MATERIALS AND METHODS
As part of an ongoing multicenter study supported by the National Institutes of Health designed to evaluate risk factors for H. pylori infection in the pediatric population, 257 patients with upper gastrointestinal symptoms have been screened and enrolled. Due to persistence of symptoms, 64 (25%) patients underwent diagnostic upper gastrointestinal fiberoptic endoscopy, during which 3 to 4 gastric biopsy specimens from the antrum and corpus were obtained. One biopsy was used for a rapid urease test, a second was cultured, and the remaining were fixed in formalin for histologic evaluation and characterization. Medical records of these 64 patients were reviewed, and the following data were obtained: sex, age, results of the rapid urease test and H. pylori serology.
H. pylori infection was considered present when one or more of the following tests was positive: rapid urease test, serology, and/or histopathology. The H. pylori infected patients were treated with triple therapy and were followed up at their original institution. The patients that had all H. pylori-related tests negative were considered uninfected and served as controls for the analysis. Human Investigation Committee, Institutional Review Board approval was obtained at each study site, and informed consent and assent were appropriately obtained on all patients enrolled in this study.
Serology and Rapid Urease Test
Serology was performed using a highly sensitive and specific (93% and 95% respectively), in-house, research-based, enzyme immunosorbent assay (ELISA) that has been validated against biopsy specimens of pediatric patients and used for various seroepidemiologic studies (16–18). Preparation of the ELISA plate included H. pylori culture, antigen extraction, and protein isolation by freeze-thaw sonication. A standard protein assay was used to determine the accurate and reproducible quantity of solid-phase antigen in the ELISA. ELISA cut-off values, sensitivity, and cross-reactivity of the assay were tested by using sera from biopsy-confirmed, H. pylori-infected patients and uninfected controls (16–19).
Rapid urease test (CLO test, Kimberly-Clark Co., Draper, UT, USA) was performed at each institution using the same assay following the manufacturers instructions. The rapid urease test assays were read at 12, 24, and 48 hours uniformly at each participating site.
Two biopsy specimens were fixed in formalin, embedded in paraffin, and stained with hematoxylin and eosin (H&E). The pathologist at the hospital provided a diagnosis for treatment purposes. A second, blinded review was performed in all cases by the study pathologist (JG). Mucosal type (corpus or antrum) of the biopsy samples was determined histologically. The visual analog scale of the updated Sydney classification for gastritis was used to grade density of H. pylori, amount of active and chronic inflammatory infiltrate, and severity of atrophy and intestinal metaplasia (10). Lymphoid aggregates were counted in each biopsy specimen. All specimens with no visible bacteria on the mucosal surface after H&E staining were stained for H. pylori, using IHC to confirm infection and grade density.
IHC:H. pylori Infection, Inflammatory Response, and Apoptosis
IHC assays were performed on available paraffin blocks (18 H. pylori-infected patients and a subset of controls). The selection of paraffin blocks from controls included uninfected cases with normal histology (3 cases), gastritis (1 case), and atrophy (4 cases).
Four-micron sections of paraffin-embedded tissues were placed on Fisher Plus slides (Fisher Scientific Co., Pittsburgh, PA.). Sections were deparaffinized, rehydrated, and digested in 0.1 mg/ml Proteinase K (Roche Molecular Biochemicals, Indianapolis, IN.) in 0.6 M Tris-HCl pH 7.5, 0.1% CaCl2 for 15 minutes. After being washed, slides were allowed to incubate at room temperature for 60 minutes with the following antibodies: anti-H. pylori, anti-CD3, anti-CD20, and anti-CD68 (Dako Corporation, Carpinteria, CA; catalogue numbers B0471, A0452, M0755, and M0876, respectively). Slides then underwent sequential application of biotinylated swine anti-mouse antibody, avidin-alkaline phosphatase complex, and naphthol/fast red substrate (Dako Corporation). Sections were counterstained in Mayer's hematoxylin (Fisher Scientific) and mounted with aqueous mounting medium (Lerner Laboratories, Pittsburgh, PA.). Positive controls consisted of lymph nodes incubated with the antibodies. Negative controls were tissue sections from each patient's sample incubated with normal mouse ascitic fluid.
