Introduction

Helicobacter (H.) pylori is the best studied gastric Helicobacter species naturally colonizing more than half of the world’s human population. It is responsible for a wide range of gastric pathologies, including cancer [1, 2]. Other spiral-shaped non-H. pylori Helicobacter species (NHPH) have been demonstrated to colonize the human gastric mucosa as well. Their prevalence in humans ranges from 0.1 to 6.2% [3,4,5], with a higher density in Asia compared to Europe [5,6,7]. These NHPHs normally colonize the stomach of animals, but some have zoonotic potential. Particularly porcine H. suis strains and H. heilmannii strains, which are mainly associated with cats and dogs, have been described in human patients with gastritis, gastric or duodenal ulcers or low grade mucosa-associated lymphoid tissue (MALT) lymphoma [5, 8, 9]. Transmission from animals to humans may occur through direct or indirect contact with infected animals, and, in case of H. suis, also via consumption of raw or undercooked pork [8, 10].

Adherence to the gastric mucosa is the initial step in gastric colonization and pathogenesis of Helicobacter spp. infections [11, 12]. To identify which bacterial genes are involved in gastric colonization, gene expression profiles of Helicobacter bacteria attached to gastric epithelial cells need to be compared with those of unbound bacteria. Cell culture experiments may provide initial and valuable information concerning specific host-regulated genes that could subsequently be verified in vivo [13].

So far, transcriptome profiles of H. suis and H. heilmannii upon adherence to the human gastric epithelium have not been described. Previous studies mainly focused on H. pylori-induced changes in host cell gene expression and potential strain differences using the human gastric adenocarcinoma AGS cell line [13,14,15,16,17,18]. This cell line, however, does not form an organized cell layer with functional tight junctions [19]. In this study, we used the human gastric epithelial MKN7 cell line as it resembles more the gastric mucosa by the formation of a contiguous polarized monolayer displaying mucin expression [19, 20]. To gain more insights into the pathogenesis of animal-associated Helicobacter infections in humans, we compared the expression of H. suis and H. heilmannii genes upon binding to human gastric epithelial MKN7 cells to unbound H. suis and H. heilmannii. RNA sequencing approaches were used as they provide a more precise, annotation-independent measurement of transcript levels, compared to other transcriptome profiling methods such as hybridization-based methods [21, 22]. Genes that are significantly up-regulated upon binding to gastric epithelial cells might play a role in gastric pathogenesis and in the ability of Helicobacter to have effects beyond the stomach as well.

Materials and methods

Bacterial strains and growth conditions

The porcine H. suis type strain HS1 (LMG 23995T; DSM 19735T) and the feline H. heilmannii type strain ASB1 (LMG 26292T, DSM 24751T) were used in this study [23, 24]. The bacteria were grown under microaerobic conditions (85% N2, 10% CO2, 5% O2) at 37 °C on biphasic Brucella agar plates (Becton–Dickinson, Erembodegem, Belgium) supplemented with 20% fetal calf serum (HyClone, Logan, UT, USA), 5 mg amphotericin B/I (Fungizone, Bristol-Myers Squibb, Epernon, France), Campylobacter selective supplement (Skirrow, Oxoid, Basingstoke, UK; containing 10 mg/L vancomycin, 5 mg/L trimethoprim lactate and 2500 U/L Polymyxin B) and Vitox supplement (Oxoid). Brucella broth (Oxoid) was added on top of the agar to obtain biphasic culture conditions. The pH of both agar and broth was adjusted to 5 by adding HCl to a final concentration of 0.05%. After incubation, the bacteria were harvested and the concentration was determined using an improved Neubauer counting chamber (Sigma-Aldrich, Saint Louis, Missouri, USA).

