair is composed of Danggu and Baishao (1:1, w/w), and the granules used in this study were acquired from Tianjiang Pharmaceuticals (Jiangsu, China). A Shimadzu LCMS-9050 system was employed for high performance liquid chromatography-quadrupole/time of flight mass spectrometry analysis to characterize the chemical composition of the herb pair. Chromatographic separation was carried out at 35 °C with the mobile phase composed of water (A) and acetonitrile (B) using a Waters ACQUITY UPLC HSS T3 chromatographic column. The injection volume was set at 1 μL and the flow rate at 0.3 mL/minute, following the gradient elution procedure below: 0-5 minutes, 10% B; 5-30 minutes, 15% B; 30-50 minutes, 30% B; 50-65 minutes, 50% B; 65-80 minutes, 70% B; 80-90 minutes, 90% B; 90-95 minutes, 95% B; 95-100 minutes, 10% B. The quadrupole/time of flight mass spectrometry detection was carried out using the positive/negative ion switch mode, and the m/z scanning range was set at 100-1200. Key instrumental parameters included: The desolvation pipeline temperature at 250 °C, the interface temperature at 300 °C, the hot block temperature at 400 °C, the atomizer gas flow rate at 1 L/minute, and the drying gas flow rate at 10 L/minute. Real-time mass axis calibration was carried out using a 400 mg/L sodium iodide solution. Finally, data collection and analysis were carried out using Labsolution software (Shimadzu, Japan).

Animal models

All specific pathogen-free male C57BL/6J mice (six-to-eight-week-old) were purchased from Sipeifu Laboratory Animal Technology Co., Ltd., Beijing, China. All mice were kept in specific pathogen-free facility with a 12-hour light/dark cycle, temperature of 22 ± 1 °C, humidity at 50% ± 15%, and free access to clean water and food. The Animal Care Ethics and Use Committee of Peking Union Medical College Hospital, No. XHDW-2024-37-1 approved the experimental design and procedures in this study.

After one week of adaptive feeding, 54 mice were randomly divided into the following six groups: Control, control + herb pair-high, DSS, herb pair-low, herb pair-high, and mesalazine (A79809, Sigma-Aldrich, MO, United States). Both the control and the control + herb pair-high groups were given clear water. The other four groups of mice received 2.5% DSS (36000-50000 MW, CAS: 216011080, MP Biomedicals, CA, United States) in their drinking water for five days, then switched to clean drinking water for two days in order to induce the acute colitis model. Mice in the herb pair groups and the mesalazine group were administered 5.2 g crude drug/kg or 10.4 g crude drug/kg and mesalazine 50 mg/kg by oral gavage daily for seven days, respectively. The same volume of normal saline was administered to mice in the control group and the DSS group by gavage.

In the inhibitor experiment, we randomly divided 24 mice into four groups: DSS, DSS + herb pair, DSS + ICG-001 (HY-14428, MCE, NJ, Unites States), and DSS + ICG-001 + herb pair. Mice in the DSS + herb pair and DSS + ICG-001 + herb pair groups received the herb pair (10.4 g crude drug/kg) by oral gavage daily for seven consecutive days. Mice in the DSS + ICG-001 and DSS + ICG-001 + herb pair groups were administered ICG-001 (20 mg/kg) by intraperitoneal injection every other day. The same volume of control saline was administered to mice in the DSS and DSS + herb pair groups by intraperitoneal injection.

Cell culture and treatment

HCT116 human colon carcinoma cells were maintained in Roswell Park Memorial Institute-1640 medium (11875093, Gibco, NY, United States) supplemented with 10% fetal bovine serum (10099141C, Gibco, NY, United States). The cells were cultured at 37 °C in a humidified atmosphere containing 5% CO₂. To establish an in vitro model, the cells were treated with 2% DSS. Benzoylpaeoniflorin (HY-N0852, MCE, NJ, Unites States) was applied at three different concentrations: Low (Ben-L, 25 μM), medium (Ben-M, 50 μM), and high (Ben-H, 100 μM), and co-treated with 2% DSS for 24 hours. In inhibitor experiments, ICG-001 (38 μM) is an inhibitor of β-catenin/T-cell factor-mediated transcription.

Histological analysis

Colon tissues were fixed with 4% paraformaldehyde, then dehydrated and embedded in paraffin. Subsequently, the sections were stained with hematoxylin and eosin, Alcian blue (AB), and AB-periodic acid Schiff (PAS) according to standard protocols, respectively.

Enzyme-linked immunosorbent assay

Enzyme-linked immunosorbent assay (CSBE13066m) kit obtained from Wuhan Huamei Biological Engineering Co., Ltd. (Wuhan, China) was used to detect the content of lipopolysaccharide (LPS) in serum samples following the manufacturer’s instructions.

Intestinal permeability analysis

The fluorescein isothiocyanate (FITC)-labeled dextran method was used to evaluate intestinal permeability. In brief, food and water were withheld for 8 hours before the mice were sacrificed. Subsequently, FITC-dextran (Sigma-Aldrich, MO, United States, 4 kDa, 600 mg/kg) was intragastrically administered. Blood was collected under isoflurane anesthesia after 4 hours, and the supernatant was centrifuged (avoiding light).

Immunohistochemical analysis

The colon tissue sections underwent antigen repair and were blocked with 5% goat serum. They were incubated with anti-CD11b (1:500, ab133357, ABCAM, United Kingdom), anti-F4/80 (1:500, ab111101, ABCAM, United Kingdom), anti-Lgr5 (1:500, ab75850, ABCAM, United Kingdom), anti-proliferating cell nuclear antigen (PCNA) (1:500, 60097-1-Ig, Proteintech, IL, United States), anti-Ki67 (1:100, MA514520, Thermo Fisher Scientific, MA, United States), and anti-β-catenin (1:500, ab32572, ABCAM, United Kingdom) overnight at 4 °C. Subsequently, secondary antibodies were applied, and the sections were stained with 3,3′-diaminobenzidine. Finally, the sections were dehydrated, rendered transparent, and sealed with neutral resin.

Quantitative polymerase chain reaction

The HiPure Total RNA Plus Kit (R4111, Magen, Beijing, China) was used to extract total RNA from colon tissues and the RevertAid First Standard cDNA Synthesis Kit (K1622, Thermo Fisher Scientific, MA, United States) was subsequently used to complete reverse transcription. Finally, the ABI 7500 system (Life Technologies, CA, United States) was used for polymerase chain reaction (PCR) detection. The quantitative analysis of all genes was compared to β-actin. Primer sequences are shown in Supplementary Table 1.

Western blotting assay

The protein expression of Occludin (1:1000, ab216327, ABCAM, United Kingdom), Claudin1 (1:1000, ab15098, ABCAM, United Kingdom), Claudin2 (1:1000, 26912-1-AP, Proteintech, IL, United States), E-cadherin (1:1000, 3195S, CST, MA, Unites States), β-catenin (1:1000, ab32572, ABCAM, United Kingdom), p-GSK-3β (1:1000, 5558S, CST, MA, Unites States), GSK-3β (1:1000, 12456S, CST, MA, Unites States), Cyclin D1 (1:1000, ab134175, ABCAM, United Kingdom), Lgr5 (1:1000, ab75850, ABCAM, United Kingdom), FXR (1:1000, 417200, Thermo Fisher Scientific, MA, United States), β-tubulin (1:5000, ab179513, ABCAM, United Kingdom), and glyceraldehyde 3-phosphate dehydrogenase (1:5000, 10494-1-AP, Proteintech, IL, United States) in colon tissues was detected by western blot analysis. The total protein in colon tissues was extracted and homogenized, and protein concentrations were then determined using the BCA Protein Assay Kit (KGP902, KeyGEN, Nanjing, China). Sodium dodecyl sulfate polyacrylamide gel electrophoresis (PG112, Epizyme, Shanghai, China) was used to separate an equal amount of sample protein and then transferred to polyvinylidene fluoride membranes (IPVH00010, Millipore, MA, United States). 5% skim milk (P1622, APPLYGEN, Beijing, China) was used to block the membranes for 90 minutes, and the indicated primary antibody was used to incubate the membranes for 60 minutes. Subsequently, all membranes underwent five washings with Tris-buffered saline with Tween (PS103, Epizyme, Shanghai, China), and then incubated with secondary antibody for 90 minutes. Binding signals were detected using the enhanced chemiluminescence system.

RNA sequencing

TRIzol^® reagent (Invitrogen, Carlsbad, CA, United States) was used to extract total RNA from colon tissues. RNA sequencing analysis was performed on an Illumina HiSeq platform. The analysis was conducted using the R package edgeR, and differentially expressed genes (DEGs) with an adjusted P value less than 0.05 were considered statistically significant.

Determination of BAs

Colon contents were resuspended, and then the sample was diluted with water. The diluted sample (100 μL) was mixed with 300 μL of acetonitrile/methanol (8:2) and placed on ice for 30 minutes. The supernatant was subsequently removed by centrifugation, and testing was carried out. The Waters ACQUITY UPLC BEH C18 column was used to complete the separation which was maintained at 50 °C. 0.1% formic acid and 5 mmol/L ammonium consisted of the mobile phase and then acetate in water (A) and acetonitrile (B), the transport flow rate was set at 0.35 mL/minute. The solvent gradient was set as follows: 0.5 minute, 5% B; 1.5 minutes, 5%-30% B; 4 minutes, 30%-37% B; 5 minutes, 37%-38% B; 5.5 minutes, 38%-39% B; 6 minutes, 39%-42% B; 6.5 minutes, 42%-43% B; 9.5 minutes, 43%-50% B; 11 minutes, 50%-60% B; 12 minutes, 60%-95% B; 13.1 minutes, 95%-5% B; 15 minutes, 5% B. The mass spectrometer operating parameters were as follows: Curtain gas (30 psi), ion spray voltage (-4500 V), ion source temperature (550 °C), ion source gases 1 and 2 (60 psi).

16S rRNA sequencing

Fresh colon contents were extracted for DNA using standard protocols. Subsequently, the V3-V4 hypervariable region of the 16S ribosomal RNA gene was amplified and sequenced in accordance with the Illumina 16S metagenomic sequencing library protocol. Data analysis was performed utilizing the online platform (https://magic-plus.novogene.com).

Metagenomic sequencing

DNA was extracted from colon contents using standard protocols. The DNA library preparation and metagenomic sequencing of all samples were performed using Illumina NovaSe™ X Plus. In brief, quality control and host filtering of the raw data were carried out to obtain clean data. Subsequently, the assembly was processed to generate scaffolds, followed by gene prediction and the removal of redundant gene catalogs. The samples were annotated with species, function, and resistance gene annotations. Finally, diversity analysis of species and function was conducted.