The TUNEL (terminal deoxynucleotidyl transferase dUTP nick end labeling) 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 ug/ml of proteinase K (Roche Molecular Biochemicals) for 15 minutes at room temperature. After being washed twice for 5 minutes in phosphate-buffered saline (PBS), the tissues were covered with 40 ul of labeling mix (25 mM Tris-HCl pH 7.2, 0.2 M potassium cacodylate, 5 mM cobalt chloride, 30 U terminal deoxynucleotidyl transferase, 0.6 nmol digoxigenin-11-dUTP, 0.15 mM dATP) and incubated in a humid chamber at 37°C for 60 minutes. The reaction was terminated by incubation in 2 X SSC (1 X SSC is 0.15 M NaCl plus 0.015 M sodium citrate) for 15 minutes at room temperature. Sections were counterstained by using Mayer's hematoxylin. TUNEL assay controls included the omission of terminal transferase (negative controls) and pre-incubation in 0.1 ug/ml DNAse for 1 hour at 23°C (positive control).
Interpretation of both IHC and apoptosis assay included: a) the location of the cells that showed the marker (superficial mucosa, glands, and/or lamina propria), and b) extent of signal graded as absent (no cells staining), low (scattered, individual cells staining in a high power field), and high (groups of cells staining).
Data were entered into Microsoft Excel 2000 and analyzed by using software SPSS (version 10.0). χ2 and Fisher exact tests were used to determine associations between groups (controls, active infection, and past infection) for pathology results (normal, gastritis, atrophy, and intestinal metaplasia), and low and high cell counts (CD20, CD68, CD3, and apoptosis). Follow-up pairwise comparisons were conducted to evaluate the difference among these proportions. P values of <0.05 were considered significant.
Nineteen (30%) children had H. pylori infection by at least one of the assay methods while 45 (70%) were not infected (controls). In the infected group there were 12 (63%) males and 7 females with a mean age of 8.9 years (range 1–16). In the uninfected control group there were 25 (55%) males and 20 females, with a mean age of 9.2 years (range 1–17). Of the 19 children with H. pylori infection, 12 had samples from the corpus and antrum while 7 had antrum only. In the control group, 27 cases had biopsy specimens from the corpus and antrum, 9 had only antrum and 9 only corpus.
A comparison of the most severe pathologic diagnoses showed that most controls had normal histology or mild chronic inflammation (34/45, 75%) while only one H. pylori-infected patient showed these mild histopathologic changes (P = 0.001). Active gastritis was primarily seen in H. pylori-infected patients, being the most severe diagnosis in 6/19 (32%) children, and accompanied by atrophy and metaplasia in 7/ 19 (36%). Active gastritis was only seen in 1/45 (2%) of controls. The mean number of lymphoid aggregates in the control group was 0.5 (range 0–3), while in the H. pylori-infected group the mean was 1.5 (range 0–9). Atrophy and/or intestinal metaplasia accompanied by varying degrees of inflammation were observed in 12/19 (63%) H. pylori-infected children compared to controls in whom atrophy was seen in 10/45 (22%, P = 0.003). None of the controls had intestinal metaplasia, while 4/19 (21%) of H. pylori-infected patients had intestinal metaplasia. The intestinal metaplasia appeared to be accompanied by atrophy in 2 patients. Intestinal metaplasia (Fig. 1) was found in the antrum (3 cases) and corpus (1 case) and consisted of 1 to 4 glands with goblet cells and brush border. Atrophy was suggested in antral biopsy specimens of 8 (42%) H. pylori infected patients; in 2 patients, only 3 glands or less (moderate atrophy) were present in the gastric mucosa (Fig. 2), while 6 patients had 4 to 6 glands (mild) in the mucosa.
Demographic data, diagnostic H. pylori test results, and pathology of the 12 H. pylori infected patients with atrophy and/or metaplasia is presented in Table I. Except for one patient with no inflammation, all others were accompanied by varying degrees of inflammation. All the patients with intestinal metaplasia were older than 9 years; however, possible atrophy was noted in a wider age range.
