2.1 Antibiotics increased the growth of chicken
The average daily gain (ADG = (the weight of day 42nd- the weight of day 1nd) / 41) in AMX + PMX groups (up to 25.9%) were highest among other antibiotic groups (14.3-23.4%) within 42 days (Supplemental Figure 1A). For feed conversion ratio (FCR = Live gain / feed consumption), the antibiotic groups especially AMX + PMX1 group showed a marked increased (17.8-34.2%) compared with the control group (P < 0.01) (Supplemental Figure 1C). No significant has been observed in food intake among all groups (3.2%–21.1%) (Supplemental Figure 1B). Therefore, the combination of AMX + PMX was employed to further experiment (Figure 1A). The results showed that at 400 mg/kg of concentration, the growth of chickens reached to the maximum compared with that of the control group (4.9%). However, after 800 mg/kg of AMX + PMX intervention, the weight was significantly decreased 2.5% than that in the control group (Figure 1B). It showed that the weight gain was increased with a certain dose of antibiotic intervention.
Antibiotics have been used as growth promoters at a subtherapeutic concentration in animals [38, 39]. Chlortetracycline and salinomycin (500 g/ton) produced a 9.8 % higher ADG than the control [40]. Oxytetracycline (100 mg/kg) and florfenicol (5 mg/kg) also showed significant increase in the final body weight, which is consistent with obtained results [41]. However, the antibiotic group exhibit 10.6 higher ADG than control, low FCR than the control, which are unsimilar with early report [40]. These data indicated that AMX and PMX consider as a growth promoter for chicken broiler. The attribution of improvement of growth and feed efficiency as a result of dietary antibiotic supplementation is not clear, and needs further investigation [42].
2.2 Health index, inflammation markers and oxidative stress indicators on chicken decreased with antibiotics
The levels of AST, ALT, MDA, CAT and T-SOD were tested, which were similar with the control group (P < 0.01) (Figure 2A-E, Supplemental Figure 1). Histopathological analysis showed that a slight nuclear shrinkage in the liver of chicks were observed after being treated with 400 g/kg of AMX + PMX (Figure 2F). Furthermore, under microscopical examination, hematoxylin-eosin staining of the adipose tissue sections in the antibiotic-treated chicken revealed that the distribution of mast cells was significantly more than that of the control group (Figure 2 G). The results showed that in the 35 days, both the 200 and 400 groups produced a 19.0 and 16.3% lower the ratio of liver/body weight than the control. There was no significant difference in the immune organ indexes of thymus, bursa and spleen (control group) as compared to the groups, only a slight decrease in chicken treated with 200 and 400 mg/kg of AMX and PMX (Figure 2I). The expression levels of IgG and IgA in the 200 group were remarkable lower (30%, 10%) than those in the control group, (Figure 2J-K).
Oxidative stress indicators (AST, ALT, T-SOD, MDA and CAT), which indicated tissue impairment caused by stress, toxicity and liver damage produced by antibiotics, revealed no significant effects occurring [25, 43, 44]. The biochemical results corroborated the demonstration to the histopathological finding of both liver and adipose tissue, which further showed that an increase in mast cells in visceral fat obesity was accompanied with antibiotics [45]. Antibiotic-treated mice were decreased liver/body weight, which is similar with ours [46]. Induction and maintenance of an appropriate level of immunological activity and inflammation markers is crucial for healthy broiler growth in poultry farms [47]. Serum IgG and intestinal secretory IgA were compared between groups (Serum IgG reflected the immune state of the system, and intestinal secretory IgA reflected the immune state of the system and the intestine, respectively) [48]. The AMX + PMX decreased colon mucosal secretory IgA (27.5%–28.6%) and serum IgG (18.2%–32.7%) concentration, which is similar with the chlortetracycline and salinomycin effect on IgA and serum IgG (15.7%)[40]. Moreover, the immune organ indices of the thymus gland and bursa indicated a weaken immunity, which was decreased 12.5% and 5.3% in 200 group and–0.57% and 4.8% in the 400 group. For example, salmonid specific and nonspecific immune responses were also depressed by the administration of tetracycline at 10 mg/ kg diet [49]. It indicated that under antibiotics, which is perhaps via reducing the nutrient consumption required for maintaining immunological activity [50, 51].
