IIIII III C A ~ E # 5 0 =~ "~ Iz z~ IS

//,pyloH - ~ i f Jc

Hum~APC

523

Four Membrane-Associated Antigens of Helicobacter pylori Suitable for Diagnosis of Both Infection and Eradication. Petra Voland, II Medical Dept, Munich Germany; David Weeks, VAGLAI-IS and Univ of CA, Los Angeles, Los Angeles, CA; Dino Vaira, Dept of Medicine, Orsola Hosp, Bologna Italy; Nina Neumayer, Dept of Medicine II, Munich Germany; George Sachs, VAGLAHS and Univ of CA, Los Angeles, Los Angeles, CA; Christian Pdnz, Oept of Medicine II, Teeh Univ, Munich Germany

Background. A serological test for diagnosis of H. pylori infection and for evaluation of eradication therapy requires antigens that are specific short half-life indicators of infection. Such a test should also detect a significant decrease of antibody titers upon eradication within a short time. Methods. Cell lysates from three H. pylori strains and one Campylobacter jejuni strain were probed with sera obtained from 10 H. pylori infected patients before, 3 and 5 months after eradication or with sera from non-infected patients. Proteins were defined using 2D gel electrophoresis and mass spectrometry. Antigens of interest were expressed as recombinant proteins using 6xHis tag expression system. Results. Western blot analysis determined that the antibody response against H. pylori antigens decreased progressively after treatment by only at MWt bands between 10 and 32 kDa. In contrast, the antibody response towards DraB or Hsp60 remained unchanged. Eight putative H. pyloriantigens were identified at these low MWts: neuraminyllactose-binding hemagglutinin precursor (Hp~) and 3-oxoacid CoA-transferaee subunit A (CoA-trans, both at 32 kDa); the elongation factor P (EF- P) and peptidoglycan-associated lipoprotein precursor (OmpfB, both at 30 kDa); hypothetical protein HP0596 and adhesin thiol peroxidase (TagD, both at 22 kDa); and 50S ribosomal protein L7/L12 (RPL7/12) and subunit b of the Fo ATP synthase (ATP-Fob', both at 14 kDa). All eight putative antigens were expressed in E. co/ias recombinant proteins and subsequently probed with patients sera. Four membrane-associated proteins (Hp~, Ompl8, HP0696 and ATP-Fob') strongly reacted with sera from H.pylori infected patients but not with non-infected patients. Reactivity against these four antigens decreased by >60% 3 months and >70% five months after eradication. The other proteins found coincident with these antigens (CoA, EF-P, TagD) either did not react with patients sara or cross-reacted (RPLT/L12) with sara from H.pylorinegative patients. Conclusion. A combination of these four membrane-associated antigens HpaA, Ompl8, HP0596 and ATP-Fob' expressed as recombinant proteins could provide the basis for a serological test of H. pyloriinfection and for assessment of successful eradication therapy.

524

Detection of Helicobactar pylori infection in Bleeding Peptic Ulcers: Are Noninvasive Tests Really Better? Albert Beyer, Juergen Barnert, Martin Wienbeck, Zentralklinikum Augsburg, Augsburg Germany

Background: The rapid grease test (RUT) is widely used to detect Helicobacter pylori (H.p.) infection in patients with bleeding peptic ulcer. It was shown (Tu ef al., Gastrointest Endosc 1999;49:302-306) that the sensitivity of the RUT is low in this clinical setting. Aims: 1. To compare the sensitivity of the invasive (RUT) and noninvasive (~30 urea breath test) in detecting H.p. infection in patients with bleeding peptic ulcers. 2. To evaluate if the presence of blood in the gastric antrum interferes with the results. 3. To find out if a short course of acid suppressive therapy in peptic ulcer patients affects the results of the 13C urea breath test. Methods: In 54 patients with bleeding peptic ulcer a two-probe-RUT (probes from gastric antrum and corpus, HUT-Test ®, ASTRA Chemicals) was done during the first upper endoscopy. 13C urea breath test followed within two hours after the emergency endoscopy. Then a high dose (240 rag/d) omeprazole therapy was started. The ~3C urea braath test was repeated 12 hours after the emergency endoscopy. Patients were classified H.p. positive if any of the two tests showed a positive result. Patients were divided into two groups: group A with and group B without blood or blood-containing material in the antrum. The data were analyzed by using Fisher's exact test and McNemar's test (a = 0.05). Results: 29/54 patients were tested positive in at least one of the diagnostic procedures. The RUT was significantly more often positive (93.1%) than was the ~3C urea breath test (79.3%, p<O.O01). There was no difference concerning the sensitivity of the RUT in patients showing blood in the antrum (sensitivity group A 93.3%, group B 92.9°/°) compared with those without. In both groups the ~C urea breath test (group A 80.0% (p = 0.001) and group B 78.6% (p<O.O01)) was less frequently positive. A short course (<12 hours) of acid suppressive therapy lowered the sensitivity of the ~3C urea breath test to 35.7% in group A and 33.3% in group B (p=O.01). Conclusion: The invasive RUT detects an H.p. infection in bleeding peptic ulcer more frequently than the noninvasive ~C urea breath test. The presence of blood in the antrum does not affect the diagnostic yield of the RUT. A short course of high-dose acid suppressive therapy (< 12 hours) lowers the sensitivity of the ~3C urea breath test dramatically.

525

Diagnostic Accuracy of Rapid Urease Test for the Diagnosis of Helicobacter pylori Infection in Patients with Acute Duodenal or Gastric Ulcer Bleeding: A Prospective Evaluation on 96 Patients Dieter Schilling, Andrea Demel, Dept of Gastroenterology, Ludwigshafen Germany; Thomas Nuesse, Dept of Pathology, Ludwigshafen Germany; Juergen F. Riemann, Dept of Gastroenterology, Ludwigshafen Germany

Introduction:Although the rapid urease teat(RUT)is well evaluated in patients with uncompli- cated peptic ulcer disease for the diagnosis of Helicobacter pylori(H, pylod)- infection the diagnostic accuracy of the commonly used RUT has not been established in patients with ulcer bleeding. We therefore evaluated the efficacy of the RUT in patients with bleeding gastric or duodenal ulcers. Mathods:96 consecutive patients with acute ulcer bleeding without proton- pump-inhibitor (ppi) therapy and without antibiotic therapy within the last 14 days before bleeding were included into the study. During index endoscopy two specimens of the antrum and corpus mucosa were obtained for histological examination of the H.pylori status and for grading gastritis (updated Sydney System). One further specimen from antrum and corpus mucosa was taken for the RUT. The same procedure was performed in 96 matched controls (age, gender, ulcer Iocation,concomittant diseases) with uncomplicated gastric or duodenal ulcer. Nearly all patients were investigated by 13C urea breath test. The diagnostic quality parameters were calculated with the Histology and the 13C urea breath test as reference. Results:The H.pylori prevalence in the bleeding group was 50 % whereas the control group showed an infection rate of 75 % (p<O.05). The mean age was 56.8 years, 73 % of the patients were male. The sensitivity of the RUT was 80 %, the specificity 100 %. The negative predictive value was 75 %. These values were significantly different from those of the control group (sensitivity 95%,specificity lO0%,negative predictive value 88%). Conclusion:The early diagnosis of H.pyled infection in ulcer bleeding is important, because all available diagnostic tools have decreased diagnostic accuracy during ppi-therapy. The only use of RUT for diagnosis of H.pylori can not be recommended in patients with ulcer bleeding.

