Totally, 640 1-day-old AA broilers (Shu-ya Poultry Co., Ltd., Tieling, China) were randomly classified into four experimental groups, and each group included 160 birds (8 replicates of 20 birds). Birds were raised in stainless steel cages (400 mm × 450 mm × 1,500 mm) in a controlled room for 42 days. This study was performed at Poultry Research Farm, Jinzhou Medical University, Liaoning, China. The temperature of room was gradually reduced by 3.0–3.5°C weekly until achieving a thermo-neutral zone ranged from 21 to 26°C by the end of the 3rd week. The experimental diets were based on corn and soybean meal. Four dietary regimes were provided as follows: control group (basal diet, CON group), Lac-L and Lac-H groups (basal diet supplemented with 50 and 100 mg/kg Lactococcus G423, respectively), and ABX group (basal diet supplemented with 50 mg/kg narasin). The basal diet was divided into two phases: the starter phase from 1 to 21 days and the growth phase from 21 to 42 days. The basal diet was formulated to meet the nutritional requirements according to the Chinese Broiler Feeding Standards (NY/T33-2004) (Table 1). LactococcusG423 (1 × 10¹⁰ CFU/g) and narasin (purity of narasin dihydrate powder 15%, Eli Lilly and Company, Indianapolis, Indiana, United States) were mixed in basal diet (Supplementary Figure S1).

Broiler performance in terms of average daily gain (ADG), average daily feed intake (ADFI), survival rate, and feed conversion ratio (FCR) was weekly recorded, in which ADG, ADFI, and FCR were calculated and presented for 6-week experimental period. Breast muscle, thigh muscle, and abdominal fat pad (including fat surrounding the gizzard, bursa of Fabricius, cloaca, and adjacent muscles) from one bird of average BW per replicate were removed and weighed at week 6. To compensate for the differences in carcass weight, these values were expressed as a percentage of carcass weight.

Content of high-density lipoprotein (HDL), LDL, TG, and TC was determined using enzyme-labeled instrument according to ELISA kit instruction (Nanjing Jiancheng Bio. Institute, Nanjing, Jiangsu, China).

Eight ileal samples per group were randomly selected for the analysis of intestinal flora. The polymerase chain reaction (PCR) amplification of the hypervariable region V3–V4 of the 16S rRNA gene was performed with the universal primers set338 F (5′-ACTCCTACGGAGGCAGCAG-3′) and 806R (5′-GGACTACHVGGGTWTC TAAT-3′) (Liu et al., 2016). The quality and concentration of DNA were determined by 1.0% agarose gel electrophoresis and a NanoDrop^® ND-2000 spectrophotometer (Thermo Fisher Scientific Inc., Waltham, MA, United States) and kept at −80°C for further experiment. All samples were amplified in triplicate. The PCR products were extracted from 2% agarose gel and were purified using the AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA, United States), according to the manufacturer’s instructions and were quantified using Quantus™ Fluorometer (Promega, Madison, WI, United States). The Illumina MiSeq platform (Illumina Inc., San Diego, CA, United States) was used for paired-end sequencing (2 × 300) of the PCR products. The raw sequence reads were quality-filtered and merged by FLASH (Tanja and Steven, 2011) before open-reference operational taxonomic unit (OTU) picking via UPARSE (Stackebrandt and Goebel, 1994; Edgar, 2013) and taxonomy classification through the SILVA 16S rRNA database (Wang, 2007).

Lactobacillus and Firmicutes were detected by qPCR. Eight ileum contents from broiler were collected. The primers used for qPCR are presented in Table 2. The conditions of PCR reaction were summarized as follows: (1) at 95°C for 5 min; (2) a: at 95°C for 30 s; b: at 60°C for 30 s; c: at 72°C for 1 min, including 35 cycles; (3) a: at 95°C for 30 s; b: at 55°C for 30 s; c: at 72°C for 1 min. The ΔCt was calculated as follows: (corrected sample) = mean value of target gene–mean value of internal reference gene (ΔΔ Ct = ΔCt–mean value of control group).

