The effect of fish stocking density and dietary supplementation of vitamin C and micronutrients (Mn, Zn and Se) on the development of systemic granulomatosis in juvenile meagre ( Argyrosomus

Systematic granulomatosis is a chronic disease that affects the majority of farmed meagre ( A . regius ). Nutritional imbalances and overcrowding can increase the risk to suffer oxidative stress, and consequently, impact the incidence of granulomatosis. In order to evaluate this, juvenile meagre were fed five isolipidic (16.7%) and isoproteic (49.6%) fish meal and fish oil- based feeds prepared by adding different levels of vitamin C, minerals (Mn, Zn, Se) with constant vitamin E and K (300 and 35 mg/kg respec-tively): Diet KEC (100 mg/kg C), Diet KEC + Mn/Zn/Se (100 mg/kg C, 40 mg/kg Mn, 200 mg/kg Zn, 1.5 mg/kg Se), Diet KECC (600 mg/kg C), Diet KECCC (1200 mg/ kg C), Diet KECCCC (3200 mg/kg C). All diets were tested at 3.20 kg/m 3 , but diets KECC and KECCCC were also tested at 6.20 kg/m 3 . Growth performance was only affected by stocking density, being lower at high density. The percentage of fish with granulomas was significantly lower in fish fed with the highest dietary vitamin C contents (KECCC and KECCCC) at low density. TBARS content was correlated with the percentage of granulomas in the liver ( R 2 = 0.9439, y = 0.003 x − 0.1242) denoting the involvement of an imbalance oxidative status in the appearance of granulomas. The present results show that high levels of vitamin C (1200– 3200 mg/kg C) and low stock density (3.20 kg/m 3 ) favours the growth of juvenile meagre, reducing the lipid peroxidation indicators and decreasing the incidence of granulomas, which confirms that this pathology is mostly triggered by the deficiency of antioxidant nutrients, particularly vitamin C.