H. suis and H. heilmannii fluorescence-based adherence assay

A quantitative fluorescence-based adherence assay was performed to gain insights into the binding capacity of the H. suis and H. heilmannii strains to the human gastric epithelial cell line MKN7 (Riken Cell Bank, Tsukuba, Japan). MKN7 cells were seeded at a concentration of 1 × 106 cells/mL in 200 µL antibiotic-free cell medium (89% RPMI medium 1640 supplemented with 1 mM 1-glutamine (Invitrogen, Waltham, MA, USA) and 10% fetal calf serum (FCS; HyClone, Logan, Utah, USA) in 96-well plates (Greiner-Bio One, Vilvoorde, Belgium) and incubated overnight at 37 °C. The HS1 and ASB1 strains were fluorescently labeled with fluorescein isothiocyanate isomer 1 (FITC, λex 492 nm, λem 518 nm; Sigma-Aldrich). Briefly, bacteria were cultivated as described above and harvested at a concentration of 3.3 × 108 viable bacteria/mL Brucella broth. Next, bacteria were washed 3 times in PBS-Tween 0.05% (2000 g for 5 min) and pellets were resuspended in 0.1 M carbonate and 0.15 M NaCl buffer (pH 9). Subsequently, 10 µL FITC (10 mg/mL dimethyl sulfoxide (DMSO)) was added to 1 mL bacterial suspension. After incubation for 30 min in the dark, FITC-labeled bacteria were washed 3 times in blocking buffer (1% bovine serum albumin (BSA) in PBS-Tween 0.05%). Viability of FITC-labeled bacteria was ascertained by checking their motility using light microscopy. After removing antibiotic-free cell medium from the 96-well plates, 150 µL of FITC-labeled bacterial suspension was added to the cells (5 replicates per strain), followed by incubation for 1 h at 37 °C under microaerobic conditions. Subsequently, cells were washed twice with HBSS + (Hank’s Balanced Salt Solution, Thermo Fisher Scientific, Waltham, Massachusetts, USA) and the emission of fluorescent light at λ = 527 nm was measured with a fluorimeter (Fluoroskan AscentTM FL Microplate Fluorometer and Luminometer, Thermo Scientific, Erembodegem-Aalst, Belgium). The adherence assay was performed immediately after fluorescent labeling of the Helicobacter strains in order to minimize the possible loss in viability of the labeled helicobacters over time. To correct for possible background signals, wells without cells (bacterial suspension only) and wells without bacterial suspension (MKN7 cells only) were included as controls. Additionally, the relative levels of FITC labeling of the HS1 and ASB1 strain were analyzed by flow cytometry (FCM) on a CytoFLEX flow cytometer (Beckman Coulter, Indiana, United States) and the mean fluorescence intensity of each labeled strain was used as a correction factor for differential FITC labeling of the strains. The ratio of the mean fluorescence intensity measured by the fluorimeter (indicating bacterial adhesion) to the mean fluorescence intensity measured by FCM (indicating the relative FITC labeling per strain) was calculated for each Helicobacter strain.

RNA isolation of bound and unbound Helicobacter bacteria

To obtain RNA from H. suis and H. heilmannii strains incubated with the gastric cell line, MKN7 cells were seeded at a concentration of 1 × 106 cells/mL in 2 mL antibiotic-free cell medium in 6-well plates (Greiner-Bio One, Vilvoorde, Belgium) and incubated for 24 h at 37 °C and 5% CO2 until a confluent monolayer was formed. Subsequently, 2 mL containing 6.6 × 108 viable H. suis or H. heilmannii bacteria were added to the wells and incubated for 8 h under microaerobic conditions. After checking the MKN7 confluency, wells were washed 3 times with HBSS + after which 1 mL of 1% trypsin solution consisting of 88% trypsin diluent (8 g NaCl, 0.2 g KCl, 0.12 g KH2PO4, 0.91 g Na2HPO4, and 4 mL 0.5% phenol red solution/1000 mL aqua dest), 10% trypsin stock solution (Invitrogen), and 2% EDTA (2 g/100 mL trypsin diluent) was added to the wells followed by an incubation of 10 min. Next, 1 mL of cell medium without antibiotics was added and the content of the wells was collected, centrifugated (3 min, 1300 g, 4 °C) after which the pellet was stored at −70 °C until RNA extraction. For the RNA isolation of H. suis and H. heilmannii in the absence of gastric epithelial cells, wells without MKN7 cells were included and 2 mL containing 6.6 × 108 viable H. suis or H. heilmannii bacteria were added to the wells. After 8 h of incubation under microaerobic conditions, bacterial suspension was collected, centrifugated (3 min, 1300 g, 4 °C), and the pellet was stored at −70 °C until RNA extraction.

From all collected samples, RNA was then extracted using the RNeasy mini kit (Qiagen, Venlo, the Netherlands) according to the manufacturer’s instructions and RNA-yield was measured using Nanodrop equipment (Nanodrop ND-1000, Fisher Scientifics, Hampton, New Hampshire, USA).

Concentration and quality of the total extracted RNA was checked by using the Quant-it ribogreen RNA assay (Life Technologies, Carlsbad, California, USA) and the RNA 6000 nano chip (Agilent Technologies, Santa Clara, California, USA), respectively.

