Helicobacter pylori dupAIs Polymorphic, and Its Active Form Induces Proinflammatory Cytokine Secretion by Mononuclear Cells

Background. Infection with Helicobacter pylori possessing a newly described virulence factor—duodenal ulcer– promoting gene A (dupA)—has been associated with duodenal ulceration and increased gastric inflammation. Methods. The dupA locus of 34 strains was sequenced. A panel of dupA mutants was generated and cocultured with human gastric epithelial cells and peripheral blood mononuclear cells; proinflammatory cytokine release was measured. IL8 expression was measured in human gastric biopsy specimens and related to the dupA and cagA status of infecting strains. Results. Most H. pylori strains had a dupA allele that was longer (1884 bp; dupA1) than previously described dupA alleles, although some had truncated versions (dupA2). Unlike the best-characterized H. pylori virulence determinant, the cag pathogenicity island (cag PaI), neither dupA type induced release of interleukin (IL)-8 from gastric epithelial cells. However, infections due to dupA-positive strains were associated with higher-level mucosal IL-8 messenger RNA expression in the human stomach than were infections due to dupA-negative strains. To explain this paradox, we found that dupA1 (but not dupA2 or the cag PaI) substantially increased H. pylori– induced IL-12p40 and IL-12p70 production from CD14+ mononuclear cells. Other T helper 1–associated cytokines were also modestly induced. Conclusion. We suggest that virulent H. pylori strains cause inflammation by stimulating epithelial cells through cag-encoded proteins and mononuclear inflammatory cells through dupA1 products. Helicobacter pylori is the main cause of peptic ulceration, gastric adenocarcinoma, and gastric lymphoma [1]. Its best-characterized virulence factor is the cag pathogenicity island (cag PaI) [2, 3]. Strains possessing the cag PaI are associated with severe gastritis, peptic ulcers, and gastric adenocarcinoma. The cag PaI encodes a type IV secretion system that translocates CagA into host cells and induces nuclear factor–kB activation and interleukin (IL)-8 secretion [4–9]. Epithelial cell IL-8 production is the major driver of gastric mucosal inflammatory cell infiltration. Other virulence factors, including BabA and OipA, also increase epithelial cell IL-8 secretion, although the mechanisms for this increase remain unclear [10, 11]. Recently, a novel H. pylori virulence factor, a virB4 homologue, was identified and named dupA (“duodenal ulcer promoting gene A”) [12]. dupA-positive strains are associated with high-level gastric inflammation [12] and, in most reports, duodenal ulceration [13–15]. H. pylori virulence genes are often highly polymorphic; more-active types have bigger effects on epithelial cells and stronger associations with disease [16–19]. C142, the candidate strain discussed by Lu et al [12], contained a single open-reading frame (ORF) of 1839 bp at the dupA locus, whereas the genome sequence strain J99 comprised 2 ORFs (jhp0917 and jhp0918). Lu and colleagues described 23 other strains as having dupA alleles similar to those of strain C142, but they did not present nucleotide sequence data. We show that dupA is commonly polymorphic; the most prevalent form has not previously been described. Contrary to previous reports, we show that dupA, unlike the cag PaI, does not influence IL-8 secretion by epithelial cells. However, we confirm the association of dupA with IL8 expression levels in the H. pylori–infected gastric mucosa. To explain this paradox, we show that dupA (but not cagA) is required for high-level H. pylori–induced proinflammatory cytokine production by human mononuclear cells. All chemicals and reagents were obtained from Sigma-Aldrich, unless otherwise stated. Strains and patients. H. pylori wild-type strains were from our international collection. Gastric antral biopsy specimens were donated by patients undergoing routine upper gastrointestinal endoscopy for investigation of dyspepsia at the Queen's Medical Centre, Nottingham, United Kingdom. All patients provided full, informed written consent, and the study was approved by the ethics committee of the Nottingham University Hospitals Trust. We studied 17 consecutively observed H. pylori–infected patients (mean age ± standard deviation [SD], 58 ± 15 years; percentage of patients who were male, 41%) and 5 uninfected control subjects (mean age ± SD, 56 ± 25 years; percentage of controls who were male, 60%). One biopsy specimen was placed in Iso-Sensitest broth (Oxoid) containing 15% (vol/vol) glycerol and was cultured within 3 h; a second biopsy specimen was placed in RNAlater solution for messenger RNA analysis; and a third was saved for direct DNA extraction and polymerase chain reaction (PCR)–based H. pylori typing [14]. PCR amplification and nucleotide sequencing of dupA. dupA was PCR amplified using primers DupA-WXF (5′-ATATACCATGGATGAGTTCYRTAYTAACAGAC-3′) and DAR1 (5′-TTAAATACTCTTCCTTATAAGTTTCTTGG-3′). A second reverse primer, JHP0919R2 (5′-GCCCACCAGTTGCAAAAACAAATGAAC-3′), was designed within jhp0919 to include the next predicted stop codon, based on the nucleotide sequence from strain J99 and used with DupA-WXF or the previously described primer DupA918F [15]. Amplified genes were sequenced on both strands with the use of BigDye Terminator version 1.1 (Applied Biosystems) on a Perkin-Elmer ABI Prism 3130 fluorescent analyzer. Mutant construction. We constructed dupA insertion mutants in several different H. pylori strains (Figure 1). Most of dupA was amplified from strain 93-67, with the use of primers DAF1 (5′-GACGATCTATTGAATTTATACGCTG-3′) and DAR1, and cloned into pGem-T Easy (Promega). Inverse PCR was used to engineer XhoI and XmaI sites into dupA with the use of primers NHDA2R (5′-CTATATCCCGGGAGCTTGAAAATTGAAAAAGTG-3′) and NHDA3F (5′-CAATACCTCGAGGCTAATACAGATGGTGGT-3′). The chloramphenicol resistance cassette from pBSC103 [21] was excised (XhoI/XmaI) and ligated into dupA to create plasmid pCDA723. H. pylori strains were transformed with this, and allelic exchange mutants were selected by marker rescue on blood-agar plates containing 30 mg/mL chloramphenicol. Disruption of dupA and confirmation of double crossover were determined by Southern blotting and restriction digestion. Comparative broth cultures of mutants and parental strain pairs demonstrated no differences in growth rates. We also constructed isogenic cagE insertion mutants in dupA-deficient mutants of strains 93–67 and AB31, as described elsewhere [22]. Gastric epithelial cell culture. AGS and ST16 cell lines [23] were grown in nutrient mixture F12 Ham (Invitrogen) containing 10% (vol/vol) fetal bovine serum; MKN45 cells were grown in RPMI 1640 medium with 10% fetal bovine serum and 200 mmol/L l-glutamine (Invitrogen); and MKN28 cells were grown in DMEM/Ham's F12 mixture with 10% fetal bovine serum and 200 mmol/L L-glutamine. Cell lines were cultured at 37°C in a humidified atmosphere of 5% carbon dioxide. Peripheral blood mononuclear cell isolation. Twenty-milliliter blood samples obtained from 6 H. pylori–negative volunteers were collected into K3 ethylenediaminetetraacetic acid vacutainer tubes (Greiner). Peripheral blood mononuclear cells (PBMCs) were immediately purified using Histopaque 1077 and cultured in RPMI 1640 medium with 10% fetal bovine serum and 200 mmol/L L-glutamine [24]. Coculture of cells with H. pylori. Gastric epithelial cell lines were seeded into 24-well culture plates and grown to 80%– 90% confluence. PBMCs were plated at 2 × 106 cells/well. H. pylori strains, which were grown on blood agar for 24 h, were harvested into 5 mL of cell culture media, and the absorbance was standardized to 0.1 at 600 nm. H. pylori was added to cultures at a multiplicity of infection of 100 for epithelial cells and 20 for PBMCs. Coculture assays were performed in 5 replicates per culture condition. Supernatants were removed after 6 h and 48 h of culture, respectively, for cytokine enzyme-linked immunosorbent assays (ELISAs). Cytokine ELISAs. Cytokine concentrations in coculture supernatants prepared from 4 replicate epithelial cell experiments or with PBMCs from 5 donors were measured using human IL8 (DuoSet; R&D Systems), IL-10, tumor necrosis factor (TNF)– a, interferon (IFN)–γ (eBioscience), and IL-12p40 (Autogen Bioclear) ELISA kits, in accordance with the manufacturers' instructions. For further investigation of IL-12p40, PBMCs from an additional donor were cocultured with and without H. pylori, with the use of 5 replicate wells per culture condition. IL-12p70 and IL-23 ELISA kits (eBioscience) were used. Optical density values were measured at 450 nm and 620 nm, and cytokine concentrations were calculated by reference to a standard curve included on each plate. Real-time PCR. Total RNA was purified using an RNeasy Mini Kit (Qiagen). Complementary DNA (cDNA) was synthesized by SuperScript II reverse transcriptase (RT) (Invitrogen) with oligo(dT) primers. RT-PCR was performed on a Rotor-Gene3000 (Corbett Research) with the use of a DyNAmo HS SYBR Green quantitative PCR Kit (New England Biolabs) [24]. No-template controls were included in each cycle. First-stage RT-PCR samples produced without RT from each RNA template were tested in parallel. Analysis used the method of Pffafl [25]. Relative expression levels of IL8 were determined by normalizing against GAPDH and using an H. pylori–negative gastric mucosal cDNA comparator. Positive control human cDNA (BD Biosciences) was included in all assays. Amplification was done over 45 cycles at 95°C for 15 s, at 61°C for 30 s, and at 72°C for 30 s. The primers used were 5′-CTCTTGGCAGCCTTCCTGA-3′ (IL8 forward), 5′-AGTTCTTTAGCACTCCTTGGCA-3′ (IL8 reverse), 5′-CCACATCGCTCAGACACCAT-3′ (GAPDH forward), and 5′-GGCAACAATATCCACTTTACCAGAGT-3 (GAPDH reverse). Flow cytometry. A total of 5 × 105 PBMCs were incubated with H. pylori at a multiplicity of infection of 20 or in medium alone for 16 h. Cells were stained with anti–CD14-phycoerythrin–cyanin 5.1 (PC5; AbD Serotec) before staining with anti– IL-12p40–phycoerythrin (PE; eBioscience) or anti–IL-12p70– PE (BD Pharmingen), as described elsewhere [24]. Data on 100,000 events were acquired using a Coulter Epics Altra flow cytometer. Isotype controls were used to ensure staining specificity. Analysis was performed using Weasel software (version 2.5; Walter and Eliza Hall Institute of Medical Research, Melbourne, Australia). Statistics. Data were analyzed using Minitab 15 software (version 15; Minitab). In box-and-whisker plots, horizontal lines denote median values, boxes denote first to third quartiles, and whiskers denote the lowest and highest observations within 1.5 times the interquartile range. Paired Student's t tests were used to compare the cytokine responses of the 5 PBMC samples in coculture supernatants, and unpaired t tests were used when PBMCs obtained from only a single donor were cultured. Unpaired RT-PCR data were compared using Mann-Whitney U tests. P ≤ .05 was considered to denote statistical significance. Nucleotide Sequence Analysis of the H. pylori dupA Locus Natural polymorphisms within dupA. To examine dupA polymorphisms, we initially sequenced dupA from 8 dupApositive H. pylori clinical isolates from the United States and the United Kingdom AB31, and and the of which only in the version of the of had that truncated the (Figure 1). All 8 had a at the of dupA that was not present in previously described strains or J99 of an insertion a which the acid sequence nucleotide and (Figure In all but strain an additional insertion in a second the predicted and the stop described in strains and the To the of the stop in strains, we PCR amplified and sequenced the of dupA and the of the gene In strains, the stop was found at for other strains from 5 bp after the of jhp0919 and the United States [15]. strains the and of the 8 strains that we initially form of dupA. and 4 AB31, and had a dupA of 2 and Strains longer than previously described dupA and had dupA ORFs of bp and had a dupA of 1839 as described for strain (Figure The most form of dupA [12] dupA in predicted strains in 25 of 34 in was the acid sequence at the and 3 had 2 ORFs in We have dupA into 2 main the dupA locus and The most prevalent form of dupA. found sequence than we further which were most Nucleotide sequence of the of dupA the of and the of was performed for other strains from and the United States [15]. strains the form of dupA. and 2 and Strains and had dupA ORFs of bp and (Figure The most form of dupA strains in 25 of 34 in was the We have dupA into 2 main We to the most within as and we to dupA with an stop as this the locus into 2 ORFs that are to and found in strain but are several different the that this is a of dupA on of IL-8 from Gastric Epithelial Cells dupA and of of IL-8 secretion from gastric epithelial cells in Lu et al [12] that dupA in strain induced IL-8 production by gastric epithelial cell having isogenic dupA we dupA1 and dupA2 alleles induced similar levels of IL-8 We selected 2 dupA1 strains and and 2 dupA2 strains and (Figure 1). [12] was not to we selected a strain with a dupA of the All were cag PaI We cocultured strains or isogenic dupA mutants with the gastric epithelial cell MKN45 used elsewhere for 48 h and IL-8 levels in (Figure cultures were performed with 2 cag wild-type and dupA strains and not no differences were observed of the and isogenic To this the strains and dupA mutants were cocultured with gastric epithelial cell lines and no effects were not The of cag PaI on gastric epithelial cell IL-8 responses to H. pylori in the and of dupA. The cag PaI is the main of H. pylori–induced epithelial cell IL-8 release To cag IL-8 responses were by dupA we constructed cagE mutants in strains and 93–67 wild-type and dupA isogenic strains. Coculture with AGS cells for 48 h that dupA had no on IL-8 as cagE (Figure AGS cell IL-8 by 20 wild-type H. pylori isolates possessing or cagE dupA was strains induced but dupA status had no (Figure of dupA with IL8 expression in We the association dupA status and IL8 expression in [12]. RT-PCR of RNA was performed on human gastric biopsy specimens obtained from 17 H. patients and 5 H. pylori–negative strains were were dupA and were dupA RNA levels were for dupA-positive infections than for dupA-negative infections and they were for uninfected patients (Figure The was present when only samples were considered were dupA and were dupA of dupA on Cytokine of Cells dupA1 and increased H. pylori–induced production of IL-8 and other cytokines by PBMCs. dupA was associated with increased gastric IL8 we that dupA with of cells other than epithelial cells. the are to have effects on and inflammatory cells we studied the of dupA on IL-8 production by and we other cytokines to epithelial cell IL-8 production PBMCs from 5 donors were cocultured with strains and and isogenic dupA 48 h, IL-8 secretion was but in to types than dupA mutants AB31, the increase was for the increase was (Figure but differences were observed in AB31, the increase was for the increase was (Figure and AB31, the increase was for the increase was (Figure However, a was on IL-12p40 secretion, the wild-type strains induced than the concentrations by the dupA AB31, the increase was for the (Figure We observed no differences in of and from PBMCs by wild-type H. pylori strains, compared with dupA-negative isogenic mutants not of dupA polymorphisms on and other cytokine responses from PBMCs. To examine dupA2 also in PBMC cytokine production and to compare this with dupA1 we IL-12p40 IL-12p40 release by PBMCs cultured with wild-type strains was than that induced by dupA2 strains (Figure in IL-12p40 secretion was observed dupA2 wild-type and isogenic that dupA2 is (Figure dupA2 had no on H. pylori–induced or IL-8 secretion from PBMCs not of dupA and the cag PaI on H. pylori–induced secretion by mononuclear cells. PBMCs were incubated with cagE and dupA single and double mutants of strains and described dupA the in of cagE had no (Figure dupA, but not an cag PaI, H. pylori–induced IL-12p40 secretion by mononuclear cells. of IL-12p40 is a of the proinflammatory cytokines IL-12p70 and and also as a We which of IL-12p40 were in to dupA (Figure The IL-12p40 of PBMCs cultured with wild-type strains was than that induced by the isogenic dupA1 mutants the increase was for AB31, the increase was (Figure The using blood from a different single were than those previously but the were IL-12p70 levels (Figure were than IL-12p40 but the in to dupA and wild-type strains was the increase was for AB31, the increase was IL-23 responses (Figure were than IL-12p40 and IL-12p70 but was a increase in the to wild-type strains, compared with dupA strains the increase was for AB31, the increase was that dupA increased IL-12p40, and IL-23 expression from PBMCs. are the most of IL-12p40, and IL-23 in our PBMC although a that IL-12p40 by T cells in to H. pylori To confirm the ELISA and to the of PBMCs were incubated with the wild-type strain and dupA before for IL-12p40, and IL-12p70 and flow 1). with the the wild-type induced IL-12p40 and IL-12p70 responses than did the dupA-negative Most events were CD14+ of responses and of dupA is a described H. pylori virulence The disease associations are different in different nucleotide sequencing of dupA that this gene is We that this to the differences and dupA in some The most form of dupA was a previously that of and the of jhp0919 in the genome sequence strain The second most form 2 although in and stop are also with or of the dupA allele in strain is a truncated no are we not which of dupA alleles we used the of dupA our in to dupA Unlike Lu et al [12], we that dupA had no on H. pylori–induced IL-8 production by epithelial cell was tested using 4 different cell lines and of isogenic dupA In of our data from et al who used clinical H. pylori isolates cocultured with AGS no association dupA status and IL-8 that strain C142, which was used by Lu et al [12], does have an not by other strains. However, strain which was similar for dupA, did not increased IL-8 our on epithelial cells in we the association of dupA-positive infection with IL-8 expression in the gastric mucosa. elsewhere [12], we found a when for cag PaI We that the major of dupA on other cell by stimulating IL-8 release or by stimulating the release of other cytokines to epithelial cell IL-8 We on mononuclear as a major of inflammatory cytokines in the gastric mucosa. our we found that dupA1 increased H. pylori–induced proinflammatory cytokine production by but dupA2 was the cag PaI, on epithelial cell cytokine production and described association with the of a had no on H. pylori–induced mononuclear cell cytokine In to on IL-12p40 expression by mononuclear dupA1 also had a on IL-12p70 and both of which an IL-12p40 proinflammatory are associated with H. pylori–induced and and with severe inflammation and disease The IL-12p40 was than the IL-12p70 and IL-23 the form was induced. was considered to an but is to as a for and cells and to associated with inflammatory The of IL-12p40 and IL-12p70 was found to CD14+ cells. experiments the by strains in cells. in that the of the T helper to H. pylori is for or of the infection and or of that dupA induces gastric inflammation and or through than epithelial and that increase the of disease by stimulating a One is H. pylori and virulence into with inflammatory cells in the gastric or they are by the epithelial However, the of H. pylori through the into in after by CagA and other of dupA data on H. pylori and with the human Other virulence factors, as the cag PaI, through proinflammatory cytokine secretion by gastric epithelial to inflammatory cell into the gastric mucosa. dupA to at the next those inflammatory cells to influence the type of this further the of the of H. pylori with which to We and for with flow the of our that the dupA allele of some strains have an To our we analyzed the of dupA by polymerase chain reaction and sequencing from strains and using primers and strains have a dupA in coculture We that the dupA allele in strains is at the to a total of the predicted in the dupA locus to the open-reading in strain J99 for the strains for which we obtained dupA sequence in and 2 previously described strains in Strains AB31, and the form and are as all other strains are as cag status was determined by polymerase chain reaction the of cagA and cagE or the of the pathogenicity by use of PCR. from which a dupA was for is the strain used in the previous experiments of Lu et al [12]. is the genome sequence strain described by et from which double mutants were for in the acid sequence at the dupA locus from acid for Helicobacter pylori strains from levels of interleukin (IL)-8 secretion from MKN45 cells cocultured with Helicobacter pylori strains and and isogenic dupA mutants for 48 h. The the wild-type and the the isogenic dupA The concentrations of IL-8 in the medium were determined by enzyme-linked immunosorbent denote standard 4 IL-8 secretion from AGS cells after 48 h of coculture with H. pylori wild-type strains and 93–67 and isogenic dupA and cagE denote wild-type dupA cagE and double and The concentrations of IL-8 in the medium were determined by denote standard different from the wild-type strain IL-8 secretion from AGS cells in to coculture with wild-type strains of dupA and cag PaI AGS cells were cocultured for 48 h with clinical isolates of H. pylori strains possessing or dupA and possessing or the cag PaI 5 per different from dupA-positive strains .05 IL8 messenger RNA expression in antral gastric biopsy specimens obtained from patients with dupA-positive and dupA-negative H. pylori strains. Real-time polymerase chain reaction was normalizing IL8 expression to the gene GAPDH and in with DNA prepared from uninfected biopsy to a specimens obtained from 17 H. patients were tested were dupA and were dupA with biopsy specimens from 5 uninfected patients as dupA and dupA Cytokine production from peripheral blood mononuclear cells (PBMCs) after 48 h of coculture with Helicobacter pylori strains and dupA1) and isogenic dupA denote wild-type strains, and denote dupA The denote obtained from PBMCs cultured in the of H. pylori. (IL)-8 interferon and tumor necrosis factor were measured by enzyme-linked immunosorbent denote the standard deviation the type and the IL-12p40 secretion from PBMCs after 48 h of coculture with H. pylori strains AB31, and and isogenic dupA denote wild-type strains, and denote dupA The obtained from PBMCs cultured in the of H. pylori. The concentrations of IL-12p40 in the medium were determined by denote the standard deviation 5 the type and the secretion from peripheral blood mononuclear cells (PBMCs) after 48 h of coculture with Helicobacter pylori strains 93–67 and isogenic dupA and cagE denote wild-type dupA cagE and double cagE and dupA The concentrations of IL-12p40 in the medium were determined by enzyme-linked immunosorbent denote standard different from the type IL-12p70 and IL-23 secretion from peripheral blood mononuclear cells (PBMCs) after 48 h of coculture with Helicobacter pylori strains and dupA1) and isogenic dupA denote wild-type strains, and denote dupA The cytokine the medium were determined by enzyme-linked immunosorbent denote standard with PBMCs from the the type and the of Peripheral Cells (PBMCs) and IL-12p70 and of CD14+ after with the Helicobacter pylori and dupA Mutant