| Japanese Journal of Clinical Oncology | Pages |
Introduction
Epidemiology and Pathology
Gene diversity and virulence
Immunoresponse and Vaccination
Eradication: Gastric lymphoma
Helicobacter species other than H.pylori
Problems to be solved
References
US-Japan Cooperative Cancer Research Program Seminar on Helicobacter and Gastric Cancer
Introduction
The bacterium now known as Helicobacter pylori (H.pylori) was first isolated in1983 from endoscopic biopsy specimens of human gastric mucosa from patients with gastritis and peptic ulcers (1). Its pathological significance has become increasingly appreciated since that time. It was identified as the principal cause of peptic ulcer at an NIH Consensus Development Conference in 1994 (2) and was evaluated by the International Agency for Research on Cancer as Group 1, carcinogenic to humans, in the IARC Monographs on the Evaluation of Carcinogenic Risks to Humans later that year, based on its association with both gastric adenocarcinoma and gastric lymphomas (3). However, many aspects of its mechanism of carcinogenic action still need to be clarified. Also, since the discovery of H.pylori, increasing numbers of additional species within this genus have been isolated and described. At least one of these, H.hepaticus, which was isolated from laboratory mice at the National Cancer Institute in 1994, is also associated with the development of neoplasia, in this case hepatocellular neoplasms (4).
Because of the longstanding interest in gastric carcinoma in Japan and increasing interest in the microbial genus Helicobacter in the USA, this seminar was held on February 12-14, 1996 at the Marriott Hotel, Maui, Hawaii organized within the framework of the US-Japan Cooperative Cancer Research Program by Dr Jerry M. Rice, National Cancer Institute, Frederick, MD, and Dr Daizo Saito, National Cancer Center Hospital, Tokyo. There were eight participants from the USA and six participants and two observers from Japan.
Following a welcome from Dr Rice, Dr Takashi Sugimura (President Emeritus, National Cancer Center, Tokyo) presented an outline of the US-Japan Cooperative Cancer Research Program, which has emphasized intensive discussions of topics of current interest in cancer research. He reminded seminar participants that there were now many well documented associations world-wide between chronic infections or inflammatory processes and neoplasia and that many of these inflammatory processes are non-viral in origin. These include the tuberculosis -> pyothorax -> B cell lymphoma sequence documented in Japan, the development of cholangiocarcinoma subsequent to infection with the liver fluke Opisthorchis viverrini in Thailand and the association between infestation with the bladder fluke, Schistosoma haematobium, and squamous carcinoma of the urinary bladder in Egypt. The addition of bacterial infection by H.pylori to this list should thus be seen as broadening an existing concept.
Dr Saito introduced the seminar by emphasizing the impact that recognition of H.pylori has had on the concept of how gastritis in humans is related to development of carcinoma. He outlined major problems in pathogenesis of human gastric neoplasia that remain to be clarified and issues related to pathogenicity of H.pylori that are not yet understood. He noted the investigational possibilities that are presented by other species of Helicobacter from non-human host species and how these may help clarify the roles of H.pylori in human cancer. Finally, he emphasized the exciting prospects for control of specific human cancers by vaccination against causative or co-causative infectious agents and the ongoing interest in Japan regarding the possibility of utilizing this approach to prevent cancers associated with H.pylori.
Epidemiology and Pathology
Dr Pelayo Correa (Louisiana State University Medical School, New Orleans, LA) then presented an overview of the subject, entitled `H.pylori and gastric carcinogenesis'. As of 1985, stomach cancer was second in prevalence only to lung cancer worldwide. There are great differences in prevalence among countries and between population groups within individual countries. Japan is a high-incidence country and the USA has been regarded as a low-incidence country, but whereas Japanese and US whites are at opposite ends of the incidence scale, within the USA there are higher incidence populations and US blacks, Hispanics, native American Indians, Hawaiians and some other groups have 3-4 times the rate of US whites. Also, a birth cohort phenomenon has been documented in both US whites and in US immigrants of Japanese ancestry: succeeding generations have successively lower lifetime risks, an observation that suggests that there are environmental exposures during childhood that can significantly affect risk for life and that these are becoming less common, at least in some populations. Such environmental exposure might result in persistent infection that begins in childhood. It is noteworthy that in Colombia, for example, [sim]90% of the population is infected with H.pylori by the age of 9.
Dr Correa reviewed the histological classification of Lauren for gastric carcinoma, noting that most human gastric cancers belong to either the Lauren intestinal or Lauren diffuse type. The distinction is important, because the intestinal type was by far the most prevalent (in the USA) during 1940-50, but its incidence is falling more rapidly than the incidence of all gastric cancers combined and, among Americans of Japanese descent, decreased incidence of the intestinal type is almost entirely responsible for the decreasing incidence in successive generations. There is a clear progression of increasingly severe pathology that precedes the intestinal type of gastric carcinoma: normal mucosa -> superficial gastritis -> multifocal atrophic gastritis (MAG) -> small intestinal metaplasia -> colonic metaplasia -> dysplasia -> carcinoma (5,6). Diet clearly plays a role in determining risk, but that role is not well defined, with the exception that high salt intake increases risk, eating fresh fruits and vegetables is protective and nitrite in the diet is under scrutiny as a possible risk factor.
