Increased Risk of Severe Gastric Symptoms by Virulence Factors vacAs1c, alpA, babA2, and hopZ in Helicobacter pylori Infection

Abstract

Two virulence factors of Helicobacter pylori, cagA and vacA, have been known to play a role in the development of severe gastric symptoms. However, they are not always associated with peptic ulcer or gastric cancer. To predict the disease outcome more accurately, it is necessary to understand the risk of severe symptoms linked to other virulence factors. Several other virulence factors of H. pylori have also been reported to be associated with disease outcomes, although there are many controversial descriptions. H. pylori isolates from Koreans may be useful in evaluating the relevance of other virulence factors to clinical symptoms of gastric diseases because the majority of Koreans are infected by toxigenic strains of H. pylori bearing cagA and vacA. In this study, a total of 116 H. pylori strains from Korean patients with chronic gastritis, peptic ulcers, and gastric cancers were genotyped. The presence of virulence factors vacAs1c, alpA, babA2, hopZ, and the extremely strong vacuolating toxin was found to contribute significantly to the development of severe gastric symptoms. The genotype combination vacAs1c/alpA/babA2 was the most predictable determinant for the development of severe symptoms, and the presence of babA2 was found to be the most critical factor. This study provides important information on the virulence factors that contribute to the development of severe gastric symptoms and will assist in predicting clinical disease outcomes due to H. pylori infection.

Introduction

Helicobacter pylori is a gram-negative, spiral-shaped capnophilic bacterium [1] which is closely associated with the epithelial surface or the surface mucus of gastric mucosa [2]. H. pylori is the causative agent of chronic gastritis and peptic ulcer, and prolonged carriage is known to be a significant risk factor for the development of gastric cancer [3-7]. Therefore, H. pylori was defined as a Class 1 carcinogen by the World Health Organization (WHO) in 1994, and estimates indicate that more than half of the global population is infected by H. pylori [7]. In 2020, the International Agency for Research on Cancer reported that H. pylori was responsible for the most infectious pathogen-related carcinogenesis in 2018 (https://gco.iarc.fr/causes/infections). In addition, most Koreans are carriers of H. pylori by early childhood [8], which is believed to be a major reason for the high prevalence of gastric disorders including gastric cancer in Korea [9].

The incidence of gastric cancer differs geographically [10], and East Asian countries have the highest incidence rates of this disease worldwide [11]. This regional discrepancy can be partly explained by different prevalence rates of H. pylori infection [11], which in East Asian countries are reported to be higher than those in Western countries [8]. However, the incidence of gastric cancer in countries located in West South Asia is similar to that in Western countries, even though H. pylori infection is more prevalent in West South Asia than in Western countries [11]. These studies suggest that other factors in addition to infection could play a role in the pathogenesis of H. pylori and the development of pestilential gastric symptoms.

H. pylori is a highly heterogeneous bacterium [12], and various virulence factors have been identified in isolates [5, 13]. The cagA and vacA genes have been extensively studied as virulence markers of H. pylori strains, and toxigenic strains of H. pylori carrying cagA and vacA are closely associated with the development of gastric diseases such as peptic ulcer and gastric cancer [2, 10, 11, 14, 15]. More cagA-positive strains of H. pylori have been recovered from East Asian populations than from Western populations [16], and more vacA-positive strains have been recovered from patients with peptic ulcer than from patients without peptic ulcer [17, 18]. These studies on cagA and vacA suggest that these virulence factors play an important role in the development of clinical entities of gastric diseases. However, although infections of H. pylori bearing cagA and vacA cause chronic gastritis in many cases, they are not always associated with peptic ulcer or gastric cancer [3]. Only approximately 10–15% of infected individuals were reported to develop peptic ulcer, gastric carcinoma, and mucosa associated lymphoid tissue (MALT) lymphoma [6, 19]. Similarly, in Korea, even though the majority of people become infected early in life with toxigenic H. pylori strains bearing cagA and vacA [20-22], only a small proportion of infected Koreans complain of severe gastric maladies [21-23]. These studies suggest that additional virulence factors of H. pylori are involved in disease outcome.

In addition to CagA and VacA, other virulence factors, including adherence proteins, endonuclease restriction/modification systems, and vir homologs have been explored to identify their relevance to the development of different outcomes of gastric diseases. AlpA/B [17, 24, 25], BabA2 [26-28], HopZ [29], OipA [30, 31], SabA [3, 5, 32], HrgA [14], HpyIII [14], IceA [33], and vir homologues [22] have been reported to be partly associated with disease outcomes, although there are many controversial descriptions. Therefore, it is necessary to evaluate which of these virulence factors, in addition to CagA and VacA, critically contribute to the development of severe symptoms of gastric diseases.

Most H. pylori isolates from East Asian countries, including Korea, contain both an East Asian repeat pattern in the 3′-region of the cagA gene and the toxigenic vacA genotype, irrespective of the presentation of gastric symptoms [11, 21, 23, 34]. Therefore, H. pylori isolates recovered from the Korean population are of value in evaluating the relevance of other virulence factors in the development of clinical outcomes in addition to cagA and vacA. Here, we assessed Korean isolates of H. pylori by genotyping for different virulence factors including vacuolating toxin, adherence proteins, endonuclease restriction/modification systems, and vir homologues. In addition, we analyzed the relevance of these genes to the clinical symptoms of gastric disorders and investigated the polymorphisms of cagA and vacuolating toxin activities.

