Introduction

With the increasing demand for aquatic feed, the booming aquaculture industry has been facing a big problem, how to replace expensive and limited animal-origin feed with cheap plant-origin feed. Carbohydrate, rich in sugar and starch, is an inexpensive and widely sourced energy substance. Adding carbohydrates appropriately in the diets can promote fish growth, save protein, reduce ammonia emissions, cut down lipid catabolism, provide metabolic intermediates, and etc. (Azaza et al. 2015; Li et al. 2015; Watanabe 2002). However, due to lower carbohydrate utilization rate of fish than mammal, long-term intake of a carbohydrate-rich diet would cause a glucose metabolic disorder to fish (Moon 2001; NRC 2011). Functional foods may have the potential to improve glycemic control (Bai et al. 2015). For example, adding specific nutrients exogenously, such as thiamine and specific amino acids, is a very simple and effective way to attenuate hyperglycemia (Andoh 2007; Hansen et al. 2015; Kamalam et al. 2017). Additionally, Chinese herbal medicine, as a natural medicine, has been showing the characteristics of mild hypoglycemic effect and low cost to solve this issue (El Kaissi and Sherbeeni 2011; Li et al. 2004).

Berberine (BBR) is a natural compound in many Chinese herbaceous plants such as Coptis chinensis and has been recognized as having various medical potentials (Tillhon et al. 2012). Traditionally, BBR is often used to treat gastrointestinal infections due to its antibacterial and anti-inflammatory properties (Chow et al. 2016; Neag et al. 2018; Tillhon et al. 2012). Nowadays, BBR is attracting much attention because of its outstanding regulation of glucose and lipid metabolism (Ilyas et al. 2020; Zhang et al. 2008), and has also been confirmed in clinical trials (Ni 1988; Wei et al. 2004; Xie et al. 2005). Not only that, BBR also shows a good potential of regulating glucolipid metabolism in fish. For example, in Ctenopharyngodon idella, BBR-supplemented feed can significantly reduce the levels of glucose, total cholesterol (TC), and triglyceride (TG) in the blood, which might be through affecting the diversity and structure of intestinal microorganisms (Pan et al. 2019). And it was also reported that BBR could also reduce lipids by activating genes related to fatty acid oxidation and genes related to reducing fatty acid intake (Lu et al. 2016; Zhou et al. 2019). However, there are limited reports about the molecular mechanism of BBR regulating glycometabolism so far.

Blunt snout bream (Megalobrama amblycephala) is a herbivorous freshwater fish of key economic and cultural importance in China. Its herbivory and relative tolerance to carbohydrate also make it possible to add more starch to its feed for saving costs, although excessive carbohydrate still causes disorder of glucose metabolism. Furthermore, adding glucose-lowering substances into the feed to alleviate the metabolic problems caused by the high-carbohydrate feed is thought to be a direct and effective strategy for blunt snout bream. Previous studies in our lab proofed that adding metformin or resveratrol can effectively reduce the blood glucose by regulating the liver glucose metabolism; however, it still causes an increase in liver glycogen and hepatosomatic index (HSI), and may elevate the liver metabolic burden in the long term (Shi et al. 2018; Xu et al. 2017). Therefore, it is very necessary to find a new candidate to reduce blood glucose as well as liver glycogen for fish. BBR has been shown great properties to anti-diabetes in rodents. However, the regulation of BBR on glucose metabolism in fish has only been limited to the study of gut microbiome currently.

Thus, the purpose of this study is to explore hypoglycemic and lipid-lowering effects of dietary BBR supplemented in high-carbohydrate diet of blunt snout bream. And it may help to understand the underlying mechanism of glycolipid metabolism control by BBR and provide some new insights for improving the carbohydrate utilization of aquatic animals.

Materials and methods

Experimental diets

Feed formulation and proximate composition were presented in Table 1. BBR (≥98%) used in this study was purchased from a local company (Spring and Autumn Biotechnology Company, Nanjing, China). The addition amount of BBR in the experimental diet is 50 mg/kg referenced from a previous study (Lu et al. 2016). A total of three isonitrogenous and isolipidic diets were formulated, including normal-carbohydrate diet (NCD, 30% carbohydrate), high-carbohydrate diet (HCD, 43% carbohydrate), and high-carbohydrate diet supplemented with BBR (HCB, 43% carbohydrate + 50 mg/kg BBR). As for the ingredients, fish meal, soybean meal, rapeseed meal, and cottonseed meal served as protein sources. Corn starch was adopted as the main carbohydrate source. Fish oil and soybean oil were applied to meet the lipid demand of this species. The experimental diets were produced by a pellet extruder (MUZL 180, Jiangsu Muyang Group Co., Ltd., Yangzhou, China), dried in a ventilation room for 24 h, and then stored in plastic bags at −20 °C until use. Proximate composition of each diet was analyzed following the protocol, which consists of moisture, crude lipid, ash, crude protein, crude fiber, nitrogen-free extract, and energy (AOAC 1995).

Table 1 Formulation and proximate composition of the different experimental diets
Full size table

Fish and the feeding trial

Juvenile blunt snout bream were purchased from the Fish Hatchery of Yangzhou (Jiangsu, China). Before feeding experiment, fish were fed with commercial aquafeed (32% protein, 7% lipid, 33% nitrogen-free extract; Shuaifeng Aquafeed Co., Ltd., Nanjing, China) three times a day to a slight excess of satiety for a week. After acclimation, 288 fish (average initial weight of 20.36 ± 1.44 g) were randomly distributed into 24 floating net cages (1 × 1 × 1.5 m, L: W: H), and each cage houses 12 fish. Fish in each cage were randomly assigned to one of the three experimental diets (each treatment had eight replicates). Fish were hand-fed to an apparent visual satiety three times a day (7:00, 12:00, and 17:00), which lasted for 10 weeks. During feeding trial, the water quality was set up as follows: Temperature varied from 27 to 29 °C; dissolved oxygen was maintained above 5.0 mg/L; pH ranged from 7.3 to 7.6; and total ammonia nitrogen and nitrite were kept < 0.4 and 0.01 mg/L, respectively.

Sample collection

After the feeding trial, fish were starved for 24 h to empty the intestine, and then all fish were counted and weighed. Sampling was performed using 100 mg/L MS-222 (tricaine methanesulfonate; Sigma, USA) for anesthetization. Blood was rapidly drawn from the caudal vein into 1.5-mL heparinized tubes and centrifuged at 3000 × rpm for 10 min at 4 °C, and the supernatant plasma was collected and kept at −80 °C. Additionally, liver, muscle, and intraperitoneal fat tissue specimens were taken and immediately frozen in liquid nitrogen, and then kept at −80 °C.

Analytical procedures

Growth parameters

The growth parameters obtained in this study were calculated using the following formulas:

$$ {\displaystyle \begin{array}{l}\mathrm{Feed}\ \mathrm{intake}\ \left(\mathrm{FI}\right)=\mathrm{total}\ \mathrm{feed}\ \mathrm{intake}\ \left(\mathrm{g}\right)/\mathrm{total}\ \mathrm{fish}\ \mathrm{number}\\ {}\mathrm{Specific}\ \mathrm{growth}\ \mathrm{rate}\ \left(\mathrm{SGR}\right)=\left(\mathrm{Ln}W\mathrm{t}-\mathrm{Ln}W0\right)\times 100/T\\ {}\begin{array}{l}\mathrm{Feed}\ \mathrm{conversion}\ \mathrm{ratio}\ \left(\mathrm{FCR}\right)=\mathrm{total}\ \mathrm{feed}\ \mathrm{intake}\ \left(\mathrm{g}\right)/\mathrm{total}\ \mathrm{weight}\ \mathrm{gain}\ \left(\mathrm{g}\right)\\ {}\mathrm{Protein}\ \mathrm{efficiency}\ \mathrm{ratio}\ \left(\mathrm{PER}\right)=\mathrm{fish}\ \mathrm{weight}\ \mathrm{gain}\ \left(\mathrm{g}\right)/\mathrm{total}\ \mathrm{protein}\ \mathrm{fed}\ \left(\mathrm{g}\right)\\ {}\begin{array}{l}\mathrm{Hepatosomatic}\ \mathrm{index}\ \left(\mathrm{HSI}\right)=\mathrm{liver}\ \mathrm{weight}\ \left(\mathrm{g}\right)\times 100/\mathrm{body}\ \mathrm{weight}\ \left(\mathrm{g}\right)\\ {}\begin{array}{l}\mathrm{Abdominal}\ \mathrm{fat}\ \mathrm{percentage}\ \left(\mathrm{AFP}\right)=\mathrm{abdominal}\ \mathrm{fat}\ \mathrm{weight}\ \left(\mathrm{g}\right)\times 100/\mathrm{body}\ \mathrm{weight}\ \left(\mathrm{g}\right)\\ {}\mathrm{Nitrogen}\ \mathrm{retention}\ \mathrm{efficiency}\ \left(\mathrm{NRE}\right)=\left[\left(W\mathrm{t}\times N\mathrm{t}\right)-\left(W0\times N0\right)\right]\times 100/\left(N\mathrm{diet}\times \mathrm{feed}\ \mathrm{intake}\right)\end{array}\\ {}\mathrm{Energy}\ \mathrm{retention}\ \mathrm{efficiency}\ \left(\mathrm{ERE}\right)=\left[\left(W\mathrm{t}\times E\mathrm{t}\right)-\left(W0\times E0\right)\right]\times 100/\left(E\mathrm{diet}\times \mathrm{feed}\ \mathrm{intake}\right)\end{array}\end{array}\end{array}} $$

where W0 is the initial body weight, Wt is the final body weight, T is the feeding period in days, N0/E0 and Nt/Et are the initial and final nitrogen/energy contents in whole body, respectively, and Ndiet/Ediet are the nitrogen/energy contents in diets.

Proximate composition analysis

The methods of approximate compositions of diets and fish were as follows: The samples were dried at 105° until constant weight to measure its moisture; crude protein contents (nitrogen × 6.25) were assessed with a Kjeltec analyzer unit; crude lipid was determined by ether extraction using a Soxtec auto extraction unit; ash content was estimated through carbonizing and combusting in a muffle furnace at 550 °C for 2 h; gross energy was measured by a Bomb Calorimeter (PARR 1281; Parr Instrument Company); and crude fiber was determined by fritted glass crucible method using an automatic analyzer (ANKOM A2000i, USA).

Measurements of plasma metabolites and tissue glycogen and lipid contents

Plasma glucose levels were determined using the glucose oxidase method described by Asadi et al. (2009). Plasma insulin level was determined by radioimmunoassay kit and used guinea pig anti-insulin antibody (Beijing North Institute of Biotechnology Co., Ltd.). Plasma pyruvate levels were estimated according to the method of Nigam (1962). Levels of advanced glycation products (AGEs) in plasma were measured following the method of Monnier et al. (1986). Plasma triglyceride levels in lipoprotein fractions were evaluated using a colorimetric method (McNamara and Schaefer 1987). Plasma lactate levels were measured by injecting 10 μL of serum into a VITROSDT60II chemical analyzer (Kodak, NY, USA) using a DT pipette. Tissue glycogen levels were determined following the method of Pfleiderer and Bergmeyer (1974). According to the method of Floch (1957), the content of tissue lipid was extracted with a mixture of chloroform:methanol (2:1, V:V).

Oil red O staining

Liver samples (n = 3) were fixed in 40 g/kg paraformaldehyde solution for 24 h. The liver was then cut into 20-μm sections and neutral lipid staining with oil red O was performed (Jiang et al. 2018). Under the microscope, the lipid is red and the nucleus is light blue. The Image-Pro Plus 6.0 software was used to measure the red area representing lipid accumulation. Three areas in each sample were randomly selected to obtain the size and number of lipid droplets.

Quantitative RT-PCR

Total RNA was extracted from liver tissue with Trizol (Invitrogen, CA, USA) following the manufacturer’s instruction. The quantity and purity of the extracted RNA were determined by NanoDrop 2000 (Thermo Fisher Scientific, Wilmington, DE, USA), in terms of its absorbance at 260 nm and 280 nm. RNA integrity assessment was performed using denatured agarose gel with 1% formaldehyde. cDNA was prepared from 500 ng/μL RNA using a PrimeScript™ RT Master Mix Kit (Cat. No. RR036A, TaKaRa, Co. Ltd., Dalian, China). The resulting cDNA was diluted with DNase/RNase-free water and then worked as a template for quantitative PCR using a TB Green™ Premix Ex Taq™ Kit (Cat. No. RR820A, TaKaRa, Co. Ltd.) and a real-time PCR detection system (ABI7300; Applied Biosystems, Foster City, CA, USA). A 20-μL real-time PCR reaction volume was adopted, consisting of 10 μL of TB Green™ Premix Ex Taq™ II (Tli RNaseH Plus) (2×), 0.4 μL of PCR forward/reverse primer (10 μM) each, 0.4 μL of ROX reference dye II (50×), 2.0 μL of cDNA template, and 6.8 μL of DNase/RNase-free water. The PCR reaction conditions were as follows: 95 °C for 30 s, then followed by 40 cycles of 95 °C for 5 s and 60 °C for 30 s. Melting curves and electrophoresis were applied to judge the specificity of primers. The relative expression of the target gene is calculated by the 2−ΔΔCt method (Livak and Schmittgen, 2001). β-Action was used as an endogenous control for normalization of the gene expression levels of target genes. All primers are presented in Table 2.

Table 2 Nucleotide sequences of primers used to assay gene expressions by real-time quantitative PCR
Full size table

Statistical analyses

Data in the present study were all analyzed by one-way ANOVA using the SPSS 25.0 statistical software package for Windows (SPSS, Michigan Avenue, Chicago, IL, USA). All data were firstly tested for the homogeneity of variances using the Levene’s test, and then analyzed by one-way ANOVA. Tukey’s honestly significant difference test was used to distinguish differences among treatments (P < 0.05). All data were expressed as mean ± S.E.M (mean standard error).

Results

Growth performance and feed utilization

To assess the effect of BBR on the growth performance of high-carbohydrate-fed fish, NCD, HCD, or HCB were fed to blunt snout bream for 10 weeks, and no mortality was observed during the trial. As was showed in Table 3, final weight, SGR, FI, FCR, PER, NRE, and HSI observed no significant difference among all groups, but the AFP value of fish-fed NCD was significantly lower than those of fish-fed HCD and HCB. HCB-fed fish had a higher ERE than that of NCD-fed fish. For the results of whole-body composition, there were no differences in moisture, ash, and crude protein among all groups, whereas HCB feeding resulted in significant increases in crude lipid and gross energy versus NCD feeding.

Table 3 Growth performance, feed utilization, and whole-body composition of blunt snout bream fed with different experimental diets
Full size table

Plasma biochemistry parameters

BBR addition significantly reversed the HCD-induced increases in plasma glucose, TG, and TC contents (Table 4). Higher AGE level was detected in HCD-fed fish than in NCD-fed fish, while BBR supplementation tended to reduce this effect. And dietary BBR inclusion was inclined to ameliorate the HCD-induced blood insulin impairment versus the HCD group. In addition, plasma pyruvate level of the HCB group was significantly lower than those of the other groups.

Table 4 Effect of diets and BBR on plasma biochemistry parameters of blunt snout bream
Full size table

Tissue glycogen and lipid contents

HCD feeding resulted in significant higher hepatic glycogen and lipid contents than NCD feeding (Table 5), whereas HCB remarkably reversed this effect versus HCD (P < 0.05). The glycogen and lipid levels of intraperitoneal fat in NCD-feeding fish were significantly lower than those in HCD-feeding or HCB-feeding fish. However, there were no significant differences in muscular glycogen or lipid content among all groups.

Table 5 Tissue glycogen and lipid contents of blunt snout bream fed with different diets
Full size table

Oil red O staining of the liver

The accumulation of liver lipid was evaluated by oil red O staining (Fig. 1). According to the photomicrographs of the liver tissue, the lipid droplet clusters of fish-fed HCD were more than that of fish-fed NCD or HCB (Fig. 1d), whereas the addition of BBR showed an obvious lipid-reducing effect.

Fig. 1
figure 1

Hepatic lipid profiles of Megalobrama amblycephala fed different experimental diets. a Fish-fed NCD, b Fish-fed HCD. c Fish-fed HCB. d The lipid accumulation (red-stained area, oil red O staining). Red point lipid droplet, white point vacuoles, blue point nucleus. Photomicrographs (×400) and scale bar (20 μm). Each data represents the average of 3 replicates. The bars assigned with different lowercase letters are significantly different (P < 0.05)

Full size image

Relative expression of genes related to insulin signaling

As can be seen from Fig. 2, to reveal the molecular mechanism associated with high-carbohydrate diet and BBR supplementation-induced insulin signaling, the relative gene expressions of insulin receptor substrate (irs), phosphatidylinositol 3-kinase (pi3k), protein kinase B (akt), and forkhead transcription factor 1 (foxo1) were detected in the liver of this species using RT-qPCR. BBR supplementation remarkably reversed the HCD-induced suppression of irs mRNA expression and the HCD-induced upregulation of foxo1 mRNA expression. The pi3k and akt mRNA expressions of HCB-fed fish were significantly higher than those of the other two groups.

Fig. 2
figure 2

Relative expressions of insulin signaling-related genes in the liver of blunt snout bream fed with different experimental diets. The mRNA levels of insulin receptor substrate (irs) (a), phosphatidylinositol 3-kinase (pi3k) (b), protein kinase B (akt) (c), and forkhead transcription factor 1 (foxo1) (d) were evaluated using real-time RT-PCR. Expression values were normalized using β-action as an endogenous control. Each data represents the average of 6-7 replicates. The bars assigned with different superscripts are significantly different (P < 0.05)

Full size image

Relative expression of glucose metabolism-related genes

As was shown in Fig. 3, there were no differences of phosphoenolpyruvate carboxykinase (pepck), glucose-6-phosphatase (g6pase), and pyruvate kinase (pk) mRNA expressions between the NCD and HCD groups. The pepck and g6pase mRNA expressions in the liver of HCB-fed fish were significantly lower than those of NCD and HCD-fed fish. HCB feeding resulted in significant increases in glucose transporter 2 (glut2) and pk mRNA expressions versus NCD and HCD. HCB remarkably reduced HCD-induced glycogen synthase (gs) expression. There was no difference of hepatic fructose-1,6-biphosphatase (fbpase) mRNA expression among all the treatments.

Fig. 3
figure 3

Relative expressions of glucose metabolism-related genes in the liver of blunt snout bream fed with different experimental diets. The mRNA levels of phosphoenolpyruvate carboxykinase (pepck) (a), glucose-6-phosphatase (g6pase) (b), fructose-1,6-biphosphatase (fbpase) (c), glucose transporter 2 (glut2) (d), pyruvate kinase (pk) (e), and glycogen synthase (gs) (f) were evaluated using real-time RT-PCR. Expression values were normalized using β-action as an endogenous control. Each data represents the average of 6-7 replicates. The bars assigned with different superscripts are significantly different (P < 0.05)

Full size image

Relative expression of lipid metabolism-related genes

In Fig. 4, HCD feeding resulted in significantly higher hepatic fatty acid synthetase (fas) and acetyl-CoA carboxylase (accα) expressions and a relatively lower hepatic carnitine palmitoyl transferase I (cpt I) expression than NCD feeding, while HCB remarkably ameliorated this effect versus HCD.

Fig. 4
figure 4

Relative expressions of lipid metabolism-related genes in the liver of blunt snout bream fed with different experimental diets. The mRNA levels of acetyl-CoA carboxylase α (accα) (a), carnitine palmitoyl transferase I (cpt I) (b), and fatty acid synthase (fas) (c) were evaluated using real-time RT-PCR. Expression values were normalized using β-action as an endogenous control. Each data represents the average of 6-7 replicates. The bars assigned with different superscripts are significantly different (P < 0.05)

Full size image

Discussion

The previous studies demonstrated that fish exposed to a high-carbohydrate diet in a long term will suffer from the postprandial hyperglycemia, the metabolic disorder, and be subjected to the impaired health status eventually (Kamalam et al. 2017; Ren et al. 2015; Ren et al. 2013). Although some medicines, such as benfotiamine, metformin, and nicotinamide, can effectively lower blood glucose and improve the glucolipid metabolism, there still exist some side effects, say the increased liver glycogen or lipid content (Shi et al. 2020; Shi et al. 2018; Xu et al. 2017; Xu et al. 2018). As a plant-origin bioactive substance, BBR has been shown an ability to reduce glucose level with almost less side effects (Na et al. 2012). Therefore, this study aimed to explore the role and mechanism of BBR applied in high-carbohydrate diet-fed blunt snout bream, from the aspects of growth, body composition, plasma biochemical indicators, and glucolipid metabolism processes.

We first determined whether BBR supplementation imposes any effects on the growth performance of HCD-fed fish. We found that either dietary carbohydrate levels or the addition of BBR revealed no significant differences of SGR, FCR, and FI among all groups. The reason might be that blunt snout bream has better utilization of carbohydrates than most carnivorous and omnivorous species (Li et al. 2013; Ren et al. 2013; Shi et al. 2018), and in a certain range, BBR inclusion does not impair growth. It was also reported that the addition of BBR did not affect the body weight gain and fatness of Ctenopharyngodon idellus compared with the control group (Pan et al. 2019). In addition, HCD feeding significantly enhanced AFP level compared with NCD feeding, while BBR supplemented in HCD did not change this effect. According to previous research, high-carbohydrates diet could increase the synthesis of glycogen and fat (Li et al. 2019; Prisingkorn et al. 2017; Ren et al. 2011; Wang et al. 2016). And the abdomen is the main part of fat deposition, which is easy to cause the increase of AFP, and can lead to an increase in whole-body lipid. With the addition of BBR, the ERE, whole-body lipid, and gross energy in the HCB group increased significantly versus the NCD group. The reason might be related to BBR’s ability to inhibit gluconeogenesis pathway, which in turn improves energy retention efficiency (especially fat). The latest research demonstrated that BBR also could enrich the intestinal flora of fish and increase beneficial microorganisms, so that these effects might promote the absorption of nutrients and enhance the ERE in fish whole body (Pan et al. 2019).

For the evaluation of fish physiological health conditions, plasma indicators are considered reliable parameters (Habte Tsion et al. 2016). After 10 weeks’ feeding trial, HCD resulted in higher levels of plasma glucose, AGEs, TG, and TC contents compared with NCD (Table 4). According to previous research in fish, long-term excessive carbohydrate intake does cause a symptom of hyperglycemia (Hemre et al. 2002; Peres and Oliva Teles 1999; Wilson 1994), and increase lipogenesis (Zhou et al. 2017), which might lead to the enhanced levels of glucose, AGEs, TG, and TC in blood. Furthermore, in this study, BBR supplementation significantly reduced the levels of glucose, TG, and pyruvate in plasma versus HCD, and insulin secretion tended to be induced by BBR although no significant difference was observed. According to previous reports, these phenomenons might be explained by the following facts: (1) BBR can directly inhibit gluconeogenesis to lower blood glucose, and does not completely depend on the action of insulin (Xia et al. 2011); (2) BBR may reduce blood glucose TG by adjusting the intestinal microbial flora (Pan et al. 2019) or downregulating stearoyl-CoA desaturase in blunt snout bream (Zhou et al. 2019); (3) BBR may accelerate glycolysis and promote the use of pyruvate (Yin et al. 2008). Regarding the change of TC level, HCD tended to increase TC content compared to NCD, and BBR inclusion had a slight TC-lowering effect (Table 4). These phenomena might be explained that the additive amount of BBR in this study was not enough to lower plasma cholesterol.

In terms of the roles in reducing tissue glycogen and lipid content, BBR has been applied to treat fatty liver and obesity in humans (Lee et al. 2006). In accordance with our results, it was also confirmed that BBR could reduce the accumulation of lipid and glycogen in the liver caused by HCD feeding in blunt snout bream. Oil red O staining of liver tissue demonstrated that BBR addition can reduce lipid droplet in hepatocyte (Fig. 1). It is generally believed that fat is the main form of energy storage, and fish lipid is generally deposited in the peritoneum, mesentery, and muscle tissues besides the liver (Tessari et al. 2009). In this study, BBR promoted the transfer of liver lipid/glycogen to peripheral tissues storing as lipid (abdominal fat and muscle) and enhanced the ERE in fish whole body (Table 3), and alleviated liver injury caused by fatty liver. This effect may be related to the increase of BBR in plasma VLDL level, the promotion of lipoprotein secretion in the liver, and the increase of liver fat output (Lu et al. 2013; Richard et al. 2006; Zhou et al. 2019). Therefore, the accumulation of peripheric fat leads to an increase in lipid in the fish body (Table 3). Hepatic glucose production accounts for more than 90% of endogenous glucose production, and mostly comes from glycogenolysis and gluconeogenesis (Petersen et al. 2017). For liver glycogen, BBR may reduce its content by increasing the expression of PPAR-α and PPAR-δ proteins, and reduce the pathological process of liver to regulate the metabolism of glycolipids (Zhou et al. 2008). Of course, the decrease in glycogen content in the liver may be also related to the decreased expression of gs. Gs is one of the important regulating enzymes for glycogen synthesis, and the reduced expression of gs mRNA in the HCB group may also explain this phenomenon (Fig. 3). Surprisingly, the addition of BBR increased the glycogen content in intraperitoneal fat. This may be because BBR is mainly enriched in the liver, and then liver glycogen synthesis is inhibited, resulting in more glucose accumulation in intraperitoneal fat (Liu et al. 2010; Yan et al. 2015).

In the regulation of glycolipid metabolism, the insulin signaling IRS/ PI3K/AKT is a major signaling pathway (Kamalam et al. 2017; Nandipati et al. 2017). After insulin binds to a receptor in its target cell, a signal cascade is triggered. Activated insulin receptor promotes tyrosine phosphorylation of IRS, thereby activating PI3K and AKT and downstream glucose and lipid metabolism (Zhang et al. 2018). In this study, the high-carbohydrate diet downregulated irs expression significantly versus NCD (Fig. 2). Pi3k and akt expressions also decreased, although the difference was not significant. This phenomenon was similar with the results in the oral glucose test of blunt snout bream (Pan et al. 2017). Moreover, high level of carbohydrates can also reduce IRS expression by stimulating S6K1 expression, leading to the phosphorylation or degradation of IRS1, which limits the transmission of insulin signal from insulin receptor to PI3K, thus weakening the effect of insulin (Manning 2004). However, the addition of BBR could significantly increase the expression of irs, pi3k, and akt (Fig. 2). Liu et al. (2018) reported that BBR prevents LPS-induced TLR4/TNF-α activation and enhances IRS expression in the liver. Thereafter, IRS could activate and stimulate PI3K and AKT activity, and then further inhibits its downstream targets, such as foxo1. Furthermore, AKT could phosphorylate FOXO1 and inhibit its expression to inhibit gluconeogenesis (Datta et al. 1999; Kaestner et al. 2000; Kops and Burgering 1999), and promote glycolysis (Zhang et al. 2006). This is also one of the main ways for BBR to lower plasma glucose.

Regarding glycometabolism, the expressions of glut2 and gs in the HCD group were significantly upregulated compared with the NCD group, and the expressions of pepck, g6pase, and fbpase in the HCD group had no effect. This indicated that long-term high-carbohydrate intake promotes liver glycogen synthesis but has no significant effect on gluconeogenesis. Glucose in the blood moves into the liver mainly through glut2, and long-term HCD feeding upregulated glut2 to promote liver uptake of glucose, which enhanced glycogen synthesis by gs upregulation (Enes et al. 2009). Pepck, g6pase, and fbpase, the key genes for gluconeogenesis, were not significantly changed in the HCD group versus the NCD group. This phenomenon could be caused by insufficient insulin secretion (Table 4), which was also observed in rainbow trout (Panserat et al. 2000; Zhou et al. 2016). However, BBR adding could reduce the expressions of pepck, g6pase, and gs, and increase the transcriptional levels of glut2 and pk (Fig. 3). This indicated that BBR can inhibit liver glycogen synthesis and gluconeogenesis and promote glucose transport and glycolysis, which may also be the cause of lower plasma glucose and reduced liver glycogen contents observed in the HCB group (Tables 4 and 5). It was also supported by a previous study that upregulated hepatic AKT can result in the upregulation of liver Glut2, a downstream substrate of AKT, which in turn promotes glucose transport (Giorgino et al. 2000). With regard to the pk expression in the HCB group, BBR may enhance the expression of pk by demethylating PK promoter and increasing the level of acetylation of histones H3 and H4 of PK (Zhang et al. 2015). BBR may also reduce ATP levels by inhibiting mitochondrial function of hepatocytes or increase the expression of hnf-4α mRNA, thereby inhibiting the expressions of pepck and g6pase (Wang et al. 2008; Xia et al. 2011). The results show that BBR not only promotes the uptake and decomposition of liver glucose, but also inhibits the production of liver glucose.

Compared to mammals, fish are more susceptible to fatty liver because the liver is the main site of fat metabolism for fish (Hemre et al. 2002). In accordance with our results, liver cpt I mRNA expression was significantly downregulated in the HCD group, while accα and fas were opposite. And these changes would increase liver fat synthesis and reduce lipolysis, eventually leading to liver lipid accumulation (Shi et al. 2018; Xu et al. 2017; Xu et al. 2018). Cpt I is considered to be a key regulator of mitochondrial β-oxidation (Kerner and Hoppel 2000). The study of Lu et al. (2016) demonstrated BBR could upregulate the expression of cpt I by activating pparα, thereby regulating mitochondrial β-oxidation. Accα and fas are two key enzymes closely related to liver fatty acid synthesis (Qian et al. 2015). The plasma TG-lowering and liver lipid-declining effects of BBR might be partially modulated by downregulating gene expressions of fas and accα through activation of the AMPK signaling pathway (Cao et al. 2013). This can explain why the hepatic lipid content of the HCB group was decreased.

Conclusion

BBR inclusion in HCD could attenuate hyperglycemia of blunt snout bream, mainly through promoting liver’s glycolysis and insulin signaling, while inhibiting liver’s glycogen synthesis and gluconeogenesis. In addition, BBR could ameliorate the metabolic burden of the liver by inhibiting fat synthesis and promoting lipid decomposition, and then enhance fat uptake in peripheral tissues.