Decreased Epithelial Cytokine Responses in the Duodenal Mucosa of Helicobacter pylori-infected Duodenal Ulcer Patients

Transcription

1 CLINICAL AND DIAGNOSTIC LABORATORY IMMUNOLOGY, Jan. 2003, p Vol. 10, No X/03/$ DOI: /CDLI Copyright 2003, American Society for Microbiology. All Rights Reserved. Decreased Epithelial Cytokine Responses in the Duodenal Mucosa of Helicobacter pylori-infected Duodenal Ulcer Patients E. Strömberg, 1 A. Edebo, 2 A.-M. Svennerholm, 1 and C. Lindholm 1 * Department of Medical Microbiology and Immunology and Göteborg University Vaccine Research Institute 1 and Department of Surgery, 2 Göteborg University, Göteborg, Sweden Received 19 July 2002/Returned for modification 4 September 2002/Accepted 1 October 2002 Helicobacter pylori colonizes the human stomach and areas of gastric metaplasia in the duodenum, but only a minority of those that are infected develop symptoms, e.g., peptic ulcers. Although most ulcers occur in the duodenum, almost all studies of mucosal immune responses against the infection have been limited to responses in the stomach. In the present study we evaluated whether there are differences in the levels of proinflammatory cytokines as well as immunoregulatory cytokines in the duodenal mucosa of duodenal ulcer (DU) patients and asymptomatic (AS) carriers which may be related to the development of duodenal ulcers. Duodenal biopsy s collected from normal mucosa as well as metaplastic mucosa of DU patients, AS carriers, and uninfected controls were analyzed for a number of cytokines by immunohistochemistry. Interestingly, the level of epithelial staining for several cytokines, e.g., interleukin-8 (IL-8), transforming growth factor (TGF- ), and gamma interferon (IFN- ), was found to be significantly lower in DU patients than in AS carriers and uninfected individuals. No differences were observed when cytokine staining in normal and metaplastic biopsy s was compared. However, larger numbers of IL-8-, IL-6-, TGF- -, and IFN- positive mononuclear cells were observed in the duodenal lamina propria of both DU patients and AS carriers than in that of the uninfected controls. Our finding that a number of cytokines that may be important for the mucosal host defense against H. pylori are strongly decreased in the duodenal epithelium of ulcer patients suggests that a down-regulated immune response plays a role in the development of duodenal ulcers. Helicobacter pylori infection is the most common gastrointestinal bacterial infection worldwide. However, even though half of the world s population is infected with H. pylori, only 10 to 15% of those infected develop peptic ulcer disease (29). The different clinical outcomes may be explained by variations in bacterial virulence factors, as well as differences in the host immune responses. H. pylori induces a strong immune response in the stomach, which nevertheless usually fails to resolve the infection. In the gastric mucosa, H. pylori antigens, together with cytokines induced by the infection, attract and activate leukocytes. During the acute phase, polymorphonuclear neutrophils (PMNs) and monocytes accumulate in the gastric mucosa (36), which leads to increased epithelial permeability and acute mucosal damage (49). After the initial acute phase the inflammation is maintained and is characterized by infiltration of lymphocytes into the mucosa. It has been suggested that the H. pylori-induced T-cell response is predominantly a Th1-type response since increased numbers of gamma interferon (IFN- )-producing T cells can be detected in the infected gastric mucosa compared to the numbers of interleukin-4 (IL-4)- and IL-5-producing cells (Th2-type cytokines), which remain unchanged (3, 18, 28). Furthermore, D Elios et al. (11) have shown that T-cell clones generated from antral biopsy s of H. pylori-infected peptic ulcer patients produce IFN- and IL-12 but usually not IL-4 or IL-5 in response to H. pylori antigen stimulation. The levels of several cytokines, both proinflammatory and immunoregulatory, have been shown to * Corresponding author. Mailing address: Department of Medical Microbiology and Immunology, Göteborg University, Box 435, Göteborg, Sweden. Phone: Fax: catharina.lindholm@microbio.gu.se. be elevated in the H. pylori-infected gastric mucosa. IL-8, a neutrophil-attracting cytokine, has been suggested to be important in H. pylori-associated disease (2, 13, 31, 39). Furthermore, the levels of expression of IL-1, IL-6, tumor necrosis factor alpha, and transforming growth factor (TGF- ) have also been shown to be increased in the antra of H. pyloriinfected subjects, both at the mrna level and at the protein level (31, 39, 55, 56). Hypersecretion of acid is a common feature in patients with antral H. pylori infection (12, 19, 41) and is also often found in patients with duodenal ulcer (DU) disease (12). The increased acid load in the duodenum may cause the duodenal epithelium to transform into duodenal gastric metaplasia (DGM), i.e., a more gastric-like epithelium, which enables H. pylori to colonize the duodenum (48, 53). In children, the presence of DGM together with H. pylori infection has been shown to be associated with an increased risk of developing DUs (17). Although most H. pylori-induced ulcers occur in the duodenum, almost all studies to date have focused on the gastric immune responses to H. pylori. These studies have failed to show any differences in the immune response to H. pylori between DU patients and asymptomatic (AS) carriers that could explain the different outcomes of infection, e.g., the development of DUs. However, we have previously shown that the bacterial load is significantly higher in DU patients than in AS carriers (20), which might reflect bacterial as well as host factors. Indeed, H. pylori strains isolated from the duodenum of DU patients have been shown to express certain virulence factors to a higher extent than strains isolated from AS carriers (51). In this study we have investigated whether these differences in duodenal colonization may be explained at least in part by 116

