Intestinal Explant Cultures from Gilthead Seabream (Sparus Aurata, L.) Allowed Determining the Mucosal Sensitivity to Bacterial Pathogens and the Impact of a Plant Protein Diet

Although high levels of sh meal replacement by alternative protein sources have been achieved without relevant alterations in terms of growth performance, negative effects on immune status were detected. The diet and sh immunity interactions at gut level have been widely discussed, although in vivo approaches have reported several limitations. In this sense, intestine explant culture system can be a valuable complementary tool to study the interactions between pathogenic bacteria and sh gut response, and the possible inuence of environmental, breeding, rearing conditions or dietary components on the responsiveness of the innate immune system in sh. The object of this study was to test the impact of total substitution of sh meal by plant protein on the intestinal health of seabream (12g) in two growth stages: phase I (90 days), up to 68 g, and phase II (305 days), up to 250g. In phase II, the effects of the long term and short exposure (15 days) to plant protein diet were determined. In order to determine the effect of plant protein feeding on the innate immune response to bacterial pathogens, and an ex vivo procedure of intestine explants culture was implemented. All qPCR assays and expression analyses were performed using the Applied Biosystems 7500 Real-Time PCR with SYBR® Green PCR Master Mix (ThermoFisher Scientic, Waltham, Massachusetts, USA). After an initial Taq activation of polymerase at 95 ºC for 10 min, 42 cycles of PCR were performed with the following cycling conditions: 95 ºC for 10 s and 60 ºC for 30 s in all genes. In order to evaluate assay specicity, a melting curve analysis was directly performed after PCR cycles by slowly increasing the temperature (1º C/min) from 60 to 95 ºC, with continuous registration of changes in uorescent emission intensity. The total volume for every PCR reaction was 20 μl, performed from diluted (1:20) cDNA template (5 μl), forward and reverse primers (10 μM, 1 μL), SYBR® Green PCR Master Mix (10 μl), ROX (2 μL, 10 nM) and nuclease-free water up to 20 μl. The analysis of the results was carried out using the 2 - ∆∆ Ct method. The target gene expression quantication was expressed relative to the expression of the selected reference gene. A cDNA pool from all the samples was included in each run and acted as a calibrator, and a non-template control for each primer pair, in which cDNA was replaced by water, was run on all plates. Reference and target genes in all samples were run in duplicate PCR reactions.


Abstract Background
Although high levels of sh meal replacement by alternative protein sources have been achieved without relevant alterations in terms of growth performance, negative effects on immune status were detected. The diet and sh immunity interactions at gut level have been widely discussed, although in vivo approaches have reported several limitations. In this sense, intestine explant culture system can be a valuable complementary tool to study the interactions between pathogenic bacteria and sh gut response, and the possible in uence of environmental, breeding, rearing conditions or dietary components on the responsiveness of the innate immune system in sh.
The object of this study was to test the impact of total substitution of sh meal by plant protein on the intestinal health of seabream (12g) in two growth stages: phase I (90 days), up to 68 g, and phase II (305 days), up to 250g. In phase II, the effects of the long term and short exposure (15 days) to plant protein diet were determined. In order to determine the effect of plant protein feeding on the innate immune response to bacterial pathogens, and an ex vivo procedure of intestine explants culture was implemented.

Results
Fish showed less tolerance to dietary plant protein in phase I than in phase II, while the ex vivo assays indicated that the intestine from sh fed at short-term plant diets showed a higher immune response than at long term feeding.
In relation to the immune response to bacterial challenge, a signi cant expression in pro-in ammatory cytokines IL-1β and IL-6 after 6 hours of exposure to V. algynoliticus, while COX-2 expression was signi cantly induced by P. damselae subsp. pisicida, showing positive high correlation between them.

Conclusions
A differential health status was observed depending of growth stage, being stricter to the plant protein inclusion the younger sh. The new experimental system based on sh intestinal explants culture has been successfully implemented, becoming an effective methodology for ex vivo studies. Under ex vivo conditions, the bacterial challenge induced in ammatory and immune intestinal response, responding stronger those intestine of sh fed during a short-term with a total substitution of sh meal.

Background
In addition to the digestion and absorption of nutrients, the sh intestine is a complex biological system that represents a major defence barrier against pathogens and plays a crucial role in osmoregulation, immune and in ammatory response [1]. Furthermore, the intestine participates in the modulation of gastrointestinal microbiota inducing in ammatory responses against pathogenic bacteria or developing immunotolerance to luminal bacteria.
The interactions between sh intestinal immunity, pathogen and commensal microbiota in the gut have been widely reviewed [2]. Bacterial challenges in vivo require specialized settings, expensive operating costs, and a high number of sh, and is di cult to perform and achieve the desired experimental working conditions [3]. In this regard, systems have been developed, based on the ex vivo maintenance of intestine fragments, to evaluate successfully the effect of different bacterial strains on the intestinal health and providing very reliable information on the interactions between the bacteria and the host. The ex vivo intestinal sack method [4] has been used to assess the histological and microbial changes in sh in response to bacteria exposure [3,[5][6][7][8][9]; however, this method is highly restricted by the tissue viability under experimental conditions [3,4]. The development of new experimental models based on tissue explants culture has allowed maintaining the tissue lifespan, as well as immune and histological features [10][11][12]. These systems have been used to register responses to exposure to speci c bacteria in human tissue explant cultures, also at gene expression level [12].
On the other hand, the necessity to replace sh meal by alternative protein sources in aquafeeds has led researchers to focus on the impact of the inclusion of alternative ingredients, such as plant protein sources. Although its use at high levels without impairing growth performance is feasible [13,14], negative effects on immune capacity have been observed [15]. Previous ndings revealed that during early growth stages the inclusion of plant protein has a great impact [16,17]; at intestinal level, the inclusion of plant sources in diets has been related to morphological alterations, changes in the intestinal bacterial community, in ammatory events and lack of capacity to regulate the intestinal epithelial integrity [18]. Similar results have been reported in vivo at molecular level, with an altered gene expression pattern [19][20][21].
Research related to the impact of sh meal replacement becomes even more relevant for carnivorous species, such as gilthead seabream. In this species, previous studies with different levels of substitution and alternative ingredients, have been assayed to evaluate zootechnical parameters and survival [16,22,23], and the immune in vivo response to sh meal replacement [24][25][26] or bacterial infection [27][28][29][30].
The ex vivo response of intestinal tissue from sh feed with different protein sources to bacterial challenge has been previously addressed in other species [3]. However, to the best of our knowledge, this is the rst study involving gene expression determination in sh intestinal explants after ex vivo bacterial exposure.
Therefore, in the current work, it was developed an intestine explant culture system to evaluate a possible differential in ammatory and immune response to ex vivo bacterial challenge in sh after a long-term of total sh meal replacement at different stages of growing. Additionally, the effect of a short-term of total sh meal replacement by a plant mixture in on growing sh was also assayed.

