Presence of cagPAI genes and characterization of vacA s, i and m regions in Helicobacter pylori isolated from Alaskans and their association with clinical pathologies

Abstract

Introduction.

Gastric cancer is a health disparity in the Alaska Native people. The incidence of Helicobacter pylori infection, a risk factor for non-cardia gastric adenocarcinoma, is also high. Gastric cancer is partially associated with the virulence of the infecting strain.

Aim.

To genotype the vacA s, m and i and cag pathogenicity island (cagPAI) genes in H. pylori from Alaskans and investigate associations with gastropathy.

Methodology.

We enrolled patients with gastritis, peptic ulcer disease (PUD) and intestinal metaplasia (IM) in 1998–2005 and patients with gastric cancer in 2011–2013. Gastric biopsies were collected and cultured and PCR was performed to detect the presence of the right and left ends of the cagPAI, the cagA, cagE, cagT and virD4 genes and to genotype the vacA s, m and i regions.

Results.

We recruited 263 people; 22 (8 %) had no/mild gastritis, 121 (46 %) had moderate gastritis, 40 (15%) had severe gastritis, 38 (14 %) had PUD, 30 (11 %) had IM and 12 (5 %) had gastric cancer. H. pylori isolates from 150 (57%) people had an intact cagPAI; those were associated with a more severe gastropathy (P≤0.02 for all comparisons). H. pylori isolates from 77 % of people had either the vacA s1/i1/m1 (40 %; 94/234) or s2/i2/m2 (37 %; 86/234) genotype. vacA s1/i1/m1 was associated with a more severe gastropathy (P≤0.03 for all comparisons).

Conclusions.

In this population with high rates of gastric cancer, we found that just over half of the H. pylori contained an intact cagPAI and 40 % had the vacA s1/i1/m1 genotype. Infection with these strains was associated with a more severe gastropathy.

INTRODUCTION

Helicobacter pylori is a common human infection with over 70 % of people infected in some countries [1]. All H. pylori-infected persons have mild to severe gastric mucosal inflammation, but in some, infection leads to chronic active gastritis or peptic ulcer disease (PUD) [2–4]. Additionally, H. pylori-infected persons have at least a twofold increased gastric cancer risk when compared with uninfected persons [5]. Because of this, H. pylori is characterized as a group I carcinogen by the International Agency for Research on Cancer (World Health Organization) and a risk factor for non-cardia gastric adenocarcinoma [6–8]. Despite H. pylori’s role in disease aetiology development, it is estimated that infected persons have only a 10 to 20% lifetime risk of developing PUD and a 1 to 2% risk of developing gastric cancer [9, 10]. Acquiring these diseases depends upon the inflammatory response to chronic colonization, which is likely determined by a combination of factors, including the virulence of the infecting strain.

Two virulence markers are the cytotoxin-associated gene pathogenicity island (cagPAI) and the vacuolating cytotoxin gene A (vacA). Not all H. pylori contain a cagPAI and it is incomplete in some strains [11, 12]. cagPAI integrity is critical in the interaction between H. pylori and its host. The majority of cagPAI genes encode proteins that form type IV secretion system (T4SS) components; inactivation of some of these genes can result in a non-functional T4SS, causing strains to function similarly to those with a completely absent cagPAI [11–19]. CagA, coded for by the cagPAI cagA gene, is an oncoprotein that is inserted into host epithelial cells via the T4SS and is associated with increased gastric cancer risk [20–22]. CagA has several Glu-Pro-Ile-Tyr-Ala (EPIYA) motifs, which undergo tyrosine phosphorylation at the SHP-2 enzyme [23]. These tyrosine phosphorylation motifs (TPMs) are described as A-, B-, C-, or d- and are associated with gastric pathology with the d- motif having the highest affinity for SHP-2 and increased gastric cancer risk; multiple C- motifs may also increase cancer risk [24, 25].

VacA is a 95kD toxin that induces vacuolization in epithelial cells in vitro by loosening tight junctions and forming pores in cell membranes [26–28]. All H. pylori have a vacA gene, although there is vacuolating activity variation due to sequence heterogeneity at the 5′ end [signal (s), intermediate (i) and middle (m) regions] [29, 30]. Many groups have found a correlation between toxin activity and pathogenicity, with strains having vacA s1, i1 and m1 genotypes being more virulent and associated with increased gastric cancer risk [22, 29–37].

Gastric cancer morbidity and mortality is a significant health disparity in Alaska Native people. Data from 1999 to 2013 demonstrated an age-adjusted annual incidence rate of 22.6 (Alaska Native) vs 7.1 (US white) cases per 100 000 population [38]. In 2007–2011, the gastric cancer mortality rate in Alaska Native people was 11.6/100 000, nearly four times higher than the mortality rate in US whites (3.0/100,000) [38]. Alaska has high rates of H. pylori infection, antimicrobial-resistant H. pylori, H. pylori treatment failure and reinfection after successful H. pylori eradication [39–44], thus it is not practical to attempt to lower gastric cancer rates in Alaska Native people by treating all H. pylori-positive persons [45]. In addition, a massive eradication campaign could result in widespread antimicrobial resistance, which would be a concern not only for H. pylori but also for other organisms with high infection rates in Alaska, such as Streptococcus pneumoniae and Haemophilus influenzae. Another strategy is to identify factors that might increase a person’s risk of developing cancer. This could lead to programmes identifying persons at highest risk and offering H. pylori treatment or regular surveillance with esophagogastroduodenoscopy (EGD) for early gastric cancer detection [46].

A previous study has shown that a majority of H. pylori collected from Alaskans contain the cagA gene and nearly half have the vacA s1/m1 genotype [47]. As described earlier, CagA protein functionality is at least partially determined by the presence of the entire cagPAI as well as the CagA EPIYA motifs. The vacA i region may also play a role in H. pylori virulence. In this study, we genotyped the vacA s, i and m regions and identified the existence of an intact cagPAI from H. pylori isolated from Alaskans with a variety of gastric pathologies.

METHODS

Participants

The participants in this study were recruited via two mechanisms. Both studies were approved by the Centers for Disease Control and Prevention and the Alaska Area Indian Health Service Institutional Review Boards as well as review boards for the participating Tribal Health Organizations.

Twenty gastric cancer cases were prospectively recruited at the Alaska Native Medical Center (ANMC) as part of a 2011–2013 study investigating gastric cancer risk factors (Fig. 1). Twelve of those patients had a biopsy sample that grew H. pylori, so they were eligible for the present analysis. Persons ≥18 years old were eligible to participate if, after EGD, they had a newly diagnosed gastric cancer and received follow-up care at the ANMC. Participants initially received a brief description of the study from their clinical provider who contacted study staff if their patient indicated interest. Study staff met with the participant and collected informed consent.

Fig. 1.

Flow diagram of enrolled participants.

Participants with gastritis, PUD and intestinal metaplasia (IM; n=251) were prospectively recruited as part of a previously described H. pylori reinfection study [44]. Briefly, persons ≥18 years old undergoing EGD from September 1998 through January 2005 were recruited at hospitals and clinics in four Alaskan communities (Fig. 1). Persons were eligible for enrolment into that study if they had a positive ¹³C-urea breath test (Meretek Diagnostics, Inc., Lafayette, CO, USA) at the time of their EGD. Persons were eligible for the present study if their H. pylori infection was also confirmed by culture and they had a histological sample read by study pathologists. Gastritis and IM diagnoses were determined by two study pathologists working independently of each other using standardized reporting criteria [48]; PUD was determined by endoscopy and chart review. For the purposes of this analysis, both acute and chronic gastritis were analysed together. Levels of acute gastritis were: none, mild (rare clustered neutrophils), moderate (neutrophils involved in several crypts) and severe (multiple extensive crypt lesions in nearly every field). Levels of chronic gastritis were: none, mild (lymphoplasmacytic infiltrates involving only the upper part of the mucosa), moderate (lamina propria involved at all levels with expiation by lymphoplasmacytic infiltrates) and severe (dense lymphoplasmacytic infiltrates at all levels with severe expansion or obliteration of epithelium). A sample was considered pathological if at least one pathologist considered it to have pathological changes. At the time of enrollment, study staff met with the participant and collected informed consent.