Materials and Methods

H. pylori Isolates

H. pylori strains were isolated at Gyeongsang National University Hospital (Korea), Kosin University Gospel Hospital (Korea), and St. Carollo Hospital (Korea) from 1987 to 2005, as previously described [35-37]. Briefly, biopsy specimens were inoculated by smearing onto Mueller-Hinton agar (Merck, USA) plates containing 10%bovine serum (Gibco, USA), vancomycin (Merck, 10 μg/ml), nalidixic acid (Merck, 25 μg/ml), and amphotericin B (Merck, 1 μg/ml). The plates were incubated at 37°C under 10% CO₂ and 100% humidity for 7 days. The bacteria were identified as H. pylori on the basis of their morphology, urease test, and 16S rRNA PCR. To amplify the 16SrRNA gene of H. pylori, primers of PV31 (5'-CGGCCCAGACTCCTACGGG-3') and PV32 (5'-TTACCGCGG CTGCTGGCAC-3') were designed based on the previous publication [38]. The PCR amplification was performed with AccuPower PCR PreMix (Cat. No. K-2016, Bioneer, Korea) containing 1 unit of Top DNA polymerase, 250 μM dNTPs, 10 mM Tris-HCl (pH 9.0), 30 mM KCl, 1.5 mM MgCl2, and 10 pmol of the primers. The PCR condition was as follows: pre-denaturation at 94°C for 5 min; followed by denaturation at 94°C for 30 sec, annealing at 55°C for 30 sec, extension at 72°C for 3 min (35 cycles), and a final extension at 72°C for 7 min. PCR-amplified products (180 bp) were detected by agarose gel electrophoresis to identify H. pylori.

All isolates were deposited at the former H. pylori Korean Type Culture Collection (HpKTCC; College of Medicine, Gyeongsang National University, Korea; 2006–2015) with anonymized clinical information. Among the deposited strains, H. pylori strains recovered from patients with chronic gastritis, peptic ulcer, and gastric cancer were selected for analysis. Frozen stocks within 50 subcultures were revitalized and grown on Brucella agar (Accumedia, USA) plates containing the same composition as Mueller-Hinton agar at 37°C under 10% CO₂ and 100% humidity.

Preparation of Genomic DNA from H. pylori

Genomic DNA was extracted from H. pylori as described previously [39]. Briefly, cultured bacterial colonies were collected and suspended in 1 ml of phosphate-buffered saline (PBS, pH 7.2). Bacterial cells were harvested by centrifugation at 12,700 ×g for 5 min and suspended in 200 μl of Tris-EDTA (TE) buffer (pH 7.0). After adding 600 μg of lysozyme (Genery, China), the cell mixture was incubated at 37°C for 1 h. Then, after adding 20 μl of 10%SDS solution and 10 μg of RNase (RBC, Taiwan), the cell mixture was incubated at 37°C for 1 h, followed by treatment with 120 μg of Protease K (GeNet Bio, Korea) and 172 μg of pronase (Merck) at 37°C for 1 h. After adding 1/10 volume of 5% cetyltrimethylammonium bromide (BDH Chemicals Ltd., England)-0.5 M NaCl solution and incubating at 37°C for 1 h, the mixture was chloroform extracted by adding an equal volume of phenol:chloroform:isoamyl alcohol (25:24:1) solution and vortexing. After centrifuging at 12,700 ×g for 10 min, the aqueous layer was collected, mixed with 1/10 volume of 3 M sodium acetate (pH 5.2), and added to an equal volume of isopropanol. After incubating at −70°C for 20 min, the sample was centrifuged at 12,700 ×g for 10 min. DNA pellets were washed with 1 mL of 70% ethanol, dried completely, dissolved in 30 μl of TE buffer, and frozen at -70°C until required.

PCR

Extracted DNA (50 ng) was PCR amplified with AccuPower PCR PreMix using 10 pmol of forward and reverse primers listed in Table 1. After pre-denaturation at 94°C for 4 min, PCR was carried out at the annealing temperature for 1 min with extension at 72°C for 30 sec, followed by a final extension at 72°C for 7 min. The annealing temperatures and number of cycles are listed in Table 1. Amplified DNA was separated by 1% agarose gel electrophoresis, stained with ethidium bromide, and visualized and analyzed with a Fluor-S MultiImager (Bio-Rad, USA).

Table 1.

Sequences of oligonucleotide primers and PCR conditions for the genotyping of H. pylori isolates.

Genes	primer F (5'⇒3')	Primer R (5‘⇒3’)	Annealing temp	Cycles	Amplified DNA size (bp)	References
cagA5’region	GATAACAGGCAAGCTTTTGATG	CTGCAAAAGATTGTTTGGCAGA	55°C	35	349	[16]
cagA EPIYA FR	ACCCTAGTCGGTAATGGG	GCAATTTTGTTAATCCGGTC	48°C	30	293~299	[72]
cagA EPIYA JSR	GCAATTTTGTTAATCCGGTC	GCTTTAGCTTCTGAYACYGC	48°C	30	222	[72]
vacA s1/s2	ATGGAAATACAACAAACACAC	CTGCTTGAATGCGCCAAAC	52°C	35	259/286	[73]
vacA s1a s1a	TCTYGCTTTAGTAGGAGC	CTGCTTGAATGCGCCAAAC	52°C	35	212	[16]
vacA s1b ss3	AGCGCCATACCGCAAGAG	CTGCTTGAATGCGCCAAAC	55°C	35	187	[73]
vacA s1c s1c	CTYCTTTAGTRGGGYTA	CTGCTTGAATGCGCCAAAC	55°C	35	213	[16]
vacA m1/m2	CAATCTGTCCAATCAAGCGAG	GCGTCTAAATAATTCCAAGG	52°C	35	570/645	[16]
vacA i1	TTAATTTAACGCTGTTTGAAG	GTTGGGATTGGGGGAATGCCG	53°C	35	426	[74]
vacA i2	GATCAACGCTCTGATTTGA	GTTGGGATTGGGGGAATGCCG	53°C	35	432	[74]
alpA	ACGCTTTCTCCCAATACC	AACACATTCCCCGCATTC	65°C	40	304	[75]
alpB	TCGCCGGGACTTTGGGCAAC	TGCGCTAAAGCGGCGTCCAA	55°C	35	505	Designed
oipA	CAAGCGCTTAACAGATAGGC	AAGGCGTTTTCTGCTGAAGC	55°C	35	450	[76]
babA2	AATCCAAAAAGGAGAAAAAACATGAAA	ATAGTTGTCTGAAAGATC	50°C	38	180	[26]
hopZ	GCCTGATATGGGTGGCATGGG	ATTTGATAGCCCGCGCTGAT	50°C	35	493	[77]
sabA	GAGCTATTGACCAGCTCAATG	TAGTTTGGATTCGTTCTCATTA	50°C	35	447	[78]
iceA1	TATTTCTGGAACTTGCGCAACCTGAT	GGCCTACAACCGCATGGATAT	55°C	30	642	[79]
iceA2	CGGCTGTAGGCACTAAAGCTA	TCAATCCTATGTGAAACAATGATCGTT	55°C	30	662	[79]
iceA1Δ94	GGTGAGTCGTTGGGTAAGCGTTACAGAATT	CACAACCATCATATTCAGCCTCCCCCTCATA	55°C	30	520	[79]
hrgA	TCTCGTGAAAGAGAATTTCC	TAAGTGTGGGTATATCAATC	55°C	30	682	[80]
hpyIIIR	CTCATTGCTGTGAGGGAT	TCTTGATAGGATCTTGCG	55°C	30	443	[80]
hpyIIIM	CTCATTGCTGTGAGGGAT	TCTTGATAGGATCTTGCG	55°C	30	562	[80]
dupA jhp0917	TGGTTTCTACTGACAGAGCGC	AACACGCTGACAGGACAATCTCCC	57°C	30	307	[22]
dupA jhp0918	CCTATATCGCTAACGCGCGCTC	AAGCTGAAGCGTTTGTAACG	57°C	30	276	[22]

