With the aim of determining whether C. jejuni is able to invade primary chicken embryo intestinal cells (CEICs), intracellular bacterial counts were performed using the gentamicin protection method 8. For this purpose two C. jejuni strains, a Danish chicken isolate SC11 and a clinical human isolate HM5040, were investigated. Prior to the infection experiment, the motility and morphology of both isolates was checked under light microscopy. The results of invasion are shown in Figure 1. The number of intracellular bacteria for both SC11 and HM5040 were gradually increased from 1 to 4 h p.i. and declined to a lower level at 24 h p.i. The percentage of internalized SC11 overall was lower than that of HM5040 from 1 to 4 h p.i., with the exception of the 24 h p.i. where the number of internalized SC11 was higher than that of HM5040. A statistically significant difference between SC11 and HM5040 was observed only at 1 and 2 h p.i. (p ≤ 0.011).

When human cells are in contact with C. jejuni a strong cytokine response is induced 101114. In order to determine if a similar response can be monitored with chicken embryo intestinal cells (CEICs), the levels of IL-1β, IL-6, CXCLi1 (K60), CXCLi2 (CAF/IL-8) 19 and TGF-β4 transcripts were measured in CEICs at 0, 1, 2, 3, 4 and 24 h p.i. with C. jejuni strains SC11 and HM5040. Mock-inoculated cells and those inoculated with LPS (5 μg/ml) were used as controls throughout. Figure 2 shows that both bacterial strains induced the production of IL-1β, IL-6, CXCLi1 and CXCLi2, but not of TGF-β4. The level of IL-1β, IL-6 and CXCLi1 transcripts increased gradually from 1 to 4 h p.i. and declined slightly at 24 h p.i., whereas CXCLi2 transcript increased steadily throughout the observation period and reached a high level at 24 h p.i. Significant differences induced by SC11 and HM5040 were observed only at some particular time points: IL-1β at 24 h p.i. (p = 0.042), IL-6 at 4 p.i. (p = 0.012) and 24 h p.i. (p = 0.046) and CXCLi2 at 1 h p.i. (p = 0.025) and 24 h p.i. (p = 0.019) while IL-2 and IL-4 transcripts were undetectable. The transcript of IFN-γ was detected only at some particular time points (data not shown).

To study whether there are inflammatory responses in CEICs, the cellular iNOS transcripts and the presence of NO product in the medium were measured. The level of iNOS transcripts was altered by both strains and increased gradually from 1 to 4 h p.i., followed by a decline at 24 h p.i (Figure 2). The production of NO from the inoculated CEICs was determined at 4 and 24 h p.i. using the Griess assay 20. The results were presented as a net nitrite production over the background of the mock-inoculated CEICs (Figure 3). Nitric oxide was produced by inoculated CEICs, while as expected, the CEICs treated with LPS also showed NO induction.

To study whether the contact between C. jejuni and CEICs induce bacterial virulence-associated genes, the expression of bacterial virulence-associated genes ciaB, dnaJ and racR was quantitatively determined during co-cultivation with CEICs, relative to 16S rRNA, and the fold change of the gene expression was determined relative to the mock-incubated bacteria in the absence of CEICs. The results are shown in Figure 4. The transcription of the virulence-associcated genes was increased following co-cultivation with CEICs, and the amount of present in the invaded/adhered bacteria was generally higher than that present in the suspended bacteria, especially the transcripts of dnaJ and racR. ciaB transcription was increased at 1 and 2 h p.i. in SC11 and HM5040, respectively, and declined to a background level at 4 h p.i., followed by a slight increase at 24 h p.i. In the suspended bacteria, the expression of ciaB was changed slightly, increased in SC11 and decreased in HM5040. A significant difference between SC11 and HM5040 was observed at 24 h p.i., where SC11 had a higher fold change than HM5040 had (p = 0.041). The transcripts of dnaJ and racR in invaded/adhered bacteria were increased in both strains throughout the course of the experiment. A statistically significant difference was observed only for racR at 1 h p.i. (p = 0.020), the level of racR and dnaJ in SC11 was higher than that in HM5040 with the exception of dnaJ at 24 h p.i. In suspended bacteria, the expression of dnaJ and racR was altered slightly during the inoculation. The racR transcript in SC11 was significantly higher than in HM5040 at 24 h p.i. (p = 0.029).

To examine whether the induction of the virulence-associated genes is contact dependent, the expression of the selected virulence-associated genes was also measured in the bacteria that had only been incubated with bacteria- and cell-free supernatant prepared from another co-cultivation experiments. The bacteria incubated with the medium from non-inoculated CEICs and mock-incubated bacteria were used as controls. The level of expression of the virulence-associated genes in mock-incubated bacteria was defined as 1, to which the expression of the genes in the bacteria incubated with bacteria- and cell-free supernatant and cultured medium was calculated (Figure 5). After 4 hours of incubation, expression of the three virulence-associated genes ciaB, dnaJ and racR of SC11 was increased slightly as compared to the mock-incubated bacteria and those bacteria incubated with the medium of non-inoculated CEICs (p ≤ 0.170). The medium of non-inoculated CEICs decreased the expression of ciaB and dnaJ in SC11 (p ≤ 0.081), but did not alter the level of racR transcripts. For HM5040, expression of the three genes showed a slight increase in both bacteria incubated with the medium of non-inoculated CEICs and bacteria- and cell-free supernatant (p ≤ 0.30).

Two C. jejuni strains SC11, a Danish chicken isolate, and HM5040, a human isolate, were used in this study. These two strains were the most common serotype (Penner serotype 2) and flaA type (flaA type 1) among isolates from broilers and human cases in Denmark 39. Both strains were taken from -80°C stock and streaked on modified CCDA (mCCDA) (blood-free agar base with cefoperazone [32 mg/liter] and amphotericin B [10 mg/liter]) (CM739 plus SR155; Oxoid, Basingstoke, UK) agar plates and incubated under microaerobic condition (10% CO2, 2–4% H2 and 86%–88% N2) at 42°C for 48 h. The bacteria were then sub-cultured on a new mCCDA plate for another 24 h prior to experimentation. The bacteria were harvested and well suspended in room temperature PBS (10 mM) to an OD₅₉₅ = 1.0 (approximately 10⁹ bacteria/ml). The bacteria were then diluted in pre-warmed Dulbecco's modified Eagle's growth medium (DMEM) (Invitrogen, UK) without antibiotics to the desired density for inoculation of cell cultures at a multiplicity of infection (MOI) of 100 bacteria per cell.

Primary chicken embryo intestinal cells (CEICs) were prepared from 19-day-old specific pathogen free (SPF) chicken embryos (Lohmann Tierzucht GmbH, Cuxhaven, Germany). Briefly, SPF chicken embryos were dissected, and intestines were separated and cut into small pieces in PBS prior to being digested with 0.25% trypsin-EDTA (Invitrogen, UK) for 3 times, 5 min for each. The cells were then collected by centrifugation at 500 × g for 10 min at room temperature (~25°C). The cells were seeded in 12-well plates at a density of 3.5 × 10⁵ cells/ml (1 ml per well) in DMEM (Invitrogen) supplemented with 10% fetal bovine serum (FBS) (Invitrogen), 0.3% (w/v) tryptose phosphate broth (containing 0.13% pancreatic digest of casein, 0.07% proteose peptone No. 3, 0.02% dextrose, 0.05% sodium chloride, 0.026% disodium phosphate), and 0.02% penicillin and streptomycin each (Invitrogen). The cells were incubated in a humidified environment at 37°C with 5% CO2 throughout the experiment. At 24 h the medium was replaced with fresh growth medium, the cells were incubated for anther 72 h prior to bacterial inoculation experiments. The cells were washed three times with pre-warmed PBS, freshly prepared 300 μl of antibiotic-free medium containing the desired number of bacteria was added and co-cultivated with cells to various time points. All inoculations were carried out at a MOI of 100:1, unless stated otherwise. Controls consisted of mock infections using antibiotic-free medium alone or positive controls of lipopolysaccharide (LPS, 5 μg/ml) (Sigma, Schnelldorf, Germany).

To prepare bacteria- and cell-free cultured medium, cells were inoculated with the desired number of bacteria for 4 h (MOI = 100), the medium was collected, the bacteria was removed by centrifugation three times at 10,000 × g, 3 min for each, or filtration through a 0.22 μm filter. Three hundred μl of this medium was added to freshly harvested bacteria of ~1.7 × 10⁷ and incubated for 4 h. As controls, the bacteria incubated with fresh and CEICs-cultured medium were included in parallel. The incubated bacteria were collected by centrifugation and subject to bacterial total RNA isolation as described below.