Many observations support a role of H.pylori in the sequence of pathological changes in gastric epithelium that lead to the intestinal type of gastric carcinoma (7-11). It is now generally accepted that H.pylori is the main cause of chronic active gastritis. H.pylori colonizes only normal foveolar epithelium, not metaplastic epithelium. Neutrophils and lymphoid cells are attracted to sites of H.pylori colonization and inducible nitric oxide synthase is histochemically demonstrable in the neutrophils, inviting the speculation that endogenous nitric oxide may play a role in the pathogenesis of H.pylori-associated lesions. Mucin disappears from the surface epithelium colonized by the bacteria, making it vulnerable to damage and a vicious circle begins: atrophy reduces acid secretion, which favors increased colonization by bacteria, which further promotes atrophy. In the population studied by Dr Correa in New Orleans, the risk of gastritis progressing to atrophy was sixfold greater in persons with H.pylori infection and the risk of metaplasia was increased fivefold. Black patients had a higher prevalence of H.pylori infection, more severe acute and chronic gastritis, higher gastric pH and lower ascorbic acid concentration in gastric fluid. These observations may help explain the higher incidence of gastric carcinoma in the black population of New Orleans, in comparison with the white population. However, there are many enigmas in the epidemiological literature on prevalence of H.pylori infection and risk of gastric carcinoma. For example, in Costa Rica, the same rates of infection prevail both in the highlands and along the coast, but highland dwellers develop 3-4 times more cases of carcinoma (12). In parts of Africa, infection is highly prevalent, but gastric carcinoma is relatively uncommon (13). This indicates that many aspects of the relationship between infection with H.pylori and risk of gastric carcinoma remain to be clarified. Dr Correa proposed, as an etiological hypothesis, that H.pylori infection may play a role at three different stages: (a) it causes chronic gastritis; (b) it interferes with normal gastric secretion of ascorbic acid; and (c) it attracts and activates neutrophils and macrophages. Both epithelial hyperproliferation that accompanies chronic active gastritis and the loss of protective mucin that guards against carcinogens in the luminal contents may be part of the carcinogenic process.
Dr Norio Matsukura (Nippon Medical School, Tokyo) discussed `The role of H.pylori infection in the chronic gastritis -> intestinal metaplasia -> gastric cancer sequence'. He measured serum pepsinogen (PG) I and II levels, PG I/II ratios, serum anti-H.pylori IgG antibody and a variety of biochemical polymorphisms in 82 gastric cancer patients matched in age and gender with 151 controls. In this specific study, H.pylori infection did not correlate significantly with gastric carcinoma. However, atrophic mucosal change, detected by abnormal serum PG levels, was significantly associated with gastric carcinoma and these markers changed significantly after eradication of H.pylori, suggesting the possibility that the superficial gastritis -> chronic atrophic gastritis -> intestinal metaplasia sequence is reversible and that the bacterial infection is important for its progression.
Dr Haruhiko Fukuda (National Cancer Center Hospital, Tokyo) presented `H.pylori infection, serum pepsinogen level and gastric cancer: a case-control study in Japan' (14). His cases were 297 gastric cancer patients selected from among 499 who underwent gastrectomy at the National Cancer Center Hospital during 1989-90 and 786 controls from 1925 outpatients without cancer detected clinically in any organ who were seen during the same period. Controls were matched to each case for sex, age (to within 3 months) and date of blood sampling (to within 3 months) and 297 matched sets with 1-16 controls were obtained. Stored serum samples from 282 cases and 767 controls were tested for anti-H.pylori IgG (by ELISA) and for serum pepsinogen (PG) I and II levels. All data together showed no significant association of H.pylori infection and gastric cancer risk (OR = 1.04; 95% CI, 0.73-1.49). When subgroups were analyzed individually, however, an association was suggested in females (OR = 1.57), in patients <50 years old (OR = 1.86), for early cancers (OR = 1.53) and for cancers <40 mm in diameter (OR = 1.55). In contrast, there was a tendency for odds ratios to decrease with increasing age of patients and increasing size and depth of invasion of the carcinomas. Recognizing that H.pylori may disappear spontaneously from infected patients as mucosal atrophy and metaplasia progress, thereby reducing the amount of remaining normal epithelium that serves as bacterial habitat, a conditional logistic regression model was applied to these data, in which adjustment was made for extent of atrophic gastritis using PG I/II ratios. The result was a significant positive association between H.pylori infection and increased risk of gastric carcinoma in all cases combined (OR = 1.69; 95% CI, 1.01-2.81), for distal cancers (OR = 1.88; 95% CI, 1.07-3.31) and for carcinomas of intestinal type (OR = 3.76; 95% CI, 1.39-10.18).
Dr Fukuda concluded that H.pylori infection does increase the risk of gastric carcinoma. Because seronegative patients include both low-risk patients without infection and very high-risk patients with extensive mucosal atrophy and metaplasia, this risk is probably often seriously underestimated in cross-sectional studies of populations because of the disappearance of the bacteria from advanced cases. He suggested that this may help explain some of the discrepancy between the high risks found in some prospective studies in Western countries and the absence of increased risk indicated by some retrospective studies in developing countries.
Dr Naomi Uemura (Kure Kyosai Hospital, Kure) presented an exceptionally interesting clinical study: `H.pylori eradication inhibits the initial stages of growth of gastric cancer of the 'intestinal" type' (15). In Japan, where the incidence of gastric cancer is high, early carcinomas are often discovered by endoscopy and removed by endoscopic mucosal resection (EMR). These carcinomas often appear on a background of chronic atrophic gastritis accompanied by advanced intestinal metaplasia and synchronous multiple primary carcinomas often arise in these patients. The incidence of metachronous multiple carcinomas that appear within 5 years after EMR is 5-10%. Residual gastric mucosa after EMR for a first carcinoma is thus a high-risk microenvironment for development of second and additional primary cancers. Preliminary evidence has been obtained that eradication of H.pylori infection following EMR can prevent or delay emergence of second primary cancers. Eradication was attempted in 65 of 132 H.pylori-infected patients (44-85 years old, mean age 69 years; 97 males and 35 females) with gastric carcinomas that had been resected by EMR. After EMR, eradication was attempted with a regimen of omeprazole (20 mg daily for 4 weeks) and clarithromycin (400 mg daily for 2 weeks). After 4 weeks, a second regimen was administered to the 35 patients in whom eradication was incomplete: this consisted of omeprazole (20 mg/day), amoxicillin (1500 mg/day) and metronidazole (500 mg/day) for 2 weeks, which successfully eradicated the organism in all 35 patients. All 65 patients in whom eradication of H.pylori was achieved were then examined by endoscopy, with biopsies conducted every 6 months. Severity of neutrophil infiltration and intestinal metaplasia was assessed in H&E-stained specimens according to the Sydney classification. In the remaining 67 patients who did not receive any antibacterial medication, endoscopy was conducted every 6 months. Progress of patients was followed for an average of 24 months (maximum 48 months). In the 65 patients in whom eradication was successful, neutrophil infiltration disappeared in the antrum and body of the stomach, significant remission of intestinal metaplasia occurred and no new gastric carcinomas were detected. On the other hand, in six of the 67 patients (9%) without H.pylori eradication, a new early stage gastric carcinoma of intestinal type was detected within 3 years after EMR for a first cancer. There were no differences between these groups in age or sex distribution, histological type of gastric cancer, site of tumor occurrence and fraction of patients with multiple carcinomas detected at the same time. Treatment therefore not only improved clinical status in regard to gastritis and intestinal metaplasia, but appeared to inhibit growth of initial stages of the intestinal type of gastric carcinoma. Eradication therapy appears essential for patients after EMR and further clinical studies to confirm these observations are urgently needed.