For nucleotide sequence analysis, PCR was performed using the pfu PCR Kit (Cat. No. EBT-11011, ELPIS, Korea), 4 ng of genomic DNA, 10 pM of primers, and 0.5 unit of pfu DNA polymerase under the conditions described above. Amplified PCR product was purified using an EZ-Pure PCR Purification Kit (Cat. No. EP201-50N, Enzynomics, Korea). Purified PCR products were sequenced using an Applied Biosystems 3730xl DNA Analyzer (Carlsbad, USA).

Vacuolation Activity Test

Vacuolation activity was measured using RK-13 (ATCC CCL-37) cells as described previously with slight modifications [40]. Briefly, RK-13 cells were grown overnight in RPMI 1640 medium (Gibco) supplemented with 5% fetal bovine serum (FBS; Gibco), adjusted to 1.5 × 10⁴ cells/100 μl, added in 96-well microplates at 100 μl/well and incubated for 4 h. H. pylori strains were grown in a thin-layer liquid culture system [1]. Briefly, bacteria were grown overnight on a Petri dish (100 mm diameter) containing 3 ml of Brucella broth (Accumedia) supplemented with 10% bovine serum (Gibco) at 37°C for 24 h in an atmosphere of 10% CO₂ and 100% humidity. Supernatants (50 μl) were harvested, serially (1:2) diluted with the culture medium, added into 96-well plates of RK-13 cells, and incubated at 37°C for 12 h. The degree of vacuolation was observed under a light microscope (Olympus, Japan). Vacuolation activity was defined as the reciprocal of the dilution factor of the culture supernatant in which 10% of cells were vacuolated. Vacuolation activity was defined as the reciprocal of the dilution factor of the culture supernatant in which 10% of cells were vacuolated.

Statistical Analysis

The two-way association between genotypes of virulent factors and clinical entities of gastric disease was examined using the Chi-square test and Fisher's exact test. The distribution of patient ages among clinical entities of gastric disorder was assessed with an independent two-sample t-test. A p-value of < 0.05 was considered statistically significant. Odds ratios (ORs) adjusted for sex were given with 95% confidence intervals to estimate the risk. All data were statistically analyzed using the Statistical Package for Social Sciences 22 (SPSS Inc., USA).

Results

Patients and H. pylori Isolates

A total of 116 H. pylori strains were recovered from patients diagnosed clinically and pathologically with chronic gastritis (38 strains; male/female, 18/20), peptic ulcers (39 strains; male/female, 32/6), and gastric cancers (39 strains; male/female, 25/14) [35-37]. Patients were from the southern area of the Korean peninsula. The average patient age with standard deviation was 48.6 ± 12.1 years for those with chronic gastritis, 48.4 ± 13 years for those with peptic ulcer, and 52.9 ± 11.1 years for those with gastric cancer. Patient ages did not significantly differ from each other, although the average age of patients with gastric cancer was higher than that of others (Fig. 1). Strains that had been subcultured less than 50 times were chosen for this study to minimize the probability of genomic mutation during in vitro maintenance.

Fig. 1.

Age distribution of patients with chronic gastritis, peptic ulcer, or gastric cancer.

Polymorphism of vacA Alleles

vacA allele polymorphisms are determined by variation in the signal sequence region (s1a, s1b, s1c, and s2), mid-region (m1 and m2), and intermediate region (i1, i2, and i3). In this study, the vacA genotype in the signal sequence region was the s1 type for all 116 H. pylori strains, whereas 109 strains (94%) were grouped as the m1 type in the mid-region, and 110 strains (95%) were grouped as the i1 type in the intermediate region (Table 2). No strains exhibited the s2 type. Only 6% and 2.5% of strains were grouped as m2 and i2 types, respectively. The s1 subtypes s1a and s1c were identified in 37 (32%) and 79 (68%) strains, respectively, whereas the s1b subtype was not detected. Eighteen strains associated with chronic gastritis (47%) were grouped into the s1a subtype, which was more than the number of strains associated with gastric cancer (eight strains, 21%) grouped into this subtype. In contrast, 31 (79%) strains associated with gastric cancer were grouped into the s1c subtype, which was significantly (p = 0.017) higher than the number (20 strains, 53%) of strains associated with chronic gastritis grouped into this subtype. No significant differences were found when vacuolating toxin activity was measured, although the average value for the gastric cancer-associated strains was higher than that of those associated with chronic gastritis or peptic ulcer (Fig. 2 and Table 2). There was no significant difference in vacuolating toxin activity between the s1a and s1c subtypes. However, 33% of gastric cancer strains were classified into the extremely strong group, which was significantly (p = 0.026) higher than that of chronic gastritis strains or peptic ulcer strains when toxin activity was classified into the following four groups: extremely strong (> 41, more than the average value plus 1 SD); strong (41–21.8, from less than the extremely strong to the average); weak (21.7–2.6, from the average value to the average minus 1 SD); and extremely weak (< 2.6, less than the value of the average minus 1 SD) (Table 3).

