2. Gastric Cancer
The stomach is a saccular organ and accordingly luminal growth of tumours does most often not cause symptoms due to obstruction of the lumen at early phases. This is one of the reasons why most gastric cancers are diagnosed at an advanced stage and the prognosis of gastric cancer still poor. Until now, tumours have curiously been divided into cardiac or proximal when occurring in the small cardiac region, and distal when localised in the entire remaining part of the stomach. A classification based upon a such localisation does not take into consideration that the acid producing oxyntic mucosa and the gastrin producing antral mucosa are completely different, which should be included into a classification
[3].
There are many histological classifications of gastric cancer, but the classification by Laurén where glandular growth pattern is classified as intestinal type and tumours without glandular growth as diffuse type, seems to represent biological differences as the types do not convert into each other
[4] and have had a marked difference in incidence trends over the past decades
[5]. The weakness of the Laurén classification is that 15–20 % of the tumours that cannot be classified to either type, and are therefore classified as intermediate. The differences between intestinal and diffuse types found microscopically are also reflected macroscopically, where carcinomas of the diffuse type often induce a more pronounced fibrosis and may be spread submucosally leading to a non-compliant stomach with reduced motor activity, a so-called linitis plastica
[6].
Whereas
H. pylori has been accepted as the major cause of gastric cancer for 30 years
[2], the mechanism by which it causes cancer has not been solved despite enormous efforts. It should also be recalled that a bacterial infection besides
H. pylori has not been claimed to have a direct carcinogenic effect
[7]. Moreover, only a small proportion of individuals infected with
H. pylori develop gastric cancer. The role of bacterial characteristics, including Cytotoxin-associated gene-A (CagA) antigen expression
[8], is poorly understood and the search for specific factors of different strains of
H. pylori predisposing to gastric cancer has not been successful. Interestingly, CagA
+ H. pylori strains seem to induce more inflammation than other
H. pylori strains
[9]. The exact mechanism for the putative role of CagA in gastric carcinogenesis is still disputed
[10], but it has been proposed that CagA produces an oncoprotein which could induce malignancy after being incorporated into a cell
[11]. Alternatively, exposure of
H. pylori carcinogen in vitro has been reported to induce DNA damage
[12]. However, these two mechanisms presently lack support from convincing in vivo data. Neither has research aiming at identifying specific traits of individuals developing gastric cancer secondary to
H. pylori infection shown any consistent results. Therefore, neither studies of properties of the infectious agent nor host characteristics have uncovered the mechanism of gastric carcinogenesis due to
H. pylori infection. The study by Uemura et al. describing that
H. pylori gastritis did not predispose to gastric cancer before having induced oxyntic atrophy
[13] was a breakthrough. Antigenic mimicry between
H. pylori and the oxyntic mucosa
[14] including parietal cells
[15] was discussed as a possibility for the development of gastritis in the late nineties, but has not been followed up later. Since
H. pylori seems mainly to be acquired in childhood
[16] and since atrophic gastritis secondary to
H. pylori requires decades to develop, as indicated by the decline in gastric acid secretory capacity with age
[17], the logic strategy to prevent
H. pylori induced gastric cancer is to test young adults for
H. pylori by serology and eradicate the bacterium from positive individuals
[18].
In addition to
H. pylori gastritis, the so-called autoimmune gastritis also causes oxyntic atrophy. The cause of autoimmune gastritis is not known, but differs from
H. pylori gastritis by antral sparing, adult onset, and more rapid progress leading to complete atrophy of oxyntic glands, finally resulting in anacidity and a marked secondary hypergastrinemia.
H. pylori gastritis, on the other hand, starts in the antrum and spreads orally, often not affecting the entire oxyntic mucosa. Therefore, patients with
H. pylori gastritis may not be completely anacidic and accordingly have a less pronounced hypergastrinemia, which also could be due to antral gastritis impairing gastrin release from the G cells
[19] (). Whereas autoimmune gastritis especially leads to ECL cell derived neuroendocrine tumours (NETs)
[20] and more seldom to gastric cancer
[21][22], the opposite is due to
H. pylori gastritis where gastric adenocarcinomas is a more frequent result
[2] and gastric NETs are infrequent
[23][24]. The difference in prevalence of gastric NETs in patients with
H. pylori gastritis versus autoimmune gastritis is most likely due to the difference in blood gastrin levels. The mechanism of gastric carcinogenesis due to atrophic oxyntic gastritis may either be an expansion of commensal bacteria colonising the stomach due to lack of gastric acidity or hypergastrinemia. Since oxyntic atrophy seems to play a less important role in the pathogenesis of cardia cancers
[25][26], it seems unlikely that commensal bacteria have an important role in gastric carcinogenesis in general. However, as stated above autoimmune gastritis causes not only ECL cell NETs, but also neuroendocrine cancers originating from the ECL cell
[27] which may have until now often been classified as adenocarcinomas
[28]. Gastrin has a trophic effect particularly on the ECL cell, but also a general and less pronounced effect on the other cell types in the oxyntic mucosa
[29] presumably by an effect on the stem cell. Interestingly, hypergastrinemia has by far a more pronounced effect on ECL cell proliferation compared with the stem cell as assessed by autoradiography in rats
[30].The general effect may be a direct effect by gastrin or secondary to an effect by a mediator secreted from the ECL cell, most likely regenerating (REG) protein
[31][32]. If the REG protein is less efficient to induce stimulation of growth of the stem cell compared with the stimulating effect of gastrin on the ECL cell, this could explain that
H. pylori gastritis is more prone to predispose to adenocarcinomas with a longer latency (). Moreover, it seems established that the reduction in gastric acid secretion recorded in older individuals decades ago is not a result of aging per se, but progression of
H. pylori induced oxyntic atrophy
[17].