2 VOL. 10, 2003 CYTOKINES IN H. PYLORI-INFECTED DUODENUM 117 differences in the local immune response. The cytokine levels in both the epithelium and the lamina propria of normal mucosa as well as metaplastic duodenal mucosa of H. pyloriinfected AS carriers and DU patients were analyzed by immunohistochemistry. MATERIALS AND METHODS Subjects. This study was approved by the Human Research Ethics Committee of the Medical Faculty, Göteborg University, Göteborg, Sweden. Six DU patients (mean age, 45.3 years; age range, 24 to 50 years; five men and one woman), six H. pylori-infected AS carriers (mean age, 53.7 years; age range, 25 to 78 years; five men and one woman), and six uninfected, healthy Swedish volunteers (mean age, 46.5 years; age range, 25 to 61 years; three men and three women) were included in the study. Informed consent was obtained from all participants. H. pylori infection was confirmed by serology and culture. The healthy individuals and the AS H. pylori-infected subjects had no history of gastrointestinal disease or symptoms. None of the H. pylori-infected subjects had taken any antibiotic medication prior to sampling of duodenal biopsy s and had not taken any antisecretory medication for at least 10 days prior to endoscopy. Specimen collection and staining technique for identification of metaplastic areas in the duodenum. The study subjects were given an antifoaming agent (Minifoam), and then gastroduodenal endoscopies were performed while the subjects were under local anesthesia with lidocain (Xylocain). A previously described method (14, 34), with some modifications (A. Edebo et al., unpublished data), was used to identify areas of DGM. The endoscope and the duodenal bulb were flushed with water, and 7 to 10 ml of methylene blue (1 mg/ml) was instilled. Two adjacent biopsy samples were taken from different areas of DGM and normal mucosa, respectively, and all individual biopsy s were cut in half. To avoid contamination, the biopsy foreceps were flamed between collection of biopsy s from different areas of the duodenum. Half of the first biopsy was fixed in formalin, embedded in paraffin, stained with hematoxylin, and examined by histopathology. The other half was embedded in O.C.T. compound (Tissue-Tek; Miles Inc., Elkhart, Ind.), immediately snap frozen in liquid nitrogen, and used for immunohistochemical staining as described below. The second biopsy was put in physiological saline and used to culture H. pylori. In addition, one biopsy was collected from the antrum and processed for culture as described below. A blood sample was drawn from all participants for serologic testing, and sera were stored at 20 C until they were assayed for H. pylori-specific antibodies. Histological examination of biopsy s. Histological examination was done by one experienced pathologist, who was blinded to the patients clinical diagnosis. The extent of DGM in each biopsy was determined as the percentage of the total epithelial surface and was graded 0, 1 to 20, 20 to 50, or more than 50%. Only biopsy s with DGM in more than 20% of the epithelium were included in the study, and biopsy s regarded as normal lacked DGM completely. The chronic inflammation score, i.e., the density of mononuclear cells (MNCs), was determined separately from the active inflammation score, i.e., PMN infiltration; and both were graded from 0 to 3 (for none, mild, moderate, and severe, respectively) according to the Sydney system (43). Diagnosis of H. pylori infection. One duodenal biopsy from each sampling area and the antral biopsy were used for culture of H. pylori. The biopsy s were homogenized in 1 ml of physiological saline and cultured on 8.5% horse blood Columbia Iso and Columbia IsoA agar plates. The plates were incubated at 37 C under microaerobic conditions (10% CO 2,5%O 2, 85% N 2 ) and screened for H. pylori-like colonies after 3 days of culture. A dot blot assay with H. pylori-specific monoclonal antibody (MAb) HP30-1:1.6 (5) was used to ascertain the growth of H. pylori bacteria. Sera collected from all subjects were screened for H. pylori-specific antibodies (21). The H. pylori-infected subjects included in the study were all positive by culture, serology, and histopathology for Heliobacter-like organisms. All uninfected subjects included in the study were negative by both culture and serology. caga PCR. The caga gene of 298 bp was detected by PCR amplification with the specific primers R008 and D008 (8). The reaction was performed in a volume of 50 l containing 10 mm deoxynucleoside triphosphate mixture, a 15 M concentration of each primer, 2.5 U of Taq DNA polymerase, Taq buffer, 10 mm MgCl 2, and the bacterial DNA. The samples were denatured at 94 C for 1 min and subsequently amplified for 40 cycles consisting of 94 C for 1 min, 60 C for 1 min, and 72 C for 1 min, followed by 72 C for 5 min. The PCR products were analyzed on a 2% agarose gel. Cytokine-specific MAbs. All cytokine-specific MAbs used were mouse antihuman antibodies of the immunoglobulin G1 (IgG1) isotype. The sources of the MAbs were as follows: anti-il-8 (NAP I), Skafte-Claesson, Mölndal, Sweden; anti-ifn- (1-D1K), MABTECH AB, Nacka, Sweden; anti-il-6 and anti- TGF-, Genzyme Diagnostics, Cambridge, Mass.; anti-il-4 (clone 8F12), ImmunoKontakt, Bioggio, Switzerland; and anti-il-1 (clone 2-D-8), Nordic Biosite, Täby, Sweden. The optimal concentrations of the MAbs were determined in initial experiments to be 5 g/ml for IL-8, IFN-, IL-4, and IL-1, 10 g/ml for TGF-, and 30 g/ml for IL-6. In each experiment, an isotype-matched mouse IgG1 primary antibody (DAKO, Glostrup, Denmark) was used as a negative control. The specificities of the different MAbs were tested by preincubation with the corresponding recombinant