Dietary Creatine Reduces Lipid Accumulation by Improving Lipid Catabolism in the Herbivorous Grass Carp, Ctenopharyngodon idella

Herein, our goal was to explore the creatine role in the fat suppression of grass carp (GC). In vivo, four semipurified diets containing 5% lipid (normal diet, ND), 8% lipid (high-fat diet, HFD), 8%lipid + 1%creatine (HFD + CR1), and 8%lipid + 2% creatine (HFD + CR2) were designed and formulated, which were fed to juvenile GC (18:51 ± 0:04 g) for 56 days. HFD significantly increased the intraperitoneal fat index as well as the serum and hepatic triglyceride (TG), total cholesterol (TC), and nonesterified fatty acid (NEFA) levels compared to ND; these effects were all attenuated following the addition of creatine in HFD (p < 0:05). Relative to HFD, HFD + CR2 markedly enhanced lipid catabolic gene expressions in the hepatopancreas. This included peroxisome proliferator-activated receptor α, adipose TG lipase (atgl), hormone-sensitive lipase, and carnitine palmitoyltransferase (cpt-1). The protein concentrations of mitochondrial respiratory chain complexes I/II/III/IV/V in the hepatopancreas were upregulated in fish fed HFD + CR1 and HFD + CR2 diets, relative to the HFD-fed fish diet (p < 0:05). In vitro, hepatocytes (L8824) preincubated with oleic acid for 24 h were exposed to varying concentrations of creatine (0, 10, 50, 100, 200, and 400μM) for another 24 h. The number of lipid droplets in each cell remarkably decreased as the concentration of creatine increased in the medium (p < 0:05). Creatine also markedly enhanced the relative transcript levels of genes related to lipolysis (such as agtl), β-oxidation (such as cpt-1 and acyl-CoA dehydrogenase very long chain), and the mitochondrial respiratory chain (such as succinate dehydrogenase complex subunit A and cytochrome c-1). Collectively, the study demonstrates that dietary creatine can suppress lipid accumulation and increase lipid catabolic capacity in GC.


Introduction
Creatine (N-methylguanidine acetic acid) is a small organic molecule that is mainly found in meat and fish [1]. Creatine can also be synthesized endogenously in a two-step process using three amino acids, namely, arginine, glycine, and methionine, and enzymes L-arginine:glycine amidinotransferase (AGAT), and guanidinoacetate N-methyltransferase (GAMT) [2]. In tissues, creatine strongly facilitates adenosine triphosphate (ATP) turnover, which is the energy source of the cell [3]. Thus, the level of creatine may influence the tissue's physiological or pathological status linked to energy homeostasis. It is widely taken by many athletes as an ergogenic aid [4]. Moreover, growing evidence supports creatine's role in cognitive enhancement and in abrogating the impact of mild sleep deprivation [5]. More recently, scientists reported that dietary creatine is critical for lowering hepatic fat accumulation in mice [6]. Mechanistically, increasing fatty acid oxidation and lipoprotein secretion are responsible for this process [7,8]. Moreover, creatine could also enhance thermogenesis in the adipocytes of mice [9,10].
Although discovered over a century ago, creatine was rarely used or mentioned in aquaculture until recent years, when researchers started to focus on this molecule. A study in gilthead seabream (Sparus aurata) shows that dietary creatine produces no improvements in muscle performance as it does in mammals but alters physiological protein expression (such as in lipid metabolism, metabolic pathways, and in cell processes) in the liver [11]. Interestingly, appropriate amounts of creatine in the diet improved growth and feeding efficiency of carnivorous red drum (Sciaenops ocellatus) [12,13], whereas there were no obvious differences in regard to the lipid accumulation in the whole body or muscle in this fish species. Dietary creatine can reduce the serum triglyceride (TG) and total cholesterol (TC) contents in red tilapia (Oreochromis sp.) and also improved robustness against stress [14].
Grass carp (Ctenopharyngodon idella, GC) is a highly prevalent and economically productive fish in China, carrying a production rate of 5.57 million tons per year, which accounts for one-fifth of the cultured freshwater fishes in 2020 [15]. In the aquaculture industry, nonprotein sources with high energy (such as lipids) are increasingly used for growth promotion in aquafeed [16]. However, GC easily accumulate excessive lipid in the mesentery and hepatopancreas during farming when eating relatively high-energy commercial feeds [17]. In commercial GC feeds, fish meal (as a source-deficient and highly priced feedstuff) has largely been replaced by plant proteins [18]. Because creatine exists in the fish meal and other animal proteins sources, it is speculated that GC cannot obtain sufficient creatine from a plant protein-based diet [12]. Moreover, the plant protein also lacks the two main precursors of creatine, arginine, and methionine, which may further reduce the production of self-synthesized creatine [2,19,20]. Thus, the relatively low level of creatine in GC feed may be one of the factors in its performance in the industry. In this study, we investigated whether increasing the creatine level could suppress lipid aggregation in the GC, as it does in mammals. To address this, we carried out both in vivo and in vitro experimentations after treatment with creatine, and lipid accumulation indices and lipid metabolism were measured.

In Vivo Experimentation
2.1.1. Experimental Diets. Based on the formulation of Lavell [21], four purified, isonitrogenous experimental diets (36.0% crude protein) were designed to satisfy the protein and energy demands of juvenile GC with some modifications [22]: normal diet (5.0% lipid, ND), high-fat diet (8.0% lipid, HFD), HFD (8.0% lipid) with 1% creatine (Sigma-Aldrich St. Louis, MO, USA) (HFD + CR1), and HFD (8.0% lipid) with 2% creatine (HFD + CR2) ( Table 1). Casein and gelatin were added as the main protein sources, and soybean oil (YiHai Kerry Co., Shanghai, China) and fish oil (Guanchen Biotechnology Co., Xi'an, China) were used as the oil sources in a ratio of 1 : 1. All raw materials were pulverized into fine powder and filtered using a 60-mesh size. They were weighed and mixed from small to large portions step by step based on the formulation. Afterward, oils were added to the ingredients and fully mixed by hand. 0.1% butylated hydroxytoluene (Sigma-Aldrich, St. Louis, MO, USA) was introduced to the feed to serve as an antioxidant. All raw materials were blended with water (added at 700 ml kg -1 ) to form a dough, which was then prepared into noodle-like pellets carrying a diameter of 2 mm using a feed processing machine, model, and producer (LETSFIND, Beijing, China), which were then forced air-dried at room temperature (25-28°C, RT) for 24 h prior to storage at -20°C until additional analysis.
2.1.2. Experimental Feeding Protocol. Juvenile GCs were acquired from Tongwei Aquaculture Co., Ltd. (Foshan, Guangdong, China) and grown in aquaria (0:73 × 0:46 × 1:0 m; with 0.27 m 3 water), with ND for 2 w to acclimatize them to the upcoming experiment. Prior to experimentation, all fish went without food for 24 h and received anesthesia via tricaine methanesulfonate (MS222, Sigma-Aldrich St. Louis, MO, USA). Overall, 264 healthy fish of similar size (18:51 ± 0:04 g) were arbitrarily separated into 12 aquaria (22 individuals/aquarium). Each diet was arbitrarily provided to three aquaria. Fish received twice daily (at 08 : 30 and 16 : 30) hand-feeding till apparent satiation for 8 w. During this process, food consumption was monitored, and dead fish were weighed and eliminated. In addition, routinely filtered water was circulated; with the daily renewal of one-third water to maintain appropriate quality. The O 2 content in water was 6:6 ± 0:8 mg/L, the pH was 7:2 ± 0:5, the water temperature was 29:6 ± 1:6°C, the ammonia content was ≤0.2 mg/L, and the room was set to a 12/12 h light/dark setting, with a light intensity of 550 lx on average.

Sampling
Procedures. Following the 8 w feeding experiment, all fish were kept without food for 24 h and anaesthetized using 0.06 g/L MS-222 to reduce discomfort in the fish. The body weights (BWs) and lengths of all fish were measured. Next, we arbitrarily picked six fish per aquarium for caudal vein blood extraction with a syringe. Blood samples were allowed to clot at 4°C for 6 h, the serum was obtained following 10 min of spinning (825 × g, 4°C), and frozen via liquid N 2 prior to storage at -80°C until further examination. Nine fish were sacrificed, and we obtained the weights of the hepatopancreas, intraperitoneal fat (IPF), and intestine. Hepatopancreatic and adipose tissue samples from three fish per aquarium underwent fixation in 4% paraformaldehyde for histological analaysis. Hepatopancreatic samples from six fish per aquarium were frozen in liquid N 2 prior to storage at -80°C for lipid content, gene expression, enzyme activity, and biochemistry analysis. The remaining fish were frozen in -80°C for whole body proximate composition analysis. All animal protocols followed the national and institutional guidelines on the ethical use and care of experimental animals. Our experimental procedure also received ethical approval from the Chinese Academy of Fishery Sciences. The viscerosomatic (VI), hepatopancreatic (HI), and intraperitoneal fat (IPFI) indices were computed as follows: The fixed adipose tissue samples were rinsed with tap water for 12 h, before twice standard dehydration using 3 Aquaculture Nutrition ethanol gradient (30,50,70,80,90,95, and 100%), with subsequent equilibration in xylene, standard embedding in paraffin, slicing into 5 μm sections with the use of a rotary microtome (RM2235, Leica, Germany), mounting on glass slides, staining with hematoxylin and eosin (H&E), and finally observing with an upright fluorescence microscope (Leica Biosystems, Germany).

Mitochondrial
Respiratory Chain Complex Assay. The hepatopancreas was homogenized for determination of the mitochondrial respiratory chain complex contents in each treatment, including complex (COX) I, COX II, COX III, COX IV, and COX V. These indices were assessed via fish enzyme-linked immunosorbent assay (ELISA) kits (MEI-MIAN, Wuhan, China). Each assay was conducted thrice (three individual samples) per aquarium sample.
2.1.9. Real-Time Quantitative RT-PCR. Total RNA was recovered from hepatopancreas (three individual samples per aquarium) with TRIzol (TaKaRa, Dalian, China), following kit directions, and quantified by an Implen NanoPhotometer (Implen, Inc., Westlake Village, CA, USA), and the integrity of the RNA was confirmed with the use of 1.2% denaturing agarose gel electrophoresis. The copy DNA (cDNA) was generated via a PrimeScript ® RT reagent kit (TaKaRa), as per kit directions. The primer sequences from our study are summarized in Table 2. According to a 6step series, we demonstrated that a 10-fold dilution of our target template yielded primer amplification efficiencies between 94.44% and 102.64% after quantitative real-time polymerase chain reaction (qRT-PCR). The determination coefficient (R 2 ) values were > 0.99. Additionally, following the recommendations of the geNorm (version 3.5) and NormFinder algorithms, β-actin was employed as the internal control. qRT-PCR assays were conducted in triplicate (three times in each sample) using the LightCycler® 96 RT-PCR apparatus (Roche, Switzerland), using 20 μL reaction mixture with 2.0 μL primer (2.5 μM), 2.0 μL diluted firststrand cDNA product, 10 μL 2× Power SYBR™ Green PCR Master Mix (Thermo Scientific, USA), and 6.0 μL sterile double-dH 2 O. The following parameters were used in the qRT-PCR reaction: 95°C for 5 min, 40 cycles at 95°C for 15 s, and 60°C for 1 min. Once this was completed, the melting curve was monitored from 72-95°C (in 1°C/20 s steps) to validate the formation of a single product. Negative references, namely, zero-cDNA and a DNase-digested, non-cDNA converted RNA sample, were employed to verify only cDNA quantification in all reactions. Relative gene expression was computed with the CT (2 -ΔΔCt ) formula, as previously reported [24,25].

2.2.3.
Real-Time qRT-PCR. We plated 1:2 × 10 6 L8824 cells/ well in 6-well plates (three replicates in each treatment). Following treatment, cells were twice rinsed with PBS, prior to total RNA extraction with TRIzol (Life Technologies Inc., USA), following kit directions. Agarose gel electrophoresis and spectrophotometry (Implen NanoPhotometer; Implen, Inc., Westlake Village, CA, USA) were employed for RNA integrity assessment. The cDNA conversion was done with a PrimeScript® RT reagent kit (TaKaRa, Otsu, Japan). Real-time qRT-PCR was carried out as reported in Section 2.1.8.

Statistical Analysis.
All data analyses were done in PASW Statistics 20 (SPSS, Chicago, IL, USA) and are shown as mean ± standard deviation (S.D.). To elucidate differences in indices, one-way analysis of variance, then post hoc least significant difference (LSD) test was employed. Different letters represent statistically significant differences (p < 0:05).

Growth Performance, Feed Utilization, and Biological
Features of Juvenile GC. The BW of fish that consumed the four diets rose from 18:43 ± 0:07 g to 39:39 ± 0:55 g, but no statistical significance was reached among treatments (p > 0:05; Table 3). In fact, the four treatments had comparable feed conversion ratios (p > 0:05). Fish fed HFD had a higher IPFI than ND-fed fish (p < 0:05), whereas this index decreased when 1% and 2% creatine were added to HFD diets (p < 0:05). No obvious changes were seen in other biological features like the condition factor, VI, and HI among the four groups (p > 0:05).  Table 4. Expectedly, HFD-fed fish had elevated TG, TC, and NEFA levels than 4 Aquaculture Nutrition fish fed ND, whereas these lipid indices were suppressed when the HFD diets were supplemented with 1% and 2% creatine (p < 0:05). Similar trends were also found in SOD activity, MDA content, ALT, and AST activities (p < 0:05).

Whole Body Proximate Composition and Lipid
Content of the Hepatopancreas in Juvenile GC. The whole-body composition of GC is summarized in Table 5. No marked alterations were observed in the crude protein and ash amounts among treatments (p > 0:05). Nevertheless, HFD-fed fish exhibited elevated crude fat, compared to ND-fed fish (p < 0:05); this was attenuated when the fish were fed HFD + CR2 (p < 0:05). Similarly, in the hepatopancreas, HFD increased the crude fat content over the ND fish, whereas dietary creatine decreased lipid content in the hepatopancreas; the HFD + CR2 group exhibited a marked alteration from the HFD group (p < 0:05).

Hepatopancreatic and Adipose Lipid Accumulation in
Juvenile GC. The GC-fed HFD had more lipid droplets than the fish fed ND, which decreased in the fish-fed creatineenriched diets, especially at 2% creatine (Figure 1(a)). Quantitatively, TG, TC, and NEFA concentrations rose in the hepatopancreas after feeding HFD, however, all these indices were decreased when creatine was added in the HFD (p < 0:05; Table 4). In adipose tissue, the HFD-fed fish displayed a larger adipocyte size than the fish fed ND, whereas a smaller adipose cell size was observed in the HFD + creatine-fed fish (Figure 1(b)).

Hepatopancreatic Mitochondrial Respiratory Chain
Complex Expression in Juvenile GC. Fish fed the HFD had higher COX I, COX II, COX IV, and COX V expression in the hepatopancreas than the ND-fed fish (p < 0:05; Figure 2(b)). Upon creatine supplementation of HFD, the COX expression level was substantially increased. Specifically, COX I, COX II, COX III, and COX V activities were elevated in both HFD + CR1 − and HFD + CR2-fed fish. COX IV activity was increased in the HFD + CR2-fed fish, relative to the HFD-fed fish (p < 0:05).

Outcome of Creatine on Lipid Catabolism-Related Gene
Expression in GC Hepatocytes. The key transcription factor of the lipid catabolic gene, pparα, showed an increased expression level, which was different when the creatine content reached 200 μM (p < 0:05; Figure 3(c)). Similarly, the relative lipolysis related gene, atgl, level had a higher expression from 100 μM to 400 μM (p < 0:05). No marked alteration was present in hsl levels (p > 0:05).
The mitochondrial β-oxidation-related gene expression showed an increase with the increase in creatine concentration (Figure 4(a)). To be specific, the relative expression of cpt-1 was increased when the creatine level was higher than 100 μM (p < 0:05). During β-oxidation, acyl-CoA dehydrogenase very long chain (acadvl) and acyl-CoA dehydrogenase short chain (acads) gene expression increased when the incubated concentration of creatine was 200 μM (p < 0:05). The transcript expression of hydroxyacyl-CoA dehydrogenase trifunctional multienzyme complex b (hadhb), a key enzyme of β-oxidation, also increased at 50 μM creatine (p < 0:05). No marked change was observed in the acyl-CoA dehydrogenase medium chain (acadm) and hydroxyacyl-CoA dehydrogenase trifunctional multienzyme complex a (hadha) expression among groups (p > 0:05 ).
In the mitochondrial respiratory chain, two genes, succinate dehydrogenase complex subunit A (sdha) and cytochrome c-1 (cyc1), from complex II and complex III, respectively, increased in response to creatine incubation. Specifically, 100 and 200 μM creatine increased the mRNA level of sdha over the control; 200 and 400 μM creatine elevated cyc1 transcript levels, as opposed to the 10 μM treatment (p < 0:05; Figure 4(b)). The levels of cytochrome c oxidase I mitochondrial (mt-co1) and ATP synthase mitochondrial (mt-atp) were comparable among all groups (p > 0:05).

Discussion
After a long period of evolution, the physiological features of fish reflect the food they consume. Studies have shown that fat accumulates in the viscera of herbivorous fish (GC) easier than in omnivorous (channel catfish, Ictalurus punctatus) and carnivorous (largemouth bass, Micropterus salmoides) fishes [27]. Because creatine and its some precursors are "carninutrients," the herbivorous fish should consume less creatine than omnivorous or carnivorous fishes [2]. In this study, we hypothesized that an increase in dietary creatine in an herbivorous fish could reduce the fat deposition, since many scientists demonstrated that dietary creatine can reduce lipid accumulation in mice and have proposed it as a possible therapy for treating fatty liver disease and obesity in humans [6,8,28]. We found that dietary creatine reduces lipid accumulation in GC both in vivo and in vitro, which may be due to increased lipid catabolism. Overall, this is a novel research investigating the significance of creatine in lipid metabolism in this type of fish and provides a reference for its use in aquatic diets.
In the in vivo experiment, the serum and hepatic TG, TC, and NEFA were all suppressed when the diets were supplemented with 1% and 2% creatine, similar to the response in red tilapia [14]. Our study further proved that hepatic lipid droplets and adipocyte size were reduced in these two groups, demonstrating that dietary creatine could ameliorate the fat deposition induced by high-fat diets. However, the minimum dose of creatine that should be added to the diets is not clear. In cell culture, we showed that lipid droplets were decreased when the concentration of creatine exceeded 100 μM, suggesting that minute amounts of creatine can directly affect the lipid accumulation in hepatocytes. Reducing lipid accumulation using creatine in GC is consistent with the response in mammals [6,28]. A previous study showed that fat accumulation increased when tissue creatine decreased in GC upon faba bean consumption [29], suggesting a negative correlation between fat accumulation and creatine. However, our results are not consistent with the red drum carnivorous fish [12,13]. The difference between these two results may be (1) fish specific; GC are herbivorous fish that may not tolerate higher levels of creatine in the body; (2) feed differences; our study use semi-purified diets (creatine-free), whereas the red drum study used commercial diets which included 29% fish meal (containing a certain amount of creatine).
PPARα transcriptionally regulates genes modulating the peroxisomal and mitochondrial β-oxidation and FA transport [30]. In this study, dietary creatine increased the transcripts of PPARα in the hepatocytes of GC both in vivo and in vitro, suggesting that dietary creatine may increase the lipid catabolic ability in this fish specifically, which is consistent with the result in mice [8]. Our study demonstrated that dietary creatine can increase the lipid degradation rate in fish. First, the gene expression of the ratelimiting enzyme of lipolysis, ATGL [31], but not HSL, increased in the hepatocytes. Second, both in vivo and in vitro data showed that dietary creatine enhanced CPT-1 transcripts, a rate-limiting event in mitochondrial β-7 Aquaculture Nutrition oxidation [32]. The in vitro results also exhibited increased relative mRNA expression of acadvl, acads, and hadhb, which are the key enzymes of β-oxidation [33]. Third, it is known that the end product of β-oxidation, acetyl-CoA forms NADPH and FADH 2 , as part of the tricarboxylic acid cycle. Moreover, NADPH and FADH 2 then enter the mitochondrial respiratory chain, whereby with the help of five enzymatic complexes ATP is produced [29]. Our in vivo study showed that the protein concentrations of complexes I-V were all increased after the fish were fed creatine, and the in vitro study showed that appropriate concentrations of creatine increased the relative transcript levels of several genes in complexes I-III. Collectively, these results demonstrate that dietary creatine can increase the rate of lipolysis, β-oxidation, and oxidative phosphorylation in GC hepatocytes. This may be one of the reasons why dietary creatine decreases lipid accumulation, as lipid catabolism is an important determinant in regulating lipid accumulation in fish [34]. The results in mammals have shown that dietary creatine induces fatty acid oxidation and lipoprotein secretion [7,8] and enhances energy expenditure [9], which is consistent with our study. No study has shown that creatine can directly bind to the enzymes for lipid catabolism. However, creatine only accelerates one enzymatic reaction  Figure 2: The outcome of a high-fat diet supplemented with creatine on the lipid catabolism-related gene expression (a) and mitochondrial respiratory chain complex expression (b) in the juvenile grass carp hepatopancreas. Data shown as means ± standard deviation (shown as error bars; n = 3). The average in each panel with absent common letter is statistically significant (p < 0:05; ANOVA). ND: normal diet; HFD: high-fat diet; HFD + CR1: HFD supplemented with 1% creatine; HFD + CR2: HFD supplemented with 2% creatine; pparα: peroxisome proliferator-activated receptor α; atgl: adipose triglyceride lipase; hsl: hormone-sensitive lipase; cpt-1: carnitine palmitoyltransferase 1; COX: mitochondrial complex. 8 Aquaculture Nutrition (creatine kinase), to interconvert creatine and phosphocreatine in tissues that have an elevated need for ATP [35]. Thus, increased creatine concentration in cells may impact the AMP/ATP ratio, which further influences the AMPK activity, thereby increasing lipid degradation [36]. However, this remains to be confirmed. A high lipid level in fish diets has been attracting increasing attention because it produces a protein-sparing effect in numerous fish species like snout bream [37], sea bream [38], and GC [16]. However, this would induce excess lipid accumulation in the whole body and carcass of these fish species. Thus, improving lipid utilization via the lipid catabolism pathway may be a promising strategy not only to reduce excessive lipid accumulation but also to reduce protein degradation. In the current study, we showed that dietary creatine promotes lipid catabolism, which may have a potential role in the 10 Aquaculture Nutrition protein-sparing function. However, we did not find differences in the whole-body protein content among treatments. Importantly, the numerical decrease in the FCR may be a clue that dietary creatine increases protein deposition. The influence of dietary creatine on protein metabolism, such as protein synthesis and degradation, still requires further investigation. Several indices related to fish health are shown in the in vivo study. The HFD increased the serum SOD, ALT, and AST activities and MDA levels, whereas these indices were all suppressed by dietary creatine. Lipid peroxidation produces MDA. Therefore, MDA levels are a direct reflection of lipid peroxidation [39]. SOD is pivotal crucial antioxidant that catalyzes O 2 to H 2 O 2 [40]. A study in red tilapia showed that dietary creatine reduced MDA level and enhanced SOD activity [14], which is similar to our study and suggests that dietary creatine can reduce oxidative stress in fish. It is believed that the tissue lipid content is responsible for lowering oxidative stress [41]. Generally, serum ALT and AST activities are strong predictors of hepatic or other tissue (like spleen, kidney, and red blood cells) health [42]. Our study demonstrated that dietary creatine can reduce cell damage and should be beneficial to fish health. It was reported that serum activities of ALT and AST are increased when lipids are excessively deposited in the liver of fish [31,43]. Thus, the reduced lipid accumulation of hepatocytes by dietary creatine should be responsible for the decreased ALT and AST activities in the serum. In mammals, creatine utilization is limited in specific circumstances as it may affect the gut health [44]. Hence, more indices related to health are therefore required to assess creatine usage in fish diets.

Conclusions
In conclusion, our study showed that dietary creatine can reduce lipid accumulation in GC hepatocytes. These alterations can be partly explained by the enhanced levels of molecules linked to lipolysis, β-oxidation, and oxidative phosphorylation. In this study, we used semipurified diets. It remains to be confirmed if creatine addition in commercial diets would also affect lipid accumulation. The optimal supplementation level, as well as its impact on the proteinsparing function and fish health should be evaluated further.