Comparative genomics of Helicobacter pylori isolates recovered from ulcer disease patients in England

Infection of the gastric mucosa with H. pylori results in a number of disease outcomes including gastritis, which precedes the development of peptic ulcer disease, gastric cancer and lymphomas of the MALT [1–3]. These diseases caused by H. pylori and their prevalence rates differ in different geographic countries and only a subset (10%) [4] of infected patients develop one of them. This raises the question as to why H. pylori causes disease in a few individuals, but not in the great majority [5].

Many studies have demonstrated the involvement of bacterial virulence factors, host genetics and environmental factors in contributing to the development of disease. Bacterial virulence factors include proteins mediating establishment/colonization, persistence of infection and finally long-term damage to the host [6]. The cag pathogenicity-island (cag PAI) is the most noteworthy among these factors. PCR analyses have suggested that this island is not intact in many strains across the world [7] and the presence of an intact PAI although not always [8, 9] is indicative of a more severe outcome [10]. The expression of various products encoded in the cag PAI is known to be involved in inducing inflammation, ulceration and carcinogenesis [11]. However, the cag A is expressed by the majority of H. pylori strains, irrespective of the geographic origin and clinical diagnosis [12].

The vacuolating cytotoxin antigen (VacA) is another virulence factor that is considered to constitute an increased risk for development of peptic ulcers and gastric cancer [13, 14]. Allelic variations in the vac A gene are found in the signal (s1, s2) and the middle region (m1a, m1b, m2) and the s1 type is associated with ulcer disease [14, 15].

More pronounced inflammation is associated with strains, which express the outer membrane protein OipA. OipA induces IL-8 secretion by epithelial cells. Active OipA protein production may be 'on' or 'off' depending on the number of CT repeats in the signal sequence of the oip A gene (HP0638). H. pylori strains may also be grouped geographically based on oip A sequence pattern [16]. Specific adhesins viz., bab A and bab B mediate the adherence of the bacterium to specific human blood group Lewis antigens and are associated with various disease outcomes [17]. Similarly, a putative E. coli restriction enzyme Nla III homologue, the ice A gene in H. pylori, which is activated on contact with the epithelium, is also shown to induce high levels of IL-8 [18].

Accordingly, strains with Oip A 'on' status, active forms of ice A and bab A [18] and particularly strains which are cag A+ and vac A s1 have been shown to cause a more severe outcome [14, 15, 19], though not in all cases [20].

Many studies have pointed out a bio-geographical variation in virulence factors; for example, the sequences of vac A and cag A differ in strains from the United States and Europe from those in China and Japan [21]. Also, the prevalence and type of H. pylori infection varies with a very high rate of occurrence (up to 70%) in Asia and the Middle East [22], in contrast to only 30–50% in Europe and the United States [23]. Further, the infection is minimal in children in the west while in the rest of the world it affects both young and the old. Active infection with H. pylori was seen in about 7.5 million people in the general population of England and Wales. This although varied from one region to the other with the highest rates recorded in London [24]. Thus H. pylori remains an important infection in the UK.

H. pylori population has been described as highly recombining, and therefore exhibits enormous strain diversity, part of it may be due to the presence of the plasticity zones [25]. Since this organism has also been shown to be transmitted within families, a greater number of epidemiological studies reveal that these strains not only show similar genotypic profiles when obtained from related patients but also show common profiles within isolates from specific countries [26]. Phylogeographic affinities were pronounced in case of European strains based on the multi locus sequence typing of seven housekeeping genes where the European strains and the Asian strains shared an ancestral relationship [27]. This observation was also recorded in other studies based on the evolutionarily conserved ERIC sequences that indicated close associations between the Irish, Spanish and the European strains and also clustering of the English strains with a few Asian strains [26]. However, the number of English strains used in that study was very small. It is noteworthy in this respect that comprehensive phylogenetic analyses in case of English strains have been rarely performed.

In this study, we aimed at a comprehensive assessment of the prevalent genetic structure of H. pylori strains infecting the English population in Nottingham, which is centrally located in the United Kingdom. The strains were analyzed to study a total of 45 different parameters pertaining to 28 informative loci including the virulence factors cag A, oip A, ice A and vac A in addition to other genes of the cag PAI and the motifs downstream to it. Composition of the plasticity region cluster including the putative gastritis (JHP0986) and gastric cancer (JHP0947) associated markers were also studied [25]. Phylogenetic analyses were performed using FAFLP markers, nucleotide sequences of the cag A, bab B and the glm M genes and the repetitive sequences interspersed in the H. pylori genome (ERIC and REP). According to our observations, phylogenetic placement of English strains shows affinities with East Asian and Amerindian strains.

Evolution of infectious microorganisms is a consequence of the genetic polymorphisms they accumulate, which in turn is the result of the long term selection pressure exerted by the host immune system in case of chronic infection as well as the environment [9, 30]. This is more pronounced in case of bacteria like H. pylori that cause multi decade long infections, wherein an acquiescent requirement to constantly keep their genomic content recasting is crucial. In H. pylori this concern is resolved by the use of restriction – modification systems and regulation at gene level by nucleotide substitution, insertion and deletion events [31]. Further, the presence of insertion elements, plasticity regions and the pathogenicity islands contribute considerably to its genetic diversity.

We attempted to analyze genetic variation and structure of H. pylori populations infecting the native people of England. The UK today is more culturally diverse than ever before with the majority of the UK population being ethnic Europeans (92 %). The remaining 4.6 million (or 7.9 %) people represent other ethnic groups. South Asians are the largest of these groups, followed by Caribbean and Black Africans [32]. Such a multiethnic presence creates an interesting genetic conundrum when we attempt to analyze incidence and healthcare-impact of any pathogen that biases itself with respect to the host genetic makeup. Also, associations of the disease outcome with the virulence factors has thus far been enigmatic and since there were no comprehensive studies involving multiple loci for genotype-phenotype assessments, the current study was envisaged in combination to generate base line data relevant in molecular epidemiology of virulent strains in England.

The cag PAI is a major virulence determinant in H. pylori and strains lacking this island are akin to commensals rather than pathogens [1]. Reports suggest that the presence of a complete set of genes within the cag PAI ensures a 5-fold increased severity of disease outcome than the intermediate PAI [10]. We have earlier showed that a higher number of strains from Japan had an intact cag PAI [7], hence it may be thought as an important factor influencing the outcome of the infection as a higher rate of gastric cancers was observed in the Japanese patients. The possible role of cag A in oncogenic mechanisms is being worked upon [33]. Most English strains in our case retained the cag T and the cag E genes. Studies showed that strains lacking the cag T gene had a defective 'molecular syringe' that is encoded by the PAI [34] reflecting thus on inability of the Type IV system to eject out the cag A protein. The cag E gene on the other hand is known to induce NFkB activation and IL8 secretion [8] in addition to mediating host-cell cytokine rearrangements in infected epithelial cells. About 25 % of the strains carried the type IIIa motif on the right end of the cag A gene. This observation also supports the hypothesis that some of the European strains share some features typical of the Asian genogroups. This is also supportive of an earlier observation [31] that a very small number of strains from the European countries also show the type III motif.

English strains mainly showed a higher number of the toxigenic s1 type of vac A allele. Interestingly, the s1m1 combination was observed in less percentage in contrast to the s1m2 genotype that was seen in 60 % of the cases. Earlier studies in a subset of European population (Mid-Essex) by RFLP of the mid-region of the vac A gene in strains originating from dyspeptic patients demonstrated that 46% of these strains had the s1m2 genotype while 40% strains had the more toxigenic s1m1 combination [35]. The s1m2 genotype was also common in strains from North Wales among clarithromycin sensitive and resistant H. pylori [36]. It was interesting to find the m1c allele [28], which is known to be prominent in the East Asians, in UK strains. None of the strains had multiple vac A genotypes, which are common in China and other East Asian countries [37]. It has been reported earlier that the s1 allele was most frequently observed in the European population with the s1a allele predominant in Northern and Eastern Europe. Also, s1a and s1b alleles were observed in France and Italy while the Spanish and Portuguese strains had the s1b subtype [13].

The virulence spectrum of the English strains was also exemplified by the observation that 70% of these strains had the oip A gene in frame. Greater than 6 CT repeats in the upstream homopolymeric tail of the oip A gene is characteristic of European strains [16]. Our results indicate that English strains most commonly displayed 6 CT repeats. Strains with 9 CT repeats were reported to have the gene out of frame owing to the deletion of the CTAA sequence present immediately upstream of the CT repeats [16]. Similar results were found with English strains with 9CT repeats. From our analyses, the ice A1 allele was more common than the ice A2 allele. Although no association between ice A subtypes and clinical outcome has been reported, strains carrying iceA1 produce higher levels of IL-8 in the gastric mucosa and are more often associated with DU in the North America and Dutch people than strains carrying iceA2 [38]. The plasticity region genes are speculated to provide the strains with survival benefits in some hosts [25]. This region extends from the ORFs JHP914 to JHP951 in the sequenced strain J99 and is shown to be unstable since some genes within the zone are lost during subsequent infections or laboratory passages [25]. It is now evident from partial sequencing of a Peruvian isolate that this plasticity region might encode yet another type IV secretion system [39] and that the 2 sequenced genomes carry incomplete sets of genes corresponding to this cluster. All the English strains we looked at were similar to either strains 26695 and J99 in that none of them harbored a complete plasticity region cluster as shown recently for a Peruvian strain [39] (data not shown). However, any role of such a putative secretory system is still enigmatic due to lack of correlation of its intactness or abrogation with disease, although, some of the ORFs in the plasticity region have been shown to be associated with a particular disease outcome [25, 40]. For English strains, we analyzed seven ORFs from the plasticity region cluster, of which HP986 was strain 26695 specific, while others excluding HP912 were J99 specific. The ORF HP912, the predicted cell division protein/septum formation protein (fts A) was PCR amplified in 93.5 % strains. JHP931, a predicted DNA topoisomerase I was amplified in 51%. None of the strains had the same regions of the plasticity zone deleted.

Our phylogenetic analyses based on the FAFLP markers showed a star like distribution on an unrooted neighbor-joining network. Four clusters were evident and the largest cluster was populated mainly by English isolates. Further, affinity to Irish and Indo-European genogroups was also observed. These observations were also substantiated by other slow evolving markers such as ERIC and REP sequences. This denotes stable associations among the genogroups. Similarly, cluster analysis of bab B and glm M genes also revealed close associations within the European strains and with the Indian strains. However, homologies with East Asian and Amerindian strains were most noteworthy and were comparable to those shown by Irish strains [41]. This reflects ancient genetic events and possible oriental influences on the evolution of H. pylori in the English population. Such kinds of non-random genetic links of H. pylori may be helpful in understanding evolution of this organism and its clinical consequences in different parts of the world. These findings are in accordance with a recent study that demonstrated that Indian and European H. pylori isolates grouped in the same subpopulation and that East Asian and a subset of European isolates share an ancestral relationship and diverged from each other recently [27, 42]. The Asian strains, however, were distinctly separated from the European and western strains based on the cag A gene sequences except for a few strains that show remote similarity to the East Asian gene pool. We found only a single English strain (N115) that diverged significantly towards the Asian cluster and was recovered from a patient of Indo-Pakistani origin and thereby denotes contribution from the Asian gene pool.

In summary, our study demonstrated certain distinctive genetic features of the H. pylori gene pool in England based on genotypes of virulence genes and neutral markers. Important among these features is the genetic affinity towards East Asian strains. This is also probably the first comprehensive study on detailed, multilocus and multi method genotyping of H. pylori from England or elsewhere. The genomic profiles generated in this study may be useful for electronic archiving and retrieval for inter-laboratory comparison and are suitable for storage in epidemiological databases for comparative analyses. However, it will be necessary to analyze additional representative strains, especially from other European populations. Also, our study has largely been an examination of a specific (peptic ulcer disease) group of patient isolates and may not be reflective of other patient isolates from different disease stages in England. Future studies are therefore clearly needed to involve other disease specific strain groups. Further characterization of associations of such informative loci as we examined, in the gene pool of diverse strain groups and with varied disease spectrum may lead to newer insights into the mechanisms of H. pylori colonization, and virulence in different hosts.