Role of Helicobacter pylori cagA EPIYA motif and vacA genotypes for the development of gastrointestinal diseases in Southeast Asian countries: a meta-analysis

Background Infection with cagA-positive, cagA EPIYA motif ABD type, and vacA s1, m1, and i1 genotype strains of Helicobacter pylori is associated with an exacerbated inflammatory response and increased risk of gastroduodenal diseases. However, it is unclear whether the prevalence and virulence factor genotypes found in Southeast Asia are similar to those in Western countries. Here, we examined the cagA status and prevalence of cagA EPIYA motifs and vacA genotypes among H. pylori strains found in Southeast Asia and examined their association with gastroduodenal disease. Methods To determine the cagA status, cagA EPIYA motifs, and vacA genotypes of H. pylori, we conducted meta-analyses of 13 previous reports for 1,281 H. pylori strains detected from several Southeast Asian countries. Results The respective frequencies of cagA-positive and vacA s1, m1, and i1 genotypes among examined subjects were 93% (1,056/1,133), 98% (1,010/1,033), 58% (581/1,009), and 96% (248/259), respectively. Stratification showed significant variation in the frequencies of cagA status and vacA genotypes among countries and the individual races residing within each respective country. The frequency of the vacA m-region genotype in patients infected with East Asian-type strains differed significantly between the northern and southern areas of Vietnam (p < 0.001). Infection with vacA m1 type or cagA-positive strains was associated with an increased risk of peptic ulcer disease (odds ratio: 1.46, 95%CI: 1.01-2.12, p = 0.046 and 2.83, 1.50-5.34, p = 0.001, respectively) in the examined Southeast Asian populations. Conclusions Both Western- and East Asian-type strains of H. pylori are found in Southeast Asia and are predominantly cagA-positive and vacA s1 type. In Southeast Asia, patients infected with vacA m1 type or cagA-positive strains have an increased risk of peptic ulcer disease. Thus, testing for this genotype and the presence of cagA may have clinical usefulness.


Background
Helicobacter pylori infection is closely associated with the development of peptic ulcer disease and gastric cancer [1][2][3][4]. H. pylori colonizes the gastric mucosa of 50% of the world's population, with infection levels exceeding 70% in developing areas, such as Latin America and Africa [5][6][7][8][9][10]. Several Southeast Asian countries also have a high prevalence of H. pylori infection, particularly Thailand, where the H. pylori infection rate ranges from 54.1%-76.1% [11,12]. Therefore, developing countries, including those located in Southeast Asia, are considered to have a higher incidence of H. pylori-related diseases. However, the estimated age-standardized incidence rate of gastric cancer in Southeast Asian countries (10.2 and 4.7/100,000 in men and women, respectively) is lower than those in East Asia (e.g., 124.63/100,000 in Japan and 48.25/100,000 in Korea) and South America (e.g., 33.22/100,000 in Colombia) (http://globocan.iarc.fr/). This phenomenon, which is characterized by high levels of H. pylori infection, but a low incidence of gastric cancer, is referred to as the "Asian paradox" [13,14].
Recently, numerous studies have examined the relationships between H. pylori virulence factors and outer membrane proteins with gastric mucosal inflammation and gastroduodenal disease development [9,10,15,16]. The cag PAI including of cagA encodes a putative type IV secretion system, which transfers a variety of multimolecular complexes, such as CagA, and across the bacterial membrane to the extracellular space or into other attached host cells [17,18]. H. pylori cagA plays important roles in gastric mucosal inflammation and injury in relation to activated inflammatory cells infiltration [19,20]. Activated neutrophils and mononuclear cells infiltrating into gastric mucosa with H. pylori infection produce several pro-inflammatory cytokines (e.g., IL-1β, IL-6, IL-8 and TNF-α) and anti-inflammatory cytokines (e.g., IL-4 and IL-10). Greater than 90% of H. pylori strains isolated from East Asian populations carries the cagA gene [7,21]. In contrast, 40% of strains isolated in Western countries are cagA-negative [7,21]. Moreover, as tyrosine phosphorylation of CagA occurs at EPIYA sites consisting of five amino acid residues, the structure and role of the cagA EPIYA motif in cagA gene have been investigated [22][23][24]. The EPIYA motif exhibits genetic variation that occurs in four distinct segments, the EPIYA-A, -B, -C, and -D segments [22][23][24]. The representative CagA of Western H. pylori strains possesses a single EPIYA-A and EPIYA-B segment, followed by a 34-amino-acid EPIYA-C segment (EPIYA-ABC type). The C-terminal regions of East Asian and Western CagA are characterized by the presence of EPIYA-ABD and -ABC segments, respectively [22][23][24]. In addition, we [7] reported that H. pylori strains isolated from East Asian and Western countries could be completely distinguished by polymerase chain reaction-based cagA 5' and 3' region genotyping, and named the East Asian type H. pylori strains as cagA 1a type and the Western type H. pylori strains as cagA 2a type. CagA 1a type is correspondent roughly to Western-type cagA (EPIYA-ABC) genotype and 2a type to East Asian-type cagA (EPIYA-ABD) genotype. However, it is unclear whether the cagA EPIYA motif is associated with the development of gastrointestinal disease in Southeast Asian populations.
Gastric epithelial cell injury associated with H. pylori infections is caused by a vacuolating cytotoxin encoded by the H. pylori vacA gene. The vacA signal (s) region encodes the signal peptide and N-terminus of processed vacA toxin. In H. pylori, the vacA s1 genotype is associated with fully active toxin, but type 2 genotype strains produce vacA with a short N-terminal extension that blocks vacuole formation [25]. The vacA middle (m) region encodes part of the 55-KDa subunit located at the C-terminus and has two genotypes (m1 and m2); the former causes stronger vacuolating activities than the latter [25]. Recently, a third polymorphic determinant of vacuolating activity located between the s-and m-regions was identified and termed the intermediate (i) region [26]. In general, the vacA s1, m1, and i1 genotypes of H. pylori are associated with an increased risk of disease due to the enhanced production of toxin with markedly higher vacuolating activity than that of vacA s2, m2, and i2 genotype strains, which are rarely associated with peptic ulcer disease and gastric cancer [7,8,15,25,[27][28][29][30]. However, it is unclear whether this association is observed in H. pylori strains commonly found in Southeast Asian populations.

Study selection
A literature search was performed using the PubMed databases for articles written in English and published before June 2011. The following search words were used: 1) cagA or EPIYA, 2) vacA or vacuolating cytotoxin, 3) pylori or Helicobacter, and 4) genotype. We did not include abstracts or unpublished articles. We conducted a combined analysis to determine the prevalence of cagA-positive strains, cagA EPIYA motif genotypes, and vacA genotypes in H. pylori strains found in Southeast Asia and their association with gastrointestinal diseases.

Inclusion Criteria
The following criteria were applied to select published case-control studies examining the relationship between cagA EPIYA motif genotype or vacA genotype and clinical outcomes in adult populations infected with H. pylori isolated in four Southeast Asian countries (Vietnam, Thailand, Singapore, and Malaysia): the presence and genotypes of the cagA EPIYA motif (ABC, ABCC, and ABD genotypes) and vacA (vacA s-, m-, and i-regions) were examined by polymerase chain reaction (PCR) and original articles published in English. The references cited in these manuscripts were also screened using the same inclusion criteria. When it appeared that the same subjects were included in multiple reports, only the earliest article was selected.

Data analysis
As several studies did not measure all parameters simultaneously (cagA status and vacA s-, m-, and i-region genotypes), the patient and strain numbers (H. pylori genotype number) did not match in the following analyses. Statistical differences in the prevalence of cagA status, EPIYA motif genotype, and vacA genotype among the individual countries and ethnic groups were determined by the chi-squared test and Fisher's exact test. The effects of cagA status, EPIYA motif genotype, and vacA genotypes on the risk of gastric cancer and peptic ulcer disease were expressed as odds ratios (ORs) with 95% confidence intervals (CIs) with reference to non-ulcer dyspepsia (NUD) subjects infected with H. pylori. NUD was defined as endoscopical gastritis with no peptic ulcer disease or gastric cancer. In this study, patients with 'NUD' were regarded as the control group. All p values were two-sided, and p values <0.05 were considered statistically significant. Calculations were carried out using the statistical software StatView 5.0 (SAS Institute, Cary, NC, USA). Meta-analyses were performed using Comprehensive Meta-Analysis software (version 2, Biostat, Englewood, NJ).

Included studies
Seventeen studies investigating Southeast Asian populations were identified using our search criteria (Table 1). Notably, a report from Yamaoka et al. [7] included two populations: Vietnamese and Thai. Van Doorn et al. [43] included strains isolated from Southeast Asia, but did not report the country of origin. Studies by Ho et al. [44] and Vivatvakin et al. [45] did not provide detailed vacA genotype information, and a report from Yamaoka Y et al. [46] investigated Vietnamese living in Houston, Texas. Finally, a total of 13 studies with a combined 1,281 patients (4 reports for Vietnam, 4 from Thailand, 5 from Malaysia, and 1 from Singapore) were included in the systematic analyses (Tables 1 and 2).
Prevalence of cagA and vacA s, m, and i-region genotypes In H. pylori strains isolated from Southeast Asia, the respective frequencies of vacA s1, m1, and i1 genotypes among the examined subjects were 98% (1,010/1,033), 58% (581/1,009), and 96% (248/259). The prevalence of the vacA m-region and combined vacA s/m genotypes differed significantly among isolates from the various Southeast Asian countries (p < 0.001) ( Table 2 and Figure 1). However, no significant difference between the prevalence of vacA s-region genotypes was detected. The frequency of the vacA s2m1 combination genotype was close to 0%, which was in agreement with most previous studies examining other geographical regions [9,25]. Interestingly, the frequency of sub-genotypes of vacA s1, s1a and s1c, which are reported to be specific for East Asian H. pylori strains [7], were similar in Thailand and Malaysia (Table 2). Reports from Nguyen TL, et al. [33] and Schmidt HMA, et al. [41] investigated vacA i-region genotypes; however, the correlation with vacA i-region genotypes at the population level was unclear.
Of the 13 included studies, 93% of H. pylori strains were cagA positive. The strains were divided into the following two groups: East Asian strains, including the EPIYA-ABD genotype [22] and 1a type according to PCR-based cagA 5' and 3' region genotyping [7], and Western-type strains, including EPIYA-ABC and -ABCC genotypes and 2a type. The prevalence of East Asianand Western-strain type cagA was 65% and 35%, respectively. In the Vietnam population, 97% (139/144) of strains was East-Asian strain type cagA, whereas H. pylori strains in Thailand and Malaysia consisted of approximately half of East Asian-and Western-type strains ( Table 2).
Three studies compared H. pylori strains isolated from the northern (Hanoi) and southern parts (Ho Chi Minh) of Vietnam [31][32][33] (Tables 1 and 3). Although the prevalence of the vacA s1 genotype was identical between both areas (100%), the vacA m1 genotype was found at a significantly higher frequency in strains isolated from the north (58%; 60/104) than from strains detected in the south (34%, 32/94; p = 0.001). Therefore, the virulence of H. pylori in Vietnam may differ between the northern and southern areas of the country, despite having the same native population.
Thailand and Malaysia consist of individuals from neighboring populations. In Thailand, although the  (48) 117 (52) 106 (48) 117 (52) (4) 'NA' (not available) indicates a lack of no data associating cagA status, EPIYA motif genotype, and vacA s-, m-, and i-region genotypes. As a report from Yamaoka Y, et al. included two populations: Vietnamese and Thai [7], the number of studies was reduced by one. *: p < 0.05 (significant differences among different countries, when compared with Vietnam.) by use of chi-square test.
prevalence of the vacA m1 genotype was higher (64%) than that of the m2 genotype (36%), vacA m-region genotypes differed among Thai, Thai-Chinese, and Chinese populations. The prevalence of the vacA m1 genotype among native Thai, who represent the majority of the population (approximately 75%), was 92%, which was significantly higher than that in Chinese (40%, p < 0.001) ( Table 3). In the analysis of the vacA s1 subtype, most strains isolated from Chinese patients were the s1c subtype (75%), which was significantly higher than that found in Thai and Thai-Chinese patients (p < 0.001). Moreover, the cagA EPIYA motif genotype significantly differed among the examined geographic populations (p < 0.001).
Although East Asian-type cagA was predominant in Chinese living in Thailand, Western-type cagA was predominant in Thai patients, suggesting that most native Thai living in Thailand are infected with Western-type H. pylori. Therefore, this observation suggests that although H. pylori infections in Thai are predominantly due to vacA s1m1 type strains, which have high virulence, the strains possess Western-type cagA, which is associated with weak virulence compared with East Asian strains [47]. In Malaysia, the prevalence of vacA m1 genotypes was similar among Malay (64%), Chinese (55%), and Indian (70%) patients (p = 0.065) ( Table 3). Most strains isolated from Indian and Malay patients were of the s1a subtype (82% and 92%, respectively). In the analysis of cagA EPIYA genotype, Chinese patients were predominantly infected with East Asian-type cagA (EPIYA-ABD), whereas Indian patients were mostly infected with Western-type cagA (EPIYA-ABC). The incidence of Western-type cagA was slightly higher in Malay patients, who formed the majority of the population. The examination of strains isolated from different ethnic groups provided an opportunity to examine the hypothesis that ethnic groups tend to retain their ancestral H. pylori genotype after migration [7,46].   'NA' (not available) indicates a lack of data associating cagA status, EPIYA motif genotype, and vacA s-, m-, and i-region genotypes. *: p < 0.05 (significant differences between Northern and Southern parts in Vietnam) by use of chi-square test, #: p < 0.05 (significant differences among Thai, Thai-Chinese, and Chinese in Thailand) by use of chi-square test, **: p < 0.05 (significant differences among Chinese, Indian, and Malay in Malaysia) by use of chi-square test.  Intermarriage among ethnic Thai and Chinese in Thailand also allowed examination of the transmission pathways in relation to the ethnic status of the parents. When the mother was Chinese, 84% of the offspring possessed H. pylori with an East Asian genotype. By contrast, when mother was Thai and father Chinese, only 29% of strains among offspring were of East Asian types. This finding is consistent with the notion that the primary caretaker of the children is the most likely source of infection. Taken together, these observations suggest that the susceptibility of different geographic populations to H. pylori strains and H. pylori-associated disease differed.
Risk of gastric cancer and peptic ulcer disease development associated with cagA status and vacA s, m, and i genotypes Of the 13 examined studies, 6 investigated the associations between gastroduodenal diseases and vacA s-and m-region genotypes [33,[35][36][37]41,42], and 2 investigated disease associations based on the vacA i-region genotype [33,41].

Discussion
Highly virulent genotypes of H. pylori are associated with gastric epithelial damage, including gastric mucosal atrophy, in infected patients [30]. Notably, the vacA s1, m1, and i1 genotypes and cagA-positive strains of H. pylori are linked to elevated inflammatory cell infiltration compared to that induced by vacA s2, m2, and i2 genotypes and cagA-negative strains [48,49]. In the present study, our meta-analyses revealed a significant relationship between H. pylori virulence factor-associated genotypes, particularly vacA m-region genotype and cagA status, and an increased risk for the development of peptic ulcer disease in Southeast Asian populations.
An important characteristic of Southeast Asian countries is that several geographic populations live together. Recently, Breurec et al. [50] reported that five major types of historical human migration patterns have occurred in Southeast Asia: i) migration from India introducing hpEurope bacteria into Thailand, Cambodia, and Malaysia; ii) migration of the ancestors of Austro-Asiatic speaking people carrying hspEAsia bacteria into Vietnam and Cambodia; and iii) migration of the ancestors of the Thai people from Southern China into Thailand carrying H. pylori of population hpAsia2; iv) migration of Chinese to Thailand and Malaysia within the last 200 years resulting in the spread of hspEAsia strains; and v) migration of Indians to Malaysia within the last 200 years distributing both hpAsia2 and hpEurope bacteria. Therefore, both Western and East Asian strains according to EPIYA motif genotyping can be observed in Southeast Asia. East Asian-type strain (EPIYA-ABD) was associated with an approximately three-fold increased risk of disease. Therefore, we concluded that the risk of developing H. pyloriassociated diseases varies among different geographic populations despite living simultaneously in the same country.
The "Asian paradox" might be explained by the widespread prevalence of weakly cytotoxic strains and correspondingly low frequency of H. pylori-associated diseases. However, here, the prevalence of strains with the vacA s1 region among Southeast Asia populations was similar to that of East Asian and Latin American populations, who have a higher risk of gastric carcinogenesis [7,9,10]; half of the isolated H. pylori strains were Western-type strains with an EPIYA-ABC genotype. The sequence of the EPIYA-D type perfectly matches the high-affinity binding sequence for SHP2 domains of SHP-2 (pY-[V/T/A/I/S]-X-[L/I/V]-X-[F/W]) [51,52]. In contrast, the sequences of the EPIYA-C type differ from the SHP-2 binding sequence by a single amino acid located in the pY + 5th position [51,52]. As a result, East Asian-type CagA exhibits stronger binding activity for SHP-2 and a greater ability to induce morphological changes in epithelial cells than the Western type. The estimated age-standardized incidence rates of gastric cancer in Vietnamese patients (24.4 and 14.6/100,000 in men and women, respectively) infected with East Asian-strains of H. pylori are relatively higher than those in Thailand (4.2 and 3.0/100,000 in men and women, respectively) and Malaysia (10.7 and 6.4/100,000), where the majority of the population is Thai and Malay, and are infected with Western-type strains (http://globocan.iarc.fr/). Therefore, the lower prevalence of higher virulence strains (e.g., ABD type) may explain the low frequency of gastric cancer observed in Southeast Asia, particularly in the southern part.
In Vietnam, H. pylori infection was detected at a rate of approximately 66% and was shown to be strongly associated with active gastritis, atrophy and intestinal metaplasia [33]. As genetic characteristics of H. pylori in Vietnam, the majority of isolates possess cagA, oipA "on", vacA s1, and vacA i1 genes, while the incidence of vacA m1 gene is less frequent [7,31,33]. Here, the prevalence of the vacA m1 genotype in patients residing in Hanoi and Ho Chi Minh were 58% and 34%, respectively. The prevalence of peptic ulcer disease in Hanoi, located in the northern part of Vietnam, is significantly higher than that in the southern city of Ho Chi Minh, despite similarities of ethnicity and diet [33]. Moreover, the age-standardized incidence rate of gastric cancer in Hanoi was approximately 1.5-fold higher than that in Ho Chi Minh (27.0 vs. 18.7 cases per 100,000 males, respectively) [33]. The higher prevalence of peptic ulcer disease and gastric cancer observed in Hanoi might be attributable to the higher prevalence of H. pylori strains carrying vacA m1 in this region. Although Vietnam is located between regions with a high and low risk of gastric cancer, the rate of gastric cancer in Vietnam is approximately three times lower than that in Japan and Korea, a finding that might also explain the low prevalence of H. pylori strains carrying vacA m1 (58% vs. nearly 100% in East Asia) [7,16,53].
Thailand has a population of 60 million people and is comprised of two major ethnic groups: Thai and Chinese. Eight types of H. pylori strains were identified in Thailand; half of the strains possessed genotypes typically found in South Asia (cagA type 2a/cag right-end junction type III and vacA s1a-m1c), while the other half consisted of genotypes typically encountered in East Asia (cagA type 1a/cag right-end junction type II and vacA s1c-m1b) based on combinations of cagA EPIYA motif genotype, genotype at the right end of the cag pathogenicity island into five subtypes according to deletion, insertion and substitution motifs (e.g., Type II strains were predominant in China and Japan and type III strains were most common in India and type III strains were typical of strains from India) and vacA s-region genotype [7,54]. Thailand was considered to be a cross roads with respect to H. pylori genotypes, as isolates from ethnic Thai were commonly either South/Central Asian or mixed genotypes (East Asian and South/Central Asian) based on the cagA, cagPAI, and vacA genes, whereas isolates from ethnic Chinese were typically the East-Asian genotype [55]. Although the incidence of gastric cancer in Thailand is lower than that of East-Asian populations, the incidence of gastric cancer is not rare among Chinese living in Thailand, as the majority of gastric cancers (82%) occurred among ethnic Chinese or Thai-Chinese. Therefore, the incidence of gastric cancer varies among different geographic populations, and the high prevalence of higher-virulence strains may be associated with the high frequency of gastric cancer in Chinese residing in Thailand, which is typically a lowincidence area.
In Malaysia, three distinct ethnic groups predominate: Malays, Chinese, and Indians. Chinese and Indians have migrated to Malaysia for nearly three consecutive generations. The vacA s1c genotype was also the predominant genotype detected among the Chinese patients residing in Malaysia, while s1a was predominant in Indians and Malays in Kuala Lumpur [37].
The present study has some limitations. First, in this study, because the prevalence of cagA-negative strains was very low in Southeast Asian countries, we did not demonstrate the relationship between risk of H. pylorirelated diseases and cagA-negative strains, which are 20-80% of H. pylori strains isolated from US and European population. Second, there was no information about biochemical role among different H. pylori virulent factors (e.g., cagA status, cagA EPIYA motif and vacA genotypes) in pathogenesis for peptic ulcer disease and gastric cancer development [56]. Recently, Mueller, et al. [56] reported association of c-Src and c-Abl kinases and phosphorylation of EPIYA motif, and differences of phosphorylation between East Asian-type strain (EPIYA-ABD) and Western type strain (EPIYA-ABC). Therefore, further studies will be required to clarify the role of H. pylori cagA EPIYA motif and vacA genotype for the development of gastrointestinal diseases.

Conclusion
We demonstrated that the prevalence of specific vacA m-region and cagA EPIYA motif genotypes was found to vary significantly among the respective Southeast Asian countries. Moreover, our present meta-analyses identified a significant relationship between vacA m-region genotype and cagA status and the development of diseases in Southeast Asian. Importantly, most of the H. pylori strains isolated from countries with high incidences of gastric cancer concurrently possessed virulent genotypes such as vacA s1/m1 and East Asian-type cagA [7]. In contrast, in countries with a low gastric cancer incidence, such as Thailand and Malaysia, a considerable proportion of H. pylori isolates exhibited less virulent genotypes, such as vacA m2 and Western-type cagA [7]. Based on the age-standardized incidence rate of gastric cancer, Asian countries can be categorized as either high-risk (e.g., Japan, Korea, and China), intermediate-risk (e.g., Vietnam), or low-risk (e.g., Thailand and Indonesia). The clinical usefulness of cagA EPIYA motif and vacA genotype testing must be evaluated in studies using a large number of individuals with cagA EPIYA motif and vacA genotypes.