Association of Helicobacter pylori vacA polymorphisms with the risk of gastric precancerous lesions in a Moroccan population

Introduction: Helicobacter pylori infection is the major risk factor of atrophic gastritis and intestinal metaplasia. The vacA gene is one of the most virulence factors of H. pylori and genetic diversity in its s, m, i, and d regions is associated with gastric lesions severity. This study aimed to investigate the association of vacA s, m, i, and d regions with the risk of atrophic gastritis and intestinal metaplasia in a Casablanca population. Methodology: A total of 210 patients suffering from gastric lesions (chronic gastritis, atrophic gastritis, and intestinal metaplasia) were enrolled. The type of lesion was diagnosed by histological examination. Detection of H. pylori infection and genotyping of vacA regions were carried out by PCR. Results: The prevalence of H. pylori was 95%. The most common vacA genotypes were s2 (51.5%), m2 (77%), i2 (60.5%), and d2 (58.5%). VacA s1, m1, and i1 genotypes were associated with a high risk of intestinal metaplasia, while the vacA d1 genotype increases the risk of atrophic gastritis and intestinal metaplasia. The most common vacA combination was s2/m2/i2/d2 (52%), and it was more detected in chronic gastritis. The moderate virulent vacA combination (s1/m2/i1/d1) increases the risk of atrophic gastritis, while the most virulent vacA combination (s1/m1/i1/d1) increases the risk of intestinal metaplasia. Conclusions: Genotyping of vacA d region might be a reliable marker for the identification of vacA virulent strains that represent a high risk of developing precancerous lesions (atrophic gastritis and intestinal metaplasia).


Introduction
Helicobacter pylori infection systematically leads to chronic gastritis (CG), which can develop into more severe pathologies such as gastric ulcer, MALT lymphoma, and gastric cancer (GC). GC is a multi-step pathology that develops through a series of lesions known as gastric carcinogenesis and includes: CG, atrophic gastritis (AG), intestinal metaplasia (IM), dysplasia, and GC [1]. The mechanisms by which AG and IM (known as precancerous lesions) develop are linked to a complex interaction between H. pylori virulence factors, human genetics, and environmental factors.
H. pylori is a genetically heterogeneous bacterium and genetic polymorphisms of its virulence factors affect its pathogenicity [2]. For instance, variations of the vacuolating cytotoxin A (vacA) gene have been proposed as a means of identifying virulent strains involved in the occurrence of gastric diseases [3].
The vacA gene is one of the most important virulence factors of H. pylori. This gene encodes for the multifunctional toxin VacA involved in several deleterious biological activities, such as vacuolization, apoptosis, tight junction disruption, and suppression of T cell activation [4]. The vacA gene is present in all H. pylori strains and comprises four polymorphic regions: signal (s), middle (m), intermediate (i), and deletion (d) region. Each vacA region is divided into two subtypes: s1, s2, m1, m2, i1, i2, d1, and d2 [5][6][7].
The vacA s region encodes for the signal peptide of the VacA protein. The vacA s1 genotype encodes for the whole signal peptide while the vacA s2 genotype encodes for a short signal peptide which results in a low vacuolating activity [8,9]. The vacA m region is responsible for the binding of the VacA protein to host cells. The vacA m1 genotype is more active and binds to a wider range of host cells than the m2 genotype [9]. The vacA i and d regions have been recently discovered [6,7]. In the vacA i region, the i1 genotype is associated with high vacuolation activity than the i2 genotype [6]. In the vacA d region, the d1 genotype is characterized by the absence of a 69 to 81 bp deletion, while the d2 genotype exhibits this deletion [7]. Several vacA combinations of these genotypes exist, and each of them is more or less associated with the risk of precancerous lesions and GC development.
GC is one of the most aggressive neoplasms and it is associated with a poor prognosis. Because of its late diagnosis, most Moroccan patients detected are at advanced stages of the disease, which results in a fiveyear survival rate of less than 15% [10]. Finding a marker for the early diagnosis of patients at high risk of developing this cancer is an important step in reducing its mortality. Our study aimed to investigate the polymorphisms of vacA s, m, i, and d regions and their association with gastric precancerous lesions in a Casablanca population, in order to use these regions as predictive markers in the identification and follow-up of patients that present high risks of developing this cancer.

Study population
A total of 210 patients consulting in the gastroenterology service of Ibn Rochd University Hospital Center of Casablanca (Morocco) and suffering from digestive pains were included in this study. From all patients, 6 biopsies (2 from antrum, 2 from fundus, and 2 from lesser curvature) have been sampled. Three biopsies (1 from antrum, 1 from fundus, and 1 from lesser curvature) were used for histological examination and the other three biopsies were used for molecular detection. All participants were informed of their inclusion in the study and agreed to it on a writing form. The study protocol has been performed under the ethical standards of Helsinki and was approved by the ethical committee of the Pasteur Institute of Morocco.

Histology
The biopsy samples were transported in 10% formalin and embedded in paraffin. Multiple histological sections were obtained from each biopsy. Biopsy sections were then obtained and stained with hematoxylin-eosin for the detection of gastric lesions. The blades were read by a pathologist.

PCR for H. pylori detection
Total DNA was extracted from gastric biopsies using a genomic DNA extraction kit (Isolate Genomic DNA Kit, Bioline, Memphis, USA). Using primers described by Lu et al. [11], the ureC gene (296 bp) was amplified to detect H. pylori infection. The PCR reaction mixture was prepared with 0.5 mM dNTPs, 1 Memphis, USA), and 300 ng of DNA in a final volume of 20 µL. All PCR thermocycling conditions for genotyping of vacA regions are listed in Table 2.
Statistical analysis R software version 3.4.0 was used to conduct statistical analysis. Chi-square and ANOVA tests were performed to assess all associations between gastric lesions, age, gender, and vacA s, m, i, and d regions.
The association between vacA d genotypes and vacA s, m, and i combinations were calculated using the Fisher test.
For the association between gastric lesions and vacA combinations, gastric lesions were considered as the dependent variable, and vacA s1/m2/i1/d1 and s1/m1/i1/d1 combinations as the predictor variables. The CG group and vacA s2/m2/i2/d2 combination were taken respectively as the control group and the reference strain. Results were expressed as odds ratio (OR), 95% confidence intervals (95% CI), and pvalues.

Population characteristics
The population was constituted of 99 (47%) males and 111 (53%) females. The mean age of the population was 49 ± 16. According to histological examination, 61% of patients were diagnosed with CG, 25% with AG, and 13% with IM.
Gastric lesions severity was increasing with age, but without being statistically meaningful (p = 0.39) ( Table  3). Concerning gender, the frequency of females and males diagnosed with CG was the same (48 and 52%, respectively). AG was more diagnosed among females, whereas IM was predominant among males (Table 3). Association between gender and gastric lesions severity was statistically significant (p = 0.04).

Prevalence of vacA genotypes
All vacA regions were determined for all the 200 H. pylori-positive patients. In all vacA regions, a dominance of the inactive form of the vacA genotype (s2, m2, i2, and d2) was observed (Table 4).

Association between vacA genotypes and gastric precancerous lesions
The frequency of the vacA s1 genotype was shown to increase with gastric lesions severity: 41% in CG, 57% in AG, and 64% in IM. The vacA m1 genotype was detected with low frequency in CG and AG (18 and 20%, respectively) while it was higher in IM (50%). The frequency of the vacA i1 genotype was found to increase with gastric lesions severity: 31% in CG, 45% in AG, and 64% in IM. Similarly, the frequency of the vacA d1 genotype increased with gastric lesions severity: 31% in CG, 49% in AG, and 71% in IM ( Table  4). Distributions of vacA s, m, i, and d genotypes according to gastric lesions severity were statistically significant (Table 4).

Association between vacA combinations and the risk of gastric precancerous lesions
By taking CG as a control group and vacA s2/m2/i2/d2 combination as a reference strain, the risks of developing AG and IM following infection with the vacA s1/m1/i1/d1 and s1/m2/i1/d1 combinations were estimated.

Discussion
Since the discovery of the vacA s, m, and i regions, numerous studies have investigated their association with the risk of precancerous lesions. However, the recently discovered vacA d region remains poorly studied. In this study, we characterized the polymorphisms of the vacA s, m, i, and d regions in order to study their association with the development of precancerous lesions.
Concerning the vacA d region, the majority of H. pylori strains were vacA d2 genotype (58.5%). Such finding has been reported by several Iranian studies [18][19][20][21], while other studies revealed a dominance of the vacA d1 genotype [22,23]. In Africa, there is no data regarding the prevalence of vacA d region, so further studies are needed to establish an accurate profile of this region.
The distribution of vacA genotypes among gastric lesions showed that the frequency of vacA s1, m1, and i1 genotypes tends to increase in AG compared to CG, but without reaching a statistically significant association (Table 4). In IM, all active forms of the vacA regions (s1, m1, and i1 genotypes) were found to be associated with the development of this lesion. Association between vacA genotypes and the development of gastric lesions varied among epidemiological studies. Some reports found that vacA s1, m1, and i1 genotypes increased the risk of both AG and IM [24,25], while others found that vacA s1, m1, and i1 genotypes were only associated with the risk of IM [14,26,27].
In the case of the vacA d region, the vacA d1 genotype was found to be associated with AG and IM. Ogiwara et al reported a positive association between vacA d1 genotype and the development of AG [7]. In addition, the vacA d1 genotype was found to increase the risk of GC by numerous studies [18][19][20]23]. However, no study has assessed the association between the vacA d region and IM.
The combination of the vacA s, m, and i genotypes allows the differentiation of the vacuolating activity of the VacA protein between H. pylori strains. It is known that the vacA s1/m1/i1 combinations induce cell vacuolation while the vacA s2/m2/i2 combinations do not. In the case of the vacA s1/m2 combinations, the presence of the i1 genotype is associated with a cellular vacuolation activity, while the presence of the i2 genotype is associated with the absence of the vacuolation activity [3,6,25].
In our population, most of our vacA d1 genotype cases were detected in the active forms of vacA combinations (s1/m1/i1 and s1/m2/i1). This observation is similar to previous studies [7,18,22]. In contrast, the inactive forms of vacA combinations (s1/m2/i2 and s2/m2/i2) were characterized by a predominance of the vacA d2 genotype. Even though the physiological role of the vacA d region remains undiscovered, it seems that the vacA d1 and d2 genotypes are highly associated with the active and inactive forms of vacA combinations, which are respectively characterized by high and low vacuolation activity.
The mosaic combination of the vacA s, m, i, and d regions can lead to several vacA combinations. Our population is characterized by the predominance of the nonvirulent vacA combination, s2/m2/i2/d2, followed by the most virulent vacA combination, s1/m1/i1/d1. In an Algerian study, the vacA s2/m2/i2/d2 was the most common combination [28]. Several African studies (Tunisia, Morocco, and Egypt) have also shown the predominance of the vacA s2/m2 combination in their population [13,15,29]. Moreover, a Moroccan and Kenyan study found a high prevalence of the vacA s2/m2/i2 combination, followed by the vacA s1/m1/i1 combination [14,16]. In contrast, the vacA s1/m1/i1 combination was predominantly detected in a Senegalese study [17].
Our finding demonstrated the association between the vacA d2 genotype with the least virulent vacA combination (s2/m2/i2). In addition, most African H. pylori strains are characterized by the predominance of the vacA s2/m2 and s2/m2/i2 combinations. Based on these observations, we might suggest that the African vacA genetic profile could belong to the s2/m2/i2/d2 combination. However, more studies are needed to confirm this hypothesis.
The results of our study suggest that patients infected with the vacA s1/m1/i1/d1 combination are more susceptible to develop IM (OR = 5.64, 95% CI = 2.09 -15.22, p-value < 0.001) than AG (OR = 1.03, 95% CI = 0.4 -2.61, p-value = 1). Winter et al showed that the vacA s1/i1 combinations are associated with a high risk of IM compared to the vacA s2/i2 combinations [26]. Moreover, a follow-up study conducted by Gonzalez et al showed that progression toward IM was more frequent in patients infected with the vacA s1/m1 combination than patients infected with the vacA s2/m2 combination [30].
The rapid evolution from a simple CG to more severe lesions is linked to the type of vacA combination. Indeed, H. pylori strains carrying the vacA s1/m1/i1 combination are known to be more virulent than H. pylori strains carrying the vacA s2/m2/i2 combination. It was shown that vacA s2/m2/i2 combinations do not cause vacuolation on epithelial cells, while vacA s1/m1/i1 combinations are characterized by a high degree of vacuolating activity [25]. In addition, the vacA s1/m1/i1 combinations are highly apoptotic and induce more intense inflammation [31]. Moreover, H. pylori strains possessing the vacA s1/m1/i1 combination are more likely to carry the cagA gene, which is another H. pylori virulence factor, and considered as an oncoprotein [32][33][34]. All these factors can explain the association between the vacA s1/m1/i1/d1 combination and the high risk of IM.
In our study, the moderate virulent vacA s1/m2/i1/d1 combination was associated with the risk of AG (OR = 2.84, 95% CI = 1.13 -7.13, p-value = 0.02). The vacA s1/m2/i1 combination is known to possess an intermediate vacuolating activity compared to vacA s1/m1/i1 combination [35]. This difference is explained by the variation encountered in the vacA m region, which influences host cell tropism between different vacA strains [36][37][38]. Indeed, vacA combinations with the m1 genotype can bind to a wider range of host cells than the vacA m2 genotype, which results in a great vacuolization effect [9]. This variation in cell tropism may explain the association between the vacA s1/m2/i1/d1 combination and the risk of AG.

Conclusions
We showed in this study that the vacA s2/m2/i2/d2 combination predominates in our Casablanca population. Compared to other vacA s, m, and i regions, the recently discovered vacA d region seems to be a better marker for the risk of AG and IM. In addition, the active form of the vacA d region was exclusively associated with the most virulent vacA combinations (s1/m1/i1 and s1/m2/i1). Moreover, patients infected with the vacA s1/m2/i1/d1 are more susceptible to develop AG, while those infected with the vacA s1/m1/i1/d1 combination are at high risk of developing IM. Taking together, our results show that the vacA d region appears to be a reliable marker for the identification of virulent vacA strains that are a risk factor for AG and IM development.