Helicobacter pylori and gastric adenocarcinoma
Corresponding author and reprint requests: V. Herrera, Division of Infectious Diseases, Department of Medicine, Stanford University School of Medicine, Grant Building, Room S125, 300 Pasteur Drive, Stanford, CA 94305, USA
Gastric cancer is the second most common cause of cancer death worldwide. A large body of evidence supports a causal role of Helicobacter pylori in the majority of gastric malignancies. Great strides have been made in understanding the pathogenesis of this relationship, but much remains to be learned. Moreover, because of the high prevalence of infection, the lack of definitive trials, and the challenges of H. pylori treatment, there remains no consensus on the role of routine screening and treatment of this infection to prevent cancer. This article reviews the current knowledge on H. pylori and gastric cancer and presents some of the clinical and public health challenges associated with this pathogen.
Gastric cancer is the second leading cause of death due to cancer worldwide. In 2002, 934 000 cases were recorded (9% of new cancers) with 700 000 deaths . Yet, despite these daunting numbers, which, in part, are consequent to increased human longevity, the rate of gastric cancer has actually been declining from the early 20th Century onward. The precipitous drop in incidence, particularly in industrialized nations, has been the topic of extensive investigation. Finally, the reason for this declining rate has now become clear; it is a result of the diminishing prevalence of Helicobacter pylori infection, an organism that, until very recently, was part of the normal flora of humans.
Interest in H. pylori as a cause of cancer began after the pioneering discoveries of Marshall and Warren in the 1980s . Prior to the discovery of the organism, it was known that gastric adenocarcinomas typically arose in areas of gastritis. When the relationship between H. pylori and chronic gastritis was established, investigators began to take interest in the causal role of H. pylori in gastric cancer. The first studies to examine the association between H. pylori and gastric cancer were ecological, comparing the regional prevalence of H. pylori with the incidence of gastric cancer . Subsequently, numerous observational studies of diverse design confirmed these findings. In 1994, the International Agency for Research on Cancer declared H. pylori to be a type I carcinogen, or a definite cause of cancer in humans .
We discuss the epidemiological and pathophysiological data that led to the consensus that H. pylori is one of the world’s most important causes of cancer.
H. pylori is a Gram-negative, spiral-shaped bacterium that is characterized by its many unipolar flagella, which give it corkscrew-like motility, and its prodigious production of urease. Unique among bacteria, it finds a niche in both the antral and fundic mucosa of the stomach under the mucus gel. The presence of infection is universally associated with chronic and acute inflammation and, more variably, with other gastric lesions, including lymphoid follicles, atrophic gastritis and intestinal metaplasia. Treatment with antimicrobial agents causes inflammation to regress over time . With relapse of infection, the gastritis is again observed .
H. pylori is typically acquired during childhood and causes lifelong infection thereafter . Although previously almost universal in humans, currently ‘only’ half of the world’s population is infected with H. pylori. Transmission is largely from person to person via the faecal–oral or the gastric–oral route within families, particularly in settings of poor sanitation and hygiene . The prevalence of infection varies worldwide, with continued hyper-endemicity in developing countries but a markedly lower prevalence in developed countries . H. pylori is now rare in native-born and middle- or upper-class children of Western Europe , North America , Oceania  and Japan .
Epidemiological Links between Gastric Cancer and H. pylori
Subsequent to its discovery in the early 1980s, over 1000 studies have been conducted on H. pylori and its association with cancer, including observational studies (ecological, case–control and cohort), clinical trials of H. pylori eradication, pathological studies and animal models. The results have been overwhelmingly in favour of a link between infection and malignancy. Among the many observational studies in humans, case–control studies indicate the lowest risk for cancer (1.8-fold increase) . It is now understood that these studies underestimate the true risk due to loss of H. pylori as the mucosa undergoes malignant transformation. Nested case–control studies indicate higher relative risks in meta-analyses . However, the most compelling observational evidence of an association between H. pylori infection and gastric cancer comes from longitudinal cohort studies. In a large prospective trial conducted in Japan, 36 out of 1246 infected individuals developed gastric cancer compared to none of 280 uninfected participants (infinite OR) . A prospective study of 1225 Taiwanese patients confirmed this ‘infinite’ OR (p 0.015) .
Not all H. pylori are alike, however, and the epidemiological story is complex. Individuals with antibodies to H. pylori’s CagA protein (a marker for the more inflammatory and virulent strain bearing a pathogenicity island of genes) have a particularly high risk of cancer. A meta-analysis of studies shows that CagA-positive strains increase the risk of noncardia gastric cancer two-fold compared to CagA-negative strains . Moreover, gastric cancer, similar to H. pylori, is heterogeneous, with two histological types predominating: the intestinal-type and the diffuse type. Although H. pylori has been linked to both histological types, CagA appears to enhance the risk of the intestinal type that arises in the setting of inflammation, atrophic gastritis and intestinal metaplasia, but not the risk of the diffuse type that appears to stem from e-cadherin mutations .
