Heterogeneity of cag genotypes and clinical outcome of Helicobacter pylori infection

Heterogeneity of cag genotypes and clinical outcome of Helicobacter pylori infection MICHELE SOZZI, MARIA LUISA TOMASINI, CARLA VINDIGNI, STEFANIA ZANUSSI, ROSAMARIA TEDESCHI, GIANCARLO BASAGLIA, NATALE FIGURA, and PAOLO DE PAOLI TRIESTE, AVIANO, AND SIENA, ITALY

Helicobacter pylori infecting strains may include colony subtypes with different cytotoxin-associated gene (cag) genotypes. We sought to determine whether the cag heterogeneity of infecting strains is related to the clinical outcome of infection. Gastric biopsies for culture and histologic study were taken from 19 patients infected with cagA-positive strains (6 with duodenal ulcer, 8 with atrophic gastritis, and 5 with nonatrophic gastritis). For each biopsy, DNA was extracted from 10 single colonies and from a sweep of colonies. Polymerase chain reaction (PCR) for cagA and cagE (both located in the right half of cag) and virB11 (located in the left half of cag) was performed. Random amplified polymorphic DNA PCR (RAPD-PCR) and sequencing of glmM PCR product were performed to verify strain identity of colonies with different cag genotypes. In all patients, PCR from sweeps were positive for cagA, showing that all specimens contained cagA-positive H. pylori subtypes. In 11 patients, PCR products from all colonies were positive for cagA, cagE, and virB11, but in 8 patients, PCR products from varying numbers of colonies were negative for 1 or more cag genes. RAPD-PCR and sequencing of glmM PCR product confirmed the strain identities of colonies with different cag genotypes. We detected cag deletions in 6 of 8, 2 of 5, and 0 of 6 patients with atrophic gastritis, nonatrophic gastritis, and duodenal ulcer, respectively (P ⴝ .02). In conclusion, changes in cag genotype in single colony isolates from subjects infected with cagA-positive H. pylori strains are more common in atrophic than in nonatrophic gastritis or duodenal ulcer. These findings are consistent with host-induced (acid secretion?) adaptive changes in cag genotype during infection. (J Lab Clin Med 2005;146:262–270) Abbreviations: cagA ⫽ cytotoxin-associated gene A; cag island ⫽ cytotoxin-associated gene island; ELISA ⫽ enzyme-linked immunosorbent assay; NSAID ⫽ nonsteroidal antiinflammatory drug; PCR ⫽ polymerase chain reaction; RAPD-PCR ⫽ random amplified polymorphic DNA PCR

H

elicobacter pylori gastritis is associated with the development of peptic ulcer, gastric adenocarcinoma, and gastric mucosabp ⫽ base pair;associated lymphoid-tissue lymphoma.1 One of the factors most closely related to the clinical outcome of H. pylori infection is the presence of the cytotoxin-associated gene (cag) island, an approxi-

mately 40-kb cluster of genes that is a determinant of H. pylori strains with enhanced interaction with gastric tissues.2 The cag island encodes a functional type IV secretion system,3 which permits the CagA protein to be translocated into gastric epithelial cells. After being delivered and tyrosine-phosphorylated, the CagA protein exerts its effects on host-cell phys-

From the Unit of Gastroenterology and Digestive Endoscopy, General Hospital, Trieste; the Unit of Microbiology, Immunology, and Virology, National Cancer Center, Aviano; the Department of Pathology and the Department of Internal Medicine, University of Siena, Siena, Italy. Submitted for publication March 23, 2005; revision submitted June 25, 2005; accepted for publication June 30, 2005.

Reprint requests: Dr. Michele Sozzi, S.C. di Gastroenterologia ed Endoscopia Digestiva, Ospedale di Cattinara, Str. di Fiume, 447, 34100 Trieste, Italy; e-mail: [email protected]. 0022-2143/$ – see front matter © 2005 Mosby, Inc. All rights reserved. doi:10.1016/j.lab.2005.06.010

262

J Lab Clin Med Volume 146, Number 5

Sozzi et al

Table I. Demographic data of the 3 groups of patients

Parameter

Duodenal ulcer (n ⴝ 6)

Nonatrophic gastritis (n ⴝ 5)

Atrophic gastritis (n ⴝ 8)

Age (yr)† Sex (M/F)

50.0 ⫾ 2.91 3:3

56.0 ⫾ 6.88 2:3

55.0 ⫾ 5.16 2:6

P*

0.109

*␹2 test for trend. † Data expressed as mean ⫾ SEM.

iology.4 The presence and integrity of the cag island appear to be critical factors in the interaction between H. pylori and its host. We have previously reported deletions of cag genes during murine infection.5 Deletions of the cag island have also been reported in human beings,6 –9 but studies aimed at assessing any relationship between cag deletions and clinical outcome of H. pylori infection have yielded conflicting results.7,10 –13 Because cagA-positive and cagA-negative H. pylori strains may coexist in a single patient and even in a single biopsy specimen14,15 and single-colony isolates that are variably cagA-positive or cagA-negative may have identical fingerprints on RAPD-PCR,15,16 detailed genetic analyses are needed to fully understand the pathogenic role of the cag genes in H. pylori infection. We recently showed that single H. pylori infecting strains may include variable proportions of colony subtypes with different cag genotypes.17 We hypothesize that the cag island is unstable in vivo and that the gastric environment (eg, intragastric pH) selects for different genotypes within the population of infecting cells. In this study we sought to determine whether patients with different clinical outcomes of infection with cagA-positive H. pylori strains harbor bacterial subtypes with different genetic compositions of the cag island. Because the serologic response against the CagA protein has been used for the study of H. pylori–associated disease,18 we also sought to assess the relationship between the serologic presence of anti-CagA antibodies and the integrity of the cag island in H. pylori isolates. METHODS Study population and endoscopy. Nineteen subjects (6 with duodenal ulcer, 5 with nonatrophic gastritis, and 8 with atrophic gastritis) infected with H. pylori strains positive on PCR for cagA were selected from a population of dyspeptic patients who had undergone upper-gastrointestinal endoscopy. Table I provides demographic data on these subjects. All subjects had given informed consent to inclusion in the study, and the protocol was approved by the ethical-review board of the National Cancer Center of Aviano. The research

263

was carried out in accordance with the principles of the Declaration of Helsinki. The criterion for exclusion was previous treatment with H. pylori– eradication therapy, protonpump inhibitors, or NSAIDs. In 14 of the 19 patients, multiple biopsy specimens for culture and histologic study were taken both from the antrum and the corpus; in the remaining 5 patients, samples were taken only from the gastric antrum because the subjects could not tolerate endoscopy. The specimens were obtained with the use of sterilized biopsy forceps that were cleansed with a detergent, disinfected with 70% ethanol, and rinsed with sterile water after each examination. Two distinct sets of forceps were used, 1 for the antrum and 1 for the corpus. Serum specimens were also collected from each patient immediately before endoscopy, then stored at ⫺80°C. Histologic study. All biopsy specimens were fixed in buffered formalin 10%, embedded in paraffin, and stained with hematoxylin and eosin and by means of the modified Giemsa method for H. pylori. The slides were evaluated by 1 pathologist in accordance with the Updated Sydney System classification.19 The pathologist—who was unaware of endoscopic, microbiologic, or serologic data—assigned a score of 0 to 3 (absent, mild, moderate, severe) to each of the following structural variables: inflammation (ie, the amount of mononuclear-cell infiltration), activity (ie, the amount of neutrophilic infiltration), atrophy (ie, the loss of glandular tissue), intestinal metaplasia, and H. pylori density. Patients were classified as having atrophic gastritis when histologic examination of the gastric antrum or corpus yielded an atrophy score of 1 or greater. H pylori culture and identification. Biopsy specimens were cultured in Pylori selective medium (Bio-Mérieux, Rome, Italy). Cultures were incubated at 37°C in a microaerophilic environment (Campygen, Oxoid Ltd, Basingstoke, Hampshire, UK) for 3 or 4 days. The cultured bacteria were identified as H. pylori on the basis of Gram-negative staining, curved or spiral shape, and positivity for catalase, oxidase, and urease production. Identification was further confirmed with the use of PCR. Several sweeps of colonies (which we obtained by harvesting isolates with sweeps in different points of the plate) and 10 single colonies from the primary cultures for each biopsy specimen were subcultured on sheep’s-blood Columbia agar (Kima, Padua, Italy). After bacterial growth had commenced, we obtained a suspension from each subculture and used it for DNA extraction. Individual colony suspensions were considered representative of each H. pylori subtype; sweep suspensions were considered representative of the entire H. pylori population. Genomic DNA extraction. Bacteria from both the 10 single colonies and from the sweeps were resuspended in 1.2 mL of 0.9% NaCl. Bacterial pellets were then obtained by means of centrifugation at 5233g for 5 minutes, after which genomic DNA was extracted with the use of the DNeasy tissue kit (Qiagen, Hilden, Germany). We used spectrophotometry to calculate the amount of DNA. PCR for glmM, cagA, virB11, and cagE. We evaluated the heterogeneity of cag genotypes by assessing the presence of some representative genes located in different segments of the

264

J Lab Clin Med November 2005

Sozzi et al

Table II. Primers used for PCR and RAPD-PCR Gene

Direction

Position*

Sequence (5= ¡ 3=)

Product (bp)

glmM

F R F R F R F R

780–805 1052–1075 1232–1259 1335–1359 914–940 1398–1422 73–95 542–564

GGATAAGCTTTTAGGGGTGTTAGGGG GCTTGCTTTCTAACACTAACGCGC ATAATGCTAAATTAGACAACTTGAGCGA AGAAACAAAAGCAATACGATCATTC TTGAAAACTTCAAGGATAGGATAGAGC GCCTAGCGTAATATCACCATTACCC TTAAATCCTCTAAGGCATGCTAC GATATAAGTCGTTTTACCGCTTC

293

cagA cagE virB11 Random primers for RAPD-PCR 1254 1281 1290

128 508 491

CCGCAGCCAA AACGCGCAAC GTGGATGCGA

F ⫽ forward; R ⫽ reverse. *From the translation initiation codon of the respective genes in H. pylori strain 26695.3

cag island. Among several genes included in the cag island, we considered the following: the cagA gene (located in the right half of the cag island), encoding the highly immunodominant protein CagA, which affects host-cell physiology after being delivered into gastric epithelial cells4,20,21; the cagE (picB) gene (located upstream from cagA in the right half of the cag island), encoding a protein involved in the process of interleukin-8 expression in gastric epithelial cells22–24; and the virB11 gene (located in the left half of the cag island), encoding, together with other genes, a type IV secretion system that permits the delivery of the CagA protein into gastric epithelial cells.25,26 To confirm the identification of the bacteria as H. pylori, we also performed PCR for glmM, a conserved gene formerly known as ureC.27–29 Primers for PCR amplification were designed on the basis of published sequences of glmM, cagA, virB11, and cagE in H. pylori strain 26695 (Table II).5 Each PCR procedure was performed with a volume of 50 ␮L containing 50 mmol/L KCl, 10 mmol/L Tris-HCl (pH 8.3), 1.5 mmol/L MgCl2 (Applied Biosystems, Monza, Italy), 200 ␮mol/L of each deoxynucleotide (Pharmacia Biotech, Milano, Italy), 1.5 U of Taq DNA polymerase (Applied Biosystems), 0.5 ␮mol/L of each primer, and 10 ␮L of DNA at the suitable concentration. Each reaction mixture for glmM was amplified as follows: denaturation at 95°C for 3 minutes, followed by 35 cycles of denaturation at 94°C for 1 minute, annealing at 52°C for 1 minute, and extension at 72°C for 1 minute. For cagA, cagE, and virB11, we used the following incubation conditions: 3 minutes at 95°C and then 50 cycles of 94°C for 1 minute, 48°C (cagA), 53°C (cagE), or 49°C (virB11) for 45 seconds and 72°C for 45 seconds.5 The number of cycles was chosen on the basis of sensitivity tests performed on dilutions ranging from 10⫺1 to 10⫺7 ng of DNA extracted from 5 different colonies. We detected an amplification signal at concentrations of as much as 10⫺5 ng of DNA for the 4 genes, making the sensitivity of the PCRs for cagA, cagE, and virB11 comparable to that of the glmM PCR with the same standardized amount of DNA. Amplifications were performed in duplicate with a DNA thermocycler (GeneAmp PCR System 9600, Perkin-Elmer, Ueberlingen, Germany; Gene Amp PCR System 2400, Perkin-Elmer,

Langen, Germany). PCR products were loaded on 8% polyacrylamide gels, subjected to electrophoresis, stained with ethidium bromide, and visualized by means of photography under ultraviolet transillumination. RAPD-PCR and sequencing of the glmM PCR product.

Colonies from samples from patients 1, 14, and 18 were analyzed with the use of RAPD-PCR to verify strain identity despite the diversity in cag genotypes. Two patients (subjects 14 and 18) were chosen as test patients on the basis of their mixed cag genotypes; patient 1, who had colonies with identical cag genotypes, was chosen as a control. RAPD-PCR was performed with the use of 3 primers (1254, 1281, and 1290; Table I), considered to be highly discriminating.5,15,30 –32 Reaction conditions were as follows: PCR buffer with 2 mmol/L (for primers 1254 and 1281) or 3 mmol/L (primer 1290) MgCl2 (all reagents were supplied by Applied Biosystems), 200 ␮mol/L each deoxynucleotide (Pharmacia), 1.5 U of Taq DNA polymerase (Applied Biosystems), 0.5 ␮mol/L of each primer, and 30 ng of DNA. The reaction volume was 50 ␮L, and the following cycle conditions were used: 4 cycles of 94°C for 5 minutes, 36°C for 5 minutes, and 72°C for 5 minutes, followed by 30 cycles of 94°C for 1 minute, 36°C for 1 minute, and 72°C for 2 minutes. Amplifications were performed with a model 9600 thermal cycler (Appied Biosystems). RAPD-PCR products were analyzed with the use of electrophoresis on 2% agarose gels and ethidium bromide staining. DNA molecular-weight marker VIII (Boehringer Mannheim, Mannheim, Germany) and Gene Ruler 100-bp DNA Ladder Plus (MBI Fermentas, St. Leon-Rot, Germany) were used as standards. To more rigorously determine whether single colonies were identical, we analyzed and compared the nucleotide sequences of an arbitrarily selected gene segment for 10 single colonies from patient 18. On the basis of previous findings,33 we chose the glmM PCR product: A stretch of 210 bp obtained from sequencing of a glmM internal 294-bp fragment was analyzed. After purification with Centricon-100 concentrators columns (Amicon, Beverly, Mass), we carried out direct sequencing of the purified PCR products for both strands by

J Lab Clin Med Volume 146, Number 5

Sozzi et al

265

Table III. PCR results for selected cag genes in single-colony isolates from patients infected by cagA positive H. pylori strains and respective serological responses for anti-CagA antibodies PCR results for the selected cag genes Antrum

Corpus

Diagnosis

cagA

cagE

virB11

No. tested single-colony

1 2 3 4 5 6 7 8

DU DU DU DU DU DU NAG NAG

⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹

⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫺

⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹

10/10 10/10 10/10 10/10 10/10 10/10 10/10 10/10

9

NAG

⫹ ⫹ ⫹

⫹ ⫺ ⫺

⫹ ⫺ ⫹

5/10 2/10 3/10

10 11 12 13 14

NAG NAG AG AG AG

15 16 17 18

AG AG AG AG

19

AG

⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫺ ⫺ ⫹ ⫹ ⫹ ⫹ ⫹ ⫺ ⫺ ⫺ ⫺ ⫹ ⫺

⫹ ⫹ ⫺ ⫹ ⫺ ⫺ ⫺ ⫺ ⫹ ⫺ ⫺ ⫹ ⫺ ⫹ ⫺ ⫹ ⫺ ⫹ ⫺

⫹ ⫹ ⫺ ⫹ ⫹ ⫺ ⫹ ⫺ ⫹ ⫹ ⫹ ⫹ ⫹ ⫺ ⫹ ⫹ ⫺ ⫹ ⫺

10/10 10/10 10/10 10/10 2/10 3/10 2/10 3/10 10/10 10/10 10/10 1/10 1/10 2/10 3/10 1/10 2/10 2/10 8/10

Patient

No. tested single-colony

Anti-CagA IgG*

⫹ ND ⫹ ⫹ ⫹ ND ⫹ ⫺ ⫹ ⫹ ⫹ ⫺ ⫹ ⫺ ⫹ ND ⫺ ⫹ ⫹ ⫹

10/10

⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫺

⫹ ⫺ ⫺ ND

⫹ ⫹ ⫹ ND

10/10 10/10 10/10

ND

ND

cagA

cagE

virB11

⫹ ND ⫹ ⫹ ⫹ ND ⫹ ⫹ ⫹ ⫺ ⫹ ⫹ ⫹ ⫺ ⫹ ND ⫹ ⫹ ⫹ ⫺

⫹ ND ⫹ ⫹ ⫹ ND ⫹ ⫺ ⫺ ⫺ ⫹ ⫺ ⫺ ⫺ ⫹ ND ⫺ ⫹ ⫺ ⫺

⫹ ⫹ ⫹ ND

ND

10/10 10/10 10/10 10/10 3/10 4/10 3/10 2/10 2/10 3/10 3/10 10/10 10/10 10/10 3/10 7/10

⫺

⫹ ⫹ ⫺ ⫹ ⫺

⫹ ⫺ ⫺ ⫺

⫹

DU ⫽ duodenal ulcer; NAG ⫽ nonatrophic gastritis; AG ⫽ atrophic gastritis; ND ⫽ not determined. *Results from ELISA and Western blotting were pooled because no discrepancies were found between the 2 assays.

means of dideoxy-chain termination with an ABI Prism dye terminator cycle sequencing ready reaction kit (Applied Biosystems) and an ABI Prism 310 automated DNA sequencer (Applied Biosystems). Detection of serum anti-CagA antibodies. We used an ELISA kit (Helori CTX IgG; Eurospital, Trieste, Italy) to detect antiCagA IgG, in accordance with the manufacturer’s instructions. To confirm the CagA results obtained on ELISA, we carried out additional experiments involving a Western-blot assay (Biotest Anti-Hp IgG Blot; Biotest, Dreieich, Germany).18 Statistical analysis. We used the ␹2 test or, when appropriate, the ␹2 test for trend, to assess the association between cag deletions and clinical outcome of infection. We considered P values of .05 or less statistically significant.

RESULTS PCR in H pylori sweeps. In all 19 patients, PCR of swept bacterial suspensions were positive for glmM, confirming the presence of H. pylori DNA, and for cagA, showing that all gastric specimens contained a proportion of cagA-positive H. pylori subtypes. Amplification of cagA, cagE, and virB11 in single col-

We isolated a total of 330 single colonies from all 19 patients. In 11 patients, PCR of all colonies showed full signals for cagA, cagE, and virB11, but in 8 patients PCR was negative for 1 or more of the studied genes in various number of colonies. Details of PCR results are shown in Table III. onies.

266

J Lab Clin Med November 2005

Sozzi et al

Table IV. Classification of the single colonies isolated from patients infected with cagA-positive H. pylori strains into 7 genotypic groups

Table V. Deletions of one or more cag genes from single colony isolates of H. pylori and clinical outcome of infection

Genotype No. of Group colonies* cagA cagE virB11

I II III IV V VI VII

200

⫹

⫹

⫹

66 30 16 15 2 1

⫹ ⫹ ⫺ ⫺ ⫺ ⫺

⫺ ⫺ ⫺ ⫺ ⫹ ⫹

⫹ ⫺ ⫺ ⫹ ⫺ ⫹

Patients with the indicated pattern

1, 2, 3, 4, 5, 6, 7, 9, 10, 11, 13, 15, 18, 19 8, 9, 14, 16, 17, 18 8, 9, 12, 14 9, 14, 18, 19 8, 14, 18 18 18

*The total number of colonies was 330.

According to the genetic composition of the cag island, the H. pylori strains analyzed in our study were divided into 7 groups (Table IV). The most frequent genotypes were I, II, and V (61%, 20%, and 9% of single colonies, respectively). RAPD-PCR and sequencing of the glmM PCR product. We compared RAPD fingerprints using the crite-

ria adopted by other authors.32,34,35 We considered fingerprints that differed by no more than 1 band highly similar; variations in band intensity were not taken into account. According to these criteria, the RAPD patterns resulted to be highly similar in all the colonies from each single patient studied (patients 1, 14, and 18). In patient 18, we confirmed strain identity by sequencing and comparing the nucleotide sequence of the glmM PCR product, which was identical in all 10 single colonies. Anti-CagA serologic study and its association with cag genotypes. ELISA and Western-blot assays for the de-

tection of anti-CagA IgG yielded positive results in 12 patients (63%) and negative results in the remaining 7 (37%) (Table III). No discrepant results were found between the 2 assays. All 7 patients negative for the CagA protein on serologic study harbored H. pylori colonies that were negative on PCR for 1 or more of the studied cag genes. In contrast, only 1 of the 12 patients with serologic findings indicative of CagA was infected with a proportion of colonies with cag deletions (patient 19). Therefore, the presence of a complete cag island in virtually all infecting H. pylori colonies seems to be necessary to elicit a serologic response in the host. As a consequence, serologic study for anti-CagA may underestimate the true prevalence of infection by cagpositive H. pylori strains. Clinical outcomes and cag genotypes. PCR amplifications of 1 or more cag genes were lacking in variable proportions of single colonies in 6 of 8 and 2 of 5

Clinical outcome

No. (%) of patients with cag deletions

Total no. of patients

P*

Duodenal ulcer Nonatrophic gastritis Atrophic gastritis

0 2 (40) 6 (75)

6 5 8

.019

*␹2 test.

patients with atrophic gastritis and nonatrophic gastritis, respectively, but this was found in none of the 6 patients with duodenal ulcers (P ⫽ .02) (Table V). The most frequent cag genotypes among the 3 groups were duodenal ulcer (100% genotype I), nonatrophic gastritis (63% genotype I, 22% genotype II, and 8% genotype III), and atrophic gastritis (31% genotype I, 33% genotype II, and 16% genotype III). The comparison of histologic features between patients with and without deletions of 1 or more cag genes from single-colony isolates of H. pylori revealed higher scores in the former group with regard to atrophy and intestinal metaplasia, statistically significant in the antrum (P ⫽ .02 for both atrophy and intestinal metaplasia) and of borderline statistical significance in the corpus (P ⫽ .06 for both atrophy and intestinal metaplasia). In contrast, we detected no differences for inflammation, activity, and H. pylori density. Among the 8 patients with gastric atrophy, 5 had atrophy in the antrum and 3 demonstrated atrophy in the antrum and in the corpus. Intestinal metaplasia was associated with atrophy in 4 of 8 patients with antral atrophy and in all 3 patients with atrophy in the corpus. All patients with intestinal metaplasia harbored H. pylori strains with cag deletions. In contrast, only 2 of 4 patients without intestinal metaplasia in the antrum and 3 of 5 patients without intestinal metaplasia in the corpus harbored strains with deletions. However, the small number of subjects in this subgroup made it impossible for us to confirm statistical significance for the differences we observed. We also tested the relationship between cag deletions and extension of atrophy (antrum only vs antrum and corpus) in the group of patients with gastric atrophy but found no significant differences. However, all 3 patients with atrophy extended to the corpus had deletions, compared with 3 of 5 patients with atrophy limited to the antrum. DISCUSSION

Studies aimed at assessing the relationship between the integrity of the cag island and the clinical outcome of H. pylori infection are based mostly on the analysis

J Lab Clin Med Volume 146, Number 5

of single- or pooled-colony isolates, which are assumed to be representative of the entire infecting bacterial population in an individual host.7,9 –11,13,22 Nevertheless, it has been shown that cagA-positive and cagAnegative H. pylori strains may coexist in a single patient and even in a single biopsy specimen,14,15,36 and we recently observed that single H. pylori infecting strains may include variable proportions of colony subtypes with different cag genotypes.17 Therefore a detailed genetic analysis of different colonies from each patient may be required to better elucidate the pathogenic role of cagA-positive H. pylori strains. Our analysis of 330 single colonies from 19 patients infected with cagA-positive strains showed that about 40% of patients carried variable proportions of colonies lacking 1 or more of the studied genes (cagA, cagE, and virB11). This rate is higher than those reported in other studies, which were based on the analysis of a single colony or a pool of colonies from each patient.7,10 –13 If we had analyzed sweeps of colonies instead of single colonies, we would have found deletions in a lesser (about 25%) percentage of patients, assuming that even the few colonies without deletions would have yielded positive PCR results from the whole bacterial population. When H. pylori colonies were considered instead of patients, partial and total deletions were present in 114 (34.5%) and in 16 (4.8%) of 330 single colonies, respectively. The gene most frequently lacking, among the 3 we studied, was cagE (38% of single colonies). It has been a matter of controversy whether cagA or other cag genes are the best markers for the presence of an intact cag island.6,7,10,13 In our study, cagE, which was associated with the other 2 studied genes in 99% of single colonies, was a better indicator of cag integrity than was cagA, which was associated with them in 61% of cases. This finding is in accordance with the observation of Ikenoue et al, who suggested that the cagEgene was a more accurate marker of an intact cag island than were the other cag genes.6 The first aim of our study was to assess whether patients with different clinic outcomes of infection with cagA-positive H. pylori strains harbor bacterial populations with different genetic compositions of the cag island. Deletions within the cag island of variable proportions of single-colony H. pylori isolates were significantly more frequent in the subjects with atrophic gastritis than in the other 2 groups. Among the subjects with duodenal ulcers, no deletions were found at all, an observation in accordance with other studies in which H. pylori strains with an intact cag island have been found more frequently in patients with severe gastroduodenal disease than in patients with nonulcer dyspepsia.6,7,11,13,37 In contrast, other studies showed no rela-

Sozzi et al

267

tionship between cag integrity and clinical outcome.9,10,12,22 Here again, we believe that these differences are partially attributable to the heterogeneity in cag genotypes within the infecting H. pylori population. Second, in our study the greatest number of deletions was found in patients with atrophic gastritis, whereas in the aforementioned studies, patients with atrophic gastritis were not considered as a distinct group. Third, in the studies by Hsu et al and Kawamura et al, a proportion of duodenal and gastric ulcers could have been related to the use of NSAID; it is not specified whether NSAID users were excluded from these studies.10,12 The serologic response to the CagA protein has been used to identify subjects carrying cag-positive strains and to investigate the relationship between such strains and disease risk.9,18,38 – 42 In our study, 12 (63%) of 19 patients carrying H. pylori strains positive on PCR for cagA from sweep isolates had detectable antibodies against the CagA protein. Notably, all 7 patients with negative serologic findings for the CagA protein harbored variable proportions of H. pylori colonies that were PCR-negative for 1 or more of the studied cag genes. In contrast, only 1 of the 12 patients with positive serologic findings for CagA was infected with a proportion of cag-negative colonies. These data suggest that the presence of a complete cag island in virtually all infecting H. pylori colonies is necessary to elicit a serologic response in the host. Five of 7 patients with negative serologic findings for CagA didn’t carry any colony with positive PCR results for all of the 3 genes studied. This observation is in accordance with the notion that all these 3 genes are essential for the secretion of the CagA protein.43 Conversely, in the other 2 patients with negative serologic findings for CagA, as many as 30% of single colonies yielded positive PCR results for all 3 of these genes. It could be speculated that these small proportions produced amounts of CagA protein insufficient to trigger a detectable antibody response, but in the single patient with a mixed bacterial population of the 12 patients with positive serologic findings for CagA, only 20% of colonies were positive for all 3 genes. An alternative explanation for these discrepant results is that cag genes other than those we have studied were lacking in the patients with negative serologic findings for CagA. If some of the components of the type IV secretory system are absent, the CagA protein will not be delivered into the host cells and an antibody response will not be elicited. Additional H. pylori–related factors, such as the colonization at an earlier stage of the infection by a strain with a different cag genotype with subsequent acquisition or loss of 1 or more cag genes, or the clonal expansion of H. pylori subtypes with different compositions of the cag island

268

Sozzi et al

in the context of a mixed infection, may account for the contrasting results between cag genotyping of the infecting strains and serologic testing of the response against the CagA protein. On the other hand, hostrelated factors, such as specific genetic traits, play a significant role in controlling the immune response against CagA, as has been shown for other microbial proteins.44,45 Overall, in our study population, 37% of patients carrying cagA-positive H. pylori strains had no detectable antibodies against the CagA protein in their sera. The highest rate was found in the subjects with atrophic gastritis (71%), whereas in the patients with duodenal ulcers, everyone was seropositive. According to these data, epidemiologic studies based on serologic findings would underestimate the true prevalence of cagA-positive strains, particularly in the subset of patients with atrophic gastritis, who are at increased risk of gastric cancer. On the other hand, our data show that the serologic presence of anti-CagA antibodies is a good marker for patients who carry a homogeneous population of H. pylori strains with an intact cag island, such as patients with duodenal ulcers. To understand whether the mixed cag genotypes observed in some of our patients were the result of multistrain H. pylori infection instead of the expression of true genomic modifications within the same strain (eg, deletions, mutations, and recombinations), which are commonly described in these species,7,32,46 – 48 we performed RAPD-PCR of single-colony isolates and sequenced the internal 294-bp fragment of the glmM gene, which has been shown to be a highly polymorphic region present in all H. pylori isolates.33 These techniques showed that each patient carrying mixed colonies was infected with a single H. pylori strain. Other studies had previously revealed that single-colony isolates that are variably cagA-positive or cagAnegative may have identical genomic fingerprints,15,16,36 suggesting that single-strain infections are more common than multiple-strain infections in developed countries, despite the observed diversity in cag genotypes. When we assessed the relationship between the integrity of the cag island and the histologic features of gastritis, we found that only atrophy and intestinal metaplasia were associated with the presence of cag deletions. In contrast, no associations were found for inflammation, activity, and H. pylori density. These findings contrast with the notion that cag-positive H. pylori strains elicit a stronger inflammatory response than do cag-negative strains. Our explanation is that even few single colonies with intact cag islands in a context of a mixed infection can induce higher scores of inflammation than can infection with cag-negative strains. An alternative explanation is that the deletions

J Lab Clin Med November 2005

we found did not affect, in most cases, the ability of the bacteria to elicit a moderate inflammatory response. It has been shown that strains with partial deletions of the cag island are still able to induce a moderate secretion of interleukin-8 in gastric epithelial cells.11 Because patients with atrophic gastritis didn’t have different inflammatory scores than patients without atrophy (data not shown), we hypothesize that the association between cag deletions and atrophy (and intestinal metaplasia) is not mediated by inflammation or that cag heterogeneity is a consequence rather than a cause of atrophy. Both antral and corpus atrophy can lead to reduced acid secretion by means of a reduction in antral G cells and a reduction in corpus parietal cells, respectively. In our study, all patients with atrophy both in the antrum and in the corpus, as opposed to about 50% of patients with atrophy limited to the antrum, harbored H. pylori strains with cag deletions. Despite a lack of statistical significance in the differences we observed, a result of the low number of subjects, this observation is consistent with a hypothetical relationship between cag heterogeneity and physiopathologic consequences of atrophy (ie, reduction in acid secretion), which are more pronounced when atrophy extends to the corpus. We have previously reported a selective deletion of elements on the right side of the cag island occurring among a population of H. pylori cells during murine infection.5 Our hypothesis is that the cag island is unstable in vivo and that passage from one host to another or changes in the microenvironmental conditions during infection in the same host might select for different genotypes within the population of colonizing cells. Among other factors, such as differences in bacterial growth48 and in host immune and inflammatory responses,2 differences in pH susceptibility between cag-positive and cag-negative strains49 could account for the selective advantage of a strain (or subclone) with respect to others. Acid-responsive genes have been identified, and expression of about 7% of the bacterial genome seems to be altered at low pH, suggesting that H. pylori has generated a multitude of acid-adaptive mechanisms.50,51 Infection with CagApositive H. pylori strains seems to be protective against the most severe acid-related complications of gastroesophageal-reflux disease, such as Barrett’s esophagus52 and esophageal adenocarcinoma.53 We therefore hypothesize that the possession of an intact cag island confers a selective advantage to H. pylori strains (or subclones) living in highly acid gastric environments, where they can in turn exert a positive effect, both for themselves and for the host, by modulating gastric acidity. The development in the course of the infection of conditions, such as gastric atrophy, leading to an increase in gastric pH could render the possession of an

J Lab Clin Med Volume 146, Number 5

intact cag island no longer necessary and consequently allow strains with partial or total cag deletions to grow. Our findings seem to support this hypothesis, assuming that patients with duodenal ulcers and atrophic gastritis have, respectively, lower and higher intragastric pH than do patients with nonatrophic gastritis. A recent study54 revealed that in individual hosts, H. pylori subclones with different naturally occurring CagA phosphorylation-site variants differed in their ability to elicit specific host responses, supporting the hypothesis that H. pylori exists in dynamic equilibrium with its host.55 In conclusion, our study seems to show that changes in cag genotype in single-colony isolates from subjects infected with cagA-positive H. pylori strains are related to the clinical outcome of the infection, these changes being more frequent in patients with atrophic gastritis than in those with nonatrophic gastritis and virtually absent in patients with duodenal ulcers. Although the mechanism for this phenomenon is unknown, it is consistent with host-induced (eg, gastric-acid secretion) adaptive changes of cag genotypes during H. pylori infection.

Sozzi et al

11.

12.

13.

14.

15.

16.

17.

REFERENCES

1. Marshall BJ. Helicobacter pylori. Am J Gastroenterol 1994;89(8 suppl):S116 –28. 2. Blaser MJ. The interaction of cag⫹ Helicobacter pylori strains with their hosts. In: Hunt RH, Tytgat GNJ, eds. Helicobacter pylori: basic mechanisms to clinical cure. Dordrecht, The Netherlands: Kluwer Academic Publishers; 1998:45– 8. 3. Tomb JF, White O, Kerlavage AR, Clayton RA, Sutton GG, Fleischmann RD, et al. The complete genome sequence of the gastric pathogen Helicobacter pylori. Nature 1997;388:539 – 47. 4. Odenbreit S, Puls J, Sedlmaier B, Gerland E, Fisher W, Haas R. Translocation of Helicobacter pylori CagA into gastric epithelial cells by type IV secretion. Science 2000;287:1497–500. 5. Sozzi M, Crosatti M, Kim SK, Romero J, Blaser MJ. Heterogeneity of Helicobacter pylori cag genotypes in experimentally infected mice. FEMS Microbiol Lett 2001;203:109 –14. 6. Ikenoue T, Maeda S, Ogura K, Akanuma M, Mitsuno Y, Imai Y, et al. Determination of Helicobacter pylori virulence by simple gene analysis of the cag pathogenicity island. Clin Diagn Lab Immunol 2001;8:181– 6. 7. Jenks PJ, Megraud F, Labigne A. Clinical outcome after infection with Helicobacter pylori does not appear to be reliably predicted by the presence of any of the genes of the cag pathogenicity island. Gut 1998;43:752– 8. 8. Occhialini A, Marais A, Urdaci M, Sierra R, Munoz N, Covacci A, et al. Composition and gene expression of the cag pathogenicity island in Helicobacter pylori strains isolated from gastric carcinoma and gastritis patients in Costa Rica. Infect Immun 2001;69:1902– 8. 9. Peters TM, Owen RJ, Slater E, Varea R, Teare EL, Saverymuttu S. Genetic diversity in the Helicobacter pylori cag pathogenicity island and effect on expression of anti-CagA serum antibody in UK patients with dyspepsia. J Clin Pathol 2001;54:219 –23. 10. Hsu PI, Hwang IR, Cittelly D, Lai KH, El-Zimaity HMT, Gutierrez O, et al. Clinical presentation in relation to diversity within

18.

19. 20.

21.

22.

23.

24.

25.

26.

269

the Helicobacter pylori cag pathogenicity island. Am J Gastroenterol 2002;97:2231– 8. Nilsson C, Sillén A, Eriksson L, Strand ML, Enroth H, Normark S, et al. Correlation between cag pathogenicity island composition and Helicobacter pylori–associated gastroduodenal disease. Infect Immun 2003;71:6573– 81. Kawamura O, Murakami M, Araki O, Yamada T, Tomizawa S, Shimoyama Y, et al. Relationship between gastric disease and deletion of cag pathogenicity island genes of Helicobacter pylori in gastric juice. Dig Dis Sci 2003;48:47–53. Maeda S, Yoshida H, Ikenoue T, Ogura K, Kanai F, Kato N, et al. Structure of cag pathogenicity island in Japanese Helicobacter pylori isolates. Gut 1999;44:336 – 41. Figura N, Vindigni C, Covacci A, Presenti L, Burroni D, Vernillo R, et al. cagA positive and negative Helicobacter pylori strains are simultaneously present in the stomach of most patients with non-ulcer dyspepsia: relevance to histological damage. Gut 1998;42:772– 8. van der Ende A, Rauws EA, Feller M, Mulder CJ, Tytgat GN, Dankert J. Heterogeneous Helicobacter pylori isolates from members of a family with a history of peptic ulcer disease. Gastroenterology 1996;111:638 – 47. Wirth HP, Yang M, Peek RM Jr, Hook-Nikanne J, Fried M, Blaser MJ. Phenotypic diversity in Lewis expression of Helicobacter pylori isolates from the same host. J Lab Clin Med 1999;133:488 –500. Tomasini ML, Zanussi S, Sozzi M, Tedeschi R, Basaglia G, De Paoli P. Heterogeneity of cag genotypes in Helicobacter pylori isolates from human biopsy specimens. J Clin Microbiol 2003; 41:976 – 80. Sozzi M, Valentini M, Figura N, De Paoli P, Tedeschi R, Gloghini A, et al. Atrophic gastritis and intestinal metaplasia in H. pylori infection: the role of the cagA status. Am J Gastroenterol 1998;93:375–9. Dixon MF, Genta RM, Yardley JH, Correa P. Classification and grading of gastritis. Am J Surg Pathol 1996;20:1161– 81. Asahi M, Azuma T, Ito S, Ito Y, Suto H, Nagai Y, et al. Helicobacter pylori CagA protein can be tyrosine phosphorylated in gastric epithelial cells. J Exp Med 2000;191:593– 602. Covacci A, Rappuoli R. Tyrosine-phosphorylated bacterial proteins: Trojan horses for the host cell. J Exp Med 2000;191:587– 92. Audibert C, Burucoa C, Janvier B, Fauchere JL. Implication of the structure of the Helicobacter pylori cag pathogenicity island in induction of interleukin-8 secretion. Infect Immun 2001;69: 1625–9. Censini S, Lange C, Xiang Z, Crabtree JE, Ghiara P, Borodovsky M, et al. cag, a pathogenicity island of Helicobacter pylori, encodes type I–specific and disease-associated virulence factors. Proc Natl Acad Sci U S A 1996;93:14648 –53. Tummuru MK, Sharma SA, Blaser MJ. Helicobacter pylori picB, a homologue of the Bordetella pertussis toxin secretion protein, is required for induction of IL-8 in gastric epithelial cells. Mol Microbiol 1995;18:867–76. Krause S, Barcena M, Pansegrau W, Lurz R, Carazo JM, Lanka E. Sequence-related protein export NTPases encoded by the conjugative transfer region of RP4 and by the cag pathogenicity island of Helicobacter pylori share similar hexameric ring structures. Proc Natl Acad Sci U S A 2000;97:3067–72. Krause S, Pansegrau W, Lurz R, de la Cruz F, Lanka E. Enzymology of type IV macromolecule secretion systems: the conjugative transfer regions of plasmids RP4 and R388 and the cag pathogenicity island of Helicobacter pylori encode structurally

270

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

J Lab Clin Med November 2005

Sozzi et al

and functionally related nucleoside triphosphate hydrolases. J Bacteriol 2000;182:2761–70. Lage AP, Godfroid E, Fauconnier A, Burette A, Butzler JP, Bollen A, et al. Diagnosis of Helicobacter pylori infection by PCR: comparison with other invasive techniques and detection of cagAgene in gastric biopsy specimens. J Clin Microbiol 1995; 33:2752– 6. Lu JJ, Perng CL, Shyu RY, Chen CH, Lou Q, Chong SK, et al. Comparison of five PCR methods for detection of Helicobacter pylori DNA in gastric tissues. J Clin Microbiol 1999;37:772– 4. Marais A, Mendz GL, Hazell SL, Megraud F. Metabolism and genetics of Helicobacter pylori: the genome era. Microbiol Mol Biol Rev 1999;63:642–74. Akopyanz N, Bukanov NO, Westblom TU, Kresovich S, Berg DE. DNA diversity among clinical isolates of Helicobacter pylori detected by PCR-based RAPD fingerprinting. Nucleic Acids Res 1992;20:5137– 42. Burucoa C, Lhomme V, Fauchere JL. Performance criteria of DNA fingerprinting methods for typing of Helicobacter pylori isolates: experimental results and meta-analysis. J Clin Microbiol 1999;37:4071– 80. Taylor NS, Fox JG, Akopyants NS, Berg DE, Thompson N, Shames B, et al. Long-term colonization with single and multiple strains of Helicobacter pylori assessed by DNA fingerprinting. J Clin Microbiol 1995;33:918 –23. Kansau I, Raymond J, Bingen E, Courcoux P, Kalach N, Bergeret M, et al. Genotyping of Helicobacter pylori isolates by sequencing of PCR products and comparison with RAPD technique. Res Microbiol 1996;147:661–9. De la Puente-Redondo VA, del Blanco NG, Gutierrez-Martin CB, Garcia-Pena FJ, and Rodriguez Ferri EF. Comparison of different PCR approaches for typing of Francisella tularensis strains. J Clin Microbiol 2000;38:1016 –22. Prewett EJ, Bickley J, Owen RJ, Pounder RE. DNA patterns of Helicobacter pylori isolated from gastric antrum, body, and duodenum. Gastroenterology 1992;102:829 –33. Enroth H, Nyrén O, Engstrand L. One stomach— one strain: does Helicobacter pylori strain variation influence disease outcome? Dig Dis Sci 1999;44:102–7. Kidd M, Lastovica AJ, Atherton JC, Louw JA. Conservation of the cag pathogenicity island is associated with vacA alleles and gastroduodenal disease in South African Helicobacter pylori isolates. Gut 2001;49:11–7. Covacci A, Censini S, Bugnoli M, Petracca R, Burroni D, Macchia G, et al. Molecular characterization of the 128-kDa immunodominant antigen of Helicobacter pylori associated with cytotoxicity and duodenal ulcer. Proc Natl Acad Sci U S A 1993; 90:5791–5. Cover TL, Glupczynski Y, Lage AP, Burette A, Tummuru MK, Perez-Perez GL, et al. Serologic detection of infection with cagA⫹ Helicobacter pylori strains. J Clin Microbiol 1995;33: 1496 –500. Figueiredo C, Quint W, Nouhan N, van den Munckhof H, Herbrink P, Scherpenisse J, et al. Assessment of Helicobacter pylori

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

vacA and cagA genotypes and host serological response. J Clin Microbiol 2001;39:1339 – 44. Weel JF, van der Hulst RW, Gerrits Y, Roorda P, Feller M, Dankert GN, et al. The interrelationship between cytotoxinassociated gene A, vacuolating cytotoxin, and Helicobacter pylori–related diseases. J Infect Dis 173:1171–5. Ching CK, Wong BC, Kwok E, Ong L, Covacci A, Lam SK. Prevalence of CagA-bearing Helicobacter pylori strains detected by the anti-CagA assay in patients with peptic ulcer disease and in controls. Am J Gastroenterol 1996;91:949 –53. Fischer W, Püls J, Buhrdorf R, Gebert B, Odenbreit S, Haas R. Systematic mutagenesis of the Helicobacter pylori cag pathogenicity island: essential genes for CagA translocation in host cells and induction of interleukin-8. Mol Microbiol 2001;42:1337– 48. De Silvestri A, Pasi A, Martinetti C, Belloni C, Tinelli C, Rondini G, et al. Family study of non-responsiveness to hepatitis B vaccine confirms the importance of HLA class III C4A locus. Genes Immun 2001;2:367–72. Musher DM, Watson DA, Baughn RE. Genetic control of the immunologic response to pneumococcal capsular polysaccharides. Vaccine 2000;19:623–7. Alm RA, Ling LS, Moir DT, King BL, Brown ED, Doig PC, et al. Genomic-sequence comparison of two unrelated isolates of the human gastric pathogen Helicobacter pylori. Nature 1999; 397:176 – 80. Han SR, Zschausch HC, Meyer HG, Schneider T, Loos M, Bhakdi S, et al. Helicobacter pylori: clonal population structure and restricted transmission within families revealed by molecular typing. J Clin Microbiol 2000;38:3646 –51. Kersulyte D, Chalkauskas H, Berg DE. Emergence of recombinant strains of Helicobacter pylori during human infection. Mol Microbiol 1999;31:31– 43. Karita M, Blaser MJ. Acid-tolerance response in Helicobacter pylori and difference between cagA⫹ and cagA⫺ strains. J Infect Dis 1998;178:213–9. Wen Y, Marcus EA, Matrubutham U, Gleeson MA, Scott DR, Sachs G. Acid-adaptive genes of Helicobacter pylori. Infect Immun 2003;71:5921–39. Merrel DS, Goodrich ML, Otto G, Tompkins LS, Falkow S. pH-regulated gene expression of the gastric pathogen Helicobacter pylori. Infect Immun 2003;71:3529 –39. Vaezi MF, Falk GW, Peek RM, Vicari JJ, Goldblum JR, PerezPerez GI, et al. CagA-positive strains of Helicobacter pylori may protect against Barrett’s esophagus. Am J Gastroenterol 2000; 95:2206 –11. Chow WH, Blaser MJ, Blot WJ, Gammon MD, Vaughan TL, Rish HA, et al. An inverse relation between cagA⫹ strains of Helicobacter pyloriinfection and risk of esophageal and gastric cardia adenocarcinoma. Cancer Res 1998;58:588 –90. Aras RA, Lee Y, Kim SK, Israel D, Peek RM Jr, Blaser MJ. Natural variation in populations of persistently colonizing bacteria affect human cell phenotype. J Infect Dis 2003;188:486 –96. Blaser MJ, Atherton JC. Helicobacter pylori persistence: biology and disease: J Clin Invest 2004;113:321–33.