Dr Daizo Saito (National Cancer Center Hospital, Tokyo) presented `H.pylori infection and gastric cancer: an introduction to the protocol of the Japanese intervention trial', which is supported by a Grant-in-Aid from the Japanese Ministry of Health and Welfare, within the Second Term Comprehensive 10-Year Strategy for Cancer Control. This study is designed to compare the development and progression of gastric mucosal atrophy and the incidence of gastric cancer in patients from 20 to 59 years of age without localized gastric lesions or ulcer scars, randomized into two groups, one that will undergo H.pylori eradication and an untreated control group. Initial eradication therapy will consist of triple therapy with lansoprazole (30 mg/day), clarithromycin (400 mg/day) and amoxicillin (1500 mg/day) for 1 week. If evaluation after 6 months indicates persistent infection after treatment, the same regimen will be repeated for 2 weeks. If this also fails, combination therapy with other drugs (e.g. plaunotol, ecabeto-Na, sofalcon) will be given. Although metronidazole is most frequently used for H.pylori eradication in the USA and in Europe, there have been reports that it is carcinogenic (16-18) and accordingly it was excluded from use in this trial. H.pylori infection will be established by histological examination of and bacterial cultures from endoscopic specimens taken from the greater curvature of the antrum and body of the stomach, without knowledge of the therapeutic regimen of the subjects. Serum samples will be assayed also for antibody to H.pylori and for pepsinogen, but these data will be only for reference, not for diagnosis. A single reference center will establish H.pylori infection, evaluate eradication results and diagnose mucosal atrophy. Considering the H.pylori infection rate in Japan, the rate of re-infection following eradication, the efficacy of eradication treatments, the incidence and rate of progression of atrophic gastritis, the risk of gastric cancer in H.pylori-infected patients and the incidence of gastric cancer in the general Japanese population, Dr Saito calculated that to achieve acceptable to ideal levels of confidence in the results, the necessary sample size is 102-736 cases when mucosal atrophy progression is the endpoint and 2342-7458 cases when the endpoint is comparison of gastric cancer incidence. Gastric mucosal atrophy will be evaluated in the fourth year after the end of the enrollment period and comparative gastric cancer incidence will be evaluated after 8 years of follow-up. Dr Saito emphasized that if this prospective, randomized intervention study can convincingly establish the causal link between H.pylori infection, atrophic gastritis (which is generally considered to be a precancerous condition) and gastric cancer, it will lead to the establishment of preventive measures for gastric cancer, which has a very high mortality in Japan and will therefore be extremely significant, both in Japan and elsewhere.
Gene diversity and virulence
The seminar then addressed the subject of determinants of virulence in H.pylori. It is now well known that many strains of the bacterium exist and that not all are equally virulent. A number of genes from H.pylori have been cloned and the gene products characterized. One gene, vacA, encodes a cytotoxin that corresponds to a protein of 87 kDa, which may be present in all strains. There is also a cytotoxin-associated protein of 120-128 kDa, the gene for which, cagA, has also been cloned and is present in 60-80% of H.pylori strains (19,20). Several presentations considered whether association of H.pylori with disease processes could be clarified by taking into consideration the presence or absence of these or other bacterial genes in specific strains of H.pylori.
Dr Martin Blaser (Vanderbilt University School of Medicine, Nashville, TN) discussed the question, `How does H.pylori infection promote the development of adenocarcinoma of the stomach?'. He referred to a case-control study of Japanese-American men in Hawaii (21), in which serum IgG to the cagA gene product had been measured in 103 H.pylori-infected men who had developed gastric cancer during a 21-year period of surveillance (1968-89) and in 103 infected men who did not develop cancer. In men with antibodies to the cagA product, the OR for developing gastric cancer was 1.9 (95% CI, 0.9-4.0), but for the intestinal type of cancer of the distal stomach, the OR was 2.3 (95% CI, 1.0-5.2) and significant. There was no significant effect on risk for the diffuse type. Further examination of that study population showed that being born in a large sibship and at higher birth order further increased cancer and gastric ulcer risk among those who were infected with H.pylori (22). This conclusion was based on a study of 102, 147 and 64 men who developed carcinoma of the distal stomach, gastric ulcer and duodenal ulcer, respectively and a matched control for each. H.pylori-infected men from larger sibships (OR = 2.06) and of higher birth order in relation to firstborn (OR = 1.67) were at increased risk for gastric cancer and for gastric ulcers (OR = 1.64) but not duodenal ulcers. These data suggest that acquiring H.pylori in early childhood increases the risk of developing gastric ulcers and gastric cancer many decades later. In another study the relationship between cagA status and development of atrophic gastritis was investigated in a cohort of 58 infected subjects during a mean follow-up period of 11.5 years (range 10-13 years), using endoscopic biopsy specimens and ELISA for anti-cagA protein IgG (23). At the initial visit, moderate to severe atrophic gastritis was observed in 8/24 (33%) cagA-positive patients and in 6/34 (18%) cagA-negative patients. During follow-up, atrophic gastritis developed in 8/16 cagA-positive patients (50%) vs 8/28 cagA-negative patients without atrophic gastritis at original presentation. In five of the former and one of the latter, atrophic gastritis was accompanied by the development of intestinal metaplasia. At the end of follow-up, 15/24 cagA-positive patients (62%) had atrophic gastritis vs 11/34 cagA-negative patients (32%), a significant difference in incidence (p = 0.02, Fisher's exact test; OR = 3.48; 95% CI, 1.02-12.18). These data support the conclusion that infection with cagA-positive H.pylori strains is associated with increased risk of development of atrophic gastritis and intestinal metaplasia.
