Effect of rearing density on nervous necrosis virus infection in Senegalese sole (Solea senegalensis)

J Fish Dis. 2021;44:2003–2012. wileyonlinelibrary.com/journal/jfd | 2003 1 | INTRODUCTION The nervous necrosis virus (NNV), genus Betanodavirus, is the aetiological agent of viral encephalopathy and retinopathy (VER), a neurological pathology affecting fish worldwide. NNV is a small nonenveloped virus containing bisegmented ssRNA (+). The largest segment, RNA1, encodes the protein A or RNA- dependent RNA polymerase (RdRp), whereas the RNA2 segment codes the capsid protein (Mori et al., 1992). NNV is clustered into four genotypes, according to a classification based on the T4 region (RNA2): barfin flounder nervous necrosis virus (BFNNV), redspotted grouper nervous necrosis virus (RGNNV), striped jack nervous necrosis virus (SJNNV) and tiger puffer nervous necrosis virus (TPNNV) (Nishizawa et al., 1997). Furthermore, reassortant strains between RGNNV and SJNNV (RGNNV/SJNNV and SJNNV/RGNNV) have been isolated in southern Europe from several fish species, including sole (Solea senegalensis and S. solea), European seabass (Dicentrarchus labrax) and gilthead seabream (Sparus aurata), among others (Olveira et al., 2009; Panzarin et al., 2012; Toffan et al., 2017; Toffolo et al., 2007; Volpe et al., 2020). Nervous necrosis virusinfected fish show abnormal swimming behaviour, provoked by the cytopathic damage (necrosis and vacuolization) of the nervous tissue, and can transmit the virus to cohabitant fish by shedding infective particles into the water column Received: 14 July 2021 | Revised: 9 August 2021 | Accepted: 10 August 2021 DOI: 10.1111/jfd.13514 RESEARCH ARTICLE Effect of rearing density on nervous necrosis virus infection in Senegalese sole (Solea senegalensis) Lucía Vázquez- Salgado | Jose G. Olveira | Carlos P. Dopazo | Isabel Bandín Departamento de Microbiología y Parasitología, Instituto de Acuicultura, Universidade de Santiago de Compostela, Santiago de Compostela, Spain Correspondence Isabel Bandín, Departamento de Microbiología y Parasitología, Instituto de Acuicultura, Universidade de Santiago de Compostela, Santiago de Compostela 15706, Spain. Email: [email protected] Funding information Interreg VA Spain- Portugal cooperation program (POCTEP) 2014- 2020, 0474_ BLUEBIOLAB project, cofunded by FEDER Abstract Intensive fish farming at high densities results in a wide range of adverse consequences on fish welfare, including pathogen spreading, stress and increased mortality rates. In this work, we have assessed whether the survival of Senegalese sole infected with the nervous necrosis virus (NNV), a pathogen responsible for severe disease outbreaks, is affected by rearing density. Based on the different fish ratios per surface area (g cm−2) and water volume (g L−1), our research showed an earlier mortality onset in the tanks containing NNV- infected fish reared at medium density (MD: 0.071 g cm−2/5 g L−1) and high density (HD: 0.142 g cm−2/10 g L−1), as well as higher cumulative mortality values. However, transcription analysis of hsp70, gr1 and pepck genes, wellknown stress biomarkers, seems to indicate that none of the challenged fish were under high stress conditions. NNV load was slightly higher both in dead and in sampled fish from MD and HD groups, and especially in the rearing water from these groups, where peaks in mortality seemed to correlate with increasing NNV load in the water. In conclusion, our results suggest that rearing NNV- infected Senegalese sole at high densities resulted in an earlier mortality onset and higher cumulative values and viral load. KEYWORDS cumulative mortality, nervous necrosis virus, rearing density, Senegalese sole, viral load This is an open access article under the terms of the Creative Commons Attribution- NonCommercial- NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is noncommercial and no modifications or adaptations are made. © 2021 The Authors. Journal of Fish Diseases published by John Wiley & Sons Ltd. 2004 | VÁZQUEZSALGADO Et AL. through the skin and gills (Souto et al., 2018). NNV horizontal transmission can severely jeopardize intensive farming strategies developed to maximize production given that density is a factor commonly related to VER outbreaks (Bandín & Souto, 2020). In addition, rearing fish above a density threshold may adversely affect fish performance, including slower growth, reduced food intake and alterations in normal behaviour and immune response (Tort, 2011; Yarahmadi et al., 2016). Therefore, in order to determine whether high rearing densities can favour VER outbreaks in the aquaculture facilities, NNV experimental infections were performed in Senegalese sole stocked at different densities. 