Ethics statement
The experiment was reviewed and approved by the Committee of Ethics and Animal Welfare of the Universitat Politècnica de València (UPV), following the Spanish Royal Decree 53/2013 on the protection of animals used for scienti c purposes [31]. Fish, rearing system conditions, diets and feeding conditions A total number of 240 juveniles of gilthead seabream (averagare weigth 7,5 g and 60 days) were obtained from the sh farm BERSOLAZ (Bersolaz Spain, S.L.U, Culmarex Group) located in Port de Sagunt (Valencia, Spain) and transported to the facilities at the Universitat Politècnica de València, where the growth trial was conducted after 15 days of adaptation to experimental conditions. Features of the system and water parameters set were described in previous growth trials carried out in these facilities [32,33]. Lighting conditions were determined by the natural photoperiod. Temperature, pH, oxygen, ammonia, nitrite and nitrate concentrations were monitored along the experiment. The sh were daily fed by hand to apparent satiation two times per day (9:00h and 17:00h). The pellets were slowly distributed, allowing sh to eat, in a weekly regime of six day of feeding and one day of fasting.
Diets were prepared by cooking extrusion process using a semi-industrial twin-screw extruder (CLEXTRAL BC-45, St. Etienne, France). A sh meal based control diet (FM), in which most of the protein was provided by sh meal (59%), and a plant protein based diet (PP), in which all the sh meal was replaced by plant sources and synthetic amino acid were added to meet the minimum amino acid requirement for gilthead seabream juveniles [34]. Ingredients and proximate composition are shown in Table 1. Prior to diet formulation, dry matter, crude protein, crude lipid, ashes and crude bre (CF) of different sources and ingredients used were analysed according to AOAC procedures [35]. All analyses were performed in triplicate. Amino acids of raw diets were also analysed by reverse phase -high performance liquid chromatography [36]. Macronutrients and essential amino acid content were determined in the experimental diets, and they are shown in Table 1. . Bacteria were grown under agitation at 26º C for 2 days (P. anguilliseptica and P. damselae subsp. piscicida) and at 30º C for 1 day (V. algynoliticus). Then, 1.5 g/L of bacteriological agar were added to these media to prepare solid medium in petri dishes. For the bacterial challenge, optical density (600 nm) of the bacterial cultures was determined and bacterial cell number was estimated using the standard curves established for each strain.
Then, bacterial cultures were centrifuged at 4.000 g for 20 min, washed once with PBS, and re-suspended in CO 2 -independent cell culture medium (Gibco, ThermoFisher) to a nal concentration of 3·10 7 ufc/mL in the case of V. algynoliticus, and 1·10 7 ufc/mL of P. damselae subsp. piscicida and P. anguilliseptica.

Experimental design
The aim of this work was to evaluate the impact of dietary sh protein substitution by plant protein on the intestinal health status and its immune response capacity. For this purpose, sh were fed with sh meal (FM) or plant protein (PP) based diets, and animals were sacri ced and processed at two critical growth phases of sea bream [16,17]: Phase I (90 days; from 12g to~ 68g) and Phase II (305 days; up to 250g). Each diet was assayed in tanks per triplicate. Phase I: Up to 68 g Fish were fed with FM or PP diets, in tanks per triplicated (40 sh per tank), up to 90 days, being scattering 2 sh per group to submit bacterial challenge (68±37.8g). For basal gene expression two fragments from foregut (FG) and hindgut (HG) from each sh were placed in an eppendorf tube containing 500 µl of RNA Later® (Qiagen, Valencia, CA) for subsequent total RNA (tRNA) extraction. Additionally, four fragments from of FG and HG fragments were used for the ex vivo assay and exposed to pathogens challenge (see below). Gene expression was determined in all samples to evaluate the in ammatory and immune status of the intestinal mucosa due to changes in the diet and the bacterial challenge.
Phase II: Up to 250 g Fish, ~30 sh per tank, were fed with the same diets, FM and PP, up to 305 days when the mean weight was 252±70.1g (Fig 1). In this second phase, in addition to the impact of long-term feeding with 100% of PP diets, also a short-term exposure (15 days) of total sh meal substitution was evaluated. For this purpose, sh bred with FM diet (n=~15 sh per tank) were changed to a PP diet two weeks before the termination of the experiment (from day 290 to day 305) (PP* group). Three sh from the FM group and two from the PP and PP* were sacri ced to obtain FG and HG explants for ex vivo assays. As in the previous assay, for basal gene expression two fragments from each sh were placed in RNA Later® for subsequent total RNA (tRNA) extraction, and four pieces for ex vivo study.