Intensity of colonization of gastric mucosa varies between cagA-positive and cagA-negative strains; the former achieve four times the bacterial density of the latter and in infections with cagA-positive strains the severity of both acute and chronic gastritis is greater. The function of cagA protein remains unknown, but cagA-positive strains of H.pylori have been shown effectively to induce cytokines including interleukin-8 (IL-8) in gastric epithelium. Knock-out of cagA has no effect on capacity to induce IL-8; H.pylori (DcagA) strains retain this ability. Further analysis of the bacterial chromosome region surrounding the cagA region has resulted in discovery of two additional genes, picA and picB (named for their function, permitting induction of cytokines). picB appears homologous to secretion factors for toxins in other genera of bacteria and knockout of either gene produces DpicA and DpicB strains of H.pylori that lack the capacity to induce cytokines. Further analysis of the cag region of the bacterial chromosome should clarify precisely which genes determine pathogenicity of specific H.pylori strains.
Dr Julie Parsonnet (Stanford University Medical School, Stanford, CA) presented `Determinants of gastric adenocarcinoma in persons with H.pylori infection'. She discussed a nested case-control study in California to evaluate cagA seropositivity, PG I and gastrin levels, ABO blood group and level of education (as an index of socio-economic status) as markers of increased risk of atrophy, metaplasia and cancer in H.pylori-infected individuals. From among 129 000 persons who had enrolled in a multiphasic health program between 1964 and 1969, 85 H.pylori-infected gastric cancer patients (62 of intestinal type, 23 of diffuse type), 90 infected individuals without cancer and 70 uninfected gastric cancer cases and corresponding controls were identified. IgG to the cagA gene product was quantified by ELISA. Low PG I serum levels (<50 mg/ml) were considered diagnostic of corpus atrophy. Gastrin in serum was determined by RIA. After adjusting for age, sex and race, infected gastric carcinoma cases were 3.2 times more likely than infected controls to have antibodies to the cagA gene product (95% CI, 1.6-6.5). This association was significant for the intestinal type (OR = 3.7; 95% CI, 1.7-8.3) but not for the diffuse type (OR = 2.2; 95% CI, 0.8-6.4). Corpus atrophy significantly increased risk for both cancer types and lessened the magnitude of association between cagA and cancer by 25%. When analyzed as a continuous, log-transformed variable, there was no association between gastrin levels and cancer. Among persons with corpus atrophy, however, gastrin levels in the lowest quartile (<26.6 pg/ml) appeared to be protective against gastric carcinoma (OR = 0.1, p value for interaction = 0.08). In persons with normal pepsinogen levels, low gastrin was not protective against cancer (OR = 1.0). Infection with the cagA-negative (Type II) phenotype of H.pylori was only modestly and not significantly associated with increased risk of gastric carcinoma (OR = 2.0; 95% CI, 0.8-4.6). Neither educational level nor ABO blood group was associated with malignancy in infected persons.
Dr Parsonnet concluded that, among persons infected with H.pylori, the cagA-positive (Type I) phenotype increases cancer risk at least threefold, whereas it is uncertain whether the cagA-negative (Type II) phenotype has any role in carcinogenesis. Since the proportion of H.pylori phenotypes varies among population groups, this may explain why high prevalences of infection with H.pylori are not always in accord with gastric cancer rates. Persons with H.pylori who develop corpus atrophy but retain normal gastrin levels appear to be particularly predisposed to malignancy. The tendency of Type I strains to cause cancer, however, can only be partially explained by their induction of corpus atrophy.
Dr Toshiro Sugiyama (Sapporo Medical University, Sapporo) presented `Diversity of H.pylori strains and gastrointestinal diseases'. He described the development of a monoclonal antibody to and cloning of the gene for a 58-60 kDa antigen, CP2, which is present in all clinical isolates of H.pylori and is partially homologous to catalase from other species (e.g. 57.9% homology to Bacillus subtilis catalase). The catalase gene is probably essential to H.pylori to protect against peroxidation by neutrophils and therefore is likely to be a good marker to follow the efficacy of antibacterial therapy.
Dr Sugiyama also devised primers to amplify non-repetitive DNA sequences from the cagA gene of H.pylori as one marker to characterize bacterial strains isolated from cases of various gastric disease processes. cagA sequence was successfully amplified from 14/24 chronic gastritis isolates (58%), 7/10 gastric carcinoma isolates (70%), 17/19 gastric ulcer isolates (90%) and 13/15 duodenal ulcer isolates (87%). He also conducted RFLP analysis of the cagA gene in HaeIII digests of H.pylori DNA. A 3.0 kb fragment (cagA-l) was observed in 9/10 duodenal ulcer strains, but a 0.9 kb fragment (cagA-s) was seen in 7/10 gastric ulcer strains. This suggested that RFLP analysis of cagA may be useful to characterize and distinguish H.pylori strains that vary in disease potential.
