Carbohydrates as the most abundant and cheapest energy sources for fish are incorporated in aquatic feeds for the potential protein-sparing effect(Reference Lin, Shi and Mu1–Reference Shiau and Peng3). Furthermore, carbohydrates can also act as stabilisers and swelling agents in diets, and appropriate concentration of dietary carbohydrates could reduce pathogen susceptibility and cause lesser environment pollution(Reference Li, Liu and Lu4).
In order to produce cost-effective aquafeeds, more and more competitively priced carbohydrates have been incorporated in feed formulation to substitute expensive marine ingredients(Reference Li, Hu and Qiao5). However, most fish species are not good at utilising carbohydrate. Excessive levels of carbohydrate in diets may induce metabolic load and stress responses, thus cause negative effects on health and increase the susceptibility to infectious diseases of fish. For instance, it has been found that high-carbohydrate diet depressed feed intake and feed utilisation efficiency, leading to poorer growth performance in gibel carp (Carassius auratus var. gibelio)(Reference Tan, Wang and Xie6). Furthermore, high levels of carbohydrate in the diet could induce negative effects on metabolic homoeostasis and immune status. It has been reported that excessive carbohydrates could increase the lipid content of fish body in Nile tilapia(Reference Boonanuntanasarn, Jangprai and Kumkhong7) and cause long-lasting hyperglycaemia in rainbow trout (Oncorhynchus mykiss)(Reference Felip, Ibarz and Fernández-Borràs8). Suppressed immune response and reduced antioxidant reactions were also found in largemouth bass (Micropterus salmoides) and Labeo rohita (Hamilton) when they were fed with high-carbohydrate diet(Reference Gou, Chen and Xiangtai9,Reference Kumar, Sahu and Pal10) . How to alleviate the negative impacts caused by high-carbohydrate diet has attracted more and more attention. It has been reported that supplementation of bile acids in high-carbohydrate diet could improve growth performance, alleviate metabolic syndromes and enhance immune ability in largemouth bass(Reference Guo, Kuang and Zhong11,Reference Yu, Zhang and Chen12) , suggesting that dietary additives supplementation have the potential to alleviate negative effects caused by high-carbohydrate diet.
Prebiotics, which are non-digestible ingredients fermented by intestinal microbiota, have been widely used in aquaculture as dietary additives(Reference Ganguly, Dora and Sarkar13). Prebiotics could change gut microbiota composition to improve growth condition and immune ability or reduce risk of diseases(Reference Gibson, Hutkins and Sanders14). Currently, the most common prebiotics used in aquaculture are oligosaccharides including inulin, fructo-oligosaccharides, short-chain fructo-oligosaccharides, mannan oligosaccharides, trans-galacto-oligosaccharides, galacto-oligosaccharides, xylo-oligosaccharides, arabinoxylo-oligosaccharides and isomalto-oligosaccharides(Reference Ringø, Dimitroglou and Hoseinifar15).
Inulin is classified as fructans or fructose polymers, and it naturally exists in many plants especially in chicory. In aquaculture, the studies have been conducted to detect the influence of inulin on growth and immune status(Reference Tiengtam, Khempaka and Paengkoum16,Reference Tiengtam, Paengkoum and Sirivoharn17) , but whether inulin could alleviate high-carbohydrate diet-induced metabolic syndrome remains unknown. A growing body of evidences showed that inulin had potential to alleviate glucose or lipid metabolism disorders in mammals, and inulin was also suggested to have the ability to alter gut microbiota composition and promote the growth of probiotics in mammals. For example, the administration of inulin decreased fasting blood glucose levels and improved insulin resistance which is correlated with the increase of Anaerovorax in rats(Reference Zhang, Yu and Xiao18). In addition, inulin could alleviate the lipid accumulation in adipocytes induced by high-fat diet, and it could modulate gut microbiota composition by decreasing abundance of Roseburia spp. and Clostridium cluster XIVa(Reference Dewulf, Cani and Neyrinck19). Evolutionarily, fish are more primitive than mammals and fish harbour a Proteobacteria-dominated microbiota, which is different from the dominant microbiota in mammals(Reference Kostic, Howitt and Garrett20), so whether inulin could regulate fish metabolism by regulating gut microbiota remains unknown.
