Association between Helicobacter pylori cagA-related genes and clinical outcomes in Colombia and Japan

Initially, candidate genes were selected from previous studies by Salama et al. [17] and Romo-González et al. [16]. Microarray data from 56 strains in our previous report was then used for the examination of the association of candidate virulence genes with the presence of cagA [18].

H. pylori strains were obtained from the gastric mucosa of H. pylori-infected patients who underwent endoscopy at Oita University Faculty of Medicine, Oita, Japan, and Universidad Nacional de Colombia, Bogota, Colombia. Presentations included gastritis, duodenal ulcer (DU), gastric ulcer (GU), and GC. DU, GU, and GC were identified by endoscopy, and GC was further confirmed by histopathology. Gastritis was defined as H. pylori gastritis in the absence of peptic ulcers or gastric malignancy. Patients with a history of partial gastric resection were excluded. Patients who received H. pylori eradication therapy or treatment with antibiotics, bismuth-containing compounds, H2-receptor blockers, or proton pump inhibitors within 4 weeks prior to the study were also excluded. Informed consent was obtained from all participants, and the protocol was approved by the ethics committees of Oita University and Universidad Nacional de Colombia.

Antral biopsy specimens were obtained for the isolation of H. pylori using standard culture methods, as previously described [19]. Chromosomal DNA was extracted from confluent plate cultures expanded from a single colony using a commercially available kit (QIAGEN Inc., Valencia, CA, USA). Two H. pylori strains with full-sequenced genomes, 26695 (ATCC 700392) and J99 (ATCC 700824) deposited in the GenBank, were used as control strains. The cagA status was determined by polymerase chain reaction (PCR) using primer pair 5'-ACC CTA GTC GGT AAT GGG-3' and 5'-GCT TTA GCT TCT GAY ACY GC-3' (Y = C+T), as described previously [20]. The vacA genotyping (s1, s2, m1, and m2) was performed by PCR, as described previously [21, 22]. Primers for the signal region yielded a fragment of 259 bp for s1 variants and one of 286 bp for s2 variants. Primers for the middle region yielded a fragment of 570 bp for m1 variants and one of 645 bp for m2 variants.

Two primer sets for each candidate gene were designed with software Primer 3 (version. 0.4.0) based on the published sequences of H. pylori (Table 1). Amplification of H. pylori genomic DNA sequences was carried out in a total volume of 25 μL containing 2.5 μL of PCR buffer, 0.2 mM of each deoxynucleotide, 0.625 U of Blend Taq DNA polymerase (Blend Taq, Toyobo Co., Ltd., Osaka, Japan), 0.2 μM of each primer, and more than 10 ng of H. pylori DNA. Each reaction mixture was amplified as follows: initial denaturation at 94°C for 5 min, which was followed by 30 cycles of denaturation at 94°C for 30 s, annealing at the indicated temperature in Table 1 for 30 s, extension at 72°C for 1 min, and then final extension at 72°C for 5 min. The amplified fragment was detected by 2.0% agarose gel electrophoresis using an ultraviolet transilluminator.

For each sample, 500 ng of total DNA was added to 100 μL of TE buffer and mixed with 100 μL of a denaturing buffer (0.5 M NaOH; 1.5 M NaCl). The denatured DNA was transferred to a Hybond-N⁺ membrane (GE HealthCare, Piscataway, NJ, USA) by means of a Bio-Dot Microfiltration Apparatus (Bio-Rad Laboratories, Inc., Hercules, CA, USA). DNA of J99 and human DNA were also transferred to the membrane and used as positive and negative controls, respectively. The membranes were hybridized at 42°C overnight in plastic bags containing ECL Gold hybridization buffer supplemented with 5% (wt/vol) blocking agent and 0.5 M NaCl. The membranes were washed 3 times in primary washing buffer (0.5× SSC [1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate] [pH 7.0], 0.4% sodium dodecyl sulfate) at room temperature for 15 min and 3 times in secondary washing buffer (2× SSC) at room temperature for 15 min. Finally, the membranes were exposed to Hyperfilm ECL film (GE HealthCare). Gene status was considered positive when at least one of the PCR reactions was positive. When gene status was considered negative by PCR, we further confirmed the results using dot-blot analyses. If PCR results yielded negative results but the dot blot showed a positive blot, we considered the samples positive.

DNA sequencing for the full length of jhp0045 (1,032 bp) was performed with several primer pairs located at jhp0044 and jhp0046. Likewise, DNA sequencing for the full length of jhp0046 (783 bp) was performed with several primer pairs located at jhp0045 and jhp0047. PCR products were purified with Centri-sep Columns (Applied Biosystems by Life Technologies, Tokyo, Japan), and the amplified fragments were sequenced with Hi-Di Formamide (Applied Biosystems by Life Technologies) using an ABI Prism 310 Genetic Analyzer (Applied Biosystems by Life Technologies, Carlsbad, CA, USA) in accordance with the manufacturer's instructions.

Variables such as gender, mean age, and the presence of each candidate gene and cagA were evaluated. The univariate association between each genotype and the clinical outcomes were quantified by the chi-square test. A multivariate logistic regression model was used to calculate the odds ratios (OR) of the clinical outcomes by including age, sex, and H. pylori genotypes. All determinants with P values of < 0.10 were entered together in the full model of logistic regression, and the model was reduced by excluding variables with P values of > 0.10. ORs and 95% confidence intervals (CIs) were used to estimate the risk. Spearman rank coefficients (r) were also determined to evaluate the association between the different genotypes of the strains. A P value of less than 0.05 was accepted as statistically significant. The SPSS statistical software package version 18.0 (IBM Corporation, Armonk, NY, USA) was used for all statistical analyses.