526

VacA And Urease Released By H. Pylori Stimulate COX-2 Expression In Macrophages Mahmood Akhtar, Jamie C. Newton, David J. McGee, Harry L T Mobley, Aiain P. Gobert, Yulan Cheng, Keith T. Wilson, Univ of Maryland Sch of Medicine, Baltimore, MD

Background: Cyclooxygenase (COX)-2 is the inducible form of the rate limiting enzyme for prostaglandin synthesis. We have shown that 1) COX-2 expression is upregulated in lamina propda mononuclear cells of H. py/ori (Hp) gastritis tissues; 2) Hp bacteria, lysates, and water extracts stimulate COX-2 expression in monocytes/macrophages in vitr#, 3) COX-2 products inhibit Thl responses to Hp. Aims: We sought to determine if secreted or released Hp products induce COX-2, and if these products include the proteins VacA, CagA, and urease. Methods: Hp strains 3401 wild-type (WT) and vacA-, cagA- and ureA- were used. Hp were grown in brucella broth, or in serum-free Ham's F12. Supematants from the latter were concentrated lO0-fold with Amicon 10 kD cut-off filters. Preparatations were added to murine RAW 264.7 macrophages. COX-2 mRNA levels (by Northern blotting) and prostaglandin (PG)E~ levels (by enzyme immunoassay) were assessed at 6 and 24 h, respectively. Recombinant VacA and urease were also used as stimuli. Results: Compared to broth alone, Hp broth culture supernatants stimulated 56-fold increases in PGE2 levels and 70-fold increases in COX-2 mRNA. When compared to Hp WT, vacA- broth supernatants lost 72% of the induction of PGE2 release and 85% of the induction of COX-2 mRNA. The supernatants from Hp grown in F12 media demonstrated large, concentration-dependent increases in PGE2 production and COX-2 mRNA in the range tested (0.1 to 5 ~ protein/ml), ureA- and vacA- strains each showed a 75% loss of induction of COX-2 mRNA and PGE2 levels (Table) in this protein range, while CagA- supernatants retained COX-2 inducing activity. VacA and urease proteins in WT supernatants, but not in the corresponding mutant supernatants, were demonstrated by Western blotting. Stimulation of macrophages with grease (1 p,g/ml) or vacA (10 ng/ml) protein resulted in large, > 50-fold increases in PGE2 and COX-2 mRNA levels. Conclusions: Soluble factors secreted or released by Hp, including VacA and urease, can stimulate large increases in COX-2 expression and activity. Luminal Hp may regulate mucosal immune responses by this induction of COX-2.

Macrophage PGE2 levels (ng/ml) in response to Hp supematants (1 pg/ml added)

Control Hp WT Hp vacA. Hp ureA- 0.1:L-O.O 27.6-~5.0" 7.9it.1~ 6.9-~0.6~

* p < O.Ol vs. conlrol, § p < O.Ol vs. Hp WT

527

IL-IO Deficiency Reduces Huli¢obacter pylori Gastric Colonization in C57BL/6 Mice Mark T. Whary, Barbara J. Sheppard, Jennifer Cline, Sandy Xu, James G. Fox, MIT, Cambridge, MA

BACKGROUND: Helicobacter pylori infection of C57BL/6 mice is an established model for human H. pylori gastritis because mice remain persistently colonized and develop significant gastdtis associated with Thl immune responses. IL-lO is an important Th2-associated regula- tory cytokine that suppresses Thl cell function and IL-IO knockout (KO) mice develop severe hyperplastic gastritis from experimental infection with H. fells. IL-IO KO mice also are prone to inflammatory bowel disease secondary to infection with either urease positive or negative helicobacters that also induce Thl responses. We examined the potential interaction between I L-IO deficiency and coinfection with H. rodentium, a urease-negative helicobacter that naturally infects the cecum and colon of mice, on host responses to gastric colonization of H. pylori (Sydney) in mice. METHODS: Nine helicobacter-free wildtype C57BL/6 and 10 IL-IO KO mice were experimentally infected with H. rodentium alone, 15 C57BI./6 and 25 IL-IO mice were infected with H. py/ori alone and 19 of each mouse genotype were super-infected with H.

A-99