Nile tilapia (Oreochromis niloticus) are an economically important fish species and is an ideal fish model for nutritional and metabolic studies due to the fast growth rate, high resistance to diseases and available genomic information(Reference He, Ning and Chen21). Although Nile tilapia have higher capacity for carbohydrate utilisation, high concentration of carbohydrate in the diet also caused metabolic syndrome in Nile tilapia, including hyperglycaemia and fat deposition(Reference Boonanuntanasarn, Jangprai and Kumkhong7,Reference Liu, He and Ning22) . In the present study, we hypothesised that inulin could alleviate metabolic syndrome induced by high carbohydrate by regulating intestinal microbiota in Nile tilapia. Fish were fed with three diets and referred as CON (35 % carbohydrate), HC (45 % carbohydrate) and HCI (45 % carbohydrate + 5 g/kg inulin) for 10 weeks. The influence of inulin on Nile tilapia fed with high-carbohydrate diet, including survival rate post Aeromonas hydrophila infection, metabolic characteristics and intestinal microbiota, was detected.
Materials and methods
This research was approved by the Committee on the Ethics of Animal Experiments of East China Normal University. All experiments were conducted under the Guidance of the Care and Use of Laboratory Animals in China (no. F20190101).
Experimental design
Juvenile Nile tilapia were collected from Shanghai Ocean University (Shanghai, China). During the 2-week acclimatisation period, Nile tilapia were fed on a commercial diet (Tongwei Co. Ltd) at a feeding rate of 3 % body weight. A total of 270 visually healthy fish (1·19 (sd 0·01) g) were divided into three treatments and fed with control diet (35 % carbohydrate, CON), high-carbohydrate diet (45 % carbohydrate, HC) and inulin-addition diet (45 % carbohydrate + inulin (5 g/kg) diet, HCI) (online Supplementary Table S1). Each treatment contained three tanks (thirty fish in each tank). Fish were hand-fed twice daily (09.00 and 18.00 hours) for 10 weeks at a feeding ratio of 4 % body weight. Feeding consumption was adjusted every 2 weeks following a 24-h starvation period and batch weighing. Water temperature was maintained at 27 (sd 1)°C with a natural photoperiodicity (12 h light–12 h dark). During the experimental period, dissolved oxygen and pH were measured daily and kept at 5–7 mg/l and 7·5–8, respectively.
Dietary ingredients (online Supplementary Table S1) were mixed with oil and distilled water to form a paste and then extruded through a Twin-Screw Extruder (TSE65; Yanggong Machine) to produce 2-mm-diameter pellets. The proximate compositions were calculated including crude lipid, crude protein and ash which are showed in online Supplementary Table S1. The prepared diets were dried at room temperature and then stored at −20°C until use.
Sample collection
At the end of the trial, all fish were deprived of feed for 12 h before sampling and twenty-one fish were randomly selected from each group (seven fish per tank). Six fish were immediately stored at −80°C to analyse body composition. Fish were euthanised with MS-222 (22 mg/l, tricaine methanesulfonate; Western Chemicals Inc.) and dissected to obtain tissues for biochemical and molecular biological analyses. Gut content was collected and promptly frozen at −80°C for SCFA and intestinal microbial composition analysis.
Growth performance
Weight gain of all fish in each tank was measured, and nine fish per diet (three fish per tank) were collected to detect the viscera index, hepatosomatic index and mesenteric fat index which were calculated according to the formulae:
$$\displaylines{
{\rm{Weight}}{\mkern 1mu} {\rm{gain}}{\mkern 1mu} {\rm{rate}}{\mkern 1mu} \left( \% \right) = {\rm{100}} \times ({\rm{final}}{\mkern 1mu} {\rm{fish}}{\mkern 1mu} {\rm{weight}} - {\rm{initial}}{\mkern 1mu} {\rm{fish}}{\mkern 1mu} {\rm{weight}}){\rm{/}} \cr
{\rm{initial}}{\mkern 1mu} {\rm{fish}}{\mkern 1mu} {\rm{weight}} \cr} $$
$$\displaylines{
{\rm{Feed}}{\mkern 1mu} {\rm{conversion}}\;{\rm{ratio}}{\mkern 1mu} (\% ) = 100 \times {\rm{dry}}{\mkern 1mu} {\rm{feed}}{\mkern 1mu} {\rm{consumed}}/ \cr
({\rm{final}}{\mkern 1mu} {\rm{fish}}{\mkern 1mu} {\rm{weight - initial}}{\mkern 1mu} {\rm{fish}}{\mkern 1mu} {\rm{weight}}) \cr} $$
$${\rm{Viscera}}\,{\rm{index}}\,(\% ) = 100 \times ({\rm{viscera}}\,{\rm{weight/body}}\,{\rm{weight}})$$
$${\rm{Hepatomatic}}\,{\rm{index}}\,(\% ) = 100 \times ({\rm{liver}}\,{\rm{weight/body}}\,{\rm{weight}})$$
$${\rm{Mesenteric}}{\mkern 1mu} {\rm{fat}}{\mkern 1mu} {\rm{index}}{\mkern 1mu} (\% ) = 100 \times ({\rm{mesenteric}}{\mkern 1mu} {\rm{fat}}{\mkern 1mu} {\rm{weight/body}}{\mkern 1mu} {\rm{weight}})$$
Biochemical analysis
Six fish from each group were collected for body composition assay. The content of moisture was measured via oven drying to a gravimetric analysis at 105°C for 12 h. Then, whole-dried fish were grounded by pulverizer (Xulang Machinery Equipment Co. Ltd) and stored in a glass desiccator at room temperature for analysing total protein and total lipid content analyses. Total protein of the whole body was determined by Kjeldahl method using a semi-automatic Kjeldahl System (Foss Tecator, Kjeltec™ 8200). Total lipid in whole fish body and in liver was calculated by a classic methanol–chloroform method(Reference Bligh and Dyer23).