Characterization of the host inflammatory response was performed for 18 specimens from H. pylori infected children (1 patient with normal histology, 6 with gastritis alone, 7 with atrophy, and 4 with intestinal metaplasia) and 8 controls (3 with normal histology, 1 with gastritis, and 4 with atrophy). Because the amount of tissue in the paraffin block was limited, the TUNEL assay was performed in only 15 patients with H. pylori infection, and 7 controls.
Table II shows the results from the inflammatory cell marker IHCs and the TUNEL assays. H. pylori infected patients had higher numbers of B lymphocytes, predominantly in lymphoid follicles, compared with controls (P = 0.03). Abundant macrophages in the lamina propria were observed in similar numbers in controls and infected patients. T lymphocytes were scattered in the lamina propria, invading the glands and the superficial mucosa in both controls and H. pylori infected patients. Apoptosis was predominantly seen in the superficial epithelium and inflammatory cells; it was rarely seen in glands. Apoptosis appeared increased in patients with H. pylori infection, while in most controls apoptosis appeared low.
In this study, a proportion of pediatric patients with H. pylori infection were found to have intestinal metaplasia and possibly mild to moderate atrophy. Intestinal metaplasia has been reported rarely in the pediatric population, while the existence of atrophy is a matter of debate (5,6,8). From a pathologic stand point, the presence of intestinal metaplasia is easy to document by using routine H&E stains since goblet cells, brush border, and Paneth cells are histologically distinct from gastric mucosal cells (special stains assess functionality and aid in the identification of the different types of metaplasia) (10,20). Since intestinal metaplasia in the stomach has a patchy distribution, the updated Sydney classification recommends obtaining at least 5 gastric biopsies in adults. In pediatric patients gastric sampling tends to be more limited. However, even with the number of biopsies collected in this cohort, we found intestinal metaplasia. The diagnosis of atrophy in adults can be difficult because it relies on having an adequate number of properly oriented gastric biopsy specimens and is subject to interpretational variability (12,13). At present, diagnostic criteria for atrophy in children are not established and studies are needed to ascertain a systematic characterization.
In adults, the association between H. pylori infection, atrophy, and intestinal metaplasia has been well documented (1–4,10,11). It has been postulated that as H. pylori-associated gastritis persists, atrophy and intestinal metaplasia develop sequentially over decades (2,3). The most common lesion found in patients between 40 and 60 years old is mild atrophy. As a result, this lesion is presumed to occur early in the natural history of H. pylori infection. In general, fewer adult patients have the combination of severe atrophy and intestinal metaplasia presumably because fewer patients evolve to this later disease stage (1–4,11). In our patients, we suggest that the entire spectrum of H. pylori-associated gastric disease can be found, beginning with atrophy and ending with atrophy and intestinal metaplasia.
Since intestinal metaplasia was observed in 4 patients, we applied the Sydney system to grade atrophy in our subjects. This is the first time in a pediatric population that atrophy has been graded by the updated Sydney system (10,15). Severe atrophy in conjunction with severe intestinal metaplasia, dysplasia, and gastric adenocarcinoma were not found in our study. The absence of these early malignant lesions may be due to the shorter duration of H. pylori infection in children, the small cohort sample, or selection bias. Mild atrophy was tentatively identified in a proportion of infected and uninfected patients in our cohort. This observation may be explained as an interpretational bias, a problem in specimen orientation, or perhaps the natural outcome of H. pylori infection in children. The latter possibility is supported by having found patients with concomitant atrophy and intestinal metaplasia. Mild gastric atrophy in patients without H. pylori infection was usually accompanied by chronic inflammation and could be attributed to past H. pylori infection that was not detected by the tests used, other gastric infections, vitamin deficiencies, or use of gastro-toxic medications (i.e. non-steroidal anti-inflammatory drugs). Since the histologic diagnosis of gastric atrophy is subject to interpretational bias and since the Sydney system was devised for adult populations, further validation of atrophy parameters needs to be obtained for pediatric biopsy samples.
