Cell Proliferation and Inflammation on Biopsy Samples With Multifocal Atrophic Gastritis Before and 1 Year After Helicobacter pylori Eradication

Helicobacter pylori infection has been associated with the presence of atrophy, intestinal metaplasia, dysplasia, and intestinal-type adenocarcinoma of the stomach.1–5 It has been postulated that persistent inflammation produced by the bacteria for several decades gives rise to atrophy and intestinal metaplasia in some patients. In a small group of susceptible individuals, this is later followed by dysplasia and eventually cancer.1,2,4 ,H pylori is now considered a type 1 carcinogen, and it is predicted that the absence of infection would result in at least 60% fewer cases of gastric cancer worldwide.3 Randomized clinical trials that evaluate gastric cancer prevention after the eradication of H pylori infection require large numbers of subjects to attain statistical significance and prolonged follow-up time to reach the cancer end point. Thus, histopathologic markers such as atrophy and intestinal metaplasia have been used as end points that reflect possible preneoplastic lesions. A study in Colombia showed a decrease in atrophy and intestinal metaplasia in biopsy samples 6 years after H pylori eradication.6 Two studies in China with 1 and 5 years of follow-up showed less definitive histopathologic changes after H pylori treatment.7,8 A recently completed Chinese study with 7.5 years of follow-up showed that, although the incidence of gastric cancer development was similar for patients receiving placebo or H pylori eradication treatment, patients without preneoplastic conditions had significantly less gastric cancer.9 

A randomized clinical trial to determine if H pylori eradication therapy resulted in histopathologic improvement of gastric lesions was performed in Chiapas, the southern-most state in Mexico, where there is a high rate of H pylori gastritis, atrophy, intestinal metaplasia, and gastric cancer.10–12 During the trial, 316 subjects were followed up for 1 year with serial endoscopies. The endoscopically obtained biopsy specimens were characterized by using the histopathologic portion of the Sydney classification.13 At 1 year, H pylori had been eradicated from 76% of patients who had partial remission of gastritis, atrophy, and intestinal metaplasia.12,14 The question remains whether clinically important lesions (ie, those that would have become malignant) or only the lesions that would have remained clinically silent go into remission. A better understanding of molecular events is necessary to address this question. If cell proliferation and inflammatory markers improve together with histopathologic lesions, the results of these clinical trials would be significantly more compelling. A pilot study using selected gastric biopsy specimens obtained from the Chiapas cohort was used with the following objectives: (1) define if immunohistochemical (IHC) assays for currently commercially available cell proliferation and type of inflammatory cell markers can be evaluated in biopsy specimens, (2) determine if different IHC staining patterns can be observed in subjects with gastritis alone compared with those with gastritis and multiple foci of atrophy and intestinal metaplasia, and (3) determine if different IHC staining patterns can be observed after H pylori eradication.

Two biopsy samples (one before treatment and another 1 year after treatment) from 12 subjects with gastritis alone and 16 with gastritis and multiple foci of atrophy and intestinal metaplasia were selected for IHC cell proliferation (tumor suppressor protein p53, proliferation marker Ki-67, cell cycle regulator cyclin D1), surface inflammatory cell marker (T and B lymphocytes and macrophages) assays, and TUNEL (terminal deoxynucleotide transferase deoxyuridine triphosphate [dUTP] nick end labeling) assay for apoptosis. The biopsy samples were selected from specimens obtained during a previously described randomized, double-blind, placebo-controlled trial that looked at the effect of H pylori eradication on gastric histopathologic lesions.12,14–18 All subjects enrolled in the double-blind, placebo-controlled trial had serologic (antibodies against Cag A), histopathologic, or culture evidence of H pylori infection and serum gastrin levels of 25 ng/ L or higher.12 All subjects were asymptomatic volunteers older than 40 years who lived in the mountainous region of Chiapas, Mexico, and had consented to undergo 1 endoscopy before starting H pylori treatment and another 1 year after. Separately, 2 pathologists (J.G., R.H.-G.) examined 7 biopsy samples of different areas of the stomach (4 from the antrum and 3 from the corpus) obtained during each endoscopy and graded the amount of H pylori, acute and chronic gastritis, atrophy, and intestinal metaplasia by using the visual analog scale of the Sydney classification.13 A final diagnosis for each endoscopy integrated the diagnoses by both pathologists for each biopsy specimen. Subjects with gastritis were defined as having acute or chronic inflammation and no atrophy or intestinal metaplasia in any of the 7 biopsy samples. Subjects with multifocal atrophic gastritis were defined as having inflammation plus evidence of atrophy and intestinal metaplasia in 2 or more of the 7 gastric samples obtained during the endoscopy. The samples used for cell proliferation and inflammatory cell markers were selected if (1) the subject had successfully completed treatment for H pylori (not placebo) during the randomized, double-blind, placebo-controlled trial and (2) the 2 time point samples were adequately oriented and showed the disease of interest (gastritis, atrophy, and intestinal metaplasia).

