Article Text

Download PDFPDF

Long-term proton pump inhibitor administration worsens atrophic corpus gastritis and promotes adenocarcinoma development in Mongolian gerbils infected with Helicobacter pylori
  1. Tadashi Hagiwara,
  2. Ken-ichi Mukaisho,
  3. Takahisa Nakayama,
  4. Hiroyuki Sugihara,
  5. Takanori Hattori
  1. Department of Pathology, Shiga University of Medical Science, Shiga, Japan
  1. Correspondence to Dr K Mukaisho, Department of Pathology, Shiga University of Medical Science, Seta-tsukinowa-cho, Otsu, Shiga 520-2192, Japan; mukaisho{at}belle.shiga-med.ac.jp

Abstract

Background We investigated whether corpus atrophic gastritis worsens in Mongolian gerbils (MGs) after long-term administration of proton pump inhibitor (PPI). MGs are an excellent model for studying Helicobacter pylori-related gastritis and adenocarcinoma.

Methods MGs were separated into four groups (n =15/group); H pylori (ATCC43504) was inoculated into the OPZ(omeprazole)+Hp (H pylori) and Hp groups, a PPI (OPZ) was administered to the OPZ+Hp and OPZ groups and the control group received no treatment. MGs had access to food containing omeprazole (100 mg/kg body weight/day) for 6 months, after which their stomachs were removed and cut into nine sections (six sections in the fundus and three sections in the antrum). Corpus atrophy was evaluated by the absence of parietal cells in the six sections in the fundus. First, we calculated a percentage of the area devoid of parietal cells in each haematoxylin and eosin-stained section, and then we scored the degree of atrophy by adding the percentages of the six sections. A full score was 600.

Results Neutrophilic and lymphoid infiltrates were greater in the OPZ+Hp group than in the other groups. The corpus atrophy score in the OPZ+Hp group was significantly higher than that in the Hp group (p<0.0048, Student t test). Significantly more adenocarcinomas were found in the OPZ+Hp (60%) than in the Hp (7%) group animals.

Conclusion Long-term PPI administration promotes development of adenocarcinoma, which is associated with the progression of atrophic corpus gastritis in MGs infected with H pylori.

  • PPI
  • adenocarcinoma
  • Helicobacter pylori, Mongolian gerbils
  • long-term
  • gastrin
  • gastritis
  • proton pump inhibition
View Full Text

Statistics from Altmetric.com

Significance of this study

What is already known about this subject?

  • It has been reported that long-term proton pump inhibitor (PPI) administration worsens atrophic corpus gastritis in patients infected with Helicobacter pylori.

  • Infection with H pylori is strongly associated with an increased risk of gastric carcinoma.

  • Long-term PPI administration leads to hypergastrinaemia, which can ultimately lead to gastric carcinoid formation.

  • Mongolian gerbils (MGs) are an excellent model for studying H pylori-related gastritis and adenocarcinoma.

What are the new findings?

  • Long-term PPI administration promotes development of adenocarcinoma in MGs infected with H pylori.

  • Long-term PPI administration worsens atrophic corpus gastritis in MGs infected with H pylori.

  • It is suggested that the direct cause of corpus atrophy leading to adenocarcinoma development was the H pylori infection rather than long-term PPI administration.

  • The risk of corpus atrophy leading to gastric cancer development under long-term PPI administration in patients infected with H pylori might depend on the H pylori strains and/or races.

How might it impact on clinical practice in the foreseeable future?

  • It is recommended that patients being considered for long-term PPI therapy should be tested for H pylori infection, and if present, this pathogen should be eradicated, especially in Asians infected with the cagA positive-H pylori strain.

