A single colony of A. salmonicida J223 (17) was grown routinely in 3 ml of Trypticase Soy Broth (TSB, Difco, Franklin Lakes, NJ) at 15°C in a 16 mm diameter glass tube and placed in a roller for 24 h. After growth, 300 μl of the overnight culture were added in 30 ml of TSB media using a 250 ml flask and incubated for 24 h at 15°C with aeration (180 rpm). The bacterial growth was monitored spectrophotometrically until O.D. 600 nm ~0.7 (1 × 10⁸ CFU ml⁻¹) using the Genesys 10 UV spectrophotometer (Thermo Spectronic, Thermo Fischer Scientific Inc., Waltham, MA, USA). Then the bacterial culture was centrifuged at 6,000 rpm at room temperature for 10 min. The pellet was washed twice with PBS and centrifuged at 6,000 rpm at room temperature for 5 min, and finally resuspended in 300 μl of PBS (~5 × 10¹⁰ CFU ml⁻¹). The concentrated bacterial inoculum was serial diluted and quantified by plating onto TSA supplemented with Congo red (50 μg ml⁻¹).

A. salmonicida J223 strain was grown in TSB media supplemented with 100 μM 2, 2′-dypyridyl at 15°C with aeration (180 rpm) up to an optical density of O.D. 600 nm ~0.7 (~1 × 10⁸ CFU ml⁻¹). The bacterial cells were washed three times with PBS and then fixed with 6% formalin for 3 days at room temperature with gentile agitation. Formalin killed cells were dialyzed (Molecular Weight cut off 3.5 kDa; Spectra/Por, Laguna Hills, CA) in PBS three times and stored at 4°C at the concentration of 6 × 10¹⁰ CFU ml⁻¹ until utilization.

Adult specimens of Atlantic cod 1.5 ± 0.2 kg (mean ± SE) were obtained from the Dr. Joe Brown Aquatic Research Building (JBARB) at the Department of Ocean Sciences, Memorial University of Newfoundland, Canada. The animals were kept in 21 m3 tanks, with flow-through (75.l × min⁻¹) of sea water (6°C) and ambient photoperiod. The individuals were fed with commercial dry pellets (Skretting: 50% protein, 18% fat, 1.5% carbohydrate, 3% calcium, 1.4% phosphorus) with a ratio of 1% of body weight 3 days per week. The experiment was performed in accordance with the guidelines of the Canadian Council on Animal Care and approved by Memorial University of Newfoundland's Institutional Animal Care Committee (protocols #17-01-JS; #17-02-JS).

Primary macrophages were isolated from Atlantic cod head kidney in accordance to the protocol established by Eslamloo et al. (28) with modifications. Briefly, head kidney tissues from six fish were aseptically removed and individually minced through 100 μm nylon sterile cell strainers (Fisher Scientific, Thermo Fisher Scientific, Waltham, MA, USA) in isolation media [(Leibovitz-15 (Gibco^®, Gran Island, NY, USA) supplemented with 2 mM L-glutamine, 4.2 mM NaHCO₃, 25 mM HEPES, 1.8 mM glucose, 20 U ml⁻¹ heparin, 100 U ml⁻¹ penicillin, 100 μg ml⁻¹ streptomycin, and 1% Fetal Bovine Serum (FBS)]. After this period, 3 ml of cell suspension were centrifuged (400 × g at 4°C) for 40 min in a 25/51% Percoll gradient (GE Healthcare, Uppsala, Sweden). Macrophages collected from the macrophage-enriched interface were washed with phosphate buffered saline [PBS; 136 mM NaCl, 2.7 mM KCl, 10.1 mM Na₂HPO₄, 1.5 mM KH₂PO₄ (pH 7.2)] (29) twice and the number and viable cells were determined using the Countness™ cell counter (Invitrogen), and trypan blue stain (Invitrogen). After determining the numbers of cells from each sample, the primary macrophages were seeded in 22 mm 12-well or 35 mm 6-well cell-culture multidishes (Thermo Scientific, Roskilde, Denmark) at a concentration of 1 × 10⁷ cells ml⁻¹. The plates were incubated at 15°C for 24 h in isolation media. After this period the cells were washed with PBS and incubated at 15°C for additional 24 h in 1 ml of culture media [Leibovitz-15 (Gibco^®), supplemented with 2 mM L-glutamine, 4.2 mM NaHCO₃, 25 mM HEPES, 1.8 mM glucose, 100 U ml⁻¹ penicillin, 100 μg ml⁻¹ streptomycin, and 1% FBS] to allow cell attachment until the infection assay.

