Benfotiamine Ameliorates High Carbohydrate Diet-Induced Hepatic Oxidative Stress, Inammation and Apoptosis in Megalobrama Amblycephala

Background: The impairment of immunity induced by high-carbohydrate diet is closely associated with the development of glucose metabolic disorders. In the study of diabetes, benfotiamine can prevent β-cell dysfunction by inhibiting inammation, thereby improving insulin resistance. However, information regarding the effects of this substance on aquatic animals is extremely scarce. Methods: A 12-week nutritional research was conducted to evaluate the inuences of benfotiamine on the growth performance, oxidative stress, inammation and apoptosis in Megalobrama amblycephala (45.25 ± 0.34 g) fed high-carbohydrate (HC) diets. Six experimental diets were formulated, containing a control diet (30% carbohydrate, C), a HC diet (43% carbohydrate), and the HC diet supplemented with four graded benfotiamine levels (0.7125 (HCB1), 1.425 (HCB2), 2.85 (HCB3), and 5.7 (HCB4) mg/kg). Results: HC diet intake remarkably decreased daily growth coecient (DGC), growth rate per metabolic body weight (GR MBW ), feed intake (FI), liver antioxidant enzymes activities, sirtuin-1 (SIRT1) protein expression as well as liver mRNA levels of SIRT1, nuclear factor erythroid 2-related factor 2 (Nrf2), catalase (CAT), manganese superoxide dismutase (Mn-SOD), interleukin10 (IL10) than those of the control group, but the opposite was true for plasma activities of alanine transaminase (AST) and aspartate aminotransferase (ALT), and contents of interleukin 1β (IL1β) and interleukin 6 (IL6), liver contents of malondialdehyde (MDA), and mRNA levels of kelch-like ECH associating protein 1 (Keap1), nuclear factor kappa B (NF-κB), tumour necrosis factor α (TNF α), IL1β, IL6, Bax, Caspase 3, Caspase 9 and P53. As with benfotiamine supplementation, HCB2 diet remarkably increased DGC, GR MBW , liver antioxidant enzymes activities, SIRT1 protein expression as well as liver mRNA levels of SIRT1, Nrf2, CAT, Mn-SOD, IL10 and Bcl2, while the opposite was true for plasma activities of AST and ALT, and contents of IL1β and IL6, liver MDA contents as well as mRNA levels of Keap1, NF-κB, TNF α, IL1β, IL6, Bax, Caspase 3, Caspase 9 and P53. Conclusion: Benfotiamine at 1.425 mg/kg can improve the growth performance and alleviate the oxidative stress, inammation and apoptosis of M. amblycephala fed HC diets through the activation of the SIRT1 pathway.


Introduction
Carbohydrates (CHO) are the most economical energy source for animals including sh, due to their abundance and relatively low price [1]. However, most sh species (especially carnivorous ones) have limited ability to utilize dietary carbohydrates for energy purposes compared with mammals [2,3]. This inadequate ability is characterized by prolonged postprandial hyperglycemia after a glucose loading or the intake of high-carbohydrate diets and even impaired growth [1]. To date, the basis for this apparent glucose intolerance in sh still remain obscure, although several hypotheses have been proposed, including a poor inhibition of postprandial gluconeogenesis, a weak ability of glucose to act as insulinotropins compared with amino acids, relatively low numbers of insulin receptors, etc [1][2][3]. Recently, researchers used several approaches such as metabolomics and transcriptomics to evaluate diet-induced metabolic syndromes in sh. Results from these studies suggested the glucose intolerance in sh is closely implicated in the impairment of immunity induced by highcarbohydrate diet [4][5][6]. However, the underlying mechanisms remain poorly understood, thus deserving our special attention.
Until now, numerous experimental studies showed that the direct link between in ammation and glucose intolerance [7,8]. In fact, the long-term postprandial hyperglycemia caused by excess intakes of dietary carbohydrates can trigger oxidative stress in tissues, which is re ected by the production of superoxide radical anions [9]. Overproduction of superoxide radical anions can stimulate the generation of multiple pro-in ammatory cytokines (like tumour necrosis factor α (TNF α), interleukin 1β (IL1β) and interleukin 6 (IL6)) by enhancing the activation of pro-in ammatory transcription factors, such as nuclear factor kappa B (NF-κB) and toll like receptors (TLRs), thereby resulting in in ammation [10,11]. The increased proin ammatory cytokines then induce insulin insensitivity of tissues by blocking the activation of some key proteins that can mediate the conduction of insulin signaling pathway, thus resulting in glucose intolerance [12][13][14]. Moreover, multiple pro-in ammatory cytokines, such as TNF α and IL6, can also trigger apoptosis by up-regulating transcription of the pro-apoptotic genes, such as Bcl2 family member Bax, thereby accelerating the programmed cell death [15,16]. This is further exacerbates glucose intolerance. These results all showed that in ammation plays an important role in glucose intolerance. Recently, researchers on sh indicated the pronounced glucose intolerance is strongly associated with the low postprandial supervision of metabolic sensor that can regulate glucose metabolism and in ammatory response [1,17]. Among them, silent information regulator 1 (SIRT1), a conserved NAD + -dependent histone deacetylase, has attracted considerable attention. Generally, the activated SIRT1 can regulate multiple biological processes including glucose metabolism, in ammation and apoptosis, such as 1) the enhancement of glucose uptake in peripheral tissues [18]; 2) the suppression of NF-κB pro-in ammatory pathway [19]; 3) the inhibition of P53 pro-apoptotic pathway [20]. However, the aforementioned ndings are mainly derived from mammals. Relevant information in aquatic animals is still quite limited. Recent evidence from studies in sh indicated that high-carbohydrate (HC) diets can cause chronic in ammation accompanied by the generation of pro-in ammatory cytokines, thus increasing the metabolic burden of liver [4,21,22]. This indicated a close link between in ammation and glucose metabolism in sh. However, the relevant physiological basis is still barely understood until now. Considering this, the biochemical and molecular investigation of these aspects will undoubtedly facilitate better understanding of the carbohydrates utilization by sh.
Benfotiamine is a thiamine prodrug with higher bioavailability and absorption than thiamine [23]. At present, it is commonly used as a food supplement for the treatment of type 2 diabetes mellitus (T2DM) to improve glucose homeostasis via enhancing the glycolytic capabilities, promoting insulin synthesis and increasing glucose oxidation in mitochondria [24]. In addition, it also has potent anti-oxidant, anti-in ammatory, anti-apoptotic and anti-carcinogenic properties [25,26]. In the study of diabetes, benfotiamine has been shown to prevent β-cell dysfunction by inhibiting in ammation, thereby improving insulin resistance in tissues [27]. However, information regarding the effects of this substance on aquatic animals is extremely scarce. Recently, our study had con rmed that long-term administration of benfotiamine signi cantly improved the glucose homeostasis of Megalobrama amblycephala (an herbivorous freshwater carp) fed with a HC diet (43% carbohydrate levels) [28]. For example, dietary benfotiamine signi cantly decreased plasma glucose levels of this sh fed HC diet [28,29]. However, the underlying mechanisms are barely understood. Here, we speculated that benfotiamine may bene t the glucose homeostasis of this species through the improvement of immune function. Bearing this in mind, our current research was conducted to investigate the effects of dietary benfotiamine on the growth performance, oxidative stress, in ammation and apoptosis of juvenile M. amblycephala fed with a HC diet. The ndings obtained here can serve to provide information regarding the impact of HC diet on immune response in sh, and promote the development of nutritional strategies for improving the carbohydrate utilization by aquatic animals.

Ethics statement
Animal experimentation within the present study was conducted in accordance with the Animal Care and Use guidelines of Nanjing Agricultural University (Nanjing, China) with the permissions obtained (permit number: SYXK (Su) 2011-0036).

Benfotiamine and the experimental diets
Benfotiamine with a purity of at least 98% was supplied by Xian Reain Biomedical Company (Xian, China). Six isonitrogenous and isolipidic diets with two levels of carbohydrate (30% and 43% carbohydrate) were prepared, containing a control diet (30% carbohydrate, C), a HC diet (43% carbohydrate), and the HC diet containing graded levels of benfotiamine [0.7125 (HCB1), 1.425 (HCB2), 2.85 (HCB3) and 5.7 (HCB4) mg/kg, respectively] ( Table 1). Fish meal, soybean meal, rapeseed meal, and cottonseed meal were used as protein sources, with sh oil and soybean oil used as the main lipid sources. Corn starch was the main carbohydrate source. Microcrystalline cellulose was used to compensate for the carbohydrate levels required.
The experimental diets were produced by the method described in detail previously [30]. Brie y, dry ingredients were grounded, weighed, then mixed with oils. An appropriate amount of water was added to produce dough. The dough was later pelleted using a laboratory pellet machine (MUZL 180, Jiangsu Muyang Group Co., Ltd., Yangzhou, China) and dried in a ventilated oven at 30 °C. After drying, the diets were broken up and sieved into proper pellet size. All diets were stored at −20°C in plastic-lined bags until use.

Feeding trial and experimental conditions
M. amblycephala were obtained from the National Fish Hatchery Station at Yangzhou (Jiangsu, China). The experiment was performed in a re-circulating aquaculture system in the laboratory. The system includes an inlet conduit, a tank, a water treatment unit, a clari er, and an outlet conduit. The tank is in uid communication with the inlet conduit to receive water displaced from the inlet conduit. The volume of each tank is 300 L. Prior to the start of the experiment, sh were acclimated to the experimental facilities and fed a commercial diet (32% protein, 6% lipids, and 33% carbohydrates) for 2 weeks. Then, 360 sh (average weight: 45.25 ± 0.34 g) were allocated to 24 indoor tanks (300 L volume) at a number of 15 sh per tank. Fish were hand-fed thrice daily (07:00, 12:00, and 17:00 h) for 12 weeks. During the experimental period, photoperiod 12: 12 h (dark: light), water conditions including temperature (27.4 ± 0.6 ℃), pH (7.4-7.5) and dissolved oxygen (>5.1 mg/L).

Samples collection
After the 12-week feeding trial, sh were starved for 24 h prior to sampling. All the sh in each tank were counted and weighed. Then, 4 sh from each tank were anesthetized by the diluted MS-222 (100 mg/L). Blood was put into heparinized tubes [28,29]. Also, liver samples were collected from these sh, and then snap frozen in liquid nitrogen and stored at −80 °C until analysis.

Analysis of western blot (WB) and RT-PCR
Total protein was extracted from liver (about 3 g) by the RIPA lysis buffer (Beyotime, China). Then, western blot analysis was performed by our previous methods [30]. Rabbit antibody SIRT1 and β-actin were purchased from Proteintech (13161-1-AP, United States) and Boster (BM3873, China), respectively. Anti-rabbit antibody was purchased from Cell Signaling Technology (#7074, United States). The signals of WB are quantitatively assayed by ImageJ 1.44p (National Institutes of Health, Bethesda, USA).

Statistical analysis
The results in this study were presented as means ± S.E.M. (standard error of the mean). The normality of distribution and the homogeneity of variances were checked prior to one-way analysis of variance (ANOVA) followed by Tukey's HSD test (IBM SPSS Statistics 22.0). Differences with P values ≤ 0.05 were considered statistically signi cant.

Growth performance and feed utilization
No statistical differences in feed intake per metabolic body weight (FI MBW ) and feed e ciency ratio (FER) were found among treatments (Table 3). Daily growth coe cient (DGC), growth rate per metabolic body weight (GR MBW ) and feed intake (FI) of the HC group were signi cantly (P < 0.05) lower than those of the C group (P < 0.05). DGC, GR MBW and FI increased signi cantly (P < 0.05) as benfotiamine levels increased from 0 to 1.425 mg/kg, but decreased signi cantly (P < 0.05) with further increasing levels.

Plasma and liver biochemical indices
Plasma activities of AST and ALT, and levels of IL1β and IL6 as well as hepatic MDA contents of the HC group were signi cantly (P < 0.05) higher than those of the C group, while the opposite was true for liver T-AOC, SOD and CAT activities ( Table 4). Dietary benfotiamine supplementation led to a decrease of plasma AST and ALT activities, and IL1β and IL6 levels as well as hepatic MDA contents, but the opposite was true for liver T-AOC, SOD and CAT activities. In addition, the lowest values of AST, ALT, IL1β and MDA, as well as the highest values of SOD and CAT were found in the HCB2 group.

Liver protein expression of SIRT1 and mRNA levels of SIRT1
The protein expression of SIRT1 and mRNA levels of SIRT1 were signi cantly (P < 0.05) lower in sh fed HC diet than in sh fed the control diet ( Figure 1). As for the HC groups, their values were signi cantly (P < 0.05) increased in sh fed the diet with benfotiamine supplementation.

The transcriptions of antioxidant-related genes in liver
No statistical difference (P > 0.05) was observed in the transcriptions of Cu/Zn-SOD among all the treatments (Figure 2). The mRNA levels of Nrf2, CAT and Mn-SOD of the HC group were signi cantly (P < 0.05) lower than those of the C group, but the opposite was true for Keap1 expression. The HC2 diets signi cantly (P < 0.05) increased the mRNA levels of Nrf2, CAT and Mn-SOD, but the opposite was true for Keap1.

The transcriptions of in ammation-related genes in liver
No statistical difference (P > 0.05) was observed in the mRNA levels of IL8 among all the treatments (Figure 3). The mRNA levels of NF-κB, TNF α, IL1β and IL6 of the HC group were signi cantly (P < 0.05) higher than those of the C group, but the opposite was true for IL10. The HC2 diets signi cantly (P < 0.05) decreased the mRNA levels of NF-κB, TNF α, IL1β and IL6, but the opposite was true for IL10 expression.

The transcriptions of apoptosis-related genes in liver
The mRNA levels of P53, Bcl2, Bax, Caspase 3 and Caspase 9 of the HC group were higher than those of the C group, but no statistical difference was observed in Bcl2 mRNA levels (P > 0.05) (Figure 4). The HC2 diets signi cantly (P < 0.05) decreased the expression of P53, Bax, Caspase 3 and Caspase 9, but the opposite was true for Bcl2 expression.

Discussion
In the present study, the DGC, GR MBW and FI of the HC group were all signi cantly lower than those of the C group, which indicating an retarded growth and an low feed consumption in M. amblycephala fed with a HC diet. These results might be attributed to the following facts: (1) the intake of high-carbohydrate diets can cause persistent hyperglycemia, which is considered as a physiological stress response, thus negatively affecting the growth performance of sh [2]; (2) highcarbohydrate levels in diet can decrease feed palatability, and increase animal satiety, as might consequently result in a decline of feed consumption [38]. In fact, a signi cant high level of plasma glucose has been found in M. amblycephala fed HC diet [28,29]. Then, DGC and GR MBW were signi cantly increased by dietary benfotiamine supplementation with increasing level up to 1.425 mg/kg, which suggesting a bene cial effect of benfotiamine at 1.425 mg/kg on the growth of M. amblycephala fed with HC diet. According to previous study, optimal levels of thiamine (the analog of benfotiamine) can enhance the activities of intestinal digestive enzymes of Jian carp (Cyprinus carpio var. Jian), thus leading to an increase in the feed e ciency [40]. This might improves the growth performance of sh. Additionally, benfotiamine administration has been demonstrated to reduce the metabolic damage induced by hyperglycemia in mammals by the inhibition of in ammation, and the decrease of advanced glycation end products (AGEs) formation [24,27]. This is possible that a similar mechanism exists in sh. However, the underlying mechanisms are poor understood. In order to characterize the corresponding mechanisms, molecular investigations were performed in certain groups (namely the C, HC and HCB2).
In this study, HC diet intake led to an increase of AST, ALT, IL1β, IL6 and MDA, but the opposite trend was true for the activities of T-AOC, SOD and CAT in the liver. These results suggested high dietary carbohydrates induced an in ammation coupled with low antioxidant ability in the liver of M. amblycephal. This result was supported by the following facts: (1) the pro-in ammatory cytokines containing IL 1β and IL 6 are the sensitive indicators of host response to in ammation [41]; (2) the protective effects against oxidative damage could be directly re ected by the activities of some antioxidant enzymes containing SOD, CAT and T-AOC in sh as in mammals [42]. According to previous study, this result may be partly due to that high-carbohydrate intake can inevitably result in the prolonged postprandial hyperglycemia, which in turn stimulates the overproduction of superoxide radical anions via increasing the activities of nicotinamide adenine dinucleotide phosphate oxidase, and inducing mitochondrial membrane hyperpolarization [43]. Then, overproduction of superoxide radical anions can induce intracellular chronic oxidative and in ammatory stresses characterized by decreasing liver SOD, CAT and T-AOC activities, as well as increasing plasma IL1β and IL6 levels [44,45], thus damaging liver metabolic functions. As for the HC groups, dietary supplementation of benfotiamine signi cantly decreased the values of AST, ALT, IL1β, IL6 and MDA, but the opposite trend was true for the activities of T-AOC, SOD and CAT in the liver. This result indicated benfotiamine supplementation signi cantly enhanced antioxidant capacity, but inhibited in ammation in the liver of M. amblycephala fed with HC diet. According to previous studies, these results may be due to that the administration of benfotiamine could promote insulin synthesis and secretion, thereby accelerating glucose disposal in peripheral insulin target tissues, as might accordingly mitigate hyperglycemia-induced intracellular oxidative and in ammatory stresses [46,47]. In addition, the lowest values of AST, ALT, IL1β and MDA, as well as the highest values of SOD and CAT were found in HCB2 group. This result indicated that benfotiamine supplementation at 1.425 mg/kg could effectively enhance the antioxidant capacity, but inhibit in ammation in the liver of M. amblycephala fed with HC diet. SIRT1 is a highly conserved nicotinamide adnine dinucleotide (NAD + ) dependent protein deacetylase, which recognized as a critical regulator in the cellular response to the metabolic, in ammatory, and oxidative stresses [48][49][50]. In this study, the intake of HC diets down-regulated hepatic SIRT1 protein and mRNA levels compared with the control group. According to previous study, the intake of HC diets can inhibit the activity of pyruvate dehydrogenase complex, thereby resulting in a decrease of NAD + content [51]. This might inevitably reduces SIRT1 protein and mRNA levels. As for the HC groups, the supplementation of benfotiamine at 1.425 mg/kg signi cantly up-regulated SIRT1 protein and mRNA levels. Previous study showed that benfotiamine could increase adenosine triphosphate (ATP)/adenosine monophosphate (AMP) ratio via the following reaction: thiamine + ATP = thiamine diphosphate (ThDP) + AMP, thereby activating AMP-activated protein kinase (AMPK) [52]. Subsequently, AMPK activation could increase SIRT1 activity by promoting the generation of NAD + [53]. This might be re ected by up-regulating the protein and mRNA levels of SIRT1.
According to previous studies, the activated SIRT1 could improve the antioxidant capacity of organisms by enhancing the activity of oxidative stress regulator-Nrf2 [54,55]. In the present study, hepatic Nrf2, CAT and Mn-SOD mRNA levels of the HC group were signi cantly lower than those of the control group, but the opposite was true for Keap1 mRNA levels. This result was in line with the antioxidant enzymes (SOD and CAT) activities in the liver, indicating that the intake of HC diets reduced hepatic antioxidant capacity of M. amblycephala. This may be attributed to that the intake of HC diets could inhibit the conduction of Nrf2 antioxidant response element signaling pathway, which might correspondingly down-regulates the activities and mRNA levels of Nrf2-modulated antioxidant enzymes (such as, SOD and CAT) [56,57], thus resulting in a decrease of hepatic antioxidant capacity. This was further supported by that high Keap1 mRNA levels was found in HC group, since Keap1 can decrease the activity of Nrf2 by mediating the ubiquitination [58]. As for the HC groups, dietary supplementation of benfotiamine signi cantly up-regulated the mRNA levels of Nrf2, CAT and Mn-SOD, but the opposite was true for Keap1 mRNA levels. This may be due to that the activated SIRT1 could enhance the activities of Nrf2 by inhibiting the Keap1-mediated ubiquitination, thus up-regulating the mRNA levels of Nrf2-modulated antioxidant enzymes [59,60].
Furthermore, the activated SIRT1 can inhibit high glucose-induced in ammation by suppressing NF-κB pro-in ammatory pathway, thus enhancing the glucose homeostasis [61]. However, such information is mainly derived from mammals. In this study, hepatic NF-κB, TNF α, IL1β and IL6 mRNA levels of the HC group were signi cantly higher than those of the control group, but the opposite was true for IL 10 mRNA levels. These results suggested that the intake of HC diets induced hepatic in ammation of M. amblycephala. According to previous study, the activation of NF-κB induced by high-glucose could stimulate the generation of multiple pro-in ammatory cytokines containing TNF α, IL1β and IL6, thus resulting in the in ammation of organisms [62]. In addition, excessive pro-in ammatory cytokines could also reduce insulin sensitivity in tissues by disturbing the conduction of insulin signaling pathway [63], thereby further aggravating glucose metabolism dysfunction. As for the HC groups, HCB2 signi cantly down-regulated the mRNA levels of NF-κB, TNF α, IL1β and IL6, but the opposite was true for IL 10. These results suggested that dietary benfotiamine supplementation alleviated hepatic in ammation induced by the intake of HC diets. According to previous studies, this result might be attributed to the following facts: (1) the activated SIRT1 by benfotiamine can deacetylate lysine 310 in the p65 subunit of NF-κB, thereby blocking NF-κB mediated pro-in ammatory pathways [64]; and (2) the activated SIRT1 can enhance the synthesis of anti-in ammatory cytokine IL10, which is bene cial to the mitigation of the in ammatory stress [65].
Previous studies indicated that high glucose levels can induce pancreatic β-cell apoptosis, thus aggravating the glucose metabolic dysfunction in tissues [66,67]. The activated SIRT1 can inhibit high glucose-induced apoptosis by the suppression of P53 pro-apoptotic pathway [68]. However, such information in sh is still barely understood. In this study, the mRNA levels of P53, Bax, Caspase 3 and Caspase 9 of the HC group were all signi cantly higher than those of the control group, which indicating HC diet intake might induce apoptosis in the liver of M. amblycephala. This might be due to the fact that P53 activation by high glucose level could increase the activity of Bax, which further activates caspase 9 and caspase 3, thus resulting in apoptosis [15,48]. As for the HC groups, HCB2 signi cantly down-regulated the mRNA levels of P53, Bax, Caspase 3 and Caspase 9, but the opposite was true for Bcl2 mRNA levels. According to previous studies, the activated SIRT1 could inhibit the activity of P53 by the deacetylation at Lys310, which in turn restrains the activation of P53 downstream target genes containing Bax, thereby reducing cell apoptosis [15,16,49,69].

Conclusions
In summary, our ndings demonstrated that dietary supplementation of benfotiamine could attenuate high-carbohydrate induced hepatic oxidative stress, in ammation and apoptosis in M. amblycephala, via the SIRT1 activation, the increase of the activities of the Nrf2-modulated antioxidant enzymes coupled with the down-regulation of the transcriptions of the NF-κB-mediated pro-in ammatory cytokines and P53-mediated pro-apoptotic genes. These results might bene t the improvement of glucose metabolism of M. amblycephala fed HC diets. In addition, sh offered 1.425 mg/kg benfotiamine have the best growth performance.