BaiRui YuPingFeng Powder consists of Thesium chinense Turcz., Saposhnikovia divaricata (Turcz.) Schischk., Astragalus membranaceus (Fisch.) Bge., and Atractylodes macrocephala Koiz. in a 2:2:1:1 ratio. Thesium chinense was provided by Anhui Jiuhua Huayuan Pharmaceutical Co., Ltd., and identified as Thesium chinense Turcz. of the Santalaceae family by Anhui Science and Technology University. The additional medicinal plants were verified to satisfy the requirements of the 2020 edition of the Pharmacopoeia of the People’s Republic of China after being acquired from Anhui Xiehecheng Pharmaceutical Co., Ltd. To obtain TCYP powder, the herbs were decocted twice using distilled water at a 1:10 (g:mL) ratio. The filtrate was then lyophilized. The chemical constituents in TCYP extracts were subsequently identified and analyzed using UHPLC-QE-MS. Detailed analytical methods and corresponding results are provided in Supplementary file S1.

The Anhui Science and Technology University Animal Ethics Committee gave its approval to the experimental protocol (Approval No. AK2024045). The animals were handled strictly in line with the Guide for the Care and Use of Agricultural Animals in Research and Teaching, guaranteeing adherence to accepted experimental and ethical norms.

Three hundred one-day-old broilers with uniform body weight and an equal number of male and female were acquired from Shandong Yisheng Livestock & Poultry Breeding Co., Ltd., a commercial hatchery. Five sets of sixty chickens each, each consisting of six replicates, 10 birds per replication. Under controlled settings, with an average temperature of 26 ± 4°C and a relative humidity of 60 ± 10%, the experiment was carried out for 42 days in the animal testing area of Anhui Science and Technology University. The broilers had unlimited access to food and water during the feeding session.

The control group (CON) received a standard basal diet, whereas the antibiotic group (ATB) was administered antibiotics through both dietary supplementation and water medication (see Table 1 for specific antimicrobial agents and dosage regimens). In accordance with established research protocols, the treatment groups were allocated TCYP-supplemented diets at graded inclusion levels of 500 mg/kg (TCYP-L), 1,000 mg/kg (TCYP-M), and 1,500 mg/kg (TCYP-H) (14). All experimental diets were formulated to meet the nutritional specifications outlined in the Chinese National Standard GB/T 5916–2020, with corn-soybean meal serving as the basal formulation (complete diet composition presented in Table 2).

On day 42, one bird per replicate was randomly selected for sample collection (n = 6). Blood samples were obtained from the wing vein and centrifuged at 3,000 rpm for 15 min at 4°C to separate serum. Concurrently, the chickens were dissected to collect 1 cm tissue segments from the mid-portions of duodenum, jejunum, and ileum. Cecal content samples were immediately transferred to sterile pre-chilled tubes and stored at −80°C for subsequent 16S rRNA sequencing and metabolomics analysis.

Each replicate’s body weight of the fasting broilers was calculated before to the experiment. The following metrics were computed: initial body weight (IBW), final body weight (FBW), average daily gain (ADG), average daily feed intake (ADFI), and feed-to-gain ratio (F/G).

Blood samples were collected from the wing vein and centrifuged at 3,000 rpm for 15 min at 4°C to obtain serum. Using commercial test kits from the Nanjing Jiancheng Bioengineering Institute, serum biochemical parameters were evaluated. The following parameters were measured: total cholesterol (TC), triglycerides (TG), albumin (ALB), alanine aminotransferase (ALT) activity, high-density lipoprotein cholesterol (HDL-C), lactate dehydrogenase (LDH), and low-density lipoprotein cholesterol (LDL-C).

For intestinal morphology analysis, six chickens were randomly selected from each group. Tissue samples (1 cm) were collected from the mid-portions of the duodenum, jejunum, and ileum. The samples were fixed in 4% paraformaldehyde for 48 h, embedded in paraffin, and sectioned at approximately 5 μm thickness. Three sections were prepared for each intestinal segment. For each section, three non-overlapping fields of view were randomly selected. The microscopic images were analyzed using ImageJ software to measure villus height, crypt depth, and the villus height to crypt depth ratio (V/C) (15).

Based on comprehensive pharmacodynamic evaluations including growth performance parameters, serum biochemical indices, and intestinal HE staining analyses, the CON group, ATB group, and the most therapeutically effective TCYP dosage group were selected for 16S rRNA sequencing of cecal contents, which represent the intestinal segment with the highest microbial density in avian species.

The cecal content samples collected from broilers as described in Section 2.2 were subjected to microbial DNA extraction following the manufacturer’s protocol of the DNA extraction kit. To determine the final DNA concentration, RT-qPCR was used, and the integrity of the DNA was evaluated using agarose gel electrophoresis. The V3-V4 variable region of the 16S rRNA gene was amplified using primers 341F and 805R (341F: 5’-CCTACGGGNGGCWGCAG-3′; 805R: 5’-GACTACHVGGGTATCTAATCC-3′). The PCR amplicons were purified using the AMPure XP kit before to detection and quantification. To sequence the purified amplicons, the Illumina NovaSeq 6,000 platform was used.

Low-quality and chimeric sequences were filtered from the raw sequencing data to obtain high-quality clean data. The clean data were then denoised using DADA2 in Qiime2 to remove PCR amplification and sequencing errors, generating an amplicon sequence variant (ASV) abundance table for subsequent alpha diversity, beta diversity, species composition, and LEfSe analysis (16).

Metabolites were extracted from the contents of the chicken’s cecum, and then the extraction solvent was added, bead-beating, and ultrasonication were performed. After centrifugation, the supernatant was collected and lyophilized for examination. A Vanquish Flex UPLC system with an ACQUITY UPLC T3 column (100 mm x 2.1 mm, 1.8 μm) was used for the separation. Acetonitrile (B) and 5 mmol/L ammonium acetate-acetic acid aqueous solution (A) were used as mobile phases, and the column temperature was kept at 40°C. The flow rate was 0.35 mL/min. In both positive and negative ion modes (±3,800/3400 V), mass spectrometric detection was carried out using an Orbitrap Exploris 120 high-resolution mass spectrometer. The ion source temperature was set at 350°C, and the gas parameters were set as sheath gas 1, auxiliary gas 15, and nebulizer gas 50. After ten injections, quality control samples were examined to guarantee the accuracy of the results.

The GEO database provided the gene expression information for chicken gut health from the GSE94095 dataset. R 4.4.2 was used for data standardization, and the WGCNA R program was used for co-expression network analysis of the GSE94095 dataset (17). By choosing a suitable soft threshold, a weighted connection network was built, and dynamic tree cut analysis was used to find gene clusters. Modules exhibiting a substantial link with Intestinal health (correlation coefficient > 0.7, p < 0.05) were chosen after examining the statistical relationship between module eigengenes (MEs) and phenotypic features.

The SwissTargetPrediction¹ database was used to predict chemical targets using TCYP chemicals derived from UHPLC-QE-MS analysis. The GeneCards database was used to find targets linked to illnesses of the chicken intestine.² The intersection of pharmaceutical targets, disease-related genes, and significant genes from the WGCNA modules was calculated in order to identify the intersecting gene set. Using the STRING database,³ a protein–protein interaction network was constructed with a confidence score criterion of ≥ 0.4. Using Cytoscape 3.7.2, the network was visualized (18).

The clusterProfiler package in R was used to perform GO and KEGG pathway enrichment analyses on the intersection gene set. A p-value of less than 0.05 was established as the significance criterion. For visualization, loops of KEGG pathway and GO enrichment analyses were produced.

After extracting RNA from chicken jejunal tissues, reverse transcription was performed. RT-qPCR was then used to determine the expression levels of MMP1, PDPK1, PPARG, Bax, and Bcl-2, using β-actin as an internal reference. The 2^-ΔΔCt method was used to determine the levels of gene expression. Table 3 contains the primer sequences for the target genes.

Statistical analyses were performed using individual broilers as independent experimental units (biological replicates, n = 6 per group). The sample size was determined through a priori power analysis using G*Power software (version 3.1.9.7). Based on a one-way ANOVA model with an effect size f = 0.25, α = 0.05, and power = 0.95, the calculated actual power was 95.21%. Continuous data are presented as mean ± standard deviation. All statistical analyses were conducted using GraphPad Prism 6 and SPSS 26.0 software. Comparisons between two groups were performed using Student’s t-test, while multiple group comparisons were analyzed by one-way ANOVA. A p-value < 0.05 was considered statistically significant.

Figures 1A–E demonstrate that compared with the CON group, low-dose TCYP showed no significant effects on the FBW and ADG of broilers (p > 0.05). In contrast, the ATB group, TCYP-M group, and TCYP-H group all significantly enhanced broiler FBW and ADG (p < 0.05). ADFI was significantly impacted by the ATB and high-dose TCYP groups (p < 0.05). The F/G ratio was considerably lower in the ATB and medium-high-dose TCYP groups than in the CON group (p < 0.05). This suggests that both ATB and medium-high doses of TCYP may enhance broiler growth performance by improving feed utilization efficiency.

Figures 1F–L, compared to the CON group, the high-dose TCYP treatment significantly affected the levels and activities of T-CHO, TG, ALB, LDH, and ALT in the serum of broilers (p < 0.05). Compared to the CON group, the high-dose TCYP group showed a significant reduction in T-CHO, TG, and LDH levels, as well as ALT activity, along with a marked increase in ALB content (p < 0.05). In contrast, no significant changes were observed in the medium-and low-dose TCYP groups relative to the CON group (p > 0.05). In contrast to the CON group, a significant reduction was exclusively observed in TG concentrations within the ATB group (p < 0.05). Compared to the CON group, there were no significant differences in HDL-C and LDL-C levels across the groups (p > 0.05).

Compared to the CON group, moderate and high dosages of TCYP significantly increased the villus height in the duodenum, jejunum, and ileum of broilers (p < 0.05), as Figure 2 illustrates. Compared to the CON group, the TCYP-H group showed a considerable reduction in crypt depth in the duodenum and ileum, along with a significantly higher villus height to crypt depth ratio in all three intestinal segments (p < 0.05). Furthermore, the antibiotic-containing diet induced significant morphological alterations in the duodenum, jejunum, and ileum of broilers compared to the CON group, manifesting as decreased villus height in all three segments and a markedly lower villus height to crypt depth ratio in the jejunum and ileum (p < 0.05).

The detected OUT, Chao1, Shannon, and Simpson indices of the cecal microbiota did not differ statistically significantly between the ATB and TCYP groups and the CON group (p > 0.05), per the results of the Alpha diversity analysis in Figure 3A. When high-dose TCYP was added to the broiler diet, the cecal microbiota OUT, Chao1, Shannon, and Simpson indices were significantly higher than those in the ATB group (p < 0.05). The results indicate that supplementation with 1,500 mg/kg TCYP not only promotes growth performance in broilers but also preserves the α-diversity of gut microbiota, thereby sustaining the ecological homeostasis of cecal microbial ecosystems.

While the CON group overlapped with the other groups (p < 0.01), the TCYP group’s coordinates differed considerably from the ATB group’s (PCoA analysis, Figure 3B). 8.55% of the variation between the treatment groups was explained by principal coordinate 1 (PCoA1), and 8.35% by principal coordinate 2 (PCoA2). For the discrepancies, the NMDS analysis (Stress = 0.18), too, produced trustworthy findings (Figure 3C). The cecal microbiota composition analysis at the genus level (Figure 3D,E) showed that Faecalibacterium, Lachnospiraceae_unclassified, Ruminococcaceae_unclassified, and Firmicutes_unclassified were the dominant species. The average proportion of Faecalibacterium in the CON group was 12.08%, which increased to 17.81% in the ATB group and decreased to 11.95% in the TCYP group. The average proportion of Lachnospiraceae_unclassified in the CON group was 9.90%, which slightly increased to 10.57% in the ATB group and decreased to 5.68% in the TCYP group. The average proportion of Ruminococcaceae_unclassified in the CON group was 5.45%, which decreased to 4.87% in the ATB group and increased to 6.72% in the TCYP group. The average proportion of Firmicutes_unclassified in the CON group was 4.86%, which decreased to 3.92% in the ATB group and further decreased to 4.45% in the TCYP group.

As shown in Figure 4A, differential metabolites were collected under both positive and negative ion modes, and the PCA plot illustrates the overall distribution of the dataset. In the PCA model, QC samples tightly clustered, indicating the stability and reliability of the experimental data. Meanwhile, a separation trend was observed among the three groups of broilers in the PCA plot, particularly between the CON and TCYP groups, suggesting potential metabolic differences between groups. This separation trend was further validated by the PLS-DA model (Figure 4B). As seen by Figure 4C, the permutation test verified that the PLS-DA model was not overfitted after 200 rounds of 7-fold cross-validation.

Based on FC ≥ 1.2 or FC ≤ 1/1.2, p value < 0.05, VIP ≥ 1, differential metabolites after ATB and TCYP treatments were identified. Comparison of the TCYP and CON groups showed 322 metabolites with significant changes, including 49 metabolites upregulated in the TCYP group, such as Esculin, Lysophosphatidylglycerol O-13:1, and D-mannose 6-phosphate, and 273 metabolites downregulated in TCYP, including L-Malic acid, (Z)-13-Docosenamide, and Diacylglycerol 8:0_16:3 (Figure 4D). Comparison between the ATB and CON groups showed 196 metabolites with significant changes, including 28 metabolites upregulated in TCYP, such as Cadaverine, Voacristine, and N-Acetylputrescine, and 168 metabolites downregulated, including Lysophosphatidylcholine O-18:2, Monoacylglycerol 19:0, and Trans-urocanate (Figure 4E). Comparison between the TCYP and ATB groups revealed 257 metabolites with significant changes, including 91 metabolites upregulated in TCYP, such as Lysophosphatidylglycerol O-13:1, Monogalactosyldiacylglycerol O-16:2_24:5, and 2-Methyladipic acid, and 166 metabolites downregulated in TCYP, including Glutamic acid, N-Acetyl-L-leucine, and Octadecanamide (Figure 4F). Figure 4G displays a heatmap of the top 30 differential metabolites, with blue denoting relatively low abundance and red denoting relatively high abundance. The distinct and similar effects of the various treatments on the metabolites are revealed by the Venn diagram of differential metabolites between the TCYP, ATB, and CON groups, which displays overlapping and particular changes in metabolites (Figure 4H).

Metabolites and gut microbiota were shown to be highly associated, according to the results of Spearman correlation analysis (Figure 4I). Notably, Lachnospiraceae_unclassified showed a significant positive correlation with Gly-Leu, Fumarate, Phenylpyruvic acid, L-Aspartate, Phe Gly, D-Lysine, and D-(+)-Malic acid (r > 0.5, p < 0.05) and a significant negative correlation with Thymine (r < −0.5, p < 0.05), indicating that this microbiota may be involved in the regulation of amino acid metabolism, energy metabolism, and organic acid synthesis.

To identify key gene modules associated with chicken intestinal health, this study utilized the WGCNA algorithm based on the GSE94095 dataset from the GEO database. 16 samples from the control and healthy chicken intestine groups were used to create a gene co-expression network (Figure 5A). Following the computation of the variance of gene expression, the top 25% of genes with the largest variation were chosen for further examination. By optimizing the soft thresholding power (soft thresholding power = 14) and combining hierarchical clustering with the dynamic tree-cutting algorithm, a gene co-expression network was successfully built, identifying 17 unique gene modules (Figures 5B–D). Among these, the MEturquoise module showed a strong correlation with the chicken intestinal health phenotype (correlation coefficient = 0.99, p < 0.001), suggesting that this module may play a key role in regulating chicken intestinal health.

The compounds derived from UHPLC-QE-MS analysis were entered into the SwissTargetPrediction database for target prediction in order to further investigate the important genes implicated in the treatment of chicken intestinal health by TCYP. This resulted in 841 compound-related targets. In the meanwhile, 15,282 targets linked to intestinal health in chickens were predicted using the GeneCards database. 46 key targets were chosen by examining the interaction of medication and disease targets with important genes in the MEturquoise module. To create a network of protein–protein interactions, these intersecting targets were then entered into the STRING database. TGFB1, PPARG, and MMP9 were selected as core targets based on the Degree value (Figures 5E,F).

Additional GO and KEGG pathway analyses were performed on the 46 significant genes using the R program. The greatest levels of cellular component (CC) enrichment were found in the plasma membrane region, receptor complex, and neuron projection, according to the GO analysis findings. The main molecular functions (MF) were endopeptidase activity, protein kinase activity, and protein tyrosine kinase activity. Biological processes (BP) were closely related to the regulation of cell population proliferation, the enzyme-linked receptor protein signaling pathway, and the transmembrane receptor protein tyrosine kinase signaling pathway (Figure 5G).

KEGG pathway enrichment analysis revealed that the FoxO signaling pathway and PPAR signaling pathway might be key pathways through which TCYP influences chicken intestinal health. These results provide important clues for revealing the molecular mechanisms by which TCYP regulates chicken intestinal health (Figure 5H).

As illustrated in Figure 6, we used real-time PCR to measure the expression levels of PPARG and its downstream genes. The ATB group’s jejunum showed no appreciable alterations in gene expression when compared to the CON group. PPARG, PDPK1, and Bcl2 gene expression in the TCYP group considerably increased in a dose-dependent manner (p < 0.05) as TCYP concentration rose, but MMP1 and Bax gene expression dramatically reduced in a dose-dependent manner (p < 0.05). The results demonstrate that TCYP potentially ameliorates intestinal barrier integrity through PPARG-mediated regulation of proliferative and apoptotic processes. This mechanistic insight is consistent with the histomorphological improvements observed in TCYP-H treated broilers, substantiating the hypothesis that TCYP’s enteroprotective efficacy stems from coordinated multi-target modulation.