2.3 Antibiotics decreased the number and changed the structure of the intestinal microflora
The fecal flora of chicken was detected in 1, 9, 16 and 23 days (Figure 3A-D). The result showed that with the increase of antibiotic concentration, the number of gut microbiota formed a concentration gradient, and it researched the lowest level on the 400 and 800 mg/kg. Simultaneously, qRT-PCR quantitation of chickens’ colonic microflora on the 35nd day showed that the content of gut microbiome was declined with the increase of antibiotic concentration, and the change was similar with fecal flora (Figure 3E). 16S rRNA gene sequencing was performed using the chicken feces collected after 35 days of antibiotic treatment. By significance tests for differences in α diversity, the gut microbiota of antibiotic mice featured an increase in Shannon index (Figure 3F). This may be due to the antibacterial effect of antibiotics, which promotes the reproduction of some nonmajor bacteria in the intestinal tract. PCoA showed that with the increase of antibiotic concentration, the distance between the intestinal flora distribution and the control group increased gradually (Figure 3G).
The antibiotic-induced difference in microbiota composition was illustrated in Figure 3. Firmicutes, Proteobacteria and Bacteroidetes species were dominant in treatment groups, as was the microbiome from the control group. It had far fewer sequences from the Firmicutes and Bacteroidetes phyla, and the microbiome of antibiotic chicken had a compositional shift to Proteobacteria (Figure 3I). The ratio of Firmicutes to Bacteroidetes (Firm/Bac ratio) was decreased upon antibiotic exposure (Figure 3H). A few bacterial genera have been identified that were significantly different in between antibiotic and control groups (Figures 3J). Lactobacillus and Bifidobacterium are widely approved probiotic genera with extensive health-promoting and immunomodulatory properties[52, 53]. Lactobacillus, Enterococcus and Romboutsia showed lower community than those in the control group (20%), while the contents of Bacteroides was significantly higher in the group treatment with 800 mg/kg of AMX+PMX as compared to that in the control group (354.67%). Additionally, few other genera including Lachnospiraceae, Clostridiales, Oscillibacter and Roseburia were altered by calorie absorption. Together, these data suggested that the total and architecture of gut microbiota had a slight altered by antibiotic intervention, likely contributing to the metabolic benefits of calorie absorption.
The chicken gastrointestinal tract is home to an ecosystem rich in microbial biodiversity, playing home to ≥ 500 phylotypes or ~1 million bacterial genes, which is integral in multiple physiological processes of the host, including being a key factor involved in host metabolism, body weight and energy homeostasis [54, 55]. Symbiotic microbes that receive their nutrition from animal and, in turn, contribute essential nutrients and play a role in immune defense [56]. Increasing evidence showed that the nutritional value of food is influenced by the structure and operation of the gut microbial community [57]. Firmicutes (Gram positive), Bacteroidetes (Gram negative), and Actinobacteria (Gram positive) represent over 90% of the phyla and dominate the gut microbiota [58]. The alteration of the proportion of Firmicutes and Bacteroidetes changes the nutrient load in the gastrointestinal tract [59]. For example, vancomycin treatment reduced the relative abundance of Firmicutes from 37% to 50% and increased in the relative abundance of Proteobacteria,which is consistent with our previous results [60]. A research found that there is a positive correlation between ratios of Bacteroidetes to Firmicutes and the plasma glucose concentration [61]. The conversation of Firmicutes and Proteobacteria phyla was decreased 29.5%-83.4%, which accelerate catalyze the conversion of choline and the progression of T2D [62]. Additionally, antibiotic such as avilamycin, reduced overall community diversity and certain bacterial species (Lactobacillus and Bifidobacterium) and predominantly alter the ileal microbiota (lactobacilli dominant) [63, 64]. The growth of Bifidobacterium is able to increase probiotic genera such as Kurthia, Lactobacillus and Bifdobacterium and meanwhile decrease harmful bacteria such as Pseudomonas (Supplemental figure 2). Lactobacillus species in the colon is crucial to produce specific fermentation products (i.e., short-chain fatty acids, predominantly acetate, propionate and butyrate) to mediate the host metabolic health [65].
2.4 The role of antibiotic and metabolomic correlation of intestinal microbiota
The fecal metabolite levels in chicken administered with different concentrations of AMX+PMX or were untreated (control) have been investigated. About 14,957 metabolic features was observed after untargeted LC-MS analysis and 5,861 features were annotated by following the standard filter criteria (features detected in < 70% of the QC samples were removed). PCA and PLS-DA indicated that the antibiotic treatment group and the control group could be separated with the increase of antibiotic concentration, and the metabolites of the group with 800 mg/kg of AMX+PMX and the control group were significantly different (Figure 4A, B). Proteobacteria, Actinobacteria, Firmicutes, and Acidobacteria were primary bacteria of gut microbioma. Regulatory 44 metabolisms were classified to phylum, among them 10 were related to Deinococcus-Thermus, 1 was related to Bacteroidetes, 1 was related to Actinobacteria, and 32 were related to Opisthokonta (Figure 5, Supplemental file 2).
The gastrointestinal tract represents the interface between ingested nutrients and the host where energy is effectively extracted [66]. In healthy individuals, indigestible carbohydrates and proteins that enter the colon represent between 10% and 30% of total ingested energy and were digested by colonic microbiota[66, 67]. Meanwhile, it could be expected to contribute anywhere from 6% to 22% of daily caloric turnover[68]. A large remove of colonic microbiota can bring down the metabolic rate and loss of the “metabolic buffering” function of the gut microbiota, leading to abrogation of CR-mediated body weight loss [69]. Proteobacteria and Actinobacteria were abundant in liver injury animals[70]. Bacteroidetes are among the major members of the microbiota of animals, which are increasingly regarded as specialists for the degradation of high molecular weight organic matter, proteins and carbohydrates [71]. Some Lactobacilli possessing lipolytic activities produce significant amounts of fatty acids with antimicrobial potential under specific conditions[72, 73]. Hence, the effect of antibiotics on the intestinal flora structure and quantity may change the produces and releases an enormous array of compounds which may further act upon host tissues modulating appetite, gut motility, energy uptake and storage, and energy expenditure.
2.5 Antibiotic treatment enhanced lipid and amino acid metabolism pathways of gut microbiome associated with obesity
Compared with the control group, 794 differential metabolites were screened in the antibiotic groups, of which 235 metabolites were significantly downregulated (P < 0.05), and 559 metabolites were significantly upregulated (P < 0.05) (Supplemental figure). According to the material information, significant differences have been observed in the top 10 pathways of different metabolites using the KEGG (P < 0.05) (Figure 4C). The top 5 as regards impact are tryptophan metabolism, ABC transporters, aminoacyl-tRNA biosynthesis, vitamin digestion and absorption, protein digestion and absorption and those linked to the immune system. Some unique metabolites, 7 amino acids (D L-(-)-Threonine, L-Phenylalanine, DL-Lysine, L-Isoleucine, Valine, Amidinoproline and Hydroxylysine), 3 steroids (tyramine sulfate and 3-(2-Acetamidoethyl)-1H-indol-5-yl hydrogen sulfate), 3 hormones (adrenaline, hexoprenaline and kinetin) and some metabolites, including 5-Hydroxy-DL-tryptophan, Fructoselysine, N6-Capryloyl lysine, dihydroxyphenylalanine, indole, Dimethyl 22'-azobis(2-methylpropionate), threonylphenylalanine, carbobenzoxyglycylphenylalanine, 2-methylhippuric Acid and methylhippuric acid overlapped among 5 groups (p < 0.05). Most of amino acids were increased in the antibiotic groups, including DL-(-)-Threonine (57.7%-141.0%), L-Phenylalanine (264.3%-149.1%), 5-Hydroxy-DL-tryptophan (196.8%-269.7%), Fructoselysine (250.3%-519.7%), DL-Lysine (49.8%-187.9%), L-Isoleucine (24.3%-64.7%). Most of them increased in the 400 group, but came close to the control group which were treated with 800 mg/kg (Figure 6).