cytokine for 3 h at 4 C, which abolished the immunostaining. Immunohistochemistry. (i) Detection of cytokines. Immunohistochemical staining of duodenal biopsy s was performed on 8- m frozen sections mounted on glass slides (SuperFrost/Plus; Menzel-Gläser, Braunschweig, Germany), as described previously (1, 5, 31). Briefly, sections were fixed in 2% paraformaldehyde, air dried, and frozen at 20 C for at least an hour. After permeabilization in phosphate-buffered saline (PBS) 0.1% saponin (Sigma, St. Louis, Mo), endogenous peroxidase activity was blocked by treating the tissue sections with 1% H 2 O % NaN 3 in PBS, and endogenous biotin was blocked with an Avidin-Biotin-Blocking kit (Vector Laboratories Inc., Burlingame, Calif.), according to the instructions of the manufacturer. The tissue sections were incubated with the cytokine-specific MAbs at 4 C overnight. After the sections were washed in PBS 0.1% saponin, they were treated with 1% normal goat serum and subsequently incubated with biotinylated goat anti-mouse IgG1 (Caltag Laboratories, South San Francisco, Calif.) diluted 1:300. Avidin-biotinhorseradish peroxidase complex (Vectastain ABC-HRP kit; Vector Laboratories Inc.) was added, and the sections were developed with the chromogen substrate 3,3-diaminobenzidine (Vector Laboratories Inc.) and counterstained with Mayer s hematoxylin (Histolab, Göteborg, Sweden). The slides were dehydrated with ethanol and xylene and subsequently mounted with Mountex (Histolab). One tissue section from one metaplastic and one normal biopsy from each individual was analyzed. The entire tissue sections were analyzed with a Leica microscope at 200 magnification. The numbers of positively stained MNCs were enumerated and expressed as the numbers of MNCs per square millimeter of tissue. Epithelial staining was determined with Qwin software (Leica, Wetzlar, Germany) and expressed as a percentage of the total epithelial area. The total area of the sections was, on average, 2 mm 2 (range, 0.7 to 3 mm 2 ). (ii) Detection of IFN- receptors. The biopsy s were analyzed for expression of IFN- receptors by the same protocol described above, but with PBS as the washing and dilution solution. Acetone-fixed sections were incubated with 5 g of a goat anti-human IFN- receptor (IFN- R1) antibody of the IgG isotype (R&D Systems, Minneapolis, Minn.) per ml and, for control purposes, with goat anti-mouse Ig (Southern Biotechnology Associates Inc., Birmingham, Ala.). A biotinylated donkey anti-goat IgG (The Binding Site, Birmingham, United Kingdom) diluted 1:600 was used as the secondary antibody, and the sections were subsequently treated as described above. (iii) Detection of apoptosis. Biopsy s from AS carriers and DU patients were sectioned and fixed in 3.7% formaldehyde, and apoptotic cells were detected with the apotacs kit (R&D Systems) according to the instructions of the manufacturer and with all recommended controls. Statistical evaluation. The statistical significance of differences in cytokine staining between the different subject groups was evaluated by the Mann-Whitney test. Differences between different tissue types within a patient group were calculated by the Wilcoxon matched pairs test. P values 0.05 were considered significant. RESULTS Collection of biopsy s from normal and metaplastic duodenal mucosa. H. pylori colonizes the duodenum only in areas of DGM (48, 53). In this study sections from one normal biopsy and one metaplastic duodenal biopsy from the same individual were analyzed in parallel. To discriminate between normal duodenal mucosa and the patchy islands of DGM, the duodenal bulb was stained with methylene blue, which stained the normal mucosa dark blue, while the areas of DGM remained unstained. The biopsy s were histologically evaluated for the degree of metaplasia, and the in vivo detection of gastric metaplasia in the duodenal bulb was found to agree well with the histological findings of DGM

3 118 STRÖMBERG ET AL. CLIN. DIAGN. LAB. IMMUNOL. (Edebo et al., unpublished). At least one metaplastic biopsy was collected from each infected individual. In one of the DU patients, only metaplastic biopsy s could be collected. H. pylori could be cultured from the antrum and from at least one biopsy from an area with DGM but could not be cultured from any of the biopsy s from the normal duodenal mucosa of any of the AS carriers and DU patients in the study. All H. pylori strains, i.e., those collected from both the AS carriers and the DU patients, were caga positive. In the continued analyses, the metaplastic biopsy with the highest degree of DGM and one normal biopsy from each subject were analyzed in parallel. In initial experiments, the variability in staining between several sections from the same biopsy was tested, and neither the number of positive MNCs nor the epithelial staining result was found to differ between different sections from the same biopsy. Decreased epithelial cytokine levels in the duodena of DU patients. The epithelial staining of IL-8, IL-6, IL-1, IFN-, IL-4, and TGF- in the normal and metaplastic mucosa of DU patients, AS carriers, and uninfected individuals was evaluated. No significant differences in epithelial cytokine staining between normal and metaplastic biopsy s could be detected for any of the study groups. However, when the epithelial cytokine staining levels for DU patients and AS carriers were compared, the level of staining was found to be decreased in the DU patients, with significantly lower levels of IL-8 (P ), IL-1 (P ), TGF- (P ), and IFN- (P ) in the metaplastic biopsy s and significantly lower levels of TGF- (P ), IL-6 (P ), and IFN- (P ) in normal duodenal biopsy s. Also, when the epithelial cytokine levels in DU patients and uninfected controls were compared, significantly lower levels of IL-8 (P 0.03), IL-6 (P 0.03), TGF- (P ), IFN- (P ), and IL-4 (P ) were found in the DU patients (Fig. 