Dr Toshio Fujioka (Oita Medical University, Oita) discussed `The role of H.pylori infection in the progression of atrophic gastritis: experimental and clinical studies'. He investigated the Japanese macaque, Macaca fuscata, as an experimental host for H.pylori and an animal model for H.pylori-associated gastric disease (24). Using adult animals, he confirmed by pre-examination that they were not already infected with H.pylori, then gave antibiotics to clear gastric infections with other spiral bacteria and finally inoculated the animals by endoscope with a mixture of four isolates of H.pylori that had been isolated from human patients, two of whom had gastric ulcers and two duodenal ulcers. Only one of these four H.pylori strains, from a patient with duodenal ulcer, could colonize these animals, as confirmed by culture and histology. This strain was cagA+, vacA+ and vacuolating cytotoxin positive. Marked leukocytic infiltration occurred in gastric mucosa of infected animals. Six infected and seven control monkeys were followed for up to 5 years after inoculation. Severity of gastritis was evaluated endoscopically and endoscopic mucosal biopsies from the antral mucosa of infected and control monkeys were taken annually for histological examination and Ki-67 immunohistochemistry. Glandular height and the thickness of the proliferative zone of Ki-67 positive cells were used as indexes of gastric mucosal atrophy and epithelial cell proliferation, respectively. Macroscopic gastritis was maximal 1 week after inoculation and then resolved by 1 month. Histologically demonstrable gastritis, however, persisted throughout the 5-year period of observation: glandular height in the antral mucosa significantly decreased to a stable level and the thickness of the proliferative zone significantly and progressively increased in infected animals in comparison with controls. The average number of parietal cells decreased progressively in infected animals during the 4 years following inoculation. Severe atrophy was seen at 3 years and reduced acid secretion was documented between 3.5 and 4.5 years after inoculation. H.pylori infection leads to chronic atrophic gastritis in M.fuscata, with reduced acid secretion and increased epithelial cell proliferation that resembles chronic atrophic gastritis in infected humans. Further observations will be required to determine whether this condition in the monkeys will progress to intestinal metaplasia and to gastric carcinoma.
Immunoresponse and Vaccination
The seminar next addressed the subject of immune responses to Helicobacter and possibilities for vaccine development. The development of effective immunization strategies to prevent or control H.pylori infection in humans has received increasing attention in recent years. This approach is especially important because antibiotic therapy of established infections is not always successful and is not often utilized in children. Moreover, it is possible that antibiotic-resistant strains of H.pylori may emerge. A number of laboratories have begun to study the feasibility of prophylactic and therapeutic vaccination for the prevention and treatment of Helicobacter infections, using animal models and both H.pylori and other Helicobacter species.
Dr Norio Matsukura correlated presence of IgA antibody to H.pylori with histological features of gastric mucosa in gastrectomy specimens (25). Areas of intestinal metaplasia were identified by the TesTape method and IgA was measured by ELISA in non-metaplastic, metaplastic and neoplastic tissue. Anti-H.pylori IgA was identified in 6/14 (43%) specimens of non-metaplastic gastric mucosa from the antrum and in 14/23 (61%) from the body. However, only 6/19 (32%) specimens of complete intestinal metaplasia and 2/7 specimens (29%) of the incomplete type were positive and anti-H.pylori IgA was identified in only 1/6 (17%) specimens of duodenal mucosa and 0/17 (0%) specimens of neoplastic tissue.
Dr Steven Czinn (Rainbow Babies and Childrens Hospital, Cleveland, OH) discussed `H.pylori: host immune response and vaccine development' (26). Using H.felis infection of mice as a model system (Note: see final presentation by Dr James Fox), he investigated prophylactic immunization by oral administration of antigen. Germ-free mice received 2 mg of sonicated H.felis cells plus 10 mg of cholera toxin, a known mucosal adjuvant, four times during one month. One week after the final immunization, control and immunized animals were challenged with viable H.felis and then killed at various intervals thereafter. Gastric tissue samples were tested for urease activity and cultured for the presence of H.felis. Serum and gastric and intestinal secretions were collected and assayed by ELISA for the presence of anti-H.felis antibodies. A significant IgA and IgG anti-Helicobacter response was generated in immunized mice, both in serum and in gastric and intestinal secretions. Moreover, significant protection against challenge with viable H.felis was seen in immunized mice [82% (28/34) vs 11% (4/36)].
Dr Czinn also briefly discussed therapeutic immunization of ferrets that were chronically infected with H.mustelae using oral immunization with H.pylori urease and cholera toxin, which succeeded in curing 7 of 23 ferrets. He suggested that if such studies are consistently effective, this approach might lead to immunotherapy for eradication of established H.pylori infection in humans. Dr Czinn noted that both natural infection and immunization result in systemic and mucosal antibody responses of comparable magnitude. However, there seems to be a qualitative difference in the immune response to immunization which results in protection. Specifically, antibodies against urease and perhaps other antigens, appear to be prominent following immunization and may play a key role in preventing or curing Helicobacter infections.
Finally, Dr Czinn described cellular immune responses following infection and immunization with Helicobacter (27). Two types of cell-mediated immune responses occurred, depending on the nature of the antigen preparation. A general, Helicobacter-independent response occurs in all mice, including those given adjuvants only, which is directed towards certain antigens, including urease and heat shock proteins. The second response is Helicobacter specific and occurs only in mice previously exposed to Helicobacter antigens either by immunization or infection and in both cases has a Th1 phenotype, which promotes a delayed-type hypersensitivity response in the stomach. However, suppression of the Th1 response unmasks the presence of a subpopulation of Th2 cells, whose role in protection against Helicobacter infection requires further study. The importance of cellular immunity for successful vaccination strategies also remains to be clarified.
