Evaluation of the Association of Nine Helicobacter pylori Virulence Factors with Strains Involved in Low-Grade Gastric Mucosa-Associated Lymphoid Tissue Lymphoma

Patients and H. pylori strains.We examined 82 H. pylori strains: 43 were isolated from patients with gastric low-grade MALT lymphoma (27 men and 16 women; mean age, 48.2 ± 13.4 years), and 39 were isolated from age-matched patients with histological gastritis only (29 men and 10 women; mean age, 48.3 ± 14.7 years) suffering from dyspepsia. Patients with gastric MALT lymphoma were enrolled from different areas of France in prospective multicenter studies carried out by the Groupe d'Etude Français des Lymphomes Digestifs of the Fédération Française de Cancérologie Digestive (GELD) (48) and the Groupe d'Etude des Lymphomes de l'Adulte (GELA) (32).

Briefly, once a lesion suspected to be a MALT lymphoma was observed during endoscopy, the patient was invited to come back for a complete exploration, including H. pylori detection within 3 weeks.

Isolation of H. pylori was carried out from two gastric biopsy specimens (one from the lesion and one from macroscopically normal mucosa) by standard culture methods as previously described (31).

All strains were cloned, and genomic DNA was extracted from culture originating from simple colonies by using a commercial kit (Qiagen SA, Courtaboeuf, France).

In addition, the two H. pylori strains for which the genome has been sequenced, 26695 and J99, were used as reference strains (3, 52, 53).

Detection of genes encoding pathogenic factors.The genes explored (Table 1) were detected by PCR, except for the determination of the functional status of the oipA, hopZ, and sabA genes, for which sequencing was also performed. All of the primers used in this study were desalted and provided by Q-BIOgene (Strasbourg, France).

cag PAI status was evaluated by PCR for two loci: cagA (with two sets of primers) (54) and cagE (59). A cagA-positive status was defined when the cagA gene was detected by at least one of the two primer pairs. Specific primers for the cag empty site were also used to confirm the absence of the cag PAI (1, 27). The vacA (s [signal] and m [middle] regions), babA, and iceA alleles were detected by PCR as previously described [5, 6, 13, 22; R. M. Peek, S. A. Thompson, J. C. Atherton, M. J. Blaser, and G. G. Miller, abstract from the Ninth International Workshop on the Gastroduodenal Pathology of Helicobacter pylori 1996, Gut 39(Suppl. 2):A71, 1996]. The hopQ type I and II alleles were determined by allele-specific primers (Table 2). The functional status of the oipA (HP638/jhp581), sabA (HP725/jhp662), and hopZ genes is regulated by a slipped strand repair based on the number of CT dinucleotide repeats in the signal coding region. The signal sequences of the oipA and sabA genes including the CT repeats were amplified by PCR with the primers listed in Table 2 with sense and antisense primers designed on the flanking gene and behind the CT dinucleotide repeats, respectively. The primer sense was also used for sequencing. For the amplification of the hopZ gene, we used the previously described primers (41), and sequencing was done with the R2-hopZ primer (Table 2).

PCR amplifications for cagA, cagE, the cag PAI empty site, vacA, iceA, babA2, hopQ, oipA, and sabA were carried out in a 25-μl volume containing 2.5 μl of 10× PCR buffer (Eurobio, Les Ulis, France), 1.5 mM MgCl₂ (Eurobio), 200 μM (each) deoxynucleoside triphosphates (dNTPs) (Eurobio), 2 U of Taq DNA polymerase (Eurobio), 1 μM (each) primers, and 10 ng of H. pylori DNA. After 4 min of denaturation at 94°C, each reaction mixture was amplified for 35 cycles as follows: 30 s at 94°C, 30 s of annealing at 60 (for cagA, iceA, hopQ, and oipA), 58 (for cagE and cag PAI empty site), 55 (for babA2 and sabA), or 53°C (for vacA); and 30 s at 72°C. After the last cycle, extension was continued for another 7 min (except for vacA [2 min]).

The hopZ gene was amplified in a 25-μl volume containing 2.5 μl of 10× PCR buffer (Invitrogen, Cergy Pontoise, France), 1.5 mM MgCl₂ (Invitrogen), 200 μM (each) dNTPs (Eurobio), 1 U of Platinum Taq polymerase (Invitrogen), 1.25 μM (each) primers, and 10 ng of H. pylori DNA. After 4 min of denaturation at 94°C, each reaction mixture was amplified for 40 cycles as follows: 30 s at 94°C then 30 s of annealing at 58°C and 2 min 30 s at 72°C. After the last cycle, extension was continued for another 7 min.

All PCR products were analyzed on a 1.5% agarose gel (except for iceA [3% agarose]) stained with ethidium bromide.

To determine the oipA, hopZ, and sabA status, amplified DNA fragments were purified with Microspin S-400 HR columns (Amersham Pharmacia Biotech, Inc., Uppsala, Sweden), and direct sequencing was performed with the ABI PRISM BigDye Terminators v3.0 cycle sequencing kit (PE Applied Biosystems, Foster City, Calif.) using an ABI 3700 analyzer DNA sequencer (PE Applied Biosystems).

Statistical analysis.Statistical analysis was performed with STATA 7.0 statistical software (Stata Corporation, Texas). A first comparison between MALT and gastritis strains was performed by univariate analysis and Fisher's exact test. Spearman rank coefficients (r) were also determined to study the association between the different characteristics of the strains.

In order to create a variable describing how the different strain characteristics were associated, a multiple correspondence analysis (MCA) was performed (47). To confirm the results, a cluster analysis was used to draw a dendrogram based on the same variables, by using the average aggregation method and Euclidian distance. Both analyses were performed with the Statbox 2.5 program (Grimmer Logiciels, Paris, France). These analyses were performed for all of the subjects included in this study for whom information on all genotypes considered was available as well as for the groups of patients with MALT lymphoma and with gastritis, separately.

The odds for association with MALT lymphoma and the 95% confidence interval (CI) were compared within groups of strains and between the different strata of the summary variables obtained by the MCA.

Relationship between each of the nine H. pylori virulence factors tested in strains of patients suffering from gastric MALT lymphoma and gastritis (Table 1). (i) cag PAI.Twenty-two MALT strains (51.2%) and 20 gastritis strains (51.3%) were cagA positive. Twenty-one MALT strains (48.8%) and 22 gastritis strains (56.4%) were cagE positive. Among the 43 MALT strains, 41 (95.3%) were either cagA positive, cagE positive, and cag PAI empty site-negative or cagA negative, cagE negative, and empty site positive, suggesting that the cag PAI is usually either complete or totally deleted. However, one MALT strain was cagA negative, cagE negative, and cag PAI empty site negative, and another one was cagA positive, cagE negative, and cag PAI empty site negative. Among the 39 gastritis strains, 34 (87.2%) were either cagA or cagE positive and cag empty site negative or cagA or cagE negative and empty site positive (data not shown). Two of these strains were cagA or cagE negative, but the cag PAI empty site PCR revealed an amplification fragment of 1,630 and 2,036 bp, respectively, suggesting that in these two strains, the cag PAI had partial deletions that allowed PCR amplification under our PCR conditions. Three other gastritis strains were cagA negative, cagE positive, and cag PAI empty site negative. At least one cagA or cagE-negative strain and one cagA-positive, cagE-negative strain were empty site negative. The distribution of cagA and cagE genes was not significantly different in H. pylori isolates from MALT patients compared to isolates from gastritis patients (P = 1.0 and 0.515, respectively).

(ii) vacA alleles.All 82 strains were classified according to the vacA gene signal region (either s1 or s2) or middle region (m1 or m2). The distributions of the different alleles s1m1, s1m2, and s2m2 among the MALT strains were 13 (30.2%), 12 (27.9%), and 18 (41.9%), respectively, and those for the gastritis strains were 22 (56.4%), 6 (15.4%), and 11 (28.2%), respectively. The type s2m1 was never detected in either group of strains. The distribution of vacA genotypes was not significantly different (P = 0.064). However, when considering the m genotype only, a slight association was found between m2 genotype and MALT strains (P = 0.025).

(iii) iceA status.The primers used allowed us to determine iceA genotypes in all strains. Twenty-four MALT strains (55.8%) were iceA1, and 19 (44.2%) were iceA2 versus 15 (38.5%) iceA1 and 24 (61.5%) iceA2 among the gastritis strains. With iceA2-specific primers, PCR products for 10 gastritis and 6 MALT strains were longer than the expected size (344 versus 230 bp), suggesting the presence of iceA2 variants in these strains. The prevalence of the two genotypes was not different in the two groups (P = 0.128).

(iv) babA2 status.Nineteen MALT strains (44.2%) and 21 gastritis strains (53.9%) were babA2 positive. The prevalence of babA2 was not significantly different between the two groups (P = 0.507).

(v) hopQ status.The primers used for this new PCR allowed us to determine hopQ alleles in all strains. Type I hopQ alleles were detected in 21 MALT strains (48.8%) and 24 gastritis strains (61.5%), and type II hopQ alleles were detected in 22 MALT strains (51.2%) and 15 gastritis strains (38.5%). The distribution of type I and II hopQ alleles was not different in both groups of strains (P = 0.274).

(vi) oipA functional status.The primers used for this new PCR amplified the oipA gene in all strains. Based on the CT dinucleotide repeats present in the signal sequence coding region of the oipA gene, 20 MALT strains (46.5%) and 21 gastritis strains (53.9%) were considered to be oipA “on,” and 23 (53.5%) and 18 (46.2%) strains, respectively, were considered oipA “off.” The number of CT dinucleotide repeats in MALT strains ranged from 5 to 11, and the number in gastritis strains ranged from 5 to 12, in agreement with Yamaoka et al. (58). CT repeats began at position +23 in 72.1% of the MALT strains and 74.4% of the gastritis strains. There was no difference in the distributions of oipA functional status between these two groups of strains (P = 0.659).

(vii) sabA functional status.The primers used for this new PCR failed to obtain sabA gene amplification for two MALT strains and four gastritis strains. For the 76 remaining strains, 25 (61.0%) MALT strains and 15 (42.9%) gastritis strains contained a functional sabA gene. The number of CT dinucleotide repeats, which began at position +18, ranged from 3 to 12 for both MALT and gastritis strains (Table 3). When three CT repeats were observed, a thymidine-thymidine was always present in the middle of the repetition in frame with the rest of the gene. Interestingly, for strains harboring seven CT repeats, the sabA gene can be in or out of frame depending on the nucleotide sequence after the CT repeats (Table 3). There was no difference in sabA functional statuses between the two groups of strains (P = 0.167).

(viii) hopZ functional status.We failed to obtain PCR hopZ gene amplification for two MALT strains and three gastritis strains. In addition, for one gastritis strain, an 800-bp PCR product, was obtained (versus products of approximately 3,000 bp according to the 26695 and J99 reference strains), and direct sequencing revealed the lack of the hopZ gene. For the remaining strains, 18 (43.9%) MALT strains and 21 (60.0%) gastritis strains contained a functional hopZ gene. Analysis of the gene sequences revealed that CT repeats began at position +18, and the number of repeats ranged from 6 to 12 as previously described (41). Interestingly, two MALT strains contained four dinucleotide repeats in frame with the residual gene (GenBank accession no. AY299993 and AY299994). Moreover, 11 (26.8%) MALT strains and 4 (11.4%) gastritis strains contained eight dinucleotide repeats out of frame (GenBank accession no. AY299995 to AY300009). There was no difference in the distributions of hopZ functional status between the two groups of strains (P = 0.176).

Relationship between the nine H. pylori virulence factors tested in strains of patients suffering from gastric MALT lymphoma and gastritis.In MALT strains, we found a positive correlation between cagA and cagE (r = 0.9545, P < 10⁻⁴) and between vacA s1m1 genotype (r = 0.6431, P < 10⁻⁴) and babA2 (r = 0.4009, P = 0.008), and oipA on (r = 0.7245, P < 10⁻⁴) and hopQ type I (r = 0.7684, P < 10⁻⁴), but there was a negative correlation between cagA and vacA s2m2 (r = 0.7742, P < 10⁻⁴). The babA2 gene was correlated with cagE (r = 0.4423, P = 0.003) and hopQ type I (r = 0.3486, P = 0.0022). Among the gastritis strains, cagA was associated with the same markers as in MALT strains, with the exception of babA2, which was correlated with vacA s1m1 alleles (r = 0.5346, P = 0.0005) and with hopQ type I (r = 0.4310, P = 0.0062).

Results of the MCA for construction of the summary variable.First, the MCA was performed on 70 strains for which complete data were available for all variables, yielding the F1/F2 plane, which accounted for 82.4% of the total information (73.9% for the horizontal axis [F1] and 8.6% for the vertical axis [F2]). Two main clusters were identified on each side of the horizontal axis: the first included strains positive for cagA, cagE, vacA s1m1 type, babA2, oipA on, and hopQ type I (group A), and the second included strains negative for cagA, cagE, vacA s2m2 type, babA1, oipA off, and hopQ type II (group B). On the vertical axis, iceA1 and sabA on strains were on one side, and iceA2 and sabA off strains were on the other. The hopZ strains were found on the third axis (F3) (data not shown).

The dendrogram drawn from the results of the cluster analysis of both populations considered together confirmed the presence of four groups: the A and B groups, which were formally identified; and two others. Group C contained iceA2, sabA off, and hopZ on strains, and group D contained iceA1 and sabA on strains (Fig. 1).

• Open in new tab
• Download powerpoint

FIG. 1.

Dendrogram of the distribution of 70 H. pylori strains from both strain populations for which complete data were available for all variables. Strains have been grouped according to cagA, cagE, vacA, babA, iceA, hopQ, oipA, hopZ, and sabA status.

The same approach was then performed for each population: i.e., MALT patients (39 strains) or gastritis patients (31 strains) from whom complete data were available for each variable. The analysis of the MCAs revealed that groups A and B were similar and that the main difference between MALT and gastritis strains was based on the distribution of iceA genotype, sabA status, and hopZ status (Fig. 2A and B). In MALT strains, a cluster absent in gastritis strains and containing iceA1, sabA on, and hopZ off was identified (Fig. 2B).

• Open in new tab
• Download powerpoint

FIG. 2.

MCA plot of 39 and 31 H. pylori strains isolated from patients suffering from gastric MALT lymphoma (A) and from gastritis (B), respectively, for which complete data were available according to cagA, cagE, vacA, babA, iceA, hopQ, oipA, hopZ, and sabA status.

Relationship between the genotype of H. pylori strains and MALT diagnosis.The possible association between the presence of MALT and the characteristics of strains cultured from these patients was tested. As shown in Table 4, the distribution of the dendrogram deduced A and B groups (based on cagA, cagE, and vacA alleles; babA and oipA status; and hopQ genotype) revealed no difference between MALT and gastritis strains (P = 0.561) (Table 4 and Fig. 3). The proportion of MALT patients was not different in those harboring strains with the sabA gene on, with iceA1, or with the hopZ gene on (62.5, 61.5, and 75.7%, respectively). The odds ratios (ORs) were in line with these results. In contrast, when analyzing the summary variables according to the MCA results, it was remarkable that the odds of having MALT lymphoma among patients harboring iceA1, sabA on, and hopZ off strains (OR, 10.3; 95% CI, 1.2 to 86.0) were 10 times higher than for the other defined groups; indeed among a total of 11 strains having these three markers, 10 (90.9%) corresponded to MALT strains (P = 0.018) (Table 4 and Fig. 4).

• Open in new tab
• Download powerpoint

FIG. 3.

Distribution of the strains harboring the group A and group B combinations among Helicobacter pylori strains isolated from patients with gastric MALT lymphoma or gastritis only.

• Open in new tab
• Download powerpoint

FIG. 4.

Distribution of strains harboring the triple combination iceA1, sabA on, and hopZ off among H. pylori strains isolated from patients with gastric MALT lymphoma or gastritis only.