H. pylori genotyping

During the EGD, gastric biopsy specimens were collected, H. pylori was cultured, and DNA was extracted as described previously [47, 49]. PCR analysis was performed to detect the presence of the right and left ends of the cagPAI, the cagA, cagE, cagT and virD4 genes and to determine the genotype of the vacA s, i and m regions (Table 1). Due to potential sequence heterogeneity, two primer sets were used for some genes. In those cases, the first set listed in Table 1 was used, followed by testing with the second set on all extracts that were negative by the first. To determine the cagA EPIYA TPMs, positive cagA PCR products (CAGTF/CAGTR or A2530S/3000AS) were sequenced using the ABI 3130 genetic analyser.

Table 1.

PCR primer sequences for amplification of cag PAI and vacA genes

Gene	Primer name	Sequence	Reference
cagA*,†	CAGTF	5’ACCCTAGTCGGTAATGGG	[69]
	CAGTR	5’GCTTTAGCTTCTGAYACYGC	[69]
	A2530S	5’GTTAARAATRGTGTRAAYGG	[70]
	3000AS	5’TTTAGCTTCTGATACCGC	[70]
cagE †	cagEAF	5’TTGAAAACTTCAAGGATAGGATAGAGC	[13]
	cagEAR	5’GCCTAGCGTAATATCACCATTACCC	[13]
	cagEBF	5’AGTGATGCTTTGAGTCGCAAGTC	[71]
	cagEBR	5’TGGGGCAATAGTGTGATGACG	[71]
cagT †	cagTAF	5’CCATGTTTATACGCCTGTGT	[54]
	cagTAR	5’CATCACCACACCCTTTTGAT	[54]
	cagTBF	5’TCTAAAAAGATTACGCTCATAGGCG	[72]
	cagTBR	5’CTTTGGCTTGCATGTTCAAGTTGCC	[72]
virD4 †	virD4BF	5’TTTCATAGGTTCTATGGCAAGCGGG	[73]
	virD4BR	5’TTAGCGCCATTCCTACCATAACC	[73]
	virD4AF	5’TTTATGATGATAATCGATCGCC	[73]
	virD4AR	5’GAACGCCTGCCCTGCGTAAGCG	[73]
Right end cagPAI†	3’F	5’GGCTCAAGCTCGGAATGAT
	emptyR	5’CTCTTTTTGTGCCTTTGATTGAA
Left end cagPAI†	emptyF	5’CCAAATACATTTTGGCTAAATAAAC
	5’R	5’GCTTATCAGTCAAATTGTTTTTG
vacA s1a	SS1F	5’GTCAGCATCACACCGCAAC	[29]
	VAIR	5’CTGCTTGAATGCGCCAAAC	[29]
vacA s1b	SS3F	5’AGCGCCATACCGCAAGAG	[29]
	VAIR	5’CTGCTTGAATGCGCCAAAC	[29]
vacA s1c	S1cF	5’CTYGCTTTAGTRGGGYTA	[74]
	VAIR	5’CTGCTTGAATGCGCCAAAC	[29]
vacA s2	SS2F	5’GCTAACACGCCAAATGATCC	[29]
	VAIR	5’CTGCTTGAATGCGCCAAAC	[29]
vacA m1a	VA3F	5’GGTCAAAATGCGGTCATGG	[29]
	VA3R	5’CTAATGCCATTGGTACCTGTAGAAAC	[29]
vacA m1b	VAMF	5’CCCCAATGCAGTCATGGAT	[75]
	VAMR	5’GCTGTTAGTGCCTAAAGAAGCAT	[75]
vacA m2	VA4F	5’GGAGCCCCAGGAAACATTG	[29]
	VA4R	5’TGTCATAACTAGCGCCTTGCAC	[29]
vacA i1	F1	5’GTTGGGATTGGGGGAATGCCG	[63]
	c1R	5’TTAATTTAACGCTGTTTGAAG	[63]
vacA i2	F1	5’GTTGGGATTGGGGGAATGCCG	[63]
	c2R	5’GATCAACGCTCTGATTTGA	[63]

^*Also used to determine the EPIYA sequence for cagA-positive samples.

^†Sample is considered to have an intact cagPAI if all of these genes are present.