RNA sequencing and downstream bioinformatics approaches

The NEBNext® UltraTM RNA Library Prep Kit for Illumina® (New England Biolabs, Ipswich, Massachusetts, USA) with rRNA depletion was used for directional total RNA library preparation. Sequencing was performed on the NextSeq 500 High output system (Illumina, San Diego, California, USA).

The obtained raw sequencing reads were trimmed for sequencing adaptors, low quality and ambiguous bases using cutadapt (v1.11) [25]. Initial and final read quality was checked with FastQC (v0.11.5). To remove contaminating reads mapping to human DNA, the trimmed reads were aligned on the human reference genome (GRCh38) using the STAR aligner (v2.5.3) [26] with recovery of unmapped reads.

Reads not mapping on the human genome were mapped against the H. suis HS1 (NCBI accession: NZ_ADGY00000000) and the H. heilmannii ASB1.4 (EMBL accession: HE984298) reference genomes using the Bowtie2 aligner (v2.2.5) [27]. Identification of mRNA features and counting was done with featureCounts from the Subread package (v1.5.3) [28] using the Helicobacter mappings and reference genome annotations as inputs.

Differential gene expression between case (i.e. H. suis and H. heilmannii attached to MKN7 cells) and control (i.e. unbound bacteria) samples was performed in the statistical programming environment R [29], using the DESeq2 package [30] and the SARTools pipeline [31]. Differentially expressed features having a padj ≤ 0.01 and a fold change of at least 2 were identified. A volcano and PCA plot were generated to visualize the differentially expressed genes.

A reciprocal BLAST was performed at the protein level. BLAStp with default parameters was used to identify homologs and to compare differentially expressed genes of H. suis and H. heilmannii. Homology was assumed when BLASTp hits of H. suis and H. heilmannii protein sequences had an e-value below 10 to avoid missing homologs. However, as this criterium alone is not very stringent, we also considered other homology quality metrics like (number of gaps, % identity and % positive match to assess the quality of the match.

Results

Binding capacity of H. suis and H. heilmannii to MKN7 cells

A quantitative fluorescence-based adherence assay was performed to compare binding capacity of H. suis HS1 and H. heilmannii ASB1 to human gastric MKN7 cells. As shown in Figure 1, the mean intensity of fluorescent light emitted after binding of FITC-labeled H. suis and H. heilmannii bacteria to MKN7 cells was not significantly different. Given that the intensity of emitted light is proportional with the quantity of cell-bound bacteria, this experiment shows a similar binding capacity of H. suis and H. heilmannii to gastric epithelial cells.

Figure 1
figure 1

Fluorescence-based adherence assay ofH. suisandH. heilmanniiisolates to gastric epithelial MKN7 cells. The binding of FITC-labeled H. suis (HS1) and H. heilmannii (ASB1) isolates to gastric epithelial MKN7 cells was quantified by measuring the emission of fluorescent light at λ = 527 nm. The relative levels of FITC labeling of HS1 and ASB1 were analyzed by flow cytometry to correct for differential labeling of both Helicobacter strains. Therefore, data are presented as the mean intensity of the emitted light normalized to the relative level of FITC-labeling of H. suis and H. heilmannii strains.

Full size image

Differential gene expression analysis of H. suis adhering to MKN7 cells

When mapping the obtained reads against the human GRCh38 genome (EMBL accession: GCA_000001405.28), we observed that 95–97% of the reads from the H. suis-MKN7 coculture samples (i.e. cases) mapped against the human genome, compared to 0.4% of the reads from unbound H. suis samples (i.e. controls). Conversely, only 3–4% of the reads from H. suis-MKN7 coculture samples (i.e. cases) mapped against the H. suis HS1, compared to 95–96% of the reads from the unbound H. suis (i.e. controls).

Using the SARTools wrapper for DESeq2, a list of 539 differentially expressed genes (padj ≤ 0.05) between case (i.e. bound bacteria) and control (i.e. unbound bacteria) groups was generated for H. suis. After refining the results to those having a padj ≤ 0.01 and a fold change of at least 2, a list of 134 (Additional files 1, 2) differentially expressed genes between case and control groups was obtained. 83 genes were up-regulated (Additional file 1) and 51 were down-regulated (Additional file 2) upon binding to MKN7 cells compared to unbound bacteria. Figures 2A and B display the fold change data of the up- and down-regulated genes (hypothetical protein genes excluded), respectively, in cases compared to controls.

Figure 2
figure 2figure 2

Fold change data of up- and down-regulated genes of MKN7-boundH. suisandH. heilmannii(i.e. cases) compared to unboundH. suisandH. heilmannii(i.e. controls). A Up-regulated H. suis genes. B Down-regulated H. suis genes. C Up-regulated H. heilmannii genes. D Down-regulated H. heilmannii genes. A, B Fold change data of up- (A) and down-regulated (B) H. suis genes (hypothetical protein genes excluded, except for H. suis 104628.16_00458) (with padj ≤ 0.01; fold change ≥ 2). A The up-regulation of 50 H. suis genes (33 hypothetical protein genes excluded) ranged from 2.01- to 3.89-fold. B The down-regulation of 42 H. suis genes (9 hypothetical protein genes excluded) ranged from 2.0- to 3.76-fold. C, D Fold change data of up- (C) and down-regulated (D) H. heilmannii genes (hypothetical protein genes excluded) (with padj ≤ 0.01; fold change ≥ 2). C The up-regulation of 43 H. heilmannii genes (17 hypothetical protein genes excluded) ranged from 2- to 4.64-fold. D The down-regulation of 54 genes (29 hypothetical protein genes excluded) ranged from 2.02- to 5.26-fold. Common up- or down-regulated genes in both MKN7-bound H. suis and H. heilmannii according to the reciprocal BLASTp analysis are displayed in orange. Interesting genes probably associated with virulence and colonization capacity are displayed in blue. 16U: general stress protein 16U gene; FlgA: flagellar basal body P-ring biosynthesis protein gene; FliW: flagellar assembly factor FliW gene; ggt: gamma-glutamyl transpeptidase gene; GrpE: heat shock protein GrpE gene; MACP4: methyl-accepting chemotaxis protein 4 gene; NikB: nickel transport system permease protein gene; omp: outer membrane protein gene; PctC: methyl-accepting chemotaxis protein PctC gene; ppi: peptidyl-prolyl cis,trans-isomerase gene; RecA: recombinase A gene; SPFM: secreted protein involved in flagellar motility gene; TerD: tellurium resistance gene; Tipα: tumor necrosis factor-α inducing protein gene; UreB: urease subunit beta gene; UreF: urease accessory protein UreF gene; UreH: urease accessory protein UreH gene; YlqC: KH domain RNA binding protein gene.

Full size image

Differential gene expression analysis of H. heilmannii adhering to MKN7 cells

When mapping the obtained reads against the human GRCh38 genome (EMBL accession: GCA_000001405.28), 86-90% of the reads from the H. heilmannii-MKN7 coculture samples (i.e. cases) mapped against the human genome, compared to 0.1% of the reads from unbound H. heilmannii samples (i.e. controls). Conversely, only 4–9% of the reads from the H. heilmannii-MKN7 coculture samples (i.e. cases) mapped against the H. heilmannii ASB1.4 genome, compared to 96–97% of the reads from the unbound H. heilmannii samples (i.e. controls).

With the SARTools wrapper for DESeq2, a list of 773 differentially expressed genes (padj ≤ 0.05) between case (i.e. bound bacteria) and control (i.e. unbound bacteria) groups was generated for H. heilmannii. After refining the results to those having a padj ≤ 0.01 and a fold change of at least 2, a list of 143 (Additional files 3, 4) differentially expressed genes between case and control groups was obtained. 60 genes were up-regulated (Additional file 3) and 83 were down-regulated (Additional file 4) when bound to MKN7 cells compared to unbound bacteria. Figures 2C and D display the fold change data of the up- and down-regulated genes (hypothetical protein genes excluded), respectively, in cases compared to controls.

Graphical display of RNA sequencing results for H. suis and H. heilmannii

The volcano plots of the comparisons (case versus control) for both H. suis and H. heilmannii are displayed in Additional file 5. The volcano plots represent the log of the adjusted p-value as a function of the log ratio of differential expression, enabling a quick visualization of those data points that display large magnitude changes that are also statistically significant. Significantly differentially expressed features (padj ≤ 0.01, fold change ≥ 2) are represented by red dots. Dots toward the top of the volvano plot represent highly significantly differentially expressed genes, and dots at either the left- or right-hand side of the volcano plot represent values that display large magnitude fold changes.