Eight ileal samples Con and Lac_H group were randomly selected for the analysis of LC–MS. The LC–MS analysis of ileal contents was conducted on a Thermo UHPLC-Q Exactive HF-X system equipped with an ACQUITY HSS T3 column (100 mm × 2.1 mm i.d., 1.8 μm; Waters Corp., Milford, MA, United States) at Majorbio Bio-Pharm Technology Co., Ltd. (Shanghai, China). The mass spectrometric data were collected using a Thermo UHPLC-Q Exactive HF-X Mass spectrometer equipped with an electrospray ionization (ESI) source operating in positive and negative modes. The pretreatment of LC–MS raw data was performed by Progenesis QI software (Waters Corp.), and a three-dimensional (3D) data matrix in CSV format was exported. This 3D matrix included the following information: sample information, metabolite name, and mass spectral response intensity. Internal standard peaks and any known false positive peaks (including noise, column bleed, and derivatized reagent peaks) were removed from the data matrix, de-redundant, and peak pooled. Moreover, the metabolites were identified by searching in the following databases: Human Metabolome Database (HMDB)¹, Metlin², and Majorbio³ (Kong et al., 2022; Li C. et al., 2022; Li Z. et al., 2022).

Between-group statistical differences were compared using one-way analysis of variance (ANOVA), followed by post-hoc multiple comparisons using Fisher’s least significant difference (LSD) t-test. The experimental data were presented as the mean ± standard error of the mean (SEM), which were analyzed using SPSS 20.0 software (IBM, Armonk, NY, United States), and p < 0.05 was considered statistically significant. The 16S rRNA genes of gut microbiota were analyzed using an online platform (see Footnote 3) (Ren et al., 2022). The multivariate statistical analysis was performed using the “ropls” (version 1.6.2) R package from Bioconductor on Majorbio Cloud Platform (see Footnote 3) (Ren et al., 2022).

Compared with control, FCR significantly decreased in the Lac_H group (p < 0.05) in 1–21d; ADG significantly increased and FCR significantly decreased in the Lac_H group (p < 0.05) in 22–42 days; weight body and ADG significantly increased in the Lac_H group (p < 0.05) in 42 days. There were no significant changes in FCR, survival rate, and ADFI (p > 0.05) among Lac_H, Lac_L, and ABX groups in 42 days (Table 3).

Compared with control, abdominal fat percentage was significantly decreased by Lactococcus G423 (p < 0.05); however, dressing percentage, thigh muscle percentage, and breast muscle percentage had no significant changes among Lac_H, Lac_L, and ABX groups (p > 0.05) (Table 4).

Comparing with control, the high dose of Lactococcus G423 significantly decreased the serum of TG, TC, and LDL level (p < 0.05), and the low dose of Lactococcus G423 significantly decreased the serum of TG and TC level (p < 0.05). ABX significantly decreased the content of TG (p < 0.05) in serum; however, HDL content had no significant changes among Lac_H, Lac_L, and ABX groups (p > 0.05) (Figure 1).

To characterize the intestinal microbiota composition of broilers in the four groups, 16S rRNA gene sequence analysis was performed. With the sequence similarity of 97%, 903 OTUs were obtained. The average good’s coverage for samples was higher than 99%, indicating that the majority of the microbial species were identified. and sequencing depth was also adequate for the robust sequence analysis.

As shown in Table 5, alpha diversity analysis of gut microbiota showed that compared with the CON group, the Chao and Ace indices in the Lac_H, Lac_L, and ABX groups significantly increased (p < 0.05); however, the Simpson index exhibited an opposite trend. The Simpson index in the Lac_L, Lac_H, and ABX groups was significantly reduced compared with that in the CON group (p < 0.05). In addition, the Sob index in the Lac_H and ABX groups was significantly higher than that in the CON group (p < 0.05). The Shannon index in the Lac_L and Lac_H groups was significantly elevated compared with that in the CON group (p < 0.05). The Coverage index in the Lac_L and ABX groups significantly increased compared with that in the CON group (p < 0.05). The effects of Lactococcus on the diversity and richness of intestinal microbiota community in broilers were evaluated based on alpha diversity (Table 5).

Based on OTU abundance, principal coordinate analysis (PCoA) showed that points in the Lac_H and ABX groups were scattered in the right, which indicated that the microbial structure in the Lac_H and ABX groups had undergone a tremendous change versus that in the CON group. In the Lac_L and CON groups, points were clustered separately from each other in the left, which showed that the low dose of Lactobacillus G423 could change the structure of gut microbiota (Figure 2A). At the phylum level, Firmicutes, Proteobacteria, and Bacteroidetes were the most of species identified in all samples (Figure 2B). At the genus level, compared with those in the CON group, the abundance of Lactobacillus was higher, whereas that of Bacteroides was lower in the Lac_L,Lac_H and ABX groups (Figure 2C). As shown in Figure 2D, qPCR showed that the proportion of Lactobacillus in the Lac_L and Lac_H groups was significantly elevated compared with that in the CON and ABX groups (p < 0.05). Additionally, the proportion of Firmicutes was significantly risen after treating with Lac_H (p < 0.05). The different effects of Lac_L and Lac_H on microbiota might justify their different number of microorganisms. Subsequent linear discriminant analysis effect size (LEfSe) revealed substantial differences in Lactobacillus_salivarius and Lactobacillus_johnsonii in the Lac_H and ABX groups (Figure 2E).

The function of the ileum microbiome was predicted using the phylogenetic investigation of communities by the reconstruction of unobserved species 2 (PICRUSt2). Then, the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis was used to divide the predicted metabolic pathways into six functional groups. The microbial communities in the CON, Lac_L, Lac_H, and ABX groups were mainly related to metabolism, genetic information processing, cellular processes, environmental information processing, human diseases, and organic systems. Their main functions were concentrated in the metabolism of amino acids, carbohydrate, vitamins, terpenoids, polyketides, and lipids (Table 5).

As shown in Figure 2F, the global and overview maps and membrane transport in the Lac_L, Lac_H, and ABX groups significantly changed compared with those in the CON group (p < 0.05). Functional predictions of differences in the mean relative abundance among groups are shown in Figures 2G,H and Table 6.

A total of 6,612 and 5,851 metabolites in ileal contents were determined in positive and negative ion modes, respectively, using LC–MS-based non-targeted metabolomics. A total of 228 metabolites were identified and named based on the HMDB and KEGG databases. Furthermore, orthogonal projection to latent structures-discriminant analysis (OPLS-DA) was employed to select the most predictive and discriminative features to assist classify cation. The loading plot showed a clear separation in metabolites between the Lac_H and CON groups (Figure 3A). The results revealed that the metabolite of broiler significantly changed after treating with Lactococcus. Then, the heat map tree of cluster analysis of metabolites (Figure 3B) was constructed, which visualized 50 significantly different metabolites. Overall, there were significant differences in metabolites between the CON and Lac_H groups.

The levels of several metabolites such as 6-hydroxy-5-methoxyindole glucuronide, 3-{3,4-dihydroxy-2-[(2Z)-4-hydroxy-3-(4-methylpent-3-en-1-yl)but-2-en-1-yl]phenyl} propanoic acid, indoxylsulfuric acid, Cinerin II, carbamazepine-O-quinone, and 4-(1-hydroxy-3-phenylpropyl)-5-methoxy-2,6-dimethylbenz ene-1,3-diol were upregulated in the Lac_H group, while the levels of 9,10-DiHOME, seryltryptophan, seryltryptophan dihydroferulic acid 4-O-glucuronide, 2,6-dimethoxy-1,4-benzoquinone, and kynurenine were downregulated (Figure 3C). Metabolites discriminated among different groups were screened using the variable importance in projection (VIP) scores obtained from the OPLS-DA model, and the ileal contents of metabolic profiles were determined. The metabolites were statistically significant if VIP score ≥ 1 and p < 0.05, and p-value was calculated by the t-test. Metabolites with VIP score > 1.0 and p < 0.05 were considered to be significantly influenced by the Lac_H. Thirty significantly affected metabolites were identified in the CON and Lac_H groups, respectively; the top 30 metabolites with the highest VIP scores are presented in Figure 3D.

Metabolic pathway enrichment analysis was performed based on the KEGG database for the differential metabolites between the CON and Lac_H groups, and the metabolic pathway with p < 0.05 was significantly enriched for the differential metabolites, including bile secretion, linoleic acid metabolism, drug metabolism-cytochrome P450, phenylalanine metabolism, tryptophan metabolism, and matching metabolites. Lactococcus G423 could significantly improve the levels of certain metabolites (6-hydroxy-5-methoxyindole glucuronide, 9,10-DiHOME, N-acetyl-l-phenylalanine, and kynurenine), and these metabolites were involved in four metabolic pathways (Table 7). Among them, the pathways of linoleic acid metabolism, phenylalanine metabolism, and pentose and glucuronate interconversions significantly varied (p < 0.05) (Figure 3E).

The variations in the intestinal microbiota could be related to the metabolic phenotype. As shown in Figure 4, correlation analysis was performed between 34 different metabolites and 44 bacteria with significantly different relative abundances at the genus level. There was a significant correlation between 2(R)-hydroxyicosanoic acid, 2(R)-hydroxydocosanoic acid, l-beta-aspartyl-l-glutamic acid, 9,10-DiHOME, TMPD (hydrochloride), kynurenine, 6-hydroxy-5-methoxyindole glucuronide, seryltryptophan, and N-acetyl-l-phenylalanine and Parabacteroides, Romboutsia, Sellimonas, Subdoligranulum, Turicibacter, Tuzzerella, Bacteroides, Lachnospiraceae, Butyricicoccus, Candidatus_Arthromitus, Eisenbergiella, Escherichia-Shigella, Faecalibacterium, Alistipes, Marvinbryantia, Monoglobus, and Negativibacillus (all p < 0.05).

The diversity and relative abundance of intestinal microbes play an important role in the health of host by participating in metabolism and immunomodulation (Zheng et al., 2021). The findings of the present study suggested that Lactococcus G423 could significantly increase ADG in broilers, which were similar to previously reported results (Faseleh et al., 2016; Chen et al., 2018). Supplementation of broilers’ diet with antibiotics could increase body weight gain (Rahman et al., 2012); however, our results showed that Lac_H significantly increase ADG compared with ABX in broiler. This improvement was explained by improved feed conversion efficiency and increased vitality and regulation of the intestinal microflora.

Different from mammals, chickens synthesize fatty acids predominantly in the liver and then export to other tissues including muscle and adipose tissue by the peripheral vascular system. Therefore, the blood lipid index is related to the carcass characteristics. The carcasses from 42-day-old Ross 308 chickens of both sexes, which received the multicomponent probiotics Pro-Biotyk (Em-15) and EMFarma^TM, did not differ significantly in the percentage of dissected carcass characteristics (Stęczny and Dariusz, 2020). The study by Ding et al. (2021) reveals that lactobacillus reduced abdominal fat deposition in broilers. Our results showed that abdominal fat percentage was lowered by Lactococcus G423, and dressing percentage, thigh muscle percentage, and breast muscle percentage had no significant changes among Lac_H, Lac_L, and ABX groups compared with CON groups in broilers. The current results are supported by previous studies on the effect of probiotics on carcasses (Liu et al., 2016; Rybarczyk et al., 2020).

Lipids mainly include triglyceride (TG), phospholipids, and cholesterol (CHO), and the contents of TG and CHO are key indicators of lipid metabolism. The administration of Paenibacillus polymyxa up to 0.4 mg/kg diet significantly reduced plasma TC, LDL, and TG (Alagawany et al., 2021).

Ding et al. (2021) found that Lactobacillus participated in the lipid metabolism of broilers by reducing the content of TC and TG. Other studies also showed that LAB had effects of blood serum levels TC, HDL-C, LDL-C, and TG on rat (An et al., 2011; Wang et al., 2023). Our results showed that Lac_H significantly decreased the content of TG, TC, and LDL, and Lac_L significantly decreased the content of TG and TC in serum, which was similar to previous studies (Abramowicz, 2019; Abdel-Moneim et al., 2020).

The gut microbiota community is consisted of diverse types of microbes. In the present study, it was found that Lactococcus G423 and ABX altered microbiome diversity in the ileum of broilers and changed the relative abundance rates of Firmicutes, Bacteroidetes, Proteobacteria, and other species. At the phylum level, Firmicutes, Bacteroidetes, and Proteobacteria were the most common phyla in the poultry intestinal samples, which is consistent with the previous findings (Shaufi et al., 2015; Qiao et al., 2018; Zheng et al., 2021). This study indicated that Firmicutes was the dominant phylum (>50%) in broilers, and similar results have been previously reported (Danzeisen et al., 2011; Mohd Shaufi et al., 2015). Moreover, this study revealed that the abundance rates of Firmicutes were relatively higher in the Lac_L, Lac_H, and ABX groups compared with those in the CON group (p < 0.05). Wang et al. (2017) reported that Lactobacillus significantly aided in altering the abundance of Firmicutes and decreased the content of TG and LDL. Firmicutes were associated with lipid metabolites (Turnbaugh et al., 2009). The phylum Bacteroidetes has influences on dissolving lipids (Kumar et al., 2018; Chen et al., 2020). Bacteroides are also positively correlated with several lipid metabolites (Saxena et al., 2016; Shulpekova et al., 2022). In the present study, it was revealed that the abundance of Bacteroides in the Lac_L, Lac_H, and ABX groups was markedly lower than that in the CON group. Meanwhile, considering the TG, LDL, and TC in this experiment, our findings also showed that Lactococcus G423 regulated lipid metabolism though regulating intestinal microflora. Although ABX has an effect on the abundance of the Firmicutes and Bacteroides, there is no effect on the level of LDL, TC, and abdominal fat percentage. Firmicutes and Bacteroidetes can also contribute to host metabolism through several mechanisms, including increased energy harvested from the diet and modulation of lipid metabolism (Greiner and Bäckhed, 2011). Some studies have suggested that a lower abundance of Bacteroidetes was associated with increased body weight (Ley et al., 2006; Arumugam et al., 2011). Previous studies also demonstrated that the Firmicutes/Bacteroidetes ratio and the growth performance were positively correlated, and this ratio could be indicative of the status of the intestinal bacteria (Haas et al., 2011; Xu et al., 2016). The results of the present study revealed that the Firmicutes/Bacteroidetes ratio was relatively higher in the Lac_L, Lac_H, and ABX groups compared with that in the CON group. Lactococcus G423 significantly increased ADG by changing the Firmicutes/Bacteroidetes ratio. Additionally, the levels of Proteobacteria phylum, including some pathogens, such as Escherichia, Salmonella, Helicobacter, and Vibrio, were slightly lower in the Lac_L group than those in the CON group, indicating that Lactococcus significantly aided in altering the abundance of opportunistic pathogens. However, at the genus level, Lactobacillus, Candidatus_Arthromitus, Romboutsia, and Bacteroides were identified as the dominant species in the ileum microbiome. Lactobacillus, belonging to the phylum of Firmicutes, had markedly higher level in the Lac_L and Lac_H groups than that in the CON group, and the abundance of Lactobacillus in the Lac_H group reached the highest rate. Lactobacillus is involved in digestive and metabolic processes and in the regulation of local and systemic immune response (Fernández et al., 2016). Moreover, Lactobacillus altered lipid metabolism (Wang et al., 2017). Similar effects were observed by Zhou et al. (2016), who studied that Bacillus licheniformis and Lactobacillus had an effect on the growth and fat deposition in broilers (Gerritsen et al., 2014). L. fermentum TSI reduces abdominal fat and improves blood lipid metabolism in HD-induced obese rats (Cho et al., 2020). Therefore, Lactococcus G423 significantly aided in altering the abundance of Lactobacillus, which participated in gut microbiota, growth, and lipid metabolism in animals (Wu et al., 2019).

Romboutsia have been identified in the human gut (Ricaboni et al., 2016), the rat gastrointestinal tract (Gerritsen et al., 2014), and the fecal of hens (Qiao et al., 2018). In the present study, a novel genus Romboutsia was found in ileum samples of broilers. However, the abundance of Romboutsia was inconsistent among the four groups. Meanwhile, it was revealed that ABX altered the relative abundance rates of other bacteria in ileum contents of broilers, negatively influencing the gut microbiota. Previous studies reported that improper uses of antibiotics have been increased antimicrobial-resistant bacteria as a public health threat (Nhung et al., 2017; Christy et al., 2018; Oniciuc et al., 2018). The results revealed that Lactobacillus G423 had more noticeable health benefits compared with antibiotics.

The Lactococcus-regulated gut microbiota led to alterations in the contents of ileum metabolites. Several significantly altered metabolites were identified in the present study, such as 6-hydroxy-5-methoxyindole glucuronide, 9,10-DiHOME, N-acetyl-l-phenylalanine, and kynurenine, which were regulated, and they were involved in four metabolic pathways (Table 7).

Among them, the bile secretion and linoleic acid metabolism were the important metabolic pathways of lipid (Hamilton and Klett, 2021; Shulperkova et al., 2022). It was revealed that 6-hydroxy-5-methoxyindole glucuronide was related to the pathway of bile acid metabolism. Bile acids are also signaling molecules and inflammatory agents that rapidly activate nuclear receptors and cellular signaling pathways, regulating lipid, glucose, and energy metabolism. To a large extent, bile salts are (>95% per cycle) absorbed in the terminal ileum, the final section of the small intestine. The bile salt hydrolase activity has been widely detected in several bacterial genera, including Bacteroides, Clostridium, Lactobacillus, and Bifidobacteria. In addition, bile acids have been found to affect glucose metabolism by activating FXR and TGR5 receptors, as well as intestinal flora (Jung et al., 2007). Lactococcus G423 upregulated 6-hydroxy-5-methoxyindole glucuronide level, suggesting that it may have some regulatory effects on the bile acid metabolism.

Linoleic acid could have beneficial effects on maintaining healthy squabs, as reflected by improved antioxidant capacity and lipid metabolism (Xu et al., 2020). Linoleic acid has shown a correlation with lipid metabolic diseases (Choque et al., 2014). Previous studies have demonstrated that linoleic acid content was associated with probiotics (Hossain et al., 2012; Sahoo et al., 2015). The metabolized product of linoleic acid is 9,10-dihydroxy-12-octadecenoic acid (9,10-DiHOME) (Felipe et al., 2023). More recent research has suggested that DiHOMEs may be important lipid mediators (Hildreth et al., 2020; Zhou et al., 2023). Propionibacterium acnes and Lactobacillus plantarum have been reported to convert linoleic acid into conjugated linoleic acid (Bo et al., 2017). In the present study, Lactococcus G423 downregulated 9,10-DiHOME level, suggesting that it may have some regulatory effects on the linoleic acid metabolism.

Lactococcus G423 regulated 9,10-DiHOME and-hydroxy-5-methoxyindole glucuronide by effecting the abundance of Bacteroides and Lactobacillus, which effected the lipid metabolic pathway of bile secretion and linoleic acid.