Statistical analysis

Persons were considered to have a particular H. pylori genotype if it was detected in ≥1 gastric biopsy specimen. We categorized the cagPAI as (1) intact, if the cagPAI left and right ends, cagA, cagE, cagT and virD4 genes were all detected; (2) partially deleted, if at least one, but not all, of the genes was detected; or (3) negative, if no cagPAI genes were detected. Analyses began with six disease categories ordered according to severity: no/mild gastritis, moderate gastritis, severe gastritis, PUD, IM, and gastric cancer. We tested for statistical significance for the prevalence of H. pylori genotypes across the entire clinical spectrum using a Mantel–Haenszel chi-square test with a modified ridit score [50]. The modified ridit score makes no assumption about the spacing between the categories of disease, only on the rank ordering. We used the likelihood ratio chi-square to test the prevalence of H. pylori genotypes between persons with gastritis, PUD, and IM/gastric cancer combined and between IM/gastric cancer combined and all others. Because gender differed between the two cohorts (1998–2005 and 2011–2013) and both gender and age differed between persons with different disease categories, we additionally ran age and gender adjusted P-values. The results remained similar and are not reported. Analyses were run within SAS software, version 9.4 (Cary, NC, USA). P-values were two-sided and exact when sample sizes necessitated and P<0.05 was considered significant.

RESULTS

Participants

Two hundred and sixty-three people were enrolled and had an H. pylori isolate available for genotyping (Table 2). Of these 263 participants, 255 (97 %) were Alaska Native, 6 (2 %) were Caucasian and 1 (0.4 %) each were African American and Asian American. Twenty-two out of 263 (8 %) participants had no/mild gastritis, 121/263 (46 %) had moderate gastritis, 40/263 (15 %) had severe gastritis, 38/263 (14 %) had peptic ulcer disease (PUD), 30/263 (11 %) had IM and 12/263 (5 %) had gastric cancer. An additional nine recruited gastric cancer patients are not included in these analyses because they either had no biopsy sample collected (n=6) or did not have H. pylori isolated from their biopsy (n=3). Older median age and male gender were associated with more severe gastric pathologies (P≤0.001 and ≤0.01, respectively, for all comparisons).

Table 2.

Study participant demographics; Alaska, 1998–2005 and 2011–2013

Characteristic	Total (n=263)	No/mild gastritis (n=22)	Moderate gastritis (n=121)	Severe gastritis (n=40)	PUD* (n=38)	IM* (n=30)	GC* (n=12)
								P-value clinical spectrum	P-value gastritis vs PUD vs IM/GC	P-value IM/GC vs others
Age median (min, max)	48 (18, 88)	50 (24, 79)	45 (18, 88)	46 (24, 77)	51 [21, 73]	51 (18, 78)	59 (38, 77)	0.005	0.001	0.003
Male gender	115 (44 %)	7 (32 %)	42 (35 %)	21 (53 %)	19 (50 %)	17 (57 %)	9 (75 %)	0.0005	0.01	0.01

Significant P-values are bolded.

^*PUD, peptic ulcer disease; IM, intestinal metaplasia; GC, gastric cancer.

cagPAI genotyping

H. pylori collected from 171 (65 %) people contained at least 1 cagPAI region. The presence of any of the six cagPAI regions was independently associated with a more severe gastric pathology regardless of how we grouped the pathologies (P≤0.04 for all but one comparison; Table 3). H. pylori collected from 150 (57 %) people had an intact cagPAI; those were also associated with a more severe gastric pathology (P≤0.02 for all comparisons). H. pylori collected from 92 (35 %) people had a completely deleted cagPAI; those were associated with less severe gastric pathologies (P≤0.02 for all comparisons).

Table 3.

cag pathogenicity island (cagPAI) genotyping of H. pylori strains isolated from Alaskans, 1998–2005 and 2011–2013

cagPAI gene		Total (n=263)	No/mild gastritis (n=22)	Moderate gastritis (n=121)	Severe gastritis (n=40)	PUD* (n=38)	IM* (n=30)	GC* (n=12)
									P-value clinical spectrum	P-value gastritis vs PUD vs. IM/GC	P-value IM/GC vs others
cagPAI left end		159 (60 %)	11 (50 %)	70 (58 %)	22 (55 %)	23 (61 %)	22 (73 %)	11 (92 %)	0.03	0.02	0.007
cagPAI right end		156 (59 %)	11 (50 %)	67 (55 %)	22 (55 %)	23 (61 %)	22 (73 %)	11 (92 %)	0.02	0.01	0.004
cagA		158 (60 %)	11 (50 %)	67 (55 %)	25 (63 %)	23 (61 %)	22 (73 %)	10 (83 %)	0.02	0.06	0.02
cagE		162 (62 %)	11 (50 %)	71 (59 %)	24 (60 %)	22 (58 %)	22 (73 %)	12 (100 %)	0.02	0.01	0.003
cagT		158 (60 %)	11 (50 %)	68 (56 %)	23 (58 %)	22 (58 %)	22 (73 %)	12 (100 %)	0.01	0.007	0.002
virD4		162 (62 %)	11 (50 %)	72 (60 %)	24 (60 %)	22 (58 %)	22 (73 %)	11 (92 %)	0.04	0.04	0.01
Intact cagPAI†		150 (57 %)	11 (50 %)	63 (52 %)	22 (55 %)	22 (58 %)	22 (73 %)	10 (83 %)	0.02	0.02	0.006
Partially deleted cagPAI‡		21 (8 %)	0 (0 %)	15 (12 %)	3 (8 %)	1 (3 %)	0 (0 %)	2 (17 %)	0.31	0.20	0.54
Completely deleted cagPAI§		92 (35 %)	11 (50 %)	43 (36 %)	15 (38 %)	15 (39 %)	8 (27 %)	0 (0 %)	0.01	0.02	0.006
EPIYA	AB	4 (2 %)	0 (0 %)	1 (1 %)	1 (4 %)	0 (0 %)	2 (9 %)	0 (0 %)	0.45‖	0.69‖	0.77‖
	ABC	133 (84 %)	11 (100 %)	56 (84 %)	22 (88 %)	19 (83 %)	17 (77 %)	8 (80 %)
	ABCC	15 (9 %)	0 (0 %)	9 (13 %)	1 (4 %)	1 (4 %)	3 (14 %)	1 (10 %)
	ABD	3 (2 %)	0 (0 %)	1 (1 %)	0 (0 %)	2 (9 %)	0 (0 %)	0 (0 %)
	ABCCC	2 (1 %)	0 (0 %)	0 (0 %)	0 (0 %)	1 (4 %)	0 (0 %)	1 (10 %)
	ACC	1 (1 %)	0 (0 %)	0 (0 %)	1 (4 %)	0 (0 %)	0 (0 %)	0 (0 %)
B-TPM¶	EPIYA	74	5 (45 %)	32 (48 %)	13 (54 %)	10 (43 %)	9 (41 %)	5 (50 %)	0.77	0.84	0.64
	EPIYT	58	3 (27 %)	28 (42 %)	5 (21 %)	8 (35 %)	11 (50 %)	3 (30 %)
	Other#	25	3 (27 %)	7 (10 %)	6 (25 %)	5 (22 %)	2 (9 %)	2 (20 %)

Significant P-values are bolded.

^*PUD, peptic ulcer disease; IM, intestinal metaplasia; GC, gastric cancer.

^†Presence of all six cagPAI regions (cagPAI left end, cagPAI right end, cagA, cagE, cagT, virD4).

^‡Presence of ≥1 but <6 cagPAI regions.

^§Absence of all cagPAI regions.

^‖P-values for AB, ABC vs all others.

^¶Tyrosine phosphorylation motif.

^#ESIYT (n=12), EDSIYT (n=6), EDPIYT (n=6) and ESIYA (n=1).

The cagA gene was present in H. pylori collected from 158 people. Of those, 133 (84 %) had a cagA gene with the ABC TPM (Table 3). When comparing cagA AB and ABC EPIYA TPMs with all others (ABCC, ABCCC, ACC, ABD), EPIYA TPMs were not associated with gastric pathologies. At the EPIYA-B TPM, the majority of the cagA genes had either an EPIYA (n=74; 47 %) or EPIYT sequence (n=58; 37 %). Other EPIYA-B TPM sequences identified were ESIYT (n=12; 8 %), EDSIYT (n=6; 4 %), EDPIYT (n=6; 4 %) and ESIYA (n=1; 1 %). The EPIYA-B TPM was not associated with gastric pathologies. There was no EPIYA heterogeneity at the EPIYA-A TPM and all but three people had an EPIYA sequence at the EPIYA-C TPM; the other EPIYA-C TPM sequences were one each of EPIYT, ESIYA and EPVYA.

vacA genotyping

Of the 263 people recruited, 250 (95 %) had H. pylori with the s region genotyped, 247 (94 %) with the i region genotyped, 257 (98 %) with the m region genotyped and 234 (89 %) with all 3 regions genotyped (Table 4). Of isolates with the s region genotyped, 140 (56 %) had an s1 subtype and 98 (39 %) an s2 subtype. Of isolates with the i region genotyped, 103 (42 %) had an i1 subtype and 127 (51 %) an i2 subtype. Of isolates with the m region genotyped, 112 (44 %) had an m1 subtype and 136 (53 %) an m2 subtype. Isolates containing multiple vacA genotypes (n=12, s region; n=17, i region; n=9, m region) are removed from further discussion and analysis. vacA subtypes s1, i1 and m1 were all independently associated with a more severe gastric pathology (P≤0.02 for all comparisons). The vacA s1/i1/m1 [94/234 40 %)] and s2/i2/m2 [86/234 (37 %)] genotypes accounted for 77 % of the H. pylori isolates. Of isolates with an s1/i1/m1 genotype, 93 (99 %) also had an intact cagPAI; of isolates with an s2/i2/m2 genotype, 77 (90 %) had a partially or completely deleted cagPAI. The vacA s1/i1/m1 genotype was associated with more severe gastric pathologies (P≤0.03 for all comparisons).

Table 4.

vacA genotyping of H. pylori strains isolated from Alaskans, 1998–2005 and 2011–2013

vacA region genotype		Total (n=263)	No/mild gastritis (n=22)	Moderate gastritis (n=121)	Severe gastritis (n=40)	PUD* (n=38)	IM* (n=30)	GC* (n=12)
									P-value clinical spectrum	P-value gastritis vs PUD vs. IM/GC	P-value IM/GC vs others
Signal (s) region (n=250)	s1	140 (56 %)	11 (50 %)	61 (52 %)	20 (53 %)	20 (57 %)	17 (68 %)	11 (92 %)	0.02	0.006	0.002
	s2	98 (39 %)	11 (50 %)	50 (42 %)	17 (45 %)	14 (40 %)	6 (24 %)	0 (0 %)
	s1/s2†	12 (5 %)	0 (0 %)	7 (6 %)	1 (2 %)	1 (3 %)	2 (8 %)	1 (8 %)
Intermediate (i) region (n=247)	i1	103 (42 %)	8 (36 %)	41 (36 %)	13 (35 %)	16 (43 %)	16 (64 %)	9 (82 %)	0.003	0.0004	0.0001
	i2	127 (51 %)	14 (64 %)	66 (57 %)	18 (49 %)	21 (57 %)	7 (28 %)	1 (9 %)
	i1/i2†	17 (7 %)	0 (0 %)	8 (7 %)	6 (16 %)	0 (0 %)	2 (8 %)	1 (9 %)
Mid (m) region (n=257)	m1	112 (44 %)	9 (41 %)	45 (38 %)	16 (41%)	16 (43 %)	17 (57 %)	9 (90 %)	0.02	0.01	0.003
	m2	136 (53 %)	13 (59 %)	69 (58 %)	21 (54 %)	20 (54 %)	13 (43 %)	0 (0 %)
	m1/m2†	9 (3 %)	0 (0 %)	5 (4 %)	2 (5 %)	1 (3 %)	0 (0 %)	1 (10 %)
vacA genotype (n=234)	s1/i1/m1	94 (40 %)	8 (36 %)	38 (34 %)	12 (34 %)	15 (45 %)	13 (57 %)	8 (80 %)	0.02	0.03	0.004
	s1/i2/m2	18 (8 %)	2 (9 %)	9 (8 %)	4 (11 %)	2 (6 %)	1 (4 %)	0 (0 %)
	s2/i2/m2	86 (37 %)	11 (50 %)	45 (41 %)	11 (31 %)	13 (39 %)	6 (26 %)	0 (0 %)
	Other‡	10 (4 %)	1 (5 %)	5 (5 %)	2 (6 %)	2 (6 %)	0 (0 %)	0 (0 %)
	mixed†	26 (11 %)	0 (0 %)	14 (13 %)	6 (17 %)	1 (3 %)	3 (13 %)	2 (20 %)

Significant P-values are bolded.

^*PUD, peptic ulcer disease; IM, intestinal metaplasia; GC, gastric cancer.

^†Mixed infections removed from statistical analyses.

^‡Other includes s1/i1/m2 (n=4) and s1/i2/m1 (n=6).

DISCUSSION

In 2011, we reported on H. pylori genotypes among Alaskans [47]. The current report enhances that data by including patients with IM and gastric cancer and by genotyping multiple cagPAI genes and three vacA regions in all isolates, including those described in the earlier report. Additionally, in that report, the physician performing the endoscopy graded the gastritis severity by visual observation of the gastric mucosa. Since then, we enlisted two pathologists to review the histological sections, allowing a more precise and objective estimate of gastritis presence and severity [48]. We found that the majority of H. pylori strains in the current study contained an intact cagPAI and either a vacA s1/i1/m1 or s2/i2/m2 genotype. The presence of any cagPAI region we investigated as well as a vacA s1, i1 or m1 subtype was associated with a more severe gastric pathology, as was the presence of an intact cagPAI and the vacA s1/i1/m1 genotype.

We detected at least one cagPAI gene in H. pylori isolates collected from nearly 2/3 of study participants; an intact cagPAI was present in just over half of the participants. Only 8 % of participants had a partially deleted cagPAI, thus the presence of any one cagPAI gene was suggestive of an intact cagPAI. The presence of H. pylori with an intact cagPAI varies geographically at least partially due to H. pylori having colonized humans for nearly 60 000 years as humans migrated around the planet [51]. In a large study of 877 isolates, Olbermann, et al. found the cagPAI intact in ≥95 % of strains associated with persons from East Asia, Asia, New Zealand, West Africa and South Africa, in 81 % of strains from persons from Northeast Africa, in 58 % of strains from persons from Europe, in 28 % of strains from persons from South American Indian populations, and in no strains from persons from some African populations [52]. In the current study, the presence of any cagPAI gene, as well as the presence of an intact cagPAI, was associated with a more severe gastric pathology. In our previous study investigating the presence of only the cagA portion of the cagPAI, the only association we discovered was one between the absence of the cagA gene and an endoscopic diagnosis of esophagitis [47]. With the exception of an ulcer diagnosis, which was evaluated using endoscopy, the clinical diagnoses described in this study were made by pathologists. We believe that is a more accurate method of diagnosing gastritis and IM and likely accounts for some of the difference in clinical associations between this study and our prior one. Furthermore, the data from this study support other reports that suggest the presence of an intact cagPAI confers an increase in the severity of a patient’s gastric pathology [13, 15, 16, 53, 54]. A possible explanation reported by Hanada, et al. is that the accumulation of DNA double-stranded breaks is significantly greater in cagPAI-positive compared with cagPAI-negative strains [55].

Although the number of gastric cancer patients recruited with H. pylori detected in their biopsy was small, we found the cagE and cagT genes in all of those isolates. This supports data from India where ≥95 and ≥90 %, respectively, of H. pylori-infected persons with gastric cancer were infected with H. pylori containing the cagE or cagT gene [53, 56]. Presence of the CagA oncoprotein in host cells and the host’s immunological response to infection contribute to a patient’s outcome from H. pylori. The cagT gene is required for a functional T4SS syringe and successful injection of CagA into the host cell [17, 57, 58]. The CagE protein is known to induce IL8 secretion and to mediate host-cell cytokine rearrangements in infected cells [15, 59]. Thus, the proteins encoded for by cagE and cagT are important in promoting inflammation and intestinal cell damage.

In this study, 151 participants were found to be infected with an H. pylori that was cagA-positive with some combination of EPIYA A-, B- and C- TPMs; these TPMs are associated with the western-type CagA [60]. Twenty-one participants were infected with strains with a single d- or multiple C- TPMs, both of which have been associated with increased virulence in some studies [24, 25]; however, that was not true in our study. Over half of our study participants were infected with H. pylori with a modified EPIYA-B TPM, with a little over a third of those having the EPIYT-B TPM. The EPIYT-B TPM has been associated with the induction of lower levels of cellular elongation and IL-8 secretion [61]. Additionally, a study of GenBank sequences found that, compared with gastritis alone, gastric cancer was significantly less associated with the EPIYT-B TPM compared with the EPIYA-B TPM [62]. Due to the small number of gastric cancer participants in our study, we were unable to perform a similar analysis.

Over 75 % of our study participants had an H. pylori with either the vacA s1/i1/m1 or s2/i2/m2 genotype. This is similar to our previous report of 83 % of participants having an isolate with the vacA s1/m1 or s2/m2 genotype [47]. In both studies, a slightly higher percentage of isolates had the genotype that produces more toxin (s1/m1 or s1/i1/m1). In our previous publication, the vacA s1/m1 genotype was associated with an increased risk of having an ulcer at enrollment or a history of PUD; in the current study, the vacA s1/i1/m1 genotype was also associated with a more severe gastric pathology. This is similar to results from our previous study in which we found an association between the vacA s1/m1 genotype and ulcer disease. Additionally, in this study the individual regions of s1, i1 and m1 were each associated with more severe disease. vacA i region data have not yet been reported from Alaska and our finding is not unique to this study. In the paper that first described the i region, Rhead, et al. reported that only infection with an i1 strain was independently associated with increased risk gastric cancer risk [63]. Other groups subsequently found similar associations, although it is not universally true, especially in isolates collected from persons living in East and Southeast Asia [64–68].

This study has three important limitations. First, there were only 12 persons diagnosed with gastric cancer from whose biopsy samples we were able to detect H. pylori. This limits our ability to detect associations between the organism’s genotype and gastric cancer specifically. However, IM is along the clinical spectrum leading to gastric cancer. When we include the 30 participants with that diagnosis along with the 12 with a cancer diagnosis, we are able to identify significant associations that seem to increase gastric cancer risk. Second, the gastric cancer patients were recruited during a different time period than the patients with the other disease categories. Due to the long-term (up to 2 years) nature of the 1998–2005 study [44] from which the majority (n=251) of participants in this analysis were originally recruited, persons with gastric cancer were ineligible for participation in that study. As it is important to understand what could be contributing to the high gastric cancer burden in the Alaska Native people, in this analysis we included persons recruited as part of the 2011–2013 study of gastric cancer. However, both of those studies were cross-sectional in nature, in that the collection of the samples from which H. pylori genotyping was performed was at the same time point that the disease status was determined. Due to the lack of a prospective study design and differential recruitment periods, time could be a confounder in our findings. Finally, the majority of participants in this study were Alaska Native, so the results may not be generalizable to other populations.

In conclusion, in this population with high rates of H. pylori infection and gastric cancer, we found that just over half of the H. pylori strains contained an intact cagPAI and 40 % had the high toxin-producing vacA s1/i1/m1 genotype. Infection with H. pylori containing an intact cagPAI or vacA s1/i1/m1 genotype was associated with a more severe gastric pathology. H. pylori infection and gastric cancer are also prevalent in indigenous populations in other Arctic countries, but there are few strain genotyping data from those countries. We would propose that similar studies to this one be conducted across the circumpolar north to give a more comprehensive picture of H. pylori infection among indigenous peoples of the Arctic.

Abbreviations:

cagPAI

cytotoxin-associated gene pathogenicity island

EGD

esophagogastroduodenoscopy

EPIYA

Glu-Pro-Ile-Tyr-Ala motifs

GC

gastric cancer

i

intermediate

IM

intestinal metaplasia

m

middle

PUD

peptic ulcer disease

s

signal

TPM

tyrosine phosphorylation motif

T4SS

type IV secretion system

vacA

vacuolating cytotoxin gene A