Ex vivo assays and bacterial challenge
Before tissue preparation, sh were sacri ced by immersion in benzocaine (60 ppm) during 15 min. Then, they were dissected and the intestine was obtained and separated in two sections (foregut and hindgut). Each section was cut with a scalpel in small pieces (4 mm x mm), which were immediately placed in culture lter plates (15 mm diameter wells with 500 µm bottom-mesh, Netwell culture systems, Costar, Cambridge, MA) with the epithelial surface facing up. Filters were placed into wells containing 1 mL of the different bacterial solutions (one of them was preserved without bacteria as control; Ex vivo Unchallenged group) in CO 2 -independent cell culture medium (Gibco, ThermoFisher). 100 µL of the corresponding bacterial solutions were nally added to epithelial surface to ensure that samples were completely submerged. At the end of the incubation time, samples were carefully collected from the culture lter plates and stored in 100 mM Tris-HCl at 4ºC or RNA later at -80º C for lactate dehydrogenase (LDH) activity evaluation or RNA isolation, respectively. Changes in pH of the explant culture medium due to different bacterial treatments were monitored. Explants of foregut (FG) and hindgut (HG) from two sh per group were incubated during 4 and 6 h at 22º C in independent CO 2 atmosphere, depending on the experiment. The bacterial species used in the pathogen challenge were: Photobacterium damselae subsp, Pseudomonas anguilliseptica and Vibrio algynoliticus. Pseudomonas anguilliseptica was discarded in phase II, because bacterial concentration could not be determined due to aggregate formation. After explant assay, the samples were placed into RNA Later (Qiagen, Valencia, CA) for subsequent tRNA extraction. All conditions ( sh/section/stimuli) were assayed in duplicate and gene expression was determined in all samples to evaluate the intestinal in ammatory and immune status based on experimental diet and bacterial challenge.
In order to determine the tissue integrity, the LDH activity was determined [37] in the tissue (U/mg protein) and explant culture medium (U/L) at different times of the incubation (0, 4, 6 and 24 h). LDH activity was analysed measuring the nicotinamide adenine dinucleotide (NADH) absorbance at 340 nm using the commercial kit (BioSystems S. A., Barcelona, Spain). Tissue was weighed, homogenised in Tris-Hcl 100 mM while maintaining the tubes on ice, centrifuged at 12.000 rpm and 4º C for 15 min and supernatant was collected for LDH assessment. Total protein in tissue extracts was determined using Bradford [38].

Gene expression assay of intestinal in ammatory and immune markers
Based on the gene expression analysis used in previous studies in this species (Sparus aurata) to evaluate the intestinal in ammatory and immune status [32] tRNA was extracted from intestinal tissue samples using the phenol/chloroform method with Trizol Reagent (Invitrogen, Spain) and treated with DNAse I (Roche) to remove DNases. Total RNA concentration, quality and integrity were assessed using a NanoDrop 2000C Spectrophotometer (Fisher Scienti c SL, Spain). The integrity of 28S/18S was also determined by gel electrophoresis. 1 µg of total RNA was used for cDNA synthesis reaction using the qScript cDNA synthesis kit (Quanta BioScience), according to the manufacturer's instructions. An Applied Biosystems 2720 Thermal Cycler was used with the following cycling conditions: 22 ºC for 5 min, 42 ºC for 30 min, and 85 ºC for 5 min. cDNA samples were stored at -20º C until gene expression was analysed.
Four housekeeping candidate genes ( Table 2) were tested to be used as reference genes, and for assessing RNA integrity along the assay. The Cq of the four genes was determined in six pooled samples from Experiment 1 (two dietary groups: FM and PP; three times: 0, 4 and 6 h). Relative gene expression of six genes was determined in the foregut and hindgut samples. The genetic markers monitored in this assay were three pro-in ammatory markers, IL1-β, IL-6 and COX-2, the main immunoglobulin, IgM, the main intestinal mucin, I-Muc, and the occludin gene, Ocl, with primers listed in Table 2.  LDH enzymatic activity in tissues and the supernatant was statistically analysed by one-way analysis of variance (ANOVA) using Newman-Keuls test to determine possible differences across the assay (0, 4, 6 and 24 hours) in FG and HG.
The expression stability of reference genes was assessed using the BestKeeper program, basing on the arithmetic means of the Cq values [39]. Lower deviation in the expression is related to better stability.
The evaluation of intestinal in ammatory and immune status was performed through the gene expression of the target genes both in vivo and ex vivo conditions. The relative gene expression was statistically analysed by ANOVA. Gene expression of cultured pieces was normalised with the expression of ex vivo unchallenged samples at 4 and 6 hours. Multifactorial analysis was used to determine the signi cance (p<0.05) of different factors considered (dietary treatment: FM/PP, intestinal section: FG/HG and bacterial stimuli: P. damselae subsp. piscicida/P. anguilliseptica/V. algynoliticus) at different times and to determine differences in normalised gene expression between dietary groups, sections and bacterial stimuli, using Newman-Keuls test. Data was expressed with the mean and the standard error of the normalised expression values, and differences were considered statistically signi cant when p<0.05.
Additionally, with the aim to evaluate if the bacterial challenge is individually inducing the target genes or a combination of a set of genes, a correlation analysis was carried out and the Pearson product-moment coe cient was obtained for each pair of genes.
Finally, in order to con rm the assay reproducibility, the gene expression of the biological replicate samples was randomly assigned to different variables (x and y). Data consistency was evaluated for each gene by simple regression analysis using the model y = ax. 95% con dence intervals for a (a±1.96σ) were obtained for each gene to validate the hypothesis a=1 (y=x).

Results
With the aim to evaluate the effect of PP diet on in ammatory and immune gene expression in the gut, 240 sh were fed on FM and PP for 305 days. The effect of diet was investigated in a rst stage at day 90 (phase I) and the experiment continued until day 305 (phase II). Two weeks before the end of the experiment, a group of animals from the FM group was introduced to the PP diet, in order to assess a possible short term effect of the PP diet in adult sh.
Intestine fragments of sh harvested in phase I and phase II were used for the ex vivo pathogen challenge assay, but also basal gene expression was determined on intestinal samples.
Optimal conditions for intestinal explant culture were set up through different approaches. Tissue and cellular integrity were monitored by the release of LDH activity at 0, 4, 6 and 24 h of incubation. No relevant differences were observed at tissue level, but a signi cant increase was registered at 24 h of incubation in the explant culture medium (Additional le Fig. 1). Candidate housekeeping genes for real time qPCR were tested at different times, indicating also that 6 h was an appropriate time of incubation (Additional le Fig. 2). On the other hand, if the effect of ex vivo procedure is evaluated, signi cant differences were observed in most of the genetic markers analysed (Additional le Fig. 3). Therefore, in the following assays, gene expression was normalised based on the ex vivo unchallenged samples.
Phase I: Up to 68 g With the aim of assessing the effect of a total sh meal substitution on intestinal in ammatory and immune status at early growth stages (68 g sh weight), basal gene expression was determined and ex vivo trials were carried out with 6 h of challenge of pathogen bacterial cultures.

Basal gene expression
After 90 days fed with PP and FM diets, sh intestines showed a signi cantly different gene expression pro le, as the PP group had a higher expression level for IL-1β and lower for COX-2 (Fig. 2).

