Aspergillus niger with degrading AFB₁ and ZEN was preserved in the Laboratory of Animal Nutrition and Feed Biotechnology in Henan Agricultural University. GA was provided by Henan Delin Biological Products Co., Ltd. The preparation of solid culture of Aspergillus niger was made by mixing the three feed ingredients of bran, soybean meal and corn in the ratio of 7:2:1, taking 10 g of the mixture in a triangular flask, after sterilization, adding 1 mL of Aspergillus niger seed solution and 5 mL of sterile water, mixing thoroughly and cultivating at 30°C for 5–7 d and then taking it out and drying and crushing the mixture (39). The mycotoxin degradation experiments in vitro confirmed that the degradation rates of AFB₁ and ZEA were 60.40 and 97.67%, respectively, when 0.04% Aspergillus niger culture was applied. The further research indicated that the degradation rate of AFB₁ was increased to 68, 71 and 63% when 0.02, 0.04 and 0.06% of GA were added. In addition, previous study in our laboratory confirmed that GA could alleviate the damage of intestinal cells caused by DON (38). Therefore, CANCGA was prepared by combining Aspergillus niger culture with GA at both 0.04% addition (mixed in ratio 1:1 ratio) for alleviating toxicity of AFB₁, ZEN and DON in the further broiler feeding experiment.

500 one-day-old male Arbor Acres (AA) broilers were divided into 10 groups, each group consisting of 5 replications with 10 broilers per replication. The broilers were reared in cages, allowing free access to diet and water. Room temperature was around 25°C, but the temperature under nurturing umbrella was 33–35°C, 29–32°C and 26–28°C for one-week-old, two-week-old and three-week-old broilers, respectively. The relative humidity was kept at 60 to 65%. The experimental period was 21 d. Daily feed intake, and dead broilers were recorded daily. Broiler body weight in each replication was weighted at the age of 1 and 22 d. Parameters such as average daily gain (ADG), average daily feed intake (ADFI), feed-to-gain ratio (F/G) and mortality were calculated. Additionally, a standard immunization program was implemented within the first one week. The experimental groups were organized as follows:

The basal diets were formulated based on the broiler feeding standards outlined in the NRC (1994) guideline. To adjust AFB₁ and ZEN contents in diets, normal corn in the basal diet was substituted with moldy corn, while DON content was adjusted with corn by-product. The diet formulation and nutrient levels were listed in Table 1.

Excreta collection of broilers in each replication was performed on days 18–20 of the feeding experiment, followed by spraying with 10% sulfuric acid solution for nitrogen fixation, dried at 65°C, and ground for further use. At day 21, blood samples (2 mL) were collected from the wing vein of five broilers in each group and stored in a refrigerator at 4°C until serum precipitation occurred in the collection tube. Subsequently, centrifugation was conducted at 1,520 × g for 10 min to obtain serum, which was then transferred to a sterilized tube and stored at −80°C for future utilization.

The crude protein (CP), ether extract (EE), calcium (Ca) and phosphorus (P) contents in diets and excreta were determined according to the methods of national standards GB/T 6432–2018, GB/T 6433–2006, GB/T 6436–2002 and GB/T 6437–2018, respectively. The calculation of nutrient metabolic rates was as follows: Nutrient metabolic rate (%) = 100 × (nutrient content in diet−nutrient content in excreta)/nutrient content in diet. Based on the growth performance of broilers in each group at 21 d, broilers in groups A, B, E and H were selected for the determination of excreta toxin residues. The contents of AFB₁, ZEN and DON in excreta were detected according to the protocol of Suwei toxin detection kits (Suwei Biological Research Co., Ltd. Jiangsu, China). The calculation of AFB₁ degradation rate was as follows: AFB₁ degradation rate (%) = (AFB₁ content in control group−AFB₁ content in test group)/AFB₁ content in control group × 100. DON and ZEN degradation rates were calculated in the same way as AFB₁.

Using a fully-automated blood biochemistry analyzer to measure the serum contents of glucose (GLU), triglyceride (TG), high-density lipoprotein (HDL), total cholesterol (TC), low-density lipoprotein (LDL), aspartate aminotransferase (AST), lactate dehydrogenase (LDH), alkaline phosphatase (ALP), alanine transaminase (ALT), total protein (TP) and albumin (ALB).

Based on the growth performance of broilers in each group at 21 d, broilers in groups A, B, E and H were selected for serum metabolomics analysis. Serum sample was thawed at 4°C and vortexed for 1 min using a vortex mixer (BE-2600, Haimen Qilin Bell Instrument Manufacturing Co., Ltd., Nantong, China) to ensure thorough mixing. And then transferred into 2 mL centrifuge tube. Subsequently, 400 μL methanol (stored at −20°C) was added and vortexed for 1 min prior to centrifugation. The sample was centrifuged at 13,680 × g and 4°C for 10 min using a refrigerated centrifuge (H1850-R, Hunan Xiangyi Laboratory Instrument Development Co., Ltd., Changsha, China). The supernatant was carefully transferred to another 2 mL centrifuge tube, then concentrated and dried before being dissolved in 150 μL 4 mg/L 2-chloro-L-phenylalanine (prepared with 80% methanol). The resulting supernatant was filtered through a 0.22 μm filter membrane before being added to the UPLC-MS vial for UPLC-MS detection.

MS/MS analysis was performed using the Thermo Vanquish Ultra High-Performance Liquid Chromatography (UHPLC) system (Vanquish, Thermo, Massachusetts, United States) coupled with an ACQUITY UPLC^® HSS T3 column (2.1 × 100 mm, 1.8 μm). In positive ion mode, the mobile phases consisted of 0.1% formic acid diluted in water (A1) and 0.1% formic acid diluted in acetonitrile (B1); in negative ion mode, the mobile phases were 5 mM ammonium formate diluted in water (A2) and acetonitrile (B2). Thermo Orbitrap Exploris 120 mass spectrometer (Orbitrap Exploris 120, Thermo, Massachusetts, United States) was used to determine the serum metabolites. Mass spectrometry parameters were as follows: positive ion spray voltage of 3.50 kV, negative ion spray voltage of −2.50 kV.

The animal experimental data were expressed as means ± standard error. The one-way ANOVA test was conducted using SPSS 26.0. Statistical significance was considered at p < 0.05.

The data from metabolomics were processed by XCMS in R package for peak detection, filtering, and alignment of raw mass spectrometry files converted by Proteowizard. Substance identification was conducted by searching and comparison with spectral databases such as HMDB, MassBank, LipidMaps, mzCloud, KEGG, and metabolite standards database of Nomi Metabolism. Differential metabolites were selected based on criteria of p < 0.05 and VIP > 1. Spearman correlation analysis was performed between differential metabolites and production performance, serum biochemistry or three kinds of toxin residues in excreta by using Nomi Metabolism platform.

Multivariate statistical methods including downscaling and categorization such as principal component analysis (PCA) and orthogonal-partial least squares discriminant analysis (OPLS-DA) were employed to analyze the data from serum metabolomics in order to identify the differential metabolites between different groups. The significant differences were determined by t-test and the variable importance projection (VIP) of the first principal component of OPLS-DA. A KEGG pathway enrichment analysis was conducted for the identified differential metabolites to evaluate their potential roles in biological responses.

As presented in Table 2, mortality was significantly increased in the low-dose or high-dose mycotoxin groups, compared to the control group. However, supplementation with CANCGA made mortality become zero to remain consistent with the control group, indicating that CANCGA was able to alleviate multi-mycotoxin toxicity for broilers. Moreover, compared to the basal diet, a significant (p < 0.05) improvement of ADG and final body weight were observed when 0.4 g/kg CANCGA was added to the basal diet, while final body weight, ADG, ADFI, and F/G were not significantly different when 0.2 g/kg and 0.6 g/kg CANCGA was added to the basal diet. Notably, final body weight and ADG were significantly higher with the addition of 0.4 g/kg CANCGA to the basal diet than with the 0.2 g/kg addition (p < 0.05). ADG and ADFI were significantly decreased, while F/G was significantly increased for broilers subjected to high-dose mycotoxin diet, compared to both basal and low-dose mycotoxin diets (p < 0.05).

Table 3 showed that high-dose mycotoxin diets without or with CANCGA addition significantly decreased CP and P metabolic rates, while increased EE metabolic rate, compared with the control group (p < 0.05). CANCGA addition in high-dose mycotoxin diet had insignificant effect on nutrient metabolic rates. P metabolic rate was significantly decreased (p < 0.05), while EE metabolic rate was significantly increased with dietary mycotoxin levels increasing (p < 0.05). Compared with the control group, low-dose mycotoxin diet had insignificant effect on other nutrient metabolic rates in spite of increasing EE metabolic rate (p < 0.05). In low-dose mycotoxin diets, 0.4 g/kg CANCGA addition significantly increased EE metabolic rate, 0.6 g/kg CANCGA addition significantly decreased Ca metabolic rate, compared with the low-dose mycotoxin diet without CANCGA addition (p < 0.05). Compared with the control group, 0.4 g/kg CANCGA addition in the basal diet significantly decreased EE and Ca metabolic rates (p < 0.05), while 0.2 g/kg and 0.6 g/kg CANCGA additions in the basal diet had insignificant effect on nutrient metabolic rates. In addition, CANCGA added at 0.4 g/kg in the low mycotoxin group had significantly higher metabolic rates of EE and P than the 0.2 g/kg addition (p < 0.05).

Based on the feeding experiment, the excreta and serum samples from groups A, B, E and H were selected as the representative ones for further analyses. As shown in Table 4, the contents of residual AFB₁, ZEN and DON in broiler excreta were significantly increased in the low-dose mycotoxin group, compared to the control group (p < 0.05). Furthermore, compared to the basal diet, residual AFB₁ content was significantly decreased when 0.4 g/kg CANCGA was added in the basal diet (p < 0.05); however, CANCGA addition had the tendency to decrease ZEN and ZON contents in excreta. Compared to the low-dose mycotoxin group, ZEN residue was significantly decreased by 0.4 g/kg CANCGA addition (p < 0.05); however, CANCGA addition had the tendency to decrease AFB₁ and ZON contents in excreta.

Figure 1 indicated that serum ALP, LDH, TP, ALB, GLU, TC and TG levels in low-dose mycotoxin group were significantly decreased (p < 0.05), while AST level was significantly increased (p < 0.05), compared with the control group. However, the addition of 0.4 g/kg CANCGA to the basal diet could decrease serum ALP, TG and LDL levels (p < 0.05), while increase HDL level (p < 0.05). Compared with the low-dose mycotoxin group, the addition of 0.4 g/kg CANCGA could decrease serum ALT, AST, ALP, GLU, TG and HDL levels (p < 0.05), while increase ALB, TC levels and ALB/GLB (p < 0.05). It was inferred that CANCGA addition was able to reduce tissue and organ damage and lipid metabolism disorders of broilers induced by multi-mycotoxins.

In order to study the mechanism of CANCGA for alleviating multi-mycotoxin toxicity, nutrient metabolism of broilers was analyzed by UPLC-MS spectrometry (Figures 2A,B). In Figure 2C, the horizontal coordinates indicated the similarity between the real grouping of the samples and the 100 random groupings, the vertical coordinates indicated the model evaluation parameters, Q2 and R2 points in the upper right corner indicated the model evaluation parameters of the real grouping. If both Q2 fell below R2, it means that the results are reliable. The results of PCA and OPLS-DA in Figure 2 revealed notable differences of metabolites in broiler serum. In the comparisons of group B vs. group A, group H vs. group B, group H vs. group A, group E vs. group A, about 26, 16, 31 and 18 differential metabolites were identified, respectively (Table 5). Nine metabolites were significantly up-regulated, and seventeen metabolites were down-regulated in the comparison of group B vs. group A (p < 0.05). Seventeen metabolites were significantly up-regulated, and fourteen metabolites were down-regulated in the comparison of group H vs. group A (p < 0.05). Six metabolites were significantly up-regulated, and twelve metabolites were down-regulated in the comparison of group E vs. group A (p < 0.05). Eleven metabolites were significantly up-regulated, and five metabolites were down-regulated in the comparison of group H vs. group B (p < 0.05).

