All bacterial strains used in this study are listed in Table 1. All Neisseria strains and Streptococcus pyogenes strains were grown on GC agar (Acumedia) containing Kellogg's supplement (Kellogg et al., 1963). Pseudomonas aeruginosa, Staphylococcus aureus, all Salmonella strains and the E. coli strains were grown on Luria agar (Acumedia). The Lactobacillus strains were grown on Rogosa agar (Oxoid). All aforementioned bacteria were cultured at 37°C and 5% CO₂ for 16–18 h before experimentation. The Helicobacter pylori strains were grown on Colombia blood agar (Acumedia) supplemented with 5% defibrinated horse blood and 5% inactivated horse serum (Håtunalab) for 3 days at 37°C under microaerophilic conditions (5% O₂, 10% CO₂). Before each experiment, the bacteria were washed once and resuspended in cell culture medium without serum that was specific to the cell line that was used.

The human pharyngeal epithelial cell line FaDu (ATCC HTB-43), the human colon epithelial cell line Caco-2 (ATCC HTB-37) and the human cervical epithelial cell line ME180 (ATCC HTB-33) were cultured in DMEM + GlutaMAX (Invitrogen) supplemented with 10% heat-inactivated fetal bovine serum (Sigma Aldrich). The human gastric epithelial cell line AGS (ATCC CRL-1739) was cultured in RPMI-1640 + GlutaMAX (Invitrogen) supplemented with 10% heat-inactivated fetal bovine serum. All cell lines were maintained at 37°C and 5% CO₂. The cells were seeded into 24-well tissue culture plates the day before the experiment to form a monolayer overnight. Before each experiment, the cells were washed twice with cell culture medium without serum.

The epithelial cells were infected with bacteria to a multiplicity of infection (MOI) of 100. The bacteria were not removed after addition to the epithelial cells. In some experiments Millicell 0.4 μm filters (Millipore) were used to prevent direct bacterial contact to the host epithelial cells. For infection with dead bacteria, a dense suspension of bacteria was heat-killed at 95°C for 10 min and diluted in cell culture medium to the desired density for infection. After 1, 2, 4, and 6 h of incubation RNA was isolated using the RNeasy Plus kit (Qiagen) according to manufacturer's instructions. The concentration of the RNA was determined using the NanoDrop 8000 (Thermo Scientific) UV-Vis Spectrophotometer. Total RNA was reverse transcribed to cDNA using SuperScript VILO Mastermix (Invitrogen). Quantitative PCR was performed using a LightCycler 480 (Roche) and the SYBRGreen I Master kit (Roche). The primers used are listed in Table 2. The thermal cycling conditions were: initial denaturation at 95°C for 10 min followed by amplification for 40 cycles with denaturation at 95°C for 10 s, annealing at 50°C for 20 s and extension at 72°C for 20 s. The melting curve analysis was as follows: 95°C for 5 s, 65°C for 1 min and then increasing to 95°C at 0.08°C/s. The expression was normalized against the housekeeping gene β-actin. The expression levels were calculated by the comparative C_T method (ΔΔ C_T method) expressed as the fold change compared to uninfected cells.

The epithelial cells were infected with bacteria to a MOI of 100. The bacteria were not removed after addition to the epithelial cells. After incubation for the indicated time points the cells were washed twice with PBS and directly placed on ice. The cells were lysed with 50 μl of 1X sample buffer (63 mM Tris-HCl pH 6.8, 25% glycerol, 1% SDS, 5% 2-mercaptoethanol), boiled for 10 min at 95°C and stored at −20°C until use. Thawed samples were centrifuged for 1 min at 10,000 × g and 15 μl was loaded on 10% acrylamide SDS-PAGE gels. After separation, the proteins were transferred to Immobilon-P PVDF membranes (Millipore) using a semi-dry transfer system (Bio-Rad). The membranes were washed in water and blocked for 1 h in 5% skim milk powder (Sigma Aldrich) in PBS at room temperature. The membranes were incubated overnight at 4°C with antibodies against EGR1 (Abcam, ab194357, 1:10,000 dilution) and β-actin (Millipore, MAB1501, 1:2000 dilution) in 1% skim milk powder in PBS. After washing 3 times with PBS, the membranes were incubated with IRDye800-conjugated goat-anti-rabbit and IRDye680-conjugated goat-anti-mouse antibodies (LI-COR, 1:10,000 dilution) for 1 h at room temperature. Bands were visualized using an Odyssey IR scanner (LI-COR).

