Comparative transcriptomics of immune response to viral and bacterial stimuli in three acanthopterygian bony fish

13 Diseases triggered by bacterial and viral infections have caused huge economic losses for three of 14 the most important European aquaculture species: turbot (Scophthalmus maximus), gilthead seabream 15 (Sparus aurata) and European seabass ( Dicentrarchus labrax). Understanding how the y respond to 16 pathogens is relevant for advancing aquaculture disease management and comprehending evolution of 17 immune response within teleosts. Since mechanisms conserved across species are assumed to perform 18 important roles, comparative analysis provides a powerful approach to pinpoint key elements of the 19 immune defence. Here, we report the first comparative immune-transcriptomic analysis of these three 20 species using bacterial and viral mimics after 20-24 hours post-stimulation with inactivated Vibrio 21 anguillarum and Poly I:C in the head kidney of live fish (in vivo), and in primary leukocyte cultures (in 22 vitro). The transcriptomic response , based on RNA -seq data, revealed a total of 50 3 differentially 23 expressed orthologous genes in response to in vitro-Poly I:C, 1,472 to in vitro-Vibrio, 920 to in vivo-24 Poly I:C, and 832 to in vivo-Vibrio. Interestingly, consistent expression patterns were identified in seven 25 genes across all species in both cell culture and live organisms in response to both pathogen stimuli. 26 Functional enrichment analysis revealed associations with immunity, DNA replication and repair, and 27 cytokine pathways, with the Toll-Like Receptor (TLR) pathway common to both conditions and stimuli. 28 Our study suggests conservation of orthologous gene expression during infection across the three 29 species for genes involved in chemokine pathways, interferon signa lling, antigen processing and 30 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint presentation, cell signalling regulators , and MAPK cascades. This study provides insights into key 31 immune defence mechanisms in acanthopterygian bony fish. 32

Immune response; RNA-seq; Vibrio anguillarum; Poly I:C; turbot; gilthead seabream; 33 European seabass. 34 35 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint 1. INTRODUCTION 36 Aquaculture is of crucial importance in addressing the growing global demand for marine 37 products [1], particularly in Europe, where species such as turbot ( Scophthalmus maximus), gilthead 38 seabream (Sparus aurata) and European seabass (Dicentrarchus labrax) are highly appreciated for their 39 rapid growth and production efficiency [2–5]. However, with the growth of the aquaculture industry, 40 bacterial and viral diseases have become increasingly severe and result in substantial economic losses 41 [6]. 42 During infection, a complex interaction takes place between hosts and pathogens, whereby the 43 host immune system activates genes within immune pathways aimed at neutralizing the infection. Fish 44 depend significantly on their innate immunity, which serves as their primary defence mechanism. 45 Physical barriers such as skin, scales and mucosal surfaces provide physical immunity. Additionally, 46 these mucosal surfaces also confer mucosal immunity through secretory immunoglobulins, which work 47 alongside antimicrobial peptides (AMPs) capable of disrupting microbial membranes [7]. Furthermore, 48 fish possess pattern recognition receptors (PRRs) that recognize pathogen-associated molecular patterns 49 (PAMPs) such as key structural components of bacterial cell walls (e.g., lipopolysaccharides (LPS) and 50 peptidoglycans) and viral RNA. These receptors initiate signalling pathways that activate pro -51 inflammatory responses in fish, which include cytokines that act as messengers to start inflammation, 52 recruit immune cells, and activate antimicrobial defence mechanisms. The head kidney is essential for 53 haematopoiesis and the production of cytokines, while the spleen contributes to the filtration of 54 pathogens and the coordination of inflammatory responses. Concerning adaptive immunity, while it is 55 not as developed as in mammals, fish possess T and B cells that generate specific antibodies (mainly 56 IgM and IgT) to neutralize pathogens [8]. 57 The great diversity of aquaculture fish species and infectious pathogens poses a challenge for 58 unravelling immune response pathways. However, it can be assumed that immune responses include 59 conserved pathways, and that different fish species will present similarities in their response to , for 60 example, bacterial or RNA-viral infections. Therefore, comparative immunology represents an efficient 61 tool to comprehend the main mechanisms underpinning immune responses in a group as diverse as fish 62 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint [9,10]. Comparative research on the immune response in fish to various pathogens has made significant 63 advances in recent years, revealing complex interactions between aquatic species and infectious agents. 64 Notably, comparative analyses are often challenged by the fact that most pathogens are either species-65 specific or exhibit species -specific serotypes, making direct comparisons difficult [11]. To overcome 66 this issue, recent studies have employed both PAMPs, such as Poly I:C (Polyinosinic:polycytidylic acid) 67 or LPS, and generalist pathogens, such as viral haemorrhagic septicaemia, V . anguillarum or 68 Streptococcus parauberis [12–14]. The use of PAMPs provides an effective method for comparing 69 immune responses across species, without the effect of pathogen specificity, thereby enabling more 70 robust interspecies comparisons [15]. In aquaculture species, a considerable number of transcriptomic 71 studies have been conducted with the objective of elucidating the immune response of fish to inactivated 72 or alive V . anguillarum [16–18] or to dsRNA viruses using Poly I:C [19–23], representing common 73 pathogens in farmed settings . In addition to comparisons between species, the difference between in 74 vivo and in vitro conditions is also of interest due to the reduced complexity of cell culture models and 75 widespread use to study infectious diseases. Saravia et al. (2022) [24] studied the transcriptomic 76 response of Harpagifer antarcticus to LPS and Poly I:C both in vitro and in vivo, while Aramburu et al. 77 (2025) [23] assessed not only the turbot transcriptome following exposure to Poly I:C and V . 78 anguillarum both in vivo and in vitro, but also epigenetic regulation using A TAC-seq and ChIP -seq. 79 However, a review of the current scientific literature reveals no studies comparing the immune 80 responses of different fish species, to Poly I:C and bacteria, while also considering both in vitro and in 81 vivo conditions. 82 Unravelling the adjustments of gene expression in the host is fundamental for the 83 comprehension of the infection process at the molecular level to develop strategies for disease control 84 [25]. In recent years, transcriptome analysis using high-throughput RNA sequencing (RNA -seq) has 85 emerged as a powerful tool for elucidating the immune response mechanisms of fish, understanding the 86 molecular basis of pathogen resistance, and examining the comparative immune response across species 87 [16,26]. A comparative functional annotation of immune responses to two categories of disease 88 pathogens (such as bacterial and viral) entails the identification and description of the roles played by 89 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint genes throughout the entire genome. As a part of the European AQUA-FAANG project, we conducted 90 an RNA-seq analysis of head kidney samples from turbot, seabream and seabass after exposure to heat-91 killed V . anguillarum and Poly I:C treatment, using both the full organ (in vivo) and primary leukocyte 92 cultures (in vitro ). The objectives of this study were first to evaluate anti -viral and anti -bacterial 93 responses by direct stimulation of primary leukocyte cultures in comparison with the head kidney of 94 intraperitoneal injected fish , and secondly, to identify conserved immune response pathways by 95 comparing expression in orthologous genes across the three species. To our knowledge, this is the first 96 comparative immune response in these species, and the findings would provide new tools for improving 97 disease resistance in aquaculture. 98 2. MATERIAL AND METHODS 99 2.1 Animals 100 Thirty specimens of S. maximus, twenty-six of S. aurata and twenty-four of D. labrax were 101 used for the study (Table S1). Sampling followed the AQUA-FAANG protocols 102 https://data.faang.org/api/fire_api/experiments/INRA_SOP_invivo.invitro.challenges_20200131.pdf 103 and details are provided in Supplementary Methods, Tables S1 and S2. 104 2.2 In vivo immunostimulation 105 Eighteen fish were intraperitoneally injected per species with Poly I:C (viral mimic , 6 106 replicates), heat-killed V . anguillarum (bacterial stimulus, 6 replicates), or PBS (control, 6 replicates). 107 Before inoculation, Poly I:C (Sigma P1530; 5 mg/ml in PBS) was heated at 55 ºC for 15 min, cooled to 108 room temperature (20 min) and administered at five µg/g of body weight. V . anguillarum (strain P0382; 109 INRA, France) was cultured in tryptic soy broth to an OD600 of 1.5. The pellet from 100 ml culture 110 was washed four times with isotonic NaCl (9 g/L), resuspended in 1 ml, heat-inactivated at 100ºC for 111 30 seconds, cooled, and stored at -80ºC. Each specimen received 138 µl of 1:10 PBS-diluted bacterial 112 extract. Controls were administered 100 µl of PBS. Following a 20-24 h period, fish were anesthetized 113 by bath (MS-222; 100 mg/L) and euthanized with an anaesthetic overdose (MS-222; 150 mg/L), and 114 head kidney samples were extracted, washed with PBS, cut (>20 mg each fragment), and either flash 115 frozen in liquid nitrogen or preserved in RNAlater (Thermofisher Scientific) for downstream RNA 116 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint extraction and sequencing protocols. Thereafter, all samples were stored at -80°C (Fig. 1A, Tables S1 117 and S2). 