1). To test whether the lower cytokine levels in the epithelium of DU patients were due to apoptosis of epithelial cells, biopsy s were also analyzed for DNA fragmentation by the terminal deoxynucleotidyltransferase-mediated dutp-biotin nick-end-labeling staining technique. However, the numbers of apoptotic cells in biopsy s from DU patients, AS carriers, and uninfected individuals were found to be similar. Since we observed staining of the epithelium for IFN-, a cytokine not produced by epithelial cells, sections were also analyzed for the expression of the IFN- receptor. The epithelial cells were positively stained for the IFN- receptor (Fig. 2), and areas with strong IFN- staining were found to be associated with higher levels of IFN- receptor staining, and vice versa. No differences in the intensities of epithelial staining for any of the cytokines were observed between AS carriers, DU patients, and uninfected individuals. H. pylori infection induces cytokine production by MNCs in the lamina propria. Tissue sections from normal and metaplastic biopsy s from AS carriers, DU patients, and uninfected controls were also examined for the presence of cytokine-specific MNCs in the lamina propria. Since the frequencies of cytokine-specific cells in the lamina propria were very similar in biopsy s from AS carriers and DU patients, the results for these two groups were combined and compared with those for biopsy s from uninfected individuals (Table 1). When normal and metaplastic biopsy s from the same H. pylori-infected individual were compared, significantly higher numbers of TGF- -specific MNCs (P ) and slightly increased numbers of IL-6-, IFN- -, and IL-1 -positive MNCs were found in metaplastic biopsy s than in normal biopsy s from DU patients (Table 1). Furthermore, significantly higher numbers of MNCs specific for IL-8 (P ), IL-6 (P ), TGF- (P ), and IFN- (P ) were found in the H. pylori-infected subjects than in the uninfected individuals (Table 1). The intensity of cytokine staining, i.e., the amount of cytokine produced, was not calculated. However, when the intensity was subjectively scored, a considerably more intense staining of MNCs was observed for all cytokines in biopsy s from DU patients than from the other subject groups (exemplified in Fig. 2B, D, and F) and was observed most markedly for TGF-. The degree of inflammation does not influence cytokine levels. In order to determine whether the degree of inflammation was related to cytokine production, the extent of active inflammation (i.e., PMN infiltration) and signs of chronic inflammation (i.e., the density of MNCs in the duodenal biopsy s) were evaluated and scored. Five of the six DU patients but only one of the six AS carriers had detectable active inflammation in the duodenum. Furthermore, the active inflammatory score was higher in the DU patients than in either the AS carriers or the uninfected controls (Table 2). It was also found to be higher in metaplastic biopsy s than in biopsy s from normal duodenal mucosa. All H. pyloriinfected individuals showed histological signs of chronic inflammation in the duodenum of at least grade 1 (Table 2), according to the Sydney system (43). No relation between the degree of active or chronic inflammation and the staining of the epithelium or lamina propria was observed for any of the cytokines studied. DISCUSSION The fact that only 10 to 15% of those who are H. pylori infected develop peptic ulcers and that most remain asymptomatic throughout their lives (29) might be explained by differences in bacterial virulence factors, as well as differences in the host immune response. Previous studies have shown that H. pylori induces strong B-cell responses (33), as well as activation of T cells (47, 50), locally in the gastric mucosa. Studies comparing the mucosal immune responses in AS carriers and DU patients have all been limited to the immune responses in the stomach, and differences that may explain the development of DUs have not been identified (23, 31, 56). However, we have previously shown that at the site of most ulcers, i.e., the duodenum, the levels of gastric metaplasia and the densities of H. pylori are considerably higher in DU patients than in AS carriers (20). Furthermore, some differences in the levels of expression of virulence factors between H. pylori strains isolated from the duodenum of DU patients and AS carriers have been identified (51). To investigate whether the observed differences in bacterial colonization can be explained, at least in part, by

4 VOL. 10, 2003 CYTOKINES IN H. PYLORI-INFECTED DUODENUM 119 FIG. 1. Epithelial cytokine staining in normal (N) and metaplastic (M) biopsy s from AS carriers, DU patients, and uninfected individuals (H.p. ). Staining of both superficial epithelium and epithelial cells of the deeper regions of the duodenal crypts was determined and expressed as a percentage of the total epithelial area in the section. Each circle represents a biopsy from one individual, and normal and metaplastic biopsy s from the same individual are connected with a dotted line. Horizontal lines represent the median for each group., P 0.01;, P 0.05.

5 120 STRO MBERG ET AL. CLIN. DIAGN. LAB. IMMUNOL. FIG. 2. Light microphotographs showing immunohistochemical detection of cytokines in duodenal tissue sections. (A and B) IL-8-specific staining (arrows indicate positive MNCs, and the arrowhead indicates a positive neutrophil) of biopsy s from the metaplastic mucosa of an AS carrier (A) and a DU patient (B); (C and D) TGF- staining (the arrow indicates a positive MNC) of biopsy s from the metaplastic mucosa of an AS carrier (C) and a DU patient (D); (D and E) IFN- receptor staining (the arrow indicates positively stained epithelia, and arrowheads indicates a positive MNC) of biopsy s from the normal mucosa of an AS carrier (E) and a DU patient (F). Magnification 97 (original magnification, 100).

