Cannabis sativa L., Rehmannia glutinosa Libosch., Citrus aurantium L., and Atractylodes macrocephala Koidz. were purchased from the Jilin Beiyao Pharmaceutical Group Co., Ltd. (Jilin, China) and identified by Professor Lili Weng at the College of Pharmacy, Changchun University of Chinese Medicine. Tiger nut oil was purchased from Kunhua Biotechnology Co., Ltd. (Henan, China). Chocolate flavoring (PR-1573) was purchased from Hasegawa Flavors and Fragrances Co., Ltd. (Shanghai, China), and crystalline fructose was purchased from Shandong Xiwang Sugar Industry Co., Ltd. (Shandong, China), and fructose syrup was purchased from Shandong Xiangchi Jianyuan Biotechnology Co., Ltd. (Shandong, China). Potassium sorbate was purchased from Nantong Aokai Biotechnology Development Co., Ltd. (Jiangsu, China), and phospholipids from Anhui Yuanchuang Technology Co., Ltd. (Anhui, China), and xanthan gum from Ordos Zhongxuan Biochemical Co. Ltd. (Inner Mongolia Autonomous Region, China). Enzyme-linked immunosorbent assay (ELISA) kits for substance P (SP), gastrin (GAS), and motilin (MTL) were purchased from Jiangsu Meimian Industrial Co. Ltd. (Jiangsu, China). Loperamide hydrochloride powder (CAS number 34552–83-5, strength ≥ 98%) was purchased from Shanghai Yuanye Biotechnology Co. Ltd. (Shanghai, China). Xiangdanqing capsules (0.4 g per capsule) were purchased from Nianqing Bao Pharmaceutical Co. Ltd. (Shanxi, China).

The MRZZ formula is based on prescriptions developed by clinical doctors at the Affiliated Hospital of the Changchun University of Chinese Medicine. To prepare MRZZ, we decocted Cannabis sativa L., Rehmannia glutinosa Libosch., Citrus aurantium L., and Atractidurides macrocephala Koidz. were decocted in a ratio of 3:5:6:6 twice with a certain amount of water under normal pressure conditions, boiling at 100°C and then reducing the heat to low and starting the timer, for 1.5 h each time. The amount of water used for the two decoctions was 10 and 8 times the weight of the herbs, respectively. The decoctions were filtered, the filtrates were combined, and the resulting supernatant was collected via centrifugation at 8000 rpm for 20 min and concentrated to 1 g/mL. The concentrate was added to crystallized fructose, fructose syrup, and potassium sorbate in the aqueous phase. Then, 50% tiger nut oil, phospholipids, and chocolate flavoring were added to the mixture, which was heated until the added contents dissolved. The oil phase was shear-mixed at 8000 rpm for 2 min. The aqueous phase was added to the oil phase while continuing shear mixing at 8000 rpm, and xanthan gum was added to obtain colostrum. The colostrum was passed through a high-pressure homogenizer (400 bar with three cycles). Filling and sterilization processes yielded the final product. MRZZ is an oil-in-water (O/W) emulsion with an average particle size of 544.6 nm.

For the activated carbon solution preparation, we accurately weighed 100 g of gum arabic, added 800 mL of water, and boiled the solution until it became transparent. Next, 50 g of powdered activated carbon was added to the above solution, which was boiled three times, cooled, diluted to 1,000 mL with water, refrigerated at 4°C, and shaken well before use.

Fifty healthy six-week-old male ICR mice (specific pathogen-free; 20–22 g) were purchased from Liaoning Changsheng Biotechnology Co., Ltd. (animal certificate number SCXK[Liao]2020–0001). The mice were acclimatized and fed for 1 week under barrier environment laboratory conditions (room temperature: 23 ± 2°C, relative humidity: 50% ± 10%, 12 h light/dark cycle). Food and water were provided ad. libitum.

When acclimatization ended, the mice were randomly divided into five groups (n = 10): normal control (NC), model control (MC), positive drug (PD), low-dose (LD), and high-dose (HD) groups. The PD group was gavaged with Xiangdanqing capsules (0.208 g/kg) once daily, whereas the LD and HD groups were gavaged with the MRZZ, with herbal dosages of 0.65 and 2.6 g/kg, respectively, once daily for 14 days. During this period, the NC and MC groups were administered 0.2 mL of saline solution. On days 15–17, except for the NC group, which was administered saline solution, the other groups were administered oral loperamide hydrochloride (4 mg/kg) daily for 3 consecutive days to establish a constipation model in mice. During this period, the PD, LD, and HD groups were administered MRZZ and Xiangdanqing capsules. After fecal sampling, blood was collected via retro-orbital bleeding, and the mice were decapitated and sacrificed. Tissue specimens were then collected.

Weight changes in mice were recorded and monitored daily. On day 17 after the last administration, the mice were transferred to individual cages and fecal samples were collected within 3 h. The wet weight of the fresh feces was then recorded. All mice were fasted for 16 h without water restriction. Except for the NC group, which was administered saline, all other mice were administered 4 mg/kg of loperamide hydrochloride. After 30 min, all mice were orally administered 0.2 mL of activated carbon solution, and the time from the completion of gavage to the first defecation of the black feces was recorded. Fecal water content was calculated using the following equation: fecal water content (%) = ([wet weight − dry weight]/wet weight) × 100%.

The gastrointestinal transit rate was assessed using the method of the first black stool test in mice. Thirty minutes after the last dose, 0.2 mL of activated carbon solution was administered via gavage. After 25 min, the mice were euthanized via cervical dislocation. The intact small intestine was dissected, and the distance from the pylorus to the end of the activated carbon solution was measured as the migration distance. The following equation was used to calculate the small intestinal transit rate (%): small intestinal transit rate (%) = (transit distance of the activated carbon/total length of the small intestine) × 100%.

Following the method described by Li et al. (2022), mice were decapitated and sacrificed, and the small intestine was dissected from the cecum to the anterior end of the anus and fixed with 4% paraformaldehyde. All samples were embedded in paraffin and sectioned to a thickness of 5 μm. The sections were deparaffinized in xylene, rehydrated sequentially with alcohol and distilled water, and stained with H&E. A microscope (Eclipse CI, Nikon Corporation, Japan) was used to observe the histological differences in the colon.

Following the method described by Li et al. (2022), paraffin sections were deparaffinized, rehydrated in xylene, and antigenically repaired using a citric acid antigen repair buffer (pH = 6). After incubation pretreatment with 10% goat serum, zonula occludens-1 (ZO-1) and occludin (prepared at a dilution ratio of 1:200 in phosphate-buffered saline [PBS; pH = 7.4]) were added dropwise, and the sections were incubated at 4°C overnight. After washing with PBS, the sections were immersed in a solution containing a secondary Alexa 488-conjugated goat anti-mouse antibody and incubated at room temperature for 50 min. The nuclei were re-stained with drops of 4′,6-diamidino-2-phenylindole (DAPI) staining solution before sealing the slices with a fluorescence-quenching sealer. Sections were observed under a fluorescence microscope (Eclipse CI, Nikon Corporation, Japan), and images were captured at 400 × magnification to observe inflammatory infiltration and tissue damage.

Blood was collected via retro-orbital bleeding of mice and left for 15–30 min before centrifuging at 3000 rpm for 15 min at 4°C. The supernatant (serum) was separated and frozen at −80°C for further biochemical analysis. SP, MTL, and GAS levels in the serum were detected using commercial ELISA kits according to the manufacturer’s instructions. Absorbance was measured at 450 nm using an enzyme marker (SER33270-1236, Molecular Devices, Shanghai, China).

The SCFA content in mice feces was determined using gas chromatography–mass spectrometry (GC–MS; Huang et al., 2020). Briefly, 50 mg of lyophilized feces was mixed with 500 μL of water, homogenized, and centrifuged (12,000 rpm and 4°C for 10 min). Then, 200 μL of the resulting supernatant was mixed with 100 μL of 15% phosphoric acid, 20 μL of internal standard solution (4-methylpentanoic acid, 375 μg/mL), and 280 μL ether. The mixture was then centrifuged at 4°C and 12,000 rpm for 10 min, and the supernatant was collected. The supernatant was analyzed using a GC system (Thermo Trace 1,300, Thermo Fisher Scientific, United States) equipped with an Agilent HP-INNOWAX capillary column (30 m × 0.25 mm inside diameter × 0.25 μm) and a Thermo ISQ 7000 mass spectrometer (Thermo Fisher Scientific, United States).

The samples were shipped on dry ice to Personalbio (Shanghai, China), as described by Cockrum et al. (2020). Polyadenylated mRNA was enriched from total RNA using oligo (dT) magnetic beads, and the mRNA was fragmented to approximately 300 base pairs (bp) in length using ionic interruption. The first strand of complementary DNA was synthesized using RNA as a template, and the second strand was synthesized using a 6-base random primer and reverse transcriptase. After library construction, polymerase chain reaction (PCR) amplification was performed to enrich the library fragments, and fragments of approximately 450 bp were selected. Libraries containing different index sequences were then mixed proportionally according to the effective concentration of the libraries and the amount of data required. After mixing, they were uniformly diluted to 2 nM, and single-stranded libraries were generated through base denaturation. After purification, the libraries were subjected to paired-end next-generation sequencing using the Illumina sequencing platform.

To validate the transcriptome sequencing data, five genes were selected for further analysis based on previously described methods (Cockrum et al., 2020). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a reference gene, and total RNA was extracted from the colon tissues using Invitrogen TRIzol reagent (15,596,018, Thermo Fisher Scientific, United States). The PrimeScript™ 1^st stand cDNA Synthesis Kit (TaKaRa Biotechnology Co., Ltd., Japan) was used to reverse transcribe the total RNA that had been tested and quantified into cDNA. Corresponding fluorescent quantitative PCR experiments were performed using the AceQ® qPCR SYBR® Green Master Kit (Vazyme Biotech Co., Ltd., Nanjing, China). For each target gene, the cDNA template of the sample was selected for the PCR reaction, which comprised an initial denaturation step of 5 min at 95°C, followed by 40 cycles of denaturation at 95°C for 15 s and annealing at 60°C for 30 s. Subsequent products were tested online using a real-time fluorescence quantifier (LightCycler480II, 384, Roche, Switzerland).

Following the method described by Zhang et al. (2024), nucleic acids were extracted using the MagBeads FastDNA Kit for Soil (116,564,384; MP Biomedicals, CA, United States) and subjected to 0.8% agarose gel electrophoresis to determine the molecular size (DYY-6C, Beijing Liuyi Biotechnology Co., Ltd., Beijing, China), DNA was quantified using a nanodrop device. Bacteria with amplicon lengths of approximately 480 bp (V3V4 region of the 16S rRNA gene) were PCR-amplified and sequenced using the Illumina platform (Illumina, San Diego, CA, United States) and specific primers 338F (5′-barcode+ACTCCTACGGGGAGGCAGCA-3′) and 806R (5’-GGACTACHVGGGTWTCTAAT-3′). Data processing (i.e., quality filtering, denoising, splicing, and chimera removal) was performed on the sequences using the DADA2 plug-in, and the data were analyzed using the GenesCloud platform.¹

Statistical analyses were performed using Prism 10.4 (GraphPad Software, San Diego, CA, United States). One-way analysis of variance and Scheffe’s multiple-comparisons technique were used to assess the differences between groups. Statistical significance was set at p < 0.05. The correlation between the gut microbiota and the transcriptome was constructed using the Spearman algorithm to create a correlation matrix. The random matrix theory was used to determine the filtering threshold for the correlation values. The igraph package was then used to construct an association network, which was visualized using the R language and ggraph data package.

The body weights of the mice were recorded daily during the feeding period. The effects of MRZZ consumption on the general physiological status and fecal characteristics of the mice in each group were evaluated by monitoring body weight gain, fecal water content, time to the first black stool, and gastrointestinal transit rate. The MC group showed lower fecal water content, longer time to the first black stool passage, and lower small intestinal motility rate than the NC group (p < 0.01 in each case), suggesting that mice modeling of constipation was successfully generated (Figures 1A–D). After treatment, the differences between the PD, HD, and LD groups were significant compared to those of the MC group (p < 0.05), demonstrating that MRZZ significantly improved all indicators and alleviated constipation.

To further investigate the efficacy of MRZZ in alleviating constipation, histological staining was performed to detect changes in intestinal morphology. The intestinal tissue structure of the NC group mice was normal, whereas that of the MC group mice was markedly abnormal (Figure 2A). Disruption of intestinal villi in the field of view, extensive desquamation of epithelial cells in the mucosal layer, and a small number of inflammatory cells infiltrated the tissue. The intestinal villi structure was clear in the PD, HD, and LD groups, with neat and tight arrangements of epithelial cells in the mucosal layer, complete crypt structures, and no edema in the submucosal layer; their intestinal tissue structure was normal compared with that in the NC group.

After immunofluorescence staining, ZO-1 and Occludin appeared red, and the nuclei appeared blue after DAPI staining (Figures 2B,C). In the MC group, the fluorescence intensities of ZO-1 and Occludin were reduced, and obvious blue nuclei were visible in the DAPI plots. Compared with that in the MC group, the fluorescence intensity was increased in the PD and HD groups, and the mean density value of ZO-1 was significantly higher (p < 0.01; Figures 2D,E).

To verify the effects of MRZZ on the neurotransmitter content of mice with constipation, serum GAS, MTL, and SP levels were detected using ELISAs. The serum levels of GAS, MTL, and SP were significantly lower in the MC group than in the NC group (p < 0.05), and these indices increased after treatment with low and high doses of MRZZ and PD treatment (p < 0.05; Figures 3A–C). These data suggest that MRZZ significantly affects gastrointestinal regulatory peptide secretion in our mice model of loperamide hydrochloride-induced constipation.

SCFAs are mainly produced as metabolites by the gut microbiota and are closely related to the species and abundance of gut microbiota (Gao et al., 2021). Therefore, we used GC–MS to determine the SCFA content in mice feces. The fecal contents of acetic acid, propionic acid, butyric acid, valeric acid, and isobutyric acid in constipated mice were significantly decreased by loperamide hydrochloride treatment compared with those in the NC group (p < 0.05; Figures 3D–H). The acetic and isobutyric acid levels in the HD and LD groups were significantly higher than those in the MC group (p < 0.05). These results suggest that MRZZ can help maintain normal SCFA levels and restore intestinal function stability.

Differences in colonic gene expression patterns between the two groups (NC vs. MC and MC vs. HD) were investigated using transcriptome analysis. Principal component analysis combined with correlation analysis between the samples showed some reproducibility within the NC, MC, and HD groups, as well as significant differences between groups (Figure 4A).

Among the top 20 differentially expressed genes in the NC and MC groups, genes such as Slc15a2, Cyp1a1, Sprr2a1, and Sstr1 in the MC group exhibited low expression, whereas genes such as Egr1, Hspa1b, Jun, Hspa1a, and Dusp1 exhibited high expression (Figure 4B). Among the differentially expressed genes between the MC and HD groups, genes such as Hspa1a, Dusp1, JunD, and Gsdma2 in the HD group showed low expression, whereas genes such as Cyp1a1, Slc15a2, Ccr4, and Slfn1 showed high expression (Figure 4C). Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis revealed that these differentially expressed genes were mainly enriched in pathways such as the NOD-like receptor signaling pathway, metabolism of xenobiotics by cytochrome P450, IL-17 signaling pathway, and estrogen signaling pathway (Figures 4D,E).

To verify the gene expression changes detected by RNA sequencing, five genes (Hspa1b, Fos, C1s1, Defa29, and Fn1) were randomly selected for RT-qPCR analysis (Figure 4F). The results were similar to those obtained by RNA sequencing. We hypothesized that MRZZ alleviates constipation in mice by regulating the expression of differentially expressed genes. Related pathways may play a key role in alleviating loperamide hydrochloride-induced constipation.

The gut microbiota is crucial for the human body, helping to protect the enteric nervous system and restore intestinal function (Li et al., 2023). We analyzed 25 fecal samples from 5 groups of mice using 16S rRNA sequencing. 5,866, 1920, 3,723, 4,819, and 4,853 operational taxonomic units (OTUs) were identified in the NC, MC, PD, HD, and LD groups, respectively, with 324 common OTUs identified (Figure 5A). The HD and LD groups showed more OTUs than the MC group, indicating that MRZZ increased the number of species within the treated groups.

The samples were used to determine the α-diversity index based on OTUs with ≥ 97% sequence identity. The α-diversity sparse curve shown in Figure 5B indicates that the curve trend tended to flatten with the increasing depth of the stretching level. The species diversities of the LD and HD of MRZZ were the closest to that of the NC group, followed by those of the PD group and the MC group, suggesting that MRZZ increased the diversity of the bacterial flora in the intestinal tract of mice. Diversity was quantified using the Chao1, Simpson, Shannon, and Observed_species indices to evaluate the diversity and richness of the gut microbiota in each mice group (Figure 5C). The species richness and diversity indices performed well, with the LD and HD groups performing well; the Shannon indices differed significantly between the groups (p < 0.05). Although the Chao1 and Observed_species indices did not reveal significant differences between groups (p > 0.05), the index values of the MC group were lower than those of the NC group, and the MRZZ intervention increased the species abundance and diversity.

β-diversity was targeted for comparative analysis of microbiota among the different groups; therefore, the samples were analyzed via principal coordinates analysis (PCoA) and non-metric multidimensional scaling (NMDS) analysis, based on OTU species richness information. PCoA (Figure 5D) and NMDS (Figure 5E) analyses showed relative separation between the different groups, indicating that all five groups had distinct floral structures. In addition, the sampling points of the HD group were closer to those of the NC group, suggesting that MRZZ treatment may have reduced community structure differences. These findings suggest that MRZZ improves the species diversity of the gut microbiota in constipated mice, potentially by regulating the gut microbiota and improving the intestinal microecological environment, thereby alleviating constipation.

The gut microbiota was tested in different groups, and the taxonomic characterization of the bacterial community was performed at different taxonomic levels (Figure 6). At the phylum level, the dominant phyla in all groups were Bacteroidota and Firmicutes, followed by Actinobacteriota, Proteobacteria, and others (Figure 6A). Compared with the NC group, the MC group had higher relative abundances of Bacteroidetes, Actinobacteria, and Proteobacteria and a decreased proportion of Firmicutes. Previous findings have shown that the proportion of Bacteroidota is higher than normal in patients with constipation, whereas that of Firmicutes is lower (Guo et al., 2020), which is consistent with the present results. In contrast, MRZZ treatment reversed this trend in the mice model, restoring the abundance of Bacteroidota and Firmicutes to levels similar to those in the NC group.

At the family level (Figure 6B), the HD and LD groups showed lower relative abundances of Muribaculaceae and Lactobacillaceae and higher proportions of Lachnospiraceae than did the MC group. Relative abundances (Figure 6C) and thermograms (Figure 6D) of the top 20 microbiota were determined at the genus level to further assess their distribution. MRZZ consumption increased the abundance of Duncaniella, Bacteroides_H, Alistipes_A, UBA3263, and Parabacteroides_B and decreased that of CAG-873, Lactobacillus, CAG-485, Paramuribaculum, Prevotella, and Muribaculum. These results indicate that MRZZ restored the composition and structure of the gut microbiota at different levels.

Differences in community composition across groups were compared using the linear discriminant analysis (LDA) effect size, and species with thresholds of > 2 were represented using taxonomic-branching diagrams (Figure 6E) and LDA histograms (Figure 6F). The NC, MC, and PD groups each had 1 taxon with a significant difference in abundance, whereas the HD group had 15 taxa with significant differences in abundance. In the NC group, g_Butyribacter was the most abundant bacterial genus, whereas p_Chloroflexota was more abundant in the MC group. In the PD group, the abundance of g_Zag111 was significantly elevated. In the HD group, the abundances of p_Cyanobacteria, c_Vampirovibrionia, and o_Gastranaerophilales were significantly higher than those in the other groups. Taxonomic-branching diagrams revealed different degrees of abundance changes in the gut microbiota of each group.

Correlations between the gut microbiota and transcriptomics were assessed using Spearman’s correlation analysis (Figure 7). The abundance of Ligilactobacillus was significantly positively correlated with differentially expressed genes such as Ccn1, Dusp1, Jund, Hbb-bt, Apold1, D630003M21Rik, Hspala, Cys1, Tbx4, Sema6c, Tulp1, and Sost, and negatively correlated with Klra5. The abundance of Lactobacillus was strongly positively correlated with Sstr1 and negatively correlated with Ccr4 and Ap4b1 expression. The abundance of Bacteroides was significantly negatively correlated with Rnf39, Sstr1, and Gsdma2 and was significantly positively correlated with Gm10591, Ccr4, Ttll6, and Ndnf.