Twelve genes that were strongly associated with cagA status in the report by Salama et al. [17] were selected. In addition, 26 genes that were associated with severe gastric diseases in the report by Romo-González et al. [16] were selected. Because hp1426 was reported in both reports, a total of 37 genes were selected as candidate genes, as shown in additional file 1.

Among the 37 genes, the status of 9 genes (hp0186, hp0713, hp0967, hp1409, hp1410, jhp0045, jhp0046, jhp0950, and jhp0951) were significantly correlated with the cagA status in our microarray data [18] (P = 0.026, 0.026, 0.014, 0.048, 0.030, 0.033, 0.033, 0.017, and 0.005 for each above gene, respectively). Among the 9 genes, 4 genes (hp0967, jhp0045, jhp0046, and jhp0951) were selected in our analyses because the functions of these genes are known. The presence of 3 genes (hp0967, jhp0045, and jhp0046) was inversely correlated with the presence of cagA, but that of jhp0951 was positively correlated with the presence of cagA. We examined the presence of these candidate genes in 28 full sequenced strains deposited in Genbank. The hp0967, jhp0045, jhp0046, and jhp0951 were found in 18, 7, 7, and 13 strains, respectively.

The distribution of the status of cagA, vacA, hp0967, jhp0045, jhp0046, and jhp0951 in the two countries is shown in Table 2. Three samples from Colombia showed the positive for both vacA m1 and m2 genotypes, which suggest the mixed infection, were excluded in the final analysis. Finally, a total of 343 patients were included in this study: 174 from Colombia (68 with gastritis, 43 with DU, and 63 with GC) and 169 from Japan (49 with gastritis, 50 with DU, 50 with GU, and 20 with GC). The results from PCR and dot blot matched well: there were no cases with negative results by PCR and only positive results by dot blot. There were significant differences in the status of cagA, hp0967, jhp0045, and jhp0046 between strains isolated from Japanese and Colombian populations. The prevalence of cagA was 100% in Japan, whereas it was 65.5% in Colombia (P < 0.0001). Higher prevalences of vacA s1 and m1 genotypes were found in Japan than Colombia (100 vs. 76.8%, P < 0.001; 100 vs. 62.7%, < 0.001, respectively). The prevalence of hp0967 was significantly higher in Japan than Colombia (62.1 vs. 48.0%, P = 0.013). However, the prevalences of jhp0045 and jhp0046 were more prevalent in Colombia than Japan (23.7 vs. 8.9%, P < 0.0001; 28.2 vs. 8.9%, P < 0.0001, respectively). There was no difference in the prevalence of jhp0951 between the 2 countries.

The prevalence of each gene was examined according to clinical outcomes. The prevalence of the vacA m1 genotype was significantly higher in strains from patients with GC than those with gastritis (75.0 vs. 55.1%, P = 0.014) (Table 2). Although it is accepted that cagA is an important virulence factor, the prevalence of cagA was not different between the strains from patients with DU, GC, and gastritis in Colombia (61.4, 70.3, and 63.8%, P > 0.05), which was in agreement with our previous study [23]. Therefore, we hypothesized that the presence/absence of novel factors that accompany the presence of cagA leads to severe clinical outcomes in the Colombian population. Based on this hypothesis, the prevalences of these 4 candidate genes were examined in the cagA-positive cases. In cagA-positive cases, vacA status was not associated with clinical outcomes. Interestingly, the prevalence of jhp0045 in cagA-positive cases from GC was significantly higher than that of gastritis (30.4 vs. 11.4%, P = 0.023) (Table 3). The prevalence of jhp0046 in cagA-positive cases from GC also tended to be higher than that of gastritis (34.8 vs. 18.2%, P = 0.06), although this did not reach statistical significance. However, there were no associations of these candidates between gastritis and DU. Table 4 shows the association determined by a multivariate analysis between clinical outcomes and the presence of jhp0045 or jhp0046 in cagA-positive cases in the Colombian population. After adjustment for age and gender, jhp0045 was an independent factor for discriminating GC from gastritis in cagA-positive cases (adjusted OR = 3.24; 95% CI = 1.00-10.42; Table 4). Likewise, jhp0046 was an independent factor for discriminating GC from gastritis in cagA-positive cases (adjusted OR = 3.16; 95% CI = 1.05-9.47). On the other hand, the prevalences of hp0967 and jhp0951 in cagA-positive cases were not associated with clinical outcomes.

All samples from Japan showed cagA-positive and vacA s1m1. The prevalences of the 4 candidate genes in cagA-positive cases were independent of clinical outcomes (Table 5).

In order to clarify whether the sequence variants in jhp0045 and jhp0046 contributed to the different outcomes in Colombia and Japan, sequences of these 2 genes were compared using 8 randomly selected strains. For jhp0045, one-point mutation in the Japanese strains was found at 643-bp position of J99 (A643G). Therefore, the amino acid was changed from Ile to Val. The sequence of jhp0045 from the Colombian strains matched with J99. On the other hand, there was no difference in the sequence of jhp0046 between the strains from the 2 countries.

The nucleotide sequences of jhp0045 and jhp0046 for 8 strains (Japanese strains: 01-401, 04-156, 05-262, and 07-238; Colombian strains: Colombia 64, Colombia 114, Colombia 174, and Colombia 229) have been deposited in the GenBank database under accession no. AB647162 to AB647169 for jhp0045 and AB647170 to AB647176 for jhp0046, respectively.