Figure 1. Patients with H. pylori (red line) gradually develop hypergastrinemia after infection in childhood, whereas patients with autoimmune chronic atrophic gastritis develop oxyntic atrophy and secondary hypergastrinemia more rapidly in adulthood (green line).
Intestinal metaplasia has been considered a separate risk factor for developing gastric cancer
[33]. However, it is difficult to discriminate the separate role of intestinal metaplasia from oxyntic atrophy in gastric carcinogenesis
[34] since intestinal metaplasia always occurs in a stomach with atrophy. Intestinal metaplasia can just be a marker and not a precursor of gastric cancer
[35][36]
Inhibitors of gastric acid secretion induce hypergastrinemia depending on their efficacy, reflecting the central role of gastrin in the regulation of gastric acidity
[37]. The risk of hypergastrinemia has particularly focused on the most efficient group, the proton pump inhibitors (PPIs). Depending of individual susceptibility, dose and duration of PPI treatment, a variable degree of hypergastrinemia develops leading to ECL cell NETs
[38][39][40] via a phase of ECL cell hyperplasia and dysplasia
[41]. Long-term profound acid inhibition by PPIs may also induce gastric neuroendocrine carcinomas originating from the ECL cell
[42]. Similarly, an inactivating mutation in the gene encoding the alpha sub-unit of the proton pump was reported to induce ECL cell NETs in patients in the third decade and a combined ECL cell NET and adenocarcinoma in the fourth decade
[43][44] demonstrating that the ECL cell has the capacity to develop into highly malignant tumours. This has been opposed by some dominating pathologists, for instance by Solcia et al.
[45], although their view was different in the late seventies
[46]. Moreover, the studies of patients with a genetic predisposition of gastric ECL-cell neoplasia illustrate that large studies following patients for a few years with respect to neoplasia may have observed patients for an impressive number of patient-years
[47] but are necessarily inconclusive concerning the long-term risk of cancers. In this context it may seem confusing that PPI treatment for gastroesophageal reflux after
H. pylori eradication was reported to induce gastric cancer
[48][49] and that PPI treatment was reported to increase the prevalence of gastric cancer the first year in a large epidemiological study based upon Swedish registries
[50]. The explanation could be that patients having been infected with
H. pylori have had a long period with hypergastrinemia before starting with PPI which has made them susceptible to a further increase in the gastrin level caused by PPI. In conclusion of this part,
H. pylori, autoimmune gastritis and PPIs all predispose to gastric cancer via gastrin. The increase in gastric cancer incidence from 1995 described in young Americans after decades of decline
[51] may be related to the increasing use of PPIs in the treatment of acid related diseases as well as mild discomfort from the upper abdomen
[52].
The interaction between
H. pylori gastritis and PPI treatment has been recognised for long. Attention was first drawn to this by the report describing progression of atrophic gastritis during PPI treatment
[53]. Subsequent studies have given ambiguous results concerning the effect of PPIs on the severity and spread of gastritis
[54]. However, there is no doubt that PPI treatment induces a more marked hypergastrinemia in individuals with
H. pylori gastritis
[55], which was to be expected in those with gastritis affecting the oxyntic mucosa. The exaggerated hypergastrinemia caused by PPI use in patients with
H. pylori infection is a strong indication for
H. pylori eradication in patients who are anticipated to start long-term PPI treatment. It is remarkable that hypergastrinemia still has not been generally accepted as a central mechanism in gastric carcinogenesis. Thus, every condition with long-standing hypergastrinemia in every species examined predisposes to gastric neoplasia of a different degree of malignancy. This has been demonstrated in rodents including rats, mice and Japanese cