Distinct Innate Immune Programs in Nile Tilapia Head Kidney During Infections with Streptococcus agalactiae, Escherichia coli and Vibrio harveyi

Abstract

Nile tilapia (Oreochromis niloticus) is a globally important aquaculture species. However, intensive farming conditions increase the risk of bacterial diseases. Despite the fact that a considerable number of transcriptomic studies have examined host responses to single bacterial infections, comparative analyses conducted within a unified experimental framework remain scarce, limiting the understanding of pathogen-specific defence mechanisms. In this study, tilapia were experimentally infected with Streptococcus agalactiae, Escherichia coli, or Vibrio harveyi via thoracic injection. Head kidney tissues were collected at 48 h post-infection for RNA sequencing. The identification of differentially expressed genes (DEGs) was conducted utilising the edgeR, and the assessment of functional enrichment was facilitated through the implementation of Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis. A comparative analysis was conducted between the bacterial infection groups and the control group. The results of this analysis revealed the identification of 2930, 3328, and 4850 DEGs were identified in the S. agalactiae, E. coli, and V. harveyi infection groups, respectively. Integrated transcriptomic analysis, combining KEGG enrichment and expression profiling of key genes, revealed distinct response patterns across pathogens. The S. agalactiae infection predominantly activated innate immune signaling pathways, including Toll-like receptor, NOD-like receptor, cytokine–cytokine receptor interaction, and NF-κB pathways. In contrast, E. coli infection induced extensive metabolic reprogramming, notably in purine and pyrimidine metabolism, carbon metabolism, and amino acid biosynthesis. Meanwhile, an infection caused by V. harveyi resulted in mucosal and lysosomal defence responses, as evidenced by an increase in lysosome, phagosome, extracellular matrix–receptor interaction, and cell adhesion molecule pathways. Collectively, this study suggests that the head kidney of Nile tilapia employs pathogen-specific defence strategies rather than a uniform antibacterial response, providing one of the first transcriptomic comparisons of distinct bacterial infections in this species. These findings provide fundamental data and theoretical insights for elucidating immune mechanisms in teleost fish and for developing targeted prevention and control strategies in aquaculture.

1. Introduction

Aquaculture plays a crucial role in global food security, and Nile tilapia is among the most widely farmed finfish species, valued for its rapid growth, environmental adaptability, and high market demand [1,2]. Global production of tilapia has expanded steadily over the past decades, making it a cornerstone of freshwater aquaculture, particularly in Asia, Africa, and Latin America [3,4]. However, the sustainability and profitability of tilapia farming are increasingly threatened by infectious diseases, with bacterial infections representing one of the most significant challenges [5,6,7]. These outbreaks have the potential to result in high mortality, growth retardation, and severe economic losses, underscoring the pressing need for a more profound comprehension of host–pathogen interactions and the development of more efficacious disease control strategies.

Nile tilapia is commonly affected by a range of bacterial pathogens that differ markedly in biology and clinical presentation [8,9]. The Streptococcus agalactiae (group B Streptococcus) is a Gram-positive pathogen repeatedly implicated in large-scale streptococcosis outbreaks in tilapia farms worldwide and is associated with high morbidity and mortality [10,11]; transcriptomic and epidemiological studies document strong inflammatory responses and substantial economic impact from S. agalactiae outbreaks [12,13]. The Escherichia coli, an opportunistic, environmental Gram-negative bacterium, has been identified in tilapia production systems and has been recovered from gills, skin and flesh of farmed fish [14,15]; its presence in aquaculture is of concern both as a potential opportunistic pathogen and as an indicator of faecal contamination and hygiene status within aquaculture systems [16]. The Vibrio harveyi is a widely distributed marine pathogen and a major cause of vibriosis in cultured fish, frequently implicated in outbreaks that produce a range of clinical signs (skin ulcers, gastroenteritis, septicaemia) and substantial economic losses in aquaculture [17,18], and it has also been isolated from diseased Nile tilapia, where it causes skin lesions, ulceration, and high mortality in farm settings [19,20]. Taken together, these three bacteria represent complementary infection contexts: (i) a frequently high-impact, host-sensitive Gram-positive pathogen (S. agalactiae), (ii) a mucosa-associated Vibrio with relevance for localized outbreaks (V. harveyi), and (iii) an opportunistic/environmental Gram-negative (E. coli). This combination renders them well-suited for a systematic comparative analysis of host transcriptional strategies.

The majority of transcriptomic investigations in the field of aquaculture immunology have focused on single-pathogen challenges, providing important insights but ultimately yielding pathogen-centered rather than comparative views of host defense mechanisms [21,22,23]. In Nile tilapia, separate transcriptomic studies have examined host responses to S. agalactiae infection, reporting altered expression of genes involved in inflammation, complement activation, and acute-phase responses in immune tissues such as the head kidney and spleen [24,25,26]. Similarly, V. harveyi challenge experiments in teleosts, including tilapia and groupers, have revealed modulation of pattern recognition receptors, cytokine networks, and apoptosis-related pathways [27,28,29]. Transcriptomic responses to E. coli in fish have been the subject of less extensive study, with available reports largely limited to innate immune activation in mucosal tissues under opportunistic infection scenarios [30,31]. Despite the proliferation of research in this area, the studies conducted thus far have been conducted independently and utilised heterogeneous infection models, tissue types and sequencing strategies. Consequently, it remains challenging to distinguish which transcriptomic changes are universally applicable to bacterial challenge and which are specific to a particular pathogen. It is evident that a number of pivotal questions have yet to be addressed. Of particular note, it remains to be ascertained which host responses are indicative of shared antimicrobial programmes, and which responses are more indicative of pathogen-specific strategies.

Consequently, direct comparative transcriptomic analyses across contrasting bacterial infections remain scarce, particularly under a unified experimental framework. In order to address this knowledge gap, the present study established a standardized challenge model in Nile tilapia and performed parallel RNA-seq profiling after infection with three biologically and epidemiologically relevant bacteria: S. agalactiae, E. coli and V. harveyi. The head kidney was selected as the target tissue for the present analysis because it is the central organ of hematopoiesis and systemic immune regulation in teleost fish. Compared with the spleen, which is more specialized in antigen trapping and the development of adaptive immunity, the head kidney plays a dominant role in early systemic innate defense and is therefore a suitable model tissue for dissecting pathogen-specific host response strategies [32,33,34]. By analyzing DEGs specific to each infection group and conducting KEGG enrichment analysis, we explored the distinction between shared versus pathogen-specific host responses and a common immune context. Overall, the findings of this work are essential for deciphering pathogen-specific host strategies and for informing targeted disease-management measures in aquaculture.

2. Materials and Methods

2.1. Experimental Fish: Collection, Maintenance and Acclimation

The Nile tilapia (Oreochromis niloticu) used in this study were obtained from the Niulanjiang Rare and Endangered Fish Breeding Base in Kunming, Yunnan, China. All fish (average weight 60 ± 10 g) were confirmed to be clinically healthy and free of external lesions prior to experimentation. After transport, fish were maintained in aerated fiberglass tanks (500–800 L) with a recirculating freshwater system at a stocking density below 10 g/L. The water temperature was maintained at 25 ± 1 °C, dissolved oxygen was kept between 7–8 mg/L, pH was maintained at 7.0–7.5, and a natural light–dark cycle was applied. During acclimation, fish were fed a commercial tilapia diet twice daily (09:00 and 17:00). Feed was administered in small portions added gradually, and feeding was terminated once all feed was consumed (typically within 1 min) to avoid overfeeding. Approximately one-third of the water volume was replaced every two days to maintain water quality. To minimize variation in physiological status prior to infection, fish were fasted for 24 h before bacterial challenge. All fish were acclimated under laboratory conditions for 15–30 days prior to experimental infection to minimize environmental stress. All experimental procedures involving fish were approved by the Animal Ethics Committee of the Kunming Institute of Zoology, Chinese Academy of Sciences (Approval No. IACUC-PA-2024-03-007).

2.2. Bacterial Strains, Culture and Storage

Three bacterial strains were used in this study, S. agalactiae, E. coli and V. harveyi, These strains were procured from Beijing Beina Chuanglian Biotechnology Co., Ltd. (Beijing, China). Upon arrival, all strains were streaked onto corresponding agar plates for activation and purity check before liquid culture. S. agalactiae was cultured in Brain Heart Infusion (BHI) broth, E. coli was cultured in Luria–Bertani (LB) broth. For V. harveyi, TSB medium supplemented with 2% NaCl was used to ensure normal growth and maintain marine bacterium’s physiological status, as salinity is required for its viability and infectivity. For routine culture, a single colony was inoculated into 50 mL of medium in a 250 mL Erlenmeyer flask and incubated at 37 °C, 120 rpm in a shaker incubator. Bacterial growth was monitored by measuring OD600, and working cultures in logarithmic phase were used for subsequent experiments.

For long-term preservation, glycerol stocks were prepared by mixing fresh logarithmic-phase cultures with sterile glycerol to a final concentration of 20% (v/v). Aliquots (1 mL) were transferred into sterile cryovials, properly labeled, snap-frozen in liquid nitrogen, and stored at −80 °C until further use. Before infection experiments, viable counts of each bacterium were determined by serial dilution and plate counting to calculate colony-forming units (CFU/mL).

2.3. Bacterial Challenge Experiments

After acclimation, healthy Nile tilapia were randomly divided into four groups: one control group and three infection groups (S. agalactiae, E. coli, and V. harveyi). Each group contained 30 fish and was maintained in independent 150 L tanks under identical conditions.