Table 2.

vacA genotypes and vacuolating toxin activity of H. pylori isolates recovered from patients with chronic gastritis, peptic ulcer, or gastric cancer.

vacA genotypes and Vac activity	Total (n=116)	Chronic gastritis (n=38)	Peptic ulcer (n=39)	Gastric cancer (n=39)
s1	116(100%)	38(100%)	39(100%)	39(100%)
s1a	37 (32%)	18 (47%)	11 (28%)	8 (21%)
s1b	0	0	0	0
s1c	79 (68%)	20 (53%)	28 (72%)a	31 (79%) b
s2	0	0	0	0
m1	109 (94%)	36 (95%)	37 (95%)	36 (92%)
m2	7 (6%)	2 (5%)	2 (5%)	3 (8%)
i1	110 (95%)	35 (89%)	38 (97%)	37 (95%)
i2	3 (2.5%)	1 (3%)	0	2 (5%)
i1,i2(-)	3 (2.5%)	2 (5%)	1 (3%)	0
Vacuolating toxin activity	21.8±19.2	20.3±17.2	18.7±18.0	26.3±21.2

^ap = 0.103 and ^bp = 0.017 compared with that in chronic gastritis with Fisher’s exact test.

Fig. 2. Vacuolating toxic effects of H. pylori culture broth on RK-13 cells.

The RK-13 cells showed vacuolation after 12 h of coculture with a two-fold diluted culture medium of each H. pylori strain. Arrows indicate well-characterized vacuolation. A, extremely strong; B, strong; C, moderate; D, weak; E, non-toxic effect control.

Table 3.

Vacuolating toxin activity of H. pylori isolates recovered from patients with chronic gastritis, peptic ulcer, or gastric cancer.

Symptoms	No	Vacuolating toxin activity
		Extremely Strong (>41)	Strong (41~21.8)	Moderate (21.8~2.6)	Weak (<2.6)
Chronic gastritis	38	5 (13%)	8 (21%)	20 (53%)	5 (13%)
Peptic ulcer	39	4 (10%)	7 (18%)	19 (49%)	9 (23%)
Gastric cancer	39	13 (33%)a	6 (15%)	17 (44%)	3 (8%)
Total	116	22 (19%)	21 (18%)	56 (48%)	17 (15%)

Parentheses indicate percentage of each group classified by vacuolating cytotoxin activity.

^ap = 0.026 compared with groups of chronic gastritis or peptic ulcer using Fisher’s exact test.

Genotypes of cagA

The cagA gene was found in almost all strains, with the exception of two strains isolated from patients with chronic gastritis. Polymorphisms in the 3'-region of cagA were analyzed by PCR and nucleotide sequencing. Five kinds of repeat patterns of EPIYA were identified in cagA-positive H. pylori isolates, and most H. pylori isolates grouped into the EPIYA-ABD pattern, which has been frequently isolated in Asia. Four patterns of variant EPIYA-ABDs were found in six strains by PCR and nucleotide sequencing (Table 4). There was no EPIYA motif in the EPIYA-B segment or deletion of amino acid residues in the downstream region of EPIYA-A and/or -D segments. Multiple EPIYA-AB motifs were found in one chronic gastritis strain and one gastric cancer strain. Multiple EPIYA-D segments (EPIYA-ABD′bD and A′bD′bD) were present in one peptic ulcer strain and in three gastric cancer strains.

Table 4.

Repeat patterns of EPIYA in the 3'-region of cagA of H. pylori isolates recovered from patients with chronic gastritis, peptic ulcer, or gastric cancer.

CagA EPIYA types a	Chronic gastritis (n=38)	Peptic ulcer (n=39)	Gastric cancer (n=39)
ABD or A′bD	35 (92.0%)	38 (97.4%)	35 (89.7%)
ABD′bD or A′bD′bD	0	1 (2.6%)	3 (7.7%)
ABABD	1 (2.7%)	0	1 (2.6%)
CagA negative	2 (5.3%)	0	0

^a, A, B, and D indicate typical EPIYA segment, b notes the EPIYA-B segment deleted in the 5’-end region including EPIYA motif, and A′ and D′ indicate the segment deleted in the downstream region of EPIYA motif. ABD segments produced DNA fragments of 293~299 bp with cagA EPIYA FR primers and 222 bp with EPIYA JSR primers. When amplified DNAs of different sizes were produced, amplified DNAs were subjected to nucleotide sequence analysis.

Genotypes of Adhesion Genes

The genotypes of the adhesion genes alpA, alpB, babA1, babA2, hopZ, oipA, and sabA were examined by PCR. Among the 116 strains, alpA, alpB, babA2, hopZ, oipA, and sabA were detected in 88 (76%), 113 (97%), 91 (79%), 96 (83%), 112 (97%), and 70 (60%) strains, respectively (Table 5). The prevalence of alpA in peptic ulcer strains (82%) and gastric cancer strains (87%) was significantly (p = 0.026 and 0.005, respectively) higher than that in chronic gastritis strains (58%), and the prevalence of babA2 in peptic ulcer (95%) or gastric cancer strains (95%) was also significantly (p = 0.000) higher than that in chronic gastritis strains (45%). The prevalence of hopZ was significantly (p = 0.036) higher in the gastric cancer strain (92%) than that in the chronic gastritis strains (73%). The prevalence of alpB, oipA, and sabA did not differ significantly among the three clinical entities of gastric diseases (Table 5).

Table 5.

Genotypes of adhesion proteins of H. pylori isolates recovered from patients with chronic gastritis, peptic ulcer, or gastric cancer.

Genotypes	Total (n=116)	Chronic gastritis (n=38)	Peptic ulcer (n=39)	Gastric cancer (n=39)
alpA	88 (76%)	22 (58%)	32 (82%) a	34 (87%) b
alpB	113 (97%)	35 (92%)	39 (100%)	39 (100%)
babA2	91 (79%)	17 (45%)	37 (95%) c	37 (95%) d
hopZ	96 (83%)	28 (73%)	32 (82%)	36 (92%)e
oipA	112 (97%)	35 (92%)	38 (97%)	39 (100%)
sabA	70 (60%)	25 (66%)	23 (59%)	22 (56%)

^ap = 0.026, ^bp = 0.005, ^cp = 0.000, ^dp = 0.000, and ^ep = 0.036 compared with those with chronic gastritis using Fisher’s exact test.

Genotypes of Genes in the Endonuclease Restriction/Modification Systems

Genes in the endonuclease restriction/modification systems were used for genotyping H. pylori isolates, including hrgA, hypIII, and iceA. Among the 116 strains, hrgA, hypIIIR, and hypIIIM were detected in 49 (35%), 79 (69%), and 116 (100%) strains, respectively. The iceA1 and iceA2 genotypes were found in 98 (84%) and 15 (13%) strains, respectively. The prevalence of hrgA in chronic gastritis (44%) strains was higher than that in peptic ulcer (21%) or gastric cancer (33%), although there was no significant difference among the three groups. The prevalence of hypIII and iceA was not significantly associated with clinical outcomes (Table 6).

Table 6.

Genotypes of endonuclease enzyme genes of H. pylori isolates recovered from patients with chronic gastritis, peptic ulcer, or gastric cancer.

Genes	Total (n = 116)	Chronic gastritis (n = 38)	Peptic ulcer (n = 39)	Gastric cancer (n = 39)
hrgA	36 (35%)	16 (44%)	8 (21%)	12 (30%)
hypⅢR	80 (70%)	22 (56%)	30 (77%)	27 (70%)
hypⅢM	116 (100%)	38 (100%)	39 (100%)	39 (100%)
iceA1	98 (84%)	30 (78%)	34 (87%)	33 (85%)
iceA1Δ94	2 (2%)	1 (3%)	0	1 (3%)
iceA2	15 (13%)	6 (16%)	4 (10%)	5 (13%)

Genotypes of dupA Genes in CagPAI

Genotyping of dupA genes was performed by PCR-based identification of jhp0917 and jhp0918; dupA was detected in 28 (24%) strains while jhp0918 was found in 2 (2%) strains. The prevalence of dupA in peptic ulcer strains (26%) and gastric cancer strains (28%) was higher than that in chronic gastritis strains (18%), although this lacked statistical significance (Table 7).

Table 7.

Genotypes of dupA in CagPAI of H. pylori isolates recovered from patients with chronic gastritis, peptic ulcer, or gastric cancer.

Genes	Total (n = 116)	Chronic gastritis (n = 38)	Peptic ulcer (n = 39)	Gastric cancer (n = 39)
jhp0918	2 (2%)	2 (5%)	0	0
jhp0917–0918	28 (25%)	7 (18%)	10 (26%)	11 (28%)

Contribution of Genotyping Factors and Vacuolating Toxicity to the Development of Severe Gastric Symptoms

Among the bacterial factors analyzed in this study, the prevalence of genotypes vacAs1c, alpA, babA2, and hopZ and the extremely strong vacuolating toxicity showed significant differences between strains associated with chronic gastritis strains and those associated with severe symptoms (peptic ulcer and gastric cancer). The presence of the babA2 genotype contributed the most to the development of peptic ulcer (OR: 27.5; 95% CI: 5.0–151.2; p = 0.000) and gastric cancer (OR: 21.7; 95% CI: 4.4–106.3; p = 0.000) (Table 8). The presence of alpA also contributed to the development of peptic ulcer (OR: 7.0; 95% CI: 2.0–24.9; p = 0.026) and gastric cancer (OR: 4.5; 95% CI: 1.2–17.6; p = 0.005). Genotypes of vacAs1c and hopZ also showed significant associations with the development of gastric cancer. Extremely strong vacuolating toxicity was significantly associated with the development of gastric cancer compared with that of chronic gastritis (OR: 3.5; 95% CI: 1.1–11.2; p = 0.009)(Table 8).

Table 8.

Analysis of the contributory role of genotyping factors and vacuolating toxicity in the development of gastric cancer and peptic ulcer.

	vacAs1c		alpA		babA2		hopZ		Extremely strong vacuolating toxin
	Peptic ulcer	Gastric cancer	Peptic ulcer	Gastric cancer	Peptic ulcer	Gastric cancer	Peptic ulcer	Gastric cancer	Peptic ulcer	Gastric cancer
OR	1.7	3.2	7.0	4.5	27.5	21.7	2.4	3.4	1.4	3.5
95% CI	0.60–5.2	1.1–9.8	2.0–24.9	1.2–17.6	5.0–151.2	4.4–106.3	0.7–8.8	0.6–18.2	0.3–6.0	1.1–11.2
p value	0.103	0.017	0.026	0.005	0.000	0.000	0.421	0.036	0.702	0.009

Combination of Genotypes Predicting the Development of Peptic Ulcer and Gastric Cancer

A total of 11 combinations generated by double, triple, and quadruple genotypes of vacAs1c, alpA, babA2, and hopZ were significantly associated with the development of severe symptoms (Table 9). Among these combinations, babA2/hopZ was present at the highest proportion (82.1%) in severe symptom strains, followed by alpA/babA2 (79.5%), alpA/hopZ (75.6%), vacAs1c/babA2 (71.8%), and alpA/babA2/hopZ (70.5%). The combination of vacAs1c/alpA/babA2 was the most predictable determinant for the development of severe symptoms (OR: 15.8; 95% CI: 4.6–53.8; p = 0.000), followed by vacAs1c/alpA/babA2/hopZ (OR: 11.6; 95% CI: 3.4–39.6; p = 0.000), alpA/babA2 (OR: 10.9; 95% CI: 3.9–30.5; p = 0.000), vacAs1c/babA2 (OR: 8.4; 95% CI: 3.0–23.6; p = 0.000), and vacAs1c/babA2/hopZ (OR: 8.6; 95% CI: 3.0–25.0; p = 0.000).