Introduction

The use of proton pump inhibitors (PPIs) for treating peptic ulcers and gastro-oesophageal reflux disease (GORD) has increased substantially over the past two decades. Furthermore, the widespread use of non-steroidal anti-inflammatory drugs in populations at a risk for gastroduodenal side effects is another common indication for PPI therapy.1 2 PPIs are very efficacious but several adverse effects, such as oxyntic cell hyperplasia, glandular cysts, hypergastrinaemia and fundic gland polyps, have been reported.3–5 However, the effect of long-term PPI administration on gastric tumour development remains largely unknown, except for a few reports on carcinoid development in female rats6 and humans.7 It has been widely accepted that gastric adenocarcinoma is associated with chronic atrophic gastritis caused by the presence of Helicobacter pylori.8 9 Also, long-term use of PPIs has been reported to worsen corpus atrophic gastritis in the patients with H pylori infection.10–12 Thus, a potential association exists between gastric neoplasia and the long-term use of PPIs, which has been a topic of debate. Further studies are required to identify the risk for gastric adenocarcinoma development under long-term PPI administration. However, available cohort studies of chronic PPI users and observational data from large population databases do not provide gastric adenocarcinoma data as an end point.13 Therefore, other methods must be considered to identify the association between PPIs and gastric neoplasia.

Animal models are good tools for understanding the pathogenesis of human diseases. However, only a few experimental studies have administered PPIs to laboratory animals for more than 1 month and the method of administration was daily per os or through subcutaneous injection,14 15 but these methods are unsuitable for long-term administration. We recently established an animal model in which long-term administration of PPIs was possible using food containing PPIs and reported the side effects on the long-term use in rats.16 However, rats are not a good species for studying gastritis and adenocarcinoma related to H pylori infection, whereas Mongolian gerbils (MGs) are an excellent model, because the gross and histological findings mimic lesions induced by H pylori infection in the human gastric mucosa.17 18

The purpose of the present study was to determine whether corpus atrophic gastritis leading to gastric adenocarcinoma in MGs infected with H pylori worsens after long-term administration of PPIs.

Materials and methods

All procedures complied with the ethical guidelines for animal experimentation and the care and use of laboratory animals at Shiga University of Medical Science, Japan.

Animals

Totally sixty male MGs were used in this experiments. Thirty MGs were infected with H pylori (ATCC43504: both cagA and vacA were positive) (MGS/Sea; Kudo, Inc., Kumamoto, Japan), and 30 were used without H pylori infection. The inoculation dose was 3.2×108 CFU/ml and 0.5 ml/head. The animals were housed in an air-conditioned biohazard room designed for infectious animals with a 12-h light/12-h dark cycle. They were provided rodent diet (CE-2, CLEA, Osaka) and water ad libitum.

PPI administration

The MGs were separated into four groups (n =15/group). One month after birth, MGs from groups OPZ(omeprazole)+Hp and Hp were inoculated with H pylori. OPZ was administered to groups OPZ+Hp and OPZ, and the control group received no treatment. MGs had access to food containing OPZ (100 mg/kg body weight/day) for 6 months, according to a recently established protocol.16 Pellet diet containing 0.1125% OPZ was provided 6 months. OPZ was administered beginning 6 months after infection, when pangastritis started to develop in the H pylori-infected MGs (figure 1). MGs were sacrificed using an overdose of diethyl ether after the OPZ treatment. They were deprived of food but allowed free access to tap water 18 h before sacrifice. Immediately after sacrifice, blood was drawn from the heart for analyses of serum gastrin levels, and the stomach was quickly removed.

Figure 1

Experimental design. Mongolian gerbils were separated into four groups (n =15/group), and Helicobacter pylori was inoculated into the OPZ+Hp and Hp groups 1-month after birth (red arrow). Omeprazole (OPZ) was administered to the OPZ+Hp and OPZ groups, starting 6 months after infection (yellow arrow). The control group received no treatment.

Histological examination

The stomach was fixed with 10% formalin in phosphate-buffered saline for 4 h and was cut into nine sections (six sections in the fundus and three sections in the antrum). The tissues were embedded in paraffin and cut into 4 μm sections. The sections were stained with haematoxylin&eosin (H&E) to examine stomach morphology.

Evaluation of atrophy

A few pyloric-type glands were present in the corpus of the normal MG stomachs immediately distal to the squamous portion. These glands might have been miscounted as corpus atrophic glands; however, there were only a few glands in normal MG stomachs, so the possibility that corpus atrophy scoring was altered was minimal. We ignored the pyloric-type glands in the corpus immediately distal to the squamous portion and evaluated corpus atrophy by the total absence of parietal cells in the six sections in the fundus. This definition was simple but very strict. We first calculated a percentage of the area devoid of parietal cells in each H& E-stained section, and then calculated the sum of the percentages of the six sections. A score of 600 was the full score in one animal (figure 2).