We placed 4-μm sections of paraffin-embedded tissues on Fisher Plus slides (Fisher Scientific Co., Pittsburgh, Pa). Sections were deparaffinized, rehydrated, and pretreated depending on the primary antibody to be applied (Table 1). After being washed, slides were allowed to incubate at room temperature for 60 minutes with the primary antibodies (Table 1). Slides then underwent sequential application of biotinylated swine anti-mouse antibody, avidin-alkaline phosphatase complex (DakoCytomation), and Vector red substrate kit (Vector Laboratories, Burlingame, Calif). Sections were counterstained by use of an aqueous-based methylene blue and mounted with resin permanent mounting medium. Positive controls were prostate and breast cancer samples and lymph nodes incubated with the antibodies. Negative controls were tissue sections from each patient's sample incubated with normal mouse ascitic fluid.

The TUNEL technique detects endonucleolytically cleaved DNA by the addition of labeled dUTP to DNA ends by terminal transferase. Tissue sections were deparaffinized, rehydrated, and permeabilized by incubation in 10 μg/mL of proteinase K (Roche Molecular Biochemicals) for 15 minutes at room temperature. After being washed twice for 5 minutes in phosphate-buffered saline, the tissues were covered with 40 μL of labeling mix (25mM Tris hydrochloride, pH 7.2, 0.2M potassium cacodylate, 5mM cobalt chloride, 30 U of terminal deoxynucleotidyl transferase, 0.6 nmol of digoxigenin-11-dUTP, 0.15mM deoxyadenosine triphosphate) and incubated in a humid chamber at 37°C for 60 minutes. The reaction was terminated by incubation in 2× SSC (1× SSC is 0.15M NaCl plus 0.015M sodium citrate) for 15 minutes at room temperature. Sections were counterstained with Mayer hematoxylin. The TUNEL assay controls included the omission of terminal transferase (negative controls) and preincubation in 0.1 μg/mL of deoxyribonuclease for 1 hour at 23°C (positive control).

The IHC and TUNEL assays were interpreted blind to time of endoscopy (initial vs 1-year posttreatment samples). For cell proliferation markers, results recorded included (1) the location of the cells that showed the marker (surface epithelium including foveola and glands) and (2) a semiquantitative distribution pattern of stained cells classified as normal (no cells staining or scattered, individual cells staining in the biopsy sample) and abnormal (groups of cells staining). For inflammatory cell markers and apoptosis, results recorded included (1) the location of the cells that showed the marker (same as for cell proliferation, lymphoid aggregates, and lamina propria) and (2) a semiquantitative distribution pattern of stained cells classified as normal (scattered, individual cells staining in the biopsy) and abundant or abnormal (groups of cells staining in the biopsy).

We used the Fisher exact test to compare initial to 1-year posttreatment samples and patients with gastritis to those with multifocal atrophic gastritis. The McNemar test, a nonparametric test for paired data (SAS statistical software, version 8.0 for Windows, SAS Inc, Cary, NC), was used to compare the marker results of the same subject before and after treatment.

Pretreatment biopsy samples of the 28 subjects studied included 12 subjects with gastritis (6 with chronic active inflammatory infiltrate and 6 with chronic inflammation) and 16 with multifocal atrophic gastritis (7 with active inflammatory infiltrate and 2 or more sites with atrophy but no intestinal metaplasia and 9 with active inflammatory infiltrate and 2 or more sites with atrophy and intestinal metaplasia). The IHC and TUNEL assays were performed in paired samples (one biopsy sample obtained before starting H pylori treatment and another 1 year after treatment) that showed the previously mentioned histopathologic findings and included 31 biopsy specimens from the antrum and 25 from the corpus.

In Table 2, we list the number of biopsy samples with abnormal distribution of cell proliferation and inflammatory cell markers before and 1 year after treatment. Of note, 1 year after treatment there was a statistically significant decrease in apoptosis of inflammatory cells (P = .001); however, groups of T lymphocytes in the lamina propria increased (P = .03) 1 year after treatment. Table 3 shows the number of biopsy samples with abnormal distribution of cell proliferation and inflammatory cell markers stratified according to histopathologic evidence of gastritis or multifocal atrophic gastritis regardless of when the sample was obtained. Significantly more samples from persons with multifocal atrophic gastritis had abnormal distribution of p53 (P = .02) and Ki-67 (P = .009) in the mucosa compared with those from subjects with gastritis. In addition, biopsy samples from persons with multifocal atrophic gastritis had significantly more apoptosis in inflammatory cells (P = .04) than subjects with gastritis.