Molecular docking

First, the three-dimensional crystal structure of the β-catenin protein (Protein Data Bank: 1JDH) was retrieved from the Protein Data Bank. Subsequently, the structure of benzoylpaeoniflorin was obtained from the Pubchem database, and then converted into a pdbqt file via Open Babel software. Molecular docking was performed using AutoDock Vina. Finally, the docking results were visualized and analyzed using PyMOL software.

Biosafety analysis

The clinical safety of the herb pair was assessed by a maximum dose test. The survival rate and body weight of 20 mice were monitored. The anatomical structures of all mice were then examined, and any alterations in their major organs were observed. Peripheral blood was collected to evaluate liver and kidney function.

Statistical analysis

GraphPad Prism software, version 10.0 (San Diego, CA, United States) was used to analyze all experimental data. Following normality and homogeneity of variance analyses, the mean ± SD was expressed as quantitative data. When the data were normally distributed, the Student’s t-test was used to compare two groups, and one-way ANOVA was used for three or more groups, and the Dunnett post-test was employed for statistical analysis. A P value less than 0.05 was defined as statistically significant.

RESULTS

Danggui-Baishao herb pair alleviated colon injury

Based on the high performance liquid chromatography-quadrupole/time of flight mass spectrometry results, eight compounds were identified: Catechin, mudanpioside E, albiflorin, paeoniflorin, ferulic acid, galloylpaeoniflorin, benzoylalbiflorin, and benzoylpaeoniflorin (Supplementary Figures 1-6, Supplementary Table 2). We performed a safety evaluation of the herb pair, and its safety was formally confirmed through a maximum dose test (Supplementary Figure 7). Then, an acute colitis model was prepared and body weight was used to evaluate the severity of the DSS model[20]. As anticipated, colitis mice exhibited substantial weight loss. However, administration of the herb pair in high doses, along with mesalazine, markedly alleviated the weight loss (Figure 1A). Furthermore, colitis mice exhibited shortened colons, while administration of the herb pair and mesalazine clearly elongated colon length (Figure 1B and C). Hematoxylin and eosin staining revealed that colon tissue sections in the control and the control + herb pair-high group exhibited normal mucosal epithelium, and no other apparent abnormalities. In contrast, the DSS group showed tissue ulcers, loss of mucosal epithelium, submucosal edema, infiltration of mucosal and submucosal inflammatory cells, and the histological scores were higher. The herb pair-high group demonstrated significant improvement in the above-mentioned pathological manifestations and had reduced histological scores (Figure 1D and E). Therefore, we used the herb pair-high group for further molecular biology analysis. In subsequent experiments, the herb pair-high group was uniformly represented as the herb pair group. Additionally, quantitative PCR (qPCR) analysis showed a significant decrease in the expression levels of tumor necrosis factor, interleukin 6, interleukin 1b, interleukin 18, and toll-like receptor 4, V-rel reticuloendotheliosis viral oncogene homolog A in the herb pair group (Figure 2A-C). Conversely, interleukin 10 expression was significantly elevated in the herb pair group (Figure 2A). However, interferon-gamma and interleukin 13 exhibited a descending trend without statistical significance (Figure 2A). Immunohistochemical (IHC) analysis showed that the DSS treatment caused a significant increase in macrophage infiltration, herb pair treatment dramatically decreased the presence of macrophage in the colon tissues (Figure 2D-F). Over, these results suggest that herb pair possessed a significant therapeutic effect in DSS-induced mice.

Figure 1 Herb pair exerted therapeutic effects on colitis mice induced by dextran sulfate sodium. A: Changes in relative body weight; B: Colon length changes in mice from each group; C: Representative images of the colon length; D: Representative images of hematoxylin and eosin (HE) staining in colon tissues; E: Histological scores of HE staining of colon tissues. For body weight and colon length measurement: n = 9, for HE staining: n = 3. ^bP < 0.01 vs dextran sulfate sodium group, ^cP < 0.001 vs dextran sulfate sodium group. DSS: Dextran sulfate sodium.

Figure 2 Herb pair inhibited intestinal inflammation in colitis mice. A: Quantitative polymerase chain reaction (qPCR) analysis of mRNA expression of tumor necrosis factor, interleukin 6, interleukin 1b, interleukin 18, interferon-gamma, interleukin 13, and interleukin 10 in colon tissues; B: QPCR analysis of mRNA expression of toll-like receptor 4 in colon tissues; C: QPCR analysis of mRNA expression of V-rel reticuloendotheliosis viral oncogene homolog A in colon tissues; D: Representative images of immunohistochemical (IHC) of CD11b and F4/80 in colon tissues; E and F: Quantitative analysis of IHC results of CD11b and F4/80. For qPCR analysis: n = 5-6, for IHC measurement: n = 3. ^aP < 0.05 vs dextran sulfate sodium group, ^bP < 0.01 vs dextran sulfate sodium group, ^cP < 0.001 vs dextran sulfate sodium group. Tnf: Tumor necrosis factor; Il6: Interleukin 6; Il1b: Interleukin 1b; Il18: Interleukin 18; Ifng: Interferon-gamma; Il13: Interleukin 13; Il10: Interleukin 10; DSS: Dextran sulfate sodium.

