Sequence analysis of the genes encoding for H⁺/K⁺‐ATPase in autoimmune gastritis

Abstract

Background. H⁺/K⁺‐ATPase is the target autoantigen in autoimmune gastritis (AIG), an organ‐specific autoimmune disease with a strong hereditary component.

Aim. To detect possible polymorphisms in H⁺/K⁺‐ATPase α‐ and β‐subunits in AIG patients.

Methods. Blood samples from 12 Finnish AIG patients were sequenced for the coding regions of genes encoding for H⁺/K⁺‐ATPase α‐ and β‐subunits; 50–52 Finnish anonymous blood donors served as controls. Additionally, parietal cell and Helicobacter pylori antibodies and serum pepsinogen I levels (PG I) were analysed.

Results. In the α‐subunit, all patients and controls had C‐allele at the non‐synonymous c.824T>C single nucleotide polymorphism (SNP) resulting in valine substitution for alanine (Val265Ala). In the β‐subunit, a previously unknown non‐synonymous SNP resulting in a substitution of alanine residue for valine (Ala248Val) was found in exon 7 in a single patient and none of the controls. All patients had low serum PG I levels and elevated parietal cell antibodies; three had positive H. pylori serology.

Conclusions. At the non‐synonymous SNP c.824T>C in the α‐subunit of H⁺/K⁺‐ATPase most Finnish individuals with or without AIG have C allele. Genetic variants of the coding regions of genes for H⁺/K⁺‐ATPase α‐ and β‐subunits are not associated with AIG in Finnish patients.

• Autoimmune gastritis
• genetic polymorphisms
• pernicious anaemia.

Abbreviations
AIG	=	autoimmune gastritis
H+/K+‐ATPase	=	H+/K+‐adenosine triphosphatase
IgG	=	immunoglobulin G
SNP	=	single nucleotide polymorphism
H. pylori	=	Helicobacter pylori
PA	=	pernicious anaemia
PCR	=	polymerase chain reaction
PG I	=	pepsinogen I

Introduction

Autoimmune gastritis (AIG) is an organ‐specific autoimmune disease that is manifested as corpus‐restricted inflammation of the gastric mucosa and results in atrophy of the corpus‐type glands, leading to achlorhydria and eventually, because of the lack of the intrinsic factor, to vitamin B12 deficiency and pernicious anaemia (PA). AIG can be recognized in histological specimens from corpus mucosa biopsies where loss of acid‐secreting glands and parietal cells is seen. Biochemical markers, such as parietal cell and H⁺/K⁺‐ATPase antibodies, low serum PG I and often very high serum gastrin concentrations, are typical for the disease. However, atrophic gastritis caused by Helicobacter pylori may be difficult to distinguish from the classical AIG.

There are several reports, especially in the older literature, showing that PA has a strong genetic component. The presence of parietal cell antibodies runs in the families, and about 20% of PA patients have a family history of the disease 1–3. As studied in the Finnish population, half of the parents and 27% of the over 50‐year‐old siblings of PA patients have corpus‐predominant atrophic gastritis 4. The mode of inheritance seems to be autosomal dominant with incomplete penetrance 1,2,4,5.

The target autoantigen in AIG has been shown to be H⁺/K⁺‐ATPase 6. H⁺/K⁺ ‐ATPase is responsible for acid secretion in the stomach and is expressed especially in human gastric parietal cells where it functions as a proton pump, transporting H⁺ in exchange for K⁺ into secretory canaliculi and, from there, into the gastric lumen. H⁺/K⁺ ‐ATPase is a heterodimer consisting of a high‐molecular‐weight catalytic α‐subunit and a smaller but heavily glycosylated β‐subunit 7. The genes encoding the α‐ and β‐subunits are located at chromosomes 19q13.1 and 13q34, respectively 8,9. The α‐subunit contains 22 exons and exhibits a catalytic function during active transport. The β‐subunit of (H⁺/K⁺)‐ATPase is tightly associated with the α‐subunit and is encoded by a gene comprising seven exons. It presumably regulates the function of the α‐subunit 9.