Table 4 presents the data with respect to time of biopsy and group after applying the McNemar test. The number of biopsy samples with abnormal distribution of Ki-67 in the glands increased 1 year after treatment for subjects with multifocal atrophic gastritis (P = .02). Groups of T lymphocytes in the lamina propria of 1-year posttreatment samples from subjects with multifocal atrophic gastritis increased significantly (P = .008). Lastly, biopsy samples from persons with multifocal atrophic gastritis had significantly less apoptosis in inflammatory cells (P = .02) 1 year after treatment.

Independent of location or time when the biopsy specimen was obtained, 17 subjects had abnormal distribution of p53: 12 (71%) had multifocal atrophic gastritis and 5 (29%) had gastritis alone. There were 23 subjects with abnormal distribution of Ki-67, 15 (65%) had multifocal atrophic gastritis and 8 (35%) had gastritis. Eleven subjects had one or more samples that showed abnormal distribution of both markers: 9 (82%) had multifocal atrophic gastritis and 2 (18%) had gastritis. Expression of cyclin D1 was observed only in 2 subjects with multifocal atrophic gastritis, one on the initial biopsy sample and the other on the 1-year posttreatment sample; in both subjects, other cell proliferation markers were also abnormal. The changes observed in subjects with abnormal distribution of cell proliferation markers during the study period stratified for gastritis or multifocal atrophic gastritis are presented in Table 5. Of the 17 subjects with abnormal distribution of p53, in 11 (65%) the distribution changed from abnormal to normal, in 2 (12%) the distribution remained abnormal, and in 4 (23%) the distribution went from normal to abnormal. Of the 23 subjects with abnormal distribution of Ki-67, in 12 subjects (52%) the distribution went from abnormal to normal: in 5 (22%) it remained abnormal and in 6 (26%) it went from normal to abnormal.

In this pilot study, abnormal distribution of p53 and Ki-67 was observed more often in subjects with multiple foci of atrophy and intestinal metaplasia compared with those with gastritis alone. Of the biomarkers, Ki-67 was most frequently abnormally distributed in surface epithelium, followed by p53 and cyclin D1. These differences in expression may suggest a sequential activation in the cell cycle. Ki-67 is an excellent indicator of proliferation, because it is expressed during the entire cell cycle with the exception of the G0 and early G1 phases.19 Ki-67 was expressed frequently in subjects with multifocal atrophic gastritis 1 year after treatment, suggesting that the continued cell proliferation gives more opportunities for genetic transformations to occur and eventually could lead to histopathologic events, such as atrophy, metaplasia, or dysplasia. One such transformation may include abnormal expression of p53. Previous studies showed accumulation of p53 in H pylori–infected individuals and possibly heightened genetic instability.20 The function of p53 is not entirely understood; however, it plays a suppressive regulatory role in the cell cycle.21 In our study, abnormal distribution of p53 was less frequently observed in subjects with gastritis compared with those with multifocal atrophic gastritis, suggesting that it could be a marker for disease progression, particularly since it has been found in areas of intestinal metaplasia and in 60% of patients with early gastric cancer.20,22 

Cyclin D1 (also known as Bcl-2) regulates the transition through the G1 phase of the cell cycle; accumulations of this protein have been demonstrated in several cancers, particularly in the gastrointestinal tract.22,23 The relationship between cyclin D1 and intestinal metaplasia has not been consistent.20 In our study, cyclin D1 was found only in subjects with atrophy and intestinal metaplasia, suggesting this may be a rare but significant event toward the sequential development of gastric neoplasia, since it may only be present in susceptible persons who will eventually develop cancer. Alternatively, cyclin D1 can be difficult to evaluate on biopsy samples because it may be focally expressed.

One year after treatment, we noted the presence of groups of T lymphocytes in the lamina propria of subjects with multiple foci of atrophy and intestinal metaplasia. A residual T-lymphocyte response may be secondary to persistence of H pylori or its antigens in the gastric lamina propria, since most T cells (T_H1) are specific for H pylori.24,25 The presence of T lymphocytes correlated with the presence of cells with abnormal distribution of cell proliferation markers in patients with gastritis and multiple foci of atrophy and intestinal metaplasia. T cells and gastric mucosa can express similar membrane antigens and could be a driving force that produces epithelial changes such as atrophy and intestinal metaplasia similar to what has been proposed for lymphoid neoplasias.26 

In this study, statistically significant reduction of B lymphocytes and macrophages 1 year after treatment did not occur; however, there was a decrease in apoptosis of inflammatory cells that may be important in reducing continued inflammatory triggers. Initially, H pylori infection produces a neutrophilic inflammatory reaction in which cytokines and oxygen-free radicals are released and cause bacterial destruction and mucosal cell death.27,28 The mucosal cell death triggers cell proliferation, possibly starting epithelial changes that may be associated with abnormal distribution of Ki-67, the most frequently encountered proliferation marker in our study. As H pylori infection becomes chronic, B and T lymphocytes increase and release cytokines that result in further apoptosis and proliferation, likely associated with abnormal distribution of other cell proliferation markers such as p53 and cyclin D1. Although reduction of the acute inflammatory response occurs shortly after finishing treatment for H pylori, a slow decrease (but not disappearance) of lymphoid follicles has been documented 1 year after treatment.6,7,29 

The samples tested for this study were specifically selected from the Chiapas cohort, because this population has a high prevalence of atrophy, intestinal metaplasia, dysplasia, and cancer, enabling an opportunity to test the efficacy of treatment 1 year after treatment ended.11,14,15 We selected well-oriented biopsy samples with gastritis and multiple foci of atrophy and intestinal metaplasia to provide adequate assessment of the localization of the markers. Although the sample selection creates an ideal setting in which to test the hypothesis of reversibility of cell proliferation and inflammatory markers 1 year after H pylori eradication, it also creates a bias. Other limitations of this study include the use of biopsy samples, because the total number of cells evaluated for the different markers is not comparable to studies of these markers performed in resection specimens. Even though before and 1-year posttreatment samples had comparable histopathologic features because they were from the same general area in the stomach, samples were not from the exact same location. Additional limitations when obtaining biopsy samples include the focal nature of these gastric lesions and misidentification of biopsy sites by endoscopists because of normal peristaltic movements, problems in identifying mucosal textures, and normal anatomic variations in each patient. Although the number of subjects studied was small because this was a pilot study, statistically significant differences were obtained in some parameters, whereas other parameters showed trends (ie, reduction of B lymphocytes 1 year after treatment, decrease in abnormal expression of p53 in subjects with multifocal atrophic gastritis 1 year after treatment). Further testing of more cases is necessary to confirm our observations and define whether markers that show trends can achieve statistical significance.

The IHC assays have been used to study the cell proliferation markers and type of inflammatory cells primarily in gastric biopsy specimens with a variety of pathologic lesions related to H pylori infection.30–32 The IHC assays have the advantage of permitting detection of the markers while preserving tissue morphologic features, allowing evaluation of topographic localization and distribution patterns; however, problems in interpretation arise because controls (ideally, gastric biopsy samples with no disease) are rarely available and there is no standard for scoring these assays (semiquantitative evaluation vs quantitative methods presented as a percentage or index).30–33 Studies that compare these markers to diseases not related to H pylori infection or normal stomachs have shown less apoptosis and expression of proliferative markers (p53 and Ki-67).31–33 Similarly, increases in expression of p53, Ki-67, and cyclin D1 have been observed in biopsy samples obtained from the gastroesophageal junction of patients with neoplastic and preneoplastic esophageal conditions.34,35 Since this was a pilot study, we elected a descriptive, semiquantitative approach that would allow us to define patterns that can be used when more subjects from the Chiapas cohort are studied. We defined an abnormal distribution pattern as groups of cells that expressed a marker in a particular location that corresponds to more than 2% of cells staining by IHC and is similar to what has been defined as increased expression of these markers and apoptosis in studies that use a quantitative evaluation.31,33 

In summary, in this pilot study statistical differences were observed by using IHC assays for cell proliferation markers and type of inflammatory cells in gastric biopsy specimens that show a variety of histopathologic lesions. The correlation found between the presence of T lymphocytes and increased expression of cell proliferation markers in subjects with gastritis and multiple foci of atrophy and intestinal metaplasia needs to be further studied.

We gratefully thank Raul Belmonte, Cecilia Limón, Juan Antonio Moguel, and Rosario Moreno for assistance with data collection. We thank Amanda Allen, Marc Welsh, and William Lee for their help in testing these samples. We are especially grateful to El Centro de Investigaciones en Salud de Comitan and El Colegio de la Frontera Sur in Chiapas, Mexico, for the use of their facilities.