Figure 2

Corpus atrophy scores. Each bold line shows the area of atrophy. Corpus atrophy was evaluated by the absence of parietal cells. We calculated the percentage of the area devoid of parietal cells in each haematoxylin and eosin-stained section (A). The figure shows samples of (i) 20%, (ii) 80% and (iii) 100% atrophy in the fundic gland section of an animal inoculated with Helicobacter pylori and administered omeprazole 6 months later (OPZ+Hp) (B).

Definition of adenocarcinoma distinguished from the heterotopic gland and intestinal metaplasia that developed in the submucosa

Franco et al reported dysplastic foci and adenocarcinoma in a model using MGs infected with the in vivo-adapted H pylori strain 7.13.19 They diagnosed dysplasia and adenocarcinoma using morphological criteria established previously for gastrointestinal neoplasia in diseased mouse models.20 We therefore also diagnosed adenocarcinoma according to these studies.19 20 The atypical glands, called dysplastic foci in the study by Franco (evident in Franco's photomicrograph), were detected frequently within the mucosa of our model in the Hp and OPZ+Hp groups. However, the lesion was difficult to distinguish from regenerative atypia due to severe active inflammation. We therefore counted only the heterotopic gland lesions and obvious adenocarcinoma in the present study. We defined submucosal dilated glands without nuclear atypia as heterotopic glands. Adenocarcinoma was characterised by marked cellular pleomorphism, cellular atypia and euchromatic nuclei that exhibited bizarre morphological features.20 In a study using MGs infected with H pylori, the histological type of adenocarcinoma was mucinous adenocarcinoma17 and well-differentiated tubular adenocarcinoma had also developed.20–22 The adenocarcinoma must accompany invasive growth and not glandular herniation. The criteria to discriminate between glandular herniation and true invasion were also determined as described previously.19

Immunohistochemical analyses for BrdU

5′-Bromo-2′-deoxyuridine (BrdU) at a dose of 100 mg/kg was injected intraperitoneally into all MGs at 60 min before sacrifice. Immunohistochemical staining was conducted using BrdU. For immunohistochemistry, the streptavidin–biotin method was performed using a Histofine kit (Nichirei, Tokyo, Japan). Peroxidase binding sites were visualised by the diaminobenzidine method, and the sections were lightly counterstained with haematoxylin. The numbers of BrdU-labelled cells in two glands of adenocarcinoma and the heterotopic gland in the Hp and OPZ+Hp groups were counted microscopically, and the mean percentages of positive cells among the total cells in the two glands of adenocarcinoma and heterotopic gland were determined.

Serum gastrin levels

The serum gastrin levels were measured by radioimmunoassay (Medic, Yasu, Japan).

Statistical analysis

The adenocarcinoma and heterotopic gland indices were compared by Welch's t test. The corpus atrophy score was expressed as mean±SE, and the groups were compared by the Student t test and Fisher's exact test. p<0.05 was considered statistically significant.

Results

Body weight

The body weights in the OPZ+Hp, Hp, OPZ and control groups were 101.3±10.6, 97.0±11.5, 104.3±5.3 and 120.0±9.9 g, respectively (table 1). No significant differences in weight were observed among the groups.

Table 1

Body weight, serum gastrin level and corpus atrophy score. No significant differences in body weight were noted among the groups. The serum gastrin level in the OPZ+Hp group was significantly higher than that in the OPZ and control groups (p <0.008). The gastrin level in the OPZ+Hp group was slightly higher than that of the Hp group (p <0.0877). The corpus atrophy score in the OPZ+Hp group was significantly higher than that in the Hp group (p <0.0008, Student t test)

Serum gastrin level

The gastrin level in the OPZ+Hp, Hp, OPZ and control groups were 349.3±181.3, 213.0±122.1, 80.1±13.5 and 97.3±24.6 pg/ml, respectively (table 1). The gastrin level in the OPZ+Hp group was significantly higher than that in the OPZ and control (p<0.008) groups, and the gastrin level in the OPZ+Hp group tended to be higher than that in the Hp group (p<0.0877, Student t test; table 1).

