Animals and experimental design
All experimental procedures were carried out in accordance with the EU directive 2010/63 for animal experiments and submitted to and approved by the C2EA 84 Poitou-Charentes Region ethics committee (decision #CE2013–1, 2013.01.21).
A divergent genetic selection experiment was conducted in purebred Large White pigs by selecting, at each generation, founders with extreme cortisolemia following ACTH stimulation as an indicator of HPA axis activity. This allowed the generation of pigs low (HPAlo) and high (HPAhi) responders to ACTH injection [9]], which were all born and raised at the INRA experimental unit of Le Magneraud (UE1372 GenESI) until euthanasia by lethal electronarcosis.
This study involved blood samples collected from 32 healthy female pigs belonging to the third generation of divergent selection (16 HPAlo and 16 HPAhi pigs). In each population, the 16 pigs were offspring of 8 sows mated with 8 different barrows, constituting 8 pairs of full-sibs.
After weaning at 4 weeks, pigs were housed by groups of 20 to 22 individuals per pen in controlled ambient conditions (27 °C, 12 h/12 h light/dark cycle) with unrestricted access to food and water. Two experiments were conducted. At week 6, hematologic and immune parameters were analysed on blood samples collected from 9 to 10 a.m (Experiment 1). Two weeks later, a kinetic response to intra-muscular LPS injection (15 μg/kg in the neck muscles) was performed on the same animals with blood samples collected just before (t0, between 9 and 10 a.m.) and 1, 4 and 24 h after LPS administration as previously described [19] (Experiment 2).
In both experiments, blood was collected from the jugular vein of pigs slightly maintained in a supine position by experienced staff, in less than 30 s per pig as previously described, with a total volume collected never exceeding 25 ml per pig [23]. All blood samples were processed by blinded investigators.
Complete blood cell counts
Complete blood cell counts (CBC) were determined from EDTA-treated blood samples using the clinical-grade Procyte Dx haematology analyser (IDEXX, Saint-Denis, France). Red blood cell parameters included red blood cell count (RBC), haematocrit (HCT), haemoglobin (HGB), mean corpuscular volume (MCV), mean corpuscular haemoglobin concentration (MCHC), reticulocyte count (Ret) and immature reticulocyte fraction (IRF). CBC also included platelet count as well as white blood cell parameters: white blood cell count (WBC), neutrophil, lymphocyte and monocyte counts.
Analysis of lymphocyte subpopulations
Lymphocyte subpopulations were identified from EDTA-treated blood samples by two types of staining using monoclonal antibodies: (1) FITC-conjugated anti-pig CD4 (Acrys Antibodies, Herford, Germany), PE-conjugated anti-pig CD8α (BD Biosciences) and PerCP-conjugated anti-pig CD3 (Bio-Techne, Lille, France) antibodies, or (2) PE-conjugated anti-pig CD21 (Acrys Antibodies) and PerCP-conjugated anti-pig CD3 in the dark for 15 min at room temperature. Erythrocytes were removed using a lysis solution (BD Biosciences, Le Pont de Claix, France). The DNA marker DRAQ5™ (1 μM final; Biostatus, Leicestershire, UK) was added to identify nucleated cells, then samples were analysed with a FACS Aria flow cytometer (BD Biosciences) and data computed using the FlowJo software (FlowJo, Ashland, Oregon). Staining allowed the identification of seven different lymphocyte subpopulations as previously described [18]: B lymphocytes (CD3− CD21+), naive CD4+ Th cells (CD3+ CD4+), antigen-experienced Th cells (CD3+ CD4+ CD8α+), CD8α− γδ T cells (CD3+ CD4− CD8α−), CD8α+ γδ T cells (CD3+ CD4− CD8αmed), CD8αhi cytotoxic T lymphocytes (CD3+ CD4− CD8αhi) and NK cells (CD3− CD4− CD8αmed). The cell count for each specific subpopulation was then determined as the product of the percentage of this specific subpopulation in the DRAQ5+ FSC/SSC white blood cell gate by the absolute number of white blood cells obtained from the Procyte analyser.
