Effects of different stocking densities and supplementation with Saccharomyces cerevisiae in juvenile Nile tilapia cultivated in a recirculating water system (RAS)

The objective of this study was to evaluate the productive performance, organosomatic indices and liver-intestinal histomorphometric parameters of Nile tilapia (Oreochromis niloticus) juveniles, fed diets containing probiotics (yeast Saccharomyces cerevisiae) and reared at different densities of storage. In this experiment, 450 juveniles of tilapia (2.9 ± 0.23 g) were distributed in 18 experimental units arranged in RAS and arranged in a randomized block design in a 2x3 factorial scheme. These animals were fed two diets, the first without yeast and the second containing 0.1% yeast inclusion and distributed in three stocking densities of 15, 25 and 35 fish / m (500, 833 and 1166 fish / m). There was an improvement in the oxygenation of the medium with the use of yeast, as well as in the lowest stocking densities evaluated (15 and 25 fish / m3). However, lower specific growth rates and higher viscerosomatic index were observed in animals supplemented with 0.1% yeast. For the average area of hepatocytes, there was an interaction between the evaluated factors (use of probiotic x storage density), where the best values were found in the lowest storage densities (15 and 25 fish / m3), regardless of use or not of the probiotic. It is recommended to use a stocking density of 15 to 25 fish / m3 and a supplementation of 0.1% with yeast as a probiotic, in order to improve liver health and oxygenation of the water used to grow Nile tilapia juveniles cultivated in RAS.


Introduction
Aquaculture is the fastest growing activity in comparison to other agricultural food production sectors (FAO, 2018).
From 2001 to 2018, world aquaculture production grew by more than 5% per year, according to data from the State of The World Fisheries and Aquaculture -SOFIA, reaching over 82.1 million tons of fish in the last year. Brazil is currently the 13th largest fish producing country in the world (FAO, 2020), and the incessant search for alternatives aimed at improvements in the aquaculture sector, such as in fish nutrition, is a constant challenge for this activity and for fish farmers (MPA, 2015). Feed costs are one of the most expensive inputs in aquaculture production making alternative ingredients that maintain or even improve the animals performance worth serious consideration (Boscolo et al., 2001;Meurer et al., 2002).
Tilapia is the second most produced species in the world but the most cultivated species in Brazil (FAO, 2020; PeixeBR, 2020) due to its favorable traits, such as good acceptance by consumers, rusticity, rapid growth and meat quality (Meurer et al., 2003;. Nevertheless, throughout the fingerling development stage, high stocking densities are usually adopted (Marengoni, 2006) with occurrences of high mortality rates, since the immune system is still underdeveloped and the natural resistance to diseases is limited (Kirkan et al., 2003;Yin et al., 2008). According to Filipetto et al. (2015), proper nutrition during this developmental stage, when fish grow at accelerated rates , can ensure an effective nutritional support for its growth and health maintenance (Kim et al., 2003). Therefore, the provision of a balanced diets containing food additives might be an efficient strategy to mitigate losses in this stage and provide additional benefits in subsequent phases .
The rational for inclusion of additives in diets for tilapia, also known as functional feeds (Ribeiro et al., 2012), are benefits to both water quality and the animal's immune system and growth performance during fingerlings development and fattening stages (Schwarz, 2009;Iwashita et al., 2014). Among several known additives, probiotics are one of the most commonly used in aquaculture. Probiotics are living microorganisms that, when supplied in adequate quantities, can benefit the Research, Society andDevelopment, v. 10, n. 16, e183101623385, 2021 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v10i16.23385 3 host and may reduce the occurrence of diseases (Reid et al., 2003). Probiotics exert their effects in several ways and have been linked to producing enzymes in the intestinal tract, improving its morphology and functionality (Heidarieh et al., 2012). In this sense, when probiotics are ingested by the fish, in addition to modifying its intestinal microbial composition, it positively affects the immune system and, as a consequence, the animals health status (Borquez et al., 2010) and its growth performance (Dharmaraj& Dhevendaran, 2010).
The use of probiotics for aquatic organisms has been gaining attention over the last few years mainly due to the need to improve sustainable management of aquaculture (Gatesoupe, 1999). Among the most used probiotics in fish nutrition, yeasts and their processing derivatives are characterized by their biosafety, easy incorporation to diets and positive effects on growth performance (Hisano et al., 2006;. According to Butolo (2001), yeasts are fundamental ingredients that must be used in diet formulations, as they act like immunostimulants and pro-nutrients -compounds that promote intrinsic nutritional values, which are demanded in lower quantities in animal diets.
The yeast of S. cerevisiae comes from the bakery industry and contains several immunostimulants compounds, such as beta-glucan, nucleic acids, mannan-oligossacharides and chitin (Gopalakannan & Arul, 2010). This yeast has high levels of proteins, amino acids, vitamins of the B complex, fatty acids, minerals and enzymes (Pardo-Gamboa et al., 2011). Studies reported improvements in growth performance (Schwarz et al., 2010) and intestinal health status of animals supplemented with this strain of yeast (Carvalho et al., 2011;Schwarz et al., 2011;Heidarieh et al., 2012); however, no studies that encompasses the hepatic health of animals were performed. Therefore, the aim of this study was to evaluate the inclusion of a probiotic composed by S. cerevisiae yeast and different stocking densities in the growth development, somatic indexes and hepaticintestinal histomorphometry of Nile tilapia juveniles, in a water recirculation system during the pre-fattening stage.