Definitive experimental proof that H. pylori causes cancer in humans lags behind the observational science, largely because the studies are so difficult to undertake. Randomized clinical trials of secondary cancer prevention through H. pylori eradication require both a large number of subjects and many years of follow-up . To obviate these problems, a number of investigators have used gastric cancer precursors (i.e. multifocal atrophic gastritis and intestinal metaplasia) as surrogates for measuring the effects of H. pylori eradication on the malignant process. Overall, these studies suggest improvement of pre-neoplastic lesions, although the change is slow and not universally observed . Indeed, in some studies, continued progression of pre-neoplastic conditions has been substantial despite H. pylori eradication. Complicating matters further, the one randomized clinical trial concerning cancer prevention completed to date did not show a significant association between H. pylori eradication and disease prevention, although there was a trend towards benefit, particularly in young patients without pre-neoplastic conditions . By contrast, a randomized, open-label trial in 544 Japanese patients treated with endoscopic resection of early gastric cancer showed significantly lower rates of secondary, metachronous cancers in those who received ‘adjuvant’H. pylori eradication therapy compared to those who did not (OR 0.35, 95% CI 0.16–0.78) . Again, not all treated patients benefited. Thus, although the experimental data in humans are thin, they do support the notion that H. pylori causes cancer and that its treatment provides some benefit.
Based on the amalgam of observational and experiment studies, the attributable risk of gastric cancer in the population (i.e. the proportion of gastric cancer in the population that would not occur were the H. pylori not to exist) has been estimated to be 75% . If this is accurate, H. pylori would be responsible for as many as 5.5% of all cancers, making it the leading infectious cause of cancer worldwide and second only to smoking as a defined cause of malignancy.
Mechanisms of Carcinogenesis
Although mechanisms of H. pylori-induced carcinogenesis are only beginning to be understood, inflammation is the most commonly cited factor in the carcinogenic process. Inflammation is thought to induce cancer by increasing production of free radicals , increasing apoptotic and necrotic epithelial cell death  and augmenting cell proliferation . To compound these pro-carcinogenic processes, H. pylori has been noted to reduce DNA repair in vivo and in vitro . The importance of inflammation as a risk factor is bolstered by three complementary observations: first, that the bacterial strains that induce the most inflammation are most closely linked to malignancy ; second, that pro-inflammatory host cytokine polymorphisms increase cancer risk , and third, that nonsteroidal anti-inflammatory agents appear to decrease the risk of cancer .
Mechanisms other than inflammation have been posited for H. pylori-related carcinogenesis. H. pylori directly interacts with epithelial cells, resulting in protein modulation and gene activation . Strains that contain the pathogenicity island of genes have a particularly intimate relationship with the host. These strains produce a type IV secretion system through which the CagA protein is injected into gastric epithelial cells. There, the protein becomes phosphorylated by members of the Sac family of kinases, implicated in other malignancies. The phosphorylated CagA then activates the eukaryotic phosphokinase, SHP2, as well as extracellular signal-regulated kinase (a member of the mitogen-activated protein kinase family). The pathogenicity island also results in a ‘hummingbird’ epithelial phenotype, epithelial motility, loss of gap junctions, and eventual epithelial cell death .
Most recently, much attention has been given to the relationship between H. pylori, stem cells and cancer. Some have proposed that H. pylori preferentially damages parietal cells, thereby altering the maturation process of epithelial stem cells . Others report that inflammation related to H. pylori recruits peripheral- or bone-marrow derived stem cells to the gastric mucosa, which then transform into the malignant clone. The strongest evidence of this comes from experiments performed by Houghton et al. , who identified bone-marrow derived stem cells as the cells of malignant origin in C57BL/6 mice. By contrast, Giannakis et al.  report identifying H. pylori inside gastric stem cells. They further observed that an isolate from a cancer patient had closer affinity to gastric stem cells, causing more profound regulation of cell function, than did an isolate from the same patient 4 years before the cancer diagnosis, suggesting that the organism evolved through crosstalk with the host epithelium. This ‘endosymbiosis’ may be, in the end, a critical factor in disease development.
Co-factors in Carcinogenesis
Although half of the world’s population is infected with H. pylori, only a minority of individuals (estimated 1–2%) progress to gastric cancer over a lifetime. This percentage is deceptively low, however, because children and young adults are included in the lifetime risk assessment. When only middle-aged adults are considered, the risk of cancer appears more substantial. For example, in prospective studies conducted in Asia, between 3% and 6% of H. pylori-infected subjects developed gastric cancer within a decade [16,37]. Yet, worldwide, the great majority of adults infected with H. pylori survive without ever suffering from malignancy. Unfortunately, the reason some develop cancer and others do not is incompletely understood.