In serum metabolism, the differential metabolites were screened using hierarchical cluster analysis (Supplemental figure 5). The KEGG enrichment analysis of the differential metabolites showed that the influence was related to the biosynthesis of unsaturated fatty acids, glycerophospholipid metabolism, retinol metabolism pathways and choline. Moreover, products including 1-arachidonoyl-sn-glycero-3-phosphocholine, 19-hydroxycholest-5-en-3-yl acetate, Vitamin A, were significantly decreased in the 400 group (Figure 7). The lipidomic further verified the effects of antibiotics in lipid metabolism. The key metabolism productions of these pathways, such as Lipolysis, TAG and cholecalciferol (Vitamin D) were changed in antibiotic group, especially in 400 group. Specific metabolites related to lipid metabolism increased when antibiotic intervention, including palmitoylcarnitine and octyl phosphate. The insulin resistance pathway was also regulated by antibiotic intervention (Supplemental Figure 6). Besides, antibiotics induced a lower fasting blood glucose level compared to that in the control group with a 16-h fast blood glucose level. However, there is a significant difference was found between fasting and fed glucose in 200 and 400 groups an hour after eating (Figure 7E, F).
The microorganisms in the gastrointestinal tract play a significant role in nutrient uptake, vitamin synthesis, energy harvest, inflammatory modulation, and host immune response [74]. For example, the weight gain occurs even when energy intake decreasd by 30% compared to common mice which remained germ-free [75]. Dysregulated fatty acid and amino acid metabolism are reported in adults with type 2 diabetes [76]. Hydroxytryptophan can help the body balance on weight, which is at a low level in obesity [77, 78]. Diabetic rats increased the content of phenylalanine (65.6%), valine (29.8%), methionine (29.6%) and aspartic acid (44.6%), but tyrosine (48.6%), alanine (16.8%), lysine (13.2%), threonine (41.4%) and histidine (19.1%) content were reduced, which is similar with our research [79]. Tryptophan / kynurenine which could confer benefits of feed utilization, body composition and antioxidative capacity, showed an upward trend (26.6%) in weight in 200 group. However, it began to decrease (64.4%) when the concentration of AMX+PMX reached 400 mg/kg [80]. It showed that amino acid levels with a significant change in body adiposity, which may be due to increased utilization of host or decreased production of amino acids by bacteria [81]. Meanwhile, some metabolites associated with the digestion and absorption of nutrients have similar response with obesity. Lysolecithin could confer benefits of feed utilization, body composition and antioxidative capacity of the channel [80]. Oxalosuccinic acid and phosphatidylcholine were reduced in serum levels of diabetes [82]. Decreased levels of 2-methylhippuric Acid and methylhippuric acid have also been related to obesity [83]. Acetyl-L-carnitine, as a long-chain fatty acid derivative, reversed the inhibition mediated by carnitine to slow the transport of fatty acids into the mitochondrial matrix where they are used for energy production [84].
Another feature of the dysbiosis in the microbiota metabolism was the levels of some health-related hormones and purine metabolites. Uric acid and hypoxanthine, as the products of the microbial metabolism of xanthines to alter gut microbiota in the presence of insulin resistance, showed an increase (41.7%-54.7% and 22.2%-120.0%, respectively) [85]. Adrenaline, the main effectors of the sympathetic nervous system, are thought to control adiposity and energy balance through several mechanisms. Kinetin induced mitotic divisions throughout the digestive tract. Both of them were high expressed in the 400 group (241.9%, 87.5%), which is different from hexoprenaline (16.1%), which may play a role in chicken weight gain. All these fluctuated were similar with the obesity, indicating that antibiotic treatment on gut microbiota largely abolishes the regulation of metabolism [86].
2.6 Antibiotic intervention altered the immune response of chicken
The alterations in the serum associated with immunity have been compared. In serum, phosphocholines changed significantly in the antibiotic groups (P﹤0.01). The results revealed an increase in phosphocholines concentration at a concentration of 200 mg/kg, returning to the lowest level at 800 mg/kg concentration, and its related derivatives (lysophosphatidylcholine and 1-arachidonoyl-sn-glycero-3-phosphocholine) have also changed (Figure 7B, H). Some unsaturated fatty acids, such as 4-phenylbutyric acid, docosahexaenoic acid, elaidolinolenic acid, pinolenic acid, L-pipecolic acid and ricinoleic acid, were fluctuated with the concentration of AMX+PMX, especially in the 800 group (Figure 8).
The immune system restricts the overgrowth of symbiotic bacteria and prevents external bacteria from entering the host internal organs[87]. Gut microbiota is necessary for the proper immunological development of the intestine and the host which is associated with proinflammatory and host immune responses that are alter the growth performance, and the change trend of metabolites is conducive to weight gain [51]. As shown in germ-free animals, which have underdeveloped mucosal immunological tissues compared to conventional animals [88]. The kynurenine pathway is the major route for tryptophan metabolism. The rate of Tryptophan/kynurenine is an important factor in the disease development, such as cardiovascular, which is significant increase with 400mg/kg of AMX+PMX (27.5%). Meanwhile, oxidative stress and immune activation as inflammation factors, showing that after antibiotic intervention, the immune response of chicken was inhibited. The changes in lipid metabolome found here display dissimilarities to some other systems involving viral infection and inflammation. For example, lysophosphocholines as a factor in immunoregulation, originate primarily in the liver and are released from larger lipids [89]. Betaine protects cells, proteins, and enzymes from environmental stress, which participates in the methionine cycle and is an important nutrient for the prevention of chronic diseases [90]. Sphingosine and sphinganine N-acyltransferase plays an important role in the pathogenesis of animal diseases,showed no significant difference in treatment groups and control group (Figure 8B-D) [91].
Significant amounts of fatty acids with antimicrobial potential under specific conditions were produced [72, 73]. For example, 4-phenylbutyric acid can suppress oxidative stress by attenuating endoplasmic reticulum stress to provide renoprotection [92]. Docosahexaenoic acid is a long-chain polyunsaturated fatty acid that worked as a ligand for the retinoid X receptor in brain [93]. Pinolenic acid, a naturally-occurring polyunsaturated fatty acid, inhibited cell metastasis by suppressing cell invasiveness and motility [94]. L-pipecolic acid was accumulated by Escherichia coli cells and protected them while growing at inhibitory osmolarity, which is able to bind the periplasmic protein, while this protein was necessary for their uptake [95]. Ricinoleic acid is the compound of antineoplastic and immunomodulatory characteristics [96]. With the increase of antibiotics, the content of all these unsaturated fatty acids showed a downward trend, which is coincident with the decrease of acidogenic bacteria (Lactobacilli et al), as shown in obesity [97]. Antibiotics restrain gut microbiota, resulting in reduced selective pressure on the host’s immune system, thus leading to the excessive growth of chickens[87]. Therefore, in the present study, the weight alteration might be due to the changes in the intestinal community and impaired host immunity.
Childhood adiposity has increased significantly over the past several decades, which has become a major challenge to public health worldwide[98]. An obvious relationship was found between antibiotic exposure in early life with a 6% increment in the risk of obesity[46]. The growth of broiler growth with the early intervention of antibiotics is the combination of the following three effects here (suppressing the growth of gut bacteria, inhibiting host immune response and regulation the metabolic pathway of intestinal flora): (i) suppressing the growth of indigenous gut bacteria and reducing its energy consumption, which results in more nutrients for chicken for greater weight gain [99], (ii) inhibiting the health-beneficial effect of the gut microbiota and host immune response by reducing the content of immunoglobulin and inflammation related factors [100, 101], and (iii) regulation of the metabolic pathway of intestinal flora including lipid , amino acid metabolism and immunity [25]. The intestinal flora is likely to have a significant impact on host physiological processes with early intervention of antibiotics, which may be similar to the factors of early childhood obesity, and it can provide a theoretical reference for the early obesity exploration of humans.