Background

Helicobacter pylori colonization of the human stomach is associated with chronic gastritis and an increased risk of peptic ulcer disease (PUD), gastric adenocarcinoma and gastric mucosa-associated lymphoid tissue lymphoma [1–3]. While some H. pylori-infected individuals remain asymptomatic, others develop severe gastric disease. Strain-dependent factors may account for differences in clinical outcome, in particular factors that modulate interactions between H. pylori and human gastric cells, such as outer membrane proteins (OMP) [4–6]. Recently, the OMP coding gene homB was associated with an increased risk of PUD in Portuguese children and young adults (age < 40 years) [7]. Moreover, in vitro assays showed that HomB contributes to the proinflammatory characteristics of H. pylori and is involved in bacterial adherence, these two phenomena being more pronounced when homB is present in two copies in a given strain, compared to one copy only [7]. The homB 90%-similar gene, designated homA, was found to be associated with non-ulcer dyspepsia (NUD) in that same population [7].

In this study, we investigated gastric disease association with homB and homA, as well as the H. pylori virulence factors cagA, vacA and babA, in a panel of H. pylori clinical strains isolated from patients from East Asian and Western countries, presenting different gastric diseases, namely NUD and PUD. The correlation between those bacterial factors was also evaluated.

Results

The presence of homB and homA in the H. pylori clinical strains was determined by PCR. Table 1 summarizes the characteristics of the study population. PCR products, corresponding to either homA, homB or both genes were obtained for all the 415 strains tested. The presence of both genes in the same genome was detected in 43 strains (10.4%) (36 PUD strains and 7 NUD strains). These strains were excluded from the analysis related to clinical outcome. Thus, a total of 372 strains were included. They comprised 197 strains isolated from PUD patients (66.3% male; 50.3 ± 14.5 years) and 175 strains isolated from NUD patients (53.7% male; 51.1 ± 13.4 years).

Table 1 Distribution of Helicobacter pylori strains included in the study (n = 415), according to geographical origin and disease status of patients.
Full size table

Distribution of homB and homA according to clinical outcome

The results comparing PUD and NUD strains (n = 372) in different countries are presented in Fig. 1. Overall, homB was significantly more prevalent in PUD than in NUD strains (75.9 vs 64.7%, p = 0.026, OR = 1.7, 95%CI [1.06–2.74]), a trend also observed in Western strains (61.3 vs 46.3%, p = 0.025, OR = 1.84, 95%CI [1.10–3.10]). East Asian strains were predominantly homB-positive regardless of the clinical outcome (90.4% in PUD and 83.1% in NUD). Considering the analysis by country (Table 2), homB was associated with PUD in strains from France, Sweden, Brazil and Colombia, although with no statistical significance.

Figure 1
figure 1

Distribution of Helicobacter pylori genotypes among 372 strains isolated from East Asian and Western countries with regard to clinical outcome. PUD (peptic ulcer disease). NUD (non-ulcer dyspepsia). *p = 0.025; OR = 1.84, 95%CI [1.10–3.10]. #p = 0.014; OR = 2.2, 95%CI [1.20–3.86]. †p = 0.039; OR = 1.8, 95%CI [1.05–3.12].

Full size image
Table 2 Univariate analysis of the relationship between Helicobacter pylori virulence genotypes and clinical outcome according to country, from patients presenting peptic ulcer disease or non-ulcer dyspepsia.
Full size table

Inversely, the hom A gene was more prevalent in gastritis than in ulcer strains (35.8 vs 25.8%, p = 0.046, OR = 1.61, 95%CI [1.06–2.74]), a trend also observed in Western strains (54.6 vs 41.9%), though not significant (Fig. 1). The analysis by country revealed that homA was more frequently detected in strains isolated from NUD than from PUD in strains from France, Sweden, Brazil and Colombia, although the difference was not statistically significant (Table 2). Previously, it had been shown that homB was strongly associated with PUD strains isolated from young adults (age < 40 years) [7]. In the present study, a total of 90 strains were isolated from this age group. In this group, homB was significantly associated with PUD (n = 47, mean age 35.7 ± 5.8 y, 47.8% men) when compared to NUD strains (n = 43, mean age 33.4 ± 5.2 y, 55.9% men) (78.7 vs 48.8%; p = 0.006, OR = 3.88, 95% CI [1.41–10.84]). When considering only Western strains (31 PUD and 28 NUD), the same association was found (74.2 vs 35.7%; p = 0.007, OR = 5.18, 95% CI [1.49–18.68]), but not when East Asian strains only were considered (data not shown).

It was previously demonstrated that homB and homA can be present in a single- or two-copy form within a genome [7]. In the present study, the homB/homA copy number was determined for all 372 strains carrying homB or homA only. All of the East Asian strains carried the single-copy genotype, and this genotype was also the most frequently found in strains isolated in Portugal (60/100, 60%), France (23/33, 69.7%), Sweden (18/22, 81.8%), Germany (15/19, 78.9%) and Colombia (11/18, 61.1%). Due to the high prevalence of the single-copy genotype, no correlation was found between homB/homA copy number and clinical outcome in these populations. Regarding strains isolated in the USA, 52.2% (12/23) were found to carry the single-copy genotype, while the remaining carried the two-copy genotype, however the distribution of these genotypes was similar among PUD and NUD strains (data not shown). Finally, concerning strains from Brazil, the two-copy genotype was the most frequently detected (28/37, 75.8%), in both PUD and in NUD strains.

When considering the group of strains isolated from Western young adults (age < 40 years, 31 PUD and 28 NUD), a correlation was observed between copy-number of a specific gene and the clinical outcome. Thus, the homB two-copy genotype was the most frequently observed among PUD strains and the rarest genotype among NUD strains (38.7% vs 14.3%, p = 0.069), while the inverse situation was observed for the homA single-copy genotype (NUD vs PUD: 35.7% vs 9.7%, p = 0.036, OR = 5.19, 95%CI [1.09–27.87]).

Distribution of cagA, vacA s and babA according to clinical outcome

Considering all strains and similarly to homB, both cagA and vacA s1 were independently correlated with PUD (78.9 vs 65.0%, p = 0.014, OR = 2.0, 95%CI [1.16–3.47] for cagA; 76.8 vs 67.1%, p = 0.070, OR = 1.6, 95%CI [0.97–2.70] for vacA), a tendency also observed in Western strains (72.2 vs 54.6%, p = 0.014, OR = 2.2, 95%CI [1.20–3.86], and 71.0 vs 57.4%, p = 0.039, OR = 1.8, 95%CI [1.05–3.12], respectively). With regard to babA, it was only slightly more prevalent in PUD than in gastritis, considering all strains (58.1 vs 54.3%), and the Western strains (47.6 vs 40.7%). East Asian strains were all vacA s1, cagA and babA-positive. Considering the analysis by country, only the cagA-positive genotype was significantly associated with PUD in strains from Portugal (Table 2).

Association of homB and homA with cagA, vacA s and babA

The association of homB and hom A with the H. pylori-virulence genotypes cagA, vacA s1 and babA was also evaluated. Considering only the Western strains, the presence of homB was associated with cagA (p = 0.043) and vacA s1 (p < 0.001), while homA was more frequently found in strains lacking cagA (p = 0.062) and with the vacA s2 genotype (p < 0.001). The East Asian strains were all cagA-positive, babA-positive and vacA s1, among which 94.2% of the Japanese isolates and 78.5% of Korean strains were also homB-positive.

Discussion

Using a large panel of H. pylori strains (n = 372) isolated from patients from East Asian and Western countries, it was possible to confirm the association of homB with PUD and homA with NUD, previously observed with H. pylori strains (n = 84) isolated from Portuguese patients [7]. Considering the distribution according to geographical region, hom B was found to be significantly associated with PUD in strains from Western countries. However, when considering each country individually, only a tendency was observed probably due to the small number of strains tested in each case. The most common H. pylori strain circulating in the East Asia was extremely virulent, harboring homB, cagA, and babA genes and the vacA s1 genotype, regardless of the clinical outcome. Consequently no association with a specific disease was found, confirming the results from previous studies [8, 9]. Indeed, exposure to risk factors must be heterogeneous to find an association, and this is not the case in Asia. In addition, other environmental factors, e.g. diet, and possibly host genetic factors, may contribute to this evolution [10].

The previously reported significant association of homB with PUD in young adults (age < 40 years, 32 patients) [7] was confirmed in the present study with a higher number of patients (n = 90).

Polymorphism in copy number of H. pylori OMPs may contribute to increase the fitness of the strain and also its virulence [4, 11, 12]. Indeed, the homB two-copy genotype was associated with an increased rate of in vitro interleukin-8 secretion as well as an increased in vitro adherence [7]. Furthermore, it was the genotype most frequently found in strains from young adults with PUD, while homA single-copy was the most frequent in NUD strains, in agreement with previous data [7]. Globally, these data suggest that in some populations, the severity of H. pylori-associated disease in younger subjects may be closely related to the virulence of the strain, irrespective of the contribution of host and/or environmental factors which play a major role in adults. On the other hand, the present study demonstrates that there is a marked geographical specificity regarding homB/homA copy number, particularly evident between East Asian and Western strains, but also amongst Western countries, suggesting that copy number of the homB/homA OMP coding genes also plays a role in adaptation to the human host.

Several H. pylori genes encoding OMP display allelic variation, as is the case of babA, babB [13], hopQ [14] and hopZ [15]. In all of these cases, a conserved profile of gene segmentation is observed, with a variable region which defines the existence of at least two highly conserved allelic variants. Regarding homB and homA, no information on allelic variation is available to date. Further sequence analysis of these coding regions using H. pylori strains with different geographical background would allow assessing the existence of allelic variation and to evaluate whether different alleles could be associated with a specific clinical outcome and/or reflect a dissimilar geographical origin.

The cagA-positive and vacA s1 genotypes were independently associated with PUD in Western strains, but not babA. Previous publications reported a significant association between the presence of babA and PUD in Western strains [16, 17], contrasting with the present result. This discrepancy may be explained by the very heterogeneous Western study group with regard to the geographical origin of the strains, and also because of a possible absence of PCR amplification due to diversity within babA [17, 18].

homB was found to co-exist with the most virulent genotypes, while homA was more frequently found in strains lacking these genotypes, in agreement with previous results [7]. Thus, it is likely that the phenotype resulting from the expression of cagA, vacA s1 and homB genes confers a biological advantage to the strain, with the cumulative action of each factor contributing at the same time to the fitness of the strains in vivo and to a more pronounced pro-inflammatory response. Another hypothesis would be that homB is linked to PUD only because of its association with other virulence factors. However, its role in H. pylori-associated inflammation and in bacterial adherence supports the hypothesis that homB contributes to disease development [7].

Globally, these results suggest that homB and homA seem to be good candidates for the pool of H. pylori factors involved in the determination of clinical outcome.

Methods

Bacterial strains

A total of 415 H. pylori strains isolated from patients from 10 different countries, suffering from NUD (n = 182), PUD (n = 233), of which 197 duodenal ulcers and 36 gastric ulcers, were included in this study (Table 1). H. pylori strains were cultured from gastric biopsies on agar supplemented with 10% horse blood, preserved in trypticase soy broth supplemented with 20% glycerol and maintained at -80°C until used. Genomic DNA was extracted from a 48 h-old culture grown in agar base supplemented with 10% horse blood, using the QIAamp DNA mini kit (Qiagen GmbH, Hilden, Germany), according to the manufacturer's instructions.

Genotyping of homB, homA, cagA, vacA s and babA by PCR and sequencing

The homB and homA genes were amplified by a single PCR with a set of primers designed on a consensus internal sequence present in both genes [19]. In order to determine the homB and homA copy number, primers targeting the respective loci where used, as previously described [19]. The presence of the vacA s allelic variants, s1 and s2, and cagA and babA genes were determined using published PCR primers [13, 16, 20, 21].

Statistical analysis

Statistical analysis was performed using the statistical software package SPSS (version 14.0; SPSS). The level of significance was set at 5%, with the null hypothesis rejected when p < 0.05.