2 | METHODOLOGY 2.1 | Production of viral inoculum The betanodavirus strain used in this study was SpSsIAusc160.03, a reassortant RGNNV/SJNNV isolated from diseased Senegalese sole (S. senegalensis; Olveira et al., 2009). Semiconfluent E- 11 cell monolayers grown with Leibovitz’s 15 medium (L- 15; Gibco, Fisher Scientific) and supplemented with 5% foetal bovine serum (FBS) were used to propagate the virus (MOI 0.1). Inoculated cells were maintained at 25°C until cytopathic effect (CPE) was extensive. The viral suspension was clarified by centrifugation at 3000 g for 20 min, and 10- fold dilutions of the supernatant were inoculated in 96- well plates for 7– 10 days at 25°C. CPE development was monitored daily for 10 days, and viral titre was expressed as the 50% tissue culture infection dose per ml (TCID50/ml) (Reed & Muench, 1938). 2.2 | Fish infection challenges Senegalese sole juveniles (N = 885; mean body weight 1 ± 0.1 g) were acclimated in opaque 300- L tanks containing recirculating sea water (salinity 33 g L−1) at 22°C and fed ad libitum at the facilities of the University of Santiago de Compostela. Fish were strictly handled according to the current regulations on animal welfare (Directive 2010/63/UE) in experimental procedures, and the experimental protocol was approved by the Bioethics and Experimental Animal Welfare Committees of the University of Santiago de Compostela and Xunta de Galicia (Permit Id. 15010/2020/004). Prior to the challenges, five fish were randomly collected and killed using an MS- 222 (Sigma- Aldrich) overdose and tested for the presence of bacterial and viral pathogens. Bacterial isolation was accomplished by inoculating samples of kidney, spleen and liver onto tryptone soy agar supplemented with 1% NaCl (TSA- 1) and incubated at 25°C for 24 h. For viral analysis, RT- qPCR was performed using specific primer sets for NNV, viral haemorrhagic septicaemia virus (VHSV), infectious pancreatic necrosis virus (IPNV) and infectious haematopoietic necrosis virus (IHNV) (Olveira et al., 2013). Fish (N = 660) were bathinfected with the SpSsIAusc160.03 strain (105 TCID50/ml) for 4 h with strong aeration. Fish were then randomly distributed into 12 flatbottomed tanks (area: 700 cm2) to assay, in triplicate, different rearing densities according to tank surface and water volume (Figure 1): high density (HD: 100 fish; 0.142 g cm−2/10 g L−1), medium density (MD: 50 fish; 0.071 g cm−2/5 g L−1), medium/low density (MLD: 50 fish; 0.071 g cm−2/1 g L−1) and low density (LD: 20 fish; 0.028 g cm−2/1 g L−1). One replicate tank was allocated for fish sampling (N = 3) on days 3, 5, 7, 10, 15, 20, 25, 30 and 35 postinfection (pi). In the remaining two replicate tanks, the development of the infection in the fish was monitored (disease signs and mortality) and viral shedding was assessed. For this purpose, an aliquot of 50 ml water was collected at 1 h postinfection (hpi) and 1, 2, 3, 5, 7, 10, 15, 20, 25, 30 and 35 days postinfection (dpi). No additional sea water was added to the experimental tanks. Fish and water samples were stored at −20°C until use. A mockinfected tank per condition was also set up. Control fish were challenged with L- 15 medium and handled as infected fish. 2.3 | Virological analysis Sole tissue samples (brain and eyes) were aseptically collected and individually homogenized (1:10 w/v) in Earle's Balanced Salt Solution (Hyclone Laboratories Inc.), supplemented with amphotericin B (200 μg ml−1), gentamycin (500 μg ml−1), penicillin and streptomycin (1000 units ml−1). Survivors were analysed in pools consisting of three fish. The homogenate was centrifuged at 4000 g for 20 min at 4°C, and the supernatants were stored at −20°C until virological analysis. Water samples were twice centrifuged at 96,000 g for 1 h in an SW32ti rotor (Beckman Coulter), and pellets were resuspended in 500 µl of saline– sodium citrate (SSC) buffer 1X. Tissue and water samples were titrated as previously described. For NNV quantification, samples were subjected to RNA extraction using a NucleoSpin ® Kit (Macherey- Nagel) following the instructions provided by the manufacturer. Total RNA was incubated with random primers (200 nM) for 5 min at 95°C and then with the reverse transcription mixture containing RevertAid First Strand cDNA Synthesis Kit (Thermo Scientific Inc.), for 1 h at 42°C in a thermocycler MyCycler™ (Bio- Rad). The reaction was terminated by inactivating the enzyme at 70°C for 10 min. RT- qPCR was performed in a CFX- 96 Real- Time PCR Detection System (Bio- Rad) amplifying the cDNA (2 µl) with 200 nM of SnodR1 F/R primers (5′- TCCAAA AGAAAGAAGCATAC- 3′/5′- TGGCATGTACCACGGAAC- 3′; Olveira et al., 2013) and iQ ™ SYBR® Green Supermix (10 µl; Bio- Rad). The protocol started with an initial denaturation/activation phase for 3 min at 95°C followed by 40 denaturation and annealing– extension cycles (15 s at 95°C and 20 s at 59°C, respectively). Viral load quantification (RNA1 copies ml−1) was estimated from a standard curve consisting of 20- fold dilutions (101– 10– 7 RNA copies ml−1) of a DNA plasmid containing RNA1 full segment of SpSsIAusc160.03. 2.4 | Stressrelated gene expression Expression of stressmodulated genes, such as the heatshock protein 70 (hsp70), the glucocorticoid receptor 1 (gr1) and the phosphoenolpyruvate carboxykinase (pepck) (Kim et al., 2013), was quantified | 2005 VÁZQUEZSALGADO Et AL. in killed and dead fish to determine whether they were subjected to any stress factors. The cDNA obtained in Section 2.3 was amplified by RT- qPCR using 200 nM of each primer set and the protocol previously described, except for the annealing– extension step (Table S1). Gene expression was normalized using the 2− ΔΔCt method (Livak & Schmittgen, 2001) using endogenous gene ribosomal 18S as reference and mockchallenged fish belonging to the LD group as control. 2.5 | Nitrogen compound measurement No water exchange was performed during the challenges, so the nitrogenderived compound concentration was monitored in the tanks allocated for water sampling and mortality assessment, at the same time points as water sampling. Total ammonia (NH3 + NH + 4 ) and nitrite ( NO − 2 ) contents were estimated using API® Ammonia and Nitrite Test Kits (MARS Fishcare), respectively. 2.6 | Statistics Data were analysed using GraphPad Prism 8.1 (San Diego). Survival curves were evaluated with the Kaplan– Meier test, and differences between groups were calculated using the logrank (Mantel– Cox) and the chisquare tests. Significant differences in viral load and hsp70 expression among groups were analysed using twoway ANOVA and Tukey's correction. p values <.05 are considered statistically significant. FIGURE 1 Schematic representation of the four rearing densities in the experimental infections: low density (LD: 0.028 g cm−2/1 g L−1), medium/low density (MLD: 0.071 g cm−2/1 g L−1), medium density (MD: 0.071 g cm−2/5 g L−1) and high density (HD: 0.142 g cm−2/10 g L−1) 2006 | VÁZQUEZSALGADO Et AL. 3 | RESULTS 3.1 | Cumulative mortality and NNV quantification The first deaths (12 and 10 fish) were observed at 1 dpi in the tanks containing the high (HD) and the medium (MD) rearing densities, respectively (Figure 2). In both groups, 50% of the fish had died by Day 17 postinfection (pi) and at the end of the experiment (35 dpi), 73% and 67% cumulative mortalities were recorded in HD and MD, respectively. In the MLD and LD challenges, mortality began on Day 12 and 26 pi, and finally reached 46% and 27%, respectively. These cumulative mortalities were significantly lower than those of the MD and HD groups (p value <.05; chisquare test). VER signs, including loop swimming, anorexia and lethargy, were observed in fish from all the density challenges from Day 10 pi onwards, except for the LD group, in which disease signs were first recorded on Day 13 pi. No VER signs were recorded in the mockinfected groups, and no viral agents were detected in the fish belonging to these groups, although two fish from HD mockchallenged groups died on the first day. The assessment of the viral load in infected fish (Figure 3) revealed that the first individuals that succumbed to death in the low- and mediumdensity (LD: 26 dpi and MLD: 12 dpi) challenges showed around 107 copies g−1/104 TCID50 g−1. In the LD tank, the NNV load increased slightly over time showing 1.59 × 1010 copies/2.1 × 104 TCID50 at 31dpi (Figure 3a), whereas in the MLD tank, after decreasing between 26 and 31 dpi, the NNV load peaked on Day 32 (5 × 1010 copies g−1 and 1.33 × 105 TCID50 g−1) (Figure 3b). Regarding the HD and MD challenges, the individuals that died at 1 dpi carried around 6 × 106 NNV copies g−1 (2 × 103 TCID50 g−1). NNV load increased over time in both groups and peaked in the MD tank on Day 26 pi (6.31 × 109 copies g−1/2 × 105 TCID50 g−1) (Figure 3c) and in the HD tank on Day 33 pi (9.25 × 1010 copies g−1/108 TCID50 g−1) (Figure 3d). In the final days of the experiment, a decrease in the viral load was observed in the fish from the HD group. 