Experimental design
This study was conducted in the Laboratory of Aquaculture and Animal Pathology, of the Barriga Verde University Center (UNIBAVE), located in the municipality of Orleans, Santa Catarina, Brazil and lasted 35 days, between the months of August and September 2019.The experiment was approved by the Ethics Committee on Animal Use (CEUA-UNIBAVE), under protocol number 003/2018. A total of 450 Nile tilapia juveniles (Oreochomis niloticus) were acquired from a commercial fish farm (lineage GIFT -EPAGRI SC02), with a mean initial weight of 2.9 ± 0.23g.
The animals were distributed in 18 experimental units in a randomized blocks design, a 2x3 factorial scheme, with the inclusion of probiotics and stocking density as factors. Fish were submitted to two diets, one without the addition of yeast (SP) and one with a 0.1% inclusion level of yeast (CP) containing 100% living cells of S. cerevisiae (strain NCYC Sc 47 -ACTISAF PWD, Phileo, France), and distributed in three stocking densities of 500, 833 and 1166 fish / m-3 (15, 25 and 35 fish per tank, respectively). Each tank was considered as an experimental unit.
The inclusion of the probiotic in the diets was performed every three days. The lyophilized yeast powder was diluted in 100 mL of water from the experimental system's supply (SAMAE -Orleans/SC, Brazil, low salinity water), and sprayed on a micro-extruded commercial diet (40% crude protein CP and 1.77 mm of thickness -BIOBASE®), with the aid of a sprayer (ULTRAJET VERDE -GUARANY®), before being provided to fish. The characterization of the commercial diet is presented in Table 1. Each days feed ration was weighed on a precision scale (Electronic Analytical Scale BG 100, Gehaka, Brazil), kept in identifiable plastic containers and stored in an aerated site. Feeding was provided three times a day at 08h00min, 13h00min and 19h00min, at a feeding rate of 15% of the animals live weight in the first week, 8% in the second week and 5% in the following weeks until the end of the experiment. The reduction of feeding rates was defined by means of weekly biometrics, with the aim of establishing a nutritional restrictive challenge.
The animals were allocated in experimental units (polypropylene tanks with 30 L of useful volume) that were oxygenated with porous stones (2 cm) coupled through individual air compressors (SUNSUN® Hp-200, China). These tanks were interlinked by a water recirculation system composed of a circular matrix tank (500 L useful volume) that served as a reservoir/sump. The filtration system was constituted by a mechanical (60-µm perlon mesh) and a biological filter (particulate material like gravel with an approximate volume of 0.1 m3), in order to provide sufficient adhesion surface to bacteria, and assist in the nitrification process. The system was equipped with a heater (2,500 W) and a submerged pump with a capacity of 4,000 liters / hora (ATMAN® model PH4000) that pumped the water to the experimental units, which returned by gravity.
Tanks were cleaned every two days with the recharge water coming from the municipal water and sewage company (SAMAE). Before using this water, it was stored in a 500-L recipient with strong aeration for 2 days, in order to remove the residual chloride. Water quality parameters were measured periodically, with temperature and dissolved oxygen being measured on a daily basis with an oximeter ALFAKIT (model AT-155, Alfakit, Florianópolis/SC, Brazil), and the pH assessed weekly, with the aid of a pH-meter SENSOGLASS (model SP1800, São Paulo/SP, Brazil). Weekly water samplings were conducted with samples kept in a freezer (-20 °C) until analysis. Nitrogenous parameters (total ammonia, nitrite and nitrate) were measured, as well as alkalinity and orthophosphate, with the aid of a photo colorimeter ALFAKIT (model AT-100P, Florianópolis/SC, Brazil). These analyses were performed in a Laboratory of Aquaculture of the Santa Catarina State University (LAQ -UDESC).
At the end of the experimental period, all the fish were captured with nets and gloves, and were manually immobilized (using moistened towels and gloves) for the final biometrics. Growth performance was evaluated according to the following variables: mean initial weight (MIW) (g), mean final weight (MFW) (g), survival rate (S) (%), mean weight gain (MWG) (g), apparent feed conversion (AFC), and specific growth rate (SGR) (% day-1), by means of the equation [(final weightinitial weight) / time in days] x 100.

Somatic indexes and hepatic-intestinal histomorphometry
Twenty percent of the animals from each experimental unit were randomly selected to measure the somatic indexes and the hepatic-intestinal morphometry, by means of collecting samples from the gastrointestinal tract and the liver. For this purpose, fish were kept in water containing ice for 10 min until loss of balance and reduction of operculum movements (Matushima & Mariano, 1996) with a subsequent medullar transection being performed on those animals (Pedrazzani et al., 2007;CONCEA, 2013). The sampled organs were weighted to calculate the hepatosomatic (HSI) and viscerosomatic (VSI) indexes, according to equations, below: Subsequently, two hepatic fractions and two portions of the intestines were sent to the Laboratory of Veterinary Pathology of the Barriga Verde University Center (UNIBAVE) for histological procedures, and later for histomorphometric analyses. The histological processing of samples was performed according to the routine methods proposed by Nunes and Cinsa (2016). Fragments of the intestines and liver were placed in sterile plastic containers that were pre-identified and filled with a formaldehyde solution (10%) (this solution was renewed and the tissues were maintained in this solution until histological slides were prepared). Tissue samples were cut in 0.3 cm portions, diaphanized and included in histological paraffin, and then cut in 3 µm sections with the aid of a semi-automatic microtome, according to the methodology of Mello et al. (2013), with adaptations. Subsequently, the slides were prepared and stained with hematoxylin-eosin (HE) (Prophet et al., 1992), and then the intestinal villi were measured, as well as the area of hepatocytes (MAH). Samples were analyzed in a light microscope (Bioval®, L-2000) coupled with a camera. Pictures were registered with the aid of the software TCapture®. The other fish remaining in the experimental units were weighted and measured for external evaluation parameters and then these were sent to the Experimental Farm of UNIBAVE.

Analysis of intestinal and hepatic histomorphometry
The slides prepared for histological evaluations were analyzed with a 10x lens and selected by means of a criterion of integrity, considering four villi per animal, totaling 360 villi. Measurements ( Figure 1) were adapted from the methodology described by Picoli et al. (2019), being: i) villi's height (VH) as the distance from the apex of the villi until the beginning of the muscle layer; ii) total height of villi (THV) as the height from the villi's apex to the end of the serosa; iii) width of villi (VW); and iv) thickness of the villus epithelium (TV). All measurements were performed with the software ImageJ®. Random images were taken from slides containing the hepatic tissue stained with HE, from five image fields (40x) for each animal.
Then, the area of 20 hepatocytes in each image captured was measured (Figure 2), totaling 100 hepatocytes per animal, according to Picoli et al. (2019). The MAH was estimated for each treatment, as recommended by Tessaro et al. (2012).

Statistical Analysis
The data obtained was analyzed for normality of residues (Shapiro-Wilk test) and homogeneity of variances (Levene test). The parameters relating to water quality, growth performance, as well as somatic and histomorphometric indexes, were evaluated by means of a factorial analysis (Myers, 1990;Quinn & Keough, 2002). In case no interaction between variables was found, a one-way variance analysis (ANOVA) was carried out separately for each factor (stocking density and use of yeast), and when significant effects were visualized, the Tukey test of multiple comparison of means was applied at a 5% probability level (Sampaio, 2010).

Water Quality
The water quality parameters evaluated in this study did not show any interaction with stocking density or with the addition of the probiotic (p > 0.05); however, significant differences were found when these were individually evaluated (Table   2). Improvement in the oxygenation (DO) of the culture water with the use of probiotic was observed in the lowest evaluated stocking densities (15 and 25 fish / m³). The other parameters (temperature, ammonium, nitrate and pH) were statistically similar (p > 0.05) when evaluated individually (Table 2).

Growth performance and somatic indexes
No significant interaction between stocking densities and/or the use of the probiotic supplement was observed in relation to the evaluated growth performance parameters or somatic indexes (p > 0.05); however, the animals supplemented with the probiotic presented a lower SGR (2.65 ± 1.11% day -1 ) and VSI (7.88 ± 2.53%) in comparison to the animals that did not receive the yeast, which had a SGR of 3.27 ± 1.12% day -1 and a VSI of 9.87 ± 2.68% (p < 0.05) ( Table 3). The other parameters (MIW, MFW, MWG, S, AFC and HSI) did not present any statistical difference relating to supplementation (p > 0.05). Similarly, no differences were observed regarding the effects of stocking densities (p > 0.05) on the growth parameters and somatic indexes evaluated (Table 3). Table 3. Growth performance and somatic indexes of Nile tilapia (O. niloticus) juveniles, supplemented with a probiotic (S. cerevisiae) at different stocking densities for 35 days.

Histomorphometric analysis
A lack of significant differences were observed in the evaluations of intestinal variables (THV, VH, VW and TV) (P > 0.05) among treatments and stocking densities, as well as the interaction between them (Table 4). Contrarily, for the average area of hepatocytes, an interaction was observed between the determination factors (use of probiotic x stocking density), where the best values were found in the lowest stocking densities (15 and 25 fish / m³), regardless of using or not the probiotic (Table   4).

Discussion
Over the years, probiotics have been used to improve water quality in aquacultural enterprises due to their beneficial effects in reducing the concentration of nitrogen compounds and competing with pathogenic microorganisms in the rearing Environment (Porubcan, 1991a;1991b;Wang et al., 2000;Wang & Xu, 2004;Zhou et al., 2009;Martínez Cruz et al., 2012). In this study, the water quality parameters evaluated remained within the adequate limits for the cultivated species (El Sayed, 2020). The improvement in the levels of DO with the probiotic inclusion (6.07 ± 0.63 mg / l) in relation to the treatment without the yeast (5.61 ± 0.59 mg / l) (p < 0.05) corroborates with the findings of Abdulrahman and Muhammad (2012), who evaluated S. cerevisiae inclusion levels for common carp (Cyprinus carpio). Those authors reported higher concentrations of DO (5.31 mg/l in the water cultivated with 7% of dry yeast in comparison to the control treatment (3.09 mg / l). This substantially higher value of DO might be related to the abundance of yeasts in the water, which increased the photosynthetic activity of the environment and led to the production of oxygen (Abdulrahman & Muhammad, 2012).
Although the levels of DO are within the adequate range for tilapia (El Sayed, 2020), the treatments with lower stocking densities (15 and 25 fish per experiment unit) suggest a higher availability of oxygen to be used by the fish, seen that at a lower density in a same volume of water, the use of oxygen in comparison to a denser environment (35 fish per unit) is lower. This result corroborate the study of Tucker et al. (1979), who observed a lower weight gain, survival and feed conversion of catfish (Ictalurus punctatus), when the mean levels of DO in the morning were reduced from 4.5 to 2.1 mg / l, due to an increased stocking density in the ponds.
In a study conducted by Ferreira et al. (2018), the authors supplied S. cerevisiae for tilapia reared in treated sludge and observed a mean value of 1.98 ± 0.51 mg / l of DO, which is considerably lower to the values observed in the present study. Taoka et al. (2006a) reported improvements in the survival of fish and in water quality parameters such as pH, N-NH4, N-NO2 and P-PO4 when rearing Japanese sole (Paralichthys olivaceus) supplemented with several probiotics in their diet (Bacillus subtilis at 1.6 × 107 CFU g-1; Lactobacillus acidophilus at 1.2 × 108 CFU g-1; Clostridium butyricum at 2.0 × 107 CFU g-1; and S. cerevisiae at 1.6 × 107 CFU g-1). In another study, Taoka et al. (2006b)  (2007) did not observe any influence on FMW, survival and HSI of the animals; however, the probiotic's inclusion efficiently reduced the number of total coliforms in the water and in the animals' intestines (Meurer et al., 2004). It was suggested that the probiotic inclusion may have colonized the gastrointestinal tract of the fish and modified the local bacterial community. In addition, Meurer et al. (2006) tested the yeast during the sex reversal stage and did not find improvements in the fish's performance.
When using 0.2% of S. cerevisiae yeast in diets for Nile tilapia juveniles, Schwarz et al. (2016) observed a higher deposition of crude protein in the carcass, as well as a higher AFC and weight at 60 days, in comparison to the fish fed a diet with an inclusion level of 0.4%, and the control. Assumpção (2011) provided diets supplemented with the same yeast (1%) to tilapia reared in brackish water ponds during the sexual reversal stage and reported improved AFC and higher weight gain up to 4 months of age, corroborating with Baccarin and Pezzato (2001), and Castro and Cervon (2004). Finally, in the study developed by Sutthi and Thaimuangphol (2020), the authors fed Nile tilapia with a 0.5% inclusion level of S. cerevisiae and observed a higher growth performance, body composition and better blood chemistry parameters in treatments where the fish were reared in low salinity (0 and 5 ppt). Meurer et al. (2009) also used S. cerevisiae at a concentration of 0.1% as a probiotic in diets for Nile tilapia; however, those authors reported low values of HSI in animals supplemented with the yeast. This index indicates if the fish is using its glycogen and lipids reserves (Young & Cech, 1994;Yogata & Oku, 2000), making a correlation between the animal's nutritional status and its growth rate (Busacker et al., 1990). Under adequate handling conditions, the animals' performance is scarcely influenced by probiotic supplementation (Lima et al., 2003), which is based on the lower probability of contact between the animals and pathogenic microorganisms in these conditions (Loddi et al., 2000;Zuanon et al., 1998), as also mentioned by Meurer et al. (2004). It is noteable that, although the animals in this study were maintained in high densities and with a restrictive decreasing feeding rate, they were submitted to adequate conditions of feed management and hygiene. This might justify the lack of clear improvements of the growth performance parameters, seen that the effect of the probiotic tends to be boosted when the animals are exposed to stress or extreme sanitary conditions (Lima et al., 2003;Macari & Furlan, 2005). Even though no significant effects (p > 0.05) were observed for the evaluated parameters (MFW, AFC, HSI and survival), regardless of stocking density and presence of the probiotic, it was evident that no harms was caused to the animals by the addition of the yeast.
Regarding the intestinal villi of animals, no significant differences (p > 0.05) were observed for the evaluated parameters. In addition, no damages were observed in the intestinal mucosa, regardless of probiotic inclusion or stocking density. According to Hisano et al. (2006), when supplementing Nile tilapia with intact yeasts of S. cerevisiae (2.0%), beneficial intestinal morphological alterations are expected, such as an increased number of villi. These benefits occur due to the presence of nucleotides and polysaccharides in the yeast's cell wall, which have an important function in the differentiation and growth of intestinal cells (Hisano et al., 2006). It is noteworthy that larger intestinal villi provide a higher absorptive area in the mucosa, and as a consequence, benefits are observed, such as a higher rate of nutrients absorption (Cechim, 2013;Hisano et al., 2006;Picoli et al., 2019).
When adopting an inclusion rate of 0.4% of yeast throughout an experimental period of 30 days (similar to the present study), Cechim (2013) observed beneficial effects in the animals' intestinal villi. Similar results were reported by Lima (2014), who supplemented diets for yellowtail Astyanax altiparanae with S. cerevisiae, which has the capacity to adhere in the bonding sites of enterocytes, and by means of competitive exclusion with pathogenic bacteria, it reduces their colonization in the intestines (Noga, 1995;Macari & Furlan, 2005;Mello et al., 2013).
Despite all the benefits that yeasts can provide to the health of intestines, a 0.1% inclusion level of S. cerevisiae during 35 days of experiment was not sufficient to positively alter the histomorphometry of intestines of supplemented fish, even though it did not harm the animals. Different food additives have beneficial results in terms of fish intestinal morphology (Schwarz et al., 2011;Lima, 2014;Honorato et al., 2014), with similar results in comparison to this study. For instance, studies with Bacillus subtilis have also demonstrated increases in villi's height and thickness of the intestinal epithelium (Mello et al., 2013;Carvalho et al., 2011;Nakandakare et al., 2013). Mello et al. (2013) attributed the results to an inhibitory action of probiotic bacteria on the multiplication of undesirable bacteria in the intestines, which leads to a higher integrity, digestibility and absorption of nutrients.
The liver plays several functions in the organism, such as accumulating substances as energy reserves, especially in the form of glycogen and lipids (Roberts, 2012;Wolf et al., 2015). Hepatocytes constitute the greatest share of the hepatic parenchyma (Cavichiolo, 2009). In conditions where there is a lack of nutrients, these reserves are used and the volume of these cells tend to reduce (Roberts, 2012;Wolf et al., 2015). In this study, an increased area of hepatocytes was observed (p < 0.05) at different stocking densities and probiotic inclusion levels, and an interaction was found between them (Table 4). This suggests an excellent effect of the yeast on the hepatic function of supplemented animals, seen that the higher areas of hepatocytes observed in the supplemented group at the lowest stocking density (15 animals per experimental unit), represented higher energy stocks to be used. In addition, it was estimated that the animals in these conditions did not have any scarcity of nutrients, which may justify the non-treated group at the densities of 25 and 35 fish per experimental unit. It was speculated that stocking density might be related to a stressful factor for the fish, seen that this limitation could interfere in the amount and quality of available nutrients on an individual level. This fact could justify the lower area of hepatocytes observed in treatments where fish were cultivated in higher densities, which also showed the beneficial effect of S. cerevisiae in relation to the control, where those effects were not observed. Hassaan et al. (2018) observed an improved structure of hepatocytes when supplementing diets for Nile tilapia with 5 g kg-1 (0.05), 10 g kg-1 (0.1%) and 15 g kg-1 (1.5) of S. cerevisiae, in addition to reporting lower levels of alanine aminotransferase (ALT) and aspartate aminotransferase (AST). Jarmolowicz et al. (2012) also reported lower activities of AST and ALT, which are hepatic enzymes liberated in the blood when hepatic cells are damaged, and its reduction might indicate an improvement of the hepatic function (Metwally, 2009). Bombardelli et al (2010) observed an increased area of hepatocytes with lipid deposition caused by a more energy-rich diet, which is common in diets containing high energy levels, since the liver is responsible for storing excess energy (Wolf et al., 2015). It is suggested that an increased area of hepatocytes, and consequently the hepatic energy reserve, as observed in this study, can be related to an increment of the nutrients use obtained with the addition of S. cerevisiae, considering that this yeast has a high protein content and good digestibility for nutrients and amino acids.
In addition, improvements in the digestibility of dry matter, crude protein, crude energy and most essential and nonessential amino acids were also observed (Fuller, 1989;Lara-Flores et al., 2003;Hisano et al., 2008;Cornélio, 2009). In accordance with this, Picoli et al. (2019) observed an increased area of hepatocytes and villi's height when using bee pollen as a feed additive for Nile tilapia. The authors justified those improvements due to the higher supply of nutrients to the fish, especially regarding amino acids coming from the pollen, which supplemented the hepatic glycogen reserves. Campos et al. (2008) suggested that such an increase can come from essential amino acids that are present in the composition of the pollen, which after meeting the animal's basic needs, were deposited in the liver as an energy reserve (Wolf et al., 2015).

Final Considerations
It is recommended to use a stocking density of 15 to 25 fish / m³ and the supplementation of 0.1% of S. cerevisiae as a probiotic to improve hepatic health and water oxygenation in a water recirculation system of Nile tilapia (O. niloticus) juveniles cultived in RAS. More studies are needed to elucidate the use of S. cerevisiae as a health promoter in Nile tilapia diets.

Ethics approval
The experiment was approved by the Ethics Committee on Animal Use (CEUA-UNIBAVE), under protocol number 003/2018.

Conflicts of interest
The authors have no conflicts of interest to declare that are relevant to the content of this article.