Livers of six fish were collected from each treatment for TAG and glycogen analysis. The contents of TAG and glycogen in sampled tissues were assayed using commercial kits (Jiancheng Biotech Co.). All experiments were conducted according to the relevant kit protocols.
Histochemical staining
Liver samples of three fish per group were embedded in an optimum cutting temperature (OCT) compound (Sakura) at −40°C in a cryostat (Cryostar NX50™; Thermo Fisher Scientific Co. Ltd), and then tissues were cut into 5 μm thickness. Then, 5 μm sections were stained with Oil red O staining solution (O1516; Sigma-Aldrich) overnight. The stained samples were washed with decreasing concentrations of propylene glycol, followed by several rinses with PBS and stored at 80 % glycerol bath. Histology images were observed by a microscope (Nikon, Eclipse, TS100).
RNA extraction and quantitative real-time PCR
Total RNA of six fish per group was extracted from liver by using Tri Pure Reagent (Takara). RNA purity and integrity parameters were measured by Nanodrop 2000 (Thermo Fisher Scientific Co. Ltd) and electrophoresis. A260/A280 of RNA ranged from 1·9 to 2·2 and A230/A260 of RNA was >2. First-strand complementary DNA was synthesised using a FastQuant RT Kit with gDNase (TIANGEN) by S1000™ Thermal Cycler (BioRad). The primers were designed in NCBI for real-time PCR, and the sequences are listed in online Supplementary Table S2. Quantitative real-time PCR (qRT-PCR) includes 2·0 μl of cDNA template, 12·5 μl of 2 × SYBR qPCR Mixture (Aidlab), 2·0 μl of PCR primers (5 μm) and 6·5 μl of nuclease-free water. qRT-PCR was performed in a CFX96 Connect Real-Time System (Bio-Rad) with 95°C for 10 min, forty cycles of 95°C for 5 s and 60°C for 15 s. In order to ensure the specificity of the primers, melt curve of the amplified products was performed at all PCR reactions. The amplification efficiency was between 95 and 105 %, and the correlation coefficient was above 0·98 for each gene. The fold difference in mRNA expression was calculated by 2−ΔΔCT method(Reference Livak and Schmittgen24).
Glucose tolerance test
At the end of the trial, thirty fish from each group (ten fish per tank) were collected for a glucose tolerance test by an i.p. injection of d-glucose (500 mg/kg BW, 20 % in 0·85 % NaCl) (Sigma) after 12 h of fasting. Blood of six fish per group was drawn from the caudal vein of fish at 0, 0·5, 1·5, 3 and 6 h after injection. Blood glucose and inulin were measured using commercial kits (Jiancheng Biotech Co.).
Intestinal microbiota analysis
Gut content (100 mg) was collected from each individual fish of seven fish per treatment for 16S RNA sequencing. Non-template control was involved during DNA extraction and PCR amplification. Total intestinal bacterial DNA was isolated, and the V3–V4 region of the 16S rRNA gene was amplified by PCR using 338F (5′ ACTCCTACGGGAGGCAGCA 3′)/806R (5′ GGACTACHVGGGTWTCTAAT 3′). The quantity of PCR products was detected by a Microplate reader (Bio Tek, Flx800) with a Quant-iT PicoGreen dsDNA Assay Kit (Invitrogen). PCR products were subjected to Illumina MiSeq/NovaSeq, generating paired-end reads (Shanghai Personal Biotechnology Co. Ltd). Microbiota sequences were available in Sequence Read Archive with the accession number PRJNA612370. The Quantitative Insights Into Microbial Ecology (QIIME, v2) pipeline was performed for quality filter, denoising and chimaera check(Reference Caporaso, Kuczynski and Stombaugh25). All the unique sequences were clustered at 98 % followed by chimaera removing. The non-chimaera sequences were re-clustered at 97 % to generate an operational taxonomic unit (OTU) table. α-Diversity indices including Chao1, Simpson and Shannon indexes were calculated by QIIME. The classification at phylum level of species composition was analysed by QIIME. Principal component analysis and heat map analysis were performed using R software. Fourteen OTU were picked for heat map analysis based on the following: (1) the abundance of these OTU was higher than 1 % at least in one group and (2) the abundance of these OTU was significantly different between CON and HC or HC and HCI.