3.2 | Viral load in sampled fish Periodic sampling indicated that fish from all challenges shared a similar viral load in the first days of infection (Figure 4), but the final values were significantly higher in the HD and MD groups (9.7 × 108 and 1.5 × 109 copies, respectively), especially regarding viable particles (HD: 3.1 × 106, MD: 2.1 × 105 TCID50), than in the LD and MLD fish (around 103 TCID50). Besides, different trends were observed between the challenges throughout the trial. Thus, in Senegalese sole belonging to the LD group (Figure 4a), a progressive increase in the viral load was observed up to the 20 dpi (1.5 × 108 RNA copies/8 × 105 TCID50) although a decrease was evident at the end of the experiment (25 dpi; 3.3 × 107 copies and 103 TCID50). In the MLD and MD challenges (Figure 4b,c), viral load peaked at around 7– 10 dpi (3.2 × 1010 copies/5 × 107 TCID50 and 2.3 × 108 copies/5.6 × 106 TCID50, respectively) to subsequently decrease until Day 25 and increase again towards the end of the experiment. In the HD tank (Figure 4d), fish sampled up to 15 dpi contained a similar number of NNV copies although a slight increase in viable particles was observed up to Day 10. At 20 dpi, a sudden 3- log increase was observed in both RNA copies and viable particles (6 × 109 and 3 × 107, respectively), but whereas the RNA values were maintained until the end of the challenge (9.8 × 108), a decrease of around 1 log was observed in the number of viable particles (3.1 × 10 6). FIGURE 2 Cumulative mortality (%) recorded in Senegalese sole reared at the following densities: low, medium/low, medium and high (LD, MLD, MD and HD, respectively) for 35 days. Data are presented as mean and standard deviation (SD) of the replicate tanks (N = 2). Asterisks indicate significant differences in cumulative mortality of the highest density challenges (MD/HD) with regard to LD and MLD (p value < .05) 01235678910 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 0 20 40 60 80 100 dpi Cumulative mortality % LD MLDMD HD * | 2007 VÁZQUEZSALGADO Et AL. 3.3 | Viral load in rearing water The initial viral load in tanks (checked at 1 hpi) was very similar in all the density challenges (average 1.5 × 105 RNA copies ml−1/1.8 × 102 TCID50 ml−1), but differences were observed between the groups over the following days. In the LD group (Figure 5a), although the number of copies and titre gradually declined up to Day 5 pi (103 and 5.6 × 101, respectively), a slight increase in NNV copies was observed at 7 dpi (1.2 × 104) and it remained similar, albeit with a few fluctuations, until the end of the challenge. However, although it was not possible to quantify the viable virus at all the sampling time points, at the end of the experiment (35 dpi) viral titre was 3.1 × 102 TCID50 ml−1. Viral load in the MLD (Figure 5b) tank showed certain oscillations with a final increase in the number of RNA copies (from the initial 5 × 104 to 106) but with TCID50 values very similar to the initial ones. NNV dynamics was similar in the water from the MD and HD tanks (Figure 5c,d), showing a gradual decline of around 1.5– 2 logs up to Day 10 pi, and an abrupt increase at 15 dpi (average 2 × 106 copies ml−1; 4.5 × 103 TCID50 ml.1) that was maintained up to Day 35 pi. 3.4 | Nitrogenderived compound concentration Nitrite (NO−2) and total ammonia nitrogen (TAN; NH3 + NH4 +) concentration was monitored during the experiments (Table S2). Nitrite content ranged from 0 to 0.5 mg L−1 (0– 1 mM) in water from the LD group, while in the remaining challenges, concentration increased up to 1 mg L−1 (20 mM) at 35 dpi. With regard to TAN, great differences were observed between the density challenges. The LD tank contained 2 mg L−1 of TAN at the end of the experiment, whereas a concentration of 8 mg L−1 was reached on days 30, 15 and 5 pi in the MLD, MD and HD groups, respectively. Further increases in TAN concentration could not be detected because of the detection limit of the test kit. 3.5 | Expression of stressrelated genes Killed fish showed hsp70, gr1 and pepck transcription values below 0.1 fold change throughout the experiment in all the challenges, except for the HD group, where hsp70 and pepck expression increased to 0.3 and 0.7 fold change, respectively, on Day 25 pi (Table S3A). FIGURE 3 Log10 of NNV copies and TCID50 in dead fish and survivors (S) from lowdensity (LD) (A), medium/lowdensity (MLD) (b), mediumdensity (MD) (c) and highdensity (HD) (d) challenges. Dead individuals were analysed individually, and survivors were analysed in pools of 3. Number of analysed fish is indicated above each bar. Error bars represent mean ± SD. Asterisks indicate significant differences in viral load with respect to the first mortalities in each group (p value < .05) 26 29 30 31 S. 0 2 4 6 8 10 12 dpi log NNV load g-1 LD TCID50 Copies 2 2 1 4 13512171826282930S. 0 2 4 6 8 10 12 dpi log NNV load g-1 MD TCID 50 Copies 10 5 11 2 2 2 2 2 3 2 12 18 23 26 31 32 33 S. 0 2 4 6 8 10 12 dpi log NNV load g-1 MLD TCID50 Copies * 1 54 3 2 3 2 12512171826333 4S . 