In most adults, and in 2 of our children, atrophy and intestinal metaplasia are found in the same patient (11). However, there is a small group of patients in whom atrophy and metaplasia were not seen concomitantly. This result may be due to sampling errors or mis-orientation of biopsy specimens, thereby preventing the diagnosis of atrophy. Alternatively, the damage caused by H. pylori-associated gastritis may be repaired differently for secretory glands and superficial epithelium (11). Secretory glands and the surrounding mesenchymal matrix are repaired by fibrosis, resulting in gastric atrophy. Damaged superficial epithelium is repaired by replacement of the surface mucosa with either normal gastric mucosal cells or, when there is persistent inflammation, with intestinal-type epithelium. Thus, we postulate that the specific location and characterization of the host inflammatory response and repair process could play an important role in determining whether a patient develops atrophy, intestinal metaplasia, or both.
Previous studies characterizing the inflammatory response in H. pylori-infected adults have demonstrated that lymphoid follicles are characteristic of the infection, and that T lymphocytes in the lamina propria combined with apoptosis are commonly seen (21–24). On the basis of these observations, it has been postulated that there is an association between degree of H. pylori colonization, chronic inflammation, apoptosis, and atrophy. Pediatric cohorts are usually small since persistent upper gastrointestinal symptoms are rare, H. pylori infection rates are lower than in comparable adult populations, and fewer biopsies are taken during diagnostic endoscopy. A study of 6 H. pylori-infected children with normal gastric biopsy specimens documented the presence of T lymphocytes in an intraepithelial location (25). Two of our cases showed intraepithelial T lymphocytes associated to intestinal metaplasia. The correlation between intestinal metaplasia and intraepithelial T lymphocytes needs to be further studied. Another published study documented that children with active H. pylori gastritis have increased apoptosis and proliferation of the gastric epithelial cells. The number of apoptotic cells decreased when H. pylori was eradicated and gastritis resolved (26). Our results were similar to these since apoptosis in the epithelium and inflammatory cells were more prominent in actively infected patients. In our cases, apoptosis correlated with a higher number of B lymphocytes in lymphoid aggregates and in the lamina propria indicating active infection, which in turn may also predispose some H. pylori-infected patients to atrophy and eventually intestinal metaplasia.
In summary, intestinal metaplasia can occur in children with H. pylori infection. We used the updated Sydney classification system for gastritis in the diagnosis and grading of atrophy in a pediatric population. This system was devised for adult populations, and clearly, atrophy parameters need to be revised for pediatric biopsy samples. We postulate that in H. pylori-infected patients, increased B lymphocytes and apoptosis of the epithelium and inflammatory cells may predispose to gastric mucosal atrophy and eventually intestinal metaplasia. To develop specific treatment strategies for children, larger studies of H. pylori-infected children are needed which better define the clinical spectrum and disease outcomes associated with childhood H. pylori infection.
1. Muñoz N, Kato I, Peraza S, et al. Prevalence of precancerous lesions of the stomach in Venezuela. Cancer Epidemiol Biomarkers Prev 1996; 5:41–46.
2. Correa P, Haenszel W, Cuello C, et al. The gastric precancerous process in a high risk population: cross-sectional studies. Cancer Res 1990; 50:4731–36.
3. Correa P, Haenszel W, Cuello C, et al. The gastric precancerous process in a high risk population: cohort follow-up. Cancer Res 1990; 50:4737–40.
4. Guarner J, Mohar A, Parsonnet J, et al. The association of Helicobacter pylori
with gastric cancer and preneoplasic gastric lesions in Chiapas, Mexico. Cancer 1993; 71:297–301.
5. Ashorn M. What are the specific features of Helicobacter pylori
gastritis in children? Ann Med 1995; 27:617–20.
6. Czinn SJ, Glassman MS. Helicobacter pylori
in infants and children. Adv Pediatr Infect Dis 1996; 11:389–401.
7. Whitney AE, Guarner J, Hutwagner L, et al. Helicobacter pylori
gastritis in children and adults: comparative histopathologic study. Ann Diagn Pathol 2000; 4:279–285.
