OUP user menu

Of microbe and man: determinants of Helicobacter pylori-related diseases

Karin Van Amsterdam, Arnoud H. M. Van Vliet, Johannes G. Kusters, Arie Van Der Ende
DOI: http://dx.doi.org/10.1111/j.1574-6976.2005.00006.x 131-156 First published online: 1 January 2006

Abstract

The human gastric pathogen Helicobacter pylori infects the human gastric mucus layer of approximately half of the world's population. Colonization with this bacterium results in superficial gastritis without clinical symptoms, but can progress into gastric or duodenal ulcers, gastric malignancies and mucosa-associated lymphoid tissue-lymphomas. Disease outcome is affected by a complex interplay between host, environmental and bacterial factors. Irrespective of disease outcome, the majority of H. pylori infected individuals remain colonized for life. Changing conditions in the human gastric mucosa may alter gene expression and/or result in the outgrowth of more fit H. pylori variants. As such, H. pylori is a highly flexible organism that is optimally adapted to its host. the heterogeneity in H. pylori populations make predictions on H. pylori-related pathogenesis difficult. In this review, we discuss host, environmental and bacterial factors that are important in disease progression. Moreover, H. pylori adaptive mechanisms, which allow its life-long survival and growth in the gastric mucosa are considered.

Keywords
  • Helicobacter pylori
  • adaptation
  • pathogenesis
  • genetic diversity

Introduction

Bizzozero (1893) was the first to report the presence of spiral-shaped organisms in biopsies of the mammalian stomach. The presence of these spiral bacteria in the gastric juice was linked to disorders of the upper gastrointestinal tract soon after (Pel, 1899). However, it was not until 1983 that these spiral bacteria were successfully isolated, and an association between their presence in the stomach and the occurrence of gastric ulcers was established (Marshall & Warren, 1984). By self-ingestion experiments, Marshall (1985) and Morris & Nicholson (1987) demonstrated that these bacteria indeed cause gastric inflammation, thereby fulfilling Koch's postulates (Koch, 1884). In addition, the bacterial load was reduced by treatment with the antimicrobial agents erythromycin and bismuth in patients with this gastric inflammation (McNulty, 1986). Current treatment of gastric inflammation is optimized to effectively eradicate the bacteria, resulting in regression of ulcers and associated pathology (Guo, 2004; Kearney, 2004).

The causative agent of gastritis and gastric ulcers was determined to be a Campylobacter-like bacterium, now named Helicobacter pylori (Goodwin, 1989). Recent analysis suggests that H. pylori already colonized the human gastric mucosa during the evolution of modern man (Falush, 2003; Wirth, 2004). In infected patients, H. pylori resides in the mucus layer covering the gastric epithelial cells (Blaser, 1997; Guruge, 1998). In an attempt to eradicate H. pylori, host immune mediators are drawn to the site of inflammation, but are unable to eliminate the bacteria (Meyer, 2000).

Helicobacter pylori colonizes the human gastric mucus layer of approximately half of the world's population (Dunn, 1997; Eaton, 2001). The majority of infected persons display a chronic superficial gastritis without clinical symptoms, although their gastric epithelium clearly shows signs of inflammation (Morris, 1991; Blaser & Parsonnet, 1994). However, in approximately 10% of infected persons, gastric or duodenal ulcers may develop in time (Parsonnet, 1995). In a small proportion of the infected population, mucosa-associated lymphoid tissue (MALT) lymphomas will develop, through a process which is not yet understood (Wotherspoon, 1993). In addition, H. pylori infection is strongly associated with the development of atrophic gastritis, which predisposes to the development of gastric cancer (Dubois, 1996; Dunn, 1997). Conversely, H. pylori infection seems to protect against the development of gastroesophageal reflux disease (GERD), although the exact mechanism of protection is unknown (Blaser, 1999). In addition, H. pylori may protect against diseases to which GERD might lead, i.e. Barrett's esophagus and esophageal adenocarcinoma (Blaser, 1999). The observed clinical differences that develop after infection with H. pylori are accounted for by host and environmental factors as well as bacterial factors (Fig. 1).

Figure 1

Schematic representation of the factors contributing to Helicobacter pylori-related gastric pathology.

The human gastric mucosa is the only known niche for H. pylori. The pH, temperature, oxygen tension and activity of immune mediators in the stomach fluctuate. Thus, H. pylori must continuously adapt if it is to persist and proliferate in its host. As an immediate response to altered conditions, H. pylori may modulate the transcription and/or translation of its genes via global gene regulators (Aras, 2003c). Spontaneous mutations, intra- and intergenomic recombination, as well as phase variation contribute to long-term adaptation and lead to an extremely diverse population structure of H. pylori (van der Ende, 1996; Suerbaum, 1998; Achtman, 1999; de Vries, 2002; Aras, 2003b). Specific conditions may select for the fittest H. pylori, resulting in an expansion of this subpopulation (van der Ende, 1996; Suerbaum, 1998; Achtman, 1999; Kersulyte, 1999). The diversity in H. pylori populations makes predictions on the outcome of H. pylori-related infection difficult. Therefore, combined genotyping of both host and bacterium may allow more accurate prediction of disease outcome. In this review, the molecular aspects of H. pylori-related pathogenicity are discussed.

Host and environmental determinants in Helicobacter pylori-related disease

Several determinants important in H. pylori-related diseases have been identified and will be discussed under the sections Host and environmental determinants and Bacterial determinants'. Host factors that have been shown to alter the risk for H. pylori-related diseases are the host's age at H. pylori acquisition and the duration of H. pylori infection, but more importantly (polymorphisms of) host immune mediators, and acid secretory status. Besides host factors, environmental factors such as smoking and diet affect the niche colonized by H. pylori, and hence might affect disease risk (Table 1).

View this table:
Table 1

Determinants important in Helicobacter pylori-associated disease outcome

DeterminantDisease associationReference(s)
Host
AgeEarly age of acquisitionPeptic ulcer and gastric cancerBlaser (1995) and Mitchell (1996)
InflammationTh1 responseMore severe inflammationKarttunen (1997), Bauditz (1999), Meyer (2000) and Guiney (2003)
IL-1β/IL-1RA/TNF-α/IL-10 polymorphismsNoncardia gastric cancerEl Omar (2000, 2003), Yea (2001) and Wu (2003)
Acid secretionHighDuodenal ulcerEl Omar (1995), Garcia-Gonzalez (2001) and Watanabe (2001)
LowGastric cancerKuipers (1995), Peek & Blaser (2002) and Blaser & Atherton (2004)
Environment
SmokingGastric cancerNomura (1990), Sasazuki (2002) and Machida-Montani (2004)
DietHigh salt/carbohydrates?Gastric cancer?Tsugane (1994), Watabe (1998), Foynes (2000) and Mathew (2000)
Low vitamin C?/B-carotene?Gastric cancer?Correa (1998)
MedicationNSAIDPeptic or duodenal ulcers?Laine (1995)
Helicobacter pylori
CagAcagA positivePeptic or duodenal ulcers, gastric cancerParsonnet (1997) and Graham & Yamaoka (2000)
cagA positive, high no. of TPMsGastric cancerArgent (2004)
VacAvacAs1positivePeptic or duodenal ulcers, gastric cancerAtherton (1995) and Figura (1997)
BabABabA2 positiveDuodenal ulcers, gastric cancerGerhard (1999) and Mizushima (2001)
OipAoipA 'on'Severe gastritis, duodenal ulcersYamaoka (2000, 2002)
SabBsabB 'off'Duodenal ulcers, gastric cancerde Jonge (2004b)
Jhp0947Jhp0947 positiveDuodenal ulcersde Jonge (2004a)
Jhp0949Jhp0949 positiveDuodenal ulcersde Jonge (2004a)
Jhp0950Jhp0950 positiveMALT lymphomaLehours (2004)
IceAIceA1 or IceA2d positivePeptic ulcers/gastric cancerPeek (1998) and Van Doorn (1998)
Jhp1462Jhp1462 positiveMALT lymphomaLehours (2004)
  • NSAID, nonsteroidal anti-inflammatory drug; MALT, mucosa-associated lymphoid tissue; Th1, T-helper cell 1; IL-1β, interleukin-1β; TNFα, tumor necrosis factor α; TPMs, tyrosine phosphorylation motifs.

Age of acquisition

Infection with H. pylori at an early age is associated with an increased risk for gastric ulcer and gastric cancer (Blaser, 1995; Mitchell, 1996). Development of gastric ulcers and gastric cancer requires long-term infection by H. pylori. Therefore, it is presumed that these diseases are seen more often in patients that have been chronically infected with H. pylori. Moreover, this association is thought to be determined by the differences in host immune responses in different age groups (Blaser, 1995).

It has been suggested that improved socioeconomic status in developed countries has resulted in an increase in the age at which H. pylori is acquired. As a consequence, gastric cancer is less likely to develop in these countries. This hypothesis also suggests that gastric cancer will occur more frequently in developing countries. However, the occurrence of gastric cancer is much lower among the African population even though the prevalence of H. pylori infection is in fact higher (the ‘African enigma’) (Holcombe, 1992). Probably the age at which H. pylori is acquired in developed countries is not different from the age of H. pylori acquisition in developing countries. Active detection and eradication of H. pylori in developing countries is not as high compared to developed countries, resulting in an increase in H. pylori incidence. Thus, differences in the age of H. pylori acquisition are not sufficient to explain the ‘African enigma’. Presumably, host protective or inhibitory factors, which are more often related to Africans, may protect against or delay the development of gastric cancer (Holcombe, 1992). Indeed, infections with helminthes, more prevalent in African countries, is presumed to prevent the progression of H. pylori-induced chronic active gastritis to cancer (Feldmeier & Krantz, 2000; Fox, 2000; Whary & Fox, 2004).

Inflammatory mediators

Upon infection with H. pylori, host gastric epithelial cells release the cytokine interleukin-8 (IL-8) (Crabtree, 1995; Sharma, 1995). This cytokine is involved in the recruitment of macrophages, neutrophils, mast cells, B-cells and T-cells to the site of inflammation. These cells enhance the immune reaction even more by the secretion of other inflammatory mediators, such as interferon-γ (IFN-γ), tumor necrosis factor-α (TNF-α) and interleukin-1β (IL-1β) (Karttunen, 1997; Bauditz, 1999; Meyer, 2000; Guiney, 2003) (Fig. 2).

Figure 2

Schematic representation of Helicobacter pylori-related pathogenesis; (A) in the human stomach, H. pylori colonizes mainly in the antrum, which lacks acid secreting parietal cells; (B) in the gastric pits, H. pylori proliferates especially near the tight junctions between the epithelial cells; (C) H. pylori factors important for colonization, persistence, and disease development. The cag PAI induces Il-8 secretion from epithelial cells, which attracts inflammatory cells. Upon translocation into the epithelial cell, CagA is phosphorylated and induces cell changes. CagA also disrupts the tight junctions, giving H. pylori access to nutrients such as Fe3+ and Ni2+. VacA causes vacuolation and degeneration of epithelial cells, and NAP attracts neutrophils to the site of inflammation. Together, these factors disrupt the epithelial barrier.

The immune reaction against H. pylori is further characterized by an oxidative burst in activated human neutrophils. This burst causes a release of reactive oxygen species (ROS), which are low molecular weight metabolites that damage essential biological molecules, including nucleic acids (de Groot, 1994). ROS may damage genes that control cell growth; and stimulate the development of cancer (Troll & Wiesner, 1985; Rautelin, 1993). Thus, the presence of immune mediators induces damage to the integrity of the gastric epithelium. This damage is believed to cause most of the pathology that is associated with H. pylori infection, rather than by direct H. pylori activity (Whary & Fox, 2004).

The immune cells present during infection with H. pylori are predominantly proinflammatory cells, and not anti-inflammatory cells (D'Elios, 1997; Karttunen, 1997; Bauditz, 1999; Meyer, 2000; Guiney, 2003). Proinflammatory mediators are secreted by T-helper 1 (Th1) cells. Naive T-cells are driven to differentiate into Th1 cells by several factors, such as TNF-α and IL-12 (D'Elios, 1997). The cytokine IL-12 is produced in high amounts by mononuclear cells upon H. pylori stimulation (Bauditz, 1999; Guiney, 2003). Thus, by stimulating IL-12 release, H. pylori augments the proinflammatory response. This results in more damage to the gastric epithelium, which predisposes to the development of gastric atrophy.

Several host genes that encode immune modulating factors harbor polymorphic regions. Polymorphisms in these genes will alter their transcription and thereby influence inflammatory processes. As a consequence, the risk for H. pylori-related diseases may alter (El Omar, 2000). So far, polymorphisms in the IL-1β, IL-1 receptor antagonist, IL-10 and TNF-α genes have been shown to be significantly linked with gastric cancer of the distal stomach (El Omar, 2000, 2003; Yea, 2001; Wu, 2003). Polymorphisms in the gene encoding TNF-α may enhance its production, and polymorphisms either in the IL-1β or IL-1 receptor antagonist may stimulate IL-1β release (Santtila, 1998; Yea, 2001; Hwang, 2002). Similarly, polymorphisms in the gene encoding the anti-inflammatory cytokine IL-10 may result in insufficient IL-10 levels (Wu, 2003). These events lead towards a Th1-driven hyper-inflammatory response against H. pylori and thus more damage of the gastric mucosa and an increased risk for H. pylori-related pathology (Berg, 1998; Eaton, 2001; Machado, 2001; El Omar, 2003).

Human leukocyte antigens (HLA) contribute to the elimination of pathogens, by presenting pathogen-derived peptide fragments to the appropriate T-cells (Lanzavecchia, 1992). HLA genes are highly polymorphic and certain HLA alleles have been associated with the development of gastric adenocarcinoma (Lee, 1996). Helicobacter pylori infection is also related to the development of gastric adenocarcinoma, but its presence is not associated with the presence of these HLA alleles. Therefore, the importance of HLA alleles in H. pylori-mediated adenocarcinoma is still controversial (Nguyen, 1999).

A genetically predetermined dominant proinflammatory Th1 response is thus associated with an increased risk for H. pylori-related disease (D'Elios, 1997). On the other hand, the intensity of the Th1 response may be limited by the production of the anti-inflammatory Th2 cytokine IL-10, thereby prolonging the infection. Th2-type responses are induced in response to parasites (Whary & Fox, 2004). As a result, the inflammation response and pathology related to H. pylori infection are tapered in individuals with a parasitic infection. Possibly, this causes a delay in the progression of chronic gastritis to gastric cancer in these individuals (Whary & Fox, 2004). Alternatively, different proinflammatory responses may direct to different pathways of H. pylori-related pathology.

Acid secretion

The Th1 response mentioned in the previous section causes gastric parietal cells to inhibit their acid secretion (Takashima, 2001; Padol & Hunt, 2004). A less acidic environment may facilitate the spread and persistence of H. pylori (Bjorkholm, 2004). Moreover, microorganisms other than H. pylori may grow in the human stomach when acid levels are low (Laine, 2000), thus enhancing the local inflammatory response. The combined action of proinflammatory cytokines and reduced gastric acid secretion increases the risk for peptic ulcer disease as well as gastric carcinoma. On the other hand, duodenal ulcers are less likely to develop when the secretion of acid is reduced. In addition, when the secretion of acid is enhanced, the risk for the development of duodenal ulcers is increased (Garcia-Gonzalez, 2001; Watanabe, 2001). However, also in low acid conditions, the risk for duodenal ulcers may be enhanced after H. pylori infection. The ability of gastric parietal cells to control their acid production is important in the development of H. pylori-related diseases. Helicobacter pylori infection is thought to be established predominantly in those parts of the human stomach that are less acidic, such as the antrum (Lee A, 1995). As a consequence, hosts with high gastric acid output display high H. pylori densities in the antrum and an develop predominantly antral gastritis (Lee A, 1995). The intact parietal cells in the corpus are subsequently stimulated to secrete acid, which induces gastric metaplasia in the duodenum. Helicobacter pylori colonizes this gastric metaplasia, which might lead to duodenal ulceration (El Omar, 1995). On the other hand, hosts with low gastric acid output are not only susceptible to H. pylori infection in the antrum, but also in the corpus. Infection with VacA-positive H. pylori in the corpus leads to a further inhibition of acid secretion from the corpus parietal cells (Kobayashi, 1996). This could stimulate gastric epithelial cells to proliferate continuously, resulting in progressive loss of gastric glands and eventually gastric atrophy and cancer (Fig. 3) (Kuipers, 1995; Peek & Blaser, 2002; Blaser & Atherton, 2004).

Figure 3

A model for the effect of host gastric acid output on Helicobacter pylori-related diseases. In hosts with high acid output, H. pylori will mainly colonize the less acidic antrum. Acid secretory cells will be stimulated to secrete more acid, which predisposes to duodenal ulceration (El Omar et al., 1995; Lee et al., 1995). In hosts with low acid output, H. pylori will also colonize the corpus. This leads to further inhibition of acid secretion, which predisposes to gastric atrophy and increases cancer risk (Kuipers et al., 1995; Peek & Blaser, 2002; Blaser & Atherton, 2004).

Environmental determinants

Several environmental factors are associated with the development of H. pylori-related pathogenesis. Smoking is a major contributor to the development of gastric cancer (Nomura, 1990; Sasazuki, 2002; Machida-Montani, 2004). Tobacco smoke contains carcinogenic nitrosamines, triggering the development of carcinoma (Mirvish, 1995). Helicobacter pylori infection enhances the risk for gastric carcinoma in smokers even more, presumably due to an increased risk of gastric atrophy (Lunet & Barros, 2003; Machida-Montani, 2004). However, other studies have not found this association (Nomura, 1990; Sasazuki, 2002).

Several dietary factors are associated with H. pylori-related diseases. A high salt diet is correlated with increased H. pylori colonization rates, gastritis and increased risk for gastric cancer (Tsugane, 1994). High salt concentrations cause atrophy of parietal cells, leading to the destruction of the mucosal barrier (Nozaki, 2002; Machida-Montani, 2004). However, only long-term high salt intake contributes to an increased risk of H. pylori-related disease (Machida-Montani, 2004). Similarly, a positive correlation between high rice intake and gastric cancer has been observed (Watabe, 1998; Mathew, 2000). Rice contains carbohydrates, which may irritate the gastric mucosa (Ji, 1998). Moreover, low fruit and vegetable intake may increase the risk of gastric cancer. Fruits and vegetables contain the antioxidants vitamin C and β-carotene (Buiatti, 1996), which are considered to protect against carcinogenesis, by neutralizing reactive oxygen species (Correa, 1998).

Use of nonsteroidal anti-inflammatory drugs (NSAIDs) is the cause of so-called NSAID gastropathy, including superficial mucosal damage, gastro-duodenal ulcers and dyspepsia (Chan, 2002; Huang, 2002). It has been suggested that the presence of H. pylori and the use of NSAIDs together increase the risk of ulcers in patients taking such drugs (Laine, 1995). In contrast, the combined effect of NSAIDs and H. pylori infection reduces the risk of mucosal damage. Infection with H. pylori increases the synthesis of prostaglandins by inducing cyclo-oxygenase 2 expression in epithelial cells (McCarthy, 1999; Akhtar, 2001). Although prostaglandins attenuate the production of reactive oxygen metabolite production by H. pylori, their importance in reducing damage caused by H. pylori infection is considered to be minimal (McCarthy, 1999).

Thus, even though several environmental factors are presumed to contribute to the development of H. pylori-related diseases, their role is often controversial. Presumably, a complex interplay between the different environmental factors, as well as host and bacterial factors hamper the interpretation of study results (Gordon, 2000; Feldmeier & Krantz, 2000; Fox, 2000; Bravo, 2002).

Helicobacter pylori-host adaptation

The human stomach is far from a stable niche. In this environment, the ability of H. pylori to adapt is indispensable for its persistence. Helicobacter pylori adapts to its environment both via short-term (see section on Gene regulation) and long-term adaptation (see sections on Spontaneous mutations, recombination and phase variation). The ability of H. pylori to induce subsequent disease is codependent on the presence or absence of certain H. pylori factors. Therefore, genetic variability affects H. pylori-related pathology.

Short-term adaptation: gene regulation

Bacterial sensing of environmental changes and subsequent adaptation is often mediated by regulatory systems (Merrell, 2003a, b; Ernst, 2005). Gene regulation may act as an immediate response to changing conditions by altering the transcription or translation of genes in the total H. pylori population. In the genome sequence of H. pylori, only a few global regulatory systems have been identified, in contrast to other bacteria (Berg, 1997; Alm, 2000). This might be the result of the lack of competition in the gastric mucosa from other bacteria, or of specialization of H. pylori for life in the human stomach (Berg, 1997).

However, extensive gene regulation has been shown to occur in response to multiple stresses (Lee IS, 1995; Bearson, 1998; van Vliet, 2004a). To deal with environmental changes promptly, H. pylori thus utilizes its few regulatory systems efficiently and/or novel regulators are involved (Barnard, 2004; van Vliet, 2004a). An efficient use of these few regulatory systems has been shown to be brought about by sensing and responding to multiple signals, and through cross-talk of the gene regulators. From this perspective, we will discuss the H. pylori ferric uptake regulator Fur, the nickel-uptake regulator NikR, and the carbon storage regulator CsrA.

Adaptation to the gastric environment includes adaptation to metal ions. Iron and nickel are essential nutrients for H. pylori growth; however, as these metal ions are potentially toxic to H. pylori, their uptake and usage should be tightly regulated. Iron is essential for maintaining energy and redox metabolism (Lee, 2004). In response to environmental iron availability, H. pylori will regulate the expression of several proteins (Szczebara, 1999; van Vliet, 2002b; Merrell, 2003b; Ernst, 2005). This iron-responsive regulation is mediated by the H. pylori Fur protein, which regulates iron-activated as well as iron-repressed promoters by binding (in complex with iron) at operator sequences upstream of Fur-regulated genes (Delany, 2001). For example, Fur regulates the expression of the iron storage protein ferritin (Bereswill, 2000; Delany, 2001). Low iron concentrations cause Fur-dependent down-regulation of the ferritin encoding gene, pfr. As a result, the concentration of free iron is increased (Bereswill, 2000).

Helicobacter pylori Fur is not only responsible for the regulation of iron metabolism, it is also involved in the response of H. pylori to gastric acidity (Bijlsma, 2002). Mutation of the fur gene affects the expression of urease (van Vliet, 2001), an important H. pylori virulence factor, involved in H. pylori acid resistance. However, the urease activities of an H. pylori fur mutant and its wild type are similar, and growth of the fur mutant to levels similar to that of the wild-type strain was restored by the addition of urea to the growth medium (Bijlsma, 2002). Therefore, the role of Fur in acid resistance of H. pylori is independent of urease and is mediated through an as yet unknown mechanism.

As with iron, nickel homeostasis is also tightly regulated in H. pylori. The H. pylori nickel regulator NikR mediates nickel-dependent repression of the nixA gene (van Vliet, 2003b). A nikR mutant, however, is not only affected in its nickel metabolism, but also its ability to metabolize other metals (iron and copper), its stress response, motility, chemotaxis, porin function and even the expression of other regulatory proteins (van Vliet, 2002a, 2004b; Contreras, 2003; Bury-Mone, 2004). NikR also regulates acid resistance. Nickel is an essential cofactor for urease (van Vliet, 2001). Regulation by NikR can act directly through the activation of urease, and indirectly through interaction with the Fur regulatory system (Fig. 4) (van Vliet, 2002a, 2004b; Bury-Mone, 2004). Thus, both NikR and Fur respond to multiple signals and cross-talk with each other and other gene regulators. An H. pylori mutant defective in both NikR and Fur is attenuated in a mouse model of infection. This infers that the responses to metal metabolism and acidity and H. pylori virulence are closely linked (Bury-Mone, 2004).

Figure 4

A schematic representation of cross-talk between regulator proteins in Helicobacter pylori. NikR regulates urease expression both directly, and indirectly, via the Fur regulatory system. CsrA regulates the expression of both the Fur and HspR regulatory systems. Through cross-talk of the different gene regulator proteins, H. pylori is able to sense and respond to multiple signals.

The ability to sense and respond to multiple signals, and cross-talk with other gene regulators is also illustrated by the actions of the global regulatory protein CsrA. This protein controls the response to environmental stress, modulates expression of virulence factors as well as motility by controlling access to the ribosome binding site and to a lesser extent also by altering mRNA stability (Liu, 1995; Liu & Romeo, 1997; Barnard, 2004). Furthermore, the expression of other global regulators, such as Fur and the heat shock protein HspR, are controlled by CsrA. Thus, H. pylori possesses a more complex network than can be anticipated from the relatively low numbers of global regulator proteins observed in its genome (Barnard, 2004; van Vliet, 2004a).

Long-term adaptation: mutation and recombination

Helicobacter pylori lacks several common mismatch repair systems, such as MutL and MutH homologs (Alm & Trust, 1999). As a result, single basepair mismatch mutations are maintained more often in H. pylori. This increases the mutation rate, thereby allowing the adaptation of H. pylori to changing conditions, of which the resistance to antibiotics has clinical relevance. For example, mutations in the frxA or rdxA genes may render H. pylori resistant to metronidazole (Marais, 2003). Antibiotic-resistant variants present in the H. pylori population survive antibiotic pressure in contrast to the original, antibiotic-sensitive population (Bjorkholm, 2001; Trieber & Taylor, 2002; Marais, 2003).

The H. pylori Nth endonuclease is involved in the repair of damaged nucleotides. Helicbacter pylori mutants with a deleted nth gene are unable to repair oxidized pyrimidines (O'Rourke, 2003). As a result, these H. pylori nth mutants display elevated spontaneous and induced mutation rates, and are more sensitive than their parental strain to killing by exposure to oxidative agents or activated macrophages (O'Rourke, 2003). However, repair mechanisms are not essential for H. pylori persistence (O'Rourke, 2003), as even in the absence of stress, subpopulations with deleterious mutations are not maintained. Also, a reduced efficiency of repair mechanisms may result in H. pylori variants with greater fitness for survival under certain conditions, which perhaps would otherwise not have been generated. Thus, in this way, H. pylori may adapt more easily to changing conditions.

Even though the frequency of mutations is relatively high in H. pylori, it is insufficient to explain the enormous diversity of this organism. In Neisseria meningitidis, the frequency of nucleotide changes is comparable to that in H. pylori, but sequence diversity in N. meningitidis is much lower than in H. pylori (Seiler, 1996). Thus, there must be one or more other mechanisms that contribute to the sequence diversity in H. pylori. Uptake or deletion of DNA fragments, through inter- or intragenomic recombination, may represent such an alternative mechanism. Recombination allows for the creation and deletion of intervening regions (Bzymek, 1999). The presence of repeats in a large number of genes encoding cell surface proteins that might directly interact with the environment emphasizes the significance of these proteins in host-adaptation (Tomb, 1997). Recombination is much more frequent in H. pylori compared to other microorganisms such as Escherichia coli or N. meningitidis (Suerbaum, 1998).

Recombination between repeat sequences may result in the loss of genetic information in the intervening region. Well studied examples are the recombination between repeat sequences flanking the cag region, but also the recombination between the 80 bp repeat sequences that flank the restriction modification (R-M) system gene hpyII, that lead to loss of either the cag region or R-M genes (Aras, 2001, 2003a, c). R-M system genes usually consist of a restriction endonuclease that cleaves DNA at a specific recognition site, and a methyltransferase that protects DNA from cleavage by the restriction endonuclease. R-M systems hereby act as barriers against transformation with foreign DNA. Subpopulations with a deleted R-M system allow the uptake of foreign DNA by recombination, thus novel genetic information may be acquired (Fig. 5A).

Figure 5

A schematic representation of recombination between repeat sequences; (A) recombination that results in an altered gene content and a different Helicobacter pylori phenotype. Repeats found upstream and downstream of the hpyII restriction modification (R-M) system encoding genes may recombine. As a consequence, the R-M system is lost (or acquired, in case another variant or strain is present that carries the R-M system); (B) recombination that results in an altered amino acid content of an expressed protein. Multiple direct repeats within the cagY gene may recombine. This results in in-frame cagY transcripts, which will be translated to intact proteins with a different size. A model for the pilus-like CagY protein of different size is given. CagY variants that differ in size from the original population minimize recognition by the host immune system, and thereby contribute to the chronicity of H. pylori infection (Aras, 2003).

Recombination may not only affect gene content, but also alter the amino acid content of an expressed protein. Both the CagA and CagY proteins may display altered expression, as a result of recombination between repeat sequences in the genes that encode these proteins. Recombination between repeats in H. pylori cagY results invariably in in-frame CagY proteins that differ in size (Fig. 5B). CagY has been described as being associated with a surface-located pilus-like appendage (Rohde, 2003). CagY variants that differ in size from the original population minimize recognition by the host immune system, and thereby contribute to the chronicity of H. pylori infection (Aras, 2003b). Within a host, the presence of heterogeneous CagY populations may allow selection of the fittest subclones under appropriate circumstances. Although there is potential for enormous size variation in CagY, only specific size ranges are detected, showing a strong selection against H. pylori CagY with sizes outside this range. Supposedly, only CagY proteins with sizes within these ranges contribute to H. pylori fitness (Aras, 2003b). The general lack of an immune response against CagY suggests that CagY variation occurs so rapidly, that the host immune system is not capable of mounting an effective immune response (Aras, 2003b).

Long-term adaptation: phase variation

Phase variation is a common phenomenon in bacteria, which enables bacteria to switch genes on and off to adapt to their environment (Henderson, 1999). Similar to the processes of mutation and recombination, phase variation results in the generation of variants in a population that may be more fit under the appropriate conditions. As a consequence, these variants may outgrow the original population. Phase variation is of importance to bacterial virulence as it provides a means of generating antigenic variants (van der Ende, 1995; Gilsdorf, 1998; Lim, 1998). One of the molecular mechanisms generating phase variation is slipped-strand mispairing mediated by short sequence repeats in the target genes. Slipped-strand mispairing during replication leads to the insertion or deletion of these short sequence repeats, resulting in the on/off switch of transcription or translation of the target gene. In H. pylori, numerous genes display phase variation, and prominent among these are genes involved in lipopolysaccharide (LPS) synthesis. Different lengths of the poly-C tract in the 5′-region of fucosyltransferase-encoding genes allow H. pylori to produce diverse LPS antigens (Appelmelk, 1999). This antigenic variation is presumed give a selective advantage to H. pylori during the changes that occur perpetually during colonization and persistent infection (Appelmelk, 1999).

Compared to other bacteria, a relatively large number of R-M systems were identified in the genome sequences of H. pylori 26695 and J99 (Tomb, 1997; Alm & Trust, 1999). Several H. pylori R-M system genes display phase variation, which may give a selective advantage by preventing or facilitating DNA uptake (Aras, 2002). For example, a putative R-M system displays phase variation at the translational level as well as at the transcriptional level. Regulation of translation of the putative R-M genes is organized via slipped-strand mispairing at a C-tract in the 5′-coding region of the restriction endonuclease encoding gene, switching translation on or off (Fig. 6). Regulation of transcription is seen in both the restriction endonuclease gene and the methyltransferase encoding gene through an as yet unidentified mechanism. Presumably, regulation at the transcriptional level adds to an efficient use of energy (de Vries, 2002).

Figure 6

A schematic representation of phase variation through slipped-strand mispairing. In this example, a gene with a homopolymeric cytosine (C) stretch in the 5′-region of its coding region is given. Depending on the number of cytosines present, the translation of this gene results either in an intact protein (in this example, 14 cytosines) or a truncated protein through the introduction of a stop codon (in this example, 13 cytosines).

Another example of a phase variable gene is the adhesin SabA. Helicobacter pylori cells expressing SabA will bind to the host Lewisx (Lex) antigen, thereby anchoring themselves to the epithelium. At sites where the local inflammatory response is very strong, those H. pylori subclones that do not express SabA due to frameshift mutation in a poly T/CT tract upstream or within the start codon, do not adhere to Lex on epithelial cells. As a consequence, these SabA-negative subclones can potentially leave a hostile environment using chemotaxis and flagellar motility, and thus evade the host immune response (Mahdavi, 2002).

Genes that show variation in repeat length and thus potentially are phase variable, are less ubiquitous among different H. pylori strains, compared to genes without such a repeat. This implies that phase variable genes cannot be essential under all conditions, as they can also be switched off (Mahdavi, 2002). In conclusion, phase variation, both at the transcriptional and translational level, adds to the extreme genetic diversity in H. pylori. This enables H. pylori to make fundamental adaptations, which thereby adds to long-term persistence in the gastric mucosa.

Bacterial determinants for the outcome of Helicobacter pylori infection

A summary of H. pylori proteins that are clearly associated with the development of H. pylori-related diseases are discussed later (see also Table 1). Factors that contribute to H. pylori colonization and persistence are also important for the pathogenesis of infection, and are also discussed later (see also Table 2).

View this table:
Table 2

Examples of Helicobacter pylori determinants for colonization and persistence

ProcessMain Helicobacter pylori factorReference(s)
Acid resistanceUreaseEaton & Krakowka (1994) and Tsuda (1994a, b)
Motility and chemotaxisFlagellaAkopyants (1995) and Eaton (1996)
CheYFoynes (2000)
AdhesionAdhesinsYamaoka (2002) and de Jonge (2004)
Nutrient acquisition
Nutrient uptakeComB4?Kavermann (2003)
FeoBVelayudhan (2000)
Degradation mucus/epitheliumPhospholipaseLangton & Cesareo (1992) and Bode (2001)
CollagenaseKavermann (2003)
NapAEvans (1995) and Obst (2000)
Immune evasion
Antigenic mimicryLe antigenLogan (2000)
Immune suppressionVacAMolinari (1998), Wang (2001), Gebert (2003) and Zheng & Jones (2003)

Helicobacter pylori disease determinants

Helicobacter pylori is unique in its ability persistently to colonize the human gastric mucosa and subsequently cause disease. So far, several H. pylori-specific proteins have been described that contribute to the development of H. pylori-related diseases. From the H. pylori point of view, these proteins are thought to contribute to its persistence. As a side-effect, these H. pylori proteins induce and alter inflammatory processes, damage the gastric epithelium, and thus determine the outcome after H. pylori infection.

CagA

Colonization with H. pylori isolates carrying the cytotoxin associated gene A (cagA) is associated with an increased risk of severe gastritis, peptic ulcer disease and adenocarcinoma of the distal stomach, when compared to patients infected with cagA-negative H. pylori isolates (Weel et al., 1996; Parsonnet, 1997; Graham & Yamaoka, 2000).

Presence or absence of the H. pylori cagA gene is often used as a marker for the presence or absence of the pathogenicity island (PAI) in H. pylori. The cag PAI consists of approximately 30 genes and is present in 50%–70% of all H. pylori isolates. Recombination between two 31-bp repeat sequences that flank the cag PAI may result in either the deletion or acquisition of the entire cag region (Censini, 1996). Eighteen genes on the cag PAI are essential for the production of a type IV secretion system (Fischer, 2001). Type IV secretion systems translocate (virulence associated) proteins from the bacterial cell into the host cell cytosol. This usually results in the initiation of a cascade of responses in the epithelial cell, such as cytoskeletal changes and secretion of the cytokine IL-8 (Christie & Vogel, 2000). The transcription of IL-8 is not only induced via the actions of the type IV secretion system, as genes outside the cag PAI are also responsible for the induction of IL-8 (Crabtree, 1995; Yamaoka, 2000). The pathway of IL-8 induction in gastric epithelial cells is not yet fully understood, but involves the extracellular signal-regulated kinase (ERK) pathway, which causes upregulation of the transcription factor NF-κB (Nozawa, 2002). Alternatively, the type IV secretion apparatus might translocate an unknown effector protein, which could activate the IL-8 signaling cascade (Fischer, 2001). Helicobacter pylori peptidoglycan, injected by the type IV secretion system, might be the key effector molecule responsible for this IL-8 induction (Viala, 2004).

In H. pylori, the components of the type IV secretion system translocate the CagA protein into gastric epithelial cells (Odenbreit, 2000). Subsequently, the CagA protein is phosphorylated. This phosphorylated CagA then interacts with the host phosphatase SHP-2, causing rearrangements of the cytoskeleton. This results in host cell morphological changes, also termed a ‘hummingbird’ phenotype. This phenotype is characterized by cell spreading, elongated growth of epithelial cells, and the presence of lamellipodia and filopodia (Segal, 1996, 1999; Backert, 2001). As a counteracting event, phosphorylated CagA interacts with the host Src kinase that attenuates signaling of the host SHP-2 phosphatase. This process decreases the cytoskeletal rearrangements and prevents further phosphorylation of CagA (Tsutsumi, 2003).

Recently, it has been demonstrated that H. pylori strains that induce higher levels of CagA phosphorylation in epithelial cells induce more cytoskeletal changes, and are more likely to be associated with gastric cancer (Argent, 2004). The differences in CagA phosphorylation levels are caused by differences in the number of tyrosine phosphorylation motifs (TPMs) within the 3′ region of the cagA gene (Yamaoka, 1998; Azuma, 2002; Argent, 2004). Recombination between repeat sequences that encode these TPMs affects the number of TPMs.

The development of atrophic gastritis diminishes local H. pylori survival rates (Karnes, 1991). In atrophic niches, acid secretion is low. To prevent further decline of the H. pylori population, minimal immune recognition is required. Therefore, in atrophic niches, subclones that have reduced numbers of these TPMs induce a less intense host immune response (Argent, 2004). On the other hand, higher levels of CagA phosphorylation induce a strong inflammation response. This inflammation causes atrophy and may potentially play a role in limiting acid stress in niches with high acid output (Argent, 2004).

In Asian populations, presence of the cagA gene is not associated with an increased risk for H. pylori-related diseases. In contrast to Western countries, where the cagA gene is found in approximately 50% of the H. pylori strains, the cagA gene is almost universally present in Asian H. pylori strains. This implies that the presence or absence of cagA alone is not enough to distinguish between virulent and nonvirulent H. pylori isolates (Pan, 1997). This is substantiated by the observation that H. pylori strains lacking the cagA gene have been isolated from patients with peptic ulcer or gastric cancer, albeit at lower frequency than cagA-positive H. pylori strains (Kuipers, 1995; Blaser & Crabtree, 1996; Van Doorn, 1998).

Thus, the association between cagA and H. pylori-related diseases varies in different geographical regions. This variation may be linked to differences in CagA expression. In East-Asian H. pylori isolates, CagA shows stronger binding to the host SHP-2 phosphatase, and CagA has a greater biological activity compared to Western CagA. This may induce greater phosphorylation levels, and a more intense immune response. The increased inflammation results in more severe atrophy, which may underlie the higher incidence of gastric carcinoma in East-Asian countries (Azuma, 2004). The presence or absence of cagA, however, is not associated with H. pylori-related cancer risk in East-Asian countries.

CagA not only induces host cell morphological changes through its interaction with SHP-2. CagA also associates with the epithelial tight-junction scaffolding protein ZO-1 and the transmembrane protein junctional adhesion molecule JAM (Amieva, 2003). This interaction prevents the assembly of tight-junction components at sites of bacterial attachment. This alters the composition and function of the apical-junctional complex, such that the epithelial barrier function is disrupted (Amieva, 2003). Presumably, by degrading the epithelial cell barrier, H. pylori increases its access to essential nutrients, such as Fe3+ and Ni2+. However, CagA is not the only H. pylori protein responsible for the barrier dysfunction, while this process is not lost upon deletion of cagA. Indeed, the VacA protein, discussed in the next section, is also important.

VacA

Eukaryotic cells were originally shown to undergo vacuolation and consecutive degeneration in an in vitro test system with H. pylori (Leunk, 1988). This phenomenon was later shown to be induced by an H. pylori toxin, now termed vacuolating cytotoxin A (VacA) (Cover & Blaser, 1992; Telford, 1994). Virtually all strains express VacA, but those strains expressing active forms (approximately 50%) are more frequently associated with disease (Atherton, 1995; Figura, 1997). This association is observed in Western countries, but not in Asian countries (Pan, 1997; Van Doorn, 1998; Yamaoka, 1999).

The expression levels of VacA depend on sequence variations in the signal sequence (s1a, s1b, s1c or s2) and the mid-region (m1 or m2) (Atherton, 1995). Vacuolating activity is higher in s1/m1 strains compared to s1/m2 strains, and absent from s2/m2 strains in Western populations (Atherton, 1995). Different H. pylori isolates thus carry different combinations of the signal sequence and mid-region, indicating recombination between vacA of different H. pylori strains (Atherton, 1995; Letley & Atherton, 2000; Falush, 2001). The H. pylori vacA s1/m1 genotype is linked with the cagA-positive genotype even though the vacA and cagA genes are not closely located on the H. pylori chromosome. Helicobacter pylori isolates positive for both cagA and vacA s1/m1 correlate with ulcer disease and gastric cancer (Eaton, 1997; Salama, 2001). Therefore, either vacA or cagA cannot be independently used as a determinant for clinical outcome (Van Doorn, 1998).

Helicobacter pylori VacA is a highly immunogenic, 95-kDa protein. VacA binds to the apical portion of epithelial cells, forming anion-selective pores. Through these pores, bicarbonate, chloride and urea are released from the cell cytosol. VacA is then endocytosed into late endosomal compartments, and alters the permeability of these compartments. This causes water influx, and hence vesicle swelling and vacuole formation, in the presence of weak bases such as ammonia. Impairment of the intracellular endocytic pathway eventually results in cell death (Cover & Blaser, 1992; Molinari, 1997; Satin, 1997; Reyrat, 2000; Montecucco & Rappuoli, 2001; Cover, 2003; de Bernard, 2004; Fischer, 2004). This causes a decrease in transepithelial cell resistance, and low molecular weight molecules such as Fe3+ and Ni2+ can thus easily traverse the epithelial cell layer. The Fe3+ and Ni2+ molecules are essential factors for H. pylori growth and urease activity, respectively (Papini, 1998; Pelicic, 1999). Thus, by degrading the epithelial cell barrier, H. pylori is presumed to more easily gain access to these essential factors.

Helicobacter pylori VacA is also associated with the degradation of parietal cells, which results in a reduction in acid secretion (Kobayashi, 1996; Neu, 2002). This event predisposes the host to the development of gastric cancer. Differences in VacA expression levels during persistent infection result in altered toxicity (Aviles-Jimenez, 2004). The waxing and waning of ulcers may hence be explained by the altered expression of VacA over time (Aviles-Jimenez, 2004).

Transcription of the vacA gene is induced near gastric epithelial cells in vitro. This finding confirms the importance of VacA interaction with host cells (van Amsterdam, 2003). In addition, VacA can escape the adaptive immune response, thereby enhancing H. pylori persistence in the gastric mucus layer (Molinari, 1998; Wang, 2001; Gebert, 2003; Zheng & Jones, 2003).

Helicobacter pylori outer membrane proteins (HOPs)

In Gram-negative bacteria, the outer membrane mediates the interaction with their surroundings. During infection, proteins present on the outer membrane of H. pylori are assumed to be altered in such a way that recognition by the host immune system is minimal. Helicobacter pylori isolates contain about 30 different outer membrane proteins (HOPs) (Alm, 2000). Several of these HOPs have been implicated as adhesins (Ilver, 1998; Odenbreit, 1999; Peck, 1999). In Gram-negative bacteria, adhesins most often form polymeric pilus structures (Remaut & Waksman, 2004). However, in H. pylori these adhesins are presumed to have adapted to the gastric environment, where acidic conditions would probably depolymerize such polymeric pilus structure (Alm, 2000). Both fucosylated glycoproteins and sialylated glycolipids have been shown to be binding sites for H. pylori in the gastric epithelium (Boren, 1993; Mahdavi, 2002).

Adherence to host cells is thought to protect H. pylori against peristalsis and mucosal shedding. Moreover, adherence is presumed to allow H. pylori better access to nutrients released from the gastric epithelium (Papini, 1998; van Amsterdam & van der Ende, 2004b) and deliver bacterial toxins to the host cells more efficiently (Gerhard, 1999). On the other hand, at sites of vigorous inflammation, loss of adherence properties may enable H. pylori to escape killing by host immune cells. Thus far, several HOPs have been associated with disease development (Guruge, 1998; Gerhard, 1999; de Jonge, 2004b; Lehours, 2004b).

Helicobacter pylori BabA (HopS) mediates H. pylori adhesion to human Leb blood group antigens, which are present on gastric epithelial cells. BabA adhesion facilitates colonization of H. pylori and increases IL-8 secretion from epithelial cells. This results in enhanced mucosal inflammation (Rad, 2002). Expression of BabA is significantly associated with duodenal ulcer and adenocarcinoma in H. pylori-infected patients in Western countries, but not in Japanese individuals (Gerhard, 1999; Mizushima, 2001). Transgenic mice expressing the human receptor for BabA, Leb, are more likely to develop gastritis and atrophy after infection with BabA-positive strains compared to wild-type, Leb-negative mice (Gerhard, 1999).

There are two distinct babA alleles: babA1 and babA2, but only the babA2 allele encodes a full sized, active bacterial adhesion protein. Helicobacter pylori babA2 contains a CT dinucleotide repeat motif in its 5′-coding region (Pride, 2001). Slipped strand mispairing during replication switches BabA expression on or off. Moreover, the babB gene, which encodes a protein with unknown function, contains a 3′ conserved region highly similar to that of babA (Pride & Blaser, 2002). Gene conversion may replace babA for babB, as was shown in the macaque model (Solnick, 2004). As a result, H. pylori lose BabA expression, and no longer adhere to host Leb. Expression of BabA is thus regulated via both phase variation (slipped strand mispairing) and antigenic variation (gene conversion). Selection of BabA variants with increased or decreased bacterial adherence may contribute to adaptation to its host, i.e. alleles that are best suited to their respective host may be selected (Aspholm-Hurtig, 2004). Both mechanisms of babA regulation contribute to the dynamic response designed to promote chronic infection.

Cycles of selection for increased or decreased adherence during chronic inflammation attains all types of Leb binding in populations where all blood group types are represented. However, in South America, almost all Amerindians have blood group O phenotype (Aspholm-Hurtig, 2004). As a consequence, binding of host Leb to the other blood group types was lost during recurrent cycles of selection for binding activities. Compared to babA-positive H. pylori from various geographic regions, babA-positive H. pylori from Amerindians bind approximately 1500-fold better to blood group O Leb compared to blood group A/B Leb (Aspholm-Hurtig, 2004). Thus, the host mucosal glycosylation patterns result in BabA evolution. This fine-tunes H. pylori strains to their individual host, which helps to avoid host responses, and contributes to the extraordinary chronicity of H. pylori infection world-wide (Aspholm-Hurtig, 2004).

Helicobacter pylori OipA (HopH) expression is associated with high H. pylori densities, increased IL-8 production from gastric cancer cell lines, more severe gastritis, and duodenal ulcer formation (Yamaoka, 2000, 2002). The oipA gene is present in virtually all H. pylori isolates, but its expression is modulated by phase variation via slipped strand mispairing of a variable number of CT-dinucleotide repeats in the 5′-region of oipA (Yamaoka, 2000). Almost all H. pylori isolates with a functional oipA possess cagA. Therefore, the observed association of H. pylori oipA with severe gastritis may be affected by the presence of cagA (Ando, 2002). However, inactivation of both the cag PAI and oipA is required to reduce IL-8 levels to baseline levels. Therefore, both the cag PAI and oipA gene products are presumed to be associated with H. pylori pathogenicity (Yamaoka, 2000).

Helicobacter pylori-induced gastritis promotes the expression of sialyl-Lex antigens in the gastric mucosa (Mahdavi, 2002). The SabA protein binds host sialyl Lex glycoproteins, and may hereby colonize inflammation-activated sites of the sialylated epithelium (Mahdavi, 2002). Levels of sialylated glycoconjugates are low in healthy gastric mucosa, but are increased in gastric mucosa of H. pylori-infected persons. It is unclear whether this is a direct effect of the infection or the associated chronic inflammation. In addition, high levels of sialylated glycoconjugates are detected in dysplastic gastric mucosa and gastric cancer tissues (Sipponen & Lindgren, 1986; Amado, 1998). Sialylated glycoconjugates may be beneficial for H. pylori attachment, as sialylation is linked with the capacity for SabA-dependent H. pylori binding in situ (Mahdavi, 2002). Slipped strand mispairing in a CT repeat in the coding region of SabA may result in the on/off switch of SabA expression. However, a possible association between SabA expression and H. pylori-related diseases was not found (de Jonge, 2004b). In contrast, the absence of H. pylori SabB expression is significantly associated with duodenal ulcer disease (de Jonge, 2004b). The function of SabB is unknown, but is presumed to be similar to that of the adhesin SabA.

If indeed SabB is an adhesin, the presence of SabB may allow a more intimate contact between H. pylori and the gastric epithelial cells. This suggests that SabB-mediated adhesion renders H. pylori more susceptible to killing by the host's immune system (de Jonge, 2004b). During severe inflammation, as seen in duodenal ulcer, absence of SabB may be advantageous to escape the immune system of its host. Alternatively, binding of SabB may induce production of anti-inflammatory cytokines leading to reduced inflammation (de Jonge, 2004b).

The adhesins AlpA or AlpB favor H. pylori colonization in a guinea pig model of infection (de Jonge, 2004). After coinfection of guinea pigs with a H. pylori alpA or alpB knockout mutant together with their wild-type H. pylori strain, H. pylori alpA and alpB knockout mutants are outcompeted. Presumably, an inability to adhere to the gastric epithelial cells clears H. pylori alpA or alpB mutants from the stomach. When an H. pylori alpA knockout mutant is given as a prophylaxic vaccine, the H. pylori load reduces significantly. However, when the same alpA knockout mutant is given therapeutically, it does not reduce the bacterial load (Sanchez, 2001). Therefore, it has been suggested that during the early stages of infection AlpA is important. Indeed, transcription of alpA is higher early in infection, supporting this hypothesis (Rokbi, 2001).

The operon carrying the alpA and alpB genes is present in virtually all H. pylori isolates (Odenbreit, 1999). However, alpA and alpB are not found in closely related Helicobacter species, such as H. felis or H. mustelae. Therefore, H. pylori AlpAB may be important determinants for colonization of the human gastric mucosa and thus for host tropism (Odenbreit, 1999).

The adhesins encoded by oipA, hopZ, hopO, hopP and BabB are all thought to be regulated by phase variation through slipped strand repair mechanisms. Individually, these adhesins are not essential for colonization. However, the expression of two or more of these proteins enhances the ability of H. pylori to colonize (Yamaoka, 2002). Still, the importance of adhesins in H. pylori colonization in vivo is difficult to determine, as different H. pylori strains express a different spectrum of adhesins, and different hosts also show variability in the expression of their adhesin receptors.

Helicobacter pylori plasticity region virulence factors

Comparison of the genome sequences from two individual strains has shown that approximately 6%–7% of the H. pylori genes present in one strain are absent from the other and vice versa (Alm, 1999). About half of the strain-specific genes are found in a hyper variable region; the plasticity zone. Genes located in such plasticity regions are often associated with increased virulence (Alm & Trust, 1999). In H. pylori, the plasticity zone indeed encodes Vir-type ATPases, which are involved in type IV secretion. In H. pylori, this secretion system is important in, for example, secretion of the CagA cytotoxin (Odenbreit, 2000). So far, several H. pylori genes located in the plasticity zone have been described that are linked to H. pylori-related diseases.

Helicobacter pylori Jhp0947 and Jhp0949 are associated with duodenal ulcer disease (de Jonge, 2004a). Both genes are located on the H. pylori strain J99 plasticity zone, and are important in H. pylori-mediated IL-12 release from monocytes in an in vitro test system (de Jonge, 2004a). The cytokine IL-12 is essential for skewing the immune reaction toward a proinflammatory Th1 response, and is strongly associated with duodenal ulcer formation (Bauditz, 1999; Trinchieri, 2003). The function of the proteins encoded by jhp0947 and jhp0949 is unknown. However, it has been hypothesized that these proteins induce the production of the proinflammatory cytokine IL-12, through the interaction with monocytes. Increased IL-12 levels cause differentiation of naive T-cells into activated Th1 cells, leading to an augmented inflammation response and increased tissue damage (Trinchieri, 2003). Santos and colleagues found an association between Jhp0947 and the presence of the cag PAI. Therefore, Jhp0947 was considered a surrogate marker for disease (Santos, 2003). However, de Jonge and colleagues (Salama, 2001) determined that neither Jhp0947 nor Jhp0949 were linked to cagA presence, and thus are independent markers for the development of duodenal ulcers. The difference in an association of Jhp0947 with cagA might relate to differences between H. pylori isolates from different geographic regions, as reported for other virulence factors (Yamaoka, 1999; de Jonge, 2004b). This underlines the notion that the overall effects of H. pylori-factors in disease development are complex and predictions on the clinical outcome of H. pylori infection may be ambiguous.

Helicobacter pylori Jhp0950, encoding a protein with unknown function, is also part of the J99 plasticity zone (Alm & Trust, 1999). The presence of this gene is associated with cagA, cagE, vacA s1m1, babA2, hopQ T1 and oipA on (Lehours, 2004b). The gene cluster comprising Jhp0950, iceA1 and sabA ‘on’ is significantly more prevalent among H. pylori isolates obtained from patients with a MALT lymphoma than among isolates from patients with gastritis (Lehours, 2004a). Jhp0950 is found adjacent to Jhp0949 in the H. pylori J99 genome. Jhp0950, unlike Jhp0949, is not associated with an increased risk for duodenal ulcer (Alm, 1999; de Jonge, 2004a; Lehours, 2004a). Still, its association with disease supports the notion that H. pylori genes located within the plasticity region may be associated with virulence (Alm & Trust, 1999).

Other Helicobacter pylori virulence factors

In the U.S.A. and the Netherlands, H. pylori iceA1, one of the two allelic variants of the iceA gene, is more prevalent among H. pylori-infected patients with peptic ulceration than among H. pylori-infected patients with gastritis only (Peek, 1998; Van Doorn, 1998). Similar to many other virulence-associated genes, a linkage between IceA expression and H. pylori-related diseases is population-dependent. In South-East Asia no association between the presence of iceA1 and H. pylori-related diseases was observed (Yamaoka, 1999). Interestingly, South-African H. pylori isolates positive for iceA1 are more prevalent among H. pylori-infected patients with gastric cancer than among H. pylori-infected patients with gastritis only (Kidd, 2001). Expression of IceA1 is linked with vacA s1, questioning the significance of IceA1 in H. pylori pathogenesis.

Helicobacter pylori iceA1 was originally identified following transcriptional up-regulation after contact with gastric epithelial cells (Peek, 1998). The iceA1 gene encodes an endonuclease with strong similarity to a restriction endonuclease NlaIIIR in Neisseria lactamica (Figueiredo, 2000; Xu, 2002). Activity of the accompanying highly conserved DNA adenine methyltransferase, encoded by hpyIM, seems to be involved in controlling gene expression in H. pylori (Donahue, 2002). Hence, associations between icA1 and adherence or long-term colonization may be a consequence of the activity of the methylase HpyIM (Dubois, 1996; Peek, 1998). In contrast to iceA1, iceA2 has no significant homology to known proteins. IceA2c strains are more prevalent among patients with asymptomatic gastritis and nonulcer dyspepsia (Van Doorn, 1998), whereas isolates from peptic ulcer patients more often harbor iceA2d (Kidd, 2001).

Helicobacter pylori Jhp1462 is associated with an increased risk for MALT lymphoma, duodenal ulcer and gastric adenocarcinoma (Lehours, 2004a). The function of the protein encoded by Jhp1462 is unknown and its importance in the development of severe gastro-duodenal diseases remains to be elucidated.

Helicobacter pylori determinants for colonization and persistence

Helicobacter pylori chronically infects the human gastric mucosa prior to disease. Therefore, factors contributing to H. pylori colonization and persistence are inherently linked to H. pylori-related pathogenesis. Upon entrance of the gastric lumen, H. pylori has to cope with gastric acidity. Helicobacter pylori survives this acidity through its acid resistance: it then traverses the mucus layer to reach its niche close to the gastric epithelial cells. Chemotaxis, motility and adhesion are important processes in the colonization of the gastric epithelial cells.

For long-term persistence in the human gastric mucosa, H. pylori also continuously requires nutrients from its host. This is achieved through degradation of the integrity of the mucus layer and the underlying gastric epithelial cells. Moreover, clearing of H. pylori is avoided by suppression of the host immune system, antigenic variation and antigenic mimicry.

Acid resistance

The gastric lumen has a pH value that is mostly below 4. However, the lower pH limit for H. pylori growth is 5.0-5.5 depending on the isolate tested (Bijlsma, 2002). In the gastric mucous layer, the pH is also acidic, varying between approximately 4 and 6.5. Moreover, the pH might occasionally drop when the mucous layer is damaged (Schade, 1994). Survival of H. pylori after acid shocks (pH<3) depends on the activity of the H. pylori protein urease, which converts urea into ammonia and bicarbonate (Burne & Chen, 2000). This results in neutralization of the periplasm or cytoplasm of H. pylori (Bury-Mone, 2004; van Vliet, 2004b). Helicobacter pylori urease activity is essential for survival at acidic pH in vitro, and for colonization of the gastric mucosa in animal models, demonstrating the importance of urease in H. pylori infections (Eaton & Krakowka, 1994; Tsuda, 1994a, b).

The urease gene cluster of H. pylori consists of the ureAB genes, which encode the two subunits of the urease enzyme, and downstream the ureIEFGH genes (Labigne, 1991). The UreI protein functions as a channel that allows passage of urease, whereas the UreEFGH proteins are presumed to function in the assembly of the UreA and UreB subunits. Another presumed function of the UreEFGH proteins is the incorporation of nickel into urease. Hereby, the activity of urease is increased, and the proton-gated urea channel UreI is opened, allowing influx of urea (Weeks, 2000; van Vliet, 2001; Scott, 2002).

Opening of the UreI channel occurs fast. This allows rapid entry of urea into the cytosol and subsequently its neutralization. More slowly, transcriptional and posttranscriptional regulation of urease assembly also contribute to H. pylori acid resistance (Akada, 2000; van Vliet, 2004b). The combination of a urease that is highly active at neutral pH, and an acid-regulated urea channel explains why H. pylori is unique in its ability to inhabit the human stomach. Effective inhibition of UreI would provide a means of eradicating H. pylori from the normal, acid-secreting stomach (Weeks, 2000).

Even though urease is essential for the survival of acid shocks, H. pylori urease is not sufficient when dealing with the chronically acidic environment near gastric epithelial cells. Treatment with acid suppressants does not restore the ability of urease-negative mutants to colonize the stomachs of gnotobiotic piglets (Eaton & Krakowka, 1994). This indicates, that in addition to acid resistance, urease serves other essential functions associated with gastric colonization. Urease is involved in H. pylori nitrogen metabolism (Williams, 1996), and the ammonia and bicarbonate produced by urease activity are thought to influence many cellular processes in the host, including cell lysis (Smoot, 1990; Kuwahara, 2000). In addition to urease, nonurease genes also contribute to acid resistance, and survival and growth at acidic pH are independent of urease (Bijlsma, 1998, 2000; Merrell, 2003a). Among those nonurease genes, three regulatory proteins (Fur, NikR, and HP0166) have been shown to contribute to H. pylori acid adaptation (Bijlsma, 2002; Bury-Mone, 2004; van Vliet, 2003a, 2004a, b). This shows the importance of tight regulation of acid resistance mechanisms.

Motility and chemotaxis

A thick layer of mucus covers the gastric epithelial cell layer. This layer is assumed to be relatively impermeable to even small molecules (Wallace & Granger, 1996). Helicobacter pylori motility and chemotaxis are essential to penetrate this mucus layer efficiently. Mutagenesis of just about any gene of the motility and chemotaxis systems abolishes the ability of H. pylori to infect the gastric mucosa in a mouse model of infection, stressing the importance of these factors for colonization (Kim, 1999; Foynes, 2000; Ottemann & Lowenthal, 2002; Kavermann, 2003).

Helicobacter pylori motility is accomplished by the presence of two to six polar, sheathed flagella. A flagellum is composed of three structural elements: a basal body, a hook and a filament. So far, over 50 proteins have been described that are involved in either structural or regulatory organization of the flagellum (Tomb, 1997; Alm, 1999; Niehus, 2004; van Amsterdam & van der Ende, 2004a). The flagellar filament is composed of two subunits, FlaA and FlaB (Suerbaum, 1993). Helicobacter pylori flaA and flaB mutants can not colonize a piglet model of infection, showing that H. pylori flagella are essential for colonization, and hence H. pylori virulence (Akopyants, 1995; Eaton, 1996). The H. pylori motB gene encodes a flagellar motor rotation protein. A H. pylori motB knockout mutant still contains flagella, but is nonmotile. This mutant is reduced in its ability to colonize mice, confirming that motility is required for full infectivity (Ottemann & Lowenthal, 2002).

Regulation of flagellar gene transcription is different from many other bacteria (Niehus, 2004). In contrast to many other flagellated Gram-negative bacteria, H. pylori does not possess a transcriptional master regulator for the regulation of flagellar gene transcription. Presumably, H. pylori motility does not require tight regulation as it is indispensable at all times. Still, some transcriptional regulation is seen in H. pylori. For example, the proteins FlhA and FlhF, components of the basal body, need to be present before transcription of middle and late flagellar genes can occur (Niehus, 2004).

Besides motility, chemotaxis is also essential for H. pylori colonization. Chemotaxis allows H. pylori to direct its movement towards the gastric epithelial cell layer (Foynes, 2000; Worku, 2004). An H. pylori chemotaxis receptor sensor encoded by HP0099 recognizes bicarbonate and arginine as attractants. This sensor protein on the H. pylori outer membrane is coupled to the response regulator CheY (Foynes, 2000; Pittman, 2001). The response regulator CheY regulates the expression of the flagellar motor proteins that enable H. pylori to swim toward its niche (Tomb, 1997; Cerda, 2003). A knockout mutant of the response regulator CheY failed to colonize gnotobiotic piglets, emphasizing the importance of chemotaxis in initial colonization (Foynes, 2000). Knockout mutants of the response regulator CheY and its histidine kinase display impaired swarming and reduced chemotactic responses to mucin.

Adhesion

Several (putative) adhesins have been associated with H. pylori-related pathogenesis and are described in a previous section.

Nutrient acquisition

Helicobacter pylori is highly limited in its capacity to oxidize carbon substrates (Doig, 1999). Access to these and other nutrients may be achieved directly by the active uptake of these nutrients. Moreover, degradation of the mucus layer as well as the underlying gastric epithelium either directly or indirectly may supply H. pylori with essential nutrients.

The acquisition of several host factors is essential for H. pylori colonization. The significance of glutamate, proline and α-ketoglutarate uptake in H. pylori colonization has been demonstrated in Mongolian gerbils as well as in vitro infection experiments (Stark, 1997; Kavermann, 2003). A putative transport ATPase, ComB4, is essential for colonization in Mongolian gerbils (Kavermann, 2003). ComB4, part of the type IV secretion cluster, is essential for the uptake of DNA (Hofreuter, 2001). However, a comB8-10 mutant strain, deficient in DNA uptake by natural transformation, was able to colonize Mongolian gerbils (Kavermann, 2003). Therefore, natural transformation competence, i.e. DNA uptake, does not seem to be essential for short-term animal colonization. This finding suggests that ComB4, in contrast to ComB8-10, has another function in addition to mediating DNA uptake, supposedly the uptake of nutrients (Kavermann, 2003).

Similar to other pathogens, the ability to acquire host iron contributes to the survival and persistence of H. pylori. Ferrous iron uptake occurs predominantly via FeoB (Velayudhan, 2000). Helicobacter pylori feoB mutants are unable to colonize the gastric mucosa of mice, showing the importance of iron uptake in H. pylori colonization (Velayudhan, 2000). This is further indicated by the importance of the ferric uptake regulator Fur in H. pylori colonization in mice (Bury-Mone, 2004). In addition to iron, nickel is also of crucial importance for H. pylori virulence, as inactivation of nikR reduces the level of colonization by H. pylori. Hence, both the Fur and NikR gene regulators play a role in H. pylori colonization and persistence (Bury-Mone, 2004; van Vliet, 2004b).

Helicobacter pylori phospholipases may damage the gastric mucus layer by degrading phospholipids, which may supply the bacteria with essential nutrients, including inorganic phosphorus (Langton & Cesareo, 1992; Bode, 2001). It is thus tempting to speculate that the increased phospholipase expression seen in H. pylori-infected patients when compared to H. pylori-negative individuals, may contribute to H. pylori-related pathogenesis. Phospholipase A activity in a number of epithelial cell lines exposed to a variety of H. pylori strains remained unchanged (Nardone, 2001). However, phospholipase C activity was significantly higher in patients with gastric ulcers compared to patients with gastritis only (Bode, 2001). The fact that phospholipase C is produced by all H. pylori strains seems to indicate that this enzyme is essential for H. pylori survival (Bode, 2001). Possibly, phospholipase C may contribute to the higher virulence of a subset of H. pylori strains in peptic ulcer patients.

Helicobacter pylori might use the degradation of collagen as a source of certain amino acids or short peptides. The gene prtC, encoding a collagenase, is essential for H. pylori colonization in Mongolian gerbils (Kavermann, 2003). Alternatively, collagenase might be important for the degradation of components of the immune system, such as IgA antibodies or components of the complement system. Hence, H. pylori could evade the immune system, similar to the method described for other bacteria (Pohlner, 1987).

The immune response is stimulated in H. pylori infection, which damages the gastric mucosa. As a result, nutrients are released from the damaged epithelium. Helicobacter pylori factors that contribute to this increased inflammation might thus be essential for H. pylori survival in the gastric niche. The neutrophil activating protein NapA is chemotactic for neutrophils and promotes their adhesion to endothelial cells (Evans, 1995). As a consequence, neutrophils are activated, leading to their apoptosis and release of reactive oxygen species (Obst, 2000). Vaccination studies in mice have shown that anti-NapA activity protects against H. pylori infection (Satin, 2000). The ability to activate neutrophils varies between different H. pylori isolates. Neutrophil activation in H. pylori infection is linked to peptic ulcer disease, but not to the expression of NapA or to differences in napA sequences (Leakey, 2000). Thus, other H. pylori proteins that may be responsible for neutrophil activation remain to be elucidated.

NapA is encoded by the napA gene, ubiquitously present in H. pylori isolates (Evans, 1995). Helicobacter pylori napA is presumed to possess a Fur-regulated promoter, indicating that NapA is important for iron storage (Evans, 1995; Dundon, 2001). By inducing inflammation and subsequently degradation of the epithelial barrier, NapA may promote nutrient (i.e. iron) availability. However, NapA is constitutively expressed in either the presence or absence of iron, indicating that NapA is not involved in the regulation of iron metabolism (Dundon, 2002).

Thus, uptake of nutrients through degradation of the gastric epithelium and mucosa both directly and indirectly via the activation of the immune system may supply H. pylori with nutrients essential for its survival and growth, and enhance its virulence.

Immune evasion

During the long course of colonization, hosts develop a robust immune response to numerous H. pylori antigens. However, these antigens are often not surface exposed. Surface exposed antigens, on the other hand, are often highly diverse, making clearing of H. pylori by the host more difficult (Wirth, 1996; Pride, 2001). Indeed, H. pylori infection seldom (if ever) results in an adequate immune response capable of removing the bacteria. In fact, H. pylori is capable to evade both the innate and adaptive immune response efficiently.

The innate immune response is unable to substantially recognize H. pylori. The E. coli lipopolysacharide (LPS) activates the innate immune response via TLR4. However, H. pylori LPS is far less potent inducer of TLR4-mediated gene expression (Backhed, 2003). TLR5-mediated activation of the immune response is also less efficient in H. pylori. Helicobacter pylori flagellins are not released and are also less potent in activating IL-8 secretion via TLR5 when compared to Salmonella enterica serovar Typhimurium flagellins (Lee, 2003; Gewirtz, 2004). Failure of the innate immune response to recognize H. pylori may result in a failure of the adaptive immune response to subsequently clear the infection.

Evasion of the adaptive immune system is mediated by at least two important mechanisms: antigenic mimicry and immune suppression. Antigenic mimicry results in an inability of the host immune system to properly recognize H. pylori. For example, Le blood group antigens on H. pylori LPS are also expressed on human epithelial cells. As a consequence, H. pylori LPS is poorly immunogenic, thus supporting H. pylori persistence (Shirai, 1998). Helicobacter pylori strains lacking the O antigen on their LPS are significantly reduced in their capacity to colonize mice (Logan, 2000). Hence, the typical structure of H. pylori LPS is important in colonization.

Besides immune evasion through antigenic mimicry, immune suppression also contributes toward the chronicity of H. pylori infection. The immune response may be either skewed towards a more anti-inflammatory response, or downregulated (VacA). Helicobacter pylori VacA can block phagosome-lysosome fusion in macrophages, resulting in failure of macrophages to kill H. pylori (Zheng & Jones, 2003). Furthermore, VacA expression blocks T-cell activation and hence proliferation, activates T-cell apoptosis, and inhibits antigen presentation in T cells. In addition, antigen processing by B lymphocytes is blocked, possibly by limiting the maturation of endosomes to MHC class II compartments where antigen loading takes place (Molinari, 1998). Hence, VacA limits the T- and B-cell response against H. pylori (Molinari, 1998; Wang, 2001; Gebert, 2003).

Concluding remarks

Helicobacter pylori chronically infects the human gastric mucosa, resulting in gastritis. Infection is linked to the development of gastric and duodenal ulcers as well as MALT lymphomas. Moreover, H. pylori infection is a risk factor for gastric adenocarcinoma. Differences between hosts and their environment, as well as differences between H. pylori isolates, determine disease outcome. As the pattern of the gastric inflammation is highly predictive, one might argue that the differences between H. pylori strains only have minor effects on disease outcome (Graham, 1997; Graham & Yamaoka, 2000). On the other hand, H. pylori factors may determine the primary site of colonization, and hence the pattern of gastric inflammation and ultimately disease outcome. Simultaneous colonization of one host with multiple strains and genetic variation of H. pylori isolates during chronic infection however, make H. pylori-related pathogenesis difficult to predict from individual factors. Still, the majority of infected individuals (80–90%) do not develop disease, and coevolution of H. pylori with their hosts enables a life-long colonization.

Footnotes

  • Editor: Mike Koomey

References

View Abstract