0 2 4 6 8 10 12 dpi log NNV load g-1 HD TCID50 Copies * * * ** * * 12 12 722 11 33 3 5 (a) (b) (c) (d) 2008 | VÁZQUEZSALGADO Et AL. Gene transcription in dead fish was slightly higher than in the killed ones, and an average of 1.7, 3.2 and 3.36 fold change of hsp70, gr1 and pepck genes, respectively, was detected on Day 26 pi in fish from the HD group (Table S3B). 4 | DISCUSSION Fish farming in aquaculture facilities is subjected to intensive culture conditions, and rearing densities are often raised to maximize the production and to meet the high demand for fish. However, increasing the fish number enhances the spread of pathogens by horizontal transmission and also induces crowding stress in fish, which can lead to a reduction in growth and an inhibition of the immune response (Iguchi et al., 2003; Salas- Leiton et al., 2010; Sánchez et al., 2010; Yarahmadi et al., 2016). Moreover, stocking density has been cited as a main factor in triggering VER outbreaks (Bandín & Souto, 2020; Munday et al., 2002). Therefore, in this work we have assessed how NNV infection in Senegalese sole juveniles, one of the species with the largest production in Mediterranean aquaculture (FAO, 2018), is affected by increasing rearing density. Firstly, one of the major consequences of fish farming at high stocking densities is the unexpected mortality, which has been shown to be densitydependent in flatfish species as Japanese flounder larvae (Paralichthys olivaceus), Senegalese sole or fine flounder (Paralichthys adspersus) (Castro et al., 2019; Costas FIGURE 4 Log10 of NNV copies and TCID50 values in fish sampled from lowdensity (LD) (a), medium/lowdensity (MLD) (b), mediumdensity (MD) (c) and highdensity (HD) (d) tanks at different time points. Data represent mean ± SD (n = 3) on each point. Asterisks indicate significant differences in each group with regard to 3 dpi samples (p value < .05) 357101520253035 0 2 4 6 8 10 12 dpi log NNV load g-1 LD TCID50 Copies * 357101520253035 0 2 4 6 8 10 12 dpi log NNV load g-1 MD *** TCID 50 Copies * 357101520253035 0 2 4 6 8 10 12 dpi log NNV load g-1 MLD * * TCID50 Copies 357101520253035 0 2 4 6 8 10 12 dpi log NNV load g-1 HD **** TCID 50 Copies (a) (b) (c) (d) | 2009 VÁZQUEZSALGADO Et AL. et al., 2008). However, in the control fish used in our study, only a 2% mortality was recorded in the first days postinfection in the HD tank. Regarding the infected fish, although mortality was expected because of NNV infection, a higher cumulative mortality was observed as rearing densities increased. In the lowest density group, 27% mortality was recorded at the end of the challenge, and a considerable increase (46%) was observed in the MLD group, which was significantly increased in the MD and HD challenges (67% and 73%, respectively). In this regard, it should be pointed out that the MLD and MD groups shared fish number per surface area (0.071 g cm−2) but differed in the individuals per water volume (MD: 5 g L−1 and MLD: 1 g L−1). These differences suggest that even though most of the aforementioned studies on flatfish provide density data based on the available surface area, the water volume is also an important parameter to take into consideration, as occurs in bilateral fish species (Sammouth et al., 2009). Although the upper limit for optimal sole rearing densities has been set at 0.9 g cm−2 (Costas et al., 2008), our results indicate that rearing VER- infected Senegalese sole juveniles at densities of 0.071 g cm−2 (MLD group) resulted in a significantly increased mortality. Moreover, high stocking density according to both surface and water volume (0.071 g cm−2 and 5 g L−1, MD group) resulted in a higher increase in the cumulative mortality. These findings may indicate that sole tolerance to high densities is reduced by the NNV infection. FIGURE 5 Viral load (log copies ml−1 and TCID50 ml−1) in rearing water from lowdensity (LD) (a), medium/lowdensity (MLD) (b), mediumdensity (MD) (c) and highdensity (HD) (d) tanks. Data represent mean ± SD (n = 3) on each point. Asterisks indicate significant differences with regard to 3 dpi samples (p value < .05) 01235710 15 20 25 30 35 0 2 4 6 8 dpi log NNV load ml-1 LD Copies TCID50 * 012357101520253035 0 2 4 6 8 dpi log NNV load ml-1 MD Copies TCID 50 * * 012357101520253035 0 2 4 6 8 dpi log NNV load ml-1 MLD Copies TCID50 ** **** * 01235710 15 20 25 30 35 0 2 4 6 8 dpi log NNV load ml-1 HD Copies TCID50 * (a) (b) (c) (d) 2010 | VÁZQUEZSALGADO Et AL. The effect of the rearing density in NNV replication was analysed by assessing the viral load (genomic copies and viral titration) at different time points. At the beginning of the challenge, fish from all groups contained a similar viral load, but there were clear differences as the challenge progressed. From Day 20 pi onwards, the highest levels of viable virus were observed in fish belonging to the HD group. Regarding RNA copy number although high values were observed in MLD and MD groups, the highest ones were also observed in the HD group. Differences between TCID50 and RNA values ranged from 1 to 5 logs depending on the challenge, but it must be noted that they were considerably lower in the HD challenge, which seems to indicate that high densities favour an effective viral production. Differences up to 6 logs between genome copies and viable NNV particles recovered from infected fish tissues have already been reported (Souto et al., 2015). This information was complemented with the viral load recorded in the dead fish, although in the first days of infection, data are only available for MD and HD (mortalities in the LD and MLD groups were first recorded at 12 and 26 dpi, respectively). In both groups, the viral load in the dead fish from 1 to 5 dpi was similar, just like in the killed fish, although the viral load in the MD group was more than 1 log higher than that in the sampled sole. At 26 dpi, when data were available from the four groups, the TCID50 values were more than 2 logs higher in the fish from the MD and HD groups than in the LD group, supporting the results obtained with the sampled fish. In addition, viable virus present in water samples was, on average, 1 log higher in the MD and HD groups. All these data suggest that high stocking density conditions favoured viral replication, which can account for the observed increased mortality levels. Similar findings have been reported in infections caused by other fish viruses, such as the ranavirus largemouth bass virus (Inendino et al., 2005), the novirhabdovirus infectious haematopoietic necrosis virus (Ogut & Reno, 2004) or the golden shiner virus (Schwedler & Plumb, 1982). To study whether the stress induced by the rearing density could be involved in the earlier deaths, we assessed the transcription of the hsp70, gr1 and pepck genes, wellknown stress biomarkers (Kim et al., 2013). Overexpression of hsp70 has been previously reported in Senegalese sole subjected to different stressors (Cordeiro et al., 2012; González- Mira et al., 2016), during cohabitation with gilthead seabream (Vázquez- Salgado et al., 2021), and in individuals maintained at high stocking densities (Salas- Leiton et al., 2010). Expression of gr1 gene could be down- or upregulated in adults and juveniles of different species subjected to increasing stocking densities (Kim et al., 2013; Terova et al., 2005; Valenzuela et al., 2020), while upregulation of pepck has been observed in river pufferfish (Takifugu obscurus) adults reared at high density (Kim et al., 2013). Unlike these reports, killed sole showed no significant differences in gene transcription with respect to the control fish (lowdensity mockinfected group), suggesting that the density conditions assayed in our study did not provoke stress in fish. This different behaviour could be due to, on the one hand, the higher stress tolerance of flatfish (all the fish used in the abovecited studies showed bilateral symmetry) and, on the other hand, the different sizes of fish used. However, dead animals collected throughout the challenge in the highdensity group showed moderate gene overexpression levels (hsp70: 1.7; gr1: 3.2; pepck: 3.36 fold change), which was not observed in the mockinfected fish subjected to the same density conditions. This slight upregulation could be caused by the increased viral load observed in this group. In fact, NNV infection has been previously linked with an hsp70 overexpression in Asian seabass (Lates calcarifer) (Liu et al., 2016) and with other stressrelated genes in European seabass (Lama et al., 2020). Because the quality of the rearing water was deteriorated by the accumulation of fish waste and uneaten food pellets, nitrite and TAN contents were assessed and a densitydependent increase in both parameters was observed. Although nitrite is generally less toxic in sea water than in freshwater, several adverse effects have been reported in marine fish exposed to high nitrite concentrations (Park et al., 2013). Furthermore, upregulation of hsp70 gene transcription has been reported in turbot (Scophthalmus maximus) and silver seabream (Sparus sarba) adults exposed to nitrite (Jia et al., 2015, 2016). However, hsp70 transcription values obtained in our study from the fish kept in the tanks with higher nitrite concentrations (20 mM; MLD, MD and HD groups) were very low, supporting the hypothesis that younger fish are more tolerant than adults to nitrite exposure, as previously reported (Ciji & Akhtar, 2020). With regard to the TAN, marine species have been reported to be more sensitive than freshwater fish, and damage in liver, gill or muscle is some of the consequences of the oxidative stress induced by ammonia (Randall & Tsui, 2002). In our study, fish did not show disease signs different to those commonly caused by NNV infection and no apparent lesions were observed in the internal organs, but since histological analysis was not performed, we cannot rule out oxidative damage in the tissues. Finally, we also evaluated NNV dynamics in the rearing water. At the beginning of the challenge, viral particles attached to the body surface of bathinfected fish were released into the water, resulting in the detection of a high NNV load (around 105– 106 NNV copies ml−1 and 103– 10 4 TCID50 ml−1) in all tanks. Both genomic copies and titre progressively decreased until Day 10 pi, especially in the HD and MD challenges, to values well below the initial ones. This drop in viral load suggests that shedding from infected fish did not seem to occur in the first few days, as previously reported (Souto et al., 2018). Afterwards, viral load increased until the end of the challenge in the highdensity groups (MD and HD), but not in the lowdensity challenges (LD and MLD). However, a different viral load was detected in the latter groups throughout the experiments, which could be explained by the fact that, although the ratio number of fish per water volume was the same, in the MLD group there were a larger number of potential shedder fish. In addition, the increase in viral load observed in the MD and HD groups coincided with the onset of clinical signs and the acute mortality phase, indicating a possible correlation of NNV load in the water with mortality peaks. This correlation has been reported in RGNNV- infected grouper (Epinephelus septemfasciatus) and in Asian seabass (Jaramillo et al., 2017; Nishi et al., 2016). | 2011 VÁZQUEZSALGADO Et AL. To conclude, our data indicated that high stocking densities not only resulted in an earlier mortality but also reduced the chances of Senegalese sole surviving an NNV infection. Moreover, the sole challenged at high densities (MD and especially at HD) carried a higher NNV load than fish from the lowdensity groups. We also reported a positive correlation between viral load in rearing water, the onset of clinical signs and mortality peaks. All these results highlight the risk posed by high stocking densities for the occurrence of VER episodes in fish farming facilities. ACKNOWLEDGEMENTS L. Vázquez- Salgado would like to thank Interreg VA Spain– Portugal cooperation Program (POCTEP) 2014- 2020, 0474_BLUEBIOLAB project, cofunded by FEDER for her research contract. The authors would like to acknowledge Stolt Sea Farm S.A. for providing the fish, A. A. Silva and J. Franqueira for their contribution to sample processing, and Native English School of Languages for the language revision. CONFLICT OF INTEREST The authors declare that there is no conflict of interest. DATA AVAILABILITY STATEMENT No data availability statement. ORCID Lucía Vázquez- Salgado https://orcid.org/0000-0002-0059-7279 Isabel Bandín https://orcid.org/0000-0002-1947-5933 REFERENCES Bandín, I., & Souto, S. (2020). Betanodavirus and VER disease: A 30- year research review. Pathogens, 9(2), 106. https://doi.org/10.3390/patho gens9 020106 Castro, A., Montes, M., Orihuela, L., Linares, J., Cota, N., Carrera, L., Toledo, P., & Lazo, J. P. (2019). Effect of stocking density on growth and survival of fine flounder Paralichthys adspersus (Steindachner, 1867) larvae. Latin American Journal of Aquatic Research, 47(1), 1– 8. https://doi.org/10.3856/vol47 - issue 1- fullt ext- 1 Ciji, A., & Akhtar, M. S. (2020). Nitrite implications and its management strategies in aquaculture: A review. Reviews in Aquaculture, 12(2), 878– 908. https://doi.org/10.1111/raq.12354 Cordeiro, O. D., Silva, T. S., Alves, R. N., Costas, B., Wulff, T., Richard, N., de Vareilles, M., Conceição, L. E. C., & Rodrigues, P. M. (2012). Changes in liver proteome expression of Senegalese sole (Solea senegalensis) in response to repeated handling stress. Marine Biotechnology, 14(6), 714– 729. https://doi.org/10.1007/s1012 6- 012- 9437- 4 Costas, B., Aragão, C., Mancera, J. M., Dinis, M. T., Conceição, L. E. C., & Refojos, B. C. (2008). High stocking density induces crowding stress and affects amino acid metabolism in Senegalese sole Solea senegalensis (Kaup 1858) juveniles. Aquaculture Research, 39(1), 1– 9. https:// doi.org/10.1111/j.1365- 2109.2007.01845.x FAO (2018). El estado mundial de la pesca y la acuicultura 2018. Cumplir los objetivos de desarrollo sostenible. CC BYNC- SA 3.0 IGO. González- Mira, A., Varó, I., Solé, M., & Torreblanca, A. (2016). Drugs of environmental concern modify Solea senegalensis physiology and biochemistry in a temperaturedependent manner. Environmental Science and Pollution Research, 23(20), 20937– 20951. https://doi. org/10.1007/s1135 6- 016- 7293- x Iguchi, K., Ogawa, K., Nagae, M., & Ito, F. (2003). The influence of rearing density on stress response and disease susceptibility of ayu (Plecoglossus altivelis). Aquaculture, 220(1– 4), 515– 523. https://doi. org/10.1016/S0044 - 8486(02)00626 - 9 Inendino, K. R., Grant, E. C., Philipp, D. P., & Goldberg, T. L. (2005). Effects of factors related to water quality and population density on the sensitivity of juvenile largemouth bass to mortality induced by viral infection. Journal of Aquatic Animal Health, 17(4), 304– 314. https://doi.org/10.1577/H04- 028.1 Jaramillo, D., Hick, P., & Whittington, R. J. (2017). Age dependency of nervous necrosis virus infection in barramundi Lates calcarifer (Bloch). Journal of Fish Diseases, 40(8), 1089– 1101. https://doi.org/10.1111/ jfd.12584 Jia, R., Han, C., Lei, J. L., Liu, B. L., Huang, B., Huo, H. H., & Yin, S. T. (2015). Effects of nitrite exposure on haematological parameters, oxidative stress and apoptosis in juvenile turbot (Scophthalmus maximus). Aquatic Toxicology, 169, 1– 9. https://doi.org/10.1016/j.aquat ox.2015.09.016 Jia, R., Liu, B. L., Han, C., Huang, B., & Lei, J. L. (2016). The physiological performance and immune response of juvenile turbot (Scophthalmus maximus) to nitrite exposure. Comparative Biochemistry and Physiology Part - C, 181– 182, 40– 46. https://doi.org/10.1016/j. cbpc.2016.01.002 Kim, J. H., Dahms, H. U., & Han, K. N. (2013). Biomonitoring of the river pufferfish, Takifugu obscurus in aquaculture at different rearing densities using stressrelated genes. Aquaculture Research, 44(12), 1835– 1846. https://doi.org/10.1111/j.1365- 2109.2012.03189.x Lama, R., Pereiro, P., Valenzuela- Muñoz, V., Gallardo- Escárate, C., Tort, L., Figueras, A., & Novoa, B. (2020). RNA- Seq analysis of European sea bass (Dicentrarchus labrax L.) infected with nodavirus reveals powerful modulation of the stress response. Veterinary Research, 51(1), 1– 22. https://doi.org/10.1186/s1356 7- 020- 00784 - y Liu, P., Wang, L., Kwang, J., Yue, G. H., & Wong, S. M. (2016). Transcriptome analysis of genes responding to NNV infection in Asian seabass epithelial cells. Fish and Shellfish Immunology, 54, 342– 352. https://doi. org/10.1016/j.fsi.2016.04.029 Livak, K. J., & Schmittgen, T. D. (2001). Analysis of relative gene expression data using realtime quantitative PCR and the 2- ΔΔCT method. Methods, 25(4), 402– 408. https://doi.org/10.1006/meth.2001.1262 Mori, K. I., Nakai, T., Muroga, K., Arimoto, M., Mushiake, K., & Furusawa, I. (1992). Properties of a new virus belonging to nodaviridae found in larval striped jack (Pseudocaranx dentex) with nervous necrosis. Virology, 187(1), 368– 371. https://doi.org/10.1016/0042- 6822(92)90329 - N Munday, B. L., Kwang, J., & Moody, N. (2002). Review article Betanodavirus infections of teleost fish: A review. Journal of Fish Diseases, 127– 142, https://doi.org/10.1046/j.1365- 2761.2002.00350.x Nishi, S., Yamashita, H., Kawato, Y., & Nakai, T. (2016). Cell culture isolation of Piscine nodavirus (Betanodavirus) in fishrearing seawater. Applied and Environmental Microbiology, 82(8), 2537– 2544. https:// doi.org/10.1128/AEM.03834 - 15.Editor Nishizawa, T., Furuhashi, M., Nagai, T., Nakai, T., & Muroga, K. (1997). Genomic classification of fish nodaviruses by molecular phylogenetic analysis of the coat protein gene. Applied and Environmental Microbiology, 63(4), 1633– 1636. https://doi.org/10.1128/ aem.63.4.1633- 1636.1997 Ogut, H., & Reno, P. W. (2004). Effects of fish density on spread of infectious hematopoietic necrosis virus (INHV) in rainbow trout, Oncorhynchus mykiss. The Israeli Journal of Aquaculture, 563(January), 218– 225. Olveira, J. G., Souto, S., Dopazo, C. P., & Bandín, I. (2013). Isolation of betanodavirus from farmed turbot Psetta maxima showing no signs of viral encephalopathy and retinopathy. Aquaculture, 406– 407, 125– 130. https://doi.org/10.1016/j.aquac ulture.2013.05.007 Olveira, J. G., Souto, S., Dopazo, C. P., Thiéry, R., Barja, J. L., & Bandín, I. (2009). Comparative analysis of both genomic segments of