Effects of anti-inflammatory and digestive-promoting fermented Chinese herbs on cecum microbiota composition and content metabolome of Chinese chickens (Gallus domesticus)

Abstract

In this study, 168 chicks were randomly divided into two groups to elucidate the effect of fermented Chinese herbs (FCHs) on the intestinal microbiota of Chinese chickens (Gallus domesticus) and provide a scientific basis for the development and application of FCHs feed additives. The chicks were fed a basal diet for 7 days during the pre-feeding period. Thereafter, the control group was fed a basal diet, and the experimental group was fed a basal diet supplemented with 1.5% FCHs for 43 days. After being weighed and euthanised, the chickens were dissected to obtain the caecum content to analyse the intestinal microbiota composition using high-throughput sequencing of the 16S rRNA gene and metabolites using ultra-high-performance liquid chromatography high-resolution mass spectrometry. Our results showed that FCHs had no effect on the growth and serum immunological indices of Chinese chickens but significantly increased the bursa of Fabricius index, and reduced the α-diversity and affected composition of the caecum intestinal microbiota. In particular, the relative abundance of Akkermansia increased, and these of Bacteroides, Megamonas, and Faecalibacterium were reduced, accompanied by a change in the composition of caecal metabolites. These results implied that the FCH probably enhance the immune capacity of Chinese chickens. This study provides a scientific basis for developing and applying FCHs as feed additives. However, more research is necessary to elucidate the mode of action, level of inclusion, and the benefits of including the FCH in chicken diets.

HIGHLIGHTS

• Fermented Chinese herbs (FCHs) did not change chick growth and serum indices.

• FCHs changed chick caecal microbiota.

• FCHs changed chick caecal metabolites.

Keywords:

• Fermented Chinese herbs
• Gallus domesticus
• gut microbiota
• non-targeted metabolomics

Introduction

In terms of the favourable balance and high bioavailability of essential amino acids, chicken meat is a good source of high-quality dietary proteins (Song et al. 2016). Dueto the rapid expansion of the scale of intensive farming, feed hygiene is one of the challenges faced by the poultry production sector, raising the risk of intestinal diseases in chickens (Qian et al. 2018; Caekebeke et al. 2020). Intestinal diseases can limit intestinal development, cause dysfunction in intestinal digestion and absorption, and induce an imbalance in the intestinal microbiota, resulting in growth restriction, disease, and even mortality in chickens (Roberts et al. 2015; Shi et al. 2018). For many years, the subtherapeutic use of dietary antibiotic growth promoters in the food animal industry has improved feed efficiency and decreased the risk of disease (Lillehoj et al. 2018; Oh et al. 2019). However, the long-term use of antibiotics has led to the development of antibiotic-resistant microorganisms, posing a threat to consumer and animal health (Mund et al. 2017; Rizzo et al. 2023). For instance, the co-selection of resistance genes in the chicken intestinal microbiota has been reported (Ferri et al. 2017; Laconi et al. 2022). Therefore, new feed additives are needed to promote intestinal health during poultry production.

A healthy and well-functioning gut microbiota is an important factor affecting digestion and nutrient absorption and serves as a barrier against pathogen invasion (Sugiharto 2016; Kogut et al. 2017). In general, the composition of the gut microbiota is affected by multiple factors such as diet, age, genotype, sex, hatching conditions, litter, and feed management (Louis et al. 2007; Pan and Yu 2014; Zaghari et al. 2020). Diet management is particularly crucial for the function and composition of the microbiota in various sections of the intestinal tract during chicken breeding. Diary amino acids such as arginine (Arg), threonine (Thr), and glutamine (Gln) have positive effects on the immune and microbiological aspects of the intestine (Bortoluzzi et al. 2018; Bortoluzzi et al. 2020). Synbiotics also showed a beneficial effect on the intestinal microbiota, their metabolism, and the performance of broiler chickens (Śliżewska et al. 2020).

Addition of fermented Chinese herbs (FCH) to feedstuffs can supplement flavonoids, saponins, alkaloids, organic acids, and other substances that have antioxidant properties, enhance immune function, promote growth, and regulate intestinal microorganisms (Wang and Zhou 2007; Li et al. 2021; Shan et al. 2018). With a wide range of raw materials, such as plant-derived substances, Chinese herbs (CHs) have the advantage of being not creating drug-resistant animals. The antidiabetic effects of FCH on some diseases have been previously explored (Oyenihi et al. 2021; Yan et al. 2018). However, due to the complex composition of the active ingredients and multiple types of fermentation microorganisms used, the interaction between the fermentation by-products of FCH and intestinal microorganisms in chickens consuming the feed remains unknown. Considering intestinal microbiota structure is significantly affected by host food compositions (Rodriguez et al. 2012), we speculated that supplying FCH could change the intestinal microorganisms and their metabolites of Chinese chickens (Gallus domesticus). To test this hypothesis, in this study, we analysed the effects of anti-inflammatory and digestive-promoting FCHs on the intestinal microorganisms of Chinese chickens (Gallus gallusdomesticus) based on microbiome analysis and metabolomics. It provides scientific evidence for developing and applying fermented Chinese herbal feed additives.

Materials and methods

Experimental materials and preparation of fermented compound Chinese herbal feed

Leonurus japonicus, Taraxacum mongolicum, Agastache rugosa, realgar, talcum powder, Polygonum cuspidatum, Portulaca oleracea, bark of Phellodendron chinense, Epimedium brevicornu, Artemisia carvifolia, Scutellaria baicalensis, Atractylodes macrocephala, Sophora flavescens, Codonopsis pilosula, Poria cocos, Glycyrrhiza uralensis and dry leaves of Artemisia argyi were purchased from a CH market in Changsha, China. Bacillus velezensis, Lactobacillus rhamnosus, and Lactobacillus fermentum were provided by the Microecology Research Laboratory of the College of Bioscience and Biotechnology, Hunan Agricultural University. The sixteen CHs were homogenised and passed through 40 mesh sieves. The particles that passed were then mixed as follows: 15% of L. japonicus, 15% of T. mongolicum, 4.0% of A. rugosa, 5.5% of talcum powder, 5.5% of P. cuspidatum, 5% of P. oleracea, 5.5% of bark of P. chinense, 5% of E. brevicornu, 7.5% of A. carvifolia, 5% of S. baicalensis, 9% of A. macrocephala, 6.5% of S. flavescens, 3.5% of C. pilosula, 3% of P. cocos, 3% of G. uralensis, and 2% of dry leaves of A. argyi. Subsequently, tap water was added to increase the water content of the mixture to 40%. The adjusted mixture was then added to the bacterial mixture (B. velezensis (1.0 × 10⁹ CFU/mL): L. rhamnosus (1.0 × 10⁸ CFU/mL): L. fermentum (1.0 × 10⁸ CFU/mL) = 1: 1: 1) at a concentration of 10% (v/w), placed into a breathing bag, and fermented at 37 °C for 5 days. The fermentation product was dried at 60 °C and pulverised before further usage. The basal diet was a powdered compound feed based on the nutritional requirements of the 1994 National Research Council (Table S1).

Experimental design and sample collection

A total of 168 1-day-old healthy chicks with similar body weights were purchased from Longyuan Agriculture and Animal Husbandry Technology Co., Ltd. (Shaoyang, Hunan, China) and randomly divided into two groups (control and treatment). Each group had six replicates comprising 14 chicks. All chicks were fed a basal diet for 7 days during the pre-feeding period, and the experimental period lasted for 43 days. During the experimental period, the chicks in the control group were fed a basal diet, whereas those in the treatment group were fed a basal diet supplemented with 1.5% FCH. The chicks were fed ad libitum and provided with sufficient water. The temperature was maintained at 32 °C for the first 3 days after hatching and then decreased by 2 °C weekly. The final temperature was maintained at 25 °C. Chickens (1–3 days old) were exposed to light for 23 h, for 20 h at 4–7 days old, for 16 h at 7–14 days old, and for 13 h thereafter. The intensity of illumination was 10 LX for 1–7 day-old chickens and 5 LX thereafter. Chickens were immunised according to the normal immunisation procedure (i.e. the Marek’s vaccine was injected at 1 day of age, the nasal drip of the new bifurcated vaccine at 7 days of age, the drinking water contained the bursal vaccine at 14 days of age, the chicken pox vaccine was injected under the wings at 27 days of age, and the bivalent vaccine against avian influenza (H5 + H9) at 30 days of age). The chickens were kept at Feilige Technology Co., Ltd. (Shaoyang, Hunan, China).

On the last day (day 51), three chickens were randomly selected from each replicate and intravenously injected with 1 mL 3% sodium pentobarbital for deep anaesthesia. They were then weighed and blood was collected for the determination of biochemical indices. The caecum contents were dissected and placed into two 2-mL sterile frozen tubes, which were quickly frozen in liquid nitrogen for 2 h and then transferred to a freezer at −80 °C until further use in intestinal microbiota and metabolite analysis. The caecal contents from the three chickens from each replicate were mixed evenly as samples for intestinal microbiota composition and metabolite analysis. Simultaneously, the spleen and bursa of Fabricius were weighed and the relative weights of the spleen (spleen index) and bursa of Fabricius (bursa of Fabricius index) were calculated as previously described (Wu et al. 2012).

High-throughput sequencing of 16S rRNA gene amplicons of cecum microbiota

Caecum microbial DNA was extracted using a NucleoSpin 96Soil kit for DNA from soil (Macherey-Nagel, Düren, Germany). The V3-V4 region of the prokaryotic 16S rRNA gene was amplified using universal primers 338 F and 806 R, as previously described (Xu et al. 2022). Amplification products were purified using an AxyPrep DNA gel extraction kit (Axygen, Hangzhou, China) and sequenced using a HiSeq system (Illumina, USA). The paired-end raw reads were filtered using fastp 0.14.1 to remove splice and low-quality sequences and then merged into raw tags, as previously described (Xu et al. 2022). The raw tags were quality-controlled using USEARCH 10.0.2440, and low-quality tags were removed. The remaining high-quality tags were clustered into operational taxonomic units (OTUs) with 97% identity using UPARSE (Edgar 2013). The taxonomy of each OTU was assigned using RDP Classifier 2.2 (Wang et al. 2007) with the Silva 128 dataset (Quast et al. 2013). Alpha-diversity indices (Ace and Shannon indices) were calculated using Mothur v.1.30 (Schloss et al. 2009).

Caecum content Metabolome analysis

A 50 mg sample of caecum content was weighed and added to 1 mL extraction solution (methanol: acetonitrile: water volume ratio = 2: 2: 1) containing 2 mg/L internal standard (500: 1) and vortexed for 30 s. Subsequently, the samples were added to porcelain beads and ground with a grinder for 10 min at 45 Hz and then ultrasonically treated for 10 min in an ice bath. The samples were then stored at −20 °C for 1 h and centrifuged at 12000 x g and 4 °C for 15 min. Finally, 120 μL supernatant in each sample was collected and added into a 2-mL injection bottle for detection. In addition, 10 μL of supernatant from each sample was mixed with a QC sample for detection.

Metabolites were analysed using the Acquity I-Class PLUS ultra-high-performance liquid chromatography system (Waters, Tauton, MA, USA) connected to Xevo G2-XS QT of high-resolution mass spectrometer (Waters, Tauton, MA, USA) with an Acquity UPLC HSS T3 chromatographic column (1.8 μm 2.1 × 100 mm; Waters, Tauton, MA, USA). The positive ion modes were mobile phases A (0.1% aqueous formic acid solution) and B (0.1% acetonitrile formate). The negative ion modes were mobile phases A (0.1% aqueous formic acid solution) and B (0.1% acetonitrile formate). The injection volume was 1 μL.

The original data were collected using MassLynx V4.2 and processed using Progenesis QI software for peak extraction, peak alignment, and other data processing. Identification was carried out based on the METLIN database of Progenesis QI software, and theoretical fragment identification was carried out simultaneously using default parameters. The mass number deviation was < 100 ppm (Slade et al. 2015).

Data analysis

Data are presented as mean ± standard error for each group. The Wilcoxon rank-sum exact test was conducted using R 4.2.0 to detect the differences of caecum microbiota and content metabolome data among different groups. Violin plots were generated using the R ggpubr package. Principal coordinate analysis (PCoA) was performed using the R vegan package (Dixon 2003). Welch’s t-test was used to screen significantly different dominant genera and was conducted using statistical analysis of metagenomic profiles (STAMP) software. Linear discriminant analysis Effect Size (LefSe) was conducted on the Galaxy platform (http://huttenhower.sph.harvard.edu/galaxy) to identify microbial biomarkers in caecum microbiota. Spearman’s rank correlation analysis was performed for the bacterial genera to analyse the relationships among the caecum microbiota. A Spearman correlation coefficient > 0.4 and p < 0.05 was considered a significant correlation. A correlation network diagram was then drawn based on the top 30 genera using the R software. According to p ≤ 0.05 and VIP ≥ 1 calculated using OPLS-DA model, significantly different metabolites in the caecum were determined, and a volcanic map was drawn. Student’s t-test was conducted using MetaboAnalyst (https://www.metaboanalyst.ca/) to screen for significantly different metabolites and to draw heatmaps. A bubble chart was drawn using R 4.2.0 based on the Kyoto Encyclopaedia of Genes and Genomes (KEGG) pathway analysis results. Results with p ≤ 0.05 were considered significant.

Results and discussion

Effects of FCH on growth and immune indices of Chinese chickens

The initial and final weights of the Chinese chickens were not significantly different between the treatment and control groups (Wilcoxon test, p > 0.05; Table 1). Therefore, the average daily weight gain did not differ significantly between the treatment and control groups (Wilcoxon test, p = 0.556; Table 1). Moreover, the feed conversion ratio did not significantly differ between the treatment and control groups (Wilcoxon test, p = 0.556; Table 1). Furthermore, intestinal lipase, tryptase, and α-amylase activities and serum total cholesterol, IgA, and IgG levels were not significantly different between the treatment and control groups (Wilcoxon test, p > 0.05; Table 1). Although the spleen index was not significantly different between the treatment and control groups (Wilcoxon test, p = 0.841; Table 1), the bursa index of Fabricius in the treatment group was significantly higher than that in control (Wilcoxon test, p = 0.008; Table 1). Considering the important role of the bursa in bird immunity (Cui et al. 2011), the result suggested that the FCH probably enhance the immune capacity of Chinese chickens.

Table 1. Growth and immune indices of Chinese chickens fed basal diet (control) and the diet added fermented compound Chinese herbs (treatment).

Group	Control	Treatment	P
Initial weight (g)	39.116 ± 0.324	39.861 ± 0.425	0.222
Terminal weight (g)	983.036 ± 7.767	989.286 ± 7.053	0.453
Average daily weight gain (g/d)	18.877 ± 0.163	18.988 ± 0.139	0.556
Feed conversion ratio	1.908 ± 0.017	1.897 ± 0.014	0.556
Lipase activity	3.270 ± 0.543	4.411 ± 0.563	0.249
Trypsase activity	144.533 ± 10.168	146 ± 11.541	1
α-amylase activity	128.225 ± 2.764	134.085 ± 5.414	0.686
Spleen index	1.814 ± 0.103	1.916 ± 0.196	0.841
Bursa of Fabricius index	2.326 ± 0.115	3.14 ± 0.125	0.008**
Total cholesterol	0.726 ± 0.040	0.702 ± 0.031	0.834
IgA	0.319 ± 0.003	0.346 ± 0.011	0.151
IgG	10.520 ± 0.300	10.493 ± 0.548	1
CAT	0.524 ± 0.005	0.525 ± 0.011	0.691

^** p < 0.01. IgA: immunoglobulin A; IgG: immunoglobulin G; CAT: catalase.

