Sample collection and study design

A total of 108 laying hens of the Lohmann Brown Plus genotype (LB, N = 108) was used for the infection experiment with A. galli and H. gallinarum. The laying hens used in this work originated from a previous study [3], where we evaluated the tolerance and resistance of laying hens of different genotypes to nematode infections. We particularly selected this high-performing genotype, as it is more sensitive to the effects of nematode infections. The hens were 24 weeks of age at the start of the experiment as described below. The total number of uninfected control hens was 42 while 66 hens were infected. The pens of infected and non-infected control hens were kept separately in two different rooms to prevent cross-contamination. In each room the LB hens were kept in 3 separate pens in a stocking density of maximum 6 hens per m2. Each hen was given a wing-tag to enable repeated measurements on the same individuals over time. The hens were fed a commercial diet (ad libitum), containing 11.2 MJ metabolizable energy, 170 g crude protein and 3.6 g calcium per kg feed [3]. The climatic conditions were optimally regulated using an automatic system to ensure similar lighting, temperature, and aeration across the pens within and between the rooms.

The infection experiment lasted for 18 weeks post infection (wpi), and randomly selected hens from each infection group were necropsied at wpi 2, 4, 6, 10, 14, and 18. Two weeks prior to necropsy hens were randomly selected from their pens (i.e., all pens were sampled with at least one hen), and transferred to individual cages (W 40 × L  45 × H 50 cm). The cages provided equipment for ad libitum water and feed intake of the hens. In each wpi, 11 infected and 7 control birds were necropsied to assess the worm burden as a direct measure of infection intensity. The hens were killed after 3-h feed withdrawal by stunning using a bolt shoot followed by bleeding to death.

Immediately after bleeding to death, blood and liver samples were collected from each bird. Blood was collected in potassium-EDTA treated tubes (Kabe Labortechnik GmbH, Nümbrecht-Elsenroth, Germany) and centrifuged for 20 min at 2500×g. The resulting supernatant was stored at − 20 °C for later analysis. The livers were macroscopically examined for typical signs of histomonosis [48]. The liver samples collected from the larger lobe (i.e., right) were snap frozen and stored at − 80 °C until use.

Experimental infection procedures and diagnosis of infections

The ethics committee for animal experimentation from the Mecklenburg-Western Pomerania State Office for Agriculture, Food Safety and Fishery, Germany, gave approval for the experiment (Permission number AZ.: 7221.3-1-080/16). The experimental procedure for infections followed the guidelines provided by the World Association for the Advancement of Veterinary Parasitology for poultry [49]. Animal handling, care, and necropsies were done by trained and authorised staff members according to the animal welfare rules. Infection material for the experiment was collected from female worms residing in intestines of free-range hens that were naturally infected with ascarids according to the procedure detailed in [50]. At 24 weeks of age, each bird was orally administered 0.4 mL containing a total of 1000 embryonated eggs of two nematodes (A. galli and H. gallinarum) using a 5-cm oesophageal cannula. The control hens were given a placebo containing 0.4 ml of 0.9% NaCl.

Determination of worm burden

Worm burdens were quantified from the hens that were necropsied at wpi 2, 4, 6, 10, 14, and 18. The hens were denied access to feed for 3 h prior to necropsies to partly empty the gastrointestinal tract (GIT). The GIT was removed immediately post-mortem, and the small intestine and caecum, the predilection sites of A. galli and H. gallinarum, respectively, were separated. The procedure for opening the intestine has been given in details in [3]. The total number of each of the A. galli and H. gallinarum worms present in the small intestine and caeca were recorded separately for the respective locations.

Measurements of plasma anti—Histomonas antibody and alpha (1)-acid glycoprotein

Because H. gallinarum is a natural vector for transmission of H. meleagridis [45] concomitant histomonosis infections cannot be excluded when the birds are infected with H. gallinarum. This also applies to experimental H. gallinarum infections, unless heterakis-infected birds are treated against histomonosis [51]. Thus, antibody titres against H. meleagridis were measured using an Enzyme Linked Immunosorbent Assay (ELISA) (52) to elucidate whether H. meleagridis was involved in the mixed infection. Briefly, ELISA plates were coated with rabbit anti-Histomonas serum at 1:10,000 dilution in carbonate buffer. The plates were treated with blocking buffer after an overnight incubation at 4 °C and a previous wash with PBS-Tween 20 (0.05 per cent PBST). Prior to the next washing step, diluted H. meleagridis antigen was added to each well and left for 1 h at room temperature. The plasma samples were then diluted (1:500) with blocking buffer and incubated for another 1 h at room temperature. Each plate included positive and negative control sera obtained from chickens infected experimentally with H. meleagridis. Goat anti-chicken IgG-horseradish peroxidase (Southern Biotech, Birmingham, AL, USA) was added for 1 h before another wash. A tetramethylbenzidine substrate solution (TMB; Calbiochem, Merck, Vienna, Austria) was used for 15 min in the dark. The optical density was measured at a wavelength of 450 nm. A predetermined the cut-off OD value of 0.54 nm suggested by Windisch & Hess, 2009 [52] was applied to differentiate birds tested for negative and positive histomonosis.

As H. meleagridis can induce damage on the caecal and liver tissues, an acute-phase protein, alpha (1)-acid glycoprotein (AGP) was measured in plasma samples using a commercial ELISA according to the manufacturer’s protocol (Life Diagnostics, West Chester, USA, Catalogue number: LD-AGP-5).

Sample preparation for 1H-NMR spectroscopy

The plasma samples were thawed on ice and filtered. Prior to filtering, empty filter tubes were initially rinsed with 500 µL of distilled water and centrifuged 3 times for (10,000 × 24 °C × 10 min). An aliquot of 500 µL each of plasma samples was transferred to filter tubes and centrifuged (14,000 × 4 °C × 120 min). For each filtered sample, 350 µL of phosphate buffer (deuterium oxide phosphate buffer 0.10 M, pH = 7.4 containing 3-(trimethylsilyl)-propionic-2,2,3,3-d4 acid, sodium salt (TSP d4) 0.04% and 0.04% of sodium azide) was added to an Eppendorf tube and 350 µL of filtered plasma was added. The tubes were mixed for 2 min at 350 rpm at a table mixer and 600 µL of the solution was transferred to a 5 mm NMR tube.

The frozen liver samples were extracted prior to NMR analysis. The samples were first lyophilized, and 20 mg of lyophilized tissue was weighed, 300 µL of ice-cold methanol (MeOH) was added to samples and whirl mixed thoroughly. Samples were placed on ice for 10 min. A 300 µL of ice-cold water was then added to samples, mixed thoroughly, and placed in 4-degree refrigerator overnight for separation.

Samples were then centrifuged for 30 min × 1400 at 4 °C. The upper MeOH phase was transferred to a new Eppendorf tube and dried for approximately 3 h. Extracted samples were re-dissolved in 575 µL of phosphate buffer (deuterium oxide phosphate buffer 0.10 M, pH = 7.4) and 25 µL of D2O with 3-(trimethylsilyl)-propionic-2,2,3,3-d4 acid, sodium salt (TSP d4) 0.05% were added, 550 µL was transferred to NMR tube containing.

1H-NMR spectrum acquisition

NMR spectroscopy was conducted at 310 K on a 14 T Bruker Avance III spectrometer (Bruker BioSpin, Rheinstetten, Germany) equipped with a 5 mm TXI probe head with gradients, automated tuning, and matching accessory (ATMATM), BCU-I for the regulation of temperature, and SampleJet robot cooling system set to 5 °C as a sample changer. The 1H NMR spectra were acquired using NOESY pre-saturation pulse sequence (Bruker 1D noesygppr1d pulse sequence), 64 K data points, spectral width of 20 ppm, acquisition time of 2.75 s, a recycle delay of 4 s, a relaxation delay of 5*T1 (19 s for blood, 26 s for liver), and 64 scans. 2D NMR experiments (JRES, 1H-13C HSQC, and 1H-1H COSY) were performed on selected samples.

The free induction decays (FIDs) were multiplied by a 0.3 Hz exponential function prior to Fourier transform. Phase and baseline corrections were carried out, and the reference standard Trimethylsilylpropanoic acid (TMSP-d4) signal was adjusted to δ 0.00. The 1D spectra were assigned using Chenomix database values, and the 2D NMR spectra.

For quantification of individual metabolites, processed spectra were imported into the Chenomx NMRSuite VX software and metabolite peaks were quantified relative to the area of TMSP-d4 signal. The Chenomx library considers metabolite information such as number of protons and molecular weight. The processing method was set with pH 7.4 and with TMSP-d4 concentrations of 0.41 mM in plasma samples, and 0.1 mM in liver samples.

Statistical analysis

Statistical analyses of AGP and histomonas antibody titers data were based on log transformation to correct for heterogeneity of variance and to produce approximately normally distributed data. Transformation was done using a natural logarithm function [Ln (y) = ln (y + 1)]. The data were analysed with one-way ANOVA using the R software version 4.1.2 [53].

For the metabolomics data, statistical significance was determined using student t-test with FDR adjustment. Data from all wpi were either pooled or analysed separately within each wpi to examine the effects of infections on metabolite concentrations. One sample with outlier metabolite results was removed from the plasma metabolite data. Differences were considered significant when p < 0.05 and a tendency for significant difference was declared when 0.05 < p ≤ 0.10.

MetaboAnalyst software (http://www.metaboanalyst.ca) was employed for univariate, hierarchical clustering and pathway enrichment analysis of the metabolomics data, and analyses were conducted separately for plasma and liver data, respectively. A hierarchical clustering heatmap showing group averages with Euclidean distance measures and ward linkage was constructed to explore the patterns among the metabolites between the two groups within each wpi. Pareto-scaling normalisation (i.e. normalised based on mean-centered and divided by the square root of the standard deviation of each variable) was applied to metabolite data before analysis. For pathway analysis, data were cross-listed with the pathways in the Gallus gallus Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway library. A global test was selected for the enrichment analysis method, while topology analysis was based on relative-betweenness centrality and scatter plot (testing significant features) was chosen for visualisation method.