Sodium salt medium-chain fatty acids and Bacillus-based probiotic strategies to improve growth and intestinal health of gilthead sea bream (Sparus aurata)

Experimental diets

The basal control diet (D1) and two different experimental diets (D2, D3) were formulated and delivered by SPAROS LDA (Portugal) (Table 1). D1 diet (45% protein and 18% fat), contained fish meal, blood proteins and poultry meal as the main source of dietary proteins. Plant ingredients as a blend of soy protein, wheat gluten, corn gluten, soybean meal, rapeseed meal, wheat meal and pea starch were included at 42%. Fish oil (SAVINOR, Portugal) was added at 13.9%. The experimental diets (D2 and D3) were prepared by supplementing D1 diet with two different commercial preparations: (i) medium-chain fatty acids in the form of a sodium salt of coconut fatty acid distillate (DICOSAN^®, Norel) rich on C-12 was added as a powder to D2 diet at 0.3%; (ii) the probiotic Bacillus amyloliquefaciens CECT 5940 was added “top coated” to D3 diet, being mixed with oil and sprayed at the appropriate rate to match the final dose at 0.1%. All diets were supplemented with antioxidants, a mineral-vitamin mix and DL-methionine.

Fish, feeding trial and sampling collection

Juvenile GSB of Atlantic origin (Tina Menor, Santander, Spain) of 4–5 g initial body weight (May 2015) were acclimatized for more than two months to the indoor experimental facilities of the Institute of Aquaculture Torre de la Sal (IATS-CSIC, Spain). During this initial period, fish were fed with standard diets (Inicio Forte 824 1.9 mm; BioMar, Dueñas, Spain). Then, fish of 29.6 ± 0.32 g initial mean body weight (±SEM) were distributed in 500 l tanks in triplicate groups of 50 fish each. Each experimental diet was offered to visual satiety (one time per day) six days per week from June to August and five days per week from September to October in a 100-days feeding trial). Feed intake was recorded weekly and fish were counted and group-weighed every 4–6 weeks. Oxygen content of outlet water remained higher than 75% saturation, and day-length and water temperature followed the natural changes at IATS latitude (40°5′N; 0°10′E), varying from 24 °C in later June to 22 °C in early October with a maximum peak of 28 °C at the end of July.

At the end of the trial, overnight fasted fish (6 fish per tank, 18 per experimental condition) were randomly sampled and anaesthetized with 3-aminobenzoic acid ethyl ester (MS-222, 0.1 g/l; Sigma, St. Louis, MO, USA) for blood and tissue collection. Blood was quickly drawn from caudal vessels with heparinized syringes. One aliquot was directly used for measurements of leukocyte respiratory burst and hematocrit/hemoglobin determinations. The remaining blood was centrifuged at 3,000 ×g for 20 min at 4 °C, and plasma samples were frozen and stored at −80 °C until biochemical and immunological assays were done. Prior to tissue collection, anaesthetized fish were killed by cervical section, and liver, viscera and mesenteric fat were weighed. Intestine was taken for weight and length measurements. Representative portions of anterior and posterior intestinal segments were rapidly taken, frozen in liquid nitrogen and stored at −80 °C until RNA isolation. Additional samples of liver and intestinal segments were taken for histological processing.

All procedures were approved by the Ethics and Animal Welfare Committee of the Institute of Aquaculture Torre de la Sal according to national (Royal Decree RD53/2013) and EU legislation (2010/63/EU) on the handling of experimental animals.

Hematological, biochemical and immunological blood analyses

Hemoglobin concentration was determined with a HemoCue B-Hemoglobin Analyser^® (Aktiebolaget Leo Diagnostics, Helsigborg, Sweden), which uses a modified azide methemoglobin reaction for Hb quantification. The hematocrit (Hc) was measured after centrifugation of blood in heparinized capillary tubes at 13,000 ×g for 10 min. Plasma glucose was determined by the glucose oxidase method (ThermoFisher Scientific, Waltham, MA, USA) according to manufacturer’s instructions. Plasma triglycerides (TGs) were determined using lipase/glycerol kinase/glycerol-3-phosphate oxidase reagent (ThermoFisher Scientific, Waltham, MA, USA). Total cholesterol was measured using cholesterol esterase/cholesterol dehydrogenase reagent (ThermoFisher Scientific, Waltham, MA, USA). Total plasma proteins were measured using BioRad protein reagent (Hercules, CA, USA) with bovine serum albumin as standard. Total antioxidant capacity was measured as Trolox activity using a microplate assay kit (709,001, Cayman Chemical, Ann Arbor, MI, USA).

Induction of the respiratory burst (RB) activity of blood leukocytes was measured directly from heparinized blood as previously reported (Saera-Vila et al., 2009). Briefly, 4 µl of heparinized blood were diluted with 96 µl Hank’s balanced salt solution pH 7.4 with calcium and magnesium (HBSS++; Gibco, ThermoFisher Scientific, Waltham, MA, USA) in duplicates in white 96 well plates (NUNC). One hundred µl of a luminol (Fluka, Sigma, St. Louis, MO, USA) solution 2 mM in borate buffer (0.2 M pH 9.2) with 2 µg/ml phorbol myristate acetate (PMA, Sigma, St. Louis, MO, USA) were added to each well and the luminescence was measured immediately every 3 min for 1 h at 25 °C in a microplate luminescence reader (Ultra Evolution; Tecan, Männedorf, Zürich, Switzerland). The integral luminescence in relative light units (RLU) was calculated using HBSS++ as a blank.

Plasma lysozyme was measured by a turbidimetric assay as previously described (Sitjà-Bobadilla et al., 2005). Briefly, 5 µl of serum were diluted with 5 µl of 50 mM sodium phosphate buffer pH 6.2 (PB) and incubated with 200 µl of a 0.3 mg/ml Micrococcus lysodeikticus (Sigma, St. Louis, MO, USA) suspension in PB. The reduction of the absorbance at 450 nm was determined in a microplate reader (Ultra Evolution; Tecan, Männedorf, Zürich, Switzerland) after 0.5 and 4.5 min and a unit of lysozyme activity was calculated as the amount of enzyme that caused a decrease in absorbance of 0.001 per min.

The alternative complement pathway (ACP) activity was determined using a modification of the method described in Sitjà-Bobadilla et al., (2005) using sheep red blood cells (SRBC; Durviz, Valencia, Spain) as a target. Briefly, 25 µl of a suspension of 2.85 ×10⁸ SRBC/ml in 10 mM EGTA, 10 mM Mg²⁺ HBSS (without calcium and magnesium; Gibco, ThermoFisher Scientific, Waltham, MA, USA) were incubated with 100 µl of different dilutions of serum (1:5, 1:10, 1:20, 1:40, 1:80 and 1:100) in duplicates for 1 h at 20 °C with constant shaking. After spinning down the remaining SRBC, 75 µl of supernatants were transferred to a new 96 well plate and the absorbance at 415 nm was measured. Finally, the dilution of serum that caused 50% hemolysis (CH50) was calculated.

Serum immunoglobulin M (IgM) was analyzed using an ELISA assay. ELISA plates were coated with 50 µl of a 1:6,000 dilution of serum in carbonate/bicarbonate buffer 0.1 M pH 9.6, incubated overnight at 4 °C and, after washing, blocked with 200 µl Tris 20 mM 0.5 M NaCl pH 7.4 (TBS) 5% non-fat dry milk (BioRad, Hercules, CA, USA) for 1 h at 37 °C. Then, plates were washed, incubated with 50 µl of rabbit polyclonal anti-GSB IgM (Palenzuela, Sitjà-Bobadilla & Álvarez Pellitero, 1996) diluted 1:20,000 in TBS, 0.05% Tween 20 and 3% non-fat dry milk (3% TTBS) for 1 h at 37 °C, washed again and incubated for another hour with 50 µl of a 1:1,000 dilution of goat anti-rabbit-horseradish peroxidase (HRP) (Sigma, St. Louis, MO, USA) in 3% TTBS. After careful washing, the reaction was developed using the TMB substrate kit (BioRad, Hercules, CA, USA) following the manufacturer’s instructions, the reaction was stopped after 20 min with 1 N H₂SO₄ and the absorbance was measured at 450 nm. Each serum was tested in triplicates and a blank with no serum was included as a background control.

Plasma growth hormone (GH) levels were determined by a homologous GSB radioimmunoassay (RIA) (Martínez-Barberá et al., 1995). The sensitivity and midrange (ED50) of the assay were 0.15 and 2.5–3 ng/ml, respectively. Plasma insulin-like growth factors (IGFs) were extracted by acid-ethanol cryoprecipitation and the concentration of IGF-I was measured using a generic fish IGF-I RIA validated for Mediterranean perciform fish (Vega-Rubín de Celis et al., 2004). The sensitivity and midrange of the assay were 0.05 and 0.7–0.8 ng/ml respectively. Cortisol levels were analyzed using an EIA kit (kit RE52061, IBL International GmbH, Hamburg, Germany) following the manufacturer’s instructions. The limit of detection was 50 pg/ml with an assay midrange of 700 pg/ml.