The distribution and alteration of distinct metabolites between the experimental groups and the control group were illustrated in Figure 2D–F. Additionally, the top five metabolites exhibiting statistically significant differences among the groups were highlighted in the volcano plot. Compared to group A, (15Z)-tetracosenoic acid and phthalic acid were downregulated, while 2,3-butanediol, N-acetylaspartylglutamate and 9-cis-retinol were upregulated in group B (Figure 2D). Compared to group B, linoleic acid and glycitein were downregulated, while hydroquinone, oxalureate and 2,3-butanediol were upregulated in group H (Figure 2E). Furthermore, compared to group A, (15Z)-tetracosenoic acid, phthalic acid and hydroxykynurenine were downregulated, while 2,3-butanediol and nicotinic acid were upregulated in group H (Figure 2F). The differential metabolites in each group were analyzed using Venn plots (Figure 2G). In four comparisons such as group B vs. group A, group H vs. group B, group H vs. group A, group E vs. group A, two common differential metabolites (2,3-butanediol and hydroxykynurenine) were identified. Additionally, seven common metabolites (dimethylglycine, 2,3-butanediol, selenocysteine, jasmonic acid, hydroxykynurenine, 3-ketosphingosine and glycitein) were found in the comparisons of group B vs. group A, group H vs. group B.

To further elucidate the metabolic pathways associated with the differential metabolites in serum, KEGG pathway enrichment analysis was conducted for the following four comparisons: group B vs. group A, group H vs. group B, group H vs. group A, group E vs. group A (Figures 3A–D). Compared with group A, the differential metabolites in group B were primarily enriched in alanine, aspartate and glutamate metabolism pathway; glycine, serine and threonine metabolism pathway; arginine and proline metabolism pathway; steroid hormone biosynthesis pathway (Figure 3A). Compared with group B, the differential metabolites in group H were primarily enriched in the pathways related to glycine, serine and threonine metabolism as well as cysteine and methionine metabolism (Figure 3B). The differential metabolites observed in group H were predominantly associated with selenocompound metabolism, arginine and proline metabolism as well as ABC transporters, compared with group A (Figure 3C).

Further KEGG enrichment analysis was conducted for the common differential metabolites in the comparisons of group B vs. group A, group H vs. group B, revealing the significant associations with three pathways such as selenocompound metabolism, alpha-linolenic acid metabolism as well as glycine, serine and threonine metabolism (Figure 3D). These findings indicated that mycotoxins impacted amino acid synthesis and metabolism as well as steroid hormone biosynthesis pathways to induce organism damage. However, supplementation with CANCGA could mitigate mycotoxin-induced damage by modulating amino acid synthesis and metabolism pathways along with selenium complex metabolism and α-linolenic acid metabolism.

The relationships between differential metabolites in serum and serum biochemical indices, ADFI, ADG, nutrient metabolic rates or residual levels of mycotoxins in excreta were illustrated in Figure 4. Serum metabolites including 2,3-butanediol, 9-cis-retinol, and Se-methylselenocysteine were significantly positively correlated with serum TG level (p < 0.05). Conversely, selenocysteine, phthalic acid, and 2-keto-glutaramic acid were significantly negatively correlated with serum TG level (p < 0.05). Serum ALP level was positively correlated with the N-acetyl-D-tryptophan (p < 0.05). Serum AST level was significantly positively correlated with oxalureate and nicotinic acid (p < 0.05). Serum N-acetylornithine level was significantly negatively correlated with F/G (p < 0.05). Serum anserine level was significantly negatively correlated with AFB₁ residue in excreta, while creatine was significantly negatively correlated with ZEN residue in excreta. Both serum creatine and L-homoserine levels showed significant negative correlations with EE metabolic rate, whereas P metabolic rate had a significant positive correlation with serum 2-methoxyestradiol level. The above results indicated that serum differential metabolites had close correlation with broiler growth performance, serum biochemical parameters or toxin residues in excreta.