As mentioned above, bacterial factors appear to play a role in disease outcome. In addition to the pathogenicity island previously mentioned, a vacuolating cytotoxin (VacA) has been linked to malignancy. This protein has genotypes of varying epithelial toxicity. The s1/mL vacA allele has been particularly linked to epithelial damage and to carcinogenesis . The vacA alleles have varying global prevalence that can explain, in part, the marked differences in the incidence of gastric cancer noted worldwide . Other microbial factors that vary among strains and may affect the pathogenicity of H. pylori include apurinic/apyrimidinic endonuclease-1, which activates transcription factors and is involved in the DNA repair of oxidative damage ; OipA, an outer membrane protein linked to inflammation; and BabA which binds H. pylori to Lewis blood group antigens .
After bacterial genetics, probably the single most important factor influencing carcinogenesis is host genetics. Polymorphisms in the interleukin (IL)-1β and IL-1 receptor antagonist are particularly well studied . IL-1β acts as both an inhibitor of gastric acid secretion and a pro-inflammatory cytokine, both of which would tend to enhance carcinogenesis. Polymorphisms in tumour necrosis factor and IL-10 have also been linked to intestinal-type cancer , whereas IL-8 promoter polymorphisms have been linked to diffuse-type cancer . Human leukocyte antigen polymorphisms have been variably linked to gastric cancer . In one of the largest study series to date, 520 H. pylor-positive patients (330 with noncardia gastric cancer and 190 non-ulcer control patients) were evaluated for the influence of genetic factors on gastric cancer susceptibility. The results suggested that the combination of human leukocyte antigen class II and IL-10-592A/C polymorphisms affected the susceptibility to gastric cancer in a synergistic manner .
There is limited evidence of associations between environmental exposures and gastric cancer in the context of H. pylori infection. A prospective study from Colombia demonstrated that H. pylori eradication and increased dietary vitamin C and β-carotene prevented progression of pre-neoplastic lesions to cancer . Long-term follow-up of this trial, however, showed that the dietary benefits disappear when participants are followed for a longer period of time . Increased consumption of salt may also increase the risk of gastric cancer. In a study of 2476 participants, followed prospectively for 14 years, there was a significant relationship between increased salt consumption and the development of gastric cancer in subjects infected with H. pylori . Several studies have also demonstrated that cigarette smoking amplifies the H. pylori-associated risk of cancer .
An intriguing area of interest is the impact of infection with helminths on the outcome of H. pylori in humans. It has been postulated that, by decreasing the Th1 inflammatory response, helminthes could theoretically reduce gastric inflammation and cancer incidence. This phenomenon has been seen in some animal models of co-infection  and has also been suggested in small human studies .
Although there is strong agreement that H. pylori causes cancer, there is little consensus on whether and when this carcinogenic process can be reversed. The clinical trial of pre-neoplastic lesions by Fukase et al.  suggests that benefits can come very late in the malignant process, even after cancer has already begun. This conflicts with other human data that demonstrate continued disease progression in many individuals after H. pylori eradication. Moreover, it conflicts with the randomized clinical trial of Wong et al.  that revealed no benefits of H. pylori treatment in those with pre-neoplastic conditions. Thus, screening for and treatment of H. pylori infection as a strategy for secondary prevention of gastric cancer remains controversial. The controversy is amplified by data indicating benefits by preventing esophageal adenocarcinoma , asthma , diarrhoea  and even tuberculosis . These data additionally call into some question the advisability of creating a vaccine.
Yet, despite the ongoing academic debate, several important consensus groups have recommended screening and treatment of those at highest risk of malignancy (i.e. those with a family history of cancer, with prior gastric surgery, or with documented atrophic gastritis) . These same guidelines do not discourage broader screening based on individual choice. This last point conforms with the view of a number of H. pylori experts who have strongly advocated screening and treatment of wide swaths of the population. Although this is an appealing strategy and numerous studies have indicated it could be cost effective , there are currently no screening and treatment programmes ongoing outside of the research sphere.
A promising field is the use of molecular testing to predict clinical outcomes. Currently there is no reliable way to identify, among individuals infected with H. pylori, those with the highest risk of developing gastric cancer. Using high-throughput proteomics, Khoder et al.  evaluated 129 H. pylori strains from individuals with gastric cancer or duodenal ulcers. After identifying 18 statistically significant biomarkers, there were three proteins that predicted progression to cancer; these included a neutrophil-activating protein NapA, a RNA-binding protein, and a DNA-binding histone-like protein. These novel biomarkers could potentially be used as diagnostic assays to predict the evolution to gastric cancer. The usefulness of this approach in patients without pre-cancerous lesions is not clear , but it is an interesting avenue for future research.
H. pylori causes the majority of the world’s gastric cancers. With this knowledge comes the power to potentially control and eliminate this highly lethal disease. To date, however, we have not yet reached a consensus on whether or not to exterminate this microbe from the human host. It is possible that the answer may vary for different populations at different levels of risk of disease. In the absence of human intervention, we are fortunate that H. pylori appears to be disappearing from many human populations spontaneously. With this natural experiment, we will undoubtedly glean further data to inform us of our best approach worldwide to infection and disease.
Both authors declare no conflict of interest.