Infections of primary macrophages with A. salmonicida were performed according to a previously established protocol (17). To remove the antibiotic present in the culture media, the isolated primary macrophages were washed once with 1 ml of PBS, and then inoculated with 1 ml of cultured media without antibiotics. After this, the primary macrophage monolayers were infected with 10 μL of bacterial suspension [~1 × 10⁶ cells ml⁻¹; Multiplicity of Infection (MOI) 1:1 (bacteria:macrophage)] and incubated at 15°C. After 1 h post infection, the A. salmonicida attached to the Atlantic cod macrophages were quantified. The infected primary macrophage monolayers were washed 3 times with PBS, and lysed with 400 μl of Triton X100 (0.01%; Sigma) during 10 min (30) and then 600 μl of PBS were added to complete 1 ml of lysed macrophage suspension. Then the lysed macrophage suspensions were serially diluted (1:10) and plate/counted on TSA plates supplemented with Congo Red to determine the number of viable A. salmonicida per monolayer. The plates were incubated at 15°C for 5 days to determine the CFU per well. In addition, samples were taken for cell viability, RNA extraction, and transmission electron microscopy (see below for details).

For the invasion assay, cell monolayers were infected for 1 h, washed 3 times with PBS, followed by the addition of 1 ml of fresh culture media supplemented with gentamicin (10 μg ml⁻¹, a higher concentration than the minimal inhibitory concentration for A. salmonicida) (31), and incubated at 15°C. Samples were taken at 2, 3, and 6 h post infection for bacterial count, cell viability, RNA extraction, and transmission electron microscopy. All the macrophages were isolated from three individual fish and triplicates were utilized for each treatment in the assays.

To determine the viability of infected primary macrophages, the cells were seeded in 12 wells plates, infected with A. salmonicida, and processed as described in the gentamicin exclusion assay. For each time point post A. salmonicida infection, the cells were washed with 1 ml of PBS and then treated with 500 μl of trypsin-EDTA (0.5%; Gibco) for 10 to 15 min. After this period, the trypsin was inactivated with 500 μl of culture media. The cells were stained with trypan blue (0.4%; Invitrogen) in a ratio of 1:1 (10 μl: 10 μl) and quantified using Countess™ Cell Counting Chamber Slides (Invitrogen) and Countess^® Automated Cell Counter (Invitrogen) according to the manufacturer's instructions. The numbers of alive and dead cells were determined at each time point post- infection. All the macrophages were isolated from three individual fish and technical triplicates were utilized in the assays.

To determine the effect of A. salmonicida on the innate immune response of Atlantic cod primary macrophages, samples of RNA were isolated from infected cells at 1, 2, and 6 h post A. salmonicida infection, using the previously described gentamycin exclusion methodology. Primary macrophages that were either mock infected with PBS or inoculated with 1 × 10⁶ CFU of formalin killed A. salmonicida were utilized as controls. Total RNA was extracted using TRIzol (Invitrogen), and purified using RNeasy (QIAGEN) following manufacturers' instruction (32). RNA samples were treated with TURBO DNA-free™ Kit (Invitrogen) for complete digestion of DNA and removal of remaining DNase and divalent cations, such as magnesium and calcium. Purified RNA samples were quantified and evaluated for purity using a Nano-quant spectrophotometer (Genway, UK), and evaluated for integrity by 1% agarose gel electrophoresis (29). cDNA was synthetized with the SuperScript™ III First-Strand Synthesis System (Invitrogen) using 500 ng of RNA per reaction and random hexamers according to the manufacturer's instructions.

Primer pair efficiencies were analyzed using a 20 ng μl⁻¹ pooled cDNA from each set of samples, which was serially diluted (dilutions starting with 1 (20 ng μl⁻¹), 1:3 (6.67 ng μl⁻¹), 1:9 (2.22 ng μl ⁻¹), 1:27 (0.74 ng μl⁻¹), 1:81 (0.25 ng μl⁻¹), 1:243 (0.08 ng μl⁻¹), 1:729 (0.03 ng μl⁻¹)). Primer pair efficiencies were calculated using the formula E = 10^(−1/slope) (33).

All qPCR reactions were done in a final volume of 20 μL, containing 10 μL of 1 × PowerUp-SYBR Master Mix (Applied BioSystems, Foster City, CA, USA), 1 μL (10 μM) of each primer, 6 μL of nuclease free water (Ambion), and 2 μL of cDNA. All samples were amplified and detected in a QuantStudio 3 (Applied BioSystems). The reaction mixtures were incubated for 2 min at 95°C, followed by 40 cycles of 1 s at 95°C, 30 s at 60°C, and finally 15 s at 95°C, 1 min at 60°C, and 15 s at 95°C. Initially, a total of five Atlantic cod genes were tested as reference gene candidates (EF-1α, ß-actin, Eif3, 18S, 60S). cDNA from a sub-set of samples (Supplementary Table 1) were utilized for evaluation of reference gene stability. The most stable gene for this set of individual samples was determined by using geNorm (M value 0.102) and BestKeeper (Value 0.101) (Supplementary Table 1). After this determination, the individual samples were analyzed. The mRNA gene expression was normalized to the Atlantic cod elongation factor 1 alpha (EF-1a) due to its stability across different treatments. Gene expression was determined using the comparative ^−ΔΔCt method (34).

The primers used in this study are listed in Table 1. In all cases, each qPCR was performed with triplicate samples and repeated with six independent fish.

Respiratory burst of primary cod macrophages infected with A. salmonicida was determined according to the protocol established by Smith et al. (41, 42) with some modifications. Briefly, isolated primary macrophages were infected with 10 μl of bacterial suspension (~1 × 10⁶ cells ml⁻¹; MOI 1:1) and incubated at 15°C for 48 h. Primary macrophages inoculated with PBS were utilized as negative control, and phorbol myristate acetate (PMA 1 mM; Sigma) dissolved in dimethyl sulfoxide (DMSO) was utilized as positive control.

After 1 h post infection, cells were washed and the culture media was replaced with respiratory burst assay buffer (Leibovitz L-15 media supplemented with 1% BSA and 1 mM CaCl₂). Then, 1 μl of dihydrorhodamine 123 (DHR, 5 mg/ml; Sigma) was diluted in 1 ml of PBS, and 50 μl of the dilution added to the macrophages for 15 min. Subsequently, 125 μl of PBS for negative control, or 125 μl of PMA (1 mM, final concentration 0.185 μM PMA) were added to the macrophages monolayers for 45 min to stimulate ROS production (43, 44). Finally, the macrophages were detached using 1 ml of trypsin-EDTA (0.5%; Gibco), washed with PBS, centrifuged for 5 min (500 × g at 4°C), and resuspended in fluorescence-activated cell sorting (FACS) buffer (1% FBS in PBS). Fluorescence was detected from 10,000 cells using a BD FACS Aria II flow cytometer (Becton Dickinson™) and analyzed using BD FACS Diva v7.0 software (BD Biosciences, San Jose, CA, USA). The PBS control cells were used to define the region of ROS negative cells, and based on this gating the FITC positive cells were identified. The mean fluorescence intensity and percentage of FITC-positive cells were determined for each condition. The experiments were conducted in macrophages isolated from six independent fish, and 10,000 events were measured for each sample.

Primary macrophages were fixed in anhydrous paraformaldehyde (4%; Electron Microscope Sciences, Hatfield, PA, USA) at 4°C until the samples were processed at the Electron Microscopy/Flow Cytometry Unit at Memorial University of Newfoundland. The cells were pelleted and resuspended in Karnovsky fixative for 20 min (45), washed in 0.1 M sodium cacodylate buffer pH 7.4 for 5 min, and post-fixed in 1% Osmium tetroxide during 15 min. After this, the fixed cells were dehydrated in increasing concentrations of ethanol and acetone followed by infiltration with EPON resin (Sigma). Cells were pelletized between incubation steps. Resin blocks were polymerized in BEEM capsules (Electron Microscope Sciences) overnight at 70°C and ultra-thin sections were cut with a diamond knife (Diatome, Hatfield, PA, USA). The ultra-thin sections were mounted on 300 copper mesh grids, stained with uranyl acetate and lead citrate, and examined in a Tecnai™ Spirit TMA with an accelerating voltage of 80 kV. Cells incubated for 3 h with PBS (control), J223 strain, and formalin-killed A. salmonicida were observed.

All data are shown as the mean ± standard error (SE). Assumptions of normality and homogeneity were tested for the detected variances. A Kruskal-Wallis nonparametric test was performed for gentamicin exclusion assay results. Gene expression and ROS data were analyzed using a repeated measures two-way ANOVA test, followed by Sidak multiple comparisons post hoc test to identify significant differences of each treatment in different times and between treatments in the same time point. Differences were considered significant at P < 0.05. All statistical analyses were performed using GraphPad Prism (GraphPad Software, La Jolla California USA, www.graphpad.com).

The viability of Atlantic cod primary macrophages infected with A. salmonicida was determined at 1, 2, 4, and 6 h post-infection. The results did not show significant differences between the time points post-infection in the number of live cells and percentage of viability. For instance, after 1, 2, 4, and 6 h post-infection 1.32 × 106 ± 8.02 × 105, 1.18 × 106 ± 7.13 × 105, 1.17 × 106 ± 6.64 × 105, and 7.77 × 105 ± 4.77 × 105 cells were quantified, respectively (Figures 1A,B). The percentage of viability during the infection process also did not show significant differences between time points post-infection. After 1, 2, 4, and 6 h the infected cells showed a viability of 89% ± 4.8, 87% ± 3.1, 80% ± 4.2%, and 92% ± 6%, respectively (Figure 1B).

Although the primary cod macrophage cells seemed to survive the A. salmonicida infection, the bacteria infected and invaded the cell monolayers. The Atlantic cod macrophages were infected with a MOI of 1:1 (bacteria: macrophage) with an initial inoculum of 9.6 × 106 CFU. After 1 h post-infection, 7.39% (6.8 × 105 CFU) was attached to the macrophage monolayer, and after 2, 4, and 6 h post-infection, 0.42% (3.87 × 104 CFU), 0.37% (3.42 × 104 CFU), and 0.21% (1.94 × 104 CFU) of A. salmonicida were intracellularly located, respectively (Figure 1C).

Quantitative real-time expression of selected genes related to Atlantic cod macrophage immunity was evaluated during A. salmonicida infection, and compared with macrophages inoculated with inactivated A. salmonicida (formalin-killed), and PBS inoculated controls. Significant increases in the expression of the pro-inflammatory cytokine interleukin 1β (IL-1β) gene were observed 1, 2, and 6 h post A. salmonicida infection compared to the time-matched PBS controls (Figure 2A). An up-regulation of the pro-inflammatory cytokine interleukin 8 (IL-8) gene, was also observed, nonetheless, this up-regulation occurred at 2 h, and 6 h post A. salmonicida infection compared to the PBS controls (Figure 2B). In contrast, the macrophages inoculated with inactivated A. salmonicida showed a higher expression of IL-1β at 2 and 6 h post-inoculation compared to their respective PBS controls (Figure 2A). IL-8 was up-regulated in cells treated with formalin-killed bacteria after 1 and 2 h post-inoculation (Figure 2B). At 6 h post-inoculation with the inactivated A. salmonicida, the expression of IL-8 in cod macrophages did not show differences compared to the PBS inoculated cells (Figure 2B). The expression of IL-1b was significantly up-regulated at 1 and 2 h after A. salmonicida infection compared with bacterin-exposed macrophages, whereas for IL-8 significant up-regulation in infected vs. bacterin-exposed macrophages was only seen at the 6 h time point (Figures 2A,B).

In contrast, the relative expression of the anti-inflammatory cytokine Interleukin 10 (IL-10) gene, was significantly down-regulated after 1 and 2 h post A. salmonicida infection compared to the non-infected control macrophages (Figure 2C). Cod macrophages inoculated with the inactivated pathogen also showed a significant down-regulation at 1, 2, and 6 h compared with PBS controls (Figure 2C).

Genes involved in antigen recognition and host defense showed different patterns of expression after macrophage exposure to the live or inactivated A. salmonicida. In the case of the major histocompatibility complex class I (MHC-I) gene, a significant down-regulation was observed after 6 h post live A. salmonicida infection and post-inoculation with the formalin-killed bacteria (Figure 2D).

A down-regulation of the relative expression of the Goose-type lysozyme (g-type lysozyme) gene was observed at 1 and 2 h post A. salmonicida infection, and at 1, 2, and 6 h post inoculation with the formalin-killed pathogen, compared to their respective controls (Figure 2E). At 6 h post-treatment, g-type lysozyme expression was significantly different in infected vs. inactivated pathogen exposed macrophages (Figure 2E).

In contrast, the bactericidal permeability-increasing protein/lipopolysaccharide-binding protein (BPI/LBP) gene, involved in the antimicrobial defense against Gram negative bacteria, showed a significant up-regulation at 2 and 6 h post-infection only in cells inoculated with the live bacteria compared with the control and formalin-killed inoculated treatments (Figure 2F).

The relative expression of AMPs encoding genes showed different expression patterns. Cathelicidin (CAMP) gene was down-regulated in macrophages infected with A. salmonicida and those inoculated with formalin-killed A. salmonicida at 1, 2, and 6 h post-infection compared with time-matched PBS controls (Figure 2G). Macrophages infected with A. salmonicida showed a significant down-regulation of CAMP 2 h post-infection compared to the cells inoculated with inactivated bacteria (Figure 2G).

Transferrin did not show variation in the level of expression of both treatments compared with the time-PBS matched controls during the assays, as well as, between macrophages inoculated with live and bacterin A. salmonicida (Figure 2H). Hepcidin (HAMP) expression was significantly up-regulated only at 6 h post-exposure to the formalin-killed A. salmonicida (Figure 2I). However, HAMP was significantly down-regulated at 2 h post-infection with A. salmonicida (Figure 2I).

Expression of genes involved in the synthesis of ROS was also evaluated (Figures 2J–N). Nuclear factor erythroid 2-related factor 2 (nrf2), a transcriptional factor that is translocated into the nucleus under oxidative stress and initiates transcription of antioxidative genes (40), did not show transcriptional variation in cells inoculated with formalin-killed A. salmonicida compared to their respective controls (Figure 2J). In contrast, macrophages infected with A. salmonicida showed a significant down-regulation in the expression of nrf2 at 2 and 6 h post-infection (Figure 2J). After 1 h of infection, nrf2 transcript was significantly lower expressed in pathogen-infected macrophages compared to bacterin-exposed macrophages.

The expression of NADPH oxidase 1 (nox1) gene, which encodes a membrane-bound pro-oxidant enzyme that catalyzes superoxide synthesis, was significantly down-regulated 1 h post-infection with A. salmonicida compared to the non-treated cells and the bacterin-exposed macrophages (Figure 2K). Furthermore, this gene was significantly lower expressed at 1 h compared with 2 and 6 h post-infection (Figure 2K).

The relative expression of the Mn superoxide dismutase (Mn-Sod) and the CuZn superoxide dismutase (Cu/Zn-Sod) genes, involved in the transformation of superoxide into H2O2 in the mitochondria and the cytosol, respectively, also was evaluated. The Mn-Sod was significantly up-regulated at 2 h post-inoculation with inactivated A. salmonicida, compared to the controls (Figure 2L). In contrast, macrophages infected with A. salmonicida showed a Mn-Sod down-regulation tendency, with a significant down-regulation at 2 and 6 h post-infection compared to the bacterin-exposed macrophages (Figure 2L).

Similar patterns were observed in Cu/Zn-Sod and catalase (cat) expression, where a higher expression was observed post-inoculation with the formalin-killed A. salmonicida and a down-regulation was observed in cells infected with A. salmonicida (Figures 2M,N). For instance, Cu/Zn-Sod show a significant down-regulation at 6 h post A. salmonicida infection compared to the PBS control inoculated cells, and 2 and 6 h post-infection compared to cells treated with the inactivated pathogen (Figure 2M).

The relative expression of cat, which encodes for the catalase enzyme that plays an important role in H₂O₂ detoxification (40), showed a down-regulation at 1, 2, and 6 h in macrophages infected with A. salmonicida compared to the control and the cells treated with the inactivated pathogen (Figure 2N). In contrast, macrophages exposed to inactivated A. salmonicida did not showed significant differential expression of cat, compared to their respective controls (Figure 2N).

ROS production was determined in Atlantic cod primary macrophages infected with A. salmonicida. Macrophages inoculated with PBS or PMA were utilized as negative and positive controls, respectively. ROS production was analyzed at 1 and 6 h post A. salmonicida infection. We did not observe significant differences in ROS production between treatments. Macrophages treated with PBS showed that 85.1% ± 6.1 and 70.4% ± 6.2 were producing ROS after 1 and 6 h, respectively. Macrophages treated with PMA showed that 75.8% ± 7.0 and 53.6% ± 12.4 were producing ROS after 1 and 6 h, respectively. Similarly, macrophages treated with PBS and infected with A. salmonicida showed that 74.5% ± 6.4 and 52.9% ± 11.9 of the cells were producing ROS after 1 and 6 h post infection, respectively. Additionally, macrophages treated with PMA and infected with A. salmonicida showed that 68.5% ± 9.5 and 53.1% ± 12.4 of the cells were producing ROS after 1 and 6 h post treatment, respectively (Figures 3A,B).

Cod primary macrophages infected with A. salmonicida were visualized 3 h post-infection using TEM. A group of non-infected cells were utilized as reference control (Figures 4a,b). These cells showed a rounded cell morphology, large nucleus, and evident presence of cell organelles (e.g., mitochondria, endoplasmic reticulum, endocytic vesicles) and pseudopodias (Figures 4a,b). In contrast, Atlantic cod macrophages infected with A. salmonicida showed poorly defined nuclei, membrane ruffling, and large vesicles containing A. salmonicida cells (average of 2–3 bacterial cells per macrophage, a maximum of 8 bacterial cells per macrophages, and 70–80% macrophages infected) (Figures 4c–e). Furthermore, secretion of A. salmonicida outer membrane vesicles (OMVs) was observed in intracellular bacterial cells (Figures 4f,g).

Macrophages inoculated with the formalin killed A. salmonicida showed a defined nucleus and a large number of secretion bodies within the cytoplasm (Figure 4h). Apoptotic-like bodies were observed in high numbers in the presence of extracellular A. salmonicida bacterin (Figure 4i).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.