118 119 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint A) B) PBS 6 fish/species Control Viral mimic stimulation Poly I:C 6 fish/species Bacterial mimic stimulation Vibrio 6 fish/species S. aurata S. maximus D. labrax Resuspended in RNAlater 18 RNA-seq libraries/ Species Sampling after 20-24h Head Kidney extraction & fragmentation S. aurata S. maximus D. labrax Resuspended in RNAlater 18 RNA-seq libraries / Species leukocyte pool/species 6 PBS ≈2x106 cells/species Control Bacterial mimic stimulation 6 Vibrio ≈2x106 cells/species Viral mimic stimulation 6 Poly I:C ≈2x106 cells/species Sampling after 20-24h Leukocyte isolation preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint Fig. 1. Experimental workflow for S. maximus, S. aurata and D. labrax stimulations: A) In vivo stimulation with PBS, Poly I:C, and Vibrio, followed by subsequent sampling of head kidney; B) In vitro stimulation with head kidney leukocyte cultures using the same agents. 120 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint 2.3 In vitro immunostimulation 121 Leukocytes were isolated from turbot, seabream, and seabass fish. The entire head kidney was 122 aseptically isolated and placed in Petri dish es containing 40 ml of cell isolation media (500 ml of 123 Leibovitz L-15 medium (L-15), 10 ml FBS (2%), and 0.02% EDTA. Subsequently, samples were cut 124 and filtered through a 100 µm nylon mesh at a constant flow of cell isolation media. Leukocytes were 125 separated by centrifugation (400 x g, 10 min, 4ºC) of a 40 -ml cell suspension that was layered in a 50 126 ml tube containing 51% Percoll. The interface layer was collected, centrifuged (400 x g, 10 min, 4°C), 127 and washed three times with L-15 medium with 0.1% FBS. After cell counting and viability assessment 128 (trypan blue exclusion test), samples were pooled and divided into aliquots of 2 x 10⁶ cells to reach six 129 replicates per treatment. Each of six technical replicates was stimulated with 20 µl of Poly I:C solution, 130 20 µl of inactivated V . anguillarum, and 20 µl PBS (controls). Leukocyte cultures were incubated for 131 20-24 h at 16 ºC. Cells were collected in 2 ml Eppendorf tubes, pelleted (500 × g for 5 minutes at room 132 temperature), resuspended in RNAlater or flash frozen in liquid nitrogen and stored at -80 °C for RNA-133 seq (Fig. 1B, Tables S1 and S2). 134 2.4 RNA isolation and sequencing 135 RNA extraction was performed according to the FAANG protocols for both in vivo and in vitro 136 stimulations (data.faang.org; see Table S2). Briefly, RNA was isolated and purified using the miRNeasy 137 Kit (QIAGEN) with specific protocol adjustments for head kidney (> 20 mg) and isolated leukocytes 138 (2 × 106 cells). RNA quality and quantity were assessed using a Bioanalyzer (Bonsai Technologies, 139 Madrid, Spain) and a NanoDrop® ND -1000 spectrophotometer (NanoDrop® Technologies Inc., 140 Wilmington, DE, USA). Libraries were prepared at Novogene (UK) using NEBNext Ultra Directional 141 RNA Library Prep Kits (Illumina). Sequencing was performed on an Illumina NovaSeq S4 platform, 142 producing 150 bp paired end reads. 143 2.5 Differential expression analysis 144 Read quality was assessed using FastQC (v0.12.1) [27] and raw reads were trimmed using FastP 145 v0.22.0, with Phred quality <15 and reads with length <30 bp [28]. Clean reads were pseudo-aligned to 146 the corresponding reference transcriptome using Kallisto (v0.46.1) [29] and gene expression was 147 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint quantified using 100 bootstraps . The reference transcriptomes used were ASM1334776v1 148 (GCA_013347765.1) for S. maximus, fSpaAur1.1(GCA_900880695.1) for S. aurata and dlabrax2021 149 (GCA_905237075.1) for D. labrax. Count data were filtered to remove genes with fewer than 5 reads 150 in all samples and represented in only one sample across all conditions . Raw counts were analysed in 151 R v4.3.3 using DESeq2 [30] for differential expression. Variance stabilizing transformation was applied 152 to normalize counts prior to quality assessment and identification of potential outliers through Principal 153 Component Analysis (PCA) plots. Genes with a false discovery rate (FDR) adjusted p-value less than 154 0.05 were identified as differentially expressed genes (DEGs). Analysis in S. maximus, S. aurata and 155 D. labrax data were all performed using the same pipeline. 156 2.6 Enrichment and gene ontology analysis 157 Functional enrichment of the lists of DEGs was performed using g:Profiler ( version 158 e111_eg58_p18_f463989d) [31] using the Ensembl reference genomes mentioned in Section 2.5 (last 159 accession in October, 2024). Enriched biological process Gene Ontology (GO) terms were estimated 160 with the Benjamini-Hochberg FDR adjusted P-value < 0.05 in Reduce + Visualize Gene Ontology 161 software (ReviGO v1.8.1) [32], applying SimRel semantic similarity (threshold : 0.4) to reduce 162 redundancy. Immune -related GO terms were identified using the online tools QuickGO 163 (https://www.ebi.ac.uk/QuickGO/annotations) [33] and AmiGO2 164 (https://amigo.geneontology.org/amigo) [34] (Last accession on October, 2024). Immune-related DEGs 165 were screened and orthologues among species were identified using the Ensembl Biomart tool via the 166 BiomaRt package ( v2.58.2 in R) [35,36]. The resulting orthology mapping table was used to map 167 immune-related DEGs across these species, facilitating comparative analysis of immune gene 168 expression patterns. Venn Diagrams [37] and heatmaps (pheatmap v1.0.12, scale: “row” in R) [38] were 169 used to visualize DEG distributions and expression patterns. 170 2.7 Identification and functional analysis of conserved orthologous genes. 171 To infer the homology of response in the three species using DEGs, five lists of potentially 172 conserved genes were compiled. Each list included DEGs in at least one species with consistent 173 regulatory trends in the others , defined as genes showing the same upregulation or downregulation 174 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint across all three species. List 1 contained genes consistently responding to both Poly I:C and Vibrio 175 under in vitro and in vivo conditions. Lists 2 and 3 included genes with conserved responses in vitro or 176 in vivo conditions and lists 4 and 5 contained genes with conserved response either to Poly I:C, or 177 Vibrio challenges. The latter four lists were further annotated by Kyoto Encyclopedia of Genes and 178 Genomes (KEGG) pathway analysis [39] using the Database for Annotation, Visualization and 179 Integrated Discovery (DA VID) tool [40]. For species comparison, 1:1 orthologs to S. maximus were 180 used to identify the biological pathways that were significantly enriched . The KEGG pathway 181 enrichment analysis was performed with FDR < 0.05. KEGG Mapper was used to provide a graphic 182 representation of interactions between genes. To further investigate the relationship between genes with 183 the same expression profile, protein interaction networks were constructed using STRING v12.0 for the 184 lists [41] based on the S. maximus proteome annotation. 185 3. RESULTS 186 3.1. Sample metadata 187 A total of 108 RNA-seq datasets were used in this study: 36 for turbot, 36 for seabream and 36 188 for seabass. These samples represented 6 experimental groups: control (mock-challenged with PBS), 189 challenged with V . anguillarum and challenged with Poly I:C, both in vivo and in vitro. Full sample and 190 metadata information is provided in Tables S1 and S2. 191 3.2. RNA-sequencing 192 For turbot, t he average number of RNA-seq raw reads per sample was 69,606,581, with an 193 average mapping of 98.9% to the turbot transcriptome (Table S3A); for seabream, the average number 194 of raw reads was 66,673,811 and 99.1% mapping to its transcriptome (Table S3B); and for seabass 195 67,880,066 raw reads on average and 99.2% mapping to its transcriptome (Table S3C). 196 Clear PCA distinctions are evident between control and stimulated samples across species and 197 conditions (Fig. S1). In in vitro conditions, Poly I:C stimulation resulted in the strongest separation in 198 S. aurata and the lowest in D. labrax, while Vibrio stimulation showed consistent differentiation across 199 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint species. In in vivo, both Poly I:C and Vibrio stimulations demonstrated non-overlapping groups, with 200 clear separations along PC1 and PC2 (see details in Supplementary Results). 201 A differential expression analysis was conducted to compare each stimulated condition with its 202 respective control for both in vitro and in vivo conditions. Overall, 55.03% of the genes were 203 upregulated in the treatments compared to the controls , and 49.96% were downregulated on average 204 (Fig S2A). Moreover, the in vitro Vibrio stimulation revealed a significant and somewhat similar number 205 of DEGs in all three species compared to the other experimental conditions with 8,783, 6,238, and 7,296 206 DEGs in turbot, seabream and seabass, respectively (Fig. S2B, Table S4). In the other three conditions, 207 the response was quite species-specific, with seabream showing the weakest response in the two in vivo 208 conditions (Poly I:C (856 DEGs) and Vibrio (430 DEGs)), while the strongest one was in the in vitro 209 Poly I:C (4,335 DEGs); turbot showed the greater number of DEGs in in vivo Poly I:C (6,101 DEGs), 210 compared to in vitro Poly I:C (1,383 DEGs), and in vivo Vibrio (2,557 DEGs); finally, seabass showed 211 a strong response in vivo both for Vibrio (7,296 DEGs) and Poly I:C (6,586 DEGs), while it showed a 212 very weak for in vitro Poly I:C (504 DEGs). In general, turbot showed a slightly higher in vitro than in 213 vivo response (10,166 vs 8,658 DEGs), seabream showed the same trend but with a much larger 214 difference (10,573 vs 1,286 DEGs), and seabass showed the opposite trend, with a higher number of 215 DEGs in vivo than in vitro response (13,881 vs 7,800 DEGs). Additionally, the UpSet plot analysis 216 revealed that shared DEGs across conditions within each species displayed species-specific patterns 217 (Fig. S 3). In S. aurata , the highest number of shared DEGs (2,619) was observed under in vitro 218 conditions (Poly I:C and Vibrio exposures). For S. maximus, 1,862 shared DEGs were found in the 219 comparison between the in vitro (Vibrio) and in vivo (Poly I:C) experiments. A high number of shared 220 DEGs was also observed between in vivo (Poly I:C and Vibrio) and in vitro (Vibrio) challenges for D. 221 labrax (2,786). 222 3.3 Functional enrichment among DEGs 223 Biological Processes (BP) Gene Ontology (GO) analysis of DEGs in the three teleost species 224 revealed conserved immune-related pathways across viral (Poly I:C) and bacterial (Vibrio) challenges, 225 with notable differences in species-specific responses. REVIGO-clustered GO terms highlighted shared 226 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint immune mechanisms, such as immune system process (GO:0002376), immune response 227 (GO:0006955), response to external stimuli/pathogens (e.g. , GO:0006952; GO:0009615), and 228 interspecies interaction (GO:00044419), among others. These results provide a framework for 229 comparative transcriptomic studies of immune evolution in teleost fish (see details in Supplementary 230 Results, Tables S5, S6 and Figure S4). 231 3.4 Turbot, Seabream and Seabass head kidney comparative transcriptome 232 A comparative analysis of immune response -associated DEGs across the three species revealed a 233 small but consistent set of shared genes in each challenge , while most responses remained species -234 specific across both in vitro and in vivo with Poly I:C and Vibrio stimulations. Notably, in vitro 235 stimulation with Poly I:C re vealed upregulation in 31 shared genes across the three species. Overall, 236 turbot exhibited a more divergent transcriptional profile, with stronger induction of interferon and 237 immune signalling genes, whereas seabream and seabass showed higher expression of genes involved 238 in antigen processing and cell regulation (see detailed in Supplementary Results; Table S7, Fig. 2). 239 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint A) In Vitro - Poly I:C Up: 14 Down: 3 U/D: 10 S. maximus (188) Up: 91 Down: 37 Up: 110 Down: 151 Up: 26 Down: 10 128 261 36 2 18 31 27 S. aurata (337) Up: 2 Down: 0 U/D: 0 D. labrax (87) Up: 16 Down: 0 U/D: 2 interferon induced protein 35 interferon regulatory factor 7-like NLR family CARD domain containing 5 suppressor of cytokine signaling 1-like lymphocyte antigen 75 bactericidal permeability-increasing protein-like lectin, galactoside-binding, soluble, 9 (galectin 9)-like 5 CD9 antigen-like ankyrin repeat and SOCS box protein 13-like interferon-induced protein with tetratricopeptide repeats 1-like C-C motif chemokine 19-like interferon-induced protein 44-like signal transducer and activator of transcription 2 secernin 3 transporter associated with antigen processing, subunit type a interferon induced with helicase C domain 1 ubiquitin-like modifier-activating enzyme 1 uncharacterized LOC115577316 interferon-induced GTP-binding protein Mx-like transporter 1, ATP-binding cassette, sub-family B (MDR/TAP) interleukin-10 receptor subunit beta-like interferon regulatory factor 3 proteasome subunit beta type-7-like proteasome activator subunit 2 proteasome activator subunit 1 transcription factor ETV6-like TAP binding protein si:dkey-85k7.12 polyubiquitin-like signal transducer and activator of transcription 1-alpha/beta-like TLR adaptor interacting with endolysosomal SLC15A4 -1 -0.5 0 0.5 1 Up: 31 Down: 0 U/D: 0 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint B) In Vitro – Vibrio DEP domain containing 1B TNF receptor-associated protein 1 dedicator of cytokinesis 5 guanine nucleotide-binding protein G(I)/G(S)/G(T) subunit beta-1-like Rho guanine nucleotide exchange factor (GEF) 1b Ral GTPase activating protein, alpha subunit 1 (catalytic) nuclear factor of activated T-cells, cytoplasmic 1-likeRho associated coiled-coil containing protein kinase 1 late endosomal/lysosomal adaptor, MAPK and MTOR activator 2 interferon regulatory factor 3 heme transporter hrg1-A-like toll-like receptor 9 suppressor of cytokine signaling 1 Kruppel like factor 3 transcription factor EB-like guanine nucleotide-binding protein subunit alpha-13 interferon stimulated exonuclease gene DH domain-containing protein TAP binding protein high affinity immunoglobulin epsilon receptor subunit gamma-like protein phosphatase 5 catalytic subunitleptin receptorBruton agammaglobulinemia tyrosine kinase suppressor of cytokine signaling 1-like phosphatidylinositol-4,5-bisphosphate 3-kinase, catalytic subunit delta heat shock protein HSP 90-alphainterferon induced protein 35 tumor necrosis factor receptor superfamily member 5-like interleukin-12 receptor subunit beta-2-like phosphatidylinositol 4-kinase, catalytic, alpha b interferon alpha/beta receptor 1b-like C-C motif chemokine ligand 17 serine/threonine-protein phosphatase 2B catalytic subunit gamma isoform-like tyrosine-protein kinase Lyn-like mitogen-activated protein kinase kinase kinase 15 TNF receptor-associated factor 2-like interferon regulatory factor 4-like growth factor receptor-bound protein 2TNF receptor superfamily member 21 lymphocyte antigen 75 ribosomal protein S6 kinase 2 alpha G protein subunit beta 1 WASP like actin nucleation promoting factor a T-lymphoma invasion and metastasis-inducing protein 1-like proteasome subunit beta type-7-like proteasome activator subunit 2 tumor necrosis factor, alpha-induced protein 8-like protein 2 B protein phosphatase 3 catalytic subunit alpha Rho GDP dissociation inhibitor alpha G protein-coupled receptor 137B phospholipase C, gamma 2 C-C motif chemokine 19-like toll-like receptor 18 protein phosphatase 3, catalytic subunit, gamma isozyme, b interleukin 17 receptor E like interleukin 19 like transporter 1, ATP-binding cassette, sub-family B (MDR/TAP) bactericidal permeability-increasing protein-like protein kinase, DNA-activated, catalytic subunit V-set immunoregulatory receptor interleukin 16 RAS p21 protein activator 3 proteasome activator subunit 1 vav guanine nucleotide exchange factor 3 cell wall protein DAN4-like inactive phospholipase C-like protein 2 signal-induced proliferation-associated 1-like protein 1 permeability factor 2-like late endosomal/lysosomal adaptor, MAPK and MTOR activator 3 mitogen-activated protein kinase kinase kinase 4 WT1 interacting protein dual specificity phosphatase 6 ras-related C3 botulinum toxin substrate 2 cyclin-dependent kinase inhibitor 1-like myosin IXb C-X-C chemokine receptor type 2-like tumor necrosis factor receptor superfamily member 6B-liketranscription factor ETV6-like KIT proto-oncogene, receptor tyrosine kinase dedicator of cytokinesis 2 Ral GTPase activating protein non-catalytic beta subunit stathmin-likemitogen-activated protein kinase kinase kinase 2tumor necrosis factor ligand superfamily member 13B-like TNF receptor associated factor 3 C-type lectin domain family 4 member M-like ubiquitin-like modifier-activating enzyme 1 si:ch211-210g13.5 transporter associated with antigen processing, subunit type a Ras protein specific guanine nucleotide releasing factor 2 Rho associated coiled-coil containing protein kinase 2 cytokine inducible SH2 containing protein Rap guanine nucleotide exchange factor 6 NFKB inhibitor alphaprotein-cysteine N-palmitoyltransferase HHAT-like protein atypical chemokine receptor 4 toll-like receptor 7 interleukin-1 receptor type 1-like uncharacterized LOC115588278 erythropoietin receptoradenylate cyclase 6 tumor necrosis factor receptor superfamily member 9-like toll-like receptor 21 RELT like 1 SMYD family member 5 C-type lectin domain family 10 member A-like NLR family member X1 interleukin-10-like G protein-coupled receptor kinase interacting ArfGAP 2b ECSIT signaling integratornegative elongation factor complex member A ral guanine nucleotide dissociation stimulator-like 2 rho-related BTB domain-containing protein 2-like phospholipase C gamma 1 platelet basic protein-like ankyrin repeat and SOCS box containing 15 rho-related GTP-binding protein RhoA-D ATP binding cassette subfamily B member 10 adenylate cyclase 9 semaphorin-4B-like C-X-C chemokine receptor type 3-2-like polyubiquitin-like interleukin-13 receptor subunit alpha-2-like MPL proto-oncogene, thrombopoietin receptorRal GTPase activating protein, alpha subunit 2 (catalytic) phospholipase C beta 4 tumor necrosis factor ligand superfamily member 11-like proliferation and apoptosis adaptor protein 15 ADP-ribosylation factor guanine nucleotide-exchange factor 1 (brefeldin A-inhibited) TGF-beta activated kinase 1 (MAP3K7) binding protein 2SH3 domain binding protein 5 C-X-C chemokine receptor type 3-like suppressor of cytokine signaling 3 nuclear factor of kappa light polypeptide gene enhancer in B-cells 2 (p49/p100)TIAM Rac1 associated GEF 2a semaphorin 4G lectin, mannose-binding, 1 peroxiredoxin-1 selectin P RasGEF domain family member 1B -1 -0.5 0 0.5 1 S. maximus (1112) S. aurata (533) D. labrax (426) Up: 293 Down: 432 Up: 68 Down: 105 Up: 35 Down: 75 U/D: 38 Up: 55 Down: 60 725 173 115 99 72 140 148 Up: 33 Down: 40 U/D: 26 Up: 29 Down: 38 U/D: 5 Up: 52 Down: 61 U/D: 27 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint C) In Vivo - Poly I:C 626 9 165 69 12 22 17 S. maximus (734) S. aurata (60) D. labrax (268) Up: 274 Down: 352 Up: 13 Down: 2 U/D: 2 Up: 5 Down: 4 Up: 36 Down: 15 U/D: 18 Up: 9 Down: 3 U/D: 0 Up: 85 Down: 80 Up: 16 Down: 1 U/D: 5 interferon-induced protein 44-like stimulator of interferon response cGAMP interactor 1 sorting nexin 14 signal transducer and activator of transcription 1-alpha/beta-like nicotinamide phosphoribosyltransferaseb toll-like receptor 5 interferon induced with helicase C domain 1 toll-like receptor 7 polymeric immunoglobulin receptor interleukin-10-like interferon induced protein 35 bactericidal permeability-increasing protein-like unc-93 homolog B1, TLR signaling regulator NLR family CARD domain containing 5 galectin-9-like cathepsin L.1 interleukin 15 receptor subunit alpha tumor necrosis factor receptor superfamily, member 1a interferon regulatory factor 1 endonuclease domain-containing 1 protein-like interferon regulatory factor 9 signal transducer and activator of transcription 5a -1 -0.5 0 0.5 1 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint D) In Vivo – Vibrio Fig 2. Venn’s diagrams showing the unique and common differentially expressed genes (DEGs) among S. maximus, S. aurata and D. labrax and heatmaps of expression-normalized DEGs across species. A) In vitro response to Poly I:C stimulation, B) In vitro response to Vibrio stimulation, C) In vivo response to Poly I:C stimulation, and D) In vivo response to Vibrio stimulation. Venn’s diagrams illustrate genes that are up- (red), down- (blue), or up/down-regulated 242 47 422 91 14 4 12 S. maximus (349) S. aurata (77) D. labrax (531) Up: 118 Down: 124 Up: 4 Down: 3 U/D: 5 Up: 41 Down: 25 U/D: 25 Up: 15 Down: 32 Up: 5 Down: 6 U/D: 3 Up: 176 Down: 246 Up: 3 Down: 0 U/D: 1 endonuclease domain-containing 1 protein-like N-acetylmuramoyl-L-alanine amidase-like sorting nexin 14 cathepsin L.1 -1 -0.5 0 0.5 1 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint (yellow) between different species. Color scale in heatmaps: -1 (blue, below mean) to 1 (red, above mean), representing deviations from the average expression. 240 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint 3.5 Conserved immune responses across fish species 241 We identified conserved genes activated by Poly I:C across in vitro and in vivo models (Table S8) 242 and Vibrio-Poly I:C challenges in cell cultures of the three species (Table S9). Seven DEGs were 243 consistently upregulated (bpifcl, ifi35, ifi44, ifih1, lgals9l, nlrc5, stat1a) in both cell cultures and live fish 244 following exposure to Poly I:C, but not in response to Vibrio (see detailed Suppl. Results, Table S8); 15 245 DEGs were detected across the three species when challenged in vitro with both Poly I:C and Vibrio, with 246 13 exhibiting the same directional expression change (ccl19, ifi35, irf3, isg15, psme1, psme2, psmb13a, 247 socs1b, tapbp2, etv7, tap1, tap2a, uba7 ). Notably, bpifcl was upregulated against Poly I:C but 248 downregulated against Vibrio, while ly75 displayed species-specific regulation. No overlapping DEGs 249 were detected in vivo for Vibrio and Poly I:C (see details in Suppl. Results, Table S9). 250 Moreover, DEGs in at least one species which showed the same direction of change in the other 251 two species were identified to look for consistent patterns of gene regulation (Table S10A). A total of 7 252 genes were detected in all species and challenges, four upregulated: endoplasmic reticulum protein 44 253 (erp44), chemokine (C -C motif) ligand 19 ( ccl19), proteasome 20S subunit beta 9a ( psmb9a), 254 immunoglobin binding protein 1 ( igbp1), while three were downregulated: mitogen-activated protein 255 kinase kinase 6 ( map2k6), IL2 inducible T cell kinase (itk) and arrestin beta 2 ( arrb2). An interaction 256 network was explored between those genes using the STRING program, which showed that all genes 257 were interconnected except for igbp1. It was also observed that itk was related to all nodes, showing a 258 weaker interaction with the errp44 gene, but a strong interaction with the remaining genes (Fig. 3). 259 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint 260 ITK IGBP1 MAP2K6 ERRP44 CCL19 PSMB9a ARRB2 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint Fig 3. Protein-protein interaction (PPI) network based on the STRING database and showing fold change profiles. The interaction pattern and fold change (FC) 261 were inspected in differentially expressed genes identified in at least one of the three species while maintaining consistent regulation direction across all species 262 under the four stimulation scenarios. The network was performed with a confidence score of 0.150. Nodes representing genes an d lines between them are the 263 edges, indicating types of evidence for interactions. Blue and pink edges are known interactions (from curated databases and experimentally determined, 264 respectively); grey, green and purple edges are interactions derived from text mining, co-expression and protein homology respectively. Gene expression changes 265 across species and challenges are described in bar charts. X-axis represents experimental conditions: VT P (In Vitro-Poly I:C), VT V (In Vitro-Vibrio), VV P (In 266 Vivo-Poly I:C), and VV V (In Vivo-Vibrio). Y-axis: FC between control and stimulated conditions. Asterisks (*) indicate significant differences (p < 0.05) 267 between control and stimulated conditions for each species in individual experiments (Table S10A). 268 269 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint In addition, genes that shared similar regulatory patterns were also analysed separately in the in 270 vitro and in vivo conditions and under Poly I:C and Vibrio stimulations. The list of identified genes is 271 presented in Table S10. For in vitro, a total of 140 genes showed the same pattern for Poly I:C and Vibrio 272 (Table S10B), while in vivo a total of 59 genes were detected (Table S10C). Meanwhile for Poly I:C a 273 total of 54 genes were revealed (Table S10D) and under Vibrio stimulation a total of 90 genes showed the 274 same pattern (Table S10E). 275 Enrichment pathway analyses for each of these lists were conducted using the Kyoto 276 Encyclopaedia of Genes and Genomes (KEGG) database (see detailed in Suppl. Results, Table S11, Fig. 277 S5). Analyses demonstrated that the toll-like receptor (TLR) signalling pathway is a common pathway of 278 all experimental conditions (Fig. S5). A detailed analysis of the TLR signalling pathway reveals a complex 279 network of interactions involving several interconnected signalling pathways for in vitro (Fig. 4A, Table 280 S12A) and in vivo (Fig. 4C, Table S12B), as well as under Poly I:C (Fig. 5A; Table S12C) and Vibrio (Fig 281 5C, Table S12D) stimulations. Further details are provided in Supplementary Results. 282 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint A) B) STAT2 MIP-1β TYK2 IL-8 IRF3 P38 CD40 IL-12B TRAF6 STAT1 PIK3CA AP-1 IL-1β TRAF3 MKK6 MIG TAB1 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint C) D) Fig. 4. An overview of the toll-like receptor (TLR) signalling pathway under different experimental conditions. A and C represent the TLR pathway in in vitro 283 and in vivo conditions, respectively. Genes upregulated in S. maximus, S. aurata and D. labrax are shown in green colour, while genes that are downregulated 284 across the three species are represented in red; grey boxes indicate no consistent regulation across species or absence of di fferentially expressed genes in any 285 species. In each gene, the left side corresponds to Poly I:C stimulation, and the right side to Vibrio stimulation. Dark blue squares represent genes that are 286 PI3K-Akt signaling pathway Ubiquitin mediated proteolysis Complement and coagulation cascade Cytokine- cytokine receptor interaction JAK-STAT signaling pathway Flagellar assembly Lipopolysaccharide biosynthesis Apoptosis NF-κ B signaling pathway MAPK signaling pathway PI3K-Akt signaling pathway Ubiquitin mediated proteolysis JAK-STAT signaling pathway Flagellar assembly Lipopolysaccharide biosynthesis Apoptosis NF-κ B signaling pathway MAPK signaling pathway Flagellar assembly Lipopolysaccharide biosynthesis PI3K-Akt signaling pathway Ubiquitin mediated proteolysis JAK-STAT signaling pathway Flagellar assembly Lipopolysaccharide biosynthesis Apoptosis NF-κ B signaling pathway MAPK signaling pathway TAK1 MKK7 CTSK TLR1 MKK6 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint regulated in the same direction in both experiments and include a statistically significant differentially expressed gene in at least one species (Table S10B and 287 S10C, respectively). These figures are modified from KEGG map04620 [39]. Figures B and D show protein-protein interaction (PPI) network analysis of genes 288 with the same regulation pattern across species and stimulations for in vitro and in vivo conditions, respectively. The network was constructed using the STRING 289 database with a confidence interval threshold of 0.150, where nodes connected by green edges indicate upregulated genes and red edges downregulated genes. 290 The lines connecting nodes represent different evidence of interactions. 291 292 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint A) B) PI3K-Akt signaling pathway Flagellar assembly Lipopolysaccharide biosynthesis Complement and coagulation cascade Cytokine- cytokine receptor interaction Ubiquitin mediated proteolysis JAK-STAT signaling pathway Apoptosis NF-κ B signaling pathway MAPK signaling pathway Upregulated Gene in 3 species Downregulated Gene in 3 species Not applicable/ not significant Legend In Vitro - Poly I:C stimulation In Vivo – Poly I:C stimulation DEGs in ≥1 species, consistent direction across all 3 species in both Poly I:C challenges LBP TLR7 MKK6 IRF3 STAT1 STAT2 IRF9 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint C) D) Fig. 5. An overview of the toll-like receptor (TLR) signalling pathway under different experimental stimulants. A and C represent the TLR pathway under Poly 293 I:C and Vibrio stimulants, respectively. Genes upregulated in S. maximus, S. aurata and D. labrax are shown in green colour, while genes that are downregulated 294 across the three species are represented in red , grey boxes indicate no consistent regulation across species or absence of differentially expressed genes in any 295 species. In each gene, the left side corresponds to the in vitro, and the right side to the in vivo condition. Dark blue squares represent genes that are regulated in 296 the same direction in both experiments and include a statistically significant differential ly expressed gene in at least one species (Table S10D and S10E, 297 PI3K-Akt signaling pathway Flagellar assembly Lipopolysaccharide biosynthesis Complement and coagulation cascade Cytokine- cytokine receptor interaction Ubiquitin mediated proteolysis JAK-STAT signaling pathway Apoptosis NF-κ B signaling pathway MAPK signaling pathway Upregulated Gene in 3 species Downregulated Gene in 3 species Not applicable/ not significant Legend In Vitro – Vibrio stimulation In Vivo – Vibrio stimulation DEGs in ≥1 species, consistent direction across all 3 species in both Vibrio challenges CTSK MKK6 MKK7 IL12B MIP-1β MIGIl-8 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint respectively). These figures are modified from KEGG map04620 [39] Figures B and D show protein-protein interaction (PPI) network analysis of genes with 298 the same regulation pattern across species and conditions under Poly I:C and Vibrio stimulants, respectively. The network was constructed using the STRING 299 database with a confidence interval threshold of 0.150, where nodes connected by green edges indicate upregulated genes and red edges downregulated genes. 300 The lines connecting nodes represent different evidence of interactions. 301 302 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint 4. DISCUSSION 303 Infectious diseases present a significant challenge to European aquaculture, impacting both 304 animal welfare and the economic growth of the sector. Exploring the genomic basis of immune function 305 and disease resistance in farmed fish has a high priority for improving health in aquaculture breeding and 306 ensuring the industry’s sustainability. This study examined the immune responses of three commercially 307 relevant fish species in Southern Europe, S. maximus, S. aurata, and D. labrax, chosen for their strong 308 research foundation, evolutionary diversity , and their relevance for comparison with other aquaculture 309 teleosts. In the current study, we investigated the impact of viral (Poly I:C) and bacterial (V . anguillarum) 310 stimulations in vitro and in vivo on transcript levels in these species. This effort represents a first step 311 towards elucidating the conserved cellular and organismal responses to pathogens in fish. Transcriptomic 312 profiling was used to analyse gene expression changes, identify DEGs specific to each pathogen type and 313 annotate their biological functions using the GO and KEGG databases. 314 4.1 Comparative transcriptomics: overview 315 To identify conserved and robust gene interaction networks, orthologous DEGs across the three 316 species were compared at two levels: in vitro vs in vivo and Poly I:C vs Vibrio. The general observations 317 suggested no common overlapping between in vitro and in vivo responses to Vibrio, whereas partial 318 overlap was observed under Poly I:C stimulation, indicating better concordance for the viral challenge. 319 This difficulty in finding overlap between the two conditions has also been documented in salmonid s, 320 where only ~25% gene overlapping was found between in vitro and in vivo responses to Poly I:C [42]. In 321 our multi-species comparison, the higher conservation of DEGs across the three species with viral stimuli 322 suggests that a potent stimulant , such as Poly I:C, induces a more evolutionarily conserved response 323 among teleosts, whereas the response to bacterial stimuli exhibits greater variability across experimental 324 systems and species. 325 In addition, a conserved pattern was observed in vitro for all three species against both pathogens. 326 A set of DEGs was identified in response to Poly I:C and Vibrio under in vitro conditions, but no 327 overlapping DEGs were found in vivo between the two stimuli and all species. This highlights the 328 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint complexity of the immune system when the DEGs identified within an entire organism are compared to 329 a simplified cell culture setting. This complexity encompasses systemic immune interactions, tissue -330 specific and species-specific responses, and the microbiome’s role in bacterial and viral infections in a 331 living organism. Beyond variations in immune responses among species to viral vs bacterial pathogens, 332 there may also be differences in their response to both pathogens on different timescales. Our study 333 focused on the 20-24 h post-challenge period, emphasizing the need for further studies at varying intervals 334 to identify conserved responses between organisms. 335 4.2 Immune pathways enriched across fish species 336 To expand the scope of our analysis, we also examined genes and pathways that exhibited 337 consistent regulatory patterns across all species and immunostimulants, focussing not only on DEGs 338 between infection stages, but also exploring similar expression trends across species. This comparative 339 approach revealed that the toll-like receptor (TLR) signalling pathway was significantly enriched across 340 both conditions and challenges, emerging as a central hub of cross-species immune activation. 341 Shared genes among species revealed that the most enriched in vitro pathways were related to the 342 main PRRs of the innate immune system. The PRRs family, which helps identify pathogens and initiate 343 immune response, includes TLR, NLR (NOD -like receptor signalling pathway), RLR (RIG -I-like 344 receptor signalling pathway), and CLR (C-type lectin receptor signalling pathway), among others [43]. 345 These receptor families exhibit functional specialization in pathogen recognition : NLRs functioning as 346 cytoplasmic sensors that identify intracellular bacterial components or stress signals; RLRs, such as RIG-347 I and MDA5, detect viral RNA within the cytoplasm and stimulate type I interferon production through 348 MA VS signalling on mitochondria, which is vital for antiviral defence in both epithelial and immune 349 cells; CLRs attach to carbohydrate structures found on fungi , bacteria and viruses , activating the Syk 350 kinase and CARD9 pathway, which promotes cytokine production and antigen presentation [8]. 351 In contrast to the PRR -dominated in vitro response, the in vivo one revealed enrichment of 352 pathways associated with DNA replication and repair , which may reflect increased proliferation and 353 activation of immune cells during the host response. This process plays a key role in restoring immune 354 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint populations following exposure to pathogens [44], but also reflects elevated haematopoietic activity in 355 general [45]. The divergence between in vitro and in vivo pathway enrichment patterns suggests that while 356 direct pathogen recognition mechanisms dominate immediate cellular response, systemic immune 357 activation involves broader cellular reprogramming and proliferative responses. 358 Analysis of stimulant -specific responses revealed distin ct functional specialization of the fish 359 immune system. Beyond the enriched TLR pathway, Poly I:C activated additional PRR-related pathways 360 such as NLR or CLC, consistent with the multi-layered viral recognition mechanisms employed by innate 361 immunity [46]. In contrast, Vibrio primarily enriched the cytokine-cytokine receptor interaction , 362 emphasizing inflammatory responses and intercellular communication networks [47]. This pattern 363 demonstrates that Poly I:C primarily triggers pathogen recognition cascades similar to those observed 364 under in vitro conditions, while Vibrio exposure activates pathways focused on inflammatory signalling 365 and immune cell coordination. These findings align with comparative studies of viral and bacterial 366 stimulation in teleosts, which demonstrate pathogen-specific pathway enrichment patterns, supporting the 367 concept that fish have evolved specialized immune recognition and response mechanisms tailored to 368 distinct classes of pathogens [23,48]. 369 4.2.1 Toll-like receptors signalling pathway 370 All experiments enriched the toll -like receptor signalling pathway involving numerous genes 371 conserved across all three species. The TLR signalling pathway involves TLRs, which are membrane -372 bound receptors located on immune cells such as dendritic cells and macrophages. These receptors 373 recognize extracellular pathogens or their components (like LPS, flagellin , and viral RNA). Upon 374 activation, they lead to NF -κB and IRF signa lling, resulting in the production of cytokines and type I 375 interferons [8]. The consistent enrichment of the TLR signalling pathway in all experiments, regardless 376 of stimulus (viral vs bacterial) or experimental condition ( in vitro vs in vivo ), reveals a functional 377 convergence toward evolutionarily conserved innate immunity mechanisms. As will be discussed later, 378 this pathway includes numerous genes that belong to other pathways and are potentially conserved across 379 all three species. Downstream signalling cascades converge toward common immune activation pathways 380 [49,50], although specific ligands and primary TLR receptors vary between conditions, as observed in 381 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint our study (e.g., conserved upregulation of TLR1 under in vivo conditions, and conservation of TLR7 and 382 TLR3 genes in response to Poly I:C) . Functional analysis of our data demonstrated this convergence 383 through three fundamental aspects: i) the upregulation of genes associated with inflammatory cytokines 384 or co-stimulatory molecules; ii) the negative regulation of the MAPK signalling pathway; and iii) the 385 upregulation of the JAK-STAT pathway. The overexpression of genes linked to inflammatory cytokines 386 or co-stimulatory molecules can result in excessive inflammation, hyperactivation of immune cells, and 387 potential tissue damage or autoimmune -like symptoms, which are often observe d in cytokine storms 388 during severe viral infections [51]. Conversely, the downregulation of the MAPK signalling pathway, 389 which plays a role in inflammation, cell survival, and proliferation, leads to reduced inflammatory 390 signalling, decreased immune cell activation, and possibly a weakened response to bacterial infections or 391 stress signals. This could serve as a protective mechanism against over-inflammation, or it might indicate 392 immune evasion by pathogens [52]. Moreover, the upregulation of the JAK -STAT pathway, which 393 transmits signals from cytokines like interferons to the nucleus to activate immune genes, plays a crucial 394 role in antiviral defence. This upregulation increases the expression of antiviral genes, improves 395 communication among immune cells, and strengthens the body’s ability to fight off viruses [53]. 396 4.3 Key conserved genes 397 In addition, the comparative transcriptomic study revealed potentially conserved differentially 398 expressed genes shared between in vitro and in vivo in responses to Poly I:C (Table S8), DEGs shared 399 between Poly I:C and Vibrio in vitro stimulation (Table S9) and genes that demonstrated a uniform 400 expression pattern in the four experimental challenges performed (DEGs and non-DEGs) (Fig. 3). This 401 integrated approach allow ed to investigate the most important genes involved in the three species ’ 402 response to viruses and bacteria, particularly those related to chemokines, interferon, antigen processing 403 and presentation, cell signalling regulators and MAPK (for details, see Supplementary Discussion). 404 4.3.1 Chemokine related genes 405 Chemokines, essential signalling molecules, play a fundamental role in modulating immune 406 responses by recruiting immune cells to sites of infection and mediating communication between innate 407 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint and adaptive immunity [54]. Notably, ccl19 emerged as a primary gene, showing consistent upregulation 408 both in vitro and in vivo in response to bacterial and viral pathogens across the three species studied (Table 409 S9, Fig. 3). This gene is known to be involved in the recruitment of T cells and maturation of dendritic 410 cells (DCs), thereby facilitating adaptive immune response. The TLR pathway identified in this study 411 links ccl19 upregulation with the production of pro -inflammatory cytokines ( il-12, il-1β) and co -412 stimulatory molecules, consistent with its role in bridging innate and adap tive immunity during host 413 defence responses in other species [55,56]. 414 4.3.2 Interferon-related genes 415 Interferons (IFNs) are a family of pleiotropic cytokines that represent a crucial part of the innate 416 immune response against invading pathogens [57] by inducing hundreds of IFN-stimulated genes (ISGs) 417 [58]. IFN regulatory factors (IRFs) are transcription factors regulating the expression of IFN and other 418 related genes [59]. This study identified several conserved IFN-related genes in response to stimulators 419 across the three species, including the viral sensor ifih1, IFN-induced genes such as ifi35, ifi44, and isg15, 420 and regulators such as irf3, etv7 and uba7 (Table S8, S9) . Infection with Poly I:C induced strong 421 upregulation of ifih1, a helicase that detects viral dsRNA and activates type I IFN signalling cascades 422 [60–62]. This process subsequently leads to the induction of isg15, a major interferon -stimulated gene 423 involved in ISGylation [63]. The E1 enzyme uba7, along with other ligases, mediates this ISGylation by 424 modulating the stability and activity of proteins during infection; uba7 was also significantly upregulated, 425 supporting its pivotal antiviral role [58,64]. Although knowledge regarding ifi35, etv7, and uba7 in 426 teleosts is still limited, ifi35 is known to regulate pro-inflammatory cytokines and ifn-β [65], while etv7 427 serves as a negative regulator that may prevent excessive inflammatory signalling [66]. Both genes 428 emerged as conserved upregulated DEGs, indicating their role in controlling interferon responses in fish. 429 Additionally, the conserved ifi44, which encodes an antiviral intracellular protein [67,68] and irf3, which 430 is essential for mediating type I IFN and ISG expression [59,69], displayed increased expression in 431 response to viral and bacterial stimuli, underlining their central roles in pathogen defence. 432 4.3.3 Antigen processing and presentation-related genes 433 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint Upon the invasion of a pathogenic antigen into the cytoplasm, the initiation of antigen processing 434 and presentation (APP) occurs [70]. We observed activation of this APP process in response to viral and 435 bacterial stimuli, as evidenced by the upregulation of several DEGs in all species for both in vitro 436 challenges: psme1, psme2, psmb13a, tapbp.2, tap1, tap2a and in Poly I:C in both conditions: nlrc5 (Table 437 S8, S9). Additionally, two other DEGs were upregulated in all four conditions: psmb9a and erp44 (Fig 438 3). These comparative findings demonstrate that the molecules involved in APP are essential for the 439 immune system, not only in cell models but also in vivo, showing their conservation in the three species 440 studied and their conserved positive regulation in response to external infections across taxa. 441 4.3.4 Cellular signalling regulators 442 In this study, we identified expression changes of several conserved cellular signalling regulators 443 20-24 h after stimulation (Tables S8, S9, Fig. 3). As they orchestrate immune responses, upregulation of 444 these conserved genes aims to enhance immune activation, reflecting their roles in the release of 445 antibacterial peptide s (bpifcl) [71], immunomodulation ( lgals9l5) [72,73], B cell receptor signalling 446 (igbp1) [74], regulation of antiviral transcription and inflammatory genes ( stat1a) [75,76] and feedback 447 inhibition of cytokine signalling (socs1b) [77,78], respectively. Conversely, downregulation of arrb2 and 448 itk may reflect pathogen strategies to modulate surface receptor signalling [79] and, in the case of itk, also 449 T cell activation for immune evasion [80]. 450 4.3.5 Mitogen-activated protein kinase 451 In the experiments, the map2k6 gene was consistently downregulated across all conditions (Fig. 452 3). Examination of the TLR-MAPK pathway revealed that most genes also exhibited downregulation. 453 Even though MAPK pathway upregulation would typically be expected due to its role in activating 454 inflammatory cytokines and immune responses, the significant downregulation observed here in response 455 to the experimental challenges suggests a distinct modulation of this pathway under these conditions. In 456 addition, our study conducted at 20-24h could reflect the early activation of map2k6 as an initial defence 457 mechanism followed by negative feedback regulation [23,81]. Although this study did not directly 458 investigate the underlying causes, it has been observed that map2k6 expression can be significantly 459 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint decreased in certain contexts, which could reflect an evolutionary strategy employed by pathogens to 460 evade host immunity [25,82]. 461 5. CONCLUSIONS 462 This study provides the first comprehensive comparative transcriptomic analysis for three 463 acanthopterygian fish species, S. maximus, S. aurata, and D. labrax, in response to viral and bacterial 464 infections, using both in vitro and in vivo conditions. The findings reveal that all three species exhibit 465 remarkable changes in the expression of immune-related genes in vitro. In contrast, in vivo, they show an 466 enhanced response related to DNA replication and repair . Additionally, viral and bacterial challenges 467 elicited differentiated immune pathway activations, reflecting specialized host defence mechanisms . 468 Moreover, a more conserved response was observed across the three species when exposed to the viral 469 mimic than to the heat-killed gram-negative bacteria. Analysis of orthologous gene sets across the three 470 species revealed conservation of genes involved in chemokines, interferons, antigen processing and 471 presentation, cell signal ling regulation, and MAPK pathways in response to external pathogens , with 472 some of these genes reported for the first time as part of a shared response to Poly I:C or Vibrio. Our 473

demonstrate not only several of the deeply conserved pathways in immune responses among 474 Acanthopterygian fish species, but also the use of the comparative method to uncover responses that may 475 have been overlooked in single-species studies, e.g. due to differences between species in response time. 476 This comprehensive approach offers valuable insights into immune responses across multiple species and 477 conditions, while significantly contributing to the critical field of comparative immunology in 478 aquaculture. 479 Funding 480 This study is part of the AQUA -FAANG project, which received funding from the European Union’s 481 Horizon 2020 research and innovation programme under grant agreement Nº 817923. Additional funding 482 was provided by Xunta de Galicia local government (Spain) (ED431C 2022/33), which also supported 483 the research fellowship of Aramburu O (refs. ED481A-2020/119). It also supported by postdoctoral 484 programme of the Xunta de Galicia (Consellería de Cultura, Educación, Formación Profesional e 485 Universidades) to R. Rodríguez-Vázquez (ED481B-2023-104). 486 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint Credit Authorship contribution statement 487 R.R.V .: Conceptualization, methodology, software, formal analysis, investigation, data curation, 488 visualization, writing - original draft, writing - review & editing; M.K.G.: Conceptualization, 489 methodology, investigation, writing - review & editing ; O.A.: methodology , data curation, investigation, 490 resources, writing - review & editing; J.R.: methodology, data curation, resources, writing - review & 491 editing; C.S.T.: funding acquisition, methodology, resources, project administration, writing - review & 492 editing; S. F.: methodology, investigation, resources, writing - review & editing; R.F.: methodology, data 493 curation, resources; L.B.: funding acquisition, methodology, resources, project administration, writing - 494 review & editing; P.M.: funding acquisition, methodology, resources, project administration writing - 495 review & editing; D.R.: methodology , investigation, supervision, resources, funding acquisition, project 496 administration, writing - review & editing ; H.J.M.: conceptualization, methodology , investigation, 497 supervision, resources, funding acquisition, project administration, writing - review & editing. 498 499 Declaration of competing interest 500 The authors declare no conflict of interest 501 Acknowledgments 502 We acknowledge the technical support and informatic resources provided by the Centro de 503 Supercomputación de Galicia (CESGA). We would also like to thank Pantelis Katharios, Tereza 504 Manousaki, Ioannis Papadakis and Elena Sarropoulou for their help in gilthead sea bream sampling and 505 preliminary analyses. Thank also certain A QUA-FAANG partners for fruitful discussions and protocol 506 optimizations. 507 Data availability 508 Raw RNA -seq datasets can be accessed through the ENA repository under the following accession 509 number: turbot (PRJEB47933), seabream ( PRJEB64880, PRJEB64877) and seabass ( PRJEB52285, 510 PRJEB52283). Detailed metadata for the samples and prepared libraries are available in Supplementary 511 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint Tables S1 and S2, respectively. Detailed experimental protocols are publicly available in the FAANG 512 repository (data.faang.org) and following the URLs facilitated in Supplementary Tables S1 and S2. 513 Additional data will be made available on request. Supplementary data has been attached. 514 515 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint

[1] FAO, The State of World Fisheries and Aquaculture, FAO, Rome, 2024. https://doi.org/10.4060/cd0683en. [2] R. Fernández-González, M.I. Pérez-Pérez, J. Correia-Da-Silva, Production strategies, productivity changes and innovation: An analysis of European turbot aquaculture from 2009 to 2020, Rev Aquac 15 (2022) 610–624. https://doi.org/10.13039/501100011033/. [3] K. Mhalhel, M. Levanti, F. Abbate, R. Laurà, M.C. Guerrera, M. Aragona, C. Porcino, M. Briglia, A. Germanà, G. Montalbano, Review on Gilthead Seabream (Sparus aurata) Aquaculture: Life Cycle, Growth, Aquaculture Practices and Challenges, J Mar Sci Eng 11 (2023) 2008. https://doi.org/10.3390/jmse11102008. [4] L. Monzón-Atienza, J. Bravo, A. Serradell, D. Montero, A. Gómez-Mercader, F. Acosta, Current Status of Probiotics in European Sea Bass Aquaculture as One Important Mediterranean and Atlantic Commercial Species: A Review, Animals 13 (2023) 2369. https://doi.org/10.3390/ani13142369. [5] FAO, Fisheries and Aquaculture. Global production by production source Quantity (1950-2022), (2024). https://www.fao.org/fishery/statistics- query/en/global_production/global_production_quantity. [6] K.D. Lafferty, C.D. Harvell, J.M. Conrad, C.S. Friedman, M.L. Kent, A.M. Kuris, E.N. Powell, D. Rondeau, S.M. Saksida, Infectious Diseases Affect Marine Fisheries and Aquaculture Economics, Ann Rev Mar Sci (2015) 471 –496. https://doi.org/10.1146/annurev-marine-010814- 015646. [7] S. Reyes-Cerpa, K. Maisey, F. Reyes-Lpez, D. Toro-Ascuy, A. Mara, M. Imarai, Fish Cytokines and Immune Response, in: New Advances and Contributions to Fish Biology, InTech, 2012. https://doi.org/10.5772/53504. [8] D.M. Mokhtar, G. Zaccone, A. Alesci, M. Kuciel, M.T. Hussein, R.K.A. Sayed, Main Components of Fish Immunity: An Overview of the Fish Immune System, Fishes 8 (2023). https://doi.org/10.3390/fishes8020093. [9] L. yun Zhu, L. Nie, G. Zhu, L. xing Xiang, J. zhong Shao, Advances in research of fish immune- relevant genes: A comparative overview of innate and adaptive immunity in teleosts, Dev Comp Immunol 39 (2013) 39–62. https://doi.org/10.1016/j.dci.2012.04.001. [10] B. Sayyaf Dezfuli, M. Lorenzoni, A. Carosi, L. Giari, G. Bosi, Teleost innate immunity, an intricate game between immune cells and parasites of fish organs: who wins, who loses, Front Immunol 14 (2023) 1250835. https://doi.org/10.3389/fimmu.2023.1250835. [11] A. Häder, S. Schäuble, J. Gehlen, N. Thielemann, B.C. Buerfent, V . Schüller, T. Hess, T. Wolf, J. Schröder, M. Weber, K. Hünniger, J. Löffler, S. Vylkova, G. Panagiotou, J. Schumacher, O. Kurzai, Pathogen-specific innate immune response patterns are distinctly affected by genetic diversity, Nat Commun 14 (2023) 3239. https://doi.org/10.1038/s41467-023-38994-5. [12] E. Holen, K.K. Lie, P. Araujo, P.A. Olsvik, Pathogen recognition and mechanisms in Atlantic cod (Gadus morhua) head kidney cells. Bacteria (LPS) and virus (poly I:C) signals through different pathways and affect distinct genes., Fish Shellfish Immunol 33 (2012) 267 –276. https://doi.org/10.1016/j.fsi.2012.05.013. [13] S.S. Rao, P.A. Nelson, H.S. Lunde, G.T. Haugland, Evolutionary, comparative, and functional analyses of STA Ts and regulation of the JAK -STAT pathway in lumpfish upon bacterial and preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint poly(I:C) exposure, Front Cell Infect Microbiol 13 (2023) 1252744. https://doi.org/10.3389/fcimb.2023.1252744. [14] G.H. Nam, A. Mishra, J.A. Gim, H.E. Lee, A. Jo, D. Yoon, A. Kim, W.J. Kim, K. Ahn, D.H. Kim, S. Kim, H.J. Cha, Y .H. Choi, C. Il Park, H.S. Kim, Gene expression profiles alteration after infection of virus, bacteria, and parasite in the Olive flounder (Paralichthys olivaceus), Sci Rep 8 (2018) 18065. https://doi.org/10.1038/s41598-018-36342-y. [15] M. Pilar Sepulcre, G. López -Castejón, J. Meseguer, V . Mulero, A Toll receptor in shrimp, Mol Immunol 44 (2007) 1999–2008. https://doi.org/10.1016/j.molimm.2006.09.022. [16] Y . Song, X. Dong, G. Hu, Transcriptome analysis of turbot (Scophthalmus maximus) head kidney and liver reveals immune mechanism in response to Vibrio anguillarum infection, J Fish Dis 45 (2022) 1045–1057. https://doi.org/10.1111/jfd.13628. [17] X. Ning, L. Sun, Gene network analysis reveals a core set of genes involved in the immune response of Japanese flounder (Paralichthys olivaceus) against Vibrio anguillarum infection, Fish Shellfish Immunol 98 (2020) 800–809. https://doi.org/10.1016/j.fsi.2019.11.033. [18] Q. Yang, X.D. Yang, M.Q. Liu, C. Zeng, H.K. Zhao, K.W. Xiang, Z.S. Hou, H.S. Wen, J.F. Li, Transcriptome analysis of liver, gill and intestine in rainbow trout (Oncorhynchus mykiss) symptomatically or asymptomatically infected with Vibrio anguillarum, Fis h Shellfish Immunol 135 (2023) 108643. https://doi.org/10.1016/j.fsi.2023.108643. [19] Y . Liu, Z.Z. Xin, D.Z. Zhang, Z.F. Wang, X.Y . Zhu, B.P. Tang, S.H. Jiang, H. Bin Zhang, C.L. Zhou, X.Y . Chai, Q.N. Liu, Transcriptome analysis of yellow catfish (Pelteobagrus fulvidraco) liver challenged with polyriboinosinic polyribocytidylic acid (poly I:C), Fish Shellfish Immunol 68 (2017) 395–403. https://doi.org/10.1016/j.fsi.2017.07.030. [20] F. Han, Y . Zhang, N. Song, T. Gao, TLR pathway signaling molecules in burbot (Lota lota): molecular characterization, basal expression, and their response to Poly(I:C), Fish Shellfish Immunol 140 (2023) 108939. https://doi.org/10.1016/j.fsi.2023.108939. [21] Q.N. Liu, Y .Y . Tang, M.J. Zhou, S. Luo, Y .T. Li, G. Wang, D.Z. Zhang, H. Yang, B.P. Tang, W.F. He, Differentially expressed genes involved in immune pathways from yellowhead catfish (Tachysurus fulvidraco) after poly (I:C) challenge, Int J Biol Macromol 1 83 (2021) 340 –345. https://doi.org/10.1016/j.ijbiomac.2021.04.167. [22] S.S. Rao, H.S. Lunde, D.W.P. Dolan, A.K. Fond, K. Petersen, G.T. Haugland, Transcriptome-wide analyses of early immune responses in lumpfish leukocytes upon stimulation with poly(I:C), Front Immunol 14 (2023) 1198211. https://doi.org/10.3389/fimmu.2023.1198211. [23] O. Aramburu, B. Gómez-Pardo, P. Rodríguez-Villamayor, A. Blanco-Hortas, J. Lamas, P. Dewari, D. Perojil-Morata, P. Boudinot, D.J. MacQueen, C. Bouza, P. Martínez, Multiomics uncovers the epigenomic and transcriptomic response to viral and bacterial stimula tion in turbot, Gigascience 14 (2025) 1–22. https://doi.org/10.1093/gigascience/giaf077. [24] J. Saravia, K. Paschke, J.P . Pontigo, D. Nualart, J.M. Navarro, L. Vargas -Chacoff, Effects of temperature on the innate immune response on Antarctic and sub -Antarctic fish Harpagifer antarcticus and Harpagifer bispinis challenged with two immunostimulants, LPS and Poly I:C: In vivo and in vitro approach, Fish Shellfish Immunol 130 (2022) 391 –408. https://doi.org/10.1016/j.fsi.2022.09.025. [25] N. Sood, D.K. Verma, A. Paria, S.C. Yadav, M.K. Yadav, M.K. Bedekar, S. Kumar, T.R. Swaminathan, C.V . Mohan, K. V . Rajendran, P.K. Pradhan, Transcriptome analysis of liver elucidates key immune-related pathways in Nile tilapia Oreochromis niloticus following infection preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint with tilapia lake virus, Fish Shellfish Immunol 111 (2021) 208 –219. https://doi.org/10.1016/j.fsi.2021.02.005. [26] Y . Song, L. Song, M.A. Soomro, X. Dong, G. Hu, Transcriptome analysis reveals immune-related differentially expressed genes in the hepatopancreas and gills of Penaeus vannamei after Vibrio anguillarum infection, Aquac Res 52 (2021) 4303–4316. https://doi.org/10.1111/are.15268. [27] S. Andrews, FastQC: A Quality Control Tool for High Throughput Sequence Data, (2010). https://www.bioinformatics.babraham.ac.uk/projects/fastqc/ (accessed April 27, 2025). [28] S. Chen, Y . Zhou, Y . Chen, J. Gu, Fastp: An ultra -fast all -in-one FASTQ preprocessor, Bioinformatics 34 (2018) i884–i890. https://doi.org/10.1093/bioinformatics/bty560. [29] N.L. Bray, H. Pimentel, P. Melsted, L. Pachter, Near -optimal probabilistic RNA -seq quantification, Nat Biotechnol 34 (2016) 525–527. https://doi.org/10.1038/nbt.3519. [30] M.I. Love, W. Huber, S. Anders, Moderated estimation of fold change and dispersion for RNA - seq data with DESeq2, Genome Biol 15 (2014) 550. https://doi.org/10.1186/s13059-014-0550-8. [31] L. Kolberg, U. Raudvere, I. Kuzmin, P. Adler, J. Vilo, H. Peterson, G:Profiler-interoperable web service for functional enrichment analysis and gene identifier mapping (2023 update), Nucleic Acids Res 51 (2023) W207–W212. https://doi.org/10.1093/nar/gkad347. [32] F. Supek, M. Bošnjak, N. Škunca, T. Šmuc, Revigo summarizes and visualizes long lists of gene ontology terms, PLoS One 6 (2011) e21800. https://doi.org/10.1371/journal.pone.0021800. [33] D. Binns, E.C. Dimmer, R.P . Huntley, D.G. Barrell, C. O’Donovan, R. Apweiler, QuickGO:a web- based tool for Gene Ontology searching, Bioinformatics 25 (2009) 3045 –3046. https://doi.org/10.1002/pmic.200800002. [34] S. Carbon, A. Ireland, C.J. Mungall, S. Shu, B. Marshall, S. Lewis, J. Lomax, C. Mungall, B. Hitz, R. Balakrishnan, M. Dolan, V . Wood, E. Hong, P. Gaudet, AmiGO: Online access to ontology and annotation data, Bioinformatics 25 (2009) 288 –289. https://doi.org/10.1093/bioinformatics/btn615. [35] R.J. Kinsella, A. Kähäri, S. Haider, J. Zamora, G. Proctor, G. Spudich, J. Almeida -King, D. Staines, P. Derwent, A. Kerhornou, P. Kersey, P. Flicek, Ensembl BioMarts: A hub for data retrieval across taxonomic space, Database 2011 (2011) 1–9. https://doi.org/10.1093/database/bar030. [36] S. Durinck, Y . Moreau, A. Kasprzyk, S. Davis, B. De Moor, A. Brazma, W. Huber, BioMart and Bioconductor: A powerful link between biological databases and microarray data analysis, Bioinformatics 21 (2005) 3439–3440. https://doi.org/10.1093/bioinformatics/bti525. [37] H. Heberle, V .G. Meirelles, F.R. da Silva, G.P. Telles, R. Minghim, InteractiVenn: A web-based tool for the analysis of sets through Venn diagrams, BMC Bioinformatics 16 (2015) 1 –7. https://doi.org/10.1186/s12859-015-0611-3. [38] R. Kolde, pheatmap: Pretty Heatmaps, (2019). https://cran.r-project.org/package=pheatmap. [39] M. Kanehisa, S. Goto, Y . Sato, M. Furumichi, M. Tanabe, KEGG for integration and interpretation of large -scale molecular data sets, Nucleic Acids Res 40 (2012). https://doi.org/10.1093/nar/gkr988. [40] B.T. Sherman, M. Hao, J. Qiu, X. Jiao, M.W. Baseler, H.C. Lane, T. Imamichi, W. Chang, DA VID: a web server for functional enrichment analysis and functional annotation of gene lists (2021 update), Nucleic Acids Res 50 (2022) W216–W221. https://doi.org/10.1093/nar/gkac194. preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint [41] D. Szklarczyk, A.L. Gable, D. Lyon, A. Junge, S. Wyder, J. Huerta -Cepas, M. Simonovic, N.T. Doncheva, J.H. Morris, P. Bork, L.J. Jensen, C. V on Mering, STRING v11: Protein -protein association networks with increased coverage, supporting functional discove ry in genome-wide experimental datasets, Nucleic Acids Res 47 (2019) D607 –D613. https://doi.org/10.1093/nar/gky1131. [42] T.C. Clark, S. Naseer, M.K. Gundappa, A. Laurent, A. Perquis, B. Collet, D.J. Macqueen, S.A.M. Martin, P. Boudinot, Conserved and divergent arms of the antiviral response in the duplicated genomes of salmonid fishes, Genomics 115 (2023) 110663. https://doi.org/10.1016/j.ygeno.2023.110663. [43] M. Stosik, B. Tokarz-Deptuła, W. Deptuła, Immunological memory in teleost fish, Fish Shellfish Immunol 115 (2021) 95–103. https://doi.org/10.1016/j.fsi.2021.05.022. [44] S. Ragu, G. Matos -Rodrigues, B.S. Lopez, Replication stress, DNA damage, inflammatory cytokines and innate immune response, Genes (Basel) 11 (2020). https://doi.org/10.3390/genes11040409. [45] J. Tong, J. Song, W. Zhang, J. Zhai, Q. Guan, H. Wang, G. Liu, C. Zheng, When DNA -damage responses meet innate and adaptive immunity, Cellular and Molecular Life Sciences 81 (2024) 185. https://doi.org/10.1007/s00018-024-05214-2. [46] S. Poynter, G. Lisser, A. Monjo, S. DeWitte-Orr, Sensors of infection: Viral nucleic acid PRRs in fish, Biology (Basel) 4 (2015) 460–493. https://doi.org/10.3390/biology4030460. [47] C. Hernández-Cabanyero, E. Sanjuán, F.E. Reyes-López, E. V allejos-Vidal, L. Tort, C. Amaro, A Transcriptomic Study Reveals That Fish Vibriosis Due to the Zoonotic Pathogen Vibrio vulnificus Is an Acute Inflammatory Disease in Which Erythrocytes May Play an Important Role, Front Microbiol 13 (2022). https://doi.org/10.3389/fmicb.2022.852677. [48] A. Krasnov, L.H. Johansen, C. Karlsen, L. Sveen, E. Ytteborg, G. Timmerhaus, C.C. Lazado, S. Afanasyev, Transcriptome Responses of Atlantic Salmon (Salmo salar L.) to Viral and Bacterial Pathogens, Inflammation, and Stress, Front Immunol 12 (2021). https://doi.org/10.3389/fimmu.2021.705601. [49] M.K. Purcell, K.D. Smith, L. Hood, J.R. Winton, J.C. Roach, J. Roach, C. Biochem Physiol, Conservation of Toll-Like Receptor Signaling Pathways in Teleost Fish, 2006. www.tigr.org. [50] Y . Palti, Toll-like receptors in bony fish: From genomics to function, Dev Comp Immunol 35 (2011) 1263–1272. https://doi.org/10.1016/j.dci.2011.03.006. [51] K.P. Michalak, A.Z. Michalak, Understanding chronic inflammation: couplings between cytokines, ROS, NO, Cai2+, HIF -1α, Nrf2 and autophagy, Front Immunol 16 (2025). https://doi.org/10.3389/fimmu.2025.1558263. [52] A. Ghara, G.S.P. Matada, R.K. Sah, S. S, S.S. Boneka, A. Paik, S. Majumder, P. Pramanick, B. Metikurki, Unravelling Immunity: Understanding Inflammatory Pathways as Key Players in Immune Response and Disease Progression, Curr Mol Biol Rep 11 (2025). https://doi.org/10.1007/s40610-025-00167-2. [53] J. Butters, Cytokine Signaling Pathways in Inflammatory Response: Mechanisms and Modulation, (2024). https://doi.org/10.4172/2576-3881.1000500. [54] A. Alejo, C. Tafalla, Chemokines in teleost fish species, Dev Comp Immunol 35 (2011) 1215 – 1222. https://doi.org/10.1016/j.dci.2011.03.011. preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint [55] M.F. Bachmann, M. Kopf, B.J. Marsland, Chemokines: more than just road signs, 6 (2006) 159– 164. [56] A.G. Shabgah, Z.M.J. Al-Obaidi, H.S. Rahman, W.K. Abdelbasset, W. Suksatan, D.O. Bokov, L. Thangavelu, A.T. Jalil, F. Jadidi-Niaragh, H. Mohammadi, K. Mashayekhi, J.G. Navashenaq, Does CCL19 act as a double-edged sword in cancer development?, Clin Exp Immunol 207 (2022) 164– 175. https://doi.org/10.1093/cei/uxab039. [57] G.M. Boxx, G. Cheng, The Roles of Type i Interferon in Bacterial Infection, Cell Host Microbe 19 (2016) 760–769. https://doi.org/10.1016/j.chom.2016.05.016. [58] F. Thery, D. Eggermont, F. Impens, Proteomics Mapping of the ISGylation Landscape in Innate Immunity, Front Immunol 12 (2021) 720765. https://doi.org/10.3389/fimmu.2021.720765. [59] Y . Sun, Z. Cao, P. Zhang, C. Wei, J. Li, Y . Wu, Y . Zhou, IFN regulatory factor 3 of golden pompano and its NLS domain are involved in antibacterial innate immunity and regulate the expression of type I interferon (IFNa3), Front Immunol 14 (2023) 1128196. https://doi.org/10.3389/fimmu.2023.1128196. [60] A.M.S. Andresen, T. Gjøen, Chitosan nanoparticle formulation attenuates poly (I:C) induced innate immune responses against inactivated virus vaccine in Atlantic salmon (Salmo salar), Comp Biochem Physiol Part D Genomics Proteomics 40 (2021) 100915. https://doi.org/10.1016/j.cbd.2021.100915. [61] C. Collins, L. Chaumont, M. Peruzzi, N. Jamak, P. Boudinot, J. Béjar, P. Moreno, D. Álvarez Torres, B. Collet, Effect of a loss of the mda5/ifih1 gene on the antiviral resistance in a Chinook salmon Oncorhynchus tshawytscha cell line, PLoS One 19 (2024) e 0311283. https://doi.org/10.1371/journal.pone.0311283. [62] T. Sun, J. Huang, L. Zhu, S. Wu, L. Zhao, Y . Kang, Integrative mRNA-miRNA interaction analysis associated with the immune response in the head kidney of rainbow trout (Oncorhynchus mykiss) after infectious hematopoietic necrosis virus infection, Fish Shel lfish Immunol 142 (2023) 109140. https://doi.org/10.1016/j.fsi.2023.109140. [63] D.J. Morales, D.J. Lenschow, The antiviral activities of ISG15, J Mol Biol 425 (2013) 4995–5008. https://doi.org/10.1016/j.jmb.2013.09.041. [64] N. V . Giannakopoulos, E. Arutyunova, C. Lai, D.J. Lenschow, A.L. Haas, H.W. Virgin, ISG15 Arg151 and the ISG15-Conjugating Enzyme UbE1L Are Important for Innate Immune Control of Sindbis Virus, J Virol 83 (2009) 1602–1610. https://doi.org/10.1128/jvi.01590-08. [65] X. Gao, Y . Zhang, J. Zheng, X. Yang, Y . Wang, Q. Qin, X. Huang, Y . Huang, Grouper interferon- induced protein 35, a CP -interacting protein, inhibits fish nodavirus replication via positively regulating host interferon and inflammatory immune response, Fish Shellfish Immunol 128 (2022) 113–122. https://doi.org/10.1016/j.fsi.2022.07.077. [66] H.M. Froggatt, A.T. Harding, R.R. Chaparian, N.S. Heaton, ETV7 limits antiviral gene expression and control of influenza viruses, Sci Signal 14 (2021). https://doi.org/10.1126/SCISIGNAL.ABE1194. [67] D.C. Busse, D. Habgood-coote, S. Clare, C. Brandt, I. Bassano, M. Kaforou, J. Herberg, M. Levin, J.-F. Eléouët, P. Kellam, J.S. Tregoning, Interferon -Induced Protein 44 and Interferon -Induced Protein 44-like Restrict Replication of Respiratory Syncytial Virus, J Virol 94 (2020) e00297. preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint [68] L.C. Hallen, Y . Burki, M. Ebeling, C. Broger, F. Siegrist, K. Oroszlan-Szovik, B. Bohrmann, U. Certa, S. Foser, Antiproliferative activity of the human IFN-α inducible protein IFI44, Journal of Interferon and Cytokine Research 27 (2007) 675–680. https://doi.org/10.1089/jir.2007.0021. [69] M. Emam, S. Kumar, K. Eslamloo, A. Caballero -Solares, J.R. Hall, X. Xue, H. Paradis, R.L. Gendron, J. Santander, M.L. Rise, Transcriptomic response of lumpfish (Cyclopterus lumpus) head kidney to viral mimic, with a focus on the interferon regulatory factor family, Front Immunol 15 (2024) 1–21. https://doi.org/10.3389/fimmu.2024.1439465. [70] H.L.P. Jatin M. Vyas, Annemartha G. Van der Veen, The know unknowns of antigen processing and presentation, Nat Rev Immunol 8 (2008) 607–618. https://doi.org/10.1038/nri2368.The. [71] F. Colombo, O. Illescas, S. Noci, F. Minnai, G. Pintarelli, A. Pettinicchio, A. Vannelli, L. Sorrentino, L. Battaglia, M. Cosimelli, T.A. Dragani, M. Gariboldi, Gut microbiota composition in colorectal cancer patients is genetically regulated, Sci Rep (202 2) 11424. https://doi.org/10.1038/s41598-022-15230-6. [72] S. Schlichtner, N.H. Meyer, I.M. Yasinska, N. Aliu, S.M. Berger, B.F. Gibbs, E. Fasler-Kan, V . V . Sumbayev, Functional role of galectin -9 in directing human innate immune reactions to Gram - negative bacteria and T cell apoptosis, Int Immunopharmacol 100 (2 021) 108155. https://doi.org/10.1016/j.intimp.2021.108155. [73] S. Kandel, P. Adhikary, G. Li, K. Cheng, The TIM3/Gal9 signaling pathway: An emerging target for cancer immunotherapy, Cancer Lett 510 (2021) 67 –78. https://doi.org/10.1016/j.canlet.2021.04.011. [74] S. Jiang, D. Li, Z. Liang, Y . Wang, X.F. Pei, J. Tang, High expression of IGBP1 correlates with poor prognosis in esophageal squamous cell carcinoma, International Journal of Biological Markers 35 (2020) 33–40. https://doi.org/10.1177/1724600819896374. [75] G.J. Fernandez, J.M. Ramírez-Mejía, S. Urcuqui-Inchima, Transcriptional and post-transcriptional mechanisms that regulate the genetic program in Zika virus -infected macrophages, International Journal of Biochemistry and Cell Biology 153 (2022) 106312. https://doi.org/10.1016/j.biocel.2022.106312. [76] F. Seif, M. Khoshmirsafa, H. Aazami, M. Mohsenzadegan, G. Sedighi, M. Bahar, The role of JAK- STAT signaling pathway and its regulators in the fate of T helper cells, Cell Communication and Signaling 15 (2017) 23. https://doi.org/10.1186/s12964-017-0177-y. [77] T. Tamiya, I. Kashiwagi, R. Takahashi, H. Yasukawa, A. Yoshimura, Suppressors of cytokine signaling (SOCS) proteins and JAK/STA T pathways: Regulation of T -cell inflammation by SOCS1 and SOCS3, Arterioscler Thromb Vasc Biol 31 (2011) 980 –985. https://doi.org/10.1161/ATVBAHA.110.207464. [78] T. Wang, B. Gorgoglione, T. Maehr, J.W. Holland, J.L.G. Vecino, S. Wadsworth, C.J. Secombes, Fish Suppressors of Cytokine Signaling (SOCS): Gene Discovery, Modulation of Expression and Function, J Signal Transduct 2011 (2011) 1–20. https://doi.org/10.1155/2011/905813. [79] Y . Zhang, M. Li, L. Li, G. Qian, Y . Wang, Z. Chen, J. Liu, C. Fang, F. Huang, D. Guo, Q. Zou, Y . Chu, D. Yan, β-arrestin 2 as an activator of cGAS -STING signaling and target of viral immune evasion, Nat Commun 11 (2020) 6000. https://doi.org/10.1038/s41467-020-19849-9. [80] W. Zhang, Y . Wang, A. Cheng, R. Liu, F. Kang, J. Zhao, J. Shao, X. Huang, X. Chen, Protective effects of dietary Astragalus polysaccharides on large yellow croaker ( Larimichthys crocea ) against Vibrio alginolyticus infection, Aquaculture 581 (2024) 7403 98. https://doi.org/10.1016/j.aquaculture.2023.740398. preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint [81] Y .L. Su, J.P. Chen, Z.Q. Mo, J.Y . Zheng, S.Y . Lv, P.H. Li, Y .S. Wei, Y .L. Liang, S.W. Wang, M. Yang, X.M. Dan, X.H. Huang, Y .H. Huang, Q.W. Qin, H.Y . Sun, A novel MKK gene (EcMKK6) in Epinephelus coioides: Identification, characterization and its response to Vibrio alginolyticus and SGIV infection, Fish Shellfish Immunol 92 (2019) 500 –507. https://doi.org/10.1016/j.fsi.2019.06.043. [82] X.D. Yang, Z.S. Hou, M.Q. Liu, C. Zeng, H.K. Zhao, Y .R. Xin, K.W. Xiang, Q. Yang, H.S. Wen, J.F. Li, Identification and characterization of mkk genes and their expression profiles in rainbow trout (Oncorhynchus mykiss) symptomatically or asymptomatically infected with Vibrio anguillarum, Fish Shellfish Immunol 121 (2022) 1–11. https://doi.org/10.1016/j.fsi.2021.12.049. preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.30.702270doi: bioRxiv preprint