Association of cagA and vacA Genotypes of Helicobacter pylori with Gastric Diseases in Estonia

Abstract

Gastric biopsy specimens from 156 adult patients from southern Estonia suffering from chronic gastritis, peptic ulcer disease, and perforated peptic ulcer were analyzed by PCR. The cagA gene was evenly distributed throughout 87% of the specimens from the patients with the different gastric diseases. The presence of the cagA gene correlated with that of vacA signal sequence type s1a (99%). However, no clear differences were found in the distribution of cagA and vacA genotypes among patients in Estonia with severe perforated peptic ulcer, uncomplicated peptic ulcer, or chronic gastritis.

Helicobacter pylori colonizes approximately half of the world’s human population (6). It is considered the etiological agent of chronic gastritis (CG) and peptic ulcer and their complications (3, 14, 15, 21). In Estonia, antibodies against H. pylori have been detected in 80% of the inhabitants (12, 13, 30). Moreover, peptic ulcer disease (PUD) (both gastric and duodenal) is widespread in Estonia (13), and the incidence of perforated peptic ulcer (PPU) in particular is very high (21 of 100,000 cases) (19).

According to the literature, virulent cytotoxin VacA- and CagA-producing strains are more common among patients with peptic ulcer and gastric cancer (2, 4). Yet in many studies, patients with a variety of clinical symptoms of gastritis, gastric ulcer, duodenal ulcer, and reflux esophagitis have been examined for similar virulence markers of H. pylori (26, 27). It has not been resolved whether differences in the pathogenesis of this spectrum of diseases may be related to particular virulence markers differentially expressed by strains of H. pylori.

The aim of the present study was to compare the distributions of well-known virulence markers (cagA and vacA) in H. pylori strains isolated from southern Estonian patients suffering from CG, PUD, and PPU.

This study was carried out at Tartu University Clinic (Tartu, Estonia) between 1995 and 2000. Samples were collected from 156 adult patients from southern Estonia with CG, PUD, and PPU. Initially, 105 patients were investigated from 1 May 1995 to 30 April 2000, following referral to the Tartu University outpatient clinic for upper gastrointestinal endoscopy. In addition, samples were collected from 51 patients who were operated on for PPUs during the period from 1 January 1997 to 31 December 1999.

For the detection of CG and PUD, H. pylori strains were isolated from biopsy samples obtained by endoscopy. Upper endoscopy was performed with an Olympus GIF 21 gastroscope. Endoscopic diagnoses of PUD were made for 69 patients (duodenal ulcer in 61 patients and gastric ulcer in 8 patients), and diagnoses of no PUD were made for 36 patients, in whom CG was diagnosed histologically according to the Sydney classification method (16).

For H. pylori isolation, the biopsy samples from 93 patients with CG and PUD were placed into Stuart transport medium and taken to the laboratory within 2 h. H. pylori was isolated on a Columbia agar base supplemented with 7% horse blood and 1% IsoVitaleX. The plates were incubated for 3 to 4 days at 37°C under microaerophilic conditions (CampyGen; Oxoid). H. pylori was identified by Gram staining and by oxidase, catalase, and urease reactions. The H. pylori suspension in brucella broth was stored at −70°C until DNA extraction. For DNA isolation from H. pylori cells, the suspension was incubated at 100°C for 30 min and the lysates were stored at −20°C prior to PCR analysis. For patient samples (n = 12) for which attempts to cultivate H. pylori were unsuccessful, biopsy samples stored at −20°C in lysis buffer (200 mM Tris-HCl [pH 8.0], 25 mM EDTA, 300 mM NaCl, 1.2% sodium dodecyl sulfate) were used for PCR analysis.

From the 51 PPU patients, gastric specimens of the antral mucosa were obtained for molecular analysis intraoperatively through the perforation or postoperatively with panendoscopy. These biopsy samples were placed directly into the aforementioned lysis buffer. For DNA extraction from the frozen gastric biopsy specimens, we used a previously described procedure (20).

The presence in each strain of cagA and vacA was determined by PCR using primers, reaction mixtures, and thermal cycling (20). PCR products were identified by electrophoresis on 2% agarose gels.

Statistical analysis was performed by the two-tailed χ² test or by Fisher’s exact test. Significance was set at a P value of <0.05.

Among the 156 patients infected with H. pylori strains, no s1b strains were found. Multiple H. pylori strains were detected in 5 (3.2%) of the 156 patients studied; these included 3 patients with duodenal ulcer disease, 1 patient with CG, and 1 patient with PPU. The results from those patients were not included in the following analysis.

Among the 151 remaining investigated samples, the cagA gene was detected in 132 (87%) of them. Concerning vacA subtypes, the s1a/m1 allelic combination was the most prevalent (65%), whereas combinations of s1a/m2 and s2/m2 were found in 24 and 11% of the cases, respectively.

The presence of the cagA gene correlated with that of vacA signal sequence type s1a, whereas type s2 was predominantly found in cagA-negative strains (P < 0.001) (Table 1). The relationships between cagA status, vacA subtypes, and disease in patients are shown in Table 2. In CG and PUD patients, all cagA-negative isolates were associated with the s2/m2 genotype. In contrast, for the PPU group, four cagA-negative isolates exhibited the vacA genotype s1a and conversely, one cagA-positive isolate exhibited the vacA genotype s2. However, statistically the PPU group results do not differ from those of the CG and PUD groups (P > 0.05), with the results of all three demonstrating the same correlation between cagA positivity and vacA s1a type.

TABLE 1.

Relationships between vacA signal alleles and cagA status of 151 H. pylori clinical isolates (P < 0.001)

vacA subtype	No. (%) of isolates
	cagA+	cagA negative
s1a	131 (99)	4 (21)
s2	1 (1)	15 (79)
Total	132 (100)	19 (100)

TABLE 2.

Relationship between H. pylori cagA status, vacA subtypes, and disease of patient group (P > 0.05)

vacA subtype	No. (%) of isolates in patients with:
	CG (n = 35)		PUD (n = 66)		PPU (n = 50)
	cagA+ (n = 29)	cagA negative (n = 6)	cagA+ (n = 62)	cagA negative (n = 4)	cagA+ (n = 41)	cagA negative (n = 9)
s1a/m1	23 (79)	0	43 (69)	0	29 (71)	3 (33)
s1a/m2	6 (21)	0	19 (31)	0	11 (27)	1 (11)
s2/m2	0	6 (100)	0	4 (100)	1 (2)	5 (56)
Total	29 (100)	6 (100)	62 (100)	4 (100)	41 (100)	9 (100)

This paper provides a comparison of the distributions of H. pylori cagA status and vacA genotypes in selected Estonian patients with confirmed diagnoses of CG, PUD, and PPU. Different consensus criteria for the role of H. pylori have been set for populations with high and low prevalences of H. pylori infections (11). We attempted to determine whether the same set of criteria was valid for a population with a high prevalence of H. pylori. In Estonia, according to serological studies, H. pylori prevalences are 87% in adults and 56% in children (28, 30).

In a study by Vorobjova et al. (29), high prevalences of the CagA protein in sample populations of Estonian adults (63%) and children (46%) have been demonstrated. In the present study, we found that the prevalence of the cagA gene among the large sample of patients with various clinical diagnoses was very high (87%). This corresponds to the findings of a previous study, which described a high (82%) prevalence of the cagA gene in cases of clinically complicated PUD (20). In addition, the present study hints that the cagA gene of H. pylori and its defined protein are important predictors of several different gastric diseases, even in populations with a high prevalence of infection.

In countries with a low prevalence of cagA positivity (1, 2, 18), a significant correlation between vacA subtypes and cagA status has been demonstrated. In our study of Estonian patients, a significant association between vacA subtypes and cagA status was also found. In particular, 87% of the cagA-positive strains had the vacA s1a/m1/m2 subtype and only 1 strain out of the 131 cagA-positive strains carried the vacA s2/m2 subtype.

According to a large number of studies (1, 2, 4, 7, 24), the vacA s1a/m1 subtype is considered more cytotoxic than the s1a/m2 and s2/m2 vacA genotypes and is therefore related to the development of PUD. The s2 strains are rarely associated with PUD and are more common in cases of CG or nonulcer dyspepsia (5, 7, 18, 22, 25). In our study, there was no difference in colonization by H. pylori cytotoxic strains between patients with peptic ulcer and those with the more severe PPU. Furthermore, few H. pylori strains were of the s2 subtype (6% in PUD patients and 10% in PPU patients). It could be speculated that, in these patients, the patchy distribution of gastric mucosa obscures the s1a strain, whereas the s2 strains are cocolonizers of the mucosa. Such colonization of the gastric mucosa with several different H. pylori strains has been described previously in numerous reports (8, 10, 17, 23). The cagA-positive strains from the CG patient samples exclusively carried the s1a subtype of the vacA gene. Within this group, however, the proportion of cagA-negative and vacA s2 alleles was the highest (20%) of all the disease groups. Therefore, the unusual result may depend on the severity of gastritis in these patients.

One important finding regarding the PPU group is that a small number of patients (8%) were apparently colonized by a cagA-negative strain that simultaneously expressed the vacA s1a subtype. Consequently, the possibility cannot be excluded that additional virulence markers of H. pylori contribute to the development of PPU in some patients with PUD, indicating a role for microbial as well as host genetic diversity.

The present study has also confirmed the findings of a previous investigation on geographic differences in the distribution of H. pylori strains (20). In the extended sample of 151 patients examined in the present investigation, no s1b subtype was detected. Our findings are similar to those for a population in Asia (Helicobacter pylori Study Group, Abstr. 9th Int. Workshop Gastroduodenal Pathol. Helicobacter, GUT 39(Suppl. 2):A73–A74, 1996) and one in Poland, for which Gošciniak et al. (9) found only 1 of 72 strains with the s1b type. Findings very different from ours have been established in several studies conducted in The Netherlands, Portugal, Brazil, Mexico, and the United States (1, 5, 17, 24). Our findings relate to the widely disparate evolutions of circulating H. pylori strains in a particular geographic area and the subsequent necessity for area-specific diagnostics for H. pylori virulence markers.

In conclusion, we found a high prevalence of the cagA gene of H. pylori within the population of strains studied. However, no clear differences existed in the distributions of cagA and vacA genotypes in Estonia between patients with severe PPU, uncomplicated PUD, or CG.