The number of intracellular bacteria per eukaryotic cell culture was assessed by the gentamicin protection assay 8. At 1, 2, 3, 4 and 24 h post inoculation (p.i.), the culture medium was removed, Hank's Buffered Salt Solution (HBSS, Invitrogen) containing gentamicin (200 μg/ml) was added, and incubated at 37°C, 5% CO2 for 2 h. The cells were then washed three times with pre-warmed PBS and lysed with 500 μl of 0.2% (v/v) Triton X-100 at room temperature for 15 min. The bacteria counts were numerated by plates counting and were present as a percentage of invaded bacteria to total inoculated bacteria.

Total cellular and bacterial RNA was extracted using TRIZOL LS Reagent (Invitrogen) and re-purified using RNeasy Mini RNA isolation kit (Qiagen, Germany) according to the manufacturers's instructions. At 0, 1, 2, 3, 4 and 24 h p.i., both the medium of the inoculation and the suspended bacteria were collected by centrifugation at 10,000 × g for 3 min. The bacteria pellets and cells were lysed in 300 μl of TRIZOL LS reagent and proceed to RNA isolation. RNA was eluted into 50 μl of RNase-free water and treated with 0.1 U/ml Dnase I Amplification Grade (Invitrogen) according to the manufacturer's protocol. The treated RNA was further tested for DNA contamination by PCR using the primer pairs of β-actin and 16S rRNA (Table 1). DNA-free RNAs were transcribed to complementary DNA (cDNA) using the iScript™ cDNA Synthesis Kit (Bio-Rad, USA) with pre-blended RNase inhibitor, oligo (dT) and random hexamer primers, according to the manufacturer's instructions. Quantitative real time PCR (qPCR) was performed in the Mx3005P thermocycler (Strategene, Denmark) using primers listed in Table 1. PCR mixtures (25 μl) contained 0.625 units of Taq DNA Polymerase (Promega, Denmark), 5 mM MgCl2, 200 μM dNTPs, 400 nM of each primer, 10 nM fluorescein (Bio-Rad), 50000 × diluted SYBR green (Invitrogen), and 2 μl of diluted cDNA. The thermal cycling conditions included an initial heat-denaturing step at 94°C for 3 min; 45 cycles of 94°C for 15 s, optimal annealing temperature for appropriate primer pair (Table 1) for 20 s and 72°C for 20 s; followed by an elongation step at 72°C for 3 min. A single fluorescence measurement was taken after each thermo cycle. The specificity of amplifications was checked by melting curve analyses and 2% agarose gel. All primers gave specific melting curves and electrophoresis bands as expected (data not shown). The efficiency of the amplification for each primer pair was assessed by a slope of the equation of a standard curve generated from the serial dilutions of the pooled samples. A slope of -3.32 indicates the PCR reaction has 100% efficiency 40. The slopes of all primer pairs are close to each other, from -3.35 to -3.12 (Table 1), suggesting a close efficiency of the amplification between reference and target genes in qPCR. The close efficiency of amplification between target and reference validates the quantification of genes by either the standard curve or the 2^-ΔΔCt methods in further experiments.

The transcripts of cytokine interleukin 1β (IL-1β), IL-6, CXCLi1, CXCLi2, tumor growth factor β4 (TGF-β4) and the inducible nitric oxide synthese (iNOS) were measured in inoculated and mock-inoculated cells using qPCR relative to β-actin. The transcripts of bacterial virulence-associated genes ciaB, dnaJ and racR in invaded/adhered, suspended and those bacteria treated with various media were quantitatively determined by qPCR relative to 16S rRNA, by use of the 2^-ΔΔCt method previously described 41. PCR was performed in duplicate including no-template controls.

Nitric oxide production was determined by measuring nitrite in cell culture media using the Griess Reagent System (Promega) assay according to the manufacture's instruction. The absorbance was read at 550 nm, using a Multiskan EX microtiter plate reader (Thermo Labsystem, Multiskan EX, Denmark).

The values were expressed as the average ± standard deviation (SD). The data were analyzed for statistical significance using one-way ANOVA (ANalysis Of VAriance, Microsoft Excel). A p-value ≤ 0.05 was considered to be statistically significant.