Danggui-Baishao herb pair improved intestinal barrier integrity

The protective effects of the herb pair on the intestinal barrier were subsequently assessed. Western blot analysis showed that DSS treatment decreased the expression levels of tight junction proteins such as Occludin and Claudin1, adherens junction protein E-cadherin, and increased the expression level of Claudin2. Conversely, the herb pair prevented all of these changes (Figure 3A-E), indicating the protective effect of the herb pair on the intestinal barrier. Subsequently, the FITC-dextran fluorescence assay and serum LPS concentration were conducted to determine any increases in intestinal permeability. Our results indicated that the FITC-dextran-positive signal and expression level of LPS were elevated in the DSS group, but reduced by treatment with the herb pair (Figure 3F and G). Furthermore, AB staining and AB-PAS staining demonstrated that mucus production decreased and goblet cells shrank in the DSS group, while treatment with the herb pair improved these manifestations (Figure 3H-J).

Figure 3 Herb pair reinforced intestinal barrier in colitis mice. A-E: Western blotting analysis of protein expression of Occludin, Claudin1, Claudin2, and E-cadherin in colon tissues; F: Fluorescein isothiocyanate-dextran fluorescence intensity in serum; G: Expression level of lipopolysaccharide in serum; H: Representative images of Alcian blue (AB) staining and AB-periodic acid schiff (PAS) staining of colon tissues; I and J: Quantitative analysis of AB staining and AB-PAS staining. For western blotting analysis: n = 3, for fluorescein isothiocyanate analysis: n = 5-7, for lipopolysaccharide measurement: n = 7, for AB staining and AB-PAS staining: n = 3. ^aP < 0.05 vs dextran sulfate sodium group, ^bP < 0.01 vs dextran sulfate sodium group, ^cP < 0.001 vs dextran sulfate sodium group. DSS: Dextran sulfate sodium; GAPDH: Glyceraldehyde 3-phosphate dehydrogenase; LPS: Lipopolysaccharide; AB: Alcian blue; PAS: Periodic acid Schiff.

Danggui-Baishao herb pair stimulated the proliferation of intestinal stem cells

An investigation into the effects of the herb pair on ISCs and proliferation-related markers found that the herb pair increased the expression of the ISCs marker Lgr5 (Figure 4A and B). Subsequent qPCR experiments indicated that colitis decreases the mRNA relative expression of stem cell and proliferation-related genes, including Lgr5, SRY-box transcription factor 9 (Sox9), homeodomain-only protein homeobox, achaete scute-like 2 (Ascl2), and telomerase reverse transcriptase (Tert). Administration of the herb pair significantly restored the expression levels of these genes (Figure 4C). Furthermore, we analyzed the relative expression of ISCs differentiation-related genes in colon tissues, and the herb pair increased the levels of genes associated with secretory cells, including Mucin 2, lysozyme 1, and Chromogranin A (Figure 4C). Additionally, IHC analysis showed that Lgr5, PCNA and Ki67 were increased in the herb pair group compared with the DSS group (Figure 4D-G). These results suggest that the herb pair stimulated the proliferation of ISCs, thereby promoting intestinal repair following DSS exposure.

Figure 4 Herb pair restored abnormal epithelial proliferation in colitis mice. A and B: Western blotting analysis of protein expression of leucine-rich repeat-containing G-protein coupled receptor 5 (Lgr5) in colon tissues; C: Quantitative polymerase chain reaction analysis of mRNA expression of Lgr5, SRY-box transcription factor 9, homeodomain-only protein homeobox, achaete scute-like 2, telomerase reverse transcriptase, Mucin 2, lysozyme 1, and Chromogranin A in colon tissues; D: Representative images of immunohistochemical (IHC) of Lgr5, proliferating cell nuclear antigen, Ki67 in colon tissues; E-G: Quantitative analysis of IHC results of Lgr5, proliferating cell nuclear antigen, Ki67. For western blotting analysis: n = 3, for quantitative polymerase chain reaction analysis: n = 5-6, for IHC measurement: n = 3. ^aP < 0.05 vs dextran sulfate sodium group, ^bP < 0.01 vs dextran sulfate sodium group, ^cP < 0.001 vs dextran sulfate sodium group. Lgr5: Leucine-rich repeat-containing G-protein coupled receptor 5; Sox9: SRY-box transcription factor 9; Hopx: Homeodomain-only protein homeobox; Ascl2: Achaete scute-like 2; Tert: Telomerase reverse transcriptase; Muc2: Mucin 2; Lyz1: Lysozyme 1; Chga: Chromogranin A; DSS: Dextran sulfate sodium; PCNA: Proliferating cell nuclear antigen.

Danggui-Baishao herb pair activated Wnt/β-catenin signaling

Through RNA sequencing, we elucidated the underlying mechanisms of the herb pair in combating colitis. An overlap analysis identified 830 overlapping DEGs between the two sets of DEGs, and their expression patterns were visualized in a heatmap (Figure 5A and B). Gene set enrichment analysis and the heatmap revealed that the herb pair activated the Wnt signal (Figure 5C and D). Subsequently, we examined the mRNA expression of Wnt signal-related genes. We found that DSS reduced the relative expression of Wnt5b, Wnt9a, and Wnt10a, while the herb pair upregulated the expression of porcupine, Wnt5b, Wnt9a, and Wnt10a (Figure 5E). Furthermore, the relative expression of Wnt4 and Wnt6 showed no statistical difference, but the herb pair exhibited a tendency to increase their expression (Figure 5E). Western blot analysis of colon tissues demonstrated that the herb pair activated the Wnt/β-catenin signaling pathway, which was manifested as the upregulation of β-catenin, p-GSK-3β, and Cyclin D1 (Figure 5F-I). Furthermore, IHC analysis showed that β-catenin was elevated following administration of the herb pair (Figure 5J and K). In summary, these data initially confirm that the herb pair activated the Wnt/β-catenin signaling pathway, thereby promoting repair of the intestinal barrier.

