Immunoregulation effects of CGA in DEX-induced broilers
Compared with the Control group, DEX treatment significantly increased (p < 0.05) the serum IL-1β, IL-6, IL-18, IL-22, TNF-α, CXCL1, and CXCL2 levels of the broilers, and significantly decreased (p < 0.01) the serum IL-4 and IFN-γ levels. In contrast, CGA supplementation significantly increased (p < 0.05) serum IgM, IL-4, and IFN-γ levels and significantly decreased serum IL-1β, IL-12, IL-18, IL-22, CXCL1, and CXCL2 levels. Moreover, CGA supplementation (CGA × DEX) significantly reversed (p < 0.05) DEX-induced changes in serum IL-1β, IL-4, IL-6, IL-10, IL-12, IL-18, IL-22, IFN-γ, CXCL1, and CXCL2 levels (Table 1).
Furthermore, DEX treatment significantly increased (p < 0.01) the expression of IL-1β, IL-4, IL-6, IL-18, IL-22, TNF-α, CXCL1, and CXCL2 in the jejunal mucosa of the broilers. In contrast, CGA supplementation significantly increased (p < 0.01) the jejunal expression of IgM and decreased (p < 0.01) the jejunal expression of IL-1β, IL-6, IL-12, IL-18, IL-22, and CXCL2. Additionally, CGA supplementation (CGA × DEX) significantly reversed (p < 0.01) DEX-induced changes in the jejunal expression of IL-1β, IL-4, IL-6, IL-12, IL-18, IL-22, CXCL1, and CXCL2 (Table 2). Moreover, gene expression analysis showed that DEX treatment significantly increased (p < 0.05) jejunal expression of IL-1β, IL-6, IL-12, IL-18, IL-22, TNF-α, Caspase-3, and Caspase-9 genes compared with the Control group. In contrast, CGA supplementation significantly decreased (p < 0.05) the jejunal expression of IL-1β, IL-6, IL-12, IL-18, IL-22, TNF-α, Caspase-3, and Caspase-9. Additionally, CGA supplementation (CGA × DEX) significantly reversed DEX-induced changes in the jejunal expression of IL-1β, IL-4, IL-6, IL-10, IL-12, IL-18, IL-22, TNF-α, Caspase-3, and Caspase-9 genes (Fig. 1).
CGA improves the jejunal morphology and barrier function of DEX-treated broilers
Compared with the Control group, histological analysis showed that DEX treatment significantly decreased (p < 0.05) villus height and villus height/crypt depth (V/C), and increased (p < 0.05) crypt depth. In contrast, CGA supplementation significantly increased (p < 0.05) villus height and V/C, and decreased crypt depth. Additionally, CGA supplementation (CGA × DEX) significantly reversed (p < 0.05) DEX-induced decrease in villus height and V/C (Fig. 2A, Table 3).
Moreover, DEX treatment significantly increased (p < 0.01) the D-LA levels of the broilers, whereas CGA supplementation significantly decreased (p < 0.01) the D-LA levels. Additionally, CGA supplementation (CGA × DEX) significantly reversed (p < 0.01) DEX-induced increase in D-LA levels. However, DAO level was not significantly affected by CGA, DEX, or their interaction (Table 4).
Furthermore, the expression of tight junction proteins was examined by Western blotting and immunohistochemical analysis. The results showed that DEX treatment significantly downregulated (p < 0.01) occludin expression, whereas CGA supplementation significantly upregulated (p < 0.01) occludin expression. Additionally, CGA supplementation (CGA × DEX) significantly reversed (p < 0.05) DEX-induced decrease in occludin and ZO-1 expression (Fig. 2B), which was confirmed by immunohistochemical analysis (Fig. 2C-D).
The gut microbiota and SCFAs were altered by CGA
High-throughput 16S rRNA sequencing was performed to determine the effect of DEX treatment and CGA supplementation on the gut microbiome of the broilers. The α-diversity indices, including chao1, goods_coverage, observed_otus, Shannon, and Simpson, were not significantly influenced (p > 0.05) by CGA or DEX treatments. However, DEX-CGA treatment significantly increased (p < 0.05) chao1 and observed_otus indices compared with the DEX group (Fig. 3A). PCoA showed that different treatments induced distinct (p < 0.05) clustering of bacterial communities (Fig. 3b), with different gut microbiota compositions at the phylum, family, genus, and species levels. At the phylum level, CGA supplementation significantly decreased (p < 0.05) the abundance of Actinobacteria. At the family level (top 20), CGA supplementation significantly decreased (p < 0.05) the abundance of Firmicutes_unclassified, Christensenellaceae, and Mollicutes_RF39_unclassified. DEX treatment significantly decreased (p < 0.05) the abundance of Clostridiales_vadinBB60_group. Compared with the DEX group, DEX-CGA treatment significantly increased (p < 0.05) the abundance of Clostridiales_vadinBB60_group. At the genus level (top 20), DEX treatment significantly decreased (p < 0.05) the abundance of Clostridiales_vadinBB60_group_unclassified and increased (p < 0.05) the abundance of Erysipelatoclostridium. Dietary supplementation with CGA significantly increased (p < 0.05) the abundance of Intestinimonas and decreased (p < 0.05) the abundances of Ruminococcaceae_UCG-014, Firmicutes_unclassified, and Ruminiclostridium_5. Additionally, DEX-CGA treatment significantly increased (p < 0.05) the abundance of Clostridiales_vadinBB60_group_unclassified compared with the DEX group. At the species level, DEX treatment significantly decreased (p < 0.05) the abundance of Clostridiales_vadinBB60_group_unclassified. Moreover, CGA supplementation significantly increased (p < 0.05) the abundance of Intestinimonas_unclassified and significantly decreased the abundance of Ruminococcaceae_UCG-014_unclassified, Firmicutes_unclassified, and Ruminiclostridium_5_unclassified. Moreover, compared with the DEX group, DEX-CGA treatment significantly increased (p < 0.05) the abundance of Clostridiales_vadinBB60_group_unclassified (Fig. 3C).
LEfSe analysis was performed to identify taxonomic biomarkers in the gut microbiota. There was an increase in the relative abundance of bacteria, including Coprobacter (genus), Coprobacter_fastidiosus (species), Anaerotruncus_unclassified (species), DTU089 (genus), and DTU089_unclassified (species), in non-treated broilers. Additionally, CGA supplementation increased the relative abundance of Intestinimonas (genus), Intestinimonas_unclassified (species), UC5_1_2E3 (genus), UC5_1_2E3_unclassified (species), and Eubacterium_unclassified (species). DEX treatment increased the relative abundances of Shuttlewothia (genus) and Erysipelatoclostridium_unclassified (species). DEX-CGA treatment increased the relative abundance of Clostridiales_vadinBB60_group (family), Clostridiales_vadinBB60_group_unclassified (genus and species), Erysipelatoclostridium (genus), Shuttleworthia_unclassified (species), and Lactobacillus_hilgardii (species) (Fig. 3D).
PICRUSt analysis was conducted to determine the potential function differences of the gut microbiota between the groups and predict their classification based on the KEGG pathways. Compared with the CGA group, there was a decrease in four terms, including methanogenesis from acetate, starch degradation V and galactose degradation I (Leloir pathway), and an increase in 26 terms, including myo-, chiro-, and scillo-inositol degradation, D-fructuronate degradation, and superpathway of sulfur oxidation (Acidianus ambivalens) in the Control group. Compared with the Control group, there was a decrease in seven terms, including L-glutamate degradation V (via hydroxyglutarate), pyrimidine deoxyribonucleotide biosynthesis from CTP, and GDP-mannose biosynthesis, and an increase in two terms, including sucrose degradation IV (sucrose phosphorylase) and sucrose degradation III (sucrose invertase) in the DEX treatment group. Compared with the DEX group, there was decrease in glycerol degradation to butanol degradation and sucrose degradation IV (sucrose phosphorylase) and an increase in nine terms, including superpathway of polyamine biosynthesis II, D-fructuronate degradation, and pyruvate fermentation to butanoate, in the DEX-CGA group (Fig. 4).
SCFAs are the main metabolites generated by gut microbiota. In the present study, DEX treatment significantly decreased acetic, propanoic, butyric, isovaleric, valeric, and, hexanoic levels. In contrast, CGA supplementation significantly increased (p < 0.05) the levels of acetic, propanoic, butyric, isovaleric, valeric, and hexanoic acid. Additionally, CGA supplementation (CGA × DEX) significantly reversed (p < 0.01) DEX-induced decrease in acetic, propanoic, butyric, isovaleric, valeric, and hexanoic acid levels (Table 5).
The jejunal protein profiles were altered by CGA
Differentially expressed proteins (DEPs) in the DEX vs Control, CGA vs Control, and DEX-CGA vs DEX comparison groups are represented using volcano plots (Fig. 5A). A total of 58 DEPs were identified in the DEX vs Control comparison group, among which 25 were upregulated and 33 were downregulated. A total of 37 DEPs were identified in the CGA vs Control comparison group, among which 27 were upregulated and 10 were downregulated. A total of 109 DEPs were identified in the DEX-CGA vs DEX comparison group, among which 61 were upregulated and 48 were downregulated. The top 10 upregulated and downregulated DEPs are presented based on fold change (Table 6-8).
GO enrichment analysis of jejunal proteins showed that there were enriched in several “biological processes (BP),” “cellular components (CC)” and “molecular functions (MF)” (Fig. 5B). Specifically, DEPs in the DEX vs Control groups were significantly enriched in BP terms, such as oxidation-reduction process, chemical homeostasis, and drug metabolic process; in CC terms, such as extracellular region, cytoskeleton, extracellular space, and extracellular; and in MF terms, such as transition metal ion binding, oxidoreductase activity, protein dimerization activity, cofactor binding, protein homodimerization activity, peptidase activity acting on L-amino acid peptides, peptidase activity, coenzyme binding, and zinc ion binding. DEPs in the CGA vs Control groups were significantly enriched in BP terms, such as cytoskeleton organization and cellular protein-containing complex assembly; in CC terms, such as cytoskeleton and plasma membrane part; and in MF terms, such as cytoskeletal protein binding and DNA binding. DEPs in the DEX-CGA vs DEX groups were significantly enriched in BP terms, such as carbohydrate metabolic process, myeloid cell differentiation, and organic anion transport; and in CC terms, such as plasma membrane part, cytoskeletal part, plasma membrane region, cell projection part, and plasma membrane bounded cell projection part; and in MF terms, such as cytoskeletal protein binding, protein dimerization activity, protein homodimerization activity, and actin binding were the dominant terms.
KEGG metabolic pathway enrichment analysis showed that DEPs in DEX vs Control groups were significantly enriched in protein digestion and absorption, peroxisome proliferator-activated receptor (PPAR) signaling pathway, and proximal tubule bicarbonate reclamation. DEPs in the CGA vs Control groups were significantly enriched in endocytosis, viral myocarditis, and type I diabetes mellitus. Additionally, the most significantly enriched pathways in the DEX vs DEX-CGA groups were protein digestion and absorption, RNA transport, and PPAR signaling pathway (Fig. 5C). MRM analysis was performed to validate the presence and levels of relevant proteins identified by proteomics. According to the KEGG results, EIF3J (accession number: Q5ZKA4) and EIF3H (accession number: Q5ZLE6) are involved in the MAPK signalling pathway. PROSC (accession number: E1C516) is involved in butanonate metabolism, APOA1 (accession number: P08250) is involved in PPAR signaling pathway, CHP1 (accession number: Q5ZM44) is involved in the apoptosis signaling pathway. The MRM results verified that EIF3J and EIF3H were downregulated (p < 0.05), whereas PROSC and APOA1 were upregulated (p < 0.05) by DEX treatment. Additionally, DEX-CGA treatment significantly downregulated (p < 0.01) APOA1 and CHP1 (Additional file: Figure S1).
Furthermore, protein-protein interaction (PPI) network was generated using the STRING database (Fig. 5D). The network diagram illustrates the interactions between the differentially expressed proteins in the screened pathways. Among the PPIs, CDK1 (accession number: F1NBD7) was the core PPI node in the DEX vs Control groups, with 8 interactions. IKZF1 (accession number: FINT33) was the core PPI node in the CGA vs Control groups, with 4 interactions. Moreover, CDK1 was the core PPI node in the DEX-CGA vs DEX groups, with 19 interactions.
The serum metabolic profiles of the broilers were altered by CGA
Broad-spectrum metabolomics was used to evaluate the serum profiles of the broilers. We observed a clear separation from the OPLS-DA score plots between the Control vs CGA groups, Control vs DEX groups, and DEX vs DEX-CGA groups (Fig. 6A). Differentially expressed metabolites between the groups were screened at a FC ≥ 2.00 or ≤ 0.50, which was illustrated using a heatmap (Fig. 6B). Compared with the Control group, CGA supplementation significantly increased the levels of 14 metabolites and decreased four metabolites, whereas DEX treatment significantly increased the levels of 37 metabolites and decreased the levels of 35 metabolites. Moreover, DEX-CGA treatment significantly increased 40 metabolites and decreased 16 metabolites compared with the DEX group (Fig. 6C). The top 20 metabolites with multiple differences between the groups are displayed in Fig. 6d. Compared to the Control group, CGA supplementation increased the levels of α-mercholic acid, phenylacetyl-L-glutamine, and cis-pentadecenoic acid and decreased the levels of 5’-deoxyadenosine, deoxyadenosine, and acetaminophen glucuronide. Additionally, DEX treatment increased the levels of α-mercholic acid, phenylacetyl-L-glutamine, and B-nicotinamide mononucleotide and decreased the levels of 20-carboxyarachidonic acid, stearidonic acid, and 9,12-octadecadienoic acid, compared with the Control group. Moreover, DEX-CGA treatment increased the levels of 3-(3-hydroxuphenyl) propionate acid, 2,4-dihydroxy benzoic acid, and homogentisic acid and decreased the levels of 23-deoxydeoxycholic acid, 2’-deoxyadenosine-5’-monophosphate, and carnitine C18:1-OH, compared with the DEX group. KEGG analysis showed that the differentially expressed metabolites in the Control vs CGA groups were enriched in purine metabolism, ABC transporters, and the cGMP-PKG signaling pathway. Differentially expressed metabolites in the Control vs DEX groups were enriched in tyrosine metabolism, biosynthesis of unsaturated fatty acids, and alpha-linolenic acid metabolism. Additionally, differentially expressed metabolites in the DEX vs DEX-CGA groups were enriched in riboflavin metabolism, tyrosine metabolism, purine metabolism, glutathione metabolism, and PPAR signaling pathway (Fig. 6E).
Effects of CGA on the PPAR and MAPK signaling pathways
Proteomic and metabolomic analyses revealed that CGA plays an important role in PPAR signaling pathway. Additionally, MRM analysis showed that CGA participates in the regulation of the MAPK signaling pathway. Thus, Western blotting was used to examine the effect of CGA on the activation of PPAR and MAPK signaling pathways. The results showed that DEX significantly decreased (p < 0.05) p-PI3K, p-JNK, P-38, p-P38, and ERK expression. In contrast, CGA treatment significantly increased (p < 0.05) JNK, p-JNK, P-38, and p-P38 expression. Additionally, CGA supplementation (CGA × DEX) significantly reversed DEX-induced decrease (p < 0.05) in JNK, p-JNK, P38, and p-P38 (Fig. 7A). Regarding the PPAR signaling pathway, DEX treatment did not significantly affect (p < 0.05) PPAR expression, whereas CGA supplementation significantly downregulated (p < 0.01) PPAR expression. Additionally, DEX × CGA interaction significantly increased (p < 0.01) PPAR expression (Fig. 7B).
Crosstalk between gut microbiota, SCFAs, and biochemical parameters
Spearman’s correlation analysis was performed to identify the relationships between biochemical parameters and differential gut bacteria, proteins, and metabolites (Fig. 8). A total of 4 bacterial genera were common between the Control vs DEX and DEX vs DEX-CGA groups (Fig. 8A). Based on this, parameters with correlation coefficient (r) > 0.70 or < -0.70 and p-value < 0.01 were selected. Among the 4 genera, Mordavella was positively correlated (p < 0.01) with villus height (r = 0.87), and negatively correlated with jejunal CXCL1 level (r = -0.71) and serum IL-6 level (r = -0.71). Coprobacter was negatively correlated (p < 0.01) with jejunal IL-18 (r = -0.87) and IL-12 levels (r = -0.70) and positively correlated (p < 0.01) with serum IL-4 level (r = 0.80) and IL-10 transcription (r = 0.71). Clostridiales_vadinBB60_group_unclassified was negatively correlated (p < 0.01) with serum IL-18 (r = -0.74), jejunal IL-18 (r = -0.74), serum CXCL2 (r = -0.71), jejunal CXCL2 (r = -0.73), serum CXCL1 (r = -0.73), and jejunal IL-12 levels (r = -0.71). Additionally, seven bacterial genera were common between the Control vs CGA and CGA vs DEX-CGA groups; however there was no significant correlation between the different genera and the biochemical parameters under the screening condition (r < -0.70 or r > 0.70). Regarding the correlation between biochemical parameters and SCFAs, results with r > 0.80 or < -0.80 and p < 0.01 were selected. A total of four SCFAs were correlated with biochemical parameters in the Control vs DEX and DEX vs DEX-CGA comparison groups, among which acetic acid was negatively correlated (p < 0.01) with D-LA level (r = -0.83), jejunal IL-6 level (r = -0.89), IL-18 transcription (r = -0.82), and serum CXCL1 (r = -0.80). Butyric acid was negatively correlated (p < 0.01) with jejunal IL-6 levels (r = -0.85), jejunal IL-22 levels (r = -0.81), and serum IL-18 levels (r = -0.81). Additionally, valeric acid was negatively correlated (p < 0.01) with jejunal IL-12 levels (r = -0.82), whereas isovaleric acid was negatively correlated (p < 0.01) with serum IL-1β levels (r = -0.82) and IL-1β transcription (r = -0.81). Regarding the Control vs CGA and CGA vs DEX-CGA comparison group, there were no significant correlations between the parameters under the screening condition (r < -0.80 or > 0.80) (Fig. 8B).
Crosstalk between proteomic and biochemical parameters
A total of 15 proteins were common between the Control vs DEX and DEX vs DEX-CGA comparison groups. Based on this, parameters with r > 0.85 or < -0.85 and p-value < 0.01 were selected. Regarding the Control vs DEX and DEX vs DEX-CGA comparison groups, LGMN (accession number: E1C958) was negatively correlated (p < 0.01) with D-LA (r = -0.87), and serum IL-12 (r = -0.98), CXCL2 (r = -0.92), and CXCL1 levels (r = -0.90). MEP1A (accession number: A0A1D5P6N4) was positively correlated (p < 0.01) with IL-10 transcription (r = 0.98) and serum IL-10 levels (r = 0.87), and CDK1 was positively correlated (p < 0.01) with Caspase-9 transcription (r = 0.93) and negatively correlated (p < 0.01) with serum IL-10 levels (r = -0.90). NASP (accession number: A0A3Q2UF99) was positively correlated (p < 0.01) with Caspase-9 transcription (r = 0.93) and negatively correlated with serum IL-10 levels (r = -0.87). Additionally, MEP1B (accession number: A0A1L1RS59) was positively correlated (p < 0.01) with serum IL-10 levels (r = 0.93), and ANO5 (accession number: F1NN74) was positively correlated (p < 0.01) with IL-10 transcription (r = 0.93). TMSB4X (accession number: R4GF71) was positively correlated (p < 0.01) with Caspase-9 transcription (r = 0.93), serum CXCL1 levels (r = 0.92), and serum IL-18 levels (r = 0.87). NOC2L (accession number: F1NV71) was positively correlated (p < 0.01) with serum IL-22 levels (r = 0.92), Dab1 (accession number: Q6XBN7) was positively correlated (p < 0.01) with IL-10 transcription (r = 0.90), and MARCKS (accession number: A0A1D5PDE6) was negatively correlated (p < 0.01) with IL-10 transcription (r = -0.87). Additionally, MGME1 (accession number: A0A1L1RXX7) was positively correlated (p < 0.01) with IL-18 transcription (r = 0.87), PROSC was positively correlated (p < 0.01) with serum IL-10 levels (r = 0.87). Furthermore, 4 proteins were common between the Control vs CGA and CGA vs DEX-CGA comparison groups, among which GAS8 (accession number: F1NLA8) and APOC3 (accession number: A0A1D5PK48) were negatively correlated (p < 0.01) with serum IL-18 levels (r = -0.93 and r = -0.90, respectively) (Fig. 8C).
Crosstalk between metabolomic and biochemical parameters
A total of 25 metabolites were common between the Control vs DEX and DEX vs DEX-CGA groups, and parameters with r > 0.80 or < -0.80 and p < 0.01 were selected. Among the 25 metabolites, α-mercholic acid was negatively (p < 0.01) correlated with V/C ratio (r = -0.87), villus height (r = -0.82), and jejunal IL-4 level (r = -0.90), and positively correlated with jejunal CXCL1 levels (r = 0.84) and serum IL-6 level (r = 0.84). Additionally, 7,8-dihydro-L-biopterin was positively correlated (p < 0.01) with villus height (r = 0.83), whereas Asp-Phe was negatively correlated (p < 0.01) with villus height (r = -0.81) and positively correlated with IL-1β transcription (r = 0.87) and jejunal IL-12 (r = 0.86) and CXCL2 levels (r = 0.81). Glycyl-l-proline was negatively correlated (p < 0.01) with jejunal CXCL2 levels (r = -0.81), whereas 2,4-hexadienoic acid was negatively correlated (p < 0.01) with IL-4 transcription (r = -0.86). Moreover, (±)5-HETE (r = 0.85), (±)9-HETE (r = 0.85), and LTE4 (r = 0.84) levels were positively correlated (p < 0.01) with jejunal CXCL2 levels. Additionally, uracil was negatively correlated (p < 0.01) with IL-4 transcription (r = -0.84). Furthermore, only 1 metabolite was common between the Control vs CGA and CGA vs DEX-CGA groups; however, there was no significant correlation between the metabolite and the biochemical parameters (Fig. 8D).