Ex vivo assay
The explant culture system was used to determine the immune response of intestinal fragments from on-growing seabream specimens fed with different diets (Fig. 3A). Additionally, the intestine was divided into two segments: FG and HG to evaluate a possible differential immune response (Fig. 3B). At 6 h, gene expression was statistically different between the PP and FM diets in most of the markers, and the induction of IL-1β and Ocl in sh under PP diet was particularly remarkable (Fig. 3A). Gene expression responses of the different intestinal sections were quite similar, with higher expression of IL-6 in FG at 4 h (Additional le Fig. 4) and at 6 h HG presented higher expression of Ocl (Fig. 3B).
After the bacterial challenge, sh belonging to PP group showed, in general, a greater response to the bacterial stimuli (Fig. 4), with signi cant differences for IL-1β and IL-6 genes. This tendency to a higher response in PP group was not observed at 4 h of incubation, not registering signi cant differences between diets or bacterial stimuli (Additional le Fig. 5).
If only the bacteria variable is considered, signi cant differences were only observed with V. alginolyticus after 6 h of incubation, speci cally with a higher expression of COX-2 and notably of IL-1β (Additional le Fig. 6B), indicating that this bacterium elicited a great response in all conditions. Finally, if a multifactorial analysis of variance is performed taking into account all the factors, IL-6 and Ocl were signi cantly altered by the section and the diet at 4 hours of incubation, respectively. Nevertheless, it was necessary 6 h of exposition to register differences caused by bacterial stimulus (Additional le Table 1).

Page 8/17
Phase II: Up to 250 g This second assay tried to evaluate the impact of a total substitution of sh meal at longer term feeding, up to 252 g (305 day; PP) on intestinal gene expression. Additionally, the differential response of total sh meal substitution with plant protein (PP) during a short-term period (15 days; PP*) was evaluated.

Basal gene expression
No gene expression differences were observed between sh fed FM diet and plant based diets at long-term (PP) as well as at short term (PP*), except to IgM in PP* group (Fig. 5).

Ex vivo assay
The expression of pro-in ammatory genes (IL-6 and COX-2 genes) increased after 6 h of incubation in the ex vivo unchallenged group respect to the basal values, con rming the results of previous assay (Additional le Figure S7). Hence, expression results in samples incubated with the different bacteria were normalised with the expression of the control samples, for each experimental factor.
The consistency of ex vivo assay for the biological replicates was checked by a correlation analysis for each biological replicate pair (Additional le Figure  S8). The adjustment to the lineal model is particularly good for pro-in ammatory genes (IL1-β, IL-6, COX-2). I-Muc expression reported a high variability between duplicates and no signi cant relationship can be established between x and y data, consequently, it was not considered in further analyses.
A multifactorial ANOVA of gene expression taking into consideration the diet, section and bacterial challenge (stimuli), underlined a signi cant linkage of PP diet with COX-2 and Ocl. As expected, the expression of pro-in ammatory genes (IL-1β, IL-6 and COX-2) was bound to the pathogen stimuli (Table 3), while again no differences were reported between sections. Therefore, due to the lack of statistical differences (p < 0.05) between intestinal sections, the following analyses were performed joining both sections. The exposition to V. algynoliticus and P. damselae subsp. piscicida induced a remarkable increase of IL-1β and IL-6 expression respect to unchallenged samples (Fig. 6). Regarding the sensitivity to the pathogen as function of diet, the PP* group showed a remarkable tendency to have higher expression values for all tested genes in response to bacteria, but only COX-2 and Ocl showed signi cant differences with respect to groups FM and PP ( Fig. 6D and Fig. 6F). Of note, V. algynoliticus induced signi cant IL-1β response in all diet groups (Fig. 6B). Finally as expected, there was a high correlation (Pearson's coe cient) between the expression of IL-1β, a known master regulator of innate immune response and in ammation, and IL-6 and COX-2 (IL-1β / IL-6 = 0.74; IL-1β / COX-2 = 0.72) (Fig. 7).