In an attempt to understand the molecular genetic background of AIG, the coding regions of the genes encoding for H⁺/K⁺‐ATPase α‐ and β‐subunits were sequenced in individuals with AIG and for controls in order to reveal possible polymorphisms associated with AIG.

Key messages

• No disease‐associated polymorphisms were found in autoimmune gastritis (AIG) patients in the coding areas of genes encoding for H⁺/K⁺‐ATPase, the target autoantigen in AIG.

• At the non‐synonymous SNP T824C resulting in substitution of amino residue Val265Ala in the α‐subunit of H⁺/K⁺‐ATPase, Finns seem to have mostly C allele. The possible significance of this amino acid change is not known.

Patients and methods

A total of 18 adult patients under 65 years of age, who had earlier undergone gastroscopy at Herttoniemi Hospital and were found to have a severe corpus‐restricted atrophic gastritis, were invited by a letter to participate in the study. Twelve patients (11 females, age range 38–63 years, median age 52 years) agreed, donated a blood sample and answered a questionnaire. All patients gave their written informed consent. The study was approved by the Ethics Committee for Medicine, Helsinki University Central Hospital.

In the questionnaire, patients were asked about vitamin B12 replacement therapy and diseases or treatment of the thyroid gland. Possible family history of vitamin B12 replacement therapy or of atrophic gastritis was also questioned. The patients' hospital records were checked for any information on autoimmune diseases. Clinical data for the patients are shown in Table I.

Table I. Clinical data, serum PG I concentrations, parietal cell and H. pylori antibodies, and histology of the antrum for the autoimmune gastritis (AIG) patients.

No	Gender	Age	Age at dg	PG 1	PCA	Hp ab	Antrum
1	F	50	43	6	257	100	0
2	F	56	53	10	172	50	0
3	F	49	47	14	165	2200	0
4	M	62	43	16	20	50	1
5	F	63	61	6	89	50	1
6	F	54	51	16	32	50	1,2
7	F	38	34	6	509	50	2
8	F	56	51	6	117	100	0
9	F	52	52	6	66	50	0
10	F	52	52	<5	48	1400	0
11	F	47	32	25	462	730	0
12	F	52	52	20	283	50	0

age at dg = age at diagnosis of atrophic gastritis or pernicious anaemia;

PG 1 = serum pepsinogen I level, µg/L;

PCA = parietal cell antibodies, U/mL;

Hp ab = H. pylori antibody titre of the Ig G class, ⩾700 considered elevated;

antrum = histological findings in the antrum (0 = normal, 1 = mild chronic inflammation, 2 = sparse intestinal metaplasia).

In addition, blood samples were available from anonymous Finnish blood donors as controls for sequencing. Their use was approved by the Ethical Committee for the Finnish Red Cross Blood transfusion service. The age range of blood donors in Finland is 18–66 years.

Of the patient serum samples, PG I levels and antibodies for H. pylori and H⁺/K⁺‐ATPase were analysed. Serum PG I concentrations were measured using an immunoenzymometric assay (Gastroset PG1, Orion Diagnostica, Espoo, Finland) and according to the manufacturer, levels below 28 µg/L were considered to indicate the presence of atrophic corpus gastritis. H. pylori antibodies of the immunoglobulin G (IgG) class were measured by an in‐house immunoassay and titres of 700 or higher were considered to be elevated. With this limit the assay has shown a specificity of 93% and sensitivity of 99% as compared with histology 10. Parietal cell antibodies in serum were determined using a commercially available enzyme immunoassay using H⁺/K⁺‐ATPase as antigen (Varelisa Parietal Cell Antibodies, Pharmacia Diagnostics, Freiburg, Germany). According to the manufacturer, concentrations <10 U/mL were normal.

Polymerase chain reaction (PCR)‐sequencing

DNA was extracted from 10 mL of peripheral blood using a genomic DNA purification kit (PureGene®, Gentrasystems, Minneapolis, MI, USA) according to the manufacturer's instructions. The coding regions as well as the flanking splice sites of the H⁺K⁺‐ATPase α‐ (ATP4A) and β‐ (ATP4B) subunits were analysed. Each exon (1‐22) of ATP4A was amplified by PCR with primers as shown in Table II. The seven exons of ATP4B were amplified in seven different PCR‐reactions as listed in Table III. PCR amplifications were performed in a 50 µL volume containing 70–100 ng genomic DNA, 10 pmol of each primer, 10 mM Tris‐HCl pH 8.8, 1.5 mM MgCl, 50 mM KCl and 0.1% Triton X‐100, 10 nmol of each nucleotide and 0,8U Dynazyme polymerase‐enzyme (Finnzymes Oy, Espoo, Finland).

Table II. Primers for the ATP4A gene listed by exon.

Exon	Forward primer	Reverse primer
1	5′‐ccaaggagcctgtaatcagc ‐3′	5′‐attccactgcaacccctgt ‐3′
2	5′‐gagttctccctccctccatc ‐3′	5′‐ctggttcccagcctaccc ‐3′
3	5′‐caggagagcccagagcaa ‐3′	5′‐cctctctcctcctccccttt ‐3′
4	5′‐ggtgtgcatgtggaggaaag ‐3′	5′‐cccatgtcctgagcctgt ‐3′
5	5′‐ctgaataaccccaccccttc ‐3′	5′‐ttggagtctctgggatctgg ‐3′
6	5′‐ccgtccacctcaactttggt ‐3′	5′‐ccagtggaggcagcagttac ‐3′
7	5′‐ccccctatccctgagtcg ‐3′	5′‐ccctcagctccctgcatc ‐3′
8	5′‐ggggaagaggactggaagc ‐3′	5′‐gtctccaccatccaccagat ‐3′
9	5′‐ggttctggtccgggagag ‐3′	5′‐cgaagccggctacaccag ‐3′
10	5′‐caggatgcagtgcctgtg ‐3′	5′‐ggagagaaggggcaaggag ‐3′
11	5′‐gtccacagcctcgcctct ‐3′	5′‐ctcctcgcacctggagtct‐3′
12	5′‐cctgctaagaaagggtgtgc ‐3′	5′‐tcttcaacccaaatcctgct ‐3′
13	5′‐ggagattgggctgatctgaa ‐3′	5′‐gtgccagatgtggggaga ‐3′
14	5′‐ggggttctccccacatct ‐3′	5′‐gagaggccggtccaagac ‐3′
15	5′‐cggcctctcactgaccac ‐3′	5′‐cctcctgtgcccacactg ‐3′
16	5′‐ctgtgaggaggctgaggtg ‐3′	5′‐tccatctgcaagcaagtgtc‐3′
17	5′‐accctcagtcccagttcctt ‐3′	5′‐tgcagacactgaggtccaga ‐3′
18	5′‐tgtccctgtctctgtgtcca ‐3′	5′‐aagttgccctgggaagga ‐3′
19	5′‐aggctggactcaggctgtt ‐3′	5′‐caggatctgatgggagttgg ‐3′
20	5′‐cacgtccccaattcctaatg ‐3′	5′‐ctacagccccctccctgt ‐3′
21	5′‐acagggagggggctgtag ‐3′	5′‐aggccaagtgttgaggatca ‐3′
22	5′‐cacccccagcactgagtc ‐3′	5′‐agagacgctgaggacagtgc ‐3

Table III. Primers for the ATP4B gene listed by exon.

Exon	Forward primer	Reverse primer
1	5′‐ tctggcagaggaaactataaagc ‐3′	5′‐ cctcctcctgagctcacct‐3′
2	5′‐ gtgtgcgttaaaatggctga ‐3′	5′‐ cagcctggcctgtcacac ‐3′
3	5′‐ tgtggtgagcatttggagac ‐3′	5′‐ ctcacctgccgtttcacac ‐3′
4	5′‐ actcgcactcagggcttg ‐3′	5′‐ gccctgactgcactgttgta ‐3′
5	5′‐ cagatctcagctccatgggt‐3′	5′‐ cctgacaagctcctttcctg ‐3′
6	5′‐ aaacgcaacacacagacacc ‐3′	5′‐ ttgctttcctttcgctaagtg‐3′
7	5′‐ gagatttttgaaacacctgtcctt ‐3′	5′‐ caaccaggagggtgtcctt ‐3′

Amplification was performed in a MJ Research thermocycler (Cambridge, MA, USA). Polymerase chain reaction conditions were as follows: 3–10 min at 95°C followed by 35–40 cycles of denaturation step: 40 sec at 95°C; annealing step: 40 sec at temperature specific for each primer (50–65°C); elongation step: 1 min at 68 or 72°C and final extension for 5 min at 72°C terminated the reaction after final annealing. Sequencing was performed using cycle sequencing with Big Dye Terminator kit (version 3) supplied by Applied Biosystems (ABI, Foster City, CA, USA), and reactions were run on an ABI 3730 capillary sequencer according to the manufacturer's instructions. Mutations were first checked manually and reactions were run on an ABI 3730 capillary sequencer according to the manufacturer's instructions. The sequence analyses were performed with Sequencher 4.0.5 software (Gene Codes Corporation, Ann Arbor, MI) and mutations were verified manually. All possible mutations were verified by new PCR and new sequence analysis from both strands.

Results

All patients had severe atrophic corpus gastritis; the antrum was normal in eight patients, and mild histological changes were seen in four. Eleven patients reported vitamin B12 replacement therapy. Altogether seven patients had either a concomitant autoimmune disease or reported atrophic gastritis in their family. All patients had serum PG I levels below normal and elevated H⁺/K⁺‐ATPase antibodies. Three patients had elevated H. pylori antibody titres of the IgG class. Clinical data and results of the blood tests for the individual patients are shown in Table I.

The gene encoding for H⁺/K⁺‐ATPase α‐subunit was sequenced for mutations for 22 exons and flanking splice sites in all 12 AIG patients. No truncating mutations were identified. Instead two polymorphisms were found in the coding area. A non‐synonymous Val265Ala (c.824T>C) was recognized in all 12 patients (11 homozygous, 1 heterozygous) and synonymous Ile503Ile (c.1539C>A) was found in 6 patients, in all of them as heterozygous. In addition, five intronic sequence changes were identified. Upstream of exon 1, at position 1‐84, a nucleotide change A/T was identified in 8 of the 12 patients. In addition in intron 14, at position 2118‐35, one polymorphism A/T in one patient was identified. There were also two nucleotide changes downstream of exon 17 in intron 17 at position 2635+51C>A and at position 2635+84C>T. These SNPs were found in three and in eight patients respectively. An intronic variant at position 2636‐44C>T in intron 17 was identified in three patients.

The gene encoding for the H⁺/K⁺‐ATPase β‐subunit was sequenced for mutations for all seven exons and flanking splice sites in all patients. One recent non‐synonymous SNP and three synonymous ones were found from that subunit. A non‐synonymous C to T transition (c.797C>T) was found in exon 7 at codon 248 resulting in a substitution of an alanine residue for a valine (p.A248V). This amino acid change was identified only in one AIG patient. In addition, a synonymous Arg36Arg (c.A160T) change was found in a homozygous form in exon 1 in all 12 patients. In exon 2 Tyr48Tyr (c.198C>T) was found in three patients in a homozygous and in seven patients in a heterozygous form, and in exon 7 Ala267Ala (c.855G>A) was found in a heterozygous form in four patients.

All exons of α‐ and β‐ subunits in which the SNPs located in coding regions were found, were sequenced in 50–52 blood donors. The identified SNPs are shown in Table IV.

Table IV. Single nucleotide polymorphisms (SNPs) of the coding area of genes encoding for H⁺/K⁺‐ATPase α‐ and β‐subunits found in autoimmune gastritis (AIG) patients and blood donors, and dbSNP id numbers for previously known SNPs.

Subunit/exon	Position	Codon change	Amino acid change	Frequency in patients (n = 12)	Frequency in controls	Controls tested	dbSNP id number
α/ex7	T824C	GTC‐>GCG	Val265Ala	11 homoz,	32 homoz,	52	rs2733743
				1 heteroz	20 heteroz
α/ex 11	C1539A	ATC‐>ATA	Ile503Ile	6 heteroz	16 homoz,	51	rs2230181
					17 heteroz
β/ex 7	C797T	GCG‐>GTG	Ala248Val	1 heteroz	None	50	–
β/ex1	A160C	AGG‐>CGG	Arg36Arg	12 homoz	51 homoz	51	rs9285616
β/ex 2	C198T	TAC‐>TAT	Tyr48Tyr	3 homoz,	23 homoz,	51	rs11164142
				7 heteroz	20 heteroz,
					8 not found
β/ex7	G855A	GCG‐>GCA	Ala267Ala	4 heteroz	11 heteroz,	51	rs17852631
					40 not found

homoz = homozygous; heteroz = heterozygous.

Discussion

Despite the strong evidence of the hereditary component of human AIG, little is known about its genetic background. The present study is to our knowledge the first attempt to search for candidate genes underlying AIG by modern molecular genetic methods. SNPs in autoantigens have been suggested to be a prerequisite for autoimmune diseases 11. However, sequencing the coding area of the genes encoding the α‐ and β‐subunits of the H⁺/K⁺‐ATPase, a well characterized autoantigen in AIG, in 12 Finnish AIG patients and 50 anonymous blood donors revealed neither any truncating mutations nor evidence for disease‐associated SNPs.

No novel SNPs were identified in α‐subunit. In the previously recognized non‐synonymous SNP (T824C) in exon 7 of subunit α, most patients and blood donors had a C allele. Thus, this SNP did not distinguish individuals with AIG in the Finnish population. This T‐>C transition at codon 265 results in substitution of a valine residue for alanine. According to the HapMap database (http://www.hapmap.org/index.html.en), the allele frequency for C is 0.676 and that for T 0.324.

In exon 7 of the β‐subunit, one non‐synonymous SNP leading to the amino acid substitution Ala248Val was identified in a single patient. This SNP has not been described before, nor was it found in any of the blood donors. Found in one case only, this SNP most probably lacks clinical significance.

Before the era of H. pylori, two forms of atrophic gastritis, the corpus‐predominant autoimmune A‐gastritis and pangastric or antrum‐predominant B‐gastritis, were recognized 12. Later it became evident that B‐gastritis is highly associated with H. pylori infection, whereas A‐gastritis is not considered to be caused by this bacterium 13. However, autoimmunity has been proposed to contribute in the development of H. pylori‐associated atrophic gastritis as well 14, 15. During the process of advancing atrophy, the number of H. pylori usually declines and in patients with advanced atrophic gastritis, detection of H. pylori antibodies may be the only means still to reveal the ongoing infection 16. Even H. pylori antibodies may disappear with advancing atrophy 17. In patients with PA, H. pylori is only seldom found histologically 18–21; the serological detection of H. pylori has varied from 0% to 89 % 20–22. Even though none of our patients was aware of a history of H. pylori infection and none was histologically H. pylori positive, three of them had elevated H. pylori antibodies; one of these had had PA for 15 years. It is not possible to determine whether these patients had AIG triggered by H. pylori or two concomitant diseases; in the patient with a 15‐year history of PA and a relatively low H. pylori antibody titre, even a non‐specific cross‐reaction cannot be excluded.

Severe atrophy of fundic glands, with loss of intrinsic factor‐secreting parietal cells, is considered an essential feature of PA. However, this is not an invariable finding in studies concerning PA patients. In two studies comprising altogether 166 PA patients, only fewer than half of the patients were reported to have severe atrophy 18,23. It has also been shown that severe food cobalamin malabsorption may occur in patients with only mild or no corpus atrophy; in those cases H. pylori gastritis is often present 24. Thus it seems that AIG and PA may develop by various mechanisms and possibly with or without H. pylori infection; however, it may be difficult to distinguish these forms of the disease clinically. One study from the era before H. pylori showed that PA patients with history of the disease in the family or a concomitant autoimmune disease had a median age of 52 years at diagnosis, while PA patients without these characteristics were diagnosed significantly later, with a median age of 66 years 25. Classical PA is most often diagnosed at the age of 30–60 years and it is more common among women than in men 13. In order to obtain patients with classical AIG, we excluded those over 65 years; the median age of our patients was 52 years. All except one of our patients were women. In addition to the histologically defined severe corpus‐restricted atrophy, the patients of the present study had low PG I levels and elevated parietal cell antibodies, and 11 of our 12 patients even had vitamin B12 replacement therapy. In addition, four patients had a family history of AIG or PA. Thus, our patients resembled closely those with classical AIG.

As a model for human AIG, an organ‐specific autoimmune disease called experimental AIG has been developed in mice. Experimental AIG is not associated with H. pylori; H. pylori infection has been shown to inhibit the development of experimental AIG in mice 26. Both human and experimental AIG seem to be mediated by CD4 positive H⁺/K⁺‐ATPase‐reactive cytotoxic T cells in the gastric mucosa 27. Peptides that stimulate the proliferation of T cells in AIG are found in both H⁺/K⁺‐ATPase subunits. In H. pylori‐negative AIG patients the T cell epitopes of the H⁺/K⁺‐ATPase overlap with those in the experimental AIG, which proposes a similar pathogenic mechanism for human and murine diseases 28. Instead, in H. pylori‐positive patients with AIG, peptide epitopes of the H⁺/K⁺‐ATPase which stimulate CD4 cell growth and cross‐react with H. pylori proteins are found 29. The parts of the peptide chain coded by the two non‐synonymous SNPs found in our patients are not located in any of the known epitopes shared by human and experimental AIG. Instead, the area coded by the SNP T824C resulting in substitution of amino residue Val265Ala is included in one (α 256‐270) of the several T cell epitopes of the α‐subunit found in H. pylori‐positive AIG patients; this particular epitope showed cross‐reaction with H. pylori VIRB4 homologue 29. Thus it seems that this polymorphism might have, if any, an impact on H. pylori‐induced AIG but not on classical AIG.

The blood donors were anonymous and thus could not be matched with the patients. The possibility that some individuals among blood donors carry genetic susceptibility for AIG cannot be excluded; nor is a stable B12 vitamin replacement therapy a contraindication for donating blood in Finland. Because of this and the relatively small number of patients, all possible associations between polymorphisms and AIG may not be detected in this study. However, our findings seem to exclude an essential role of changes in the exons of genes for H⁺/K⁺‐ATPase α‐ and β‐subunits in the development of AIG.

In conclusion, when the exons of the genes encoding H⁺/K⁺‐ATPase α‐ and β‐subunits were sequenced for 12 Finnish patients with AIG, no truncating nor missense mutations nor disease associated SNPs were identified. At the non‐synonymous SNP c.824T>C in the α‐subunit of H⁺/K⁺‐ATPase most Finns seem to have C allele.

Acknowledgements

The study was partially supported by grants from Helsinki University's Research Funds, and Helsinki University Central Hospital Research Funds.