Figure 5 Herb pair promoted the Wnt/β-catenin signaling activation in colitis mice. A: The overlap analysis visualized in the Venn diagram; B: The overlap analysis visualized in the heatmap; C: Gene set enrichment analysis; D: The overlap analysis visualized in the heatmap of the Wnt signal; E: Quantitative polymerase chain reaction analysis of mRNA expression of porcupine, Wnt5b, Wnt6, Wnt9a, Wnt10a, and Wnt4 in colon tissues; F-I: Western blotting analysis of protein expression of β-catenin, p- glycogen synthase kinase-3β, glycogen synthase kinase-3β, and Cyclin D1 in colon tissues; J: Representative images of immunohistochemical (IHC) of β-catenin in colon tissues; K: Quantitative analysis of IHC results. For RNA sequencing and western blotting analysis: n = 3, for quantitative polymerase chain reaction analysis: n = 5-6, for IHC measurement: n = 3. ^aP < 0.05 vs dextran sulfate sodium group, ^bP < 0.01 vs dextran sulfate sodium group, ^cP < 0.001 vs dextran sulfate sodium group. DSS: Dextran sulfate sodium; GSK-3β: Glycogen synthase kinase-3β; Porcn: Porcupine.

Danggui-Baishao herb pair influenced the microbiota and BA metabolism

The gene set enrichment analysis based on colonic RNA-seq revealed that the herb pair activated the FXR pathway (Figure 6A). Western blot analysis demonstrated that the herb pair increased the expression of FXR (Figure 6B and C). The qPCR indicated that the relative expression of nuclear receptor subfamily 1 group H member 4, fibroblast growth factor 15, and nuclear receptor subfamily 0 group B member 2 were increased in the herb pair group (Figure 6D). To elucidate the metabolomic phenotypes, we employed targeted metabolomics to determine the BAs in colon contents. Compared to the DSS group, the herb pair did not significantly alter primary BAs. However, there were notable changes in secondary BAs, including tauroursodeoxycholic acid and 7-ketolithocholic acid (7-keto-LCA) (Figure 6E and F). To ascertain whether the therapeutic effects of the herb pair involved regulating the microbiota, 16S rRNA sequencing and metagenomic sequencing were used to analyze the microbiota. To further understand the composition of the gut microbiota, we visualized the relative abundance of the top 10 bacterial genera (Figure 6G). In addition, linear discriminant analysis coupled with effect size analysis identified bacterial clades that consistently exhibited biologically and statistically significant differences from phylum to genus levels (Figure 6H and I). Metagenome sequencing was performed to obtain genome-wide information. The results showed that DSS reduced bacterial diversity, while herb pair supplementation restored bacterial diversity (Figure 7). Detection of the microbiota at the species level distinguished a consortium of bacteria that were enriched in the control, DSS, and herb pair groups (Figure 7). In summary, 16S rRNA and metagenomic analysis demonstrated that the herb pair restored the gut microbiota.

Figure 6 Herb pair altered the composition of bile acids metabolism and gut microbiota in colitis mice. A: Gene set enrichment analysis; B and C: Western blotting analysis of protein expression of farnesoid X receptor in colon tissues; D: Quantitative polymerase chain reaction analysis of mRNA expression of nuclear receptor subfamily 1 group H member 4, fibroblast growth factor 15, and nuclear receptor subfamily 0 group B member 2 in colon tissues; E and F: Influenced bile acids in colitis mice colon contents; G: Abundance of differential bacteria in colitis mice colon contents; H: Taxonomic cladogram analyzed by linear discriminant analysis coupled with effect size; I: Histogram of the linear discriminant analysis coupled with effect size analysis. For western blotting analysis: n = 3, for quantitative polymerase chain reaction analysis: n = 6, for bile acids measurement: n = 7, for 16S rRNA sequencing: n = 5. ^aP < 0.05 vs dextran sulfate sodium group, ^bP < 0.01 vs dextran sulfate sodium group, ^cP < 0.001 vs dextran sulfate sodium group. FXR: Farnesoid X receptor; DSS: Dextran sulfate sodium; GAPDH: Glyceraldehyde 3-phosphate dehydrogenase; Nr1h4: Nuclear receptor subfamily 1 group H member 4; Fgf15: Fibroblast growth factor 15; Nr0b2: Nuclear receptor subfamily 0 group B member 2; BAs: Bile acids.

Figure 7 Herb pair ameliorated gut microbiota dysbiosis in colitis mice. A: Linear discriminant analysis coupled with effect size analysis of latent Dirichlet allocation score histogram in control mice and colitis mice; B: Heatmap of the gut microbiota in control mice and colitis mice; C: Linear discriminant analysis coupled with effect size analysis of latent Dirichlet allocation score histogram in colitis mice with normal saline or herb pair; D: Heatmap of the gut microbiota in colitis mice with normal saline or herb pair. For Metagenomic sequencing: n = 5. DSS: Dextran sulfate sodium.

β-catenin inhibitor abolished the therapeutic effects of the Danggui-Baishao herb pair

To understand the potential biological mechanism of the herb pair in alleviating intestinal epithelial barrier damage, the β-catenin inhibitor ICG-001 was utilized. The experimental results demonstrated that ICG-001 significantly diminished the therapeutic action of the herb pair in the colitis model, this was primarily manifested as colon length and histological scores that were not improved (Figure 8A-D). Furthermore, ICG-001 was demonstrated to inhibit the relative mRNA expression of several genes involved in epithelial differentiation, including Lgr5, Sox9, Ascl2, Tert, and Mucin 2. ICG-001 administration resulted in elimination of the relative mRNA expression of Occludin (Figure 8E and F). Additionally, ICG-001 treatment eliminated the function of the upregulated relative protein expression of Occludin, and E-cadherin (Figure 8G-I).

Figure 8 The β-catenin inhibitor abolished the protective effect of the herb pair on dextran sulfate sodium-induced colitis. A: Colon length changes in mice from each group; B: Representative images of the colon length; C: Representative images of hematoxylin and eosin staining of colon tissues; D: Histological scores of hematoxylin and eosin staining of colon tissues; E: Quantitative polymerase chain reaction analysis of mRNA expression of Occludin in colon tissues; F: Quantitative polymerase chain reaction analysis of mRNA expression of leucine-rich repeat-containing G-protein coupled receptor 5, SRY-box transcription factor 9, achaete scute-like 2, telomerase reverse transcriptase, and Mucin 2 in colon tissues; G-I: Western blotting analysis of protein expression of Occludin and E-cadherin in colon tissues. For colon length measurement: n = 6, for western blotting analysis: n = 3, for quantitative polymerase chain reaction analysis: n = 5-6. ^aP < 0.05 vs dextran sulfate sodium group, ^bP < 0.01 vs dextran sulfate sodium group, ^cP < 0.001 vs dextran sulfate sodium group. ns: No significance compared with the dextran sulfate sodium + ICG-001 group. DSS: Dextran sulfate sodium; GAPDH: Glyceraldehyde 3-phosphate dehydrogenase; Lgr5: Leucine-rich repeat-containing G-protein coupled receptor 5; Sox9: SRY-box transcription factor 9; Ascl2: Achaete scute-like 2; Tert: Telomerase reverse transcriptase; Muc2: Mucin 2.

Benzoylpaeoniflorin in Danggui-Baishao herb pair protected DSS-induced injury by activating Wnt/β-catenin signaling in vitro

We performed molecular docking to evaluate the potential binding of the main compounds to the β-catenin protein. A potential binding pocket was obtained from a previous study which identified notoginsenoside r1 as an agonist of the Wnt/β-catenin pathway and which bound to β-catenin protein with relatively low binding energies of -7.29 kcal/mol[21]. The present study showed that benzoylpaeoniflorin, galloylpaeoniflorin, benzoylalbiflorin, mudanpioside E, paeoniflorin, catechin, albiflorin, and ferulic acid can bind to the β-catenin protein with relatively low binding energies of -8, -8, -7.8, -7.3, -6.8, -6.6, -6.5 and -5.1 kcal/mol, respectively. In addition, benzoylpaeoniflorin interacted with amino acid residues, including GLU462, GLU458, HIS503, LEU506, PRO505, and LYS508 (Figure 9A). We further assessed the effects of benzoylpaeoniflorin on the β-catenin signaling pathway using the 2% DSS-induced in vitro model. Compared with the DSS group, the middle (50 μM) and high dose (100 μM) of benzoylpaeoniflorin significantly upregulated the expression of Claudin1 and Cyclin D1, but, did not upregulated the expression of β-catenin in vitro (Figure 9B-F). In the presence of ICG-001, the DSS + ICG-001 group and the DSS + ICG-001 + benzoylpaeoniflorin group were observed no significant differences in the expression of Claudin1 and Lgr5 (Figure 9G-I). These findings confirm that the therapeutic effects of benzoylpaeoniflorin in colitis are associated with activating the Wnt/β-catenin pathway.

Figure 9 Benzoylpaeoniflorin protected against dextran sulfate sodium-induced injury through activating β-catenin in vitro. A: Molecular docking of benzoylpaeoniflorin and β-catenin protein and their binding sites; B and C: Western blotting analysis of protein expression of Claudin1; D-F: Western blotting analysis of protein expression of β-catenin and Cyclin D1; G-I: Western blotting analysis of protein expression of Claudin1 and leucine-rich repeat-containing G-protein coupled receptor 5. For all the experiments: n = 4. ^aP < 0.05 vs dextran sulfate sodium group, ^cP < 0.001 vs dextran sulfate sodium group. ns: No significance compared with dextran sulfate sodium + ICG-001 group; DSS: Dextran sulfate sodium; Lgr5: Leucine-rich repeat-containing G-protein coupled receptor 5; Ben: Benzoylpaeoniflorin.

DISCUSSION

UC is a chronic non-specific intestinal inflammatory disease with diarrhea accompanied by blood or pus, abdominal pain and tenesmus as the main clinical manifestations. However, improvement of clinical symptoms does not alter disease progression. Approximately 39% to 60% of patients with a clinical response still exhibit endoscopic active status, underscoring the progressive importance of mucosal barrier healing as the primary therapeutic objective[22]. Given that ISCs can differentiate into multiple cell types, the management of dysfunctional ISCs to promote mucosal barrier healing represents an effective alternative therapy. The herb pair has been extensively utilized in TCM for over a millennium to alleviate symptoms such as diarrhea and abdominal discomfort. The majority of medical practitioners believe that this herb pair facilitates the regeneration of injured intestinal tissue and restores its functionality. We established a colitis model to investigate the efficacy of the herb pair treatment in reversing weight loss, colon shortening, alleviating histopathological damage, and reducing the inflammation in colon tissues. It was observed that the herb pair significantly enhanced the intestinal barrier, as assessed by a fluorescence assay using FITC-dextran and by measuring the expression of serum LPS and colonic junction proteins. The AB staining and AB-PAS staining showed that the herb pair enhanced the goblet cell count. Goblet cells are highly expressed and have normal secretory function in physiological conditions. They are mainly responsible for producing and secreting glycoproteins, which are the main components of the mucous layer, anti-microbial peptides are enriched in the mucosal layer[5]. Both clinical data and experimental evidence have shown a correlation between elevated intestinal tight junction permeability and adverse outcomes in active UC[23], and has been shown to generate vital therapeutic effects in relieving intestinal inflammation by reducing intestinal permeability[24].

Our study verified that the herb pair elevated the expression of Lgr5, which serves as a crucial marker gene for ISCs. Additionally, we assessed other ISCs markers, including Sox9, homeodomain-only protein homeobox, Ascl2, Tert, Chromogranin A, and lysozyme 1. The herb pair exhibited a consistent upregulation of these genes, suggesting an elevated expression of ISCs. This finding was corroborated by IHC for Lgr5, PCNA, and Ki67. The Wnt pathway is a signal transduction route that is activated by the binding of Wnt proteins to membrane protein receptors, triggering a cascade reaction from upstream to downstream and regulating various biological behaviors. During embryogenesis and adult tissue homeostasis, it is crucial for the proliferation, differentiation and renewal of stem cells[25]. The Wnt signal target genes act as intercellular signals and regulate a diverse range of cellular processes, such as stem cell maintenance and proliferation[26]. Through RNA-seq and qPCR analyses, we identified Wnt5b, Wnt9a, and Wnt10a as upregulated genes in response to the herb pair treatment. Five publicly available IBD datasets were integrated to identify a molecular subtype associated with the WNT5B⁺ cell type[27]. It has been well documented that the self-renewal activities of mesenchymal stem-like cells were stimulated by recombinant Wnt5a[28]. Signal transducer and activator of transcription 6 knockdown results in a decrease in Wnt10a and nuclear β-catenin protein levels, as well as the downregulated of Lgr5 and c-Myc mRNA expression, furthermore, this treatment delayed wound healing after 2,4,6-trinitrobenzenesulfonic acid administration[29]. Our data revealed that the herb pair treatment substantially elevated cellular levels of total β-catenin and concurrently increased the p-GSK-3β/GSK-3β ratio. This suggests a significant reduction in the non-phosphorylated form of GSK-3β, which plays an important role in the phosphorylation of β-catenin and its subsequent degradation. To elucidate the precise role of Wnt/β-catenin activation in the herb pair treatment, we administered ICG-001 by intraperitoneal injection[21]. The results showed that ICG-001 diminished the protective effects of the herb pair in terms of junction protein expression and ISCs marker gene levels. These findings indicated a Wnt signal-dependent mechanism through which the herb pair alleviated colon shortening and facilitated intestinal mucosal healing. Our findings also echo previous research, where berberine and notoginsenoside R1 were also highlighted as novel and effective strategies in anti-colitis by restoring mucosal barrier and regulating the Wnt/β-catenin signaling pathway[21,30].

The close connection between IBD and intestinal ecological imbalance has received extensive attention from researchers[31]. Compared with healthy controls, clinical rectal biopsies revealed that subclinical UC patients had a higher abundance of Bacteroidetes[32]. Furthermore, Bacteroides proteins, particularly serine proteases, exhibited a positive correlation with UC disease activity following an analysis of six fecal or serum-based omics datasets[33]. In the current study, we observed a significant enrichment of the Bacteroides phylum in the DSS-treated group, as determined by 16S sequencing analysis. This finding suggests that the herb pair treatment may decrease Bacteroides levels. Furthermore, we identified Bacteroides thetaiotaomicron and Bacteroides uniformis as decreased in the treatment group using metagenomic sequencing of stool samples. Bacteroides species is a vulgaris gram-negative bacteria that maintains a complex symbiotic relationship with its host. Although some researchers have proposed that Bacteroides has the potential to alleviate intestinal inflammation[34], conflicting fecal and metagenomics data suggest that Bacteroides strains play intricate roles in maintaining intestinal health. Bacteroides uniformis and Bacteroides thetaiotaomicron are gram-negative bacteria that possess multiple genes encoding BA-altering enzymes[35,36]. The administration of Bacteroides uniformis resulted in a substantial increase of 7-keto-LCA in the colon of mice treated with DSS[37]. Our study revealed a significant reduction in 7-keto-LCA levels following the herb pair treatment. Furthermore, colonic RNA-seq data provided unequivocal evidence that the herb pair treatment activated the colonic BA nuclear receptor FXR. This activation was subsequently validated by western blot analysis, confirming the protein levels of FXR. Although in vivo experiments suggested that 7-keto-LCA protected against aspirin-induced intestinal injury, our experiment revealed that 7-keto-LCA appeared to exacerbate intestinal mucosal injury by inhibiting FXR. Consequently, further research is warranted to elucidate these aspects. In vitro experiments demonstrated that the activation of FXR by chenodeoxycholic acid, deoxycholic acid, and GW4064 resulted in a dose-dependent increase in the secretion of Wnt ligands[12]. There are also some opposing opinions suggesting that FXR inhibition induces Wnt/β-catenin pathway activation, although most of these results originate from studies on colonic carcinogenesis[38,39]. As per our unpublished data, FXR knockdown mice exhibited a higher degree of colitis symptoms, a reduced body weight, and a shortened colonic length induced by DSS.

Our study confirmed that the herb pair treatment can effectively modulate the structure of the gut microbiota, promote the proliferation and activity of ISCs and activate the Wnt/β-catenin pathway. A key question is whether the gut microbiota mediates the pharmacological effects of the herb pair treatment. To investigate the microbiota-dependent effects, we recently designed an experiment where mice treated with an antibiotic cocktail were concurrently administered the herb pair. These unpublished data indicated that while depleting the microbiota did not affect colon length by itself, it at least partly abrogated the herb pair’s efficacy in ameliorating body weight loss, colon shortening, and histological damage. Total glucosides of peony extracted from Paeonia lactiflora, showed a microbiota-dependent effect against colitis[40]. Angelica sinensis oil significantly downregulated the abundance of pathogenic bacteria (Bacteroidetes, Proteobacteria, and Desulfobacteriaceae) while increasing the abundance of beneficial bacteria (Firmicutes, Blautia, Akkermansia, and Lachnospiraceae) in a colitis model[41]. Our data indicated that the herb pair treatment activated the Wnt/β-catenin pathway partly by modulating the microbiota. It is known that certain gut microbiota can influence the Wnt/β-catenin pathway and the niche of ISCs through their metabolites. We found that the herb pair remodeled the gut microbiota and reduced BAs and therefore activated the Wnt/β-catenin pathway. This potential action intimately links the processes of microbial modulation and mucosal repair. Interestingly, previous studies showed that compounds in the herb pair could directly modulate host signaling pathways and inhibited colitis. Some of the main components of the herb pair, such as paeoniflorin[42] and Angelica sinensis polysaccharide[18], have been reported to protect against colitis conditions. In addition, the early growth response factor 1 was identified as a target of ligustilide and overexpression of early growth response factor 1 promoted cell proliferation and the cell cycle and upregulated the expression of the Wnt/β-catenin signaling pathway[43,44]. We found that benzoylpaeoniflorin, a main compound of the herb pair, directly bound to β-catenin and activated the Wnt/β-catenin pathway in vitro. Benzoylpaeoniflorin exhibited significant anti-colitis activity, leading to a reduction in both the disease activity index and histological damage[45]. Our in vitro experiments showed that benzoylpaeoniflorin protected against DSS-induced injury and upregulated Claudin1 and the β-catenin-target gene Cyclin D1. It is well established that activated β-catenin signal leads to upregulated Claudin1 expression[46]. Therefore, we used ICG-001 to determine whether the protective effect of benzoylpaeoniflorin is dependent on β-catenin. Our previous data showed that ICG-001 treatment significantly down-regulated the expression of Cyclin D1, which proved no off-target effect in this experiment. And in the presence of ICG-001, benzoylpaeoniflorin failed to influence the expression of Claudin1 and Lgr5. These data indicated that benzoylpaeoniflorin protected against DSS-induced injury by activating the β-catenin. We hypothesized that the herb pair treated colitis through a dual mechanism: On the one hand, by modulating harmful microbiota/BAs axis, and on the other hand, by directly activating the Wnt/β-catenin pathway. The herb pair exhibited multiple effects on the Wnt/β-catenin pathway, and thereby enhancing stem cell function, inhibiting inflammation and facilitating mucosal healing. However, this hypothesis warrants further investigation, such as utilizing germ-free animal models or fecal microbiota transplantation experiments to directly verify the critical role of microbial changes, identifying the specific compounds inhibiting BAs metabolic enzymes in microbiota, and validate the key mechanisms linking the main compounds to the Wnt/β-catenin pathway and ISCs proliferation.

CONCLUSION

The herb pair effectively alleviated experimental colitis through the following mechanisms: Modulating the microbiota, facilitating the restoration of Lgr5⁺ ISCs, and safeguarding epithelial integrity. Furthermore, the beneficial impact of the herb pair on colitis was, at least in part, facilitated by activation of the Wnt/β-catenin signaling pathway (Figure 10).

Figure 10  Schematic diagram of the anti-colitis mechanism of the herb pair. Herb pair exerts its alleviates experimental colitis by the activation of the Wnt/β-catenin signaling pathway. LRP5/6: Low-density lipoprotein receptor-related protein 5/6; GSK-3β: Glycogen synthase kinase-3β; APC: Adenomatous polyposis coli; CK1α: Casein kinase 1α; TCF: T-cell factor; LEF: Lymphoid enhancer-binding factor.