6 VOL. 10, 2003 CYTOKINES IN H. PYLORI-INFECTED DUODENUM 121 TABLE 1. Comparison of cytokine expression by lamina propria MNCs in metaplastic and normal duodenal biopsy s from H. pyloriinfected and uninfected individuals Mean no. of cytokine-specific MNCs/mm 2 in metaplastic and normal biopsy s a P value b Cytokine Metaplastic H. pylori positive AS carriers (n 6) DU patients (n 6) Normal Metaplastic Normal H. pylori negative (n 6), normal H. pylori positive, c normal vs metaplastic d H. pylori positive, c normal vs H. pylori negative e IL (3.8 16) 14.7 (7.5 37) 10.2 (2.5 22) 5.7 ( ) 3.2 ( ) NS IL (4.0 38) 9.5 (1.5 31) 11.6 (5.0 54) 10.0 (4.8 15) 4.3 ( ) NS IL (2.5 16) 7.5 (1.0 14) 20.6 (3.3 56) 8.0 (3.3 27) 4.3 (3.2 11) NS NS TGF (4.0 29) 11.5 (4.0 24) 18.3 (8.0 39) 8.8 (3.0 28) 3.5 ( ) IFN- 7.6 ( ) 8.3 (5.5 20) 14.0 (10 42) 6.3 (5.0 29) 2.8 ( ) NS IL (0 16.0) 2.6 (1.0 14) 4.8 (2.0 10) 5.0 ( ) 2.3 (0.5 10) NS NS a Values are given as medians (interquartile ranges). b, P 0.001;, P 0.01;, P 0.05; ns, not significant. c AS carriers and DU patients combined. d Determined by Wilcoxon matched pairs test. e Determined by Mann-Whitney test. differences in the local immune responses, we compared the cytokine responses in the duodenum of H. pylori-infected DU patients and AS carriers, and for comparison, we also evaluated the cytokine responses in uninfected controls. Since H. pylori colonizes the duodenum only in areas of gastric metaplasia (48, 53), we studied the cytokine responses in both normal and metaplastic duodenal mucosa. Even though H. pylori colonization was restricted to the metaplastic areas, the levels of all cytokines analyzed were found to be comparable in the normal and metaplastic mucosa. This might be explained by the bacterial factors that are released, e.g., urease and other enzymes or toxins, which may also affect the surrounding epithelium. In contrast to our results, Noshiro et al. (40) have reported that cultured metaplastic duodenal biopsy s from H. pylori-infected DU patients produce higher levels of IL-8 than normal biopsy s from the same individuals. However, in the study by Noshiro and coworkers (40), the IL-8 levels in culture supernatants were measured by enzyme-linked immunosorbent assay, which reflects the accumulated IL-8 produced over a period of time and by several different cell types, i.e., also by neutrophils, which have been shown to be present in increased numbers in the metaplastic areas (20). Furthermore, we have previously found that culturing of mucosal biopsy s itself induces the spontaneous production of IL-8 (32). H. pylori has been shown to induce a strong cytokine response in both human gastric epithelial cells (31) and gastric epithelial cell lines (10, 25, 26). Whether this is also true for the duodenal epithelium of H. pylori-infected subjects has not previously been analyzed. Here we demonstrate a substantial down-regulation of several cytokines, especially IL-8, TGF-, and IFN-, in the duodenal epithelium of DU patients compared to the levels in AS carriers and uninfected individuals. There are several possible explanations for why DU patients may have decreased epithelial cytokine responses, e.g., apoptosis of the epithelium, bacterial factors, genetic host factors, or possibly, down-regulation by other immune cells. Since apoptosis of epithelial cells has previously been demonstrated in the H. pylori-infected stomach (37, 52), we evaluated whether the lower cytokine levels observed in DU patients simply could be explained by an inability of the epithelial cells to produce cytokines due to apoptosis. However, the epithelium did not show any signs of disruption microscopically, and when epithelial cells were analyzed for DNA fragmentation, the numbers of apoptotic cells were shown to be similar in AS carriers and DU patients. Apoptosis is therefore not a likely explanation for the marked differences in cytokine staining. The considerably lower level of epithelial cytokine staining observed in DU patients could also be explained by differences in the infecting bacterial strains (51), i.e., some bacterial factors that may interfere with cytokine production. caga is one of the most TABLE 2. Comparison of chronic and active inflammation in duodenal biopsy s from H. pylori-infected and uninfected individuals Study group Metaplastic Active inflammation Normal Inflammation score a Metaplastic Chronic inflammation Normal AS carriers (n 6) 0.4 (0 2) 0.2 (0 1) 1.3 (1 2) 1.0 (1 1) DU patients (n 6) 1.6 (0 2) 0.8 (0 2) 1.8 (1 2) 1.4 (1 2) H. pylori-negative subjects (n 6) b 0.0 (0 0) 0.3 (0 0) a Chronic and active inflammation is scored from 0 to 3 according to the Sydney system (43). Values are given as means (interquartile ranges)., P 0.01;, P 0.05 for biopsy samples from AS carriers or DU patients versus normal biopsy samples from H. pylori-negative subjects. b Metaplastic biopsy s were not obtained from the uninfected subjects.

7 122 STRÖMBERG ET AL. CLIN. DIAGN. LAB. IMMUNOL. extensively studied virulence factors of H. pylori and has been associated with cytokine expression (42, 54) and has also been linked to DU disease (20, 51). However, the differences in epithelial cytokine levels that we observed could not be explained by differences in caga expression since all strains, i.e., those collected from both AS carriers and DU patients, were caga positive. Various genetic host factors have been suggested to be of importance for the development of DUs (15, 30). It is possible that the down-regulated epithelial cytokine levels that we observed in the DU patients in this study may be due to differences in genetic host factors. However, this needs to be further investigated. Another possible explanation for the deceased level of cytokine staining of the epithelium in DU patients is down-regulation by other immune cells. Indeed, we have recently observed higher numbers of down-regulatory T cells in the lamina propria of the duodenal mucosa of DU patients compared to the numbers in AS carriers (48a). Furthermore, regulatory T cells have recently been shown to suppress H. pylorispecific immune responses in humans (A. Lundgren et al., submitted for publication), as well as in an experimental mouse model of H. pylori infection (S. Raghavan et al., submitted for publication). However, whether these T cells may also have a down-regulatory effect on the cytokines produced by or bound to epithelial cells remains to be elucidated. IFN- is a cytokine that has been shown to be involved in intestinal epithelial cell growth (4) and barrier function (27), as well as in the induction of major histocompatibility complex class II expression on human intestinal epithelial cells (7, 24). In mice, IFN- has been shown to play a major role in H. pylori infection by increasing the level of gastric inflammation and reducing the level of bacterial colonization (35, 46). Epithelial cells are not believed to produce IFN- but can express the IFN- receptor and thereby respond to this cytokine. The substantial epithelial IFN- staining of biopsy s from AS carriers and uninfected subjects was found to reflect receptor-bound IFN-, produced by MNCs. Accordingly, the decreased level of epithelial IFN- staining in DU patients corresponded to lower levels of IFN- receptor staining, as previously described for the numbers of IFN- and TGF- receptors in Shigella-infected rectal epithelium (44). We believe that the down-regulated epithelial cytokine responses in the duodenum may be indirectly involved in the development of DUs. The decreased cytokine responses in the epithelium of DU patients may allow a higher degree of H. pylori colonization, which is indeed observed in the duodenum of DU patients compared to the degree of colonization in AS carriers (20). A higher bacterial load may lead to increased levels of production of enzymes, toxins, and other bacterial factors that may be harmful, which in turn may cause the formation of DUs. Another consequence of a decreased epithelial cytokine response might be defective signaling from the epithelium to the lamina propria. Thus, the epithelial levels of IL-8 have been demonstrated to be of major importance for neutrophil transport over the epithelium (16). Also, mice deficient in the IL-8 receptor have been shown to have impaired transepithelial neutrophil transport, leading to neutrophil entrapment within the lamina propria and subsequent tissue destruction (22). It is possible that the lower IL-8 levels in the epithelium of DU patients may in part explain the increased numbers of neutrophils that are observed in the lamina propria of DU patients (20). As previously shown for the antrum (2, 13, 31, 39), H. pylori infection was found to significantly induce production of the proinflammatory cytokines IL-8, IL-6, and IFN- and the anti-inflammatory cytokine TGF- by MNCs in the lamina propria of the duodenum of both AS carriers and DU patients. The higher levels of IL-8 in the lamina propria of the duodenum of H. pylori-infected individuals compared to those in uninfected controls that were observed in this study are consistent with previous findings (9, 40). The numbers of cytokine-positive MNCs in the lamina propria were similar in AS carriers and DU patients. However, substantially more intense staining was observed in the ulcer patients, most markedly for TGF-. Considering that we have also observed an increased number of cytotoxic T-lymphocyteassociated antigen 4-positive cells, i.e., possibly regulatory T cells, in the duodenal lamina propria of DU patients (48a), we believe that the TGF- -positive cells are at least partly regulatory T cells (6, 38, 45). This further supports the hypothesis of a down-regulated duodenal immune response in DU patients. The mucosal epithelium plays an important role as a first line of defense against gastroduodenal infections, and an impaired epithelial barrier may play a role in the development of H. pylori-induced ulcers. In this study we have shown that the epithelial cytokine responses in the duodenum of DU patients differ from those in AS carriers and uninfected subjects, with significantly lower levels of IL-8, IFN-, and TGF- in the ulcer patients. Together with our previous observations of higher numbers of down-regulating T cells and, as shown here, more intense staining of the anti-inflammatory cytokine TGF- in the lamina propria, these results indicate that DU patients have a down-regulated duodenal immune response. The lower epithelial cytokine response in the duodenum may explain the higher bacterial load observed in ulcer patients and is therefore likely to be of importance for the pathogenesis of H. pylori-induced DUs. ACKNOWLEDGMENTS This study was financially supported by grants from Astra Research Center, Boston, Mass.; the Swedish Medical Research Council (16X ); the Göteborg Medical Society; and Adlerbertska Forskningsfonden. REFERENCES 1. Andersson, J., J. Abrams, L. Bjork, K. Funa, M. Litton, K. Agren, and U. Andersson Concomitant in vivo production of 19 different cytokines in human tonsils. Immunology 83: Ando, T., K. Kusugami, M. Ohsuga, M. Shinoda, M. Sakakibara, H. Saito, A. Fukatsu, S. Ichiyama, and M. Ohta Interleukin-8 activity correlates with histological severity in Helicobacter pylori-associated antral gastritis. Am. J. Gastroenterol. 91: Bamford, K. B., X. Fan, S. E. Crowe, J. F. Leary, W. K. Gourley, G. K. Luthra, E. G. Brooks, D. Y. Graham, V. E. Reyes, and P. B. Ernst Lymphocytes in the human gastric mucosa during Helicobacter pylori have a T helper cell 1 phenotype. Gastroenterology 114: Boismenu, R., and W. L. Havran Modulation of epithelial cell growth by intraepithelial gamma delta T cells. Science 266: Bolin, I., H. Lonroth, and A. M. Svennerholm Identification of Helicobacter pylori by immunological dot blot method based on reaction of a

