2.1 Experimental design
The experiment was carried out at the Aquaculture Laboratory of the State University of Santa Catarina, Campus Chapecó, Santa Catarina, Brazil, for 48 days. The Committee for Ethics in the Use of Animals approved the study under protocol number 6983180521. Four diets were formulated: 0% or control (without Spirulina biomass) and 33, 66 and 100% replacement level of fishmeal by Spirulina biomass. The design was completely randomized with four replications, in a 4 x 2 factorial scheme, with four treatments (0, 33, 66, and 100%) and two cultivation systems (clear water x BFT).
2.2 Experimental diets
The experimental diets were isoenergetic (4,300 GE kcal kg− 1 – Hafedh 1999) and isoproteic (28% crude protein [CP]). They were formulated using SuperCrac® software to meet protein levels approximately 20% below the required for the species and life stage (35% CP [FAO 2017]). The proximal composition of the diets (Table 1) was analyzed according to methods described in the Association of Official Analytical Chemists (AOAC 2000).
The cultivation of Arthrospira platensis was carried out in raceway tanks arranged inside a greenhouse in a semi-continuous system at the Laboratory of Algae Cultivation and Biotechnology – LCBA UDESC, Laguna, Brazil (Table 2). The culture medium contained fresh water and was composed of 10 g L− 1 of NaHCO3, 1 g L− 1 of commercial soluble fertilizer with a Nitrogen:Phosphate:Potassium ratio of 18:6:18 and a high amount of NaCl (30 g L− 1). The cultivation was carried out under natural temperature and lighting conditions (not controlled) and the Spirulina biomass was collected three times a week by filtration with a 20 µm mesh. Subsequently, it was dried in a laboratory oven at 50 ° C for 24 hours and ground in a cryogenic spray mill (MA775, Marconi) before use in the experimental diets.
In addition to this ingredient, the formulated diets contained fish and soybean meal as protein sources and soybean and corn oil as energy sources (Table 1). All ingredients were ground in an industrial processor and sieved through a 0.71-mm mesh. Subsequently, they were mixed with the addition of water (50%), pelleted in sizes 0.1 (mash), 0.5 and 1-mm in “spaghetti” mixer (Malta®, Model 1036, nozzle no. 8), and placed in an oven (Cienlab®, Model CE-220/150) at 55°C for 48 h. Diets were stored in plastic packages and placed in refrigerators (4°C) until use.
Table 1
Composition of experimental diets (g kg-1) based on dry matter.
Ingredients | Replacement percentage of fish meal |
0% | 33% | 66% | 100% |
Fish meal | 36.45 | 23.26 | 10.08 | - |
Spirulina biomass | - | 26.25 | 52.50 | 72.57 |
Soybean meal | 10.00 | 10.00 | 10.00 | 10.00 |
Corn | 41.59 | 27.86 | 14.02 | 3.43 |
Soy oil | 10.73 | 9.00 | 7.29 | 5.98 |
Premix | 0.40 | 0.40 | 0.40 | 0.40 |
Limestone | 0.64 | 1.65 | 2.53 | 3.20 |
Dicalcium phosphate | - | 1.15 | 2.53 | 3.58 |
Table salt | 0.06 | 0.30 | 0.53 | 0.72 |
Antioxidant | 0.10 | 0.10 | 0.10 | 0.10 |
Total (%) | 100 | 100 | 100 | 100 |
Analyzed composition (%) |
DM | 84.33 | 83.78 | 84.56 | 83.10 |
MM | 8.07 | 8.55 | 8.80 | 8.89 |
EE | 10.32 | 7.09 | 6.67 | 6.74 |
CP | 27.64 | 28.76 | 27.65 | 28.26 |
GE (kcal kg− 1) | 4,596 | 4,380 | 4,395 | 4,425 |
Composition analyzed (AOAC 2000). DM – Dry Matter; MM – Mineral Matter; EE – Ethereal Extract; CP – Crude Protein. GE – Gross Energy; Premix - Folic Acid – 2,400 mg, Nicotinic Acid – 48 g, Pantothenic Acid – 24 g, Biotin – 96 mg, Vit. A – 2,400,000 IU, Vit. D3–400,000 IU, Vit. E – 24,000 IU, Vit. B1–9,600 mg, Vit. B2–9,600 mg, Vit. B6–9,600 mg, Vit. B12–9600 mg, Vit K3–4800 mg, Vit. C – 96 g, Iron – 100 g, Manganese – 40 g, Zinc – 6,000 mg, Cobalt – 20 mg, Iodine – 200 mg, Selenium – 200 mg, Antioxidant – 19.6 g.
Table 2
Proximal composition of fish meal and Spirulina biomass used in experimental diets.
Proximal composition | Fish meal | Spirulina biomass |
DM (%) | 93.33 | 84.74 |
MM (%) | 25.79 | 16.56 |
EE (%) | 7.52 | 2.60 |
CP (%) | 55.02 | 32.01 |
Composition analyzed (AOAC 2000). DM – Dry Matter; MM – Mineral Matter; EE – Ethereal Extract; CP - Crude Protein.
2.3 Animals and facilities
Nile tilapia juveniles with an average initial weight of 0.23 ± 0.01 g were purchased from a commercial fish farm. The animals were distributed at seven fish per tank in circular polyethylene tanks (useful volume of 70 L), divided into two independent water recirculation systems (clear water x BFT). Each system consisted of 16 circular boxes interconnected with circular tanks (usable volume: 850 L) called “macrocosms.” The tanks were individually aerated through porous stone (1 x 2 cm) connected to a radial air compressor. Each macrocosm contained a peripheral water pump (Intech® model BP500), responsible for the distribution and recirculation of water in its respective system. Each macrocosm was equipped with a heater resistance (Nobre Brasil®, 3000 W) and connected to a digital temperature controller with a sensor (XH-W3001®, 12V, China), aiming to maintain the appropriate temperature for the species (El-Sayed 2006). A biomass of 2.25 kg m− 3 of tilapia was also maintained in the BFT macrocosm, fed with commercial feed (Supra® Tilapia 32% CP) once a day (1:00 pm; 32% CP). The macrocosm of the clear water system was composed of a filtering system combined with a mechanical filter (60-µm perlon mesh) and biological particulate material (0.2 m³ of analog of low-density polyethylene bio-balls and 0.05 m³ of expanded clay stone). These materials served as a substrate for the nitrifying bacteria to carry out nitrification of the culture water.
The animals were fed throughout the experimental period at a rate of 10% of the biomass per day (Workagegn et al. 2014), divided into three feedings (08:00 am, 1:00 pm, and 5:00 pm). The biomass monitoring was performed every 16 days using biometrics of all experimental units.
At the beginning of the experiment, all tanks in the BFT system received a previously matured biofloc inoculum (10 L) to prevent unwanted variations in water quality (Krummenauer et al. 2014; Martins et al. 2020). Subsequently, dechlorinated water was added to complete the useful volume of 70 L. The same proportion (1:6) was used in the macrocosm. For the formation and maintenance of the microbial community and the adequate conditions of the culture medium, a C:N ratio of 10–15:1 was maintained (Pérez-Fuentes et al. 2016). Sugarcane molasses was used as an external source of carbon.
The water quality parameters of both cultivation systems were measured daily, including temperature, dissolved oxygen (ALFAKIT® oximeter, model AT170), and pH (pHmeter ALFAKIT®, model AT315). Three times a week, the measurement of total ammoniacal nitrogen (TAN) was performed for both systems, and the measurement of total sedimentable solids (TSS) using an Imhoff cone for the BFT system. According to the results obtained in these measurements (TAN > 1 mg L− 1), molasses was or was not added to the BFT system. Weekly analyses of nitrite, nitrate, orthophosphate, and alkalinity of the culture medium were performed using a photocolorimeter (ALFAKIT®, model AT100P). The parameters remained within the recommended range for tilapia (El-Sayed 2006; Azim and Little 2008; Luo et al. 2014; Lima et al. 2015; 2018) (Table 3).
Table 3
Water quality parameters (mean ± standard deviation) of juvenile Nile tilapia grown in two culture systems (clear water x BFT) after 48 days of feeding with Spirulina biomass.
Parameters | Macrocosm |
AC | BFT |
Tª (°C) | 25.40 ± 1.28 | 25.90 ± 1.26 |
OD (mg L− 1) | 7.97 ± 1.10 | 7.72 ± 1.05 |
pH | 7.53 ± 0.59 | 7.36 ± 0.30 |
N-NH3 (mg L− 1) | 0.24 ± 0.64 | 0.60 ± 0.67 |
NO2− (mg L− 1) | 0.14 ± 0.15 | 0.31 ± 0.08 |
NO3− (mg L− 1) | 1.82 ± 0.09 | 1.32 ± 0.14 |
ALK (mg L− 1 CaCO3) | 25.8 ± 20.62 | 41.83 ± 20.18 |
ORT (PO4) | 0.64 ± 0.03 | 3.41 ± 1.67 |
TSS (mg L− 1) | - | 13.35 ± 13.82 |
Tª = Temperature; °C = Degrees Celsius; pH = Hydrogenionic Potential; N-NH3 = Ammonia Nitrogen; mg L− 1 = milligram per liter; NO2- = Nitrite; NO3- = Nitrate; ALK = Alkalinity; mg L− 1 CaCO3 = milligram per liter of calcium carbonate; ORT = Orthophosphate; PO4 = Phosphate; TSS = total sedimentable solids; AC = clear water culture system; BFT = Biofloc culture system.
2.4 Zootechnical performance and hepatosomatic index
After the experimental period, the animals remained fasting for 24 hours and were individually weighed (0.01 g precision scale, Marte model ML 600, São Paulo, Brazil) and measured for the evaluation of zootechnical parameters: final weight; average weight gain = final weight (g) – initial weight (g); specific growth rate = ln final weight (g) – ln initial weight (g) / experimental period x 100 (% day-1); apparent feed conversion = feed intake/weight gain; feed intake = feed consumed during the experimental period (g) and survival (S) = number of dead animals / total fish x 100 (%). Subsequently, 12 fish per treatment were sedated with eugenol and sacrificed by medullary section to collect the liver to measure the hepatosomatic index (HSI = liver weight / total weight x 100) and the gastrointestinal tract for further analysis.
2.5 Intestinal and liver histomorphometry
We collected fractions of the proximal intestine and liver for the histomorphometric analysis. The histological processing of the samples was performed at the UDESC Histology Laboratory, Campus Lages (Santa Catarina, Brazil), according to routine methods for preparing histological slides (Nunes and Cinsa 2016). The intestinal and liver fragments were placed in sterile plastic containers with a 10% formalin solution, identified, and fixed for 12–16 hours in 70% alcohol until the evaluation of the morpho-histological characteristics. The fractions were sectioned at 0.3 cm using the histology technique, diaphanization, and embedding in histological paraffin. Fractionation into 3-µm sections was performed using a semiautomatic microtome (Mello et al. 2013). Longitudinal sections were semi-serialized and stained using periodic acid-Schiff (Tolosa et al. 2003).
The samples were observed using light microscopy (Bioval L-2000ª microscope, São Paulo, Brazil). Using a camera attached to the microscope, the intestinal (10x) and liver (40x) images were digitized using “TCapture” software (Brand Laborana®, São Paulo, Brazil). After the scans were completed, ten villi per animal were selected by integrity criteria (Picoli et al. 2019), as follows: i) VH – villus height (i.e., distance from the apex of the villi to the beginning of the muscle layer); ii) TVH – total height of the villi (i.e., height from the apex of the villi to the end of the serosa); iii) VW – villi width; iv) VT – villus epithelium thickness, according to Mello et al. (2013). All measurements were performed using Image J software (National Institute for Health, Bethesda, Maryland, United States of America). An area including 100 hepatocytes per animal was measured. The mean area of hepatocytes (MAH) per treatment was estimated using the same software, according to Picoli et al. (2019).
2.6 Serum biochemical and antioxidant analyses
After anesthesia of 12 animals per treatment, blood was collected by puncture of the caudal vein into syringes containing 10% EDTA. Due to the size of the fish in clear water, a "pool" of the animals' blood was analyzed. The collected blood was centrifuged for 10 min at 3,500 rpm, and supernatants (serum) were stored in microtubes at − 20 ºC until analysis. Total proteins (PTNA) and albumin (ALBU) were evaluated in a biochemical analyzer (Bio Plus® BIO-2000) using commercial kits (Gold Analisa®). Globulin levels (GLOBU) were calculated using the formula:
\(GLOBU{\text{ }}={\text{ }}PTNA{\text{ }}--{\text{ }}ALBU\)
Liver samples of 0.5 g were ground/homogenized (Turratec®) with saline solution (0.9%) and centrifuged at 7,000 rpm for 10 min. The supernatants were stored in microtubes at − 20 ºC until analysis of the following parameters: thiobarbituric acid reactive substances (TBARS), glutathione S-transferase (GST), and protein thiol (PSH). Lipid peroxidation was estimated in samples by measuring malondialdehyde (MDA). Tissue samples were incubated at 95°C for 60 minutes in an acid medium containing 8.1% sodium dodecyl sulfate, 0.5 ml acetic acid buffer (500 mM, pH 3.4), and 0.6% TBA (Ohkawa et al. 1978). Absorbances were measured at 532 nm and compared to the malondialdehyde standard curve. The results were expressed in nmol MDA/ml (Jentzsch et al. 1996).
Glutathione-S-transferase activity was measured kinetically using 1-chloro-2,4-dinitrobenzene (CDNB) as a substrate. The reaction mixture (2 ml) contained 0.1 M phosphate buffer, GSH (1 mM), CDNB (1 mM), and TPM (10%). The absorbance was determined at 340 nm at 37°C, and enzymatic activity was expressed as µmol CDNB/min/mg of protein (Habig et al. 1974).
Protein thiols (PSH) were measured using a method based on the use of DTNB (5,5-dithiobis-2-nitrobenzoic acid) (Ellman, 1959; Sedlak and Lindsay 1968), measured after deproteinization of the homogenates with trichloroacetic acid. The sediment formed by the precipitated protein was resuspended with a homogenization buffer to determine the PSH content. Results were expressed as µmol SH g-1 of tissue.
2.7 Statistical analysis
Normality (Shapiro-Wilk Test) and homogeneity of variances (Bartlett Test) were verified, and the data were subjected to analysis of variance for a factorial scheme. When not significant, means were compared using Tukey's test at a 5% probability level. In the case of interaction, factorial splitting and means were compared using Tukey's test (5%). The analyses were performed using the R® 4.1.1 statistical analysis program.