Discussion
It has been extensively proven that the inclusion of alternative plant proteins can lead to nutritional imbalances [23,40] and immune dysfunctions [24,41], particularly in carnivorous sh, increasing their susceptibility to pathogenic invasion, disease and nally death. Findings of present study suggest that the sh gut response to the total dietary substitution of sh meal by plant protein meals can differ at short term and long-term feeding and the sh size.
Previous studies in seabream demonstrated that the use of plant proteins induced signi cant alterations of the gut microbiota, gut gene expression and gut proteomic pro le [32,42,43]. Nonetheless, these alterations can be affected to different extent by the feeding period or sh stage. In this study, the effect of a total sh meal (FM) substitution by plant protein (PP) was evaluated in two stages of sh growth: from 12 to 68 g (phase I) and up to 252 g (phase II). In phase II, the effect on intestinal health status of FM substitution by PP throughout the whole period (305 days) was compared with that of a shorter term feeding (15 days).
An explant culture assay has been implemented in this work to evaluate the intestinal health status of sh exposed to PP diet. Ex vivo approaches based on explants culture proved to be useful to analyse pro-in ammatory responses [12]. In sh, several works have been attempted to evaluate the pathogen-host [44], especially by the intestinal sack method [3,[7][8][9]. In the present work, signi cant differences were clearly observed with incubations of the intestine explants of 6 h, and these experimental conditions showed to preserve tissue integrity. Furthermore, explants obtained from the same section of each sh demonstrated a sound consistency in the response to pathogens. Although a different immunological performance has been attributed to the foregut and hindgut [45], using this experimental set up no signi cant differences could be found between gene expression in both segments. This allowed increasing the number of explant fragments from each single intestine, then improving the experimental e ciency and reducing the number of sh and, hence, reducing individual variability.
In the present experiment, three pro-in ammatory markers have been monitored to assess the in ammatory status (IL-1β, IL-6, COX-2) [46,47]. IgM is considered as the most abundant immunoglobulin in plasma, high levels in sh fed with plant sources based diets have been reported [48] and its expression was induced in mucosal tissues as response to pathogen infection [49]. I-Muc is the main intestinal mucin, and it is involved in epithelial protection, bacteria adhesion and growth, and has been suggested as a resilience biomarker to in ammation in sh [50]. Finally, Ocl is a key protein in the regulation of tight junctions between enterocytes, and therefore in the permeability of the epithelial barrier [51].
The stage of sh growth and the feeding period have a clear in uence on the response of sh performance to dietary plant protein and previous studies reported that juvenile sh tolerates less dietary plant protein than commercial size sh [16,52]. In phase I, intestine explants of seabream fed with plant protein for 90 days showed a high basal in ammatory response, particularly of pro-in ammatory genes IL-1β, IL-6 and also COX-2, as shown before. IL-1β is a known canonical master regulator of pro-in ammatory processes, it is secreted in response to gram negative bacteria and its release is followed by the production of IL-6 and other cytokines in the pro-in ammatory cascade [46,53]. Modi cation in diet composition changed the expression of IL-1β, in agreement with previous studies, thus indicating that after plant protein diet rendered an already sensitized intestine to in ammatory/stress stimuli [32,[53][54][55].
During infection, IL-1β is induced, and increased expression of IL-1β has been reported in the intestine of different species, including gilthead seabream [29], after intraperitoneal challenge with gram negative bacteria. Increased expression of IL-1β and COX-2 has been reported before after in vitro challenge of gilthead seabream immune cells with bacteria or commercial pathogen associated molecular pattern (PAMP) solutions [56][57][58][59]. Although COX-2 is a typical oxidative stress marker, its expression is also induced by in ammatory mediators [58,60], including IL-1β [61], as it takes part in the production of reactive oxygen species (ROS) and NO that have antibacterial activity and are also part of the innate immune system in higher vertebrates and carp macrophages [62,63]. Also genes related to the maintenance of the epithelial tissue integrity, such as Ocl (expressing occludin), are in uenced by in ammatory processes [64,65], and its regulation depends on several cytoskeletal, scaffolding, signalling and polarity proteins [51] and it is de nitely related to epithelial barrier functions in vivo and in vitro [65]. In mice, all these genes are possibly connected for the maintenance if intestinal barrier [66].
In phase II (up to 250 g), there were no signi cant differences between experimental groups (FM and PP), thus supporting that older sh are more tolerant to diet plant proteins than sh in earlier growth stages [16]. It is noteworthy the lack of in ammatory response in PP* group (short PP exposure), supporting that PP long term feeding has deeper alterations in sh gut and contributes to changes in the microbiota and an increase of sh mortality [32], in agreement with previous studies [33,67,68], In so far as de cient diets could be considered a stress factor, long term feeding could determine suppressive or depressive effects on the immune mechanisms [24,28,32,33]. The down-regulation of mRNA expression of some immune related genes with the increased plant proteins in the diet has also been reported in other species [69]. On the other hand, lower gene expression values reported in FM group might be related to a higher protection in host from bacterial adhesion and growth. Nevertheless, results should be analysed with caution since a wide individual variation of in ammatory and immune genes expression has been reported in other species [70] and level of expression before the ex vivo trial conditions the in ammatory and immune capacity registered after bacterial exposure [41].
Nevertheless, a longer stimulation with PP may lead to a degree of tolerance. This would explain why in the ex vivo challenges with bacterial pathogens, stimulation in seabream samples from PP* group was higher than in PP group.
Finally, bacterial pathogens have been selected for the challenge in the ex vivo assay due to their proven pathogenic activity in farmed gilthead seabream. Photobacterium damselae subsp. piscicida is the causal agent of pasteurellosis (Romalde, 2002), Pseudomonas anguilliseptica is related to 'winter disease' [71] and Vibrio algynoliticus has been described as the causal agent of vibriosis [72], also associated to other Vibrio species in high mortality outbreaks [73]. This work showed that Vibrio algynoliticus CECT 521 (ATCC 17749) had a very powerful in ammatory effect displaying a very high induction of IL-1β. This agrees with the fact that the published genome of this strain displayed a great toxigenic potential [74], with at least a gene encoding a pore forming RTX family toxin, among others (KEGG Genomes, toxin search on assembly GCA_000354175.2).

Conclusion
Plant protein diets showed to alter the mucosal immune homeostasis. In early stages of development, sh are very sensitive to plant diets, while adult sh seem to become tolerant to a constant PP diet, although maintaining a high threshold of in ammatory signals. The exposure to PP fed for short-term in adults led to a greater response to bacterial challenge. V. algynoliticus triggered the highest immune and in ammatory response. The successful evaluation of in ammatory and immune responses and pathogen challenge has been achieved by means of a new experimental system that implemented sh intestinal explants culture in gilthead seabream, a system that might be easily adapted to other teleost species. The use of such ex vivo methods constitutes an amenable technique that renders reliable results and it helps in reducing the number of animals per assay. Availability of data and materials The datasets during the current study are available from the corresponding authors on reasonable request

Competing interests
The authors declare that they have no competing interests that could in uence the content of the paper.

Funding
The research was supported by a grant nanced by the Spanish Ministerio de Economía y Competitividad AGL2015-70487-P. The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.
Author's contributions CB, MJC, GPM, DSP and SML designed the assay. GE, CB and ATV carried out the experiments. GE analysed the data. GE, DSP, GPM and SML prepared the manuscript and discussed the results. All authors read and approved the nal manuscript.
Summary of the experimental design. The impact of dietary sh protein substitution by plant protein in sh immune response to ex vivo bacterial challenge was evaluated at two on-growing phases: 12-68 g (90 days) and up to 250g (305 days). Additionally, a group was included at 305 days to estimate the effect of short-term sh meal substitution (15 days; PP*).

Figure 2
Intestinal basal expression of sh exposed to PP diet in Phase I (68 g). Relative gene expression (A. U.) of the different genes is expressed by the mean and standard error. Different superscripts on the bars indicate signi cant differences (p<0.05).  Intestinal basal expression of sh exposed to long (PP) or short term (PP*) diet in Phase II (250 g). Relative gene expression (A. U.) of the different genes is expressed by the mean and standard error. Asterisk on the bars indicates signi cant differences (p<0.05).