Macroscopic findings

No macroscopic abnormalities were found in the control group. The stomach wall was mildly thickened in the OPZ group compared to that in the control group. Ruggedness of the superficial mucosa in the border lesion between the corpus and the antrum was detected in the Hp group. Polypoid lesions and a marked reduction in the normal corpus area were observed in the OPZ+Hp group (figure 3).

Figure 3

Macroscopic findings. Polypoid lesions and a marked reduction in the normal corpus area were observed in the OPZ+Hp group (A). The polypoid lesions were milder in the Hp group than in OPZ+Hp group (B). A mild thickening of the stomach wall was observed in the OPZ group (C) compared to that in the control group (D). The OPZ+Hp and Hp groups were inoculated with Helicobacter pylori (Hp), and the OPZ and OPZ+Hp groups were administered the proton pump inhibitor omeprazole (OPZ) 6 months later. The control group received no treatment.

Microscopic findings

Inflammatory cell infiltrates

We evaluated the histological features in the centre of the fundus. The mucosa was thicker in the OPZ than in the control group, and infiltration of neutrophils and lymphocytes into the mucosa was detected in the OPZ+Hp and Hp groups. Corpus atrophy included replacement of the oxyntic cells by mucous-type cells in the OPZ+Hp and Hp groups. Marked foveolar hyperplasia was also noted in these atrophic glands. Macroscopical extensive polypoid formations were caused by foveolar hyperplastic changes, cystic dilated glands and severe infiltration of neutrophils and lymphocytes. The degree of glandular atrophy, defined as the total loss of parietal cells and of neutrophilic and lymphocytic infiltration, was much higher in the OPZ+Hp group than in the control group (figure 4).

Figure 4

Inflammatory cell infiltrate. We evaluated the histological features in the same area of the fundus centre. Group OPZ+Hp animals had corpus atrophy and marked neutrophilic and lymphocytic infiltrates (A). Small numbers of neutrophils and lymphocyte infiltrates were observed in the mucosa of the Hp group (B). Mild hyperplasia of the superficial mucosa was observed in the OPZ group (C). No pathological findings were observed in the control group (D).

Incidence of intestinal metaplasia, heterotopic glands and adenocarcinoma

Based on the analyses of the sections, we counted the animals with intestinal metaplasia, heterotopic glands and adenocarcinoma (table 2). Intestinal metaplasia was detected in the OPZ+Hp (14 of 15) and Hp (13 of 15) groups, but not in the OPZ or control groups (table 2). Heterotopic glands were identified in the submucosal layer from the OPZ+Hp (15 of 15), Hp (14 of 15) and OPZ (1 of 15) groups but not in the control group (figure 5A–C table 2). Adenocarcinomas were detected in the OPZ+Hp (9 of 15) and Hp (1 of 15) groups, but not in the OPZ or control groups. The adenocarcinomas detected were exclusively of the mucinous type; detached mucous columnar epithelia floating in mucin lakes showed structural and nuclear atypism and several mitotic cells (figure 5D–I). The incidence of adenocarcinomas was greater in the OPZ+Hp group than in the Hp group (p<0.0052, Fisher's exact test) (table 2).

Table 2

Cases of intestinal metaplasia, heterotopic glands and adenocarcinoma. Significantly more adenocarcinomas were detected in the OPZ+Hp group than in the Hp group

Figure 5

Comparison between heterotopic glands and adenocarcinomas. (A–C) Heterotopic glands, (D–I) adenocarcinomas, (B, C) Higher magnification of figure 5A, (E, F) Higher magnification of figure 5D, (I) Higher magnification of figure 5H. Heterotopic glands were found in the submucosal area in the OPZ+Hp, Hp and OPZ groups. These glands had no nuclear pleomorphism or structural atypia (A, B, C). The gland was covered with a muscular mucosae component, and a desmoplastic response was absent (B, C). In contrast, adenocarcinomas were detected in the OPZ+Hp (9 of 15) and Hp (1 of 15) groups. The adenocarcinoma revealed nuclear pleomorphism and several mitotic cells (D–I). Obvious desmoplastic reaction was observed around the atypical glands (D, E, H, I). Many atypical glands were detected in the submucosal layer (D–I).