For bacterial challenge, each fish in the infection groups received a thoracic injection of 200 µL PBS containing 1 × 10⁷ CFU of live bacteria, while the control group was injected with the same volume of sterile PBS. To balance the need for a strong immune provocation with animal welfare, we referred to preliminary LD50 data for S. agalactiae. Accordingly, a dose of 1 × 10⁷ CFU (a non-lethal concentration) was administered to consistently trigger infection and immune activation without causing high mortality within the 48-h experimental window. To maintain comparability among infection models, the same challenge dose was used for E. coli and V. harveyi. After challenge, fish were maintained under the same rearing conditions as described above and monitored regularly.

At 48 h post-infection, five individual fish per group were sampled. Head kidney tissues were aseptically collected from each fish (no pooling was performed), immediately frozen in liquid nitrogen, and stored at −80 °C for RNA extraction. RNA-seq libraries were initially prepared for five biological replicates from each treatment; however, samples failing RNA integrity or sequencing quality control were excluded, resulting in 3 replicates for the Control group, 3 for the S. agalactiae group, 4 for the E. coli group, and 5 for the V. harveyi group for downstream transcriptomic analysis.

2.4. Infection Validation

To confirm successful infection, bacterial reisolation and 16S rRNA identification were performed. To isolate bacteria, head kidney tissue samples were aseptically processed by homogenization in sterile PBS. Subsequently, the homogenates were plated onto their respective culture media solidified with 15% agar and incubated overnight at 37 °C. For each fish, 3–5 colonies were randomly selected and subjected to colony PCR using universal 16S rRNA primers 27F (5′-AGAGTTTGATCCTGGCTCAG-3′) and 1492R (5′-GGTTACCTTGTTACGACTT-3′). Sequencing results were compared to NCBI GenBank to confirm bacterial identity. Colonies from the infection groups matched the inoculated bacteria, while control group samples yielded no bacterial growth under the same conditions.

2.5. Total RNA Extraction and Transcriptome Sequencing

Total RNA was extracted from head kidney tissues using TRIzol reagent (TaKaRa, Dalian, China) according to the manufacturer’s protocol. Briefly, frozen tissues were homogenized in 1 mL TRIzol using a high-speed tissue grinder and incubated at room temperature for 5 min. After centrifugation at 12,000× g for 5 min at 4 °C, the supernatant was transferred to a new tube, followed by the addition of 200 μL chloroform. Samples were vigorously shaken for 15 s, incubated for 5 min, and centrifuged at 12,000× g for 15 min at 4 °C. The upper aqueous phase containing RNA was carefully transferred to a new tube and mixed with an equal volume of pre-chilled isopropanol. After 10 min incubation, RNA was precipitated by centrifugation at 12,000× g for 10 min at 4 °C. The RNA pellet was washed with 1 mL of 75% ethanol, air-dried for 2–5 min, and dissolved in RNase-free water. RNA concentration and purity were measured using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Wilmington, DE, USA), and RNA integrity was assessed by agarose gel electrophoresis. High-quality RNA samples were stored at −80 °C until sequencing.

For RNA-seq, 1 μg of total RNA from each sample was used for library construction. rRNA was removed by poly(A)+ mRNA enrichment using Oligo(dT) magnetic beads (Thermo Fisher Scientific, Waltham, MA, USA). Poly(A)+ mRNA was purified using Oligo(dT) magnetic beads and fragmented at high temperature in the presence of divalent cations. First-strand and second-strand cDNA were synthesized using random hexamer primers. After end repair and A-tailing, sequencing adapters were ligated, and cDNA fragments were size-selected using DNA Clean Beads. Each library was enriched by PCR amplification with index primers (P5 and P7), validated, pooled, and sequenced on the Illumina NovaSeq 6000 platform using a 2 × 150 bp paired-end configuration.

Raw reads were processed using Cutadapt (v1.9.1) to remove adapters and low-quality sequences. Clean reads were aligned to the Nile tilapia reference genome (GCA_001858045.3, NCBI GenBank, https://ftp.ncbi.nlm.nih.gov/genomes/all/GCF/001/858/045/GCF_001858045.2_O_niloticus_UMD_NMBU/; accessed on 5 June 2025) using HISAT2 (v2.0.1) [35,36]. Only uniquely mapped reads were used for gene expression quantification and subsequent analysis.

2.6. Principal Component Analysis (PCA)

The PCA was performed using MATLAB R2023a to evaluate the overall similarity and variation in gene expression profiles among different treatment groups. Each point in the PCA plot corresponds to one biological replicate, and samples from the same group are marked in the same color.

2.7. Differential Expression Analysis and KEGG Enrichment Analysis

Gene expression levels were quantified as fragments per kilobase of transcript per million mapped reads (FPKM). Differentially expressed genes (DEGs) between each infected group and the control group were identified using the edgeR package 3.38.1 in R [37]. Raw count data were normalized using the trimmed mean of M-values (TMM) method, and a negative binomial generalized log-linear model was fitted to estimate gene-wise dispersions. Genes with |log2 fold change| ≥ 1 and adjusted p-value (padj) < 0.05 (Benjamini–Hochberg correction) were defined as significantly differentially expressed.