8. Meining A, Behrens R, Lehn N, et al. Different expression of Helicobacter pylori
gastritis in children: Evidence for a specific pediatric disease? Helicobacter 1996; 1:92–97.
9. Gold BD, van Doorn LJ, Guarner J, et al. Helicobacter pylori
genotypic, clinical and demographic characteristics of infected children. J Clin Microbiol 2001; 39:1348–52.
10. Dixon MF, Genta RM, Yardley JH, Correa P, and the Participants of the International Workshop on the Histopathology of Gastritis, Houston 1994: Classification and grading of gastritis: The updated Sydney system. Am J Surg Pathol 1996; 20:1161–81.
11. Guarner J, Herrera-Goepfert R, Mohar A, et al. Gastric atrophy and extent of intestinal metaplasia in a cohort of Helicobacter pylori-infected patients. Hum Pathol 2001; 32:31–35.
12. Guarner J, Herrera-Goepfert R, Mohar A, et al. Interobserver variability in application of the Revised Sydney Classification for Gastritis. Hum Pathol 1999; 30:1431–34.
13. Zimaity, HM, Graham DY, al-Assi MT, et al. Interobserver variation in the histopathologic assessment of Helicobacter pylori
gastritis. Hum Pathol 1996; 27:35–41.
14. Luzza F, Pensabene L, Imeneo M, et al. Antral nodularity and positive CagA serology are distinct and relevant markers of severe gastric inflammation in children with Helicobacter pylori
infection. Helicobacter 2002; 7: 46–52.
15. Cohen MC, Cueto RE, Balcarce N, et al. Assessment of the Sydney System in Helicobater pylori-associated gastritis in children. Acts Gastroenterol Latinoam 2000; 30:35–40.
16. Khanna B, Cutler A, Israel NR, et al. Use caution serologic testing for Helicobacter pylori
in children. J Infect Dis 1998; 178:460–65.
17. Parkinson AJ, Gold BD, Bulkow L, et al. High prevalence of Helicobacter pylori
in the Alaska native population and association with low serum ferritin levels in young adults. Clin Diagn Lab Immunol 2000; 7: 885–88.
18. Hisada M, Lee MG, Hanchard B, et al. Characteristics of Helicobacter pylori
infection in Jamaican adults with gastrointestinal symptoms. J Clin Microbiol 2001; 39:212–16.
19. Pei Z, Ellison RT, Blaser MJ. Identification, purification, and characterization of major antigenic proteins of Campylobacter jejunii.
J Biol Chem 1991; 266:16363–69.
20. Segura DI, Montero C. Histochemical characterization of different types of intestinal metaplasia in gastric mucosa. Cancer 1983; 52:408–503.
21. Genta RM, Hamner HW, Graham DY. Gastric lymphoid follicles in Helicobacter pylori
infection: frequency, distribution, and response to triple therapy. Hum Pathol 1993; 24:577–83.
22. Hatz RA, Meimarakis G, Bayerdorffer E, et al. Characterization of lymphocytic infiltrates in Helicobacter pylori
associated gastritis. Scand J Gastroenterol 1996; 31:222–28.
23. Moss SF, Sordillo EM, Abdalla AM, et al. Increased gastric epithelial cell apoptosis associated with colonization with cagA+Helicobacter pylori
strains. Cancer Res 2001; 61:1406–11.
24. Vorobjova T, Maaroos HI, Sipponen P, et al. Apoptosis in different compartments of antrum and corpus mucosa in chronic Helicobacter pylori
gastritis. An 18-year follow-up study. Scand J Gastroenterol 2001; 36:136–43.
25. Cohen MC, Drut R. Immunophenotype of inflammatory infiltrates in symptomatic gastric infection by Helicobacter pylori.
Analysis of a series of pediatric cases. Patologia (Mex) 1997; 36:23–30.
26. Jones NL, Shannon PT, Cutz E, et al. Increase in proliferation and apoptosis of gastric epithelial cells early in the natural history of Helicobacter pylori
infection. Am J Pathol 1997; 151:1695–1703.