Table 9.

The multiple correspondence analysis of the genotypes vacAs1c, alpA, babA2, and hopZ with gastric symptoms.

Combinations of genotypes
Genotypes	Double						Triple				Quadruple
vacAs1c	+	+	+				+	+	+		+
alpA	+			+	+		+		+	+	+
babA2		+		+		+	+	+		+	+
hopZ			+		+	+		+	+	+	+
No. (%) in 38 chronic gastritis strains	10 (26.3)	8 (21.1)	15 (39.5)	11 (29.0)	18 (47.4)	13 (34.2)	4 (10.5)	6 (15.8)	8 (21.1)	10 (26.3)	4 (10.5)
No. (%) in of 78 severe symptoms strains	52 (66.7)	56 (71.8)	52 (66.7)	62 (79.5)	59 (75.6)	64 (82.1)	49 (62.8)	49 (62.8)	46 (59.0)	55 (70.5)	43 (55.1)
Odds ratio	7.0	8.4	3.3	10.9	4.0	8.9	15.8	8.6	6.5	7.4	11.6
95% CI	2.7–18.01	3.0–23.6	1.4–8.0	3.9–30.5	1.5–10.6	3.4–23.2	4.6–53.8	3.0–25.0	2.4–17.1	2.8–19.3	3.4–39.6
p value	0.000	0.000	0.008	0.000	0.002	0.000	0.000	0.000	0.000	0.000	0.000

Among the four genotype markers, the OR average of double and triple combination groups containing babA2 showed the highest value when compared with combinations containing other genotype markers, demonstrating that babA2 might be the most critical determinant in the development of severe symptoms of gastric diseases (Tables 8 and 9).

Discussion

Toxigenic strains of H. pylori carrying cagA and vacA are known to be closely associated with the development of gastric diseases such as peptic ulcer and gastric cancer [2, 10, 11, 14, 15]. Notably, although infections of H. pylori bearing cagA and vacA cause chronic gastritis in many cases, they are not always associated with peptic ulcer or gastric cancer [3]. Only approximately 10–15% of infected individuals were reported to develop peptic ulcer, gastric carcinoma, and MALT lymphoma [6, 19]. In Korea, most people are infected early in life with toxigenic H. pylori strains bearing cagA and vacA [20-22], but only a small proportion of these individuals have complained of severe gastric maladies [21-23]. These studies suggest that, along with cagA and vacA, other virulence factors of H. pylori are involved in disease outcomes.

A variety of candidate virulence factors have been identified for their role in provoking gastric pathogenesis. Unfortunately, their relevance to severe gastric disorders has proved epidemiologically controversial [3]. The reported discrepancies in the relationship between virulence factors and disease outcomes may be due to differences in prevalence rates and the types of cagA and vacA virulence factors present in different regions [3,5,26-28,41,42]. In contrast, most Korean H. pylori isolates contain cagA and vacA, thereby simplifying the analysis of the influence of other virulence factors and the understanding of their roles in the development of severe gastric symptoms. The risk prediction of severe gastric symptoms according to virulence factors is critical in the management and treatment of H. pylori infections. Therefore, in this study we investigated the relevance of virulence factors to clinical symptoms and sought to establish risk rates for the development of gastric diseases caused by H. pylori infections based on previously identified virulence factors.

Difference in vacuolating ability of H. pylori is conferred by vacA type variations in the signal sequence (s1a, s1b, s1c, and s2), mid-region (m1 and m2), and intermediate region (i1, i2, and i3). Recently, a deletion-region (d1 and d2) and c-region (c1 and c2) have also been identified [3, 43]. Genotypes and subtypes of vacA in H. pylori strains have been reported to be geographically distributed. H. pylori strains in Northeast Asia have been predominantly grouped into the s1, m1, and i1 types based on the vacA signal sequence, middle, and intermediate regions, respectively [16, 34]. Consistent with the results of previous studies [16, 34], most strains in this study were grouped as s1/m1/i1, irrespective of gastric symptoms, with a third of strains being classified as s1a subtype and two-thirds of strains being classified as s1c subtypes. Several studies have reported that subtypes s1a and m1 are associated with peptic ulcer, intestinal metaplasia, or gastric cancer [15, 44, 45]. However, the proportion of s1c subtype was significantly higher in the gastric cancer group in this study (Table 2). The proportion of s1c was > 2.5-fold higher in peptic ulcer strains (72%) and > 3.5-fold higher in gastric cancer strains (79%) than those (28% in peptic ulcer strains and 21% in gastric cancer strains) of s1a, demonstrating that the s1c subtype may increase the risk of severe symptoms. Since all H. pylori strains carry different actual vacuolating abilities [3], the risk rates based on the expressed toxicity were analyzed. Values for the vacuolating cytotoxin activities of all 116 strains were widely distributed from 0 to 191 with no significant difference among gastric symptoms (Table 2). However, when vacuolating toxin activities were divided into four levels as shown in Table 3, gastric cancer strains were noticeably more classified into the extremely strong group than strains in other symptom groups (p < 0.05). This result might reflect the extent to which vacuolating toxin activity has a critical role in the development of gastric cancer.

The vacAs1/i1/m1 gene is closely associated with cagA [3, 46], and therefore the repeat pattern of EPIYA of cagA in all 116 isolates was also investigated. Most East Asian H. pylori strains carry the cagA gene that encodes CagA proteins of the EPIYA-ABD type (Asian type CagA), irrespective of gastric symptoms [3, 21, 41]. Our results are consistent with these previous studies that demonstrated a low-level association of cagA with gastric cancer and peptic ulcer in Asian strains [6, 21]. In this study, since most isolates were found to be EPIYA-ABD type, the difference in risk rate between EPIYA type and multi-EPIYA type could not be identified.

The gastric mucus moves rapidly outward and H. pylori has to overcome the shedding flow of mucus by using bacterial factors to attach to the epithelium of the gastric mucosa. Therefore, H. pylori adherent factors play important roles in the initial colonization, long-term persistence, and gastric pathogenesis [47]. Several H. pylori adhesins belonging to the H. pylori outer membrane prion (HOP) subfamily, including BabA (HopS), SabA (HopP), AlpA (HopC), AlpB (HopB), HopZ, and OipA (HopH), have been extensively analyzed to understand bacterial interaction with the epithelial surface of the gastric mucosa [32].

In this study, the presence of alpA, babA2, and hopZ in H. pylori strains with severe gastric symptoms was significantly higher than that in chronic gastritis strains (Table 5). alpA is known as a virulence factor involved in signal transduction of host epithelial cells during H. pylori infection [4, 48]. Previous studies also demonstrated that an H. pylori ΔalpA/B mutant poorly colonized the stomachs of guinea pigs, mice, and Mongolian gerbils, indicating that AlpA and AlpB play an important role in bacterial colonization [24, 25, 48]. AlpA and AlpB have not yet been clearly identified as a virulence factor in humans and their role remains controversial [4]; however, in this study, we confirmed that the presence of alpA was significantly correlated with the development of gastric cancer in H. pylori infection (Table 5, p < 0.05).

Previous studies showed that the presence of the babA2 gene substantially increased the risk of peptic ulcer development [28, 49]. In addition, a genome-wide association study on European H. pylori isolates demonstrated that, compared with isolates from gastritis patients, the gastric cancer phenotype was associated with the babA2 gene [42]. Conversely, a meta-analysis study reported a lack of significant correlation in Asian populations [49]. This controversial result of the Asian type is probably due to the high overall prevalence of the babA2 gene and significant heterogeneity in Asian isolates [3]. Limiting factors, such as the distinct genotypic profile of Western/Asian isolates and poor correlation between the presence of babA2 gene and actual expression of BabA2 protein also hamper demonstrating the risk of the babA2 gene on disease outcome [28, 50]. Although there is still controversy as to whether the babA2 gene is related to the development of severe gastric symptoms, we observed the presence of the babA2 gene with high frequency in patients with peptic ulcer and gastric cancer in this study (Table 5, p < 0.05).

hopZ was found to be regulated at the transcriptional level according to pH change [51], and hopZ expression depends on an on/off switch during early bacterial colonization [29]. These results indicate that hopZ has a strong selectivity in vivo, plays an important role in adaptation to the host environment [52], and has an essential role in colonization of the gastric mucosa during early infection [29]. Nevertheless, it is unclear whether hopZ is related to other virulence factors or is linked to other clinical diseases [29, 53]. However, we confirmed that the presence of the hopZ gene had a significant correlation with the disease outcome of gastric cancer (Table 5, p < 0.05).

Unlike alpA, babA2, and hopZ, the difference in gene prevalence of alpB, opiA, and sabA was difficult to determine according to the clinical outcomes. alpB has high homogeneity with alpA and is known to play a similar important role in adhering to gastric tissues and colonization [24, 25, 48]. In this study, most isolates from each clinical symptom group were found to have alpB (97%), and therefore alpB may not be a useful marker for predicting the clinical outcome of H. pylori infection. Previous studies reported that oipA is closely linked to the expression of cagA [31, 54], which was also confirmed in this study with only 2 out of 112 oipA(+) strains being cagA(-) strains. Yamaoka Y. et al. reported that oipA can be functionally turned “on” or “off ” by a slipped strand mispairing mechanism [6]. Several studies of prevalence and meta-analysis demonstrated that the oipA “on” function increased the risk of development of peptic ulcer and gastric cancer [55]. Conversely, an induced oipA “on” status is reported to have no relation to the risk of severe gastric symptoms [27, 54, 56]. This study focused on discovering virulence factors of H. pylori that could be detected by PCR and then used as markers of development of gastric symptoms, which would be more useful in clinical approaches than sequencing to identify the functional status of oipA. This limitation made it more difficult to find any link between oipA prevalence and disease outcome in this study. Moreover, oipA does not appear to be a useful marker for predicting the clinical outcome of H. pylori infection because most H. pylori isolates in Korea were identified as virulent strains [56].

sabA is known to be associated with chronic infection establishment of H. pylori [57]. The proportion of the sabA gene was slightly higher in chronic gastritis strains than in others, although this lacked statistical significance (Table 5). A survey in Japan reported that sabA was linked to gastric cancer [58]. However, although severe neutrophil infiltration and epithelial atrophy are associated with sabA, there are reports that sabA does not affect clinical outcome, and therefore the relationship between sabA and clinical diseases remains controversial [58, 59].

Alleles of restriction and modification systems, including iceA and hrgA, are reportedly predictive of gastric symptoms in East Asia [14, 33, 60]. However, the relationship between H. pylori iceA/hrgA and clinical outcomes is still controversial [5, 13, 61]. There are two main allelic variants, iceA1 and iceA2, of which iceA1 is reported to be predominant in East Asia [62]. Several studies have suggested that iceA1 is frequently found in strains isolated from patients with peptic ulcer and gastric cancer [33, 60, 62], whereas others have shown different findings [16, 63], even though iceA1 is presumed to facilitate neutrophil filtration and inflammation [62, 64]. In this study, the prevalence of iceA1 was higher in peptic ulcer- and gastric cancer-associated strains in comparison with that of iceA2, and significant discrepancies were not found. H. pylori has a highly heterogeneous and variable type II R-M system, whereas the hypIII R-M system contains two genes, hypIIIR and hypIIIM [12]. The hrgA gene, which is considered an important virulence determinant in H. pylori-associated gastric diseases, can replace hypIIIR [14]. Therefore, the correlation between hrgA/hypIIIR status and clinical outcome has been used to determine H. pylori toxicity [14]. Subgroup analysis of possible correlations between clinical outcome and hrgA/hpyIIIR status suggested that the prevalence of the hrgA gene was increased among gastric cancer patients (42%) in East Asian countries compared with that in patients without gastric cancer (17%) [14]. In contrast, another study reported that hrgA/hypIIIR status and iceA genotypes are not related to gastric outcomes, although regional differences in the prevalence between Asia and the West were observed [61]. In addition, these authors suggested that the prevalence of the hrgA gene was not related to other putative virulence factors, such as cagA, vacA, or iceA, in either East Asia or Western countries [61]. In this study, the prevalence of hrgA gene was higher, though not significantly, in chronic gastritis strains than in other strains, showing different results from the previous study. Thus, the results of studies on hrgA are contradictory. This may be because hrgA prevalence was not related to other putative virulence factors (cagA, vacA, or iceA), which play an important role in the development of pestilent gastric disease caused by H. pylori infection [22, 61].

Members of the pathogenic island region (CagPAI) of H. pylori containing the type IV secretion system have been proposed as playing a role in the pathogenesis of gastric diseases [12, 19]. dupA is a component gene of the type IV secretion system, which encompasses both jhp0917 and jhp0918. dupA is the first identified disease-specific H. pylori virulence factor that induces duodenal ulcer and has a suppressive action on gastric cancer [22]. Despite the reported gastric cancer inhibitory function, a pooled analysis of data from three Western countries (the USA, Belgium, and South Africa) linked the presence of the dupA gene to peptic ulcer and gastric cancer [65]. In addition, although there is a large regional difference in prevalence of the dupA gene, this may be a risk factor for gastric cancer along with duodenal ulcer [66]. However, a meta-analysis using 11–12 studies on dupA (9 countries) reported that dupA had no correlation with the development of peptic ulcer and gastric cancer [67]. Other studies also reported no correlation between dupA and gastric disease outcome, and this association remains controversial [68, 69]. In this study, the prevalence of dupA (jhp0917) was higher, but not significantly, in peptic ulcer- and gastric cancer-associated strains than that in chronic gastritis-associated strains. The dupA gene may be mutated and protein expression may be inhibited. Therefore, a study based on DupA protein expression should be considered to accurately analyze the correlation between dupA expression and clinical outcome [6].

Prediction of the severe gastric disease outcome of H. pylori using one virulence factor is difficult. Many virulence factors have been correlated with each other. The oipA “on” status has been found to be associated with cagA and vacA [27, 30], and other major virulence factors have also been reported to be correlated [3]. Therefore, severe symptoms may originate from the complex results of several linked virulence factors, rather than the function of a single factor [3, 52, 53, 57]. Multiple correspondence analysis has been shown to serve as a better approach for the prediction of peptic ulcer, gastric cancer, and MALT lymphoma [26, 57]. Therefore, this study was conducted using the four identified virulence factors, vacAs1c, alpA, babA2, and hopZ, that showed significant discrepancies in their prevalence between chronic gastritis strains and severe symptoms, to investigate whether clinical outcome could be predicted with each combination. Eleven combinations of genotypes using the four virulence factors were generated (Table 9), all of which showed an OR value of 3.3 or higher (p < 0.001). These results confirmed that all four factors could affect gastric symptoms and be used as markers for disease outcome. Among the combinations, the triple genotype vacAs1c/alpA/babA2 was the most predictable for the development of severe symptoms (OD: 15.8). The average OR of double and triple combination groups containing babA2 was significantly higher than that of the other groups. Taken together, these data suggest that the genotype factor babA2 might contribute more to the development of severe gastric symptoms than others.

Various other factors, as well as genotype factors, play an important role in the development of severe clinical symptoms, including age, sex, genetic characteristics, diet conditions, and underlying diseases [3, 54]. The link between gene presence and functional protein expression must also be considered [28]. In addition, the possibility of complex H. pylori infections in which different genotype combinations coexist cannot be ruled out [70]. In this case, only one strain of H. pylori was isolated from each patient, making it difficult to analyze the correlation between virulence factors and disease outcome. The large regional differences in the prevalence of virulence factors also impose limitations, causing bias in the analysis results. In this study, 116 strains of H. pylori isolates from patients with no significant age difference were analyzed (Fig. 1): all strains were vacAs1 genopositive while 114 strains were cagA genopositive. The OR value for the development of gastric symptoms and the statistical significance in the relevance of genotype factors to gastric symptoms were calculated by the adjustment of sex using the Mantel-Haenszel method [71] to overcome gender imbalance. However, other limiting factors in this study remain a challenge.

This study attempted to identify candidate virulence factors of H. pylori that can predict severe gastric symptoms via PCR screening. Since this study analyzed H. pylori isolated over a limited period of time, it could not provide information on the current prevalence of the virulence factors. However, we found that the presence of vacAs1c, alpA, babA2, and hopZ genes could increase the risk of disease outcome in infections with toxigenic H. pylori that also harbor the cagA and vacA genes. Moreover, we confirmed that severe gastric symptoms could be predicted at a high level of possibility through multiple correspondence analysis using a combination of these four genes. This could improve our understanding of the role of these virulence factors of H. pylori, which will assist in early prediction of disease outcome through the simple PCR method and thereby enable appropriate treatment.