Dr Bruce Dunn (C.J. Zablocki Veterans Administration Hospital, Milwaukee, WI) described an approach to vaccine development against H.pylori in humans. Recombinant urease apoenzyme is being used in Phase I clinical trials and was chosen because much of this urease activity is located on the bacterial surface. Urease and HspB (a GroEL heat shock protein homolog) are located strictly within bacterial cytoplasm in early logarithmic phase cultures, but in late logarithmic phase cultures, urease and HspB become associated with the bacterial surface in a novel manner (28). These cytoplasmic proteins are released by bacterial autolysis and become adsorbed on the surface of intact bacteria. This strategy represents a unique mechanism for surface localization of proteins. Immunoelectron microscopy of infected human gastric biopsies shows that a fraction of both urease and HspB is associated with the bacterial surface, so this novel mechanism does function in vivo. Such `altruistic lysis', in which autolysis of a fraction of the bacterial population presumably benefits the remaining viable bacteria, would (a) allow adsorption of the essential enzyme urease on the surface of viable bacteria, helping promote survival in the harsh gastric milieu; (b) help to explain how the non-invasive bacterium H.pylori can present virulence factors and antigens to the gastric mucosa and the host immune system; (c) possibly explain why the immune system is unable to eradicate H.pylori, due in part to `immune overload' resulting from continual release of cytoplasmic proteins, and (d) potentially influence the choice of antigens for oral immunization of humans against H.pylori.
Eradication: Gastric lymphoma
Turning from gastric carcinoma to gastric lymphoma, Dr Hiroyuki Ono (National Cancer Center Hospital, Tokyo) discussed `Association between H.pylori infection and lymphoproliferative disorders of the stomach', including regression of MALT (mucosa-associated lymphoid tissue) lymphoma after eradication of H.pylori infection. The gastrointestinal (GI) tract is the most common site of origin of primary extranodal non-Hodgkin lymphoma (NHL), accounting for up to 50% of all cases and within the GI tract the stomach is the most commonly affected organ (29). Since the stomach lacks lymphatics, specialized lymphoid tissue (MALT) has evolved to protect the GI mucosa. Low-grade gastric NHL, characterized by an indolent clinical course and prolonged confinement to the site of origin, arises from MALT (30). Diagnostic criteria in use at the National Cancer Center Hospital distinguish four categories of lymphoid lesion in the stomach: high-grade lymphoma (HG), with mucosal infiltration; low-grade MALT (LG-MALT) lymphoma, with a lymphoepithelial histological pattern that is easily identified; atypical lymphoid proliferation, a borderline lesion between reactive lymphoid tissue and low-grade neoplasia; and gastritis, with purely inflammatory/reactive lymphoid tissue. Several epidemiological studies have associated H.pylori infection with gastric lymphoma and regression of LG-MALT gastric lymphoma after eradication of H.pylori has been reported (31,32).
To confirm these reports, a case-control study was conducted in Japan and a prospective study on the effects of H.pylori eradication on the clinical course of LG-MALT lymphoma was carried out to determine the utility of this approach in the clinical management of this disease. In the case-control study, 35 patients diagnosed with primary gastric NHL during the period 1988-95 were selected as the case pool and 1295 outpatients with no prior history or clinical diagnosis of cancer who were seen for the first time during this period served as the control pool. Thirty case-control pairs matched by age and sex showed a suggestive positive association between H.pylori infection and gastric NHL (OR = 2.76; 95% CI, 0.81-9.44). In nine cases of LG-MALT lymphoma and in one case of atypical lymphoid proliferation, endoscopic gross and histological changes and clonality of the lesions were followed after H.pylori eradication. MALT lymphomas were defined as lesions with diffuse and abundant centrocyte-like cells, infiltration and prominent lymphoepithelial lesions according to the criteria of Wotherspoon et al. (32) and Isaacson and Wright (30). Patients received triple therapy with lansoprazole (30 mg/day), plaunotol (240 mg/day) and amoxicillin (1500 mg/day) in combination for 2 weeks and double therapy with lansoprazole and plaunotol was continued for 2 weeks more. At each endoscopic examination before and after treatment, two biopsy specimens were taken from the greater curvature of the antrum and the body of the stomach for H.pylori cultures and at least five specimens from the area of the lymphoid lesion were taken for histology and two for molecular genetic analysis. After conclusion of treatment, this procedure was repeated every 3-6 months during follow-up. Efficacy of eradication therapy was assessed by bacterial cultures. In the single patient with atypical lymphoid proliferation and in five patients with LG-MALT lymphoma, histological diagnosis changed to `chronic gastritis', and endoscopic findings improved in four of these six cases. In two cases, the previous monoclonal pattern of immunoglobulin gene rearrangement was shown by PCR to change to a polyclonal pattern after H.pylori eradication. Although one case with LG-MALT lymphoma deteriorated by endoscopy and was treated by gastrectomy, that patient had no lymph node involvement or involvement of other organs.
Dr Ono also retrospectively reviewed the clinical course of eight cases of MALT lymphoma, one HG and seven LG-MALT and four cases of atypical lymphoid proliferation that had been followed for periods ranging from 2 months to more than 7 years. None of these cases showed histological progression to higher-grade disease; four did not change, while the one HG and three LG-MALT lymphomas and all four cases of atypical lymphoid proliferation improved to histological gastritis during this period. In another study to clarify the malignant potential of HG and LG-MALT lymphomas, Dr Ono found that 0/13 LG-MALT and 6/17 HG cases metastasized to regional lymph nodes. He concluded that the favorable clinical behavior of gastric LG-MALT lymphomas suggests that they are not true malignancies, but `pseudo-lymphomas' or a reactive process in response to H.pylori infection. They do appear highly associated with H.pylori infection, however, and he proposed that the following guideline should be adopted for clinical management of MALT lymphomas: for H.pylori-infected patients with lymphoproliferative disease but no metastasis beyond the stomach, eradication therapy should be given and if subsequent endoscopic biopsies show only gastritis, then no further treatment is indicated beyond close follow-up. However, since the natural course of MALT lymphomas is slow and unpredictable, even in patients whose MALT lymphoma persists in spite of successful H.pylori eradication or in those who are negative for H.pylori from the beginning, close observation should be considered in preference to surgery and gastrectomy should be performed only in cases of MALT lymphoma.
