Article Text

Download PDFPDF

H pylori and Lewis antigens
  1. Vrije Universiteit, Medical School
  2. Department of Medical Microbiology
  3. van der Boechorststraat 7
  4. 1081 BT Amsterdam, Netherlands
  5. Email:{at}

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

See article on page 18

Lipopolysaccharide (LPS) of manyHelicobacter pylori strains expresses Lewis antigens (Lex, Ley, Lea, Leb) which are similar to those expressed by gastric epithelial cells (“molecular mimicry”).1 In addition,H pylori LPS displays phase variation in these antigens—that is, the high frequency, reversible switching of phenotype2; for instance, a strain that expresses Lex may yield variants that express Ley. As yet, no definite role has been assigned to these Lewis antigens, nor to phase variation, in the pathogenesis of gastric disease.

In this issue of Gut (see page 18), Zheng and colleagues3 report that H pylori strains isolated from Asian peptic ulcer patients express two or more Lewis antigens more often than strains from non-ulcer dyspepsia patients (89.6 v 73.2%; p=0.035). What could be the link between H pyloriLewis antigen expression and the development of pathology in the host?

Lewis antigens induce pathogenic antibodies. On infection, H pylori LPS may induce anti-Lex/y antibodies that bind not only to bacteria but also to host gastric epithelium, followed by complement fixation and tissue destruction.1 Indeed, H pylori infection of mice induces autoreactive anti-Lex/y antibodies.1 However, although high titres of antibodies to H pylori LPS are found in the serum of infected patients,1 these antibodies are not autoreactive and not directed against Lex/y (see Yokota and colleagues4).
Lewis antigen mimicry provides persistence through immune evasion. Analogous to ABO blood group antigens, a host that expresses Lex would be able to form anti-Leyantibodies but not antibodies directed at Lex. Consequently, Lex positive bacteria infecting a Lex positive host would escape immune attack and be able to persist while an Ley positive strain would not be able to persist. Experimental infection in monkeys confirms this concept: aH pylori strain isolated from Ley positive monkeys expressed mainly Ley; thesame strain expressed mainly Lexafter colonisation of Lex positive animals.5This indicates that expression of the H pylori Lewis phenotype depends on the host; adaptation can occur by means of phase variation2 followed by selection through anti-Lex/y antibodies. Two of three studies in humans, however, failed to demonstrate the existence of a correlation between the phenotypes of the host and pathogen.6 Moreover, it has been shown that strains expressing Lex, and strains expressing Ley, can be isolated from the same patient.7 Finally, a shift in H pylori Lex/y antigen expression would be driven by anti-Lex/y antibodies and there is no evidence that these are formed in infected patients.4
Lewis antigens are involved in adhesion and colonisation. Expression of Lex/y is crucial for in vivo colonisation of mice: mutants with inactivated β1,4-galactosyltransferase8 or α3-fucosyltransferase genes (S L Martin, submitted) expressed no Lex/y and colonised less well than their Lex/ypositive parent strains.

How would Lex/y expression affect colonisation? Recent data suggest that Lex/y plays a role in adhesion. A monoclonal antibody (Mab) specific for H pylori LPS inhibited adhesion to gastric epithelial cells9; this Mab was specific for Lex (B J Appelmelk, H Yamaguchi, unpublished). Strains knocked out in galE orrfbM, genes involved in the biosynthesis of Lex, did not adhere to gastric tissue sections.10 In addition, synthetic Lex bound to human gastric epithelial cells from some hosts10 but not from all (T Boren, unpublished). Studies in gastritis patients demonstrated that H pylori strains that strongly expressed Lex/y caused a higher colonisation density than strains that expressed Lex/yweakly.6 These data suggest that expression of Lex enhances colonisation through increasing adherence. They also predict the existence of a gastric Lex binding lectin. Experimental studies confirmed this: Lex binding polypeptides of 16–29 kDa are found in gastric epithelial cells11; the identity of these proteins is unknown. Independent studies have shown that surfactant protein D, a 120 kDa lectin belonging to the innate defence system and expressed in the stomach, can also interact with H pylori LPS12; which part of the LPS is recognised is unknown.

Thus a role for LPS/Lex/y in adherence seems likely. This role is not absolute: Lex/y negative mutants can adhere as strongly as their Lex/y positive parents (T Boren, unpublished), and Lex/y negative strains colonise human hosts well.13 Thus an Lex/y-lectin interaction may contribute to adhesion for only some strains and only in part of the hosts. For example, the adhesin BabA14 is important for adhesion of H pylori and recognises Leb expressed by gastric epithelium; onlyH pylori strains that do not express BabA or strains that colonise humans that do not express Leb might need their Lex/y antigens for adhesion. In this concept, LPS phase variation allows detachment of bacteria not expressing Lex/y and hence transmission to another host; subsequently, switch back variants expressing Lex/y adhere and colonise a new host.

Adhesion of H pylori has clinical relevance: strains from ulcer patients more often express BabA compared with strains from gastritis patients.14 What is the link between adherence and development of host pathology? Firstly, increased adherence may lead to an increased bacterial burden. Secondly, studies in mice showed that increased adherence did not necessarily lead to increased colonisation density but to closer contact between bacteria and gastric epithelial cells.15 A more intimate contact enhances the cross talk between microorganism and host and may lead to activation of transcription factor NF-κB and host signal transduction pathways. This induces IL-8 production and inflammation, and finally, ulceration. This sequence of events is in agreement with data that show that increased Lex expression in H pylori is associated with increased neutrophil infiltration.6

In summary, current data, including those provided by Zheng and colleagues,3 are in agreement with the hypothesis thatH pylori LPS Lewis antigens play a role in adhesion and inflammation; LPS phase variation may be essential for host-to-host transmission. To conclude, after several years of intensive research on H pylori LPS structure, genetics, and biosynthesis we may finally start to understand the biological role of H pyloriLewis antigens.


We thank N J High, S M Logan and D E Taylor for providing unpublished data.


Linked Articles