To check whether the main variability within each experiment came from biological differences between case and control samples, the first principal components (PC1 and PC2) of the PCA (Principal Component Analysis) were checked (Additional file 6). The first principal component (PC1) separates samples from different biological conditions. In both the PCA-plot for H. suis. (Additional file 6A) and the PCA-plot for H. heilmannii (Additional file 6B), the sample groups (i.e. case and control) were well separated, indicating that biological variability was the main source of variance in the data.

Comparative analysis of the differentially expressed H. suis genes with the H. heilmannii genome

Reciprocal BLASTp analysis of the differentially expressed H. suis genes with the H. heilmannii genome showed that 11 out of 83 up-regulated H. suis genes did not have homologs (i.e. BLASTp hits of H. suis and H. heilmannii protein sequences having e-values below 10) present in the H. heilmannii genome (Additional file 7). The 11 H. suis genes with no H. heilmannii homologs were mainly genes encoding unknown function, with the exception of one gene involved in oxidation–reduction (YdgJ) and one in DNA modification (Additional file 7). Seventy-two out of the 83 up-regulated H. suis genes had homologs present in the H. heilmannii genome (Additional file 7). Only 2 commonly up-regulated genes (i.e. the urease accessory protein gene (UreF) and ferrochelatase gene (displayed in orange in Figures 2A, C, and in Figures 3 and 5) among both species were identified (Additional files 1 and 3). One up-regulated H. suis gene (hypothetical protein gene) had a down-regulated H. heilmannii homolog (Figure 5).

Figure 3
figure 3

Commonly up-regulatedH. suisandH. heilmanniigenes according to reciprocal BLASTp.

Full size image

Furthermore, 6 out of 51 down-regulated H. suis genes did not have homologs present in the H. heilmannii genome (Additional file 8). Although the function of these 6 genes is mostly unknown, 1 gene (i.e. 60 kDa chaperonin 1) is likely to be involved in protein folding (Additional file 8). Forty-five out of 51 down-regulated H. suis genes had H. heilmannii homologs (Additional file 8). Four genes (i.e. recombinase A gene (RecA), tellurium resistance gene (TerD) of H. heilmannii with its H. suis homolog general stress protein 16U gene, KH domain RNA binding protein gene (YlqC) of H. heilmannii with its H. suis homolog hypothetical protein gene, and heat shock protein gene (GrpE)) were commonly down-regulated in H. suis and H. heilmannii and are shown in orange in Figures 2B and D, and in Figures 4 and 5 (Additional files 2 and 4). Two down-regulated H. suis genes (i.e. 30S ribosomal protein S10 gene and UreB) had an up-regulated H. heilmannii homolog (Figure 5).

Figure 4
figure 4

Commonly down-regulatedH. suisandH. heilmanniigenes according to reciprocal BLASTp.

Full size image
Figure 5
figure 5

Comparative analysis between theH. suisandH. heilmanniidifferentially expressed genes.

Full size image

Comparative analysis of the differentially expressed H. heilmannii genes with the H. suis genome

Reciprocal BLAST analysis of the differentially expressed H. heilmannii genes with the H. suis genome showed that 13 out of 60 up-regulated H. heilmannii genes did not have homologs present in the H. suis genome (Additional file 9). The function of most H. heimannii genes not having homologs in the H. suis genome is currently unknown, and only 2 genes involved in translation (S16e, L13Ae) were identified (Additional file 9). Of these 60 up-regulated H. heilmannii genes, 47 genes had homologs present in the H. suis genome (Additional file 9). Apart from the 2 earlier mentioned commonly up-regulated genes among both Helicobacter species, 2 up-regulated H. heilmannii genes (i.e. SSU ribosomal protein S10p (S20e) gene and UreB) had a down-regulated H. suis homolog (Figure 5).

Finally, 9 of the 83 down-regulated H. heilmannii genes did not have homologs present in the H. suis genome (Additional file 10) and encoded hypothetical proteins, except for one gene which is involved in metal ion transport (i.e. Mn2+/Fe2+ transporter gene). Seventy-four out of 83 down-regulated H. heilmannii genes had H. suis homologs, with, as previously mentioned, 4 genes commonly down-regulated among both Helicobacter species (Figures 4, 5 and orange in Figure 2).

Comparative analysis between differentially expressed H. suis and H. heilmannii genes according to gene function

The differentially expressed genes in H. suis and H. heilmannii upon adherence to MKN7 cells compared to unbound bacteria were also classified by function, as shown in Figure 6 and Additional files 11, 12, 13, 14. In general, H. suis and H. heilmannii had both up- and down-regulated genes within the same functional class (Figure 6). However, the up-regulation of DNA replication related genes and down-regulation of cell envelope related genes and genes involved in response to stress were unique for H. heilmannii.