Helicobacter species other than H.pylori
Turning to Helicobacter species other than H.pylori, Dr Bruce Dunn presented `Induction of gastric dysplasia by H.felis in p53-deficient mice'. As up to 50% of gastric carcinomas are associated with loss of function of the p53 tumor suppressor gene, the availability of genetically engineered mice that lack normal p53 allows testing the hypothesis that gastric epithelial cell hyperproliferation induced by Helicobacter infection coupled with lack of p53 gene activity would lead to the development of significant premalignant or malignant gastric lesions. Homozygous p53 gene knock-out mice (TSG-p53) were infected with H.felis or given culture medium (Brucella broth) alone. Infected mice killed as early as 17 days after infection showed foci of active gastritis. Mice killed 60-130 days after infection showed inflammation and the presence of H.felis within gastric antral glands and some mice developed focal atrophic changes. In some of them, foci of moderate dysplasia were identified. No such alterations were observed in uninfected mice. H.felis infection can induce gastric atrophy and dysplasia in at least a fraction of TSG-p53 mice as early as 60 days post-infection.
Dr Jerry Rice presented `H.hepaticus and its association with hepatitis and hepatocellular tumors in mice' (4). In the autumn of 1992, a novel form of chronic active hepatitis of unknown etiology was discovered in mice at the National Cancer Institute's Frederick Cancer Research and Development Center (NCI-FCRDC) at Frederick, MD. A remarkably high incidence of hepatocellular neoplasms including both adenomas and carcinomas, often multiple, developed in affected animals by 64-77 weeks of age. The disease entity was originally identified in mice that were untreated controls in a long-term carcinogenicity study. These mice were of strain A/JCr that historically had a very low incidence of both liver tumors and any form of hepatitis. By use of the Steiner modification of the Warthin-Starry silver impregnation for bacteria in formalin-fixed tissue sections, it was found by light microscopy that liver tissue from mice with hepatitis contained bodies the size of bacteria that were singly distributed within hepatic parenchyma. A previously unknown microaerophilic bacterium was cultured from fresh liver tissue from mice with hepatitis and was described as a novel species of Helicobacter on the basis of its morphological and biochemical characteristics and the base sequence of its 16S rRNA gene. The bacterium is motile and Gram negative, 0.2-0.3 mm in diameter, 1.5-5.0 mm in length and spirally curved in shape with one to several spirals. It has bipolar sheathed flagella, one at each end, but lacks the periplasmic fibrils that envelop the bacterial cells in other murine Helicobacter species. Because of its association with inflammatory and neoplastic liver disease, it was named H.hepaticus. Transmission electron microscopy has revealed that it is localized preferentially in hepatic bile canaliculi in infected livers; it has not been observed within hepatocytes. It has been isolated thus far only from mice and not from other rodent species, e.g. rats, guinea pigs and Syrian hamsters. It appears to cause liver disease only in mice of certain strains, including A, DBA and C3H, and to colonize the cecum and colon without invading the liver in certain other strains such as C57BL. Male mice of susceptible strains are much more highly susceptible than females. While single-agent antibiotic therapy is ineffective in eradicating the bacteria from the gastrointestinal tract, either amoxicillin or tetracycline, in combination with metronidazole and bismuth, has been found to eradicate H.hepaticus from the gastrointestinal tract when given orally by gavage for a period of 2 weeks. H.hepaticus-associated chronic active hepatitis and its progression to hepatocellular neoplasia represent a new model to study mechanisms of carcinogenesis by this genus of bacteria. Dr Rice emphasized that H.hepaticus infection in mice constitutes the only other association that is known to exist naturally between a persistent infection by a specific bacterial species and tumor development in the infected tissue, other than the association of H.pylori with human gastritis and gastric neoplasia. Study of the mechanism by which chronic H.hepaticus-associated hepatitis proceeds to development of hepatocellular tumors in mice may provide insight into the etiologic role of H.pylori in human stomach cancer and gastric lymphoma.
Continuing discussion of this newly discovered Helicobacter species, Dr Lucy Anderson (National Cancer Institute, Frederick, MD) presented `Liver carcinogenesis by H.hepaticus: investigations of possible mechanisms'. She described several experimental hypotheses for tumor development in H.hepaticus-infected mice with hepatitis. No mutagenic substance produced by the bacteria has been found. The possibility was investigated that nitric oxide is released in response to infection and might cause genotoxic damage, either directly by deamination of 5-methylcytosine residues at CpG sites in DNA or indirectly by nitrosation of endogenous substrates to generate carcinogenic nitrosamines. Several tests of this hypothesis have given negative results. The most promising line of investigation has involved the production of reactive oxygen species (ROS) in response to infection, with resulting oxidative genotoxic damage to DNA. The promutagenic ROS product, 8-hydroxydeoxyguanosine (8-OH-dG), was found to be significantly elevated in infected livers from an early stage, with further increases as hepatitis progressed. A significant threefold increase in 5-hydroxy-5-methylhydantoin, an oxidation product of thymine, was also detected. ROS might in principle be produced by hepatocytes or by activated Kupffer cells or leukocytes. Certain cytochrome P450 isoforms that are known to be major sources of intrahepatocellular ROS are known to increase in liver tissue in other hepatitis/tumorigenesis systems. Significant increases in P450 1a1 and 1a2 have been found in H.hepaticus-infected mouse liver by both enzyme assays and immunohistochemistry early in the infection and in P450 2a5 late in infection, when grossly visible tumors have appeared. As ROS have also been implicated as contributing to nongenotoxic mechanisms of tumor promotion in other rodent model systems, including hepatocarcinogenesis, the ability of H.hepaticus infection to promote nitrosamine-initiated liver tumors was tested in strain A/JCr male mice and was found to depend on the nitrosamine used for initiation, with positive results in mice given N-nitrosodimethylamine (NDMA) but not in mice given N-nitrosodiethylamine. Among NDMA-initiated mice, 10/60 infected mice were moribund before 36 weeks, versus 0/60 non-infected mice (p = 0.001) and eight of the 10 moribund mice presented non-hepatocellular neoplasms, including five lymphomas, in addition to liver tumors. For the liver, at 36 weeks the multiplicity of liver tumors was significantly greater in infected mice (4.6 ± 0.8) than in controls (2.4 ± 0.6, p = 0.04). These results suggest that H.hepaticus infection can promote hepatocarcinogenesis and that this effect may not be restricted to hepatocytes.
Finally, Dr James Fox (Massachusetts Institute of Technology, Cambridge, MA) presented `The expanding genus of Helicobacter: pathogenic and zoonotic potential'. Since the description of H.pylori, the type species of the genus, in 1983, an additional 12 species had at the time of this seminar been formally described and named and more are constantly being reported. He described in detail some additional species of Helicobacter that have been isolated from various mammalian host species and their association with disease and emphasized that spiral bacteria are common in the stomach and widespread among mammals (29). Although there are dramatic differences in morphology from species to species, all have in common their microaerophilic growth habit, certain biochemical characteristics that often include strong urease activity and varying degrees of homology in the 16S rRNA gene sequence. Toxin production has been well characterized only in H.pylori, but is now being studied in other species. Several Helicobacter species other than H.pylori have been isolated from the human intestine and three of these, H.cinaedi, H.fennelliae and H.canis, have been isolated from human blood. H.mustelae was isolated from ferrets with gastric ulcers and the parasite-host relationship has been found to resemble closely H.pylori infection in humans: infection causes the sequence diffuse antral gastritis -> duodenal ulcer, MAG -> peptic ulcer. H.mustelae establishes a persistent infection in ferrets, which naturally develop adenocarcinoma of the stomach; virtually all ferrets kept as pets in the USA are infected. Cheetahs, which are a severely inbred species, have been found to suffer from severe chronic gastritis with ulceration and to harbor a number of Helicobacter species including one specific to cheetahs, H.acinonyx. Although H.pylori was originally found to colonize monkeys and gnotobiotic pigs, but not rodents or carnivora, under experimental conditions, it has more recently been shown that cats are easily infected with cagA-positive strains of H.pylori, with persistent colonization of gastric mucosa and gastritis. Cats, dogs and non-human primates harbor spiral microorganisms in their stomachs, including H.heilmannii, which occurs naturally in dogs and cats, in which peptic ulcers are rare, and H.felis in cats. However, cats from one commercial breeder were found to have gastritis and MALT tissue with lymphoid follicles in the stomach and were infected with an organism determined to be H.pylori, which was present in saliva of 6/12 cats and in gastric washings from 11/12. How general such infection is in domestic cat populations is not known, but this observation does suggest that zoonotic transmission of H.pylori from domestic animals to man is possible and the likelihood is great that as yet undiscovered species of Helicobacter exist whose host range may include human beings.
Problems to be solved
In conclusion, evidence of a causal association between infection of human beings with H.pylori and increased risk of gastric carcinoma and lymphoma continues to increase and gives additional support to the concept that these cancers may be preventable by timely intervention with antibiotic therapy or, possibly in the future, by prophylactic or therapeutic vaccination. The large-scale intervention trial now under way in Japan may provide conclusive evidence for a role of H.pylori infection in causing these cancers. However, many aspects of the host-parasite relationship are not understood and require clarification, including better definition of host factors that increase susceptibility of some human populations and further definition of bacterial genes that confer pathogenicity. Moreover, the rapidly increasing number of recognized Helicobacter species and the association of some of these with neoplasms in tissues other than the stomach in non-human hosts raises the possibility that some human diseases besides gastric carcinoma and lymphoma may turn out to be related to infections with species of Helicobacter that have not yet been discovered. The increasing recognition that the host range for individual Helicobacter species may extend to several mammalian species and that humans may be among these also indicates the possibility of zoonotic transmission of these pathogens from animals to humans. Future research on the bacterial genus Helicobacter and its role in human disease must recognize these complex host-parasite relationships.
All participants strongly endorsed Dr Sugimura's suggestion that American and Japanese researchers should mutually exchange research material, including cultures of specific strains of H.pylori and of other Helicobacter species, DNA probes and antisera, and should also exchange both reprints of published research and research personnel to give increased impetus to further development of cooperative Japanese-American research on Helicobacter and diseases related to infection with these bacteria.
References
For reprints and all correspondence: Daizo Saito, Division of Gastroenterology, National Cancer Center Hospital, 1-1, Tsukiji 5-chome, Chuo-ku, Tokyo 104, Japan
Abbreviations: MAG, multifocal atrophic gastritis; PG, pepsinogen; EMR, endoscopic mucosal resection; vacA, vacuolating cytotoxin; cagA, cytotoxin associated protein; IL-8, interleukin-8; CP2, catalase protein 2; MALT, mucosa-associated lymphoid tissue; LG-MALT, low-grade MALT; HG, high-grade lymphoma; NHL, non-Hodgkin lymphoma; ROS, reactive oxygen species; 8-OH-dGh, 8-hydroxydeoxyguanosine; NDMA, N-nitrosodimethylamine
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 19 May 1998
Copyright© Japanese Journal of Clinical Oncology, 1998.
CiteULike 
Connotea 
Del.icio.us What's this?
This Article 
Extract
Alert me when this article is cited
Alert me if a correction is posted
Services
Email this article to a friend
Similar articles in this journal
Similar articles in ISI Web of Science
Similar articles in PubMed
Alert me to new issues of the journal
Add to My Personal Archive
Download to citation manager
Search for citing articles in:
ISI Web of Science (1)
Request Permissions
Google Scholar
Articles by Saito, D
Articles by Rice, J.
Search for Related Content
PubMed
PubMed Citation
Articles by Saito, D
Articles by Rice, J.
Social Bookmarking
What's this?