Article Text


Association of peptic ulcer with increased expression of Lewis antigens but not cagA, iceA, andvacA in Helicobacter pylori isolates in an Asian population


BACKGROUND Studies in Western populations suggest that cagA, iceA, and vacAgene status in Helicobacter pylori isolates is associated with increased virulence and peptic ulcer disease.

AIM To investigate the relationship between peptic ulcer and expression of Lewis (Le) antigens as well as cagA, iceA, and vacA inH pylori isolates in Singapore.

METHODS Expression of Le antigens in H pylori isolates obtained from patients with dyspepsia was measured by enzyme linked immunosorbent assay. The cagA, iceA, and vacAstatus was determined by polymerase chain reaction.

RESULTS Of 108H pylori isolates, 103 (95.4%) expressed Lex and/or Ley, while Lea and Leb were expressed in 23 (21.3%) and 47 (43.5%) isolates, respectively. Expression of two or more Le antigens (Lex, Ley, Lea, or Leb) was significantly higher in H pylori isolated from ulcer patients than in non-ulcer patients (89.6%v 73.2%, p=0.035). There were no significant differences in the prevalence ofcagA or iceA1 inH pylori isolates from peptic ulcer and non-ulcer patients (86.6% v 90.2% forcagA; 70.1% v68.3% for iceA1), and no association of peptic ulcer with any specific vacAgenotype.

CONCLUSIONS The present study indicates that peptic ulcer disease is associated with increased expression of Lewis antigens but notcagA, iceA, orvacA genotype in H pylori isolates in our population. This suggests thatcagA, iceA, andvacA are not universal virulence markers, and that host-pathogen interactions are important in determining clinical outcome.

  • Lewis blood group antigens
  • cagA; iceA
  • vacA; Helicobacter pylori
  • peptic ulcer

Statistics from

Helicobacter pylori is the major aetiological agent of chronic active gastritis and is generally accepted as having a causative role in the pathogenesis of peptic ulcer (PU) disease. H pylori infection has also been aetiologically linked to the development of gastric carcinoma.1 2 It is estimated that more than 50% of the world's population are infected with H pylori. However, only a minority of H pylori infected subjects develop PU or gastric cancer. The reasons for this are not well understood.

Vacuolating cytotoxin gene (vacA) s1a genotype and the cytotoxin associated gene (cagA) have been demonstrated to be related to the virulence of H pylori infection and the development of peptic ulcer.3 4 However, there are also reports to the contrary.5-7 A novel geneiceA (induced by contact with epithelium gene) has been reported and two allelic variants of the gene (iceA1 and iceA2) described.8 Studies based on Western populations suggested that iceA1 is associated with PU.9 10 Recent studies showed that the lipopolysaccharides (LPS) of most H pyloriisolates express Lewisx (Lex) and/or Ley blood group antigens,11 and these antigens are also expressed on human gastric mucosa.12 It is postulated that this molecular mimicry may play a role in the pathogenesis of H pyloriinfections.13 Peptic ulcer disease has been suggested to be associated with H pylori expression of Lex/Ley, and an association between cagA gene and expression of Lex/Ley has also been reported.14Expression of Le antigens and the prevalence oficeA have not been fully investigated in Asian countries where the prevalence of thecagA gene is high regardless of the presence of the disease.15 H pyloristrains may differ in various geographical regions7and studies in different populations may clarify the importance and universality of putative virulence factors. In the present study expression of Le antigens and the prevalence ofcagA as well asiceA and vacAwere investigated in 108 H pyloriisolates in Singapore.

Materials and methods


H pylori strains were isolated from the gastric biopsies of 108 patients undergoing upper gastrointestinal endoscopy for dyspepsia at the National University Hospital, Singapore. Informed consent was obtained from all patients for gastroscopy and biopsies. All patients included in the study were H pylori positive as assessed using the rapid urease test and culture. The patient population comprised 88 Chinese, 13 Indians, and seven Malays. Of these, 67 were males and 41 females. Mean age was 46 years (16–78). Based on history and endoscopic examination, patients were classified into the following groups: duodenal ulcer (36), gastric ulcer (31), and non-ulcer dyspepsia (41). Non-ulcer dyspepsia (NUD) was defined as patients with neither a history of ulcer disease nor endoscopic evidence of ulcer disease. Upper gastrointestinal endoscopy was performed, and after completion of mucosal examination two biopsies were obtained from the gastric antrum within 2 cm of the pylorus. Our previous study16 on multiple antrum biopsies showed a predominance of a single strain of H pylori. Bacteria isolated from gastric biopsies were grown on blood agar plates (blood agar base 2 supplemented with 5% horse blood) in a humidified incubator supplied with 5% CO2 (Forma Scientific, USA) at 37°C for three days. Isolates were identified asH pylori based on Gram stain morphology and positive tests for urease, oxidase, and catalase.


Expression of various LPS epitopes in the H pylori isolates was measured using an enzyme linked immunosorbent assay (ELISA) as described by Simoons-Smit and colleagues.17 The following murine monoclonal antibodies (Mabs) were used: Mab 54-1F6A, specific for Lex 18; Mab 1E52, specific for Ley 19; Mab 7Le, specific for Lea (Bioprobe, Netherlands); Mab 225Le, specific for Leb (Bioprobe); Mab 3-3A, specific for blood group A antigen17; and Mab 3E7, specific for blood group B antigen (Dako, USA). Optical density (OD) was read at 490 nm. An OD value of 0.2 was chosen as the cut off value because the sum of non-specific background binding values for Mabs and for the conjugate never exceeded an OD of 0.1. Synthetic protein linked Le antigens (that is, Lex, Ley, Leaand Leb (IsoSep, Sweden)) were used as positive controls for the Mabs.


Detection of cagA and iceA

DNA of each H pylori strain was isolated by chloroform-phenol extraction.20 The polymerase chain reaction (PCR) was carried out in an amplification thermal cycler (Perkin-Elmer, USA) programmed according to the protocol of Zheng and colleagues.21 The primers used forcagA, 22 iceA1, andiceA29 and the expected PCR fragment lengths are listed in table 1.

Table 1

Polymerase chain reaction (PCR) for amplification of H pylori cagA, iceA, and vacA genes

Genotyping of vacA alleles

The combination of primers described by Atherton and colleagues23 and Wang and colleagues6 were used to amplify the vacA gene fragments of all H pylori isolates. Typing ofvacA for signal sequence region alleles and middle region alleles was carried out using the primers and methods as described by Atherton and colleagues23 for s1a, s1b, s2, m1, or m2 and Wang and colleagues6 for m1T or m1Tm2. These seven pairs of primers and the expected PCR fragment lengths are listed in table 1.


Frequencies were compared using two tailed Fisher's exact test (SPSS 9.0, Chicago, USA). OD values were expressed as mean (SD), and the distributions of ODs were compared by the Student'st test for comparison of means of independent samples. A p value <0.05 was considered statistically significant.



Of the 108 H pylori isolates, 86 were found to express both Lex and Ley, while 11 and six isolates expressed only Lex and Ley, respectively. Three isolates expressed only Leb while two other isolates were non-typeable. The study revealed that 23 (21.8%) isolates expressed Lea while 47 (43.5%) isolates expressed Leb (table 2). Only one isolate expressed blood group A antigen while none expressed blood group B antigen.

Table 2

Relationship between H pylori expression of Le antigens, cagA, iceA1, vacA, and peptic ulcer


Table 2 shows that of 67 isolates from patients with PU, 64 (95.5%) were positive for Lex and/or Leyantigens compared with 39 (95.1%) of 41 isolates from patients with NUD (p=1.000). Furthermore, the mean OD level of Lex or Ley expression was not significantly different between isolates from patients with or without ulcer disease (1.0150 (0.6515)v 1.0813 (0.6496) for Lex(p=0.608); 1.3549 (1.1200)v 1.3774 (1.1917) for Ley(p=0.890)). Sixty (89.6%) of 67 isolates from PU patients expressed two or more Le antigens (Lex, Ley, Lea, or Leb) compared with 30 (73.2%) of 41 isolates from NUD patients (p=0.035). Furthermore, 32/67 (47.8%) isolates from PU patients expressed three or more Le antigens compared with 11/41 (26.8%) isolates from NUD patients (p=0.043).

Levels of expression of Lea and Leb were not significantly different in the H pyloriisolates from patients with peptic ulcer compared with those without ulcers (26.9% v 12.2% (p=0.091) for Lea and 50.7% v 31.7% (p=0.072) for Leb, respectively). The results also showed that Lea coexpressed with Leb.


The cagA gene was positive in 95 (88%) of 108 H pylori isolates. As shown in table2, the cagA gene was found in 58 (86.6%) of 67 PU isolates compared with 37 (90.2%) of 41 non-ulcer isolates (p=0.763). This shows that there was no significant difference for thecagA gene in H pylori isolates from PU and NUD patients.

Of the 95 cagA positive isolates, 91 (95.8%) expressed Lex and/or Ley antigens compared with 12 (92.3%) of 13 cagAnegative isolates (p=0.480). Furthermore, the mean OD level of Lex or Ley expression was not significantly different between cagA positive andcagA negative isolates (1.0490 (0.6125)v 0.9862 (0.8928) for Lex(p=0.745); 1.2860 (1.0970)v 1.8615 (1.3687) for Ley(p=0.088)).


Of 108 isolates, iceA1 was positive in 75 isolates and iceA2 was detected in 26 isolates. Four isolates were positive for bothiceA1 and iceA2, while three isolates did not yield eithericeA1 or iceA2 fragments. There was no significant difference in the presence oficeA1 in H pyloriisolates from PU and NUD patients (70.1% v68.3%; p=0.833).iceA1 was not associated withcagA status (p=0.531). Similarly, there was no significant difference in the presence oficeA2 in H pyloriisolates from PU and NUD (23.9% v 24.4%; p=1.000).

Of 75 iceA1 positive isolates, 72 (96.0%) expressed Lex and/or Ley antigens compared with 31 (93.9%) of 33 iceA1 negative isolates (p=0.640). The mean OD level of Lex or Leyexpression was not significantly different betweeniceA1 positive andiceA1 negative isolates (1.0106 (0.6242)v 1.0699 (0.7026) for Lex(p=0.663); 1.2570 (1.0831)v 1.4669 (1.2140) for Ley(p=0.373)).


Of 108 isolates, 107 were typed as s1a of thevacA genotype and the other isolate was typed as s2. Four vacA genotypes (s1a/m1T, s1a/m1Tm2, s1a/m2, and s2/m2) of H pyloriisolates were identified (table 3), with the distribution as follows: 39 s1a/m1T, 4 s1a/m1Tm2, 64 s1a/m2, and 1 s2/m2. There was no significant difference between PU and NUD patients for infection by s1a/m1T genotype H pylori isolates (37.3%v 34.1%; p=0.837).

Table 3

Relationship between H pylori vacA genotype and peptic ulcer


In this study we observed the occurrence of Lex, Ley, Lea, Leb, and blood group A antigen in H pylori isolates. Of 108 isolates, 106 (98.1%) were typeable with Mabs specific for Lewis and other blood group antigens. The two isolates which were non-typeable showed O side chain (data not shown), indicating the existence of other serotypes that were not reactive with the Mabs used as described by Simoons-Smit and colleagues.17 One strain ofH pylori expressed blood group A antigen which has so far been reported in one otherHelicobacter species, H mustelae. 24 However, the low prevalence of this antigen in H pylori isolates suggests that it does not have an important role in the pathogenesis of gastric diseases. The chemical structures of Lex, Ley, and Lea of H pylori were elucidated earlier,25 while the chemical structure of Leb has recently been determined.26

Lex and Ley were frequently encountered in our local H pylori isolates. Expression of Lex and Ley was similar to the finding in Canada27 but was higher than the findings in the USA14 and Europe.17 Expression of Lea and Leb in H pylori isolates in our study was much higher than that found in the USA,14 Netherlands,17 and Canada.27 Broadberry and Lin-Chu reported that the Le(a+ b+) phenotype is frequent in Chinese patients but rare or absent in Caucasians.28 The relationship between Le antigen expression by H pylori and host phenotype is not clear.27 29 Our observation of higher expression of Lea and Leb in our population of predominantly Chinese patients lends support to the suggestion by Wirth and colleagues29 that H pylori Le antigen expression is related to the host phenotype. However, Taylor and colleagues did not find such a correlation in their study.27

Wirth and colleagues14 suggested that the risk of peptic ulcer increases with expression of Le antigens in H pylori isolates. They found that expression of Lexor Ley in H pylori isolates was significantly higher in PU than in NUD patients. It is important to note that only three of 96 subjects in their multicentre study were of Chinese origin. In contrast, in our predominantly Chinese patients (88/108 (81.5%)), we found equally high expression of Lexand Ley in H pylori isolates from both PU and NUD patients. Furthermore, the mean OD level of Lex or Ley expression was not significantly different between the two groups.

In this study we showed that the presence of two or more Le antigens (Lex, Ley, Lea, or Leb) was significantly higher in H pyloriisolates from PU than from NUD patients. Expression of Leaand Leb antigen in H pyloriisolates from patients with PU was not significantly different from NUD patients.

As this was an exploratory study on the expression of Le antigens inH pylori strains and indeed the first such study in our population, statistical analyses were performed on multiple parameters. We cannot exclude the possibility that some findings may be due to chance. However, the finding of higher expression of Le antigens in peptic ulcer associatedH pylori strains is novel in our population because it holds true on testing for ⩾2 Le antigens and ⩾3 Le antigens.

Several studies in Caucasian populations suggest an association between infection by cagA positiveH pylori and PU disease.30 31Our study indicates that cagA status is not predictive of gastroduodenal disease in the Singapore population, and adds to the growing evidence32 33 that thecagA gene should not be regarded as a universal virulence marker of peptic ulcer disease.

An association of peptic ulcer and infection byiceA1 positive H pylori isolates was described in two previous studies of Western patients.9 10 In accordance with van Doorn and colleagues,9 we found that the presence oficeA1 was independent of thecagA gene. Similar tocagA status,iceA1 of H pyloriwas not associated with PU risk in our population. This constitutes further evidence that distinct H pylorigenotypes circulate in Western and Asian countries.15

Four vacA genotypes (s1a/m1T, s1a/m1Tm2, s1a/m2, and s2/m2) of H pylori isolates were found in the Singapore population. In the present study almost all of the H pylori isolates were typed as s1a, which is similar to reports from China5 and Taiwan.6 The majority (95.4%) of H pylori isolates in Singapore were typed as s1a/m1T or s1a/m2, suggesting that there is less mosaicism invacA alleles of H pylori in Singapore where the population is of mainly Chinese origin. The s1a/m1 genotype H pylori that was reported to be associated with PU3 was not detected in our population. In contrast with the finding by Wang and colleagues,6 there was no association between infection of s1a/m1T H pylori isolates and peptic ulcer in the present study.

The present study showed that peptic ulcer disease was not associated with cagA status oriceA or vacAgenotypes but there was an association with increased expression of a combination of Le antigens. This suggests that the pathogenesis ofH pylori induced gastric diseases may be due to host-pathogen interactions rather than H pylori itself. Expression of Le antigens may favour development of disease in the host by two mechanisms. Firstly,H pylori strains expressing Le antigens may adapt more readily to a host possessing similar antigens.29 Expression of host related antigens on the surface of bacteria could allow bacteria to elude elimination by the host immune response and thus facilitate persistence of infection in the host.34 Chronicity of infection may lead to the development of disease by increasing inflammation and promoting mucosa atrophy. Secondly, expression of host related antigens onH pylori may increase the pathogenic potential of the bacterium via induction of autoantibodies that cross react with gastric mucosa.13 However, it is noted that expression of Le antigens in H pylori was also high in NUD patients, which indicates that additional unidentified factors also contribute to the pathogenesis of H pylori associated peptic ulcer in our population.


The project was supported by the National University of Singapore (NUS) grant GR6431. The authors thank Dr BJ Appelmelk (Amsterdam Free University, Netherlands) for generously providing all the Mabs except 3E7. We thank A/Prof JL Ding (NUS) for helpful discussion. ZPY is a Research Scholar of NUS.


View Abstract


  • Abbreviations used in this paper:
    cytotoxin associated gene
    induced by contact with epithelium gene
    vacuolating cytotoxin gene
    Lewis blood group antigen
    polymerase chain reaction
    peptic ulcer
    non-ulcer dyspepsia
    optical density

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.

Linked Articles

  • Commentary