Functional enrichment analysis of DEGs was performed using Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis. KEGG enrichment was conducted using the clusterProfiler R package against the KEGG database (https://www.kegg.jp/, accessed on 5 June 2025) [38,39]. Enriched pathways with padj < 0.05 were considered statistically significant. The visualization of enrichment results was performed using ggplot2 in R [40].

3. Results

3.1. Overview of Transcriptomic Responses in Nile Tilapia Under Pathogenic Bacterial Infections

A total of 15 high-quality transcriptome libraries were retained for analysis, including Control (n = 3), S. agalactiae (n = 3), E. coli (n = 4), and V. harveyi (n = 5). All libraries were derived from individual fish (non-pooled samples), and samples with insufficient RNA integrity or low sequencing output were removed during library QC. Each sample generated between 39.06 and 57.91 million raw reads, from which 38.00 to 56.49 million clean reads were retained after quality control. The Q20 and Q30 values of the clean reads were consistently higher than 97% and 91%, respectively, indicating high sequencing quality. The GC content ranged from 45.90% to 57.82%, which is consistent with the genomic characteristics of teleost fish. Mapping analysis revealed that the proportion of mapped reads ranged from 90.53% to 97.06% and uniquely mapped reads from 61.09% to 87.32% for most samples, suggesting generally high alignment efficiency. In addition, the proportions of reads mapped to non-coding regions (including intronic and intergenic regions) were comparable among all groups, indicating no notable differences in genome annotation coverage or non-coding read distribution across treatments (Supplementary Table S1). The rRNA reads were removed during preprocessing, and no sample showed abnormal rRNA contamination, confirming overall high RNA integrity and library quality. However, a few samples (e.g., from the control and S. agalactiae groups) exhibited relatively lower mapping ratios, which may be attributed to the incomplete annotation of the tilapia genome, the presence of numerous unannotated transcripts, and the contribution of pathogen-derived transcripts following infection. Overall, the sequencing data were of good quality and suitable for subsequent differential expression and functional analyses.

A substantial number of differentially expressed genes (DEGs) were identified in the head kidney across all three bacterial infection groups (Figure 1). Specifically, S. agalactiae infection resulted in 2930 DEGs, with slightly more up-regulated than down-regulated genes. In the E. coli group, 3328 DEGs were detected, with comparable numbers of up- and down-regulated genes. In contrast, V. harveyi infection induced the largest transcriptional response, with 4850 DEGs and a marked increase in both up- and down-regulated genes. These findings indicate that all three pathogens triggered significant transcriptional alterations in the tilapia head kidney, with V. harveyi eliciting the most pronounced response.

Principal component analysis (PCA) was conducted to assess the global transcriptional differences among the control and three bacterial infection groups (Supplementary Figure S1). The first two principal components (PC1 and PC2) accounted for 27.36% and 15.43% of the total variance, respectively. Samples within each treatment group clustered closely together, indicating high reproducibility and consistent gene expression patterns within groups. In contrast, clear separations were observed among the four groups, reflecting distinct transcriptomic responses to the different bacterial challenges. Specifically, S. agalactiae-infected samples were separated along the positive axis of PC2, E. coli-infected samples along the negative axis of PC1, and V. harveyi-infected samples along the positive axis of PC1, whereas control samples were located near the coordinate origin. These findings suggest that each bacterial pathogen triggered a unique and characteristic transcriptional response in the head kidney of Nile tilapia.

3.2. Comparative Analysis of DEGs Among Pathogenic Bacterial Infection Groups

KEGG enrichment analysis of DEGs in the S. agalactiae infection group revealed significant enrichment in multiple pathways associated with immune responses and metabolic processes (Figure 2A). Notably, several key immune-related pathways, such as the Cytokine-cytokine receptor interaction, Toll-like receptor signaling pathway, and C-type lectin receptor signaling pathway, were enriched, suggesting their critical roles in initiating innate immunity against streptococcal infection. In addition, pathways related to Glycolysis/Gluconeogenesis, Carbon metabolism, Fructose and mannose metabolism, and Pyruvate metabolism were significantly enriched, indicating a substantial metabolic reprogramming during the infection process. Furthermore, signaling pathways such as MAPK, FoxO, p53, and processes like Necroptosis and Ferroptosis were also enriched, which may be involved in regulating stress responses, cell survival, and death. Collectively, these results suggest that S. agalactiae infection triggers a pronounced transcriptomic response in the tilapia head kidney, characterized by coordinated immune activation and metabolic adaptation.

KEGG enrichment analysis of DEGs in the E. coli infection group revealed significant enrichment in pathways predominantly related to metabolic regulation and xenobiotic detoxification (Figure 2B). Notably, broad metabolic pathways, along with specific ones such as Purine metabolism, Pyrimidine metabolism, and Carbon metabolism, were significantly enriched, underscoring a high demand for energy production and nucleotide biosynthesis during infection. Amino acid metabolism pathways, including Glycine, serine and threonine metabolism, were also prominent. Furthermore, a strong emphasis on detoxification processes was observed, evidenced by the enrichment of Glutathione metabolism, Metabolism of xenobiotics by cytochrome P450, and Drug metabolism-cytochrome P450. Additional enriched pathways like Sphingolipid metabolism, Porphyrin metabolism, and One-carbon pool by folate suggest alterations in specialized metabolic and signaling processes. Collectively, these results indicate that E. coli infection induces extensive transcriptomic remodeling in the tilapia head kidney, primarily characterized by a comprehensive reprogramming of core metabolic and detoxification pathways.

In the V. harveyi infection group, KEGG enrichment analysis revealed a distinct profile, with differentially expressed genes significantly enriched in a limited number of specific pathways (Figure 2C). The most significantly enriched pathway was Lysosome, which directly indicates the activation of intracellular degradation and pathogen clearance mechanisms. Glycosaminoglycan degradation was also enriched, suggesting alterations in the metabolism of structural components of the extracellular matrix. Furthermore, enrichment in the Spliceosome pathway points to potential widespread changes in RNA processing and post-transcriptional regulation in response to infection. Collectively, these results suggest that V. harveyi infection triggers a focused transcriptomic response in the tilapia head kidney, primarily involving lysosomal function, extracellular matrix remodelling, and RNA splicing.

The DEGs produced following infection by three species of bacteria were compared, in order to reveal that infection with different pathogens induced both distinct and shared transcriptional responses in the tilapia head kidney (Figure 3). Specifically, 1664, 1695, and 2922 unique DEGs were identified in the S. agalactiae, E. coli, and V. harveyi groups, respectively. A total of 337 DEGs (approximately 4%) were identified as common among all three infection groups, suggesting the presence of shared molecular mechanisms underlying tilapia responses to bacterial infection. Additionally, 317 DEGs were shared between S. agalactiae and E. coli, 612 between S. agalactiae and V. harveyi, and 979 between E. coli and V. harveyi. These results indicate that while each pathogen triggers distinct transcriptional responses, a subset of genes is commonly regulated across infections.

3.3. Shared Immune Responses Induced by Pathogenic Bacterial Infections

A total of 337 DEGs were identified as common to all three bacterial infection groups, annotated to 240 distinct KEGG pathways (Supplementary Table S2). KEGG enrichment analysis revealed that significantly enriched categories were primarily related to metabolic processes, suggesting broad host metabolic reprogramming during bacterial infection. Notably, within these annotated genes, 38 DEGs were distributed across 16 immune-related pathways (Figure 4 and Supplementary Table S2). The most represented pathways included antigen processing and presentation (11 genes), phagosome (9 genes), cytokine–cytokine receptor interaction (8 genes), leukocyte transendothelial migration (7 genes), and Toll-like receptor signaling pathway (5 genes). Furthermore, the analysis revealed the presence of additional pathways, including chemokine signaling, RIG-I-like receptor signaling, and B/T cell receptor signaling were also observed, albeit with fewer genes. Although these immune pathways were not statistically overrepresented in the enrichment analysis, the consistent involvement of key immune processes across all infection models highlights a conserved set of immune genes that may underpin the broad-spectrum defense strategies of Nile tilapia against bacterial pathogens.

3.4. Pathogen-Specific Transcriptomic Signatures and KEGG Pathway Enrichment

KEGG enrichment analysis was performed on the unique DEGs identified in each bacterial infection (Figure 5). In the S. agalactiae group (Figure 5A), 1664 specific DEGs were significantly enriched in multiple pathways. Compared to the full set of DEGs, the unique profile showed a sharper focus on core immune signaling such as the MAPK signaling pathway, cytokine-cytokine receptor interaction, NOD-like receptor signaling pathway, Toll-like receptor signaling pathway, and C-type lectin receptor signaling pathway, along with novel enrichments in cytoskeletal regulation, highlighting a refined anti-streptococcal defense strategy. In the E. coli group (Figure 5B), 1695 specific DEGs were predominantly enriched in broad metabolic pathways, including purine metabolism, carbon metabolism, amino acid biosynthesis, glycine/serine/threonine metabolism, pyrimidine metabolism, and tryptophan metabolism. The unique DEGs presented a more concentrated impact on fundamental metabolic and detoxification functions compared to the broader transcriptomic changes, indicating a strong, specialized bias toward metabolic reprogramming. By contrast, in the V. harveyi group (Figure 5C), 2922 specific DEGs showed enrichment in a more limited set of pathways. The unique response was distinctly streamlined, with a pronounced emphasis on cell adhesion and a specialized IgA-mediated immune response not as evident in the full DEG set, with the most significant categories including cell adhesion molecules (CAMs), lysosome, intestinal immune network for IgA production, and glycan degradation pathways. Collectively, these results suggest that the three bacterial infections induce distinct host responses. The pathogenic bacterium S. agalactiae infection has been observed to trigger broad changes in pathways and immune-related signalling responses. In contrast, E. coli primarily alters host metabolic regulation, while V. harveyi infection has been shown to affect cell adhesion, lysosome function and IgA-mediated mucosal immunity in a prominent and unique manner.

To further validate and illustrate the pathway-level differences revealed by KEGG enrichment (Figure 5), we examined the expression patterns of a selection of representative genes associated with metabolism, innate immunity, and mucosal/lysosomal function (

Table 1. Representative genes associated with metabolic reprogramming, immune activation, and mucosal/lysosomal pathways in the three infection groups.

Category	Gene ID	log2FC (S. agalactiae)	log2FC (E. coli)	log2FC (V. harveyi)	Note
Metabolism	pdk3	−4.75	−1.42	0.35	Pyruvate dehydrogenase kinase, energy metabolism
	ldhd	−0.52	−1.08	−0.41	Lactate dehydrogenase D, glycolysis
	guk1	−0.47	2.21	0.22	Purine metabolism
	nt5c3a	−0.74	1.30	−0.15	Nucleotide metabolism
	dck	−0.30	1.79	0.95	Nucleotide salvage
	nme6	−0.78	1.31	0.03	Nucleoside diphosphate kinase
	ak3	−2.11	1.13	−0.10	Adenylate kinase, purine metabolism
	chdh	−2.07	2.54	1.02	Choline metabolism
	shmt1	−1.09	1.08	0.98	Serine/glycine metabolism
	cyp1a	0.91	3.15	5.20	Cytochrome P450 1A, xenobiotics
Immunity	tlr5	−3.76	−1.58	−2.38	TLR5, innate immune sensing
	traf3	1.59	0.01	0.00	TNF receptor signaling
	traf2	1.42	0.56	−0.21	TNF receptor signaling
	nfkbia	1.12	−0.15	−0.01	NF-κB inhibitor
	irf7	−1.56	−1.42	−0.86	IFN regulatory factor
	myd88	1.12	−0.53	−0.36	MyD88 adaptor
	il-8	4.74	−0.66	0.62	Pro-inflammatory cytokine
	ripk1	−2.24	−0.59	0.14	Apoptosis/immune signaling
	nod1	−1.37	−1.04	−0.08	NOD-like receptor
	il6st	1.49	0.01	0.51	IL-6 signaling
	il1rap	−2.32	0.41	2.03	IL-1 receptor accessory
	il10	3.16	0.24	−0.35	Anti-inflammatory cytokine
	tnfaip3	2.28	−0.47	0.35	TNF-induced protein
	tnfsf13b	−3.15	0.59	1.92	TNF superfamily member
	tnfrsf1a	−2.71	−0.09	1.81	TNF receptor 1A
	tap1	3.07	1.78	1.58	Antigen processing
	psme2	1.17	0.29	−0.35	Proteasome activator
Mucosal/Lysosomal defense	itga9	0.43	0.28	2.06	Integrin α9, adhesion
	itgb1	0.32	0.76	1.30	Integrin β1, adhesion
	jam2	0.40	0.34	2.17	Junctional adhesion molecule
	cldn5	1.66	1.05	2.28	Tight junction protein
	cdh15	1.62	−0.72	−3.46	Cadherin-15, barrier
	lgmn	1.64	1.40	2.28	Legumain, lysosomal protease
	ids	−1.34	0.61	1.82	Iduronate 2-sulfatase, lysosomal enzyme
	ctsh	−0.33	0.64	1.21	Cathepsin H precursor
	tpp1	0.26	1.07	2.05	Lysosomal peptidase
	man2b1	0.20	0.62	1.14	Lysosomal α-mannosidase
	glb1	−0.84	0.01	1.04	β-galactosidase, lysosome
	lamp3	−1.56	−2.04	1.01	Lysosomal membrane glycoprotein
	galnt7	−1.39	−1.56	−1.70	Mucin O-glycosylation enzyme

Note: Log₂FC values indicate differential expression in each bacterial infection relative to the NC group. Genes with padj < 0.05 were considered statistically significant and are shown in bold. Gray shading highlights the predominant trends: metabolic reprogramming in E. coli, immune activation in S. agalactiae, and mucosal/lysosomal regulation in V. harveyi.

). At the gene level, in the S. agalactiae group, multiple widely recognized immune effectors and signaling modulators (e.g., TRAF2/3, MYD88, NFKBIA, IL8, IL10, TNFAIP3, TAP1, PSME2) were significantly regulated, supporting the immune activation; however, some pattern-recognition and signaling components (notably TLR5, NOD1, RIPK1, IRF7) were downregulated in this group, indicating a complex and possibly feedback-regulated immune response rather than uniform upregulation across every PRR or pathway node. The E. coli group showed pronounced regulation of metabolic enzymes and nucleotide metabolism components (e.g., GUK1, NT5C3A, DCK, NME6, CHDH, SHMT1, CYP1A), consistent with the pathway-level indication of metabolic reprogramming. In the V. harveyi group, genes related to cell adhesion and mucosal/lysosomal activity (e.g., ITGB1, JAM2, CLDN5, LGNM, CTSH, TPP1, LAMP3) were predominantly upregulated, in line with enrichment for CAMs, lysosomal function and IgA-related pathways. In conclusion, the gene-level data provide a valuable complement to the KEGG enrichment results by offering specific molecular examples. They corroborate the major trends (metabolic bias for E. coli, immune activation for S. agalactiae, mucosal/lysosomal responses for V. harveyi) while also revealing locus-specific complexity.

4. Discussion

Nile tilapia is one of the most widely cultured fish species and a cornerstone of global aquaculture production, making disease control in this kind of fish a priority for food security and industry resilience [41,42]. To enhance comprehension of the molecular mechanisms underlying Nile tilapia responses to different bacterial infections, we performed a comparative transcriptomic analysis of head kidney samples from Nile tilapia infected with S. agalactiae, E. coli, or V. harveyi. The results reveal three recurring host response modes: an immune-activation mode in response to S. agalactiae, a metabolic-remodeling mode associated with E. coli, and a barrier/lysosomal-focused mode for V. harveyi. These pathogen-specific patterns indicate that tilapia deploy tailored defense strategies rather than a uniform response, a principle that aligns with comparative studies showing host responses tuned to pathogen lifestyle and tissue tropism [21,43,44]. Although a uniform challenge dose (1 × 10⁷ CFU) was used to enable direct cross-pathogen comparison, we acknowledge this as a methodological consideration. The inherent differences in virulence among S. agalactiae, E. coli, and V. harveyi could be a contributing factor to the distinct transcriptomic profiles observed. Therefore, future studies should consider incorporating LD₅₀ alongside additional refined metrics (e.g., clinical symptom scoring, bacterial load dynamics, host tissue injury indices) and, where possible, employ dose–response designs to better dissect the relationship between pathogen virulence intensity and host transcriptional dynamics.

Upon S. agalactiae infection, the Nile tilapia head kidney transcriptome exhibited a pronounced activation of innate immune signaling pathways, a response consistent with previous transcriptomic studies in tilapia and other teleosts [13,24,45] that highlight the robust innate immune signature triggered by this pathogen. Key components involved in pathogen recognition, signal transduction, and effector responses were differentially regulated in this group, highlighting the complexity of the Nile tilapia’s immune orchestration against S. agalactiae. For instance, canonical signaling adapters and modulators such as MYD88, TRAF2/3, NFKBIA, and TNFAIP3 [46,47,48] were significantly upregulated, likely amplifying downstream NF-κB and MAPK cascades that mediate pro-inflammatory gene expression [49,50]. In parallel, cytokines such as IL-8 and IL-10 also exhibited strong transcriptional regulation, indicating that both the pro-inflammatory and regulatory arms of the innate response were activated [51,52]. Interestingly, not all pattern recognition receptors (PRRs) followed the same regulatory direction. Genes such as TLR5, NOD1, RIPK1, and IRF7 were downregulated, suggesting that the Nile tilapia may simultaneously engage negative feedback loops or regulatory checkpoints to prevent excessive inflammation [53,54]. This regulatory tuning is consistent with findings in teleost fish, where microRNAs (e.g., miR-148) have been shown to suppress MyD88 and NF-κB signaling during bacterial infection as a mechanism to restrain uncontrolled inflammation [55]. Thus, the mixed pattern of up-regulation and down-regulation among PRRs and downstream nodes is indicative of a finely balanced immune modulation, as opposed to a unidirectional activation.

In contrast, the E. coli response was characterised by alterations in metabolic pathways (purine/pyrimidine metabolism, carbon metabolism, amino-acid biosynthesis, one-carbon metabolism), supported by differential regulation of representative enzymes (GUK1, NT5C3A, DCK, AK3, SHMT1, CYP1A). Such transcriptional reprogramming likely reflects altered Nile tilapia demands for nucleotides, energy and redox balance during infection, and may represent either host compensatory responses (to support repair and immune function) or pathogen-driven metabolic manipulation. Such metabolic reprogramming in aquatic hosts has been observed in multiple studies [56,57,58]. For example, in iridovirus-infected fish, integrated transcriptome-metabolome analyses revealed perturbations in carbohydrate, nucleotide and lipid metabolism during viral infection [59]. In bacterial infection models, fish transcriptomic studies also report co-regulation of immunity and metabolism (e.g., turbot with Aeromonas) [60]. These observations are consistent with our own findings, underscoring the notion that metabolic adaptation is a pivotal dimension of Nile tilapia–pathogen interaction. In the future, integrating transcriptomic and metabolomic analyses, or performing targeted perturbation of key metabolic nodes, could help disentangle host-orchestrated responses from pathogen-driven effects.

The DEG profile of V. harveyi infection revealed a distinct signature, centered on mucosal/lysosomal axis functions rather than broad metabolic or inflammatory signaling. Key adhesion molecules and barrier components—such as ITGB1, JAM2, and CLDN5—were upregulated, suggesting remodeling or reinforcement of epithelial and endothelial interfaces [61,62,63]. Concurrently, lysosomal enzymes and proteins such as CTSH, LGMN, TPP1 and LAMP3 were significantly induced, indicating enhanced intracellular degradation and potential antigen presentation capacity [64,65]. This mucosal-oriented transcriptome response is mirrored in prior studies of Vibrio infections in fish. In grouper immunized with V. harveyi, KEGG enrichment also flagged antigen processing, lysosome, and intestinal immune network pathways among the top hits [66,67]. Additionally, transcriptome studies of fish vibriosis often report elevated mucosal/lysosomal genes as an early defense mechanism to limit pathogen penetration [68]. The convergence of external evidence and our findings collectively suggests that V. harveyi infection triggers a predominant mucosal-like defense strategy in the Nile tilapia. This response, characterized by remodeling of the epithelial/vascular interface, enhanced intracellular degradation and antigen processing, and mucin reorganization, appears to function in concert to limit bacterial adhesion, invasion, and long-term colonization, rather than relying solely on systemic immune activation. It is recommended that future studies employ targeted analyses of mucosal tissues and spatial localisation of lysosomal markers in order to provide a more robust substantiation of these inferences. It should be noted that NaCl concentration may influence V. harveyi virulence factor expression, and we did not compare different salinity conditions in this study. Future work evaluating salinity-dependent pathogenicity could further refine the infection model and mechanistic interpretation.

Although our study provides insight into pathogen-specific transcriptional responses in the tilapia head kidney, the analysis was based on a single sampling time (48 h post-infection). This time point was selected because all three pathogens elicited visible clinical symptoms without inducing overwhelming mortality, allowing a controlled comparison across treatments. Nonetheless, 48 h may not represent the peak immune activation or pathological severity for each bacterium, as their infection kinetics and virulence strategies differ. Future studies incorporating multiple time points, particularly covering early innate activation, mid-infection progression, and later resolution or tissue repair phases, will be essential to map the full temporal trajectory of immune responses and to determine species-specific peak response periods more precisely.

From a holistic perspective, our findings illustrate three distinct host defense strategies: an immune-activation mode under S. agalactiae, a metabolic-remodeling mode under E. coli, and a barrier/lysosomal defense mode under V. harveyi. The divergence in transcriptomic responses suggests that tilapia does not mount a “one-size-fits-all” defense, but rather tailors its response to pathogen type. In aquatic environments where fish continuously face diverse microbial challenges, such adaptive specificity may minimize collateral tissue damage while ensuring effective pathogen control [69]. These observations echo broader concepts in comparative immunology: pathogens often manipulate host metabolism or barrier interfaces to aid their survival, and successful hosts evolve countermeasures tuned to pathogen tactics [70,71,72]. For example, in mammalian contexts, metabolic reprogramming fuels immune cell activation (so-called immunometabolism) and dictates inflammatory potential [73,74,75]. Similarly, in teleosts, mucosal immunity is a frontline barrier in fish, particularly at epithelial surfaces like gills and gut, where adhesion molecules, mucus glycosylation, and lysosomal degradation pathways play crucial roles [76,77,78].

From a pragmatic standpoint, the identification of these discrete patterns can facilitate the development of bespoke interventions within the context of aquaculture. These interventions may encompass immunostimulatory or adjuvant approaches for pathogens that principally trigger PRR-mediated inflammation, metabolic support or modulation for pathogens associated with host metabolic disturbance, and mucosal-targeted vaccines or barrier-protective measures for pathogens that interact at epithelial surfaces. In aquaculture, combining immune, metabolic and barrier-targeted strategies may present a promising integrated approach to bolster disease resilience. To translate these mechanistic insights into potential applications, future work should focus on the functional validation of key regulatory nodes (e.g., MYD88, GUK1, CTSH) via techniques such as CRISPR/Cas9 knockout or overexpression. Furthermore, localized tissue analyses (e.g., gills, intestinal mucosa) and integrated multi-omics approaches (e.g., metabolomics, proteomics) will be essential to elucidate the causal interactions and identify promising therapeutic targets.

5. Conclusions

In summary, our transcriptomic dissection highlights that tilapia deploys pathogen-specific defense strategies when challenged by distinct bacterial species. S. agalactiae primarily elicits robust immune activation, E. coli drives metabolic reprogramming, and V. harveyi engages mucosal barrier remodeling and lysosomal responses. This diversity underscores how pathogen identity fundamentally shapes host defense logic, revealing not a uniform immune response but a spectrum of tailored strategies optimized for different pathogenic pressures. These insights not only deepen our mechanistic understanding of host–pathogen interactions in teleosts but also point toward integrated immune–metabolic–barrier interventions as promising avenues for disease prevention and control in aquaculture.