Histological analyses

For histological examination, pieces of liver, anterior (immediately after the pyloric caeca) and posterior (immediately before the rectum) intestinal segments were fixed in 10% buffered formalin, embedded in paraffin, 4 µm-sectioned and stained with hematoxylin-eosin (H&E), Giemsa and periodic acid-Schiff (PAS) following standard procedures. For each dietary group, a total of 9 fish were examined (three fish per tank replicate).

Immunohistochemical analyses

In order to further characterize the observations performed in the histological analyses, 4 µm thick paraffin sections of anterior and posterior intestine samples were collected on Super-Frost-plus microscope slides (Menzel-Gläser, Braunschweig, Germany) and dried overnight. To stain histamine positive cells we used the protocol described in Estensoro et al. (2014) with some modifications. Briefly, samples were deparaffinized and hydrated and the endogenous peroxidase activity was blocked by incubation in methanol:hydrogen peroxide 0.3% v/v in a 9:1 proportion for 40 min. All incubations were performed in a humid chamber, at room temperature and the washing steps consisted of 5 min immersion in TTBS (20 mM Tris–HCl, 0.5 M NaCl, 0.05% Tween 20, pH 7.4) and 5 min immersion in TBS (without Tween 20) unless stated otherwise. Slides were washed and blocked twice for 30 min, first with TBS 5% bovine serum albumin (BSA) and secondly with 1.5% normal goat serum (Vector Laboratories, Burlingame, CA, USA). After washing, they were incubated with a rabbit anti-histamine antibody (Sigma, St. Louis, MO, USA) 1:50 in TBS overnight at 4 °C, washed again and incubated with a biotinylated goat anti-rabbit antibody 1:200 in TBS (Vector Laboratories, Burlingame, CA, USA) for 1 h. The slides were subsequently washed, incubated for 1 h with the avidin-biotin-peroxidase complex (ABC, Vector Laboratories, Burlingame, CA, USA.), washed and developed by incubating with 3,3′-diaminobenzidine tetrahydrochloride chromogen (DAB; Sigma, St. Louis, MO, USA) for 2 min. The reaction was stopped with deionized water and the slides were counterstained for 5 min with Gill’s hematoxylin before being dehydrated and mounted for light microscopy examination.

RNA extraction, reverse transcription and gene expression analyses

RNA from anterior and posterior intestine samples was extracted using a MagMAX-96 total RNA isolation kit (Life Technologies, Madrid, Spain) as described before (Bermejo-Nogales, Calduch-Giner & Pérez-Sánchez, 2015). The RNA yield was 50–100 µg and RIN (RNA integrity number) values were 8–10 with the Agilent 2100 Bioanalyzer. Reverse transcription (RT) of 500 ng total RNA was performed with random decamers, using the High-Capacity cDNA Archive kit (Applied Biosystems, Foster City, CA, USA) following manufacturer’s instructions. Negative control reactions were performed excluding the reverse transcriptase.

A 96-well PCR array layout was used for the simultaneous profiling under uniform cycling conditions of 35 selected markers of intestinal epithelial barrier, enterocyte mass and nutrient absorption, mucin production and goblet cell differentiation and immunological status (Table 2). The primers were designed to obtain amplicons of 50–150 bp in length (Table S1). Each PCR reaction of 25 µl contained the equivalent of 660 pg of total input RNA, 12.5 µl of 2 × SYBR Green Master Mix (BioRad, Hercules, CA, USA) and 0.9 µM of specific primers. The pipetting and liquid manipulations required to construct the arrays were performed by use of an EpMotion 5070 Liquid Handling Robot (Eppendorf, Hamburg, Germany), and the real-time quantitative PCR was carried out on an Eppendorf Mastercycler Ep Realplex Real-Time PCR Detection System (Eppendorf, Hamburg, Germany). The PCR amplification program consisted of an initial denaturation step at 95 °C for 3 min, followed by 40 cycles of denaturation for 15 s at 95 °C and annealing/extension for 60 s at 60 °C. The efficiency of PCRs (>90%) was checked, and the specificity of reactions was verified by analysis of melting curves (ramping rates of 0.5 °C/10 s over a temperature range of 55–95 °C) and linearity of serial dilutions of RT reactions. Fluorescence data acquired were analyzed using the delta-delta Ct method (Livak & Schmittgen, 2001). Ct values of β actin were consistent among groups (with a minimum of 21.88 ± 0.14 and a maximum of 22.21 ± 0.19) in spite of the dietary treatment or the intestine section, and it was used as housekeeping gene. To compare mRNA expression levels, all values were referenced to the expression level of sirt1 (with a Ct value corresponding to the median of Ct values from all genes analyzed) in the anterior intestine of D1 fish, for which a value 1 was arbitrarily assigned.

Statistical analyses

Data on growth performance and blood biochemistry of the different experimental diets were analyzed by one-way analysis of variance followed by Holm-Sidak test. Two-way analysis of variance, followed by the Student-Newman-Keuls (SNK) test, was carried out to analyze the effect of dietary treatment on the intestine gene expression (with intestine segment and diet as variable factors). The significance level was set at P < 0.05. All analyses were made using the SigmaPlot version 13.0 (Systat Software, San Jose, CA, USA). Hierarchical Clustering to assess the gene expression pattern across intestine between experimental diets was carried out by means of Genesis software (Sturn, Quackenbush & Trajanoski, 2002).

Growth performance

To assess the effects of experimental diets on growth performance, data on feed intake, body weight and organo-somatic indexes were calculated (Table 3). Fish fed D1 and D3 diets showed equally good growth performance across the feeding trial, growing from an initial body weight of 29.66 ± 0.49 g to 146.56 ± 1.11 g with overall feed efficiencies of 0.82–0.84 that were not significantly altered by dietary treatment. Conversely, fish fed D2 diet showed increased feed intake and grew significantly faster when compared to the fish fed D1 with a final body weight (167.7 ± 2.26 g) and weight gain (468 ± 6.10%) that were significantly improved by a 12% and 15%, respectively. This yielded specific growth rates that increased significantly from 1.61 ± 0.01 (D1) and 1.58 ± 0.04 (D3) to 1.74 ± 0.01 in fish fed D2 diet.

Organo-somatic indexes calculated as the ratio of tissue to body weight were determined for liver, viscera, mesenteric fat and intestine (Table 3). The resulting viscerosomatic (VSI) and mesenteric (MSI) indexes were not altered by dietary treatment, whereas hepatosomatic index (HSI) of fish fed D2 (2.03 ± 0.06) was significantly higher than those of D1 (1.70 ± 0.06) and D3 (1.77 ± 0.07) fed fish. Dietary intervention did not alter the intestine weight index calculated as the ratio of organ weight to fish weight (IWI). However, when it was determined as the quotient of fish weight to intestine length (ILI), this relative intestine length index was significantly increased from 10.9 ± 1.78 in D1 fish to 19.9 ± 2.12 in fish fed D2. Intermediate ILI values (12.3 ± 2.78), that did not differ significantly from D1 values, were achieved in fish fed D3 diet.

Blood biochemistry and immunological parameters

Different biochemical and immunological analyses were performed on fish blood to assess the effects of experimental diets (Table 4). Dietary intervention did not alter the measured hematological parameters yielding values in the range of 39–43% for Hc and 9.6–10.3 g/dl for Hb concentration. Likewise, plasma levels of triglycerides, cholesterol and proteins were not altered by diet, whereas plasma glucose concentration was significantly increased in fish fed D2 in comparison to D1. Plasma total antioxidant capacity remained unaltered by dietary treatment. A significant diet effect was not found on circulating levels of growth factors (GH, IGF-I), although the highest IGF-I/GH ratio was coincident with the enhanced feed intake and growth found in fish fed D2.

Regarding blood immunological parameters, D3 fed fish showed an anti-inflammatory profile characterized by significantly decreased respiratory burst on PMA stimulated leukocytes and reduced levels of circulating IgM and cortisol when compared to D1 diet. D2 diet yielded intermediate values in all cases. Serum lysozyme values were not significantly different among groups. The alternative complement pathway (ACH50) was completely unaffected by the different diets.

Histological observations

The effect of the different diets was evaluated by histology. Liver and intestinal samples of all groups were observed to determine differences in general appearance, tissue integrity and cell types present. The liver did not show any pathological effect or significant difference among groups (Figs. 1A–1C). Fat accumulation was similar in the three groups and no steatosis was observed in hepatocytes. Glycogen deposition was similar and moderate in all groups.

In the intestine, goblet cells showed similar histochemical characteristics in all groups and their abundance was slightly higher in the anterior intestine of D3 fed fish (Figs. 1G–1L). Interesting differences were observed in the intestinal architecture of the different dietary groups (Figs. 1D–1F). Anterior intestines of D3 fed fish showed clearly higher intestinal folds or villi, whereas D2 diets induced denser and more complex folds. Another remarkable difference was the presence of eosinophilic granulocytes in the submucosa of the different groups. The irregular shape and size of these cells make exact quantification difficult, thus we only refer to abundance based on overall observations. Eosinophilic granulocytes appeared in almost double amounts in the submucosa of the anterior intestine of fish fed with D2 and were clearly less abundant in the posterior intestine of the D3 group, when compared to the other two groups (Figs. 1G–1L).

Immunohistochemistry

The presence of a differential number of eosinophilic granulocytes in the intestinal submucosa of fish from the different dietary challenges required a further characterization of the cell types. Fish granulocytes differ from those of mammals in their staining properties, which also vary among fish species (Reite & Evensen, 2006). GSB eosinophilic granulocytes can be either acidophils or mast cells, and from these two, only mast cells contain histamine (Mulero et al., 2007). Thus, an anti-histamine antibody was used to further characterize these cells. This allowed detecting histamine-positive cells with similar shape, size and location to those of the observed eosinophilic granulocytes in Figs. 1G–1L (Fig. 2), and therefore we conclude that at least part of the observed eosinophilic granulocytes were histamine-positive mast cells. As stated before, reliable quantification of these cells could not be performed, thus we only refer to significant changes that can be observed by microscopic observation. The anterior intestine of D1 and D2 fed fish had large numbers of mast cells, whereas at least half the amount was found in D3 fed fish (Figs. 2A–2C). In the posterior intestine, mast cells were more abundant in D2 fed fish and clearly lower in D3 fed fish when compared to the control animals (Figs. 2D–2F). Although the presence of another type of eosinophilic granulocytes (acidophils) cannot be discarded, it is clear that, at least, the mast cell population in the intestinal submucosa is affected by the different dietary treatments.

Gene expression profiles

The gene expression profile of selected markers was studied in samples of anterior and posterior intestines of fish fed the different experimental diets in order to characterize the general intestinal status of these fish. For most studied genes, a clear expression pattern across intestinal segments was found, regardless of the diet. The structural genes cdh1, cdh17, had higher expression in anterior intestine, whereas cldn15, ocln and cldn12, were mostly expressed in posterior intestine (as indicated by the low P values in the intestinal segment column of Table 5). Markers of enterocyte mass and nutrient absorption were also differentially expressed across the intestine, with higher expression of alpi in the anterior intestine and fabp6 in the posterior intestine. Some markers of mucus production (muc13, i-muc) also showed a pronounced spatial pattern. Markers of immunological status (il1 β, il8, il10, cd4, cd8 α, cd8 β, lgals1, mIgM and mIgT) were significantly higher expressed in posterior intestine.

Regarding the diet effect, fish fed D2 and D3 diets showed a reduced expression of energy sensing molecules (sirt1, sirt2, sirt3), cdh1, cd4 and lgals8 in anterior intestine, compared to fish fed D1 diet. Fish fed D2 diet showed the highest expression of alpi, significantly in the anterior intestine when compared to D3 fed fish. The expression of fabp2 was higher in D2 than in D1 fish in both anterior and posterior intestinal segments, whereas intermediate values were found in D3 fish. Overall, the expression level of mucin genes was not significantly altered by diet, although the diet × intestine section interaction was statistically significant for muc13, and its expression was significantly down-regulated in D3 fish in comparison to D1 fish. Regarding the named i-muc gene, a diet effect was not statistically significant for this exclusive fish gene that was mostly expressed in the posterior intestine of GSB. D2 induced a significant down-regulation of il8 in anterior intestine, and both D2 and D3 diets down-regulated the expression of cd4 and lgals8 in the anterior intestine segment. D3 diet did not modify significantly the expression of membrane and secreted IgM isoforms, although this diet significantly increased the expression of the membrane IgT (mIgT) in posterior intestine.

Globally across the intestine, the dietary-mediated effects affected the expression of 15 out of the 35 genes studied (43%). The heatmap in Fig. 3 was constructed with the expression values of those differentially regulated genes in order to obtain a visual overview of the changes. Fish fed D2 diet revealed an overall down-regulation of genes related to intestinal integrity (cdh1) and energy sensing (sirt1-3) mainly affecting the anterior intestine. Fish fed D3 diet showed more prominent effects on posterior intestine with up-regulation of genes related to immune response (IgT, il6).