Phagocytosis
Ex vivo phagocytosis was assessed using the Phagotest™ Kit (BD Biosciences) according to manufacturer’s instructions. Briefly, heparinized blood samples were incubated for 10 min at 37 °C with opsonized FITC-labelled E. coli bacteria. For each analysis, ice-incubated negative controls were included. Samples were immediately analysed by flow cytometry using FSC/SSC dot plot to discriminate granulocytes from mononuclear cells.
Whole blood assay (WBA)
Heparinized blood samples were five-fold diluted in RPMI 1640 medium supplemented with 2 mM L-glutamine, 100 IU/ml penicillin and 100 mg/ml streptomycin and stimulated in duplicate using 10 ng/ml Escherichia coli O111:B4 LPS (Sigma-Aldrich, St-Quentin-Fallavier, France). After an 18-h incubation at 37 °C, in 5% CO2, 95% relative humidity, supernatants were collected, centrifuged at 500 g for 5 min at 4 °C and stored at − 80 °C before cytokine concentration measurements. Porcine TNFα, IL-8 and IL-10 were quantified by ELISA with respective detection limits of 15, 62 and 12 pg/ml (Bio-Techne). For statistical analyses, we attributed to undetected samples half of the corresponding threshold value.
Kinetic response to LPS
Cortisol, glucose and free fatty acid plasma levels were assessed after LPS administration as previously described [23].
Plasma cytokine levels in response to LPS injection were quantified in heparin-treated samples by ELISA (porcine IL-6 and TNFα, Bio-Techne).
In order to investigate divergent selection effects on gene expression response to LPS, we tested expression of 34 genes modulated by LPS. These genes were provided by the literature and from our own data ([23] and unpublished data). Gene expression analysis was performed as previously described [23]. Total RNA extraction was done using the Nucleospin RNA Blood kit (Macherey-Nagel, Hoerdt, France) followed by DNase treatment. The quality of each RNA preparation was verified through the Bioanalyser Agilent 2100 (Agilent Technologies, Massy, France) and low-quality RNA samples were discarded (RIN < 8). One μg of total RNA was reverse-transcribed. The transferrin receptor and glutamyl-prolyl-tRNA synthetase genes were used as housekeeping genes, according to previous data. Pre-amplified samples were analysed with a 96 × 96 Dynamic Array™ IFC (Fluidigm) following the protocol defined by Spurgeon et al. [24]. Each gene was tested twice for each sample. Four dilution points containing a pool of all samples were used to determine PCR efficiency. Data were analysed using BioMark Gene Expression Data Analysis software (Fluidigm) to obtain Ct values. The Pfaffl method was applied to compute the relative expression of each gene [25]. A linear model was performed to reveal the genes differentially expressed according to the population (high, low) and the time point (0, 1, 4, 24 h):
$$ {expr}_i={\beta}_0+{\beta}_1t+{\beta}_2{population}_i+{\beta}_3{t}^{\ast }{population}_i+{\varepsilon}_i $$
Statistical analysis
A PCA was performed on raw data obtained on 6-week-old piglets using the FactoMineR package in the R software (version 3.2.1, R Core Team, 2016) [26]. Redundant parameters were excluded from the analysis, i.e. MCV, MCHC, % Ret, WBC and lymphocyte counts. The first two components explained more than 40% of the total inertia and each contributed to more than 15% of the total variance.
Univariate statistical analyses for experiment 1 were made using Graph Pad Prism (version 6.0). For each parameter measured, 16 HPAlo and 16 HPAhi pigs were compared using non-parametric Mann-Whitney tests; statistical significance appears as stars on charts (*: p < 0.05; **: p < 0.01; ***: p < 0.001); each point represents a single animal and lines denote medians.
For experiment 2, cortisol, TNFα, IL-6, white blood cells and platelets data were transformed into their logarithmic scores for normalisation. The square root transformation was used for NEFA (non-esterified fatty acids) levels [19]. For each parameter, the effect of HPA activity, time, and their interaction was tested in a two-way analysis of variance with repeated measures for the time factor. The effects of LPS on tympanic temperature, cortisolemia, plasma TNFα and IL-6 levels were graphed, each point representing a single pig.