SCFA analysis
Gut content (200 mg) of three fish per treatment was suspended with 0·5 ml distilled water. 0·4 ml 50 % sulphuric acid was added to the mixture for acidifying. Then, SCFA were extracted with 1 ml diethyl ether. The samples were centrifuged at 12 000 g for 10 min, and the supernatants were collected for the detection. The analysis was conducted with GC by using GC7900 (TianMei Scientific Instrument). The temperature of the injector port and detector was 180°C. The oven temperature increased from 80 to 155°C at the rate of 5°C/min and hold on 80°C for 2 min. The content of SCFA was measured according to the external standard curve.
Aeromonas hydrophila challenge trial
To investigate whether inulin could protect Nile tilapia fed with high-carbohydrate diet from pathogen infection, thirty fish were randomly collected from each group (ten fish per tank) and challenged with A. hydrophila which was cultured in a Luria Broth at 37°C for 16 h. A. hydrophila was collected and washed with PBS. The quantity of bacteria was counted by a standard dilution and plating method. The bacteria were added to cultural water and diluted to the density of 9·2 × 105 colony-forming units/l. Mortality of Nile tilapia was recorded per d, and the challenge assay lasted for 11 d.
Statistical analysis
Data were presented as mean values with their standard errors. Normal distribution was tested using the Shapiro–Wilk test. All analyses were performed using GraphPad Prism 8.0 (GraphPad). The significant difference was determined using one-way ANOVA followed by Duncan’s test. Differences were statistically significant at P < 0·05 (*), P < 0·01 (**) and P < 0·001 (***).
Results
Addition of inulin in high-carbohydrate diet did not influence the growth performance or body composition of Nile tilapia
After 10-week feeding trial, weight gain showed an increased tendency in the HC group compared with the CON group, but no significant difference has been detected (P = 0·07). Compared with HC treatment, inulin did not markedly affect weight gain rate (Table 1). High carbohydrate or addition of inulin had no influence on feed conversion ratio or viscera index (Table 1). Hepatosomatic index was significantly higher in the HC group than in CON (P < 0·05), whereas no obvious difference was found between the HC and HCI groups (Table 1). The body moisture, total lipid and total protein were detected, and no significant difference in body compositions was found among three treatments (Table 1). These results indicated that the supplement of inulin in high-carbohydrate diet had no obvious effect on growth performance and body composition of Nile tilapia.
Influence of high carbohydrate and a supplement of inulin on growth condition and body composition of Nile tilapia
(Mean values with their standard errors)
CON, fish fed with 35 % starch; HC, fish fed with 45 % starch; HCI, fish fed with 45 % starch supplemented with 5 g/kg inulin.
a,b Mean values within a row with unlike superscript letters were significantly different (P < 0·05).
Addition of inulin in high-carbohydrate diet improved pathogen resistance of Nile tilapia
The results of challenge assay indicated that both the CON and HC groups had similar cumulative mortality, but addition of inulin in high-carbohydrate diet protected the fish against pathogen infection as evidenced by higher survival rate (Fig. 1(a)). The expression level of immune-related genes was detected in liver, and the results indicated that the HC group had significantly lower expression levels of transforming growth factor β and il-10 compared with the CON group (P < 0·01), while HCI treatment up-regulated these genes markedly (P < 0·01) (Fig. 1(b) and (c)). Furthermore, high-carbohydrate diet significantly up-regulated the expression of pro-inflammatory factors, including tnf-α and cyclo-oxygenase-2 (P < 0·01), but these genes were down-regulated in HCI diet (Fig. 1(d) and (e)). Expression level of il-8 was not affected by excessive carbohydrate, but it was lower in inulin-addition diet (P < 0·01, Fig. 1(f)). These results suggested that the supplement of inulin in high-carbohydrate diet could improve pathogen resistance of fish which may be caused by the up-regulation of anti-inflammatory factors and down-regulation of pro-inflammatory factors.
Inulin improved pathogen resistance of Nile tilapia by increasing anti-inflammatory factors and decreasing pro-inflammatory factors. (a) Survival rate of Nile tilapia after Aeromonas hydrophila infection. 
, Fish fed with 35 % starch (CON); 
, fish fed with 45 % starch (HC); 
, fish fed with 45 % starch supplemented with 5 g/kg inulin (HCI). The gene expression of transforming growth factor-β (tgf-β) (b), il-10 (c), tnf-α (d), cyclo-oxygenase-2 (cox2) (e) and il-8 (f). Data are mean values with their standard errors. Mean values were significantly different: ** P < 0·01, *** P < 0·001 (one-way ANOVA).
Addition of inulin in high-carbohydrate diet alleviated lipid accumulation in liver of Nile tilapia
In order to investigate whether addition of inulin could alleviate the adverse effects caused by high-carbohydrate diet, the content of mesenteric fat and the total lipid in liver was measured in these three groups. High-carbohydrate diet distinctly increased the mesenteric fat index compared with the control group (P < 0·05), but no significant difference was found between HC and HCI treatments (Fig. 2(a)). The content of lipid in liver was measured, and the results indicated that the HC group had a higher tendency of total lipid compared with the CON group (P > 0·05), while addition of inulin to high-carbohydrate diet significantly decreased the content of this parameter (P < 0·05, Fig. 2(b)). Compared with the control diet, the level of TAG in liver was significantly increased in high-carbohydrate diet (P < 0·001), but inulin obviously reduced its level (P < 0·05, Fig. 2(c)). These results were further confirmed by oil red staining (Fig. 2(d)). Lipid droplets accumulation was significantly reduced in the HCI group compared with the HC group, suggesting that inulin could alleviate hepatic steatosis caused by high-carbohydrate diet.
Inulin alleviated lipid accumulation in liver induced by high-carbohydrate diet. (a) Abdominal adipose factor; (b) total lipid; (c) TAG contents in liver; (d) liver histology images with oil red O staining. The scale bar is 100 μm. Data are mean values with their standard errors. Mean values were significantly different: * P < 0·05, *** P < 0·001 (one-way ANOVA). CON, fish fed with 35 % starch; HC, fish fed with 45 % starch; HCI, fish fed with 45 % starch supplemented with 5 g/kg inulin.
Addition of inulin in high-carbohydrate diet influenced the expression level of genes related to lipogenesis and β-oxidation in liver of Nile tilapia
Considering the effects of inulin on hepatic lipid accumulation, the expression level of genes related to lipid metabolism was detected. Compared with CON, HC had markedly higher gene expression level of sterol-regulatory element binding proteins, which plays an important role in up-regulating lipogenic (P < 0·001), but the expression level of this gene was down-regulated in the HCI group (P < 0·01, Fig. 3(a)). The similar trend was also found for the expression level of ATP citrate lyase, which encodes a key enzyme of hepatic lipogenesis (P < 0·05, Fig. 3(b)). Excessive carbohydrate had no influence on gene expression of acetyl-CoA carboxylase α, but inulin distinctly down-regulated the expression level of this gene in liver (P < 0·05, Fig. 3(c)). High carbohydrate significantly up-regulated the expression level of fatty acid synthase (P < 0·01), and no difference was found between HC and HCI treatment (Fig. 3(d)). Consistent with the reduced TAG level in liver, expression level of glycerol-3-phosphateacyl transferase, which encodes key enzyme in TAG synthesis, was not influenced by high-carbohydrate diet, but it was distinctly suppressed in the HCI group (P < 0·01, Fig. 3(e)).
Inulin influenced the gene expression of lipogenesis and energy expenditure. The gene expression of sterol-regulatory element binding proteins (srebp) (a), ATP citrate lyase (acly) (b), acetyl-CoA carboxylase α (accα) (c), fatty acid synthase (fas) (d), glycerol-3-phosphateacyl transferase (gpat) (e), carnitine palmitoyltransferase 1a (cpt1a) (f), carnitine palmitoyltransferase 1b (cpt1b) (g) and pparα (h) in liver of Nile tilapia. Data are mean values with their standard errors. Mean values were significantly different: * P < 0·05, ** P < 0·01, *** P < 0·001 (one-way ANOVA). CON, fish fed with 35 % starch; HC, fish fed with 45 % starch; HCI, fish fed with 45 % starch supplemented with 5 g/kg inulin.
Genes related to fatty acid oxidation, including carnitine palmitoyltransferase 1a and carnitine palmitoyltransferase 1b, were markedly decreased in HC diet compared with CON (P < 0·01), while these two genes were significantly up-regulated in HCI diet (P < 0·01) (Fig. 3(f) and (g)). The gene expression level of pparα, a transcription factor to active β-oxidation, was not affected by high-carbohydrate diet, but notably up-regulated in inulin-added diet (P < 0·01, Fig. 3(h)). Collectively, addition of inulin in high-carbohydrate diet could alleviate the hepatic lipid accumulation by down-regulating genes related to lipogenesis and up-regulating genes related to β-oxidation.
Addition of inulin in high-carbohydrate diet increased glycogen synthesis in liver
In order to investigate whether the beneficial impacts of inulin also acted on glucose homoeostasis in Nile tilapia, an oral glucose tolerance test was performed. Compared with CON treatment, plasma glucose was significantly higher at 1·5 and 3 h despite the excessive content of plasma insulin in HC treatment (P < 0·05, Fig. 4(a) and (b)). Although the HCI group had lower plasma insulin than the HC group (P < 0·05), the plasma glucose was lower at 1·5 and 3 h post-glucose injection compared with the HC group (P < 0·05, Fig. 4(a) and (b)). Because liver is an important metabolic organ for blood glucose regulation, the expression level of genes related to glucose metabolism was also detected. The result showed that HC diet obviously reduced the gene expression of insulin receptor compared with CON (P < 0·05), but the supplement of inulin distinctly up-regulated the expression level of this gene (P < 0·05, Fig. 4(c)). We also found that the glycogen content in HC was similar with CON, but it was higher in HCI (P < 0·05, Fig. 4(d)). In line with the tendency, gene expression of glycogen synthase (gs), coding a key enzyme for glycogen synthesis, was not influenced by high level of carbohydrate, but the expression level of this gene was significantly higher in HCI compared with HC (P < 0·05, Fig. 4(e)). The above data suggested that inulin could improve high-carbohydrate-induced glucose metabolism disorder by increasing insulin sensitivity and liver glycogen content.
Addition of inulin influenced glycogen synthesis and gluconeogenesis of Nile tilapia. (a) Glucose tolerance test. 
, Fish fed with 35 % starch (CON); 
, fish fed with 45 % starch (HC); 
, fish fed with 45 % starch supplemented with 5 g/kg inulin (HCI); (b) insulin concentrations after glucose load (500 mg/kg body weight); (c) gene expression level of insulin receptor (ir); (d) glycogen content in liver; (e) gene expression level of glycogen synthase (gs) of Nile tilapia. Data are mean values with their standard errors. Mean values were significantly different: * P < 0·05, ** P < 0·01 (one-way ANOVA).
Addition of inulin in high-carbohydrate diet altered intestinal microbiota composition and their metabolites
A total of 1 798 795 high-quality sequences were obtained from twenty-one samples (seven individuals in each group) with an average of 85 656 sequences per fish. The α-diversity indices displayed that Chao1, Simpson and Shannon indexes were not influenced by high carbohydrate and inulin (Fig. 5(a)–(c)). The microbial community composition analysis showed that the most dominant phylum in all samples was Proteobacteria, representing 70, 55 and 44 % of the total sequences in CON, HC and HCI, respectively. Actinobacteria was accounted for 27 % in HC which was higher than that in CON (3 %), and the proportion of this phylum was 11 % in HCI. The relative abundance of Fusobacteria (9 %) was lower in HC treatment than in CON (19 %), but it increased to 30 % in HCI treatment (Fig. 5(d)). The microbiota composition at the phylum level of each sample was similar within the same treatment, but was different among three diets (Fig. 5(e)). Principal component analysis was conducted according to the proportion of OTU in different groups, and the results showed that intestinal microbiota in these three groups were obviously distinguished, suggesting that either high-carbohydrate diet or addition of inulin could influence gut microbiota (Fig. 5(f)). Compared with CON, six OTU including OTU1142 (Microbacteriaceae), OTU971 (Microbacteriaceae), OTU624 (Rhodobacteraceae), OTU449 (Rhodobacteraceae), OTU228 (Rhodobacteraceae) and OTU1135 (Xanthobacteraceae) were significantly increased in HC, while they were decreased in HCI (P < 0·05). High-carbohydrate diet increased the abundance of OTU327 (Microbacteriaceae) which was further increased in inulin supplementation diet (P < 0·05). In contrast with this tendency, the abundance of OTU1249 (Rhodobacteraceae) was lower in the HC group than in the CON group and was lowest in the HCI group (P < 0·05). No significant difference was found in the abundance of six OTU, including OTU1329 (Mycobacteriaceae), OTU2450 (Fusobacteriaceae), OTU116 (Fusobacteriaceae), OTU571 (Brucellaceae), OTU816 (Comamonadaceae) and OTU364 (Oxalobacteraceae), between CON and HC treatment, but addition of inulin significantly increased the abundance of these OTU (P < 0·05) (Fig. 5(g)), suggesting that inulin did not influence microbial complexity of Nile tilapia, but altered the gut microbiota composition.
Intestinal microbiota composition was altered by inulin. (a) Chao1; (b) Simpson; (c) Shannon indexes; (d) proportion of dominant phyla (
, Verrucomicrobia; 
, Actinobacteria; 
, Thermomicrobia; 
, Fusobacteria; 
, Proteobacteria; 
, others); (e) microbiota composition at phylum level of each sample (
, others; Bacteroidetes; 
, Verrucomicrobia; 
, Proteobacteria; 
, Planctomycetes; 
, Fusobacteria; 
, Firmicutes; 
, Thermomicrobia; 
, Actinobacteria); (f) principal component analysis (PCA). 