BrdU labelling index (%)

The BrdU labelling index of the adenocarcinoma (21.2±2.06) was significantly higher than that of the heterotopic gland (1.21±0.26) (p <0.0001) (figure 6).

Figure 6

Immunohistochemical stainings of BrdU in adenocarcinomatous and heterotopic glands. (A) Heterotopic glands, (B) adenocarcinomas. Each photo of these glands was taken from the OPZ+Hp group. The BrdU adenocarcinoma labelling indices (21.2±2.06) (A) were significantly higher than those of the heterotopic gland (1.21±0.26) (B) (p <0.0001).

Corpus atrophy scores

More severe neutrophilic and lymphoid infiltrates were detected microscopically in the OPZ+Hp group than in the other groups (figure 4). No glandular atrophy was noted in the OPZ and control groups. The corpus atrophy score in the OPZ+Hp group was significantly higher than that in the Hp group (p<0.0008, Student t test) (table 1).

Discussion

Kuipers et al11 hypothesised that inflammation of the corpus worsens because stomach pH rises in patients infected with H pylori who are undergoing long-term PPI therapy. This may cause an extension of H pylori infection from the pylorus to the corpus. Furthermore, interleukin (IL) 1β and tumour necrosis factor (TNF) α are proinflammatory cytokines with potent acid-suppressive properties, and highly expressed polymorphisms within the human IL-1β and TNFα gene promoters heighten the risk for gastric adenocarcinoma among subjects infected with H pylori.23 24 However, a mechanism for this change remains unclear. In the present study, we confirmed Kuipers' hypothesis using a well-characterised MG model of H pylori-associated gastritis and gastric adenocarcinoma; long-term use of PPIs worsens corpus atrophic gastritis in patients with H pylori infection.11 Furthermore, the results also suggest that long-term PPI administration promotes adenocarcinoma development in the corpus. However, this study has two main limitations. One concerns the strains of MGs and H pylori, and the other concerns the PPI dose.

Long-term H pylori infection in MGs can eventually progress to gastric adenocarcinoma25 26; therefore, this prolonged time course precludes large-scale analyses to evaluate the effects of both pathogen and host in a carcinogenic cascade. H pylori-induced cancer in this model has never been reported outside Japan or China.22 Watanabe et al17 used the TN2GF4 strain and the other groups used the ATCC43504 strain. These strains are positive for both cagA and vacA. One strain-specific H pylori constituent that augments cancer risk is the cag pathogenicity island,9 a genetic locus that encodes a type IV secretion system. Upon delivery into host cells by the cag secretion system, the terminal gene product in the island, CagA, undergoes Src-dependent tyrosine phosphorylation and activates a eukaryotic phosphatase (SHP-2), leading to dephosphorylation of host cell proteins and cellular morphological changes.27–29 Translocation, but not phosphorylation, of CagA also disrupts the apical–junctional complexes, resulting in a loss of cellular polarity.30 Therefore, the cagA-positive strain is likely critical for gastric cancer development. However, in the previous study, adenocarcinoma did not develop among any of the infected MGs using the cag+vacA s1a human H pylori gastric ulcer strain B128.19 Thus, the cause of cancer development may depend on cagA as well as on host factors that influence gastric carcinogenesis. Moreover, we also considered the MG strains. In the literature, it has been reported that the MG strain in Europe and the US were separated from those in Japan in 1954.31 The MG strains in Asian countries may be different from those in Europe and the US. From these findings, even if a patient is infected with H pylori, the risk of corpus atrophy leading to gastric cancer development under long-term PPI administration might depend on the H pylori strains and/or races.

The other potential problem concerns the OPZ dose. In the literature, the method of OPZ administration is daily per os or subcutaneous injection until we established a method for long-term PPI administration using OPZ-containing food.14 15 The OPZ dose in the present study was 100 times more than the usual therapeutic dose used in humans. However, it is unclear how much of the dose is inactivated during the process of incorporating OPZ in food. A dose of 100 mg/kg/day was determined from the results of a previous preliminary study using a rat model.16 The MGs were healthy in the OPZ group, and neither glandular atrophy nor tumour development was detected in their stomachs. These findings suggest that the dose was appropriate for long-term PPI administration and that the cause of gastric atrophy leading to cancer development is not a large OPZ dose.

We understand the difficulty of diagnosing adenocarcinoma in animal models. Elfvin et al25 stated that adenocarcinoma was not found in any MGs when they examined sequential morphological changes in the stomach for up to 62 weeks using MGs infected with TN2GF4 and SS1. They suggested that the glands were buried in the submucosal layer, that these changes might be misinterpreted as adenocarcinoma and that glands in the muscularis propria were insufficient to diagnose gastric adenocarcinoma, because the deranged glandular structures can grow in and below the submucosa. Therefore, we diagnosed adenocarcinoma and distinguished it from heterotopic glands. Adenocarcinomas found in the present study were mostly of the mucinous type; mucin lakes contained detached mucous columnar epithelia with nuclear and structural atypia and several mitotic cells. Such a mucinous lesion, quite different from those of the heterotopic glands and intestinal metaplasia developed in submucosa, were seldom detected in previous studies after a 1-year infection. We also found that the BrdU labelling indices of mucinous and tubular adenocarcinoma were significantly higher than those of the heterotopic glands. Adenocarcinomatous glands have a much higher proliferative capacity. These findings suggest that adenocarcinoma diagnosed in the present study was completely different from both heterotopic glands and intestinal metaplasia developed in the submucosa.

Profound acid suppression therapy leads to hypergastrinaemia in nearly all patients receiving long-term administration of PPIs.26 Furthermore, prolonged hypergastrinaemia leads to hyperplasia of enterochromaffin-like cells, which can ultimately lead to gastric carcinoid formation.32 In humans, diffuse, linear or micronodular hyperplasia of enterochromaffin-like cells is observed in 10–30% of chronic PPI users, particularly in H pylori-infected patients with markedly increased gastrin levels.33 Dysplasia or invasive carcinoid formation has not been described in long-term PPI users; thus, it is not an indication for surveillance in PPI maintenance users. Insulin–gastrin transgenic mice, which show hypergastrinaemia with H felis or H pylori SS1 infection, progress to gastric cancer.34 35 In the present study, the serum gastrin levels in the OPZ+Hp group, which was administered OPZ and infected with H pylori, were higher than those of the other groups, and adenocarcinomas were significantly more common in the OPZ+Hp group than in the other groups. These findings suggest that hypergastrinaemia might promote the development of gastric cancer.

One of the most important findings in the present study was that long-term OPZ administration without H pylori infection did not induce corpus atrophy. No neoplastic lesions developed in a previous study using male rats administered OPZ for 1 year.36 These findings suggest that the direct cause of corpus atrophy leading to the development of adenocarcinoma was the H pylori infection rather than long-term PPI administration. H pylori eradication induces regression of gastric mucosal inflammation and atrophy in chronic PPI users without affecting the efficacy of PPI therapy for GORD.37 Therefore, the advice of the Maastricht consensus panel was to eradicate H pylori in infected subjects requiring long-term maintenance treatment with a PPI.38 H pylori eradication makes long-term PPI treatment a relatively safe therapy for patients with peptic acid disorders and GORD.

Taken together, although the present study had several limitations, we were able to show that long-term PPI administration worsens atrophic corpus gastritis and promotes the development of adenocarcinoma in MGs infected with H pylori. The main cause of corpus atrophy was not the PPI but H pylori infection. H pylori might move to the oral side under decreased gastric acid conditions, induce corpus atrophic gastritis and promote adenocarcinoma development.13 Hypergastrinaemia may also be associated with gastric cancer development. Thus, it is recommended that patients being considered for long-term PPI therapy should be tested for H pylori infection, and if present, this pathogen should be eradicated, especially in Asians infected with the cagA positive H pylori strain.

References

View Abstract

Footnotes

  • See Commentary, p 567

  • Linked articles 229286.

  • Competing interests None.

  • Ethics approval All procedures complied with the ethical guidelines for animal experimentation and the care and use of laboratory animals at Shiga University of Medical Science, Japan.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles