Therapeutic mechanism of Liangxue-Guyuan-Yishen decoction on intestinal stem cells and tight junction proteins in gastrointestinal acute radiation syndrome rats

Drug preparation

LGYD comprised Panax ginseng C.A.Mey. (Renshen), Hedysarum Multijugum Maxim. (Huangqi), Cornu Bubali (Shuiniujiao), Moutan officinalis (L.) Lindl. & Paxton(Mudanpi), Salvia miltiorrhiza Bunge (Danshen), Atractylis macrocephala (Koidz.) Hand.-Mazz. (Baizhu), Pueraria lobata (Willd.) Ohwi (Gegen), Rehmannia glutinosa (Gaertn.) DC. (Dihuang), Coptis chinensis Franch. (Huanglian) and Paeonia officinalis L. (Chishao). All TCM materials were purchased from Beijing Tong Ren Tang Group (Beijing, China) and decocted twice. Both decoctions were mixed, and the decocted liquid was placed in a rotary evaporator (Hei-VAP Core, Heidolph Instruments, Schwabach, Germany) and concentrated according to the corresponding proportions. The medium-concentration decoction (MD) and high-concentration decoction (HD) treatments were concentrated to 2.73 and 5.46 g/ml crude concentrations at double and quadruple the equivalent doses administered to adult humans, respectively. They were sealed and stored at 4°C, and heated to 30–40°C before intragastric administration.

The positive control treatment was glutamine (Glu) in capsule form (Jiangsu Shenhua Pharmaceutical Co., Jiangsu, China). From d1 after irradiation to before death, the above experimental animals were respectively given medium and high concentration LGYD at a dose of 10 ml/kg; the Glu group was given 0.3 g/ml Glu suspension with 0.9% (w/v) physiological saline at a dose of 10 ml/kg; the Control group and Radiation group were given 0.9% (w/v) physiological saline at a dose of 10 ml/kg. All of the above drugs were administered once a day.

Untargeted liquid chromatography-mass spectrometry analysis

The components in LGYD were analyzed using liquid chromatography-mass spectrometry (LC–MS). Methanol, acetonitrile, formic acid, and water were purchased from Merck KGaA (Darmstadt, Germany). The 2-chlorophenylalanine was obtained from GL Biochem (Shanghai, China). The devices used were: vortex oscillator (TYXH-1; ZOLLO, Shanghai, China), high-speed benchtop refrigerated centrifuge (TGL-16MS; LUXIANGYI, Shanghai, China), high-performance liquid chromatography (HPLC) system (Acquity UPLC/HPLC; Waters, USA), HPLC column (Acquity UPLC HSS T3; Waters), mass spectrometer (Q Exactive Plus Hybrid Quadrupole-Orbitrap; Thermo Fisher Scientific, Waltham, MA, USA), and centrifuge (Sigma 1-16 K, Sigma-Aldrich, Darmstadt, Germany).

Compound Discoverer v. 2.0 (Thermo Fisher Scientific) was used to extract and preprocess LC–MS detection data. The data were retrieved using the Orbitrap TCM Library (Thermo Fisher Scientific), normalized and post-edited in Microsoft Excel 2010 (Microsoft Corp., Redmond, WA, USA), and organized into a 2D data matrix comprising the molecular weight, mass/charge ratio, retention time, peak strength, and database-matching results. A TIC diagram of LC–MS in the positive and negative ion modes is shown in Supplementary Fig. S1.

Network pharmacological analysis

The absorption, distribution, metabolism and excretion model in TCMSP database was used to screen out the components with the highest content in the samples, and the components were compared and queried. Unique molecular ID and relative target information were obtained. Duplicate components and those with no corresponding ID were removed. Ingredients with a unique molecular ID in the top 50 positive and negative ion modes were obtained. Compounds were retained only if their oral bioavailability ≥ 30% and drug-likeness ≥ 0.18 based on the criteria set by the TCMSP v. 2.3 database (https://tcmsp-e.com). A TCMSP target prediction model was used to forecast the putative target proteins of each active component. The names of the target proteins were retrieved from the UniProt database (https://www.uniprot.org). The qualifier used was ‘confirmed’ and species was ‘human’. The generic names of the genes corresponding to the target proteins were identified. The data were imported into Cytoscape v. 3.8.2 (https://cytoscape.org/download.html) to construct a drug-component-target-pathway network. The name of the disease was searched in the GeneCards (https://www.genecards.org) and Online Mendelian Inheritance in Man (OMIM) (https://www.omim.org) databases to obtain the genes associated with acute intestinal radiation injury. The online Wayne figure tools (https://bioinfogp.cnb.csic.es/tools/venny/) were used to screen for intersections among the target genes of the active ingredients in the compound and genes related to acute intestinal radiation damage. The data were downloaded from the STRING database into Cytoscape v. 3.8.2 to plot a target-path network graph. The Metascape (ttps://metascape.org/gp/index.html#/main/step1) database was used for gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses.

Animal groupings and radiation source

Sprague–Dawley (SD) rats (SPF; male; body weight: 280–300 g) were purchased from Beijing Biotechnology Co., Ltd (Beijing, China). They were randomly assigned to the following five groups: control (Control), model (Radiation), positive control drug (Glu), MD and HD. We used 53 rats for survival analysis, with 12 rats in each experimental group and 5 in the Control group. The subsequent experiment included 85 rats, with 20 rats in each experimental group and 5 in the Control group. After adaptive feeding for 3 days, the rats fasted for 8 h before irradiation. All experiments complied with the relevant experimental animal welfare principles, approved by the Ethics Committee of the Academy of Military Medical Science (No. IACUC-DWZX-2020-783). The γ-radiation source was ⁶⁰Co and was procured from the Institute of Academy of Military Medical Science. Excluding the Control group, the rats will be placed in a irradiation box and subjected to a whole-body one-time irradiation with a 100-cm source–skin distance, 300-cGy/min radiation dose rate, and 10-Gy dose in order to prepare the rat GI-ARS model. Rats with very poor survival status and abdominal aorta aneurysm were euthanized by a peritoneal overdose of pentobarbital sodium (100 mg/kg, IV). Five rats in each group were randomly selected on Days 3, 5 and 10 postirradiation. Small intestine tissues from each rat were frozen and sliced, respectively.

Western blotting

The rat small intestine tissues were homogenized at 4°C with radioimmunoprecipitation assay buffer, protease inhibitor cocktail and phosphorylase inhibitors (Servicebio, Wuhan, China) and centrifuged at 4°C and 12 000 × g for 5 min. The proteins were separated by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to a nitrocellulose membrane. The membrane was incubated with anti-β-Catenin (8480, Cell Signaling Technology, Danvers, MA, USA), anti-C-MYC (185 835, Cell Signaling Technology, Danvers, MA, USA), anti-phospho-MEK (9127S, Cell Signaling Technology, Danvers, MA, USA), anti-MEK (4694S, Cell Signaling Technology, Danvers, MA, USA), anti-phospho-ERK (4370S, Cell Signaling Technology, Danvers, MA, USA), anti-ERK (4695S, Cell Signaling Technology, Danvers, MA, USA), anti-GAPDH (5174, Cell Signaling Technology, Danvers, MA, USA), anti-WNT3A (ab219412, Abcam, UK) and anti-phospho-β-Catenin (ab75777, Abcam, UK) antibodies at 4°C overnight. The membrane was then rinsed with Tris/Tween-20 buffer (Servicebio) and incubated with horseradish peroxidase-bound secondary antibodies. The bands were analyzed using a ChemiDoc Imaging System (12003153-S; Bio-Rad, USA).

Reverse transcription quantitative polymerase chain reaction

The intestinal tissue was homogenized using RNA extraction compound (G3013; Servicebio). Reverse transcription was performed using a Servicebio RT First-Strand cDNA Synthesis Kit (G3330; Servicebio) according to the manufacturer’s instructions. Primers are listed in Supplementary Table S3. The mRNA levels were normalized to glyceraldehyde-3-phosphate dehydrogenase levels.

Hematoxylin–eosin staining and immunohistochemistry

About 3 cm of the upper segment of each small intestine was immersed in 4% (v/v) paraformaldehyde for 72 h. The small intestine were sectioned into 5-μm slices, embedded in paraffin and subjected to hematoxylin–eosin (HE) staining. The small intestine paraffin sections were rehydrated in 100%, 95% (v/v) and 80% (v/v) ethanol for 10 min per treatment and rinsed with phosphate-buffered saline (PBS). The samples were incubated with 1% (v/v) bovine serum albumin, incubated with primary antibodies at 4°C overnight and washed thrice with PBS. The secondary antibody was applied at 37°C for 30 min. The samples were rinsed thrice with PBS, rinsed with distilled water and re-stained with hematoxylin. The samples were then dehydrated and sealed. Thirty intact crypts or villi were counted per section. The numbers of positive cells per crypt or villus are reported as means ± standard deviation (SD). At least five rats were used per group. The antibodies used were anti-Lgr5 (NLS1236, Novusbio, USA), anti-CyclinD1 (GB111372, Servicebio, China), anti-PCNA (GB11010, Servicebio, China), anti-Olfm4(DF13440, Affinity Biosciences, Jiangsu, China), anti-claudin-1 (GB11032, Servicebio, China) and anti-occludin (GB111401, Servicebio, China). HE staining and immunohistochemistry (IHC) staining of the small intestine were observed under a microscope (Leika, Germany). Villus length (μm) and crypt depth (μm) were measured using ImagePro Plus v. 6.0 (Media Cybernetics Inc, Rockville, MD, USA) for ≥5 sections per time point per group. The lower right corner scale was the standard. The IHC and IF results were analyzed using ImageJ (Java 1.8.0_172; National Institutes of Health, Bethesda, MD, USA). The measurements were made by three observers blinded to the treatments.

Immunofluorescence staining

The intestinal slides were dewaxed, rehydrated and immersed in EDTA antigen recovery buffer (pH 8.0). After blocking, the slides were incubated with anti-CyclinD1 (sc8396, Santa, USA) and anti-GPR49/LGR5 (bs1117R, Bioss, China). After incubating with the above antibodies overnight and washing, the intestinal section will be incubated with CY3-labeled Goat anti-mouse IgG (GB21301, Servicebio, Wuhan, China) and Alexa Fluor 488 labeled Goat Anti-Rabbit IgG (GB25303, Servicebio, Wuhan, China). After washing, DAPI reagent was added to stain the nuclei. The images were collected after the self-fluorescence quencher was added and washed. The excitation wavelength of DAPI was 330–380 nm, and the emission wavelength was 420 nm. The excitation wavelength of 488 is 465–495 nm, and the emission wavelength is 515–555 nm. The excitation wavelength of CY3 is 510–560 nm, the emission wavelength is 590 nm, the nucleus of DAPI channel is blue, the positive channel of 488 is green and the positive channel of CY3 is red. The images were analyzed using ImageJ (Java 1.8.0_172, National Institutes of Health).

Enzyme-linked immunosorbent assay

Blood (≥3 ml) was drawn from each abdominal aorta of five randomly selected rats per group and per time point. The blood was stored at room temperature and centrifuged at 4000 × g for 15 min at 4°C. The supernatants were collected and stored at −80°C until the enzyme-linked immunosorbent assay (ELISA) was performed (Rat D-Lactate ELISA Kit; Shanghai Xin Fan Biotechnology Co., Ltd, Shanghai, China). The D-lactate standard was diluted and added to the test kit plate according to the manufacturer’s directions.

Statistical analysis

SPSS v. 19.0 (IBM Co., Armonk, NY, USA) was used for statistical analysis. Data are expressed as means ± SD. P < 0.05 indicated a statistically significant difference. One-way ANOVA was used to compare the differences among the five groups at each time point. The LSD post hoc test was used when the variances were homogeneous among the groups (In order to reduce the probability of type I error in LSD method, we recalibrated the LSD test level), and the Games-Howell test was used to compare the differences between every two groups when the variances were uneven. Survival rate was determined using a Kaplan–Meier survival curve, and pairwise comparisons were made using log-rank (Mantel–Cox) and Gehan–Breslow–Wilcoxon tests. Survival analyses and graphs were plotted using GraphPad Prism v. 7 (GraphPad Software, La Jolla, CA, USA). Rat body weights were repeated-measurement data and subjected to a correlated-sample ANOVA, which was used to analyze the data at various time points per group.

LGYD improved the short-term survival rate and general health status of rats subjected to a lethal dose of total body irradiation

LGYD significantly prolonged short-term survival in irradiated rats (Fig. 1a; Table S1). Only one rat in each of the Radiation and Glu groups survived until Day 10, indicating a high fatality rate. Zero and two rats in the MD and HD groups, respectively, survived after 2 weeks.

LGYD administration improved the general health status of the irradiated rats, except for their overall survival rates. On Days 1–3, the rats were relatively lethargic, and their food and water consumption decreased sharply. Pus and blood appeared in the feces, and some animals demonstrated bleeding from the eyelids. These symptoms were severe in the Radiation group. The general health status of HD rats was marginally better than that of rats in other groups. In all groups, the average body weight decreased significantly after irradiation and reached its lowest by Day 3; it began to rise thereafter, without returning to the initial values. It continued to increase until Days 9–10. On Day 3 postirradiation, the average body weights of HD and MD rats were significantly higher than those of Radiation rats (P < 0.01). By Day 9, the average body weight of HD rats was significantly higher than that of Radiation rats (P < 0.05; Fig. 1b; Supplementary Table S2). On Days 4–8, although diarrhea was ameliorated, the rats became lethargic and bled from their eyelids or nasal cavities.

LGYD promoted small intestinal tissue repair in rats subjected to a lethal dose of radiation

HE staining revealed that the epithelial cells of the intestinal villi of rats in the Radiation and Glu groups were enlarged and had a pale cytoplasm on Days 3, 5 and 10 postirradiation (Fig. 1d; black arrow). Fibrous tissue proliferation was observed in the submucosa with widened gaps (purple arrow). Extensive necrosis of the intestinal mucosa and diffuse inflammatory cell infiltration were observed (red arrow). After radiation, the number of intestinal glands in the mucosal layer decreased significantly (P < 0.01), and fibrous tissue proliferated in the glandular interstitium (blue arrow). Fibrous tissue proliferation was observed in the submucosa with widened gaps (purple arrow). However, in the group treated with TCM, pathological damage of the small intestine was not obvious, the epithelial cells of the intestinal villi were edematous with a light cytoplasm (Fig. 1d; black arrow) and inflammatory cell infiltration was observed in certain areas (red arrow). A few intestinal villi exhibited microvascular hyperplasia and dilation of the lamina propria (yellow arrow).

We also analyzed pathological changes at various time points after irradiation (Fig. 1e–g). On Days 3, 5 and 10 postirradiation, the number of villi, crypt depth and villus length was significantly lower in Radiation rats than in Control rats (P < 0.01). Pathological damage concerning the number of villi, crypt depth and villus length was most and least severe on Days 3 and 5, respectively. Damage to the intestinal mucosa was gradually reduced in the Radiation group by Day 10. Hence, the intestine, damaged by irradiation at the experimental dose, underwent self-repair to a certain degree; however, the tissue was not restored to its normal structure.

The impact of irradiation was less severe in Glu rats than in Radiation rats, and those in the former group gradually recovered from the damage. The recovery was better in MD and HD rats than in Radiation rats. The pathological changes that occurred in the intestinal tract of MD rats gradually reversed over time. Nevertheless, some extent of inflammatory cell infiltration and vasodilation persisted. In contrast, HD rats exhibited less inflammatory cell infiltration and no apparent tube wall thickening. Vasodilation and intestinal epithelial edema were observed in a few cases. Overall, the treatment efficacy concerning the number of villi, crypt depth and villus length was high in Glu, MD and HD groups (P < 0.05); HD rats fared best among these three groups.

Active ingredients and network analysis of LGYD

The crude drug (evaporated decoction) was subjected to LC–MS (Fig. 2a); total ion chromatograms are shown in Supplementary Fig. S2. The result of LC–MS was then analyzed by network pharmacology (Fig. 2b), and the top 100 compounds identified by the semi-quantitation of positive and negative ions were selected for preliminary screening. In total, 13 effective components were obtained (Supplementary Fig. S1).

We obtained 126 LGYD targets from the TCMSP database, and correlations between these targets and the core molecules were demonstrated using Cytoscape (Fig. 2c). The keywords ‘radiation-induced intestinal injury’ and ‘acute radiation syndrome’ were used to query disease-related genes using the GeneCards and OMIM databases. We obtained 478 targets after the elimination of duplicates and normalization of gene names, and the targets were converted into common gene names through the UniProt database. Common genes were identified using a Venn diagram (Fig. 2d).

We analyzed the intersecting genes corresponding to the target of the active LGYD component and genes contributing to radiation-induced intestinal damage. Fifty-one common genes were found to be potential LGYD targets (Fig. 2e), which were enriched by the STRING database and imported into Cytoscape for association analysis (Fig. 2f) and the Metascape database for GO analysis (Fig. 2g). GO BP showed 51 overlapping targets associated with the radiation-related biological processes cell cycle, differentiation, migration and inflammatory response and apoptosis. GO CC demonstrated that the targets focused on the cellular composition of enzymes regulating cyclin, cell membrane raft, transcription factors (TFs), nuclear envelope, basement membrane, spindle and extracellular matrix. GO MF revealed that the intersecting targets focused on TF, phosphorylase, ubiquitin-specific protease, cytokine receptor, NF-κB binding, protein kinase and nuclear receptor activation. KEGG enrichment analysis (Fig. 2h) revealed that at least 20 significant pathways were enriched, including PI3K, NF-κB, Wnt and TJ pathways.

LGYD might induce LGR5 ⁺ stem cell proliferation in the ileal crypts of rats via the Wnt pathway

The network pharmacological analyses suggested that the Wnt pathway played a role in intestinal crypt regeneration in irradiated rats. However, biological data are required to validate these findings. Results from IHC staining (Fig. 3a and b) showed that the expression of CyclinD1, a key protein in the WNT/β-Catenin pathway, was significantly lower in the intestinal crypts in Radiation rats than in Control rats (P < 0.01) on Day 3 postirradiation. CyclinD1 expression was higher in HD rats than in MD or Glu rats (P < 0.01) and nearly 3-fold higher than in Radiation rats (P < 0.01). CyclinD1 expression in MD rats was 2-fold higher than that in Radiation rats (P < 0.01). No significant difference between Radiation and Control rats regarding CyclinD1 expression on Day 5 postirradiation was observed. However, the former exhibited shorter intestinal villi and disordered intestinal epithelial morphology. The order of CyclinD1 expression was MD > HD > Glu. CyclinD1 expression was significantly lower in Radiation rats than in Control rats on Day 10 postirradiation (P < 0.01); however, it did not differ significantly from that in Glu rats. CyclinD1 expression significantly increased in MD and HD rats, although it was higher in HD rats than in MD rats (P < 0.01).

On Days 3, 5 and 10 postradiation, the number of PCNA-positive cells in Radiation rats was significantly lower than that in Control rats (P < 0.01; Fig. 4a and b). Compared with that in the Radiation group, that in all other treatment groups improved to varying degrees. In general, HD had the most significant effect, which exceeded that in the Control group on Days 3 and 5 postradiation (P < 0.01).

Lgr5 is involved in the Wnt pathway and is a pivotal ISC marker. IHC (Fig. 3c and d) showed that the number of Lgr5-positive cells in the intestinal crypts of Radiation rats was significantly lower than that in Control rats on Day 3 postirradiation. In contrast, Lgr5 expression was significantly upregulated in the intestinal crypts of MD and HD rats (P < 0.01). Lgr5-positive cells appeared mainly at the bottom of the crypt, and the cell numbers were similar in the MD and Control groups. Nevertheless, the number of cells in HD rats was <1.5-fold higher than that in Control rats. Radiation rats had approximately half as many Lgr5-positive cells as Control rats on Day 5 postirradiation (P < 0.05). Lgr5 expression in HD rats was 3-fold higher than that in Radiation rats. The number of Lgr5-positive cells was significantly lower in Radiation rats than in Control rats on Day 10 postirradiation (P < 0.01). MD and HD rats had significantly more Lgr5-positive cells than Radiation rats (P < 0.01), and the number of Lgr5-positive cells in HD rats was similar to that in Control rats (P > 0.05). Additionally, the expression levels of olfactomedin 4 (Olfm4), another protein involved in the Wnt pathway, were significantly decreased and remained at a relatively low level in intestinal tissues after radiation (Fig. 4c and d). The number of Olmf4-positive cells was increased to a certain extent in all three treatment groups but the recovery effect was different. In general, HD had the best effect (P < 0.05), and the number of Olmf4-positive cells in the short radiation period (0–5 days) was higher than that in the Control group (P < 0.05).

After radiation, the number of cells labeled with both CyclinD1 and Lgr5 increased significantly in the small intestine tissues (P < 0.01; Fig. 3e and f). All treatments showed a significantly increased number of labeled cells to different degrees (P < 0.01), and the effect of HD was the best (P < 0.01), followed by MD (P < 0.01); that of Glu was worse than that of MD (P < 0.01).

To verify the involvement of the Wnt pathway in intestinal repair, we measured the expression of key proteins involved in the classical Wnt pathway. At each time point, WNT3A, C-MYC, p-β-catenin and β-catenin expression were higher in irradiated rats than in Control rats (P < 0.01; Fig. 4e and f). The stimulatory effects of MD and HD on the Wnt pathway were stronger than those of Glu immediately after radiation. The stimulatory effect of HD was most evident on Days 5 and 10 postirradiation and was ~5-fold stronger than that in the Radiation group on Day 5 (P < 0.01). HD promoted the Wnt pathway more strongly than MD, especially on Days 5 and 10 postirradiation (P < 0.01).

The reverse transcription quantitative polymerase chain reaction (RT-qPCR) results were consistent with those of western blotting (WB) (Fig. 4g). The WNT-3A, C-MYC and β-catenin transcription levels were significantly higher in Radiation rats than in Control rats on Day 3 postirradiation (P < 0.05). Glu did not significantly promote WNT-3A, C-MYC or β-catenin transcription compared with that in the Radiation group; MD and HD were significantly more effective than Glu (P < 0.05). Moreover, HD was the most effective, especially on Days 5 and 10 postirradiation (P < 0.01).

LGYD promoted TJ protein expression in intestinal epithelial cells via MEK/ERK pathway activation

We measured D-lactic acid concentrations in the peripheral blood to reflect its intestinal barrier function. The D-lactic acid concentration in rats in Radiation, Glu, MD and HD groups was significantly higher than that in those in the Control group (P < 0.01; Fig. 5b). The D-lactic acid concentration in Glu, MD and HD groups was significantly lower than that in the Radiation group (P < 0.01); it was lowest in the HD group.

We performed IHC staining to assess the effects of LGYD on intestinal epithelial cell TJ proteins: occludin and claudin-1 (Fig. 5a, c–e). Occludin content was significantly lower in the radiation treatment groups than in the Control group (P < 0.01). Occludin expression was significantly higher in the LGYD treatment group than in the Radiation group on Day 3 postradiation (P < 0.01); it was significantly higher in MD, HD and Glu groups on Day 5 postirradiation (P < 0.05) and in HD and MD groups on Day 10 postirradiation (P < 0.05) than in the Radiation group on these respective days.

Claudin-1 expression was significantly lower in Radiation rats than in Control rats (P < 0.01), whereas it was significantly higher in MD rats than in rats in other groups (P < 0.01) on Day 3 postirradiation. However, no difference was found in claudin-1 expression between Glu and Radiation groups (P > 0.05). Claudin-1 expression in HD rats was higher than that in MD, Glu and Radiation rats on Day 5 postirradiation (P < 0.01). HD rats had the highest claudin-1 expression, which was 2-fold greater than that in Glu rats. Claudin-1 expression was significantly higher in all treatment groups than in the Radiation group on Day 10 postirradiation (P < 0.05). At that point in time, claudin-1 expression in HD rats was approximately 4-fold higher than that in Radiation rats (P < 0.01).

The expression of key proteins involved in the MEK/ERK pathway was measured (Fig. 5f and g). No differences were found in ERK expression among the groups on Days 3, 5 and 10 postirradiation. Phosphorylated ERK expression was significantly higher in HD rats than in MD, Glu and Radiation rats (P < 0.01). MEK expression was significantly higher in MD and HD rats than in Radiation rats (P < 0.05). No significant difference was found in MEK expression between MD and HD groups. The p-MEK expression level was nearly zero in Control rats and considerably higher in HD rats than in MD, Glu and Radiation rats on Days 5 and 10 (P < 0.01).