8 VOL. 10, 2003 CYTOKINES IN H. PYLORI-INFECTED DUODENUM 123 species-specific monoclonal antibody with a surface-exposed protein. J. Clin. Microbiol. 33: Chen, W., W. Jin, and S. M. Wahl Engagement of cytotoxic T lymphocyte-associated antigen 4 (CTLA-4) induces transforming growth factor beta (TGF-beta) production by murine CD4( ) T cells. J. Exp. Med. 188: Colgan, S. P., C. A. Parkos, J. B. Matthews, L. D Andrea, C. S. Awtrey, A. H. Lichtman, C. Delp-Archer, and J. L. Madara Interferon-gamma induces a cell surface phenotype switch on T84 intestinal epithelial cells. Am. J. Physiol. 267:C402 C Covacci, A., R. Rappuoli PCR amplification of the gene sequences from H. pylori strains. In Helicobacter pylori: techniques for clinical diagnosis and basic research. The W. B. Saunders Co., London, United Kingdom. 9. Crabtree, J. E., J. I. Wyatt, L. K. Trejdosiewicz, P. Peichl, P. H. Nichols, N. Ramsay, J. N. Primrose, and I. J. Lindley Interleukin-8 expression in Helicobacter pylori infected, normal, and neoplastic gastroduodenal mucosa. J. Clin. Pathol. 47: Crowe, S. E., L. Alvarez, M. Dytoc, R. H. Hunt, M. Muller, P. Sherman, J. Patel, Y. Jin, and P. B. Ernst Expression of interleukin 8 and CD54 by human gastric epithelium after Helicobacter pylori infection in vitro. Gastroenterology 108: D Elios, M. M., M. Manghetti, M. De Carli, F. Costa, C. T. Baldari, D. Burroni, J. L. Telford, S. Romagnani, and G. Del Prete T helper 1 effector cells specific for Helicobacter pylori in the gastric antrum of patients with peptic ulcer disease. J. Immunol. 158: el-omar, E. M., I. D. Penman, J. E. Ardill, R. S. Chittajallu, C. Howie, and K. E. McColl Helicobacter pylori infection and abnormalities of acid secretion in patients with duodenal ulcer disease. Gastroenterology 109: Fan, X. G., A. Chua, X. J. Fan, and P. W. Keeling Increased gastric production of interleukin-8 and tumour necrosis factor in patients with Helicobacter pylori infection. J. Clin. Pathol. 48: Futami, H., M. Takashima, T. Furuta, H. Hanai, and E. Kaneko Relationship between Helicobacter pylori infection and gastric metaplasia in the duodenal bulb in the pathogenesis of duodenal ulcer. J. Gastroenterol. Hepatol. 14: Garcia-Gonzalez, M. A., A. Lanas, S. Santolaria, J. B. Crusius, M. T. Serrano, and A. S. Pena The polymorphic IL-1B and IL-1RN genes in the aetiopathogenesis of peptic ulcer. Clin. Exp. Immunol. 125: Godaly, G., A. E. Proudfoot, R. E. Offord, C. Svanborg, and W. W. Agace Role of epithelial interleukin-8 (IL-8) and neutrophil IL-8 receptor A in Escherichia coli-induced transuroepithelial neutrophil migration. Infect. Immun. 65: Gormally, S. M., B. M. Kierce, L. E. Daly, B. Bourke, R. Carroll, M. T. Durnin, and B. Drumm Gastric metaplasia and duodenal ulcer disease in children infected by Helicobacter pylori. Gut 38: Haeberle, H. A., M. Kubin, K. B. Bamford, R. Garofalo, D. Y. Graham, F. El-Zaatari, R. Karttunen, S. E. Crowe, V. E. Reyes, and P. B. Ernst Differential stimulation of interleukin-12 (IL-12) and IL-10 by live and killed Helicobacter pylori in vitro and association of IL-12 production with gamma interferon-producing T cells in the human gastric mucosa. Infect. Immun. 65: Hamlet, A., and L. Olbe The influence of Helicobacter pylori infection on postprandial duodenal acid load and duodenal bulb ph in humans. Gastroenterology 111: Hamlet, A., A. C. Thoreson, O. Nilsson, A. M. Svennerholm, and L. Olbe Duodenal Helicobacter pylori infection differs in caga genotype between asymptomatic subjects and patients with duodenal ulcers. Gastroenterology 116: Hamlet, A. K., K. I. Erlandsson, L. Olbe, A. M. Svennerholm, V. E. Backman, and A. B. Pettersson A simple, rapid, and highly reliable capsulebased 14 C urea breath test for diagnosis of Helicobacter pylori infection. Scand. J. Gastroenterol. 30: Hang, L., B. Frendeus, G. Godaly, and C. Svanborg Interleukin-8 receptor knockout mice have subepithelial neutrophil entrapment and renal scarring following acute pyelonephritis. J. Infect. Dis. 182: Hida, N., T. Shimoyama, Jr., P. Neville, M. F. Dixon, A. T. Axon, T. Shimoyama, Sr., and J. E. Crabtree Increased expression of IL-10 and IL-12 (p40) mrna in Helicobacter pylori infected gastric mucosa: relation to bacterial cag status and peptic ulceration. J. Clin. Pathol. 52: Horie, Y., M. Chiba, T. Suzuki, T. Kudo, A. Kamata, M. Iizuka, and O. Masamune Induction of major histocompatibility complex class II antigens on human colonic epithelium by interferon-gamma, tumor necrosis factor-alpha, and interleukin-2. J. Gastroenterol. 33: Huang, J., P. W. O Toole, P. Doig, and T. J. Trust Stimulation of interleukin-8 production in epithelial cell lines by Helicobacter pylori. Infect. Immun. 63: Jung, H. C., J. M. Kim, I. S. Song, and C. Y. Kim Helicobacter pylori induces an array of pro-inflammatory cytokines in human gastric epithelial cells: quantification of mrna for interleukin-8, -1 alpha/beta, granulocytemacrophage colony-stimulating factor, monocyte chemoattractant protein-1 and tumour necrosis factor-alpha. J. Gastroenterol. Hepatol. 12: Kaoutzani, P., S. P. Colgan, K. L. Cepek, P. G. Burkard, S. Carlson, C. Delp-Archer, M. B. Brenner, and J. L. Madara Reconstitution of cultured intestinal epithelial monolayers with a mucosal-derived T lymphocyte cell line. Modulation of epithelial phenotype dependent on lymphocytebasolateral membrane apposition. J. Clin. Investig. 