Helicobacter pylori cagA, vacA, iceA and babA Genotypes from Peruvian Patients with Gastric Intestinal Metaplasia

Simple Summary Virulence factor genes in Helicobacter pylori (H. pylori) strains promote changes in the gastric epithelial mucosa that are associated with the risk of developing neoplastic lesions, and in Peru, H. pylori infection prevalence is greater than 45% and the gastric cancer incidence rate is the highest in the Americas. The aim of our cross-sectional study was to explore the association between clinical strain virulence genotypes of H. pylori from Peruvian patients with gastric intestinal metaplasia compared to chronic non-atrophic gastritis cases. We observed that the prevalence of the genotypes cagA+/vacAs1m1 and cagA+/vacAs1am1 was 2.42 and 1.67 times higher in cases with intestinal metaplasia compared to chronic non-atrophic gastritis cases, respectively. Our findings revealed that H. pylori strains circulating in our environment have a higher frequency of genotypes documented as risk variants for neoplastic lesions, highlighting that in Peruvian patients with H. pylori infection, the risk genotypes are related to intestinal metaplasia clinical stage. Abstract We explored the clinical-stage association of gastric intestinal metaplasia (IM) compared to cases of chronic non-atrophic gastritis (CNAG) and its relationship with virulence genotypes of Helicobacter pylori (H. pylori) clinical isolates from patients with dyspepsia in Peru. This study was cross-sectional and included 158 H. pylori clinical isolates; each isolate corresponded to a different Peruvian patient, genotyped by polymerase chain reaction to detect cagA gene and EPIYA motifs, the vacA gene (alleles s1, s2, i1, i2, d1, d2, m1, m2 and subtypes s1a, s1b and s1c), the iceA gene (alleles 1 and 2), and the babA gene (allele 2). We observed that 38.6% presented with IM and that all clinical isolates were CagA positive. The EPIYA-ABC motif was predominant (68.4%), and we observed a high frequency for the vacA gene alleles s1 (94.9%), m1 (81.7%), i1 (63.9%), and d1 (70.9%). Strains with both iceA alleles were also detected (69.6%) and 52.2% were babA2 positive. In addition, it was observed that the cagA+/vacAs1m1 (PR: 2.42, 1.14 to 5.13, p < 0.05) and cagA+/vacAs1am1 (PR: 1.67, 1.13 to 2.45, p < 0.01) genotypes were associated with IM. Our findings revealed the cagA and vacA risk genotypes predominance, and we provided clinically relevant associations between Peruvian patients with H. pylori infection and IM clinical stage.


Introduction
Gastric cancer (GC) is the sixth most common cause of death due to malignancy worldwide, and Helicobacter pylori (H.pylori) infection is the GC leading cause among infection-related cancer diagnoses.In the Americas, Peru has the GC highest incidence for both sexes (age standardized rate per-100,000 people: 10.9), followed by Colombia (10.2),Chile (9.8) and Haiti (9.7) [1][2][3].GC in Peru represents the second leading cause of death by malignancy [4], H. pylori infection prevalence ranges from 45% to 83%, and it is estimated that 60.8% of patients with a GC diagnosis are infected by H. pylori [5,6].
Gastric adenocarcinoma development is initiated by a series of changes in the gastric epithelial mucosa with highly differentiated patterns, which is described as the Correa Cascade [7,8], and it is known that persistent infection and the presence of genes related to virulence factors in H. pylori are associated with the risk of developing preneoplastic lesions.However, not all patients with chronic infection develop severe lesions on the gastric epithelium [9,10].
Genetic variability in genes encoding H. pylori virulence factors may modulate risk for the development of malignant lesions [11].Gene-A-associated cytotoxin is a protein encoded by the cagA gene and can translocate by means of a type IV secretion system and induce a pro-inflammatory response in gastric epithelial cells that may promote precancerous and cancerous lesions.Vacuolating cytotoxin A gene encodes the vacA cytotoxin and can induce an acute inflammatory response in the gastric epithelium, generating important promoter activity for the development of neoplastic lesions [12,13].In addition, virulence factors linked to H. pylori colonization and adhesion on the gastric epithelial mucosa, such as the epithelial-contact-induced gene (iceA) and the blood group antigen-binding adhesin gene (babA), are associated with the presence of ulcerative or atrophic gastric lesions, promoting the risk of preneoplastic lesions [14][15][16].
Each geographic region has its own profile of genetic variants based on the distribution of virulence factors among its circulating strains, and in Peru, a variable frequency of H. pylori strains positive for the cagA gene (79.9% to 100%) and in the detection of genotypes s1 (41.6% to 100%) and m1 (60.7% to 75%) genotypes of the vacA gene has been reported [17,18], In addition, a frequency of 33.3% for the vacAs1m1 genotype was found in cases with a diagnosis of GC and H. pylori infection [6].
Despite the high incidence of CG and the high frequency of H. pylori infection in Peru, information on the virulence genotypes of H. pylori strains is still scarce and little is known about their relationship with the clinical stage of Peruvian patients.Therefore, the aim of this study was to determine the association between genes associated with H. pylori virulence factors [cagA status, EPIYA motifs, vacA gene polymorphisms (s, m, i and d), iceA gene alleles 1 and 2, and babA gene allele 2] and the clinical stage of preneoplastic lesion gastric intestinal metaplasia observed in a sample of Peruvian patients.

Study Population and Design
This study was descriptive, cross-sectional, and included 158 H. pylori clinical isolates from adult outpatients with dyspepsia referred for endoscopy at the gastroenterology service of a university hospital in Lima, Peru, and diagnosed with H. pylori infection for the first time (between March 2016 and August 2017).The strains were obtained from a previous study conducted by researchers at the Helicobacter pylori Sentinel Laboratory at the Alexander von Humboldt Institute of Tropical Medicine in Lima, Peru.The protocol was approved by the Institutional Human Ethics Committee of the Universidad Peruana Cayetano Heredia (File 059-03-20 with registration code 221182).Strains were isolated from gastric biopsies obtained according to the Maastricht protocol [19], and each isolate was from a different patient.
Of the patients from whom the strains were isolated, the histopathological report was obtained, and the cases were classified into clinical stages of gastric intestinal metaplasia (IM) (N = 61, 38.6%) and chronic non-atrophic gastritis (CNAG) (N = 97, 61.4%).Histopathological analysis was performed using the Sydney system by hematoxylin-eosin staining of gastric samples [20].The sample size was estimated by power calculation based on a pilot trial with a subsample of the analyzed cohort, where a prevalence of cases with IM-presenting isolates with the cagA/vacAs1am1 genotype of 58.3% and a prevalence of 34.5% of cases with other IM-presenting genotypes were considered.The POWER command in Stata/SE version 17.0 software was used to determine the statistical power, resulting in a power of 84.3%.

Statistical Analysis
The association between the preneoplastic lesion types of the gastric epithelial mucosa and the genotype of H. pylori isolates was evaluated through a bivariate analysis using the chi-squared test, and in cases of violation of assumptions, Fisher's exact test was used.A binomial-log generalized linear regression analysis was performed, and the magnitude of the association was expressed as a prevalence ratio (PR) as an estimate relative to the risk with a 95% confidence interval (95% CI) regression.A statistical significance level was considered for values of p < 0.05.All analyses were performed using the statistical package Stata/SE version 17.0.

Patient Demographics
In the population of 158 cases of dyspepsia with H. pylori infection, a higher frequency of infection was observed in the female sex (70.3%) and in the age group over 50 years (58.8%).In addition, 38.6% of cases were histopathologically diagnosed as IM, with a higher frequency of IM in the age group over 50 years (p < 0.05) (Table 2).

cagA Genotypes and EPIYA Motifs
All isolates were positive for the cagA gene, and the EPIYA motifs of the cagA protein corresponded to the EPIYA-A, EPIYA-B, and EPIYA-C segments, where there were single and multiple cagA-EPIYA genotypes and a predominance for the cagA-ABC genotype.It was observed that 114 cases had no co-infection (6 EPIYA-AB cases (3.8%); 108 EPIYA-ABC cases (68.4%)).In addition, among the cases with co-infection, 33 cases were observed with a dual EPIYA genotype with AB/ABC (3.2%) and ABC/ABCC (17.7%) profiles and 11 cases with a triple EPIYA genotype (ABC/ABCC/ABCCC) (Table 3).

vacA Genotypes
A high frequency was observed for the vacAs1 allele (94.9%) and the vacAs1a subtype (48.1%) of the signal region of the gene.The findings in the middle region, intermediate region, and deletion of the intermediate region of the vacA gene showed a high frequency for the genotypes vacAm1 (81.7%), vacAi1 (63.9%), and vacAd1 (70.9%).It was also observed that the vacAs1 subtypes (vacAs1a genotype: 43.4%), the middle region (vacAm1 genotype: 43.4%), and the intermediate region (vacAi1 genotype: 47.5%) were statistically significant regarding IM diagnosis.In addition, the subtyping of the vacA gene signal region revealed co-infection with multiple genotypes (s1a/s1b, s1a/s1c, and s1a/s1b/s1c) (Table 3).

iceA and babA2 Genotypes
H. pylori strains presented both alleles for the iceA gene in 69.6% of cases, and 52.5% of isolates were positive for the babA2 gene.In addition, H. pylori strains were positive for allele 1 of the iceA gene in 51.5% of cases with IM, which was statistically significant (Table 3).

Relation of Virulence Genotypes with IM
The EPIYA-ABC motif was predominant, and EPIYA-C segment duplication was statistically associated with and prevalent in IM with respect to the detection of a single EPIYA-C segment (prevalence ratio (PR): 1.56, 95% confidence interval (95% CI): 1.02 to 2.41, p < 0.05).In relation to the vacA gene, m1 and i1 alleles had significant associations with IM with respect to m2 and i2 polymorphisms (PR: 2.52, 95% CI: 1.11 to 5.74, p < 0.05 for m1; PR: 2.32, 95% CI: 1.25 to 4.32, p < 0.01 for i1).Clinical isolates carrying both alleles (allele 1 and allele 2) of the iceA gene were also observed and were significantly associated with a lower frequency of IM cases than iceA2 clinical isolates (PR: 0.46, 95% CI: 0.29 to 0.73, p < 0.01).The allele 2 of the babA gene did not show a significant association with clinical-stage IM.In addition, it was observed that the co-detection of both the positive status of the cagA gene and the presence of s1, s1a, and m1 polymorphisms of the vacA gene were statistically associated with and prevalent in IM with respect to other documented genotypes with lower virulence (cagA+/vacAs1m1, PR: 2.42, 95% CI: 1.14 to 5.13, p < 0.05) (cagA+/vacAs1am1, PR: 1.67, 95% CI: 1.13 to 2.45, p < 0.01) (Table 3).

Discussion
IM can be considered a point of no return in the GC cascade, and it is believed that patients with IM are at high risk for GC even after H. pylori eradication [34,35].We observed that preneoplastic condition IM presented a significant association with age and that its prevalence was 2.4 times higher in patients older than 50 years compared to cases younger than 35 years (95% CI: 0.97 to 5.88, p = 0.05).This finding is similar to those regarding the age-related preneoplastic lesions reported in patients infected with H. pylori [36][37][38] and may be a chronic effect of the infection due to cumulative damage to the epithelial mucosa.
Genetic and epigenetic changes on gastric cells are induced by gene coding for virulence factor genes in H. pylori [39].CagA-positive H. pylori strains are known to promote toxicity mechanisms that increase the risk of GC [23,[40][41][42].In Peru, high frequencies of strains positive for the cagA gene have been reported (79.9% to 100%) [6,17,18], and we confirm these findings.In our study, we observed that all clinical isolates were positive for the cagA gene, suggesting a predominance of the cagA-positive genotype among circulating H. pylori strains and predisposing infected cases to a higher degree of chronic inflammation with a risk of preneoplastic gastric lesions [43].
We also characterized the EPIYA segments of the C-terminal region of the cagA gene; as in many reports [44][45][46][47], we observed that an increase in EPIYA-C segment repeats is a risk factor for preneoplastic gastric lesions.The analysis showed that the IM condition was prevalent and associated with the presence of H. pylori strains with two EPIYA-C segments (PR: 1.56, 95% CI: 1.02 to 2.41, p < 0.05), which may be related to the pro-oncogenic effects of the EPIYA-C segment on gastric cells [48].In this study, H. pylori strains with three EPIYA-C segments were detected, and their frequency was lower than detections with two EPIYA-C segments (6.9% and 17.7%, respectively), but the association with IM condition was absent.In addition, we observed multiple cagA EPIYA genotypes in 27.8% of cases, suggesting the presence of different H. pylori strains from the same patient or the presence of new strains arising from the high recombination frequency and microevolution of H. pylori [24], which could confound the potential effect of the virulence marker and lead to a Type II error [44].
Vacuolating cytotoxin gen (vacA) is an important virulence marker in H. pylori and is known to induce vacuolization of gastric epithelial cells in a variable manner due to its ability to assume different polymorphic rearrangements due to the genetic diversity of the vacA gene [25].We confirm the findings reported in Peru [17,18], and we observed a predominance of H. pylori strains with vacAs1, vacAm1, and vacAi1 polymorphisms.The analysis showed no association of the vacAs1 polymorphism with the IM condition, but in similarity to what was reported in clinical isolates from Colombia.[49], the frequency of the vacAs1 polymorphism and its vacAs1a variant was predominant.In addition, we observed that when characterizing the vacAs1 polymorphism, 51.9% of the cases presented multiple genotypes and that co-infection by vacAs1a and vacAs1b strains was inversely associated with the IM condition (PR: 0.40, 95% CI: 0.18 to 0.88, p < 0.05).It has been suggested that the presence of co-infection could reduce the potential risk of a virulence factor causing preneoplastic injury, and therefore warrants a thorough evaluation [50].
In our study, we also reported that the vacAd1 polymorphism was predominant but was not associated with the IM condition.According to Ogiwari et al., the vacAd1 polymorphism could be a predictor of gastric neoplastic lesions and reports frequencies in populations from the United States (74.1%) and Colombia (74%) that are comparable to our study.Furthermore, it reports significant risks for cases with GC (OR: 8.04; 95% CI: 2.67 to 24.16; p < 0.05) but no association in patients with preneoplastic lesions (OR: 1.87; 95% CI: 0.91 to 3.87; p > 0.05) [31].
We have studied the allelic variants of the iceA gene, and although its function is still uncertain, it is known that its expression is induced by contact between H. pylori and gastric cells [54].Our study showed that the frequency of iceA1 was higher than that of iceA2 (20.9% and 9.5%, respectively), and that both polymorphisms were prevalent among IM cases (51.5% for iceA1 and 66.7% for iceA2).This finding differs completely from that reported for Western countries [14,16], but shows similarity to the distribution of iceA1 strains [55] and with the frequencies observed among cases with neoplastic lesions in Colombia (52.1% for iceA1 and 69.5% for iceA2) [49].This suggests that the allelic distribution of iceA may be related to the geographical distribution of H. pylori strains [16].Likewise, several studies have reported co-infection by iceA1 and iceA2 strains [14,16,56,57]; co-infection was frequent (69.6%) in our study and prevailed in the clinical condition CNAG (69.1%).Moreover, it was significantly inversely associated with IM (PR: 0.46; 95% CI: 0.29 to 0.73, p < 0.05).
On the other hand, we characterized allele 2 of the babA gene and observed a frequency of 52.5% of babA2 strains, which is comparable to the frequencies observed (40.4% to 82.3%) in populations of the Americas [55,[58][59][60].Adhesin babA is a determinant in H. pylori gastric colonization [15,61] and it has been reported that the presence of the babA2 gene is related to GC [33,62]; however, it is suggested that the risk relationship with neoplastic lesions may be especially associated with Asian populations but not with South American populations [63], which is similar to our finding (PR: 1.13; 95% CI: 0.76 to 1.69, p > 0.05).
In addition, the risk of IM or GC is known to be increased by infection with H. pylori strains that simultaneously display multiple virulence genes [49,64].We observed that both the cagA+/vacAs1m1 genotype (PR: 2.42, 95% CI: 1.14 to 5.13, p < 0.05) and the cagA+/vacAs1am1 genotype (PR: 1.67, 95% CI: 1.13 to 2.45, p < 0.01) were prevalent in the IM condition.It has been reported that the risk of gastric neoplastic lesions in cases with vacAs1m1 genotype infection increases from 1.75 (95% CI: 1.04 to 2.96)-fold to 4.8 (95% CI: 1.71 to 13.5)-fold given the detection of a cagA+ gene-positive status.[12,65].In our study, the simultaneous detection of virulence polymorphisms for the cagA and vacA genes was associated with the IM condition; however, it is suggested that the strength of association might have been underestimated given the absence of cagA-negative H. pylori genotypes among our circulating strains; so, these findings should be taken with caution and should be considered significant.
Finally, our study was subject to limitations inherent to its cross-sectional nature and its ability to differentiate cause-and-effect relationships in the genotype-phenotype association.Furthermore, the sample size, the absence of healthy cases and the absence of cases with clinical stages such as dysplasia or gastric adenocarcinoma prevent the generalization of the results.However, in our study, H. pylori characterization has important clinical significance, but its relationship to the preneoplastic clinical stage may not reveal its potential risk given the influence of other variables such as the genetic variability of H. pylori or the presence of coinfection, the genetic susceptibility of the host, the responsiveness of the host immune system, and environmental factors [66,67].

Conclusions
Our findings revealed that the H. pylori strains circulating in our environment present a higher frequency of genotypes documented as risk variants for neoplastic lesions, and that their distribution was observed both for cases with CNAG and for cases with IM, highlighting that in Peruvian patients with dyspepsia who present H. pylori infection, the presence of risk polymorphisms for the cagA and vacA genes is related to the clinical stage of preneoplastic IM lesion.

Table 1 .
Protocols used for genotyping H. pylori clinical isolates.

Table 2 .
Characteristics of adult patients with a diagnosis of dyspepsia and H. pylori infection in Lima during 2016 and 2017.

Table 3 .
Frequency and association of virulence genotypes with histopathological findings in adult patients diagnosed with dyspepsia and H. pylori infection in Lima during 2016 and 2017.