The epithelial cells were infected with bacteria to a MOI of 100. The bacteria were not removed after addition to the epithelial cells. After incubation, the cells were washed three times with PBS to remove unbound bacteria. Bacterial adherence was estimated from viable counts by lysing the host epithelial cells with 1% saponin in cell culture medium for 10 min and plating serial dilutions. Viable counts for P. aeruginosa were performed on Luria agar plates, incubated at 37°C and 5% CO₂ and the colony forming units (cfu) were counted the following day. Viable counts for H. pylori were performed on blood agar plates incubated at 37°C under microaerophilic conditions for 4–7 days. Viable counts for all other bacteria were performed on GC agar and incubated at 37°C and 5% CO₂ for 1–2 days.

PD153035 (125 nM), PD184352 (125 nM), SP600125 (25 μM), p38 MAP kinase inhibitor IV (1 μM) and Protein Kinase A inhibitor fragment 14:22 (10 μM) are chemical inhibitors of the EGFR, ERK1/2, JNK, p38 and PKA signaling molecules, respectively. All inhibitors used were purchased from Sigma Aldrich and resuspended in DMSO. The cells were pre-treated with the inhibitors for 1 h and then co-incubated with the inhibitors and bacteria (MOI 100) for 2 h at 37°C and 5% CO₂. In the experiments with N. gonorrhoeae the cells were co-incubated for 4 h. Following incubation, adhesion assays and qPCR analysis was performed as described in the sections above.

The epithelial cells were seeded into 24-well tissue culture plates to a confluency of 50–80% prior to transfection. For RNA silencing, AGS cells were replaced by MKN-45. The cells were washed twice with serum-free cell culture medium and transfected with 25 nM ON-Target Plus SMARTpool siRNA (Dharmacon) in Opti-MEM (Invitrogen) using Lipofectamine RNAiMAX (Invitrogen) according to manufacturer's recommendations. Following an overnight incubation cell culture medium supplemented with 10% FBS was added. The cells were maintained for a further 48 h at 37°C and 5% CO₂ before use in experiments for bacterial adhesion and qPCR analysis as described in the sections above. The efficiency of knockdown was determined using qPCR.

All experiments were performed in triplicate and repeated three times. Analysis of variance (ANOVA) and the Student's t-test were employed to analyze the difference between the groups for statistical significance. P < 0.05 was considered statistically significant. The data is represented as the mean ± standard deviation. The asterisk in the bar graph denote statistical significance. The significance level is represented in the graphs as ^***P < 0.001, ^**P < 0.01, ^*P < 0.05, NS-non significant.

Only a few studies have investigated the bacterial induction of EGR1 in host cells. To determine whether the induction of EGR1 is a general stress response of the host cells to bacterial colonization, we performed a screen using 25 different bacterial strains including both Gram-positive and Gram-negative pathogens and non-pathogens. We divided the bacterial strains into groups depending on the site of isolation. Bacteria isolated from the upper respiratory tract were added to FaDu pharyngeal epithelial cells (Figures 1A,B), gastric isolates were added to AGS gastric epithelial cells (Figure 1C), and intestinal isolates were added to intestinal epithelial Caco-2 cells (Figure 1D). Bacterial isolates from the cervix were added to cervical ME180 cells and are shown in Figure 1C.