Effects of FCH on cecum microbiota compositions of Chinese chickens

Totally, 149,250 high-quality sequences were obtained. To exclude the impact of sequencing depth on the results, 8,375 sequences were randomly resampled from each sample for subsequent analysis. Totally, 863 OTUs were obtained, of which 685 were shared between the control and treatment groups, 74 were unique to the treatment group, and 104 were unique to the control group (Figure 1A). The rarefaction curves indicated that with an increase in sequencing depth, the measured feature number of samples in the control and treatment groups gradually approached saturation (Figure 1B). This result was also confirmed by Goods’ coverage, as the Goods’ coverage of sequenced species in both the control and treatment groups was > 98% (Figure 1C). Moreover, rarefaction curves showed that biodiversity in the treatment group was significantly lower than that in the control group (Figure 1B). This result was confirmed by the feature number and ACE, Shannon, and Simpson indices (Figure 1D–G). These results indicate that the FCHM supplementation significantly reduced the alpha-diversity of the caecum microbiota in Chinese chickens (Wilcoxon test, p < 0.05).

Figure 1. Effects of fermented compound Chinese herbal medicine on caecum microbiota compositions of Chinese chickens. (A) Venn diagram shows the difference in species number between treatment and control groups; (B) Rarefaction curve; (C) Goods’ coverage; (D) Feature number; (E) ACE index; (F) Shannon index; (G) Simpson index; (H) PCoA; (I) compositions of dominant phyla in the caecum microbiota; and (J) compositions of main genera in the caecum microbiota.

PCoA based on the Bray-Curtis distance with ANOISN showed that FCH significantly altered the intestinal microbiota composition of Chinese chickens (ANOISN, R = 0.972, p < 0.01; Figure 1H).

The caecum microbiota of the Chinese chickens was mainly composed of Bacteroidetes and Firmicutes at the phylum level (Figure 1I), which was consistent with previous reports (Rychlik 2020; Artdita et al. 2021; Liu et al. 2022a). However, the average relative abundance of Verrucomicrobiota in the caecum microbiota of the treatment group was 49.33 ± 3.12%, which was significantly higher than 7.66 ± 1.83% in the control group (Wilcoxon rank-sum exact test, p < 0.05; Figure S1). The average relative abundance of Bacteroidetes in treatment caecum microbiota was 23.33 ± 1.60%, which was significantly lower than that of the control group (41.96 ± 2.74%) (Wilcoxon rank-sum exact test, p < 0.05; Figure S1). The average relative abundance of Firmicutes in the control group (44.03 ± 2.39%) was significantly higher than that in the treatment group (23.27 ± 2.22%) (Wilcoxon test, p < 0.05; Figure S1). Considering that multiple Gram-negative pathogens come from Bacteroidetes, while Firmicutes contributes more Gram-positive probiotics, these results indicate that FCH can improve the structure of the caecum microbiota. Although previously reported that the relative abundance of Verrucomicrobiota and its main representative genus Akkermansia is low in chicken intestinal microbiota (Rychlik 2020; Artdita et al. 2021; Liu et al. 2022a), our results indicated that FCH treatment enhanced the relative abundance of Verrucomicrobiota.

The similarity in the top 30 genera between the control and treatment groups was 96.6%. Simultaneously, 10 genera had relative abundances of more than 1% in the treatment group, among which Akkermansia accounted for 48.62 ± 2.90% and Bacteroides accounted for 9.19 ± 1.30%. In the control group, there were 19 genera with relative abundances of more than 1%, among which Bacteroides accounted for 25.03 ± 2.76% and Akkermansia for 6.93 ± 1.88% (Figure 1J).

Among the 37 dominant genera, the relative abundance of 16 was significantly different between the treatment and control groups (Figure 2A). The relative abundance of Akkermansia in the treatment group was significantly higher than in the control group, whereas that in the treatment group was significantly lower than in the control group (Figure 2B). LEfSe results also showed that only the relative abundance of Akkermansia in the treatment group was significantly higher than that in the control group, whereas the relative abundances of Bacteroides, Phascolarctobacterium, and Megasphaera in the treatment group were significantly lower than those in the control group (Figure 2C). In general, the reported abundance of Akkermansia in the chicken caecum is at most 5%. However, our results showed that the average relative abundance of Akkermansia in the control caecum microbiota accounted for 6.93 ± 1.88%, and supplementation with 1.5% FCH significantly increased the relative abundance of Akkermansi and resulting in it being the dominant genus. These results were consistent with the previous report that active substances in CH can increase the relative abundance of Akkermansia in the gut (Liu et al. 2022b; Bu et al. 2020; Zhao et al. 2019). Akkermansia muciniphila is widely distributed in the intestinal tract of animals and is currently reported as a potential probiotic (Cani and de Vos 2017; Zhang et al. 2021). A. muciniphila can stimulate mucin synthesis and maintain the intestinal barrier to some extent (Derrien et al. 2017; Zhang et al. 2019; Ansaldo et al. 2019), although its adverse effects on the host have also been reported (Ganesh et al. 2013). Moreover, low concentrations of A. muciniphila in the intestine indicate a thinner mucus layer, leading to a weakened intestinal barrier function (Pierre et al. 2013; Ottman et al. 2017). Therefore, our results implied that FCH probably increased intestinal barrier and innate immunity by upregulating the relative abundance of Akkermansia in the caecal microbiota.

Figure 2. Differences in dominant genera in caecal microbiota between the treatment and control groups of Chinese chickens. (A) Heatmap; (B) Extended error bar; (C) LEfSe. *p < 0.05; **p < 0.01; ***p < 0.001.

Effects of FCH on cecal metabonomics of Chinese chickens

A total of 3451 metabolites were identified. The OPLS-DA results showed that the metabolites in the treatment group were significantly different from those in the control group (Figure 3A). A total of 295 significantly differential metabolites were detected based on a volcano plot, of which 202 were significantly upregulated, and 93 were significantly downregulated in the treatment group (Figure 3B). Among 295 differential metabolites, 53 were annotated using CATEGORY (lipidomics). There were 8 sterol lipids [ST], 2sphingolipids [SP], 9polyketides [PK], 10 prenol lipids [PR], 18 fatty acids [FA], and 6glycerophospholipids [GP]. As seen in Table S2, seven fatty acids and conjugates in the fatty acyls [FA], (i.e. stearic acid, 9Z,12Z,15Z-octadecatrienoic acid, C12-2-dodecenoic acid, C12:2n-3,5, alpha-linolenic acid, 18-methyl-5Z,8Z,11Z,14Z-nonadecatetraenoic acid, and adrenic acid) were significantly downregulated in caecum metabolites from the treatment group (Table S2). The pathways enriched in significantly different metabolites caused by dietary supplementation with 1.5% FCH were mainly lipid metabolism pathways. The downregulation of fatty acyl metabolites and upregulation of taurodeoxycholic acid in caecal metabolites indicated that FCH added to the diet may promote the digestion and absorption of fatty acyls in Chinese chickens. Furthermore, six flavonoids in the polyketides (i.e. daidzin, glycitein, peonidin 3-O-glucoside, luteolin, isorhamnetin, and kaempferol) were significantly upregulated in the treatment group (Table S2). These plant flavonoids are likely unable to be digested and absorbed by the small intestine in CH, which is likely the reason for the changes in the intestinal microbiota. Moreover, higher levels of organic acids, such as ferulic and p-anisic acids, were found in the treatment group than in the control group. Among these, ferulic acid has been reported to have a strong antioxidant effect and to reduce the risk of various diseases, such as cancer and diabetes (Boz 2015).

Figure 3. OPLS-DA model (a), volcano plot (B), bubble chart (C) and heatmap (D) showing the significantly different metabolites in the Chinese chicken caecum induced by fermented compound Chinese herbal medicine. Each point in the volcano plot represents a metabolite. The x-axis represents the multiple changes of the metabolites, and the y-axis represents the P-value of the test (logarithm with the base of 10). the sizes of the scatter points represent the VIP values of the OPLS-DA model. The data had been standardised before the heatmap drawing.

MetaboAnalyst and KEGG pathway analysis showed that the metabolites that differed significantly between the treatment and control groups are involved in the α-linolenic acid metabolism, and linoleic acid metabolism were the most impacted, whereas the differences of purine metabolism, and steroid hormone biosynthesis were the maximum (Figure 3C). Heatmap combined cluster analysis of the top 30 significantly different metabolites showed that terpinen-4-ol, N,N-dimethylsphingosine, arachidonoyl ethanolamide-d4, and adrenic acid significantly decreased in the treatment group, whereas N-acetyldemethylphosphinothricin tripeptide, holyrine A, MM 42842, deisopropylhydroxyatrazine, melilotigenin, L-N2-(2-carboxyethyl)arginine, 8-demethyl-8-(methylamino)riboflavin, haematimmic acid, p-anisic acid, berberine, gibberellin A34-catabolite, PG(18:1(11Z)/19:0), elliptinol, γ-glutamyl-γ-aminobutyraldehyde, L-ala-γ-D-Glu-DAP-D-Ala, 6-demethylsterigmatocystin, 5,10-dihydrophenazine, luteolin, duartin(-), dihydrophloroglucinol, isorhamnetin, His Pro Cys, 2,3-dihydroxy-p-cumate, kaempferol, emodin, and avenalumin III were significantly increased compared with that of the control (Figure 3D).

Correlation between cecal microorganisms and metabolites

Pearson correlation analysis showed that the relative abundance of Akkermansia was significantly negatively correlated with luteolin, berberine, 2-(1-aziridinyl)ethanol, 5′-methylthioformycin, (3Z)-phycocyanobilin, 5,10-dihydrophenazine, avenalumin III, holyrine A, L-γ-glutamyl-(3R)-L-β-ethynylserine, 6-demethylsterigmatocystin, N-acetyldemethylphosphinothricin tripeptide, 2,3-dihydroxy-p-cumate, savinin, elliptinol, kaempferol, PG(18:1(11Z)/19:0), dihydrophloroglucinol, and isorhamnetin in caecal contents, whereas it was significantly positively correlated with O-ureido-L-serine and PAF C-16 (p < 0.05; Figure 4). However, the correlations between the caecal metabolites and the relative abundances of Bacteroides, Desulfovibrio, Lachnoclostridium, and Megasphaera were almost the opposite of result of Akkermansia (p < 0.05; Figure 4).

Figure 4. The bubble chart showing the correlations between caecal microorganisms and metabolites. Results with Pearson correlation coefficient > 0.6and p < 0.05 were considered significant. *p < 0.06; **p < 0.01; ***p < 0.001.

In conclusion, supplementation with 1.5% FCH significantly decreased the α-diversity and changed the structure of the caecal microbiota in Chinese chickens. FCH can improve the structure of the caecum microbiota through reduced Bacteroidetes and increased Firmicutes. It significantly increased the relative abundance of Akkermansia but reduced these of Bacteroides, Megamonas and Faecalibacterium. FCH probably increased intestinal barrier and innate immunity by upregulating the relative abundance of Akkermansia in the caecal microbiota. Simultaneously, the metabolite composition of the chicken caecum was significantly altered, while the growth performance and serum immunological indices of chickens were not significantly affected.

Ethics statement

The animal study was reviewed and approved by the Animal Ethics Committee of the Hunan Agricultural University with approval number Lunshenke-2022-No. 110.

Acknowledgements

This work was supported by the Special Fund for Innovative Province Construction in Hunan Province (2021SK2010), and the Hunan Natural Science Changde Joint Fund (2021JJ50130). We would like to thank Jiajia Ni from Guangdong Meilikang Bio-Science Ltd., China, for his help with data analysis and manuscript revision.

Data availability statement

The sequences were delivered to the NCBI SRA database with accession number PRJNA906741.

Disclosure statement

The authors declare that they have no competing interests.