| INTRODUC TI ON
The Mediterranean production of gilthead sea bream (Sparus aurata) and European sea bass (Dicentrarchus labrax) is consolidated, and their production is saturating the fish market. In the last years, there has been an increase in the relevance of meagre, (Argyrosomus regius, Asso, 1801), as a farmed species in the Mediterranean aquaculture. However, the intensive farming of meagre is impacted by systemic granulomatosis. The pathology of this chronic disease demonstrates a large impact in farmed meagre, causing low mortalities, but with high prevalence and intensity (Ghittino et al., 2004). Although some infectious agents like the bacteria Nocardia spp. can produce granulomatosis in meagre (Elkesh et al., 2012), there are evidences that this pathology can be caused by a nutritional imbalance, given that an infectious agent could not be associated with the appearance of granulomas (Katharios et al. 2011;Cotou et al., 2016;Ruiz et al., 2018Ruiz et al., , 2019aRuiz et al., ,2019bCarvalho et al. 2018).
Granulomas can firstly appear at very early stages in meagre larvae (20 days post hatching; dph), although its development could be avoided by the modulation of the feeding sequence (Ruiz et al., 2019a) and by using a microdiet with high enough levels of the antioxidant vitamins E and C (Ruiz et al., 2019b). Inadequate levels of these two vitamins could cause a nutritional imbalance between pro-oxidants and antioxidants, which could in turn lead to a status of oxidative stress. Indeed, a previous study demonstrated that the dietary supplementation with vitamin E and C (450 mg/kg vitamin E, 230 mg/kg vitamin C) in feeds for on-growing meagre the incidence and severity of granulomas was lower in liver together with a promotion of fish growth (Ruiz et al., 2018). The effects of dietary inclusion of vitamin D 3 have also been studied in the development of systemic granulomatosis in juvenile meagre, although the dietary inclusion of this vitamin did not seem to prevent the appearance of this disease (Cotou et al., 2016). A deficiency of antioxidants in the diet can cause primary lesions in the tissues and these lesions could potentially lead to a status of oxidative stress (Abdel-Haimed et al. 2012;Betancor et al., 2012). Oxidative stress is caused by the presence of reactive oxygen species (ROS) which are continuously produced in various metabolic pathways in all organisms. A consequence of ROS attack is an oxidative modification of membranebound lipids resulting in membrane bilayer damage, together with alterations in the regulation of cellular metabolism (Farooqui & Horrocks, 1998). Indeed, the oxidative stress and ROS have been related to the appearance of granulomas in humans (Facchetti et al., 1999). Nevertheless, ROS action can be inhibited or delayed by the effect of the antioxidant nutrients and enzymes such as superoxide dismutase, glutathione peroxidase and catalase (Pokorny & Korczak, 2001). These antioxidant enzymes can prevent the cascade of oxidant reactions, intercepting and inactivation the reactive intermediates of oxygen and their expression could be related to the appearance of granulomas.
Additionally, there are other micronutrients that can act as antioxidants. Zinc (Zn) has antioxidant properties and protects tissue from oxidative damage (Ho & Ames, 2002). Selenium (Se) is an exogenous antioxidant involved in the prevention of oxidative stress (Biller-Takahashi et al., 2015;Felton et al., 1996;Silva-Brito et al., 2016) being one of the components of glutathione peroxidase (GPX) reducing hydroperoxides at the expense of reduced glutathione (Arteel & Sies, 2001). Manganese (Mn) is another mineral involved in many cellular processes including lipid, protein and carbohydrate metabolism and acts as a cofactor or activator for many enzyme systems, such as Mn superoxide dismutase (Mn-SOD) (Andreini et al., 2008;De Rosa et al., 1980). One strategy to increase the benefits of intensive aquaculture is to optimize the space with high stock densities. However, overcrowding can be a stressful factor (Ashley, 2007;Wendelaar Bonga, 1997) and negatively affect fish growth, reproduction and immune system (Barton & Iwama, 1991;Di Marco et al., 2008;Pickering, 1998;Shubha & Reddy, 2011). Dietary supplementation of vitamin C and E has been shown to modulate the stress response (Kolkovski et al., 2001;Montero et al., 1998) including in fish reared under different stocking density (Belo et al., 2005;Montero et al., 1999), but little is known about its effects on meagre reared at different stocking densities. The effect of the stock density has been studied in meagre larvae, observing that high rearing density could reduce growth and survival, and also decrease the resistance to a challenge (Estévez et al., 2007;Roo et al., 2007). Additionally, negative effects have been associated with high stock density such as reduced growth and food conversion in juvenile rainbow trout (Oncorhynchus mykiss) (Procarione et al., 1999), or increased plasma cortisol and free amino acids in Senegalese sole (Solea senegalensis) (Costas et al., 2008). In this sense, high densities can activate the fish stress response affecting negatively different metabolic pathways (Costas et al., 2008;Laiz-Carrión et al., 2012). Therefore, non-optimal culture conditions could affect the normal growth of meagre and even be a source of stress, which in turn could potentially facilitate the appearance of disease.
The objective of the present study was to elucidate the involvement of the dietary vitamins E and C and the addition of the antioxidant minerals Mn, Zn and Se on the appearance and incidence of systemic granulomatosis in meagre. Additionally, the effect of the stock density on granulomas appearance was tested in fish fed two different levels of vitamin C levels (600 and 3200 mg/ kg respectively). A basal dietary vitamin E level of 300 mg/kg was used given that it has been previously demonstrated that potentiates the growth of meagre (Ruiz et al., 2018). To reach this objective, diets containing graded levels of vitamin C were fed to juvenile meagre and growth, survival, biochemical composition (total lipids, fatty acid profiles and thiobarbituric acid reactive substances), histopathological evaluation and gene expression of antioxidant enzymes (catalase, superoxide dismutase and glutathione peroxidase) were determined. Additionally, another experimental feed was formulated to contain higher levels of Mn, Zn and Se. The relationship between the presence of granulomas and the oxidative status in the different tissues affected with this condition is evaluated in the present trial.

| Fish and feeding
The experiment was carried out at the ECOAQUA facilities (Taliarte, Canary Islands, Spain). The juvenile meagre were obtained from induced spawns of F2 broodstock at the ECOAQUA facilities. Meagre larvae were fed with a standard feeding protocol. From 3 to 21 days after hatching (dah) were fed with enriched rotifer with Ori-Green (0.15-0.25 g/million rotifers) twice a day. From 12 to 30 dah were fed with enriched Artemia with Ori-Green (0.8 g/million Artemia) twice a day (8:00 a.m. and 15:00 p.m.). During the co-feeding period, microdiets Gemma Micro 150 and 300 (Skretting, France) were used (15% biomass/day). Juveniles were fed a commercial diet until the start of the experiment. Juveniles were acclimated to the experimental conditions and the basal diet for 2 weeks. Five isolipidic (16.7% lipid) and isoproteic (49.6% protein) fish meal and fish oil-based feeds were prepared by adding different levels of vitamin C, Mn, Zn and Se and the addition of vitamin E and K (300 and 35 mg/kg respectively) by Skretting ARC (Stavanger, Norway). Diet KEC (100 mg/kg C), Diet KEC+Mn/Zn/Se (100 mg/kg C, 40 mg/kg Mn, 200 mg/kg Zn, 1.5 mg/ kg Se), Diet KECC (600 mg/kg C), Diet KECCC (1200 mg/kg C), Diet KECCCC (3200 mg/kg C). The inclusion level of vitamin E was based in a previous trial (Ruiz et al., 2019) where 250 mg/kg of vitamin E proved to maximize meagre growth without potentiating the appearance of granulomas. The formulation, proximate composition and fatty acid content of each feed are shown in Table 1 and Table S1.
The juvenile meagre were fed three times per day (8:00, 11:30, 15:00), 6 days per week for 90 days with the different experimental diets. All the uneaten feed was daily collected 1 h after feeding from each tank with a strainer through the tank drain and dried in order to calculate the daily feed intake. Dead fish were recorded daily and survival was determined. The experiment was carried out in 21 fibre glass tanks of 500 L with 100 fish per tank (3.20 kg/m 3 ) for all the diets.
All tanks were covered with a net to prevent escapes. During the feeding trial the temperature and dissolved oxygen concentration were measured twice a week with values ranging from 18.2 to 21.5℃ and from 5.5 to 6.5 mg/L respectively. Fish were reared under natural light conditions throughout the feeding trial.
All procedures were conducted in accordance with the regula-

| Sample collection
At the beginning (day 0, prior to the start of the feeding trial; n = 50 fish from the stock tank) and at the end of the experimental trial (90 days of feeding; n = 30 fish per treatment, 10 fish per tank) fish were sacrificed with an overdose of anaesthetic (clove oil; Guinama, Valencia, Spain) and samples of liver, kidney, heart and spleen were collected and fixed in 4% buffered formalin for histological analysis. Additionally, 5 fish per tank (n = 15 per treatment) were sacrificed in ice and liver, heart and kidney removed and frozen at −20℃ for biochemical analysis. Four fish per tank (n = 12 per treatment) were also sacrificed and the same tissues collected, pooled, stabilized in RNA later (Sigma, Poole, UK) and stored at −80℃ until RNA extraction.

| Growth performance
At the beginning (day 0, prior to the start of the feeding trial) and end of the trial, fish were anaesthetized with clove oil (50-70 mg/L) and individual whole body weight and standard length recorded. All fish in the tanks were measured/weighed. Fish were unfed for 24 h before all samplings. Means and standard deviations of each triplicate were calculated for each treatment.

| Histopathology
Samples, previously fixed in 4% buffered formalin, were dehydrated in a series of different concentrations of ethanol and embedded in a paraffin block. The samples were cut at 4 μm, fixed to the microscope slide, heated and finally stained with haematoxylin and eosin (H&E), Ziel-Neelsen (ZN) (Martoja & Martoja-Pearson, 1970), Fite-Faraco (Fite et al., 1947) and Gram stain (Gregersen, 1978). Then, the samples were used for histopathological evaluation. At the sampling, all the tissues were analysed to see if macroscopic granulomas were present. The H&E stain was used to analyse the presence of granulomas in the target tissues, while the stainings of ZN, Fite-Faraco and Gram stain were used to detect the presence of a possible bacteria that could produce the presence of granulomas.

| Biochemical analysis
The analysis of feed and fish biochemical composition was conducted following standard procedures. Total lipids in liver, heart, kidney and feeds were extracted with a chloroform-methanol (2:1 v/v) mixture as described by Folch et al. (1957). Protein content (Kjeldahl method), dry matter and ash were determined according to AOAC (2010).
Fatty acids from total lipids were prepared by transmethylation as described by Christie (1982). Fatty acid methyl esters (FAMES) were separated and quantified by gas-liquid chromatography following the conditions described by Izquierdo et al. (1992).  phase was composed of a phosphate buffer at a flow rate of 0.8 ml/ min. It was used a wavelength of 254 nm and a Gemini C18 column, 5 µm particular size and 150 × 4.6 mm fitted with a Gemini pre-column of the same material to determinate the vitamin C concentration and was achieved by comparison with tris (cyclohexylammonium) ascorbic acid-2-phosphate (Sigma-Aldrich) as the external standard.
The mineral analysis of zinc, selenium and manganese of the experimental diets was performed at NIFES (Bergen, Norway).
Approximately 0.2 g of diet was digested in 2 ml of HNO 3 (69% w/w) and 0.5 mL of H 2 O 2 (20% w/w) in a microwave system (Julshamn et al., 2007). The digested sample was diluted to a final volume of 25 ml with Milli-Q water. The analysis was done by inductively coupled plasma mass spectrometry (ICP-MS; iCAP-Q and FAST SC-4Q DX autosampler, both Thermo Fisher Scientific Inc, Waltham, Massachusetts, USA) after acidic digestion of the feeds.
TBARs were measured in triplicate from extracted total fatty acids (10 mg/ml) of the liver, kidney and heart according to Burk et al. (1980). Thiobarbituric acid (TBA) is reacted with malondialdehyde (MDA, a product from the oxidation) which is resulting in a colour compound, which can be determined spectrophotometrically.
Briefly, 50 µl of 0.2% (w/v) butylated hydroxytoluene (BHT) in ethanol was added to 200 µl of lipid. Next, freshly prepared 0.5 ml of 1% (w/v) TBA and 0.5 ml of 10% (w/v) trichloroacetic acid were added to the sample. All reagents were mixed in a stoppered test tube and heated in darkness at 100℃ for 20 min. Then, samples were cooled in ice for 5 min and particulate matter removed by centrifugation at 2000 g (Sigma 4K15; Osterode am Harz, Germany) for 5 min.
The supernatant was read in a spectrophotometer (Evolution 300; Thermo Scientific, Cheshire, UK) at 532 nm and recorded against a blank sample. The concentration of TBA-malondialdehyde (MDA) was expressed as μmol MDA per g of tissue and was calculated using the extinction coefficient 0.156 μM −1 cm −1 .

| Histopathology scoring
The severity of granulomatosis was individually scored in each organ as described in a previous study (Ruiz et al., 2018). The severity of the granulomas was classified in each organ depending on the number of granulomas observed during the microscopy evaluation. The average severity was classified in liver, kidney and heart according to the criteria shown in Table S2.

| Gene expression
Kidney, liver and heart were aseptically collected from four fish per tank at the final sampling and stored at −80℃ until further analysis. Total RNA was extracted from, ~100 mg of sample using TRI  Table 2).
The relative transcript abundance of gpx, sod and cat was determined by quantitative real-time PCR (qPCR). Primer efficiency for each gene was previously evaluated to ensure that it was close to 100%.
The relative expression values were normalized against the geometric mean of the three housekeeping genes, β-actin (bact), elongation factor 1α (ef1α) and tubulin (tub), using the method described by Pfaffl (2001). All PCRs were performed using a Biometra TOptical Thermocycler (Analytik Jena, Goettingen, Germany) in 96-well plates in duplicate using 10 μl Thermo Scientific Luminaris Color Higreen qPCR Master Mix (Bio-Rad Hercules, California), 1 μl of forward and reverse primers (100 pmol/μl), 6 μl water nuclease-free and 5 μl of a 1:10 dilution of the cDNA, with the exception of the housekeeping genes, which were determined using 2 μL of cDNA, in a final volume of 20 μl. In addition, amplifications were carried out with a systematic negative control (NTC-non-template control) containing no cDNA.
The PCR conditions were an uracil-DNA glycosylase pretreatment at 50℃ for 2 min, a denaturation at 95℃ for 10 min, TA B L E 2 Sequences of the forward and reverse primers (5′-3′) used for real-time quantitative-PCR  The tissue presenting the highest number of granulomas was the liver followed by kidney and heart, whereas no granulomas were observed in neither spleen nor muscle. The lowest number of fish with     hepatic granulomas was observed when the high level of vitamin C was added to the feeds (KECCC-KECCCC diet) ( Table 5).

|
The severity score did not show significant differences among fish fed the different dietary treatments in any tissue after 90 days of feeding the experimental diets (Table 6). However, there was a tendency towards a decrease in the severity of granulomatosis in the liver along with an increase in vitamin C content, for instance 1.37 in diet KEC versus 0.93 in diet KECCCC.

| Effect of stocking density and vitamin C
The appearance of granulomas was affected by the diet, stock density and the interaction between them, being significantly lower in fish fed with the highest addition of vitamin C (KECCCC) at low density (3.20 kg/m 3 ) (Figure 2).
The only organ with differences in the incidence of granulomas among fish fed the different dietary treatments was the liver. The incidence of granulomas was only affected by the diet, being lower in the fish fed the diet with higher addition of vitamin C, regardless of the density (63.3% of affected fish with granulomas in liver) ( Table 7). The scored severity in liver, kidney and heart was not affected neither the diet nor the density (Table 8).

| Effect of vitamins and minerals supplementation
There were no differences in the tissue proximate composition of liver, kidney and heart among the fish fed the different experimental diets (Table S3). The tissue fatty acid profile (Table S4) reflected the dietary fatty acid content (Table S1). The highest levels of total monounsaturated fatty acids were observed in the liver, followed by the kidney and heart; however, the total omega-3 (n-3) and total polyunsaturated fatty acid (PUFA) were higher in the heart, followed by the kidney and liver. All the other fatty acids were similarly distributed in the three tissues. Eicosapentaenoic (20:5n-3; EPA), docosahexaenoic (22:6n-3; DHA) and arachidonic acid (20:4n-6, ARA) contents were similar between the diets, being in average higher in the heart (8.9%, 27.8% and 3.34% respectively) compare with the kidney (9.8%, 17.4% and 2.12% respectively) and liver (4.2%, 6.9% and 0.54% respectively).

TA B L E 6
Average granuloma severity scores in liver, kidney and heart of meagre (Argyrosomus regius) fed diets with different levels of C and Mn, Zn and Se at the beginning and after 90 days of feeding the experimental diets

TBARs content
The level of lipid peroxides, as indicated by TBARS content (μmol/g tissue), was significantly lower in liver, heart and kidney of fish-fed diets with high addition of vitamin C (KECCC and KECCCC) (

TA B L E 9
Thiobarbituric reactive substances (TBARS) content in liver, kidney and heart of juvenile meagre (Argyrosomus regius) after 90 days of feeding the experimental diets appearance of granulomas was observed. The supplementation of Mn, Zn and Se did not reduce this indicator of lipid peroxidation.

| Effect of stocking density and vitamin C
Density did not affect the proximate composition or fatty acid profile of the liver, kidney and heart between fish fed the different dietary treatments (Tables S5, S6). All fatty acids were similarly distributed among all treatments.

TBARS content
TBARS values were affected by the different stock density. The two-way ANOVA revealed that high stock density (6.20 kg/m 3 ) had a stronger influence in the TBARS content than the supplementation of vitamin C, increasing the lipid peroxidation in the liver, kidney and heart, being the TBARS content in the heart also influenced by the interaction between diet and density (Table 10).

| Effect of vitamins and minerals supplementation
There were significant differences in the expression levels of cat, gpx and sod in the liver (Figure 3a).

| DISCUSS ION
In the present study, the inclusion of vitamin E, Mn, Zn, Se and different levels of vitamin C did not affect growth parameters. The  the larger initial size of the fish (62.9 g) in the former study. In previous studies, it was observed that the addition of vitamin E and C in the diet (300 and 70 mg/kg respectively) could affect final growth in on-growing meagre (Ruiz et al., 2018). However, in the present study, the concentration of vitamin E was lower, which suggests that to promote growth in meagre higher levels of this vitamin are required. Significant differences in growth, SGR and FCR were observed between the two stocking densities, being better in the treatments with low density (100 fish per tank). The negative effect of stocking density has also been observed in other fish species, such as juvenile rainbow trout (Procarione et al., 1999), Dover sole (Solea solea) (Schram et al., 2006), Atlantic cod (Gadus morhua L.) (Lambert & Dutil, 2001) and Amur sturgeon (Acipenser schrenckii) (Li et al., 2012). The reduction in growth at high stock density, could be due to the heterogeneous size of the fish as a consequence of the dominance of some fish in the tank. In this sense, larger sized fish exercise their dominance in the areas where the food is supplied preventing an adequate food access to those of smaller size (Alanärä & Brännäs, 1996;Grand & Grant, 1994;McCarthy et al., 1999). Moreover, it has been shown that the high stock density produces stress and, in consequence, the elevation of cortisol levels in different fish species (Ellis et al., 2002;North et al., 2006;Ortuño et al., 2001;Turnbull et al., 2005;Wuertz et al., 2006). When the levels of cortisol are elevated in plasma, a variety of secondary physiological responses take place, including an increase in the metabolic rate (Lankford et al., 2005), which in turn leads to a decrease in growth (Heath, 1995).
In this experiment, fish exposed to high stocking densities may have been experiencing chronic stress, as could be observed by the increase in the MDA content in all the analysed organs. A similar effect was observed by  in Atlantic salmon (Salmo salar) and by Jia et al. (2016) in turbot, where the fish farmed at a high stock density showed higher levels of TBARS compared with those in lower densities. It has been shown that high stock density in fish can disturb the balance between the production and removal of ROS, inducing lipid peroxidation (Andrade et al., 2015;Sahin et al., 2014).
Therefore, the higher stock density might induce a higher stress resulting in an increase in lipid peroxidation and cell degradation.
Additionally, the level of lipid peroxides was lower in all the an-  (Harats et al., 1990;de Zwart et al., 1999). Similar results were obtained by Betancor et al. (2012), where the vitamin C dietary supplementation, markedly improved the protection against peroxidation, decreasing MDA contents to less than one-third in sea bass larvae.
In previous studies in meagre larvae, a strong positive correlation between the TBARS content and the appearance of granulomas was observed (Ruiz et al., 2019a(Ruiz et al., ,2019b, suggesting a possible relationship between lipid peroxidation and the appearance of granulomas. In the present study, a modest correlation between the TBARS levels and the appearance of granulomas in liver (R 2 = 0.651, y = 487.63x + 43.914) and heart (R 2 = 0.648, y = 206.75x + 11.087) was observed. Probably this correlation is more attenuated than the observed in previous studies in meagre larvae at 30 dph (R 2 = 0.948, y = 0.084x − 4.3924) (Ruiz et al., 2019a) and 44 dph (R 2 = 0.892, y = 0.0446x + 0.0756) (Ruiz et al., 2019b), due to the higher initial incidence of granulomas in juvenile meagre in the present trial (61%).
In relation to the presence of granulomas, the increase in the dietary level of vitamin C from 100 mg/kg to 1200-3200 mg/kg, the percentage of fish with granulomas in any tissue at low density was significantly lower, from 87% in diet KEC to 80 and 76% in diet KECCC and KECCCC respectively. The liver was the main affected organ (71.5%) followed by kidney (62.2%) and heart (7.6%), and the incidence of granulomas in liver was significantly lower in fish-fed diet KECCC and KECCCC at low density, and in fish-fed diet KECCCC*2. Indeed, the percentage of fish with granulomas in any tissue was affected by the diet, but also by the stock density. The high stock density can disturb the balance between the production of ROS and its removal (Andrade et al., 2015;Sahin et al., 2014), this imbalance potentially leading to a status of oxidative stress which may cause different diseases and lesions (Betancor et al., 2012;Cowey et al., 1984;Kawatsu, 1969;Lewis-McCrea & Lall, 2007;Sakai et al., 1989;Watanabe et al., 1989). In the present study, the TBARS levels were affected by the stock density, being higher in fish farmed at high density. Therefore, the present results suggest that the high stock density is a stressful factor that can lead to oxidative stress, lipid peroxidation and promote the appearance of granulomas.
On the other hand, in the present study the supplementation of Se, Mn and Zinc did not seem to prevent lipid peroxidation, as indicated by TBARS levels. This is probably due to the fact that this species needs a higher supplementation of these minerals in order to see an effect on the antioxidant system. For instance Mansour et al. (2017) observed that the supplementation of 3.98 mg/kg of selenium improved the antioxidant balance and inmate immune status of juvenile meagre, while in the present experiment it was supplemented with 1.6 mg/kg. Little is known about the requirements and its effects on the antioxidant system of these three minerals.

F I G U R E 4
Expression levels of cat, sod and gpx measured by real-time PCR in liver, kidney and heart of meagre (Argyrosomus regius) after 90 days fed different levels of vitamin C and cultured under two different densities (Diets KECC and KECCCC at 3.20 kg/m 3 and diets KECC*2 and KECCCC*2 at 6.20 kg/m 3 ). Di, diet; De, density; DI*DE, diet-density interaction. Values are normalized expression ratios, corresponding to an average of six individuals (n = 6) with standard errors (SEM). Different superscript letters denote differences between treatments identified by one-way ANOVA. n.s., not significant. *p < 0.05. **p < 0.01 (n = 6) The lower percentage of granulomas and TBARS values in fish fed with a high addition of vitamin C (3200 mg/kg 3 ) at low densities was also accompanied by the increase in the expression of some antioxidant enzymes capable to neutralize ROS, named sod, gpx and cat. In the present study, the expression of these enzymes was affected by the addition of different levels of dietary vitamin C.
Expression of sod and gpx was increased in liver and heart of fish fed the diet supplemented with 3200 mg/kg vitamin C.  (Dawood et al., 2016) and sea bass (Betancor et al., 2012). The present results seem to indicate that dietary vitamin C has antioxidant potential by enhancing the expression of sod and gpx in liver and heart, and cat in liver and kidney, being influenced by the organ where they are acting. Moreover, the reduction in the TBARS levels and in the percentage of granulomas in fish fed high addition of vitamin C, suggests that this vitamin has a protective effect against oxidative stress, therefore, preventing the appearance of granulomas.
Nevertheless, not only the supplementation of vitamin C had an effect on the expression of antioxidant genes, but also the density. In liver sod expression was affected by the stock density, with fish-fed diet KECCCC and reared at low density (3.20 kg/m 3 ) displaying a higher expression of sod than fish fed with the same diet but reared at a high density (6.20 kg/m 3 ). The expression of gpx was also up-regulated in heart of fish-fed diet supplemented with 3200 mg/kg of vitamin C at low density but was significantly lower in the same diet but at high density. Similar to these results, Jia et al. (2016) observed that juvenile turbot reared under three different stock densities (5.1, 7.7 and 10.8 kg/m 2 ), showed a lower expression levels of sod, gpx and cat after 80 days when reared at the higher stock density. The high stock density has been established as a source of stress that increases the production of ROS (Braun et al., 2010;Vijayan et al., 1990). These ROS could damage cellular components and lead to autoxidation (Sayeed et al., 2003). It has been observed that the increase in ROS can induce a variety of secondary physiological responses, such as a decrease in the antioxidant capacity and immune function (Aksakal et al., 2011;Sadhu et al., 2014). In our experiment fish fed under high stock density, presented a high content of MDA together with down-regulation of antioxidant genes, which indicates a higher oxidative stress in those fish reared at 6.20 kg/m 3 .
To conclude, meagre farmed at a high stock density (6.20 kg/ m 3 ) for 90 days showed reduced growth, SGR and a higher FCR, regardless of the dietary content. The addition of different levels of vitamin C and the supplementation of Mn, Se and Zn did not affect growth parameters. TBARS values in kidney, liver and heart at low stocking density were lower when fish were fed with high addition of vitamin C (1200-3200 mg/kg). Also, the percentage of granulomas in any tissue and particularly in liver, where the TBARS values were highly correlated with the percentage of granulomas (R 2 = 0.9439, y = 0.003x -0.1242) was lower in fish farmed at low density. In addition, fish-fed KECCCC showed the highest expression of cat in liver and kidney as well as the highest expression of sod and gpx in liver and heart at low stock density. The present results show that low stock density (3.20 kg/m 3 ) favours the growth of juvenile meagre and that high levels of vitamin C (1200-3200 mg/kg C) can reduce the lipid peroxidation indicators as well as decrease the incidence of granulomas. The results seem to suggest that this pathology is mostly triggered by a deficiency of antioxidant nutrients, as high levels of vitamin C led to a lower percentage of granulomas in the liver. However, the percentage of fish with granulomas increased along the experimental trial regardless of the dietary treatment, which opens the possibility that other factors may be involved in the appearance of the disease.
Given that systemic granulomatosis has been diagnosed in very early life stages, a genetic origin of the pathology could be speculated. Nevertheless, this pathology has also been described in several other locations such as Italy or Greece, it is highly unlikely that the genetic background of the fish plays a role in the appearance of the disease.

DATA AVA I L A B I L I T Y S TAT E M E N T
Research data are not shared.