Figure 6
figure 6

Classification by function ofH. suisandH. heilmanniidifferentially expressed (DE) genes.

Full size image

Distinct gene expression patterns associated with virulence and colonization capacity in H. suis and H. heilmannii adhering to MKN7 cells

Subsequently, when comparing the up- and down-regulated genes of H. suis and H. heilmannii upon adhesion to MKN7 cells, and more specifically differentially expressed genes implicated in bacterium–host interactions, several differences between both species were found and described below (Figure 2 and Additional files 1, 2, 3, 4).

Ure genes & nikB

The urease protein is encoded by two urease subunit genes (i.e. UreA and UreB) and several urease accessory genes (i.e. UreE, UreF, UreG, and UreH) which are necessary for enzyme activity and thus survival of Helicobacter in the acidic pH of the stomach [32, 33]. Here, UreF was up-regulated in both H. suis and H. heilmannii (fold change of 2.74 and 2.73, respectively), whereas UreB was up-regulated in H. heilmannii (fold change of 2.18) but down-regulated (fold change of 2.0) in H. suis. Furthermore, UreH was uniquely up-regulated (fold change of 2.05) in H. suis. In addition, urease requires nickel for enzyme activity [34]. In our study, the nickel transport system permease protein gene (NikB) was significantly upregulated in H. suis (fold change of 2.03), but not in H. heilmannii upon adhesion to MKN7 cells.

Flagella encoding genes

Helicobacter suis and H. heilmannii are spiral shaped bacteria containing 4 to 10 bipolar flagella, enabling them to move to and into the mucus layer of the gastric epithelium [8, 35]. Helicobacter movement is mainly driven by chemotaxis, whereby pH, CO2, arginine, urea/NH4, and molecules that lead to bacterial energy generation are detected [36]. Two genes involved in chemotaxis (i.e. the methyl-accepting chemotaxis protein 4 gene and the methyl-accepting chemotaxis protein gene (PctC)) and 1 gene encoding a flagellar basal body P-ring biosynthesis protein (FlgA) were significantly up-regulated in H. suis (fold changes of 3.65, 2.30, and 3.13, respectively) upon binding to MKN7 cells, whereas 2 flagella related genes (flagellar assembly factor gene (FliW) and a gene encoding a secreted protein involved in flagellar motility) were significantly up-regulated in H. heilmannii (fold changes of 2.20 and 2.86, respectively). Remarkably, an FliW homolog gene was significantly down-regulated in H. heilmannii (fold change of 3.01) when bound to the human gastric epithelial cells.

Tipα

Tumor necrosis factor-α inducing protein (Tipα) is considered to be involved in carcinogenesis as well as in colonization of the gastric mucosa by Helicobacter [37,38,39]. In our study, Tipα was significantly up-regulated (fold change of 2.59) in H. suis, but not in H. heilmannii.

Ggt

The gamma-glutamyl transpeptidase (GGT), contributing to gastric colonization, inhibition of T-cell proliferation, and gastric epithelial cell death, was significantly up-regulated in H. heilmannii (fold change of 2.17), but not in H. suis upon binding to the gastric MKN7 cells [40,41,42].

OMPs

Helicobacter species contain a large set of outer membrane proteins (OMPs) as an adaptation to the hostile environment of the stomach [43]. Their outer membrane proteome takes an important role in the colonization of the gastric mucosa. Several outer membrane protein (OMP) genes (omp4, omp13, omp27, omp30, and 1 putative omp gene) were significantly down-regulated (fold changes of 2.46, 2.11, 2.22, 5.26, and 2.58, respectively) in H. heilmannii, but not in H. suis.

Ppi

A gene encoding peptidyl-prolyl cis,trans-isomerase (i.e. ppi), which might play a role in gastric inflammation and apoptosis of gastric epithelial cells, was significantly down-regulated (fold change of 2.53) in H. heilmannii but not in H. suis upon binding to MKN7 cells [44].

Discussion

The human gastric epithelial MKN7 cell line, derived from a well differentiated gastric tubular adenocarcinoma, was used as it resembles a normal, polarized gastric epithelium with mucin production [19, 20]. Nevertheless, keeping in mind the neoplastic background, the effects of receptor phenotypes on H. suis and H. heilmannii could be different from the effects of either in vivo gastric cells or primary cell lines [13]. Therefore, potential differences in gene regulation of adherent H. suis and H. heilmannii might appear when comparing our results with a healthy, non-neoplastic gastric mucosa in vivo.

Quantitative fluorescence-based adherence assays demonstrated an equal, albeit low, binding capacity of the H. suis and H. heilmannii strains to human gastric epithelial cells. In contrast with H. pylori, humans are not commonly infected with H. suis and H. heilmannii, which may have contributed to the low binding capacity observed in this study. For example, it is possible that non-natural host cells lack essential receptors for H. suis and H. heilmannii. Still, the use of human gastric epithelial cells in our study was justified since we were particularly interested in the pathogenesis of H. suis and H. heilmannii in human patients, and less in their natural host.

The low binding level of H. suis and H. heilmannii was translated in a low percentage of identified reads that mapped against the H. suis and H. heilmannii genome. Despite this low percentage, it was justified to further analyze the differential expression of the identified features as sequencing depth was shown to be good. Saturation curves reached a near horizontal plateau, indicating that the gain of new features was small compared to the number of added reads. Since the case groups in our study had almost two orders of magnitude less reads compared to controls, we used DESeq2 as a statistical model as it is claimed to be somewhat more robust when library sizes are very different [45].

Reciprocal BLASTp analysis showed only 2 commonly up-regulated and 4 commonly down-regulated genes among the H. suis and H. heilmannii strains tested here. The low correspondence in differentially expressed genes among both Helicobacter species indicates a distinct transcriptome of the adherent H. suis and H. heilmannii strains. Furthermore, 2 down-regulated H. suis genes (i.e. S10 and UreB) had an up-regulated H. heilmannii homolog. These opposite gene regulations in the H. suis and H. heilmannii strains demonstrated in vitro might indicate that different virulence-associated factors are involved in the pathogenesis of both Helicobacter species upon adherence to the human gastric epithelium in vivo.

When classifying the differentially expressed H. suis and H. heilmannii genes according to their function, it is clear that genes of many different functional classes are expressed upon contact of the bacteria with the human gastric epithelium. The involvement of diverse functional classes upon adhesion was also described for H. pylori and indicates that adhesion to the human gastric epithelium initiates pleiotropic adaptive responses in the bacterium [13]. Discrepancies in functional classes, but also at individual gene expression level between the H. suis and H. heilmannii strains in vitro further underline the probability of distinct pathogenic pathways between both species upon adherence to the human gastric mucosa in vivo.

In general, prediction of the function of H. suis and H. heilmannii genes is based on extrapolation of what is known for H. pylori [24, 46]. It would have been interesting to have included a H. pylori strain in this study, as it would allow comparison of gene regulation of H. suis and H. heilmannii with H. pylori. Still, several genes associated with gastric colonization and virulence of H. pylori were differentially expressed in H. suis and H. heilmannii upon binding to MKN7 cells, indicating that the pathogenesis of animal-associated Helicobacter spp. infections clearly differs from H. pylori.

Gastric Helicobacter species produce urease for their survival and colonization in the acidic environment of the stomach. Urease is a multi-subunit complex which hydrolyzes host-derived urea to ammonia and carbonic acid. The released ammonia neutralizes stomach HCl, creating a neutral microenvironment around the bacterium [32, 47]. The metallocenter of urease consists of two Ni2+ ions, which is important for urease enzyme activity [34]. In our study, UreB and UreF were differentially expressed in adherent H. suis and H. heilmannii, but expression of UreH and NikB was only altered in adherent H. suis. In H. pylori, the function of UreF is still unknown, whereas UreH probably facilitates proper assembly of the urease metallocenter [48]. Interestingly, changes in expression of urease gene subunits have not been described for H. pylori upon attachment to human gastric adenocarcinoma cell lines [13]. These in vitro findings may indicate that urease activity is differently regulated among different Helicobacter species when adhering to human gastric epithelium in vivo.

Apart from urease, motility and chemotaxis are essential for gastric Helicobacter colonization and persistence [49, 50]. Upon adhesion to human gastric epithelial cells, FlgA, PctC, and a methyl-accepting chemotaxis protein 4 gene of H. suis and FliW and a gene encoding a secreted protein involved in flagellar motility of H. heilmannii were significantly up-regulated. Additionally, a FliW homolog was down-regulated in adherent H. heilmannii. In H. pylori, however, a different set of genes encoding flagellar motility have been shown to be up-regulated (i.e. flgB) or down-regulated (i.e. flgM, flgG, and flaA) upon adherence to epithelial cells [13]. These results reinforce previous observations that Helicobacter species present different mechanisms of chemotaxis and motility regulation [51].

Another virulence factor is the tumor necrosis factor-α inducing protein (Tipα). This protein binds to cell-surface nucleolin, which transports Tipα into the cytosol and nuclei of gastric cells, thereby activating NFκB and inducing the expression of TNF-α, IL-1β, IL-8, the up-regulation of Bcl-2 and the down-regulation of p53 [38, 39]. Besides its carcinogenic role, Tipα is also considered to be involved in the stimulation of macrophages and colonization of the gastric mucosa by Helicobacter spp. [37]. Based on our results, the H. suis and H. heilmannii strains used here are able to colonize the human gastric epithelium in similar levels (Figure 1). However, Tipα was only significantly up-regulated in the H. suis strain, indicating that gastric colonization of the H. heilmannii strain is probably regulated by other genes.

Gamma-glutamyl transpeptidase (GGT) has been described to promote gastric inflammation, to contribute to peptic ulcer development, to play a role in the colonization of the gastric mucosa and to modulate the host immune response [42, 52]. Since ggt was only up-regulated in the H. heilmannii strain upon adherence, this enzyme could represent an essential factor for human gastric colonization by H. heilmannii. Previously, it was demonstrated that H. suis ggt mutant strains were capable of colonizing the stomach at levels comparable to H. suis wild type strains [53], confirming that GGT is not essential for H. suis colonization. The role of GGT in H. pylori colonization remains inconclusive as no differences in ggt gene expression were shown upon binding to epithelial cells [13], although an impaired colonization capacity of a H. pylori ggt mutant strain was demonstrated in mice [53,54,55].

Bacterial OMPs are directly involved in the interactions of pathogenic bacteria with their environment. However, the role of the different H. suis and H. heilmannii OMPs in gastric colonization and pathogenesis remains largely unknown [43]. While H. suis and H. heilmannii lack all known H. pylori adhesins described so far [24, 46, 56], H. heilmannii HofE and HofF OMPs have been shown to be important for adhesion to the mouse gastric mucosa, with a higher affinity for gastric epithelial cells than for mucins [57]. Furthermore, two OMP families (i.e. the eight-stranded β-barrel outer membrane protein W (ompW) family and the Borrelia burgdorferi outer membrane spanning (BP-oms28) porin family) seem to be uniquely present in H. suis [43]. In our study, no significant changes in expression could be demonstrated neither for the H. heilmannii HofE and HofF genes, nor for members of the ompW and BP-oms28 family of H. suis. However, several other OMP genes were significantly down-regulated upon MKN7 cell adhesion in the H. heilmannii strain, but not in the H. suis strain. Such down-regulation of OMP genes has also been demonstrated in H. pylori [13] and might be explained by bacterial cell wall modification as a result of proximity and binding of the bacteria to the host cells. For the H. suis strain, no OMP genes were differentially expressed upon epithelial cell adhesion, which suggests a difference in OMP regulation between H. suis and H. heilmannii upon human gastric colonization.

A final virulence factor which was differentially expressed in the H. heilmannii strain upon adherence to human gastric epithelial cells was the secreted peptidyl-prolyl cis,trans-isomerase (PPI). In H. pylori, this protein is known to induce IL-6 release from macrophages and to induce apoptosis of gastric epithelial cells by a cascade of mechanisms initiated through its interaction with toll-like receptor 4 [44]. Indeed, mutant H. pylori strains with inactivated PPI have a very low apoptosis-inducing capacity [44]. The down-regulation of the PPI-encoding gene in the H. heilmannii strain, but not in H. suis strain, might indicate that H. heilmannii is less likely to induce apoptosis of gastric epithelial cells compared to H. suis. However, the role of PPI in H. suis and H. heilmannii needs further elucidation.

Taken together, a large set of H. suis and H. heilmannii genes were significantly up-regulated or down-regulated upon adhesion to human gastric epithelial MKN7 cells. Genes with altered expression profiles belonged to many different functional classes and only few genes were commonly up-regulated or down-regulated in both H. suis and H. heilmannii, indicating that these pathogens follow distinct pathways upon adherence to the human gastric epithelium in vivo.

To further analyze the role of these differentially expressed genes in human gastric colonization and virulence, H. suis and H. heilmannii mutants lacking these genes should be created and tested using in vitro and in vivo models.