, Fish fed with 35 % starch (CON); 
, fish fed with 45 % starch (HC); 
, fish fed with 45 % starch supplemented with 5 g/kg inulin (HCI). Heat map analyses of operational taxonomic units (OTU) showing significantly different among groups. Green indicated that the abundance of OTU was higher in the HC group, and yellow indicated that abundance of OTU was lower in the HC group (g). Data are mean values with their standard errors. Mean values of CON and HC groups were significantly different: * P < 0·05, ** P < 0·01. Mean values of HC and HCI groups were significantly different: † P < 0·05, †† P < 0·01.
The production of SCFA was also detected. Compared with the control diet, high-carbohydrate diet did not affect production of acetic acid and propionic acid (Fig. 6(a) and (b)). But the supplement of inulin significantly increased the content of acetic acid and propionic acid in gut (P < 0·05) (Fig. 6(a) and (b)). In addition, the production of butyric acid was markedly decreased in HC treatment (P < 0·01), while there was no difference between HC and HCI treatments (Fig. 6(c)).
Concentration of SCFA of three treatments. (a) Acetic acid; (b) propionic acid and (c) butyric acid in the gut of Nile tilapia. Data are mean values with their standard errors. Mean values were significantly different: * P < 0·05, ** P < 0·01 (one-way ANOVA). CON, fish fed with 35 % starch; HC, fish fed with 45 % starch; HCI, fish fed with 45 % starch supplemented with 5 g/kg inulin.
Discussion
Inulin is widely used in aquaculture, but the influence of inulin on growth condition of fish remains controversial. Recent studies have found that grass carp (Ctenopharyngodon idella) and rainbow trout (O. mykiss) fed with inulin resulted in higher weight gain, specific growth rate and feed condition rate(Reference Mo, Cheng and Choi26,Reference Ortiz, Rebole and Velasco27) . Ibrahem et al. and Tiengtam et al. also found that supplementation of 5 g/kg inulin in diet enhanced the final weight of Nile tilapia(Reference Tiengtam, Khempaka and Paengkoum16,Reference Ibrahem, Fathi and Mesalhy28) . While some studies showed that addition of inulin did no influence on the growth condition of juvenile beluga (Huso huso), Indian white shrimp (Fenneropenaeus indicus) and common carp (Cyprinus carpio)(Reference Reza, Abdolmajid and Abbas29–Reference Hoseinifar, Zare and Merrifield32). Host phylogeny, the concentration of inulin, the duration of the feeding experiment and the diet composition may account for the inconsistency. In order to identify the influence of inulin on fish fed with high-carbohydrate diet and to reveal the possible mechanism, we detected the growth condition and intestinal microbial composition in Nile tilapia. We found no difference on growth performance or body composition among three treatments, but addition of inulin changed the metabolic characteristics of fish and altered intestinal microbiota composition and their metabolites.
It has been reported that high-carbohydrate diet could suppress innate immunity and thus affect the health status and disease resistance(Reference Lin, Shi and Mu1,Reference Li, Hu and Qiao5) . Dietary inulin always acts as immunostimulant to improve immune status in aquaculture. For example, inulin supplementation improved the resistance of Pacific white shrimp (Litopenaeus vannamei) and Nile tilapia to Vibrio alginolyticus or A. hydrophila and the enhancement of immunity response including higher nitroblue tetrazolium activity and higher gene expression of TLR were observed(Reference Ibrahem, Fathi and Mesalhy28,Reference Li, Liu and Dai33) . Sheikholeslami et al. also found that rainbow trout fed with 5 and 20 g/kg inulin could increase lysozyme activity and leucocytes to resist against Streptococcus sp.(Reference Sheikholeslami, Yusefian and Yavari34). However, the connection between inulin and immunostimulant effect in aquatic animals remains unknown. In mammals, the beneficial effects of inulin treatment on immune dysfunction were attributed to the increase in SCFA formation(Reference Valcheva, Koleva and Martinez35). In the present study, the survival rate was higher in the HCI group than that in the HC group post a pathogen challenge. We also found the addition of inulin altered the composition of gut microbiota and increased the content of acetate and propionate. It has been shown that addition of acetate in high-carbohydrate diet could alleviate intestinal inflammation by inhibiting MAPK and NF-κB signalling pathways in Nile tilapia(Reference Li, Hu and Qiao5). Hoseinifar et al. found that dietary sodium propionate improved mucosal and non-specific immune responses of Caspian white fish (Rutilus frisii kutum) including total Ig, lysozyme activity and alternative haemolytic complement activity(Reference Hoseinifar, Zoheiri and Caipang36). These results suggested that in spare of the different composition of gut microbiota in fish and mammals, inulin could improve the immune response and pathogen resistance in fish by modulating the intestinal composition and their metabolites.
Excessive carbohydrate in the diet induced lipid accumulation and glucose intolerance in fish(Reference Zhang, Liu and Tan37,Reference Kamalam, Medale and Panserat38) . In the present study, we found that supplement of inulin in high-carbohydrate diet alleviated the lipid accumulation by down-regulating genes related to lipogenesis and up-regulating genes related to β-oxidation. And inulin also regulated glucose homoeostasis by improving insulin sensitivity. Supplement of inulin could alleviate lipid accumulation by reducing lipogenesis and increasing energy expenditure which was closely related to the alteration of intestinal microbiota composition and the production of SCFA in mammals(Reference Weitkunat, Stuhlmann and Postel39,Reference Sugatani, Sadamitsu and Tadashi40) . However, how inulin regulates the metabolic characteristics in fish remains elusive. In order to identify the mechanism by which inulin regulates the metabolic characteristics in fish, we detected the gut microbiota composition and their metabolites. We found that inulin did not influence gut microbiota diversity, but changed the gut microbiota composition. In addition, we found that addition of inulin obviously increased the abundance of Fusobacteria in Nile tilapia, especially Cetobacterium. This genus commonly exists in the intestinal tracts of fishes and whales(Reference Brennan and Garrett41), and it could produce acetic and propionic acids by fermenting carbohydrates(Reference Edwards, Logan and Gharbia42). 16S rRNA sequencing, which provides valuable information on the comparison of microbiota composition among different groups, has been widely used to identify the intestinal microbiota in aquatic animals(Reference Ortiz-Estrada, Gollas-Galván and Martínez-Córdova43). However, 16S rRNA sequencing cannot reveal the bacterial function directly(Reference Rezasoltani, Ahmadi Bashirzadeh and Nazemalhosseini Mojarad44). In order to identify the bacterial function, we analysed the content of SCFA, which are produced by microbial fermentation in the gut. We found that inulin supplementation increased the production of acetic acid and propionic acid but did not influence the concentration of butyrate in fish gut. Geraylou et al. found that addition of prebiotics enhanced growth performance and immune responses in Siberian sturgeon (Acipenser baerii) and giant freshwater prawn (Macrobrachium rosenbergii) and increased production of acetic acid and propionic acid may account for these changes(Reference Geraylou, Souffreau and Rurangwa45,Reference Chen, Romano and Ebrahimi46) . It has been reported that diet supplemented with acetate or propionate enhanced the growth and immune parameters in Nile tilapia and zebrafish (Danio rerio)(Reference Li, Hu and Qiao5,Reference Safari, Hoseinifar and Kavandi47) . Although the influence of SCFA on fish health is just starting, SCFA showed multiple beneficial effects on mammals. Zou et al. found that inulin prevented dysglycaemia and hepatosteatosis induced by high-fat diet in a microbiota-dependent manner and enhanced the production of acetic acid and propionic acid of C57BL/6 mice(Reference Zou, Chassaing and Singh48). Weitkunat et al. found that addition of propionic acid in high-fat diet improved insulin resistance and prevented hepatic steatosis by decreasing hepatic TAG in mice(Reference Weitkunat, Stuhlmann and Postel39). The above results suggested that the main mechanism by which inulin regulates fish metabolism may attribute to the change of intestinal microbiota and their metabolites.
Conclusions
In conclusion, the present study indicated that inulin could ameliorate inflammation status induced by high-carbohydrate diet to increase the pathogen resistance in Nile tilapia. We also found that supplement of inulin altered the gut microbiota composition and their metabolites, which may account for the alleviation of metabolic syndromes induced by high-carbohydrate diet in fish. To our knowledge, this is the first study revealing the mechanism of inulin in regulating metabolic disorders induced by high-carbohydrate diet. In order to improve the efficiency and sustainability of aquaculture production, the potential beneficial effects of prebiotics and their underlying mechanisms should be identified deeply.
Acknowledgements
We thank all authors for their contribution to the present study.
This work was supported by the National Key R&D program (grant number 2019YFD0900200) and the National Natural Science Foundation of China (grant number 31972798).
M.-L. Z. and Z.-Y. D. designed the research. T. W., N. Z. and X.-B. Y. conducted all experiments. T. W. analysed the data. F. Q., L.-Q. C. and M.-L Z. provided the assistance in analysing data. T. W. wrote the manuscript. M.-L. Z. revised the manuscript. M.-L. Z. had primary responsibility for final content. All authors read and approved the final manuscript.
The authors declare no conflicts of interest.
Supplementary material
For supplementary materials referred to in this article, please visit https://doi.org/10.1017/S000711452000402X