Interestingly, the majority of strains tested were able to upregulate EGR1 irrespective of their pathogenicity or Gram staining type (Figures 1A–D). N. meningitidis strains of different serogroups, P. aeruginosa, as well as the non-pathogenic N. lactamica and N. subflava all triggered EGR1 expression at 2 h post-inoculation (Figure 1A). A similar induction of EGR1 occurred at 2 h post-inoculation with different serogroups of S. pyogenes, S. aureus, and the non-pathogenic oral isolates L. reuteri and L. salivarius (Figure 1B). The gastric pathogen H. pylori triggered EGR1 upregulation at 2 h after infection, whereas the non-pathogenic gastric isolate L. rhamnosus did not (Figure 1C). None of the tested E. coli strains induced EGR1 in intestinal Caco-2 cells (Figure 1D). Salmonella enterica serovar Enteritidis triggered EGR1, whereas S. enterica serovar Typhimurium did not (Figure 1D). Of the cervical isolates N. gonorrhoeae, but not L. crispatus, could induce the transcription of EGR1 (Figure 1C). Western blot analysis showed that the increase in EGR1 expression also occurred at protein level (Figures 1E–H). EGR1 is an early response transcription factor that can be induced rapidly. We detected EGR1 upregulation at 1 h post-inoculation for several strains and a peak in its transcriptional activity at 2 h. Only Neisseria gonorrhoeae displayed different time kinetics with the highest EGR1 mRNA levels at 4 h after infection (Figure 1C).

We also examined the gene expression levels of factors that are known to be involved in either the upstream [β1-integrins, epidermal growth factor receptor (EGFR)] or downstream (fibronectin, amphiregulin) signaling of EGR1. No change was detected for β1-integrins and fibronectin (Supplementary Figure S1). Upregulation of the transcript levels of EGFR was observed only upon an infection with H. pylori (Supplementary Figure S1). The expression of amphiregulin was induced by a few bacteria, i.e., P. aeruginosa, S. pyogenes M5, N. gonorrhoea and L. crispatus, but not by all bacteria that upregulated EGR1 (Supplementary Figure S1). Moreover, the induction of EGFR and amphiregulin occurred at later time points than that for EGR1, i.e., at 4–6 h post infection. This result suggests that the upregulation of EGFR and amphiregulin are bacterial species specific and independent of EGR1 induction.

We hypothesized that difference in EGR1 induction might be dependent on the amount of bacteria in contact with the host cells. We therefore determined the level of adhesion for each bacterial strain (Figures 1I–L). However, we could not find any correlation between attachment and EGR1 upregulation. For example, S. enterica serovar Enteritidis and S. enterica serovar Typhimurium adhered to the host epithelial cells at similar levels, but only S. Enteritidis induced EGR1 whereas S. Typhimurium did not (Figures 1D,L).

Taken together, EGR1 expression can be induced at both the transcriptional and protein level by several species of bacteria and is independent of the level of bacterial adherence, Gram-staining type and pathogenicity.

Different bacteria colonize different sites within the human body. To determine the role of site specificity in the upregulation of EGR1 several cell lines representing different body sites were infected by the same bacterial species. We selected a representative strain from each group; N. meningitidis serogroup C (Nm-C), S. pyogenes serogroup M6 (Sp-M6), H. pylori J99 (Hp-J99), N. gonorrheae (Ng), and S. enterica serovar Enteritidis (SE-3934). Remarkably, most bacteria could induce the strongest EGR1 response in the cell type of their natural niche and upregulation was low or absent in host cells that did not represent the natural colonization site (Figures 2A–E). An exception is the intestinal pathogen S. Enteritidis that induced a very strong response in the gastric epithelial cell line AGS (Figure 2E). Upregulation in other cell types sometimes showed different time kinetics, such as that observed for N. meningitidis where EGR1 levels peaked at 2 h in FaDu cells, which represents the natural colonization site of the nasopharynx, EGR1 expression peaked at 4h in ME180, a cervical cell line, and in the gastric epithelial cell line AGS (Figure 2A). Interestingly, infection with N. gonorrhoeae showed the same pattern (Figure 2D). This finding is not surprising, because these bacteria are closely related species that colonize both the pharynx and urogenital tract and therefore are likely to induce similar host responses. At 2 h post infection, S. pyogenes specifically triggered EGR1 expression in target FaDu cells, whereas after 6 h, EGR1 was unexpectedly induced in AGS cells (Figure 2B).

Similar to the screening experiments, we compared the binding of the bacteria to the different cell lines (Figures 2F–J). Again, no correlation between EGR1 upregulation and adherence could be established. This result was illustrated by H. pylori, which adhered equally to all cell types at early time points, but induced different EGR1 expression levels (Figures 2C,H).

Because we detected upregulation of EGFR and amphiregulin for some bacterial strains in initial screening experiments, we also investigated the expression of these genes in the different cell lines (Supplementary Figure S2). Amphiregulin was upregulated by N. gonorrhoeae and S. Enteritidis in ME180 cells, and some induction could also be detected in Caco-2 cells for S. Enteritidis (Supplementary Figure S2). Similarly, EGFR was induced only by H. pylori and most pronounced in AGS cells (Supplementary Figure S2). These results strengthen the findings of the screening experiments, which showed that the induction of EGFR and amphiregulin are cell type and species specific, and are possibly not related to EGR1 upregulation.

In conclusion, bacteria-mediated induction of EGR1 is mostly cell type specific and a precise match between bacterium and host cell is necessary to elicit the maximum response.

We hypothesized that active interaction between bacteria and host cells was necessary to trigger EGR1 induction. To address this, we infected host epithelial cells with live, heat-killed or live bacteria separated from the host cells by a filter. The maximum induction in the EGR1 was observed with the live bacteria (Figure 3). Heat-killed N. meningitidis, S. pyogenes, H. pylori and N. gonorrhoeae were not able to induce EGR1, suggesting that bacterial viability is important (Figures 3A–D). However, S. Enteritidis could upregulate EGR1 after heat treatment (Figure 3E). After the heat treatment, the bacteria were not washed indicating a possible role of surface molecule or component that is released upon heat treatment might be involved in the upregulation of EGR1 by heat-killed S. Enteritidis. Separation of the bacteria from host cells by using a Millicell 0.4 μm filter inhibited EGR1 induction in cells infected with N. meningitidis, S. pyogenes, H. pylori, N. gonorrhoeae, and S. Enteritidis. This observation indicates that contact between bacterium and the host cell is necessary (Figure 3).

The data indicate that upregulation of EGR1 in host epithelial cells is dependent on the viability of the bacteria and is usually mediated through contact between the bacterium and host cell.

Next, we aimed to identify the molecular mechanisms through which bacteria induce EGR1. There are several signaling pathways upstream of EGR1 and therefore we used different chemical inhibitors to specifically block each of these pathways. Interestingly, inhibition of ERK1/2 and EGFR completely abolished the EGR1 upregulation for all the 5 strains tested (Figures 4A–E). Inhibition of signaling through JNK reduced EGR1 upregulation for H. pylori, N. gonorrhoeae and S. Enteritidis (Figures 4C–E). Inhibition of PKA resulted in a reduction of EGR1 induction specifically for S. Enteritidis (Figure 4E). P38 pathway played no role in the bacteria mediated induction of EGR1. Treatment of the host cells with the inhibitors did not affect the adhesion of the bacteria, indicating that EGR1 upregulation is not important for bacterial attachment to host cells (Supplementary Figure S3 online). Since, the blocking of EGFR and ERK1/2 exhibited the strongest inhibition of EGR1 upregulation for all the bacteria tested, we propose that the EGFR–ERK1/2 signaling pathway is a commonly used route for bacterial induction of EGR1.

Growth factor receptors can cooperate with integrins to induce signaling and/or enhance the response upon activation by their ligands (Giancotti and Tarone, 2003). Integrin and EGFR cross-talk has also been implicated in EGR1 expression (Cabodi et al., 2009). We therefore used siRNA to silence β1-integrins and integrin-linked kinase (ILK) in host epithelial cells to investigate whether integrin signaling is involved in EGR1 upregulation by bacteria. Silencing β1-integrins completely inhibited the induction of EGR1 expression after infection with N. meningitidis, S. pyogenes, H. pylori, N. gonorrhoeae, and S. Enteritidis (Figures 5A–E). ILK played role in the upregulation of EGR1 only for S. Enteritidis (Figure 5E). The adherence level of bacteria to host cells was not affected by silencing of EGR1, β1-integrins or ILK (Supplementary Figure S4). Downregulation of the transcription of the target genes by siRNA treatment in each cell line was confirmed using qPCR (Supplementary Figure S5). Therefore, signaling through integrins is important for bacteria-mediated EGR1 induction.

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.