94: Karttunen, R., T. Karttunen, H. P. Ekre, and T. T. MacDonald Interferon gamma and interleukin 4 secreting cells in the gastric antrum in Helicobacter pylori positive and negative gastritis. Gut 36: Kuipers, E. J., J. C. Thijs, and H. P. Festen The prevalence of Helicobacter pylori in peptic ulcer disease. Aliment. Pharmacol. Ther. 9: Lanas, A., M. A. Garcia-Gonzalez, S. Santolaria, J. B. Crusius, M. T. Serrano, R. Benito, and A. S. Pena TNF and LTA gene polymorphisms reveal different risk in gastric and duodenal ulcer patients. Genes Immun. 2: Lindholm, C., M. Quiding-Jarbrink, H. Lonroth, A. Hamlet, and A. M. Svennerholm Local cytokine response in Helicobacter pylori-infected subjects. Infect. Immun. 66: Lindholm, C., M. Quiding-Jarbrink, H. Lonroth, and A. M. Svennerholm Induction of chemokine and cytokine responses by Helicobacter pylori in human stomach explants. Scand. J. Gastroenterol. 36: Mattsson, A., M. Quiding-Jarbrink, H. Lonroth, A. Hamlet, I. Ahlstedt, and A. Svennerholm Antibody-secreting cells in the stomachs of symptomatic and asymptomatic Helicobacter pylori-infected subjects. Infect. Immun. 66: Mertz, H., T. Kovacs, M. Thronson, and W. Weinstein Gastric metaplasia of the duodenum: identification by an endoscopic selective mucosal staining technique. Gastrointest. Endosc. 48: Mohammadi, M., S. Czinn, R. Redline, and J. Nedrud Helicobacterspecific cell-mediated immune responses display a predominant Th1 phenotype and promote a delayed-type hypersensitivity response in the stomachs of mice. J. Immunol. 156: Mooney, C., J. Keenan, D. Munster, I. Wilson, R. Allardyce, P. Bagshaw, B. Chapman, and V. Chadwick Neutrophil activation by Helicobacter pylori. Gut 32: Moss, S. F., J. Calam, B. Agarwal, S. Wang, and P. R. Holt Induction of gastric epithelial apoptosis by Helicobacter pylori. Gut 38: Nakamura, K., A. Kitani, and W. Strober Cell contact-dependent immunosuppression by CD4( )CD25( ) regulatory T cells is mediated by cell surface-bound transforming growth factor beta. J. Exp. Med. 194: Noach, L. A., N. B. Bosma, J. Jansen, F. J. Hoek, S. J. van Deventer, and G. N. Tytgat Mucosal tumor necrosis factor-alpha, interleukin-1 beta, and interleukin-8 production in patients with Helicobacter pylori infection. Scand. J. Gastroenterol. 29: Noshiro, M., K. Kusugami, T. Sakai, A. Imada, T. Ando, K. Ina, K. Nobata, K. Morise, H. Kaneko, M. Ito, and Y. Nishio Gastric metaplasia in the duodenal bulb shows increased mucosal interleukin-8 activity in Helicobacter pylori-positive duodenal ulcer patients. Scand. J. Gastroenterol. 35: Olbe, L., A. Hamlet, J. Dalenback, and L. Fandriks A mechanism by which Helicobacter pylori infection of the antrum contributes to the development of duodenal ulcer. Gastroenterology 110: Peek, R. M., Jr., G. G. Miller, K. T. Tham, G. I. Perez-Perez, X. Zhao, J. C. Atherton, and M. J. Blaser Heightened inflammatory response and cytokine expression in vivo to caga Helicobacter pylori strains. Lab. Investig. 73: Price, A. B The Sydney system: histological division. J. Gastroenterol. Hepatol. 6: Raqib, R., A. A. Lindberg, L. Bjork, P. K. Bardhan, B. Wretlind, U. Andersson, and J. Andersson Down-regulation of gamma interferon, tumor necrosis factor type I, interleukin 1 (IL-1) type I, IL-3, IL-4, and transforming growth factor beta type I receptors at the local site during the acute phase of Shigella infection. Infect. Immun. 63: Read, S., V. Malmstrom, and F. Powrie Cytotoxic T lymphocyteassociated antigen 4 plays an essential role in the function of CD25( )CD4( ) regulatory cells that control intestinal inflammation. J. Exp. Med. 192: Sawai, N., M. Kita, T. Kodama, T. Tanahashi, Y. Yamaoka, Y. Tagawa, Y. Iwakura, and J. Imanishi Role of gamma interferon in Helicobacter pylori-induced gastric inflammatory responses in a mouse model. Infect. Immun. 67: Seifarth, C., A. Funk, K. Reich, I. Dahne, M. Classen, and K. Deusch Selective increase of CD4 and CD25 T cells but not of gamma delta T cells in H. pylori associated gastritis. Adv. Exp. Med. Biol. 371B: Steer, H. W Surface morphology of the gastroduodenal mucosa in duodenal ulceration. Gut 25: a.Stromberg, E., A. Lundgren, A. Edebo, B. S. Lundin, A. M. Svennerholm, and C. Lindholm. Increased frequency of activated T cells in the Helicobacter pylori-infected antrum and duodenum. FEMS Immunol. Med. Microbiol., in press. 49. Telford, J. L., A. Covacci, R. Rappuoli, and P. Chiara Immunobiology of Helicobacter pylori infection. Curr. Opin. Immunol. 9: Terres, A. M., and J. M. Pajares An increased number of follicles

9 124 STRÖMBERG ET AL. CLIN. DIAGN. LAB. IMMUNOL. containing activated CD69 helper T cells and proliferating CD71 B cells are found in H. pylori-infected gastric mucosa. Am. J. Gastroenterol. 93: Thoreson, A. C., A. Hamlet, J. Celik, M. Bystrom, S. Nystrom, L. Olbe, and A. M. Svennerholm Differences in surface-exposed antigen expression between Helicobacter pylori strains isolated from duodenal ulcer patients and from asymptomatic subjects. J. Clin. Microbiol. 38: Wagner, S., W. Beil, J. Westermann, R. P. Logan, C. T. Bock, C. Trautwein, J. S. Bleck, and M. P. Manns Regulation of gastric epithelial cell growth by Helicobacter pylori: offdence for a major role of apoptosis. Gastroenterology 113: Wyatt, J. I., B. J. Rathbone, M. F. Dixon, and R. V. Heatley Campylobacter pyloridis and acid induced gastric metaplasia in the pathogenesis of duodenitis. J. Clin. Pathol. 40: Yamaoka, Y., M. Kita, T. Kodama, N. Sawai, and J. Imanishi Helicobacter pylori caga gene and expression of cytokine messenger RNA in gastric mucosa. Gastroenterology 110: Yamaoka, Y., M. Kita, T. Kodama, N. Sawai, K. Kashima, and J. Imanishi Expression of cytokine mrna in gastric mucosa with Helicobacter pylori infection. Scand. J. Gastroenterol. 30: Yamaoka, Y., T. Kodama, M. Kita, J. Imanishi, K. Kashima, and D. Y. Graham Relation between clinical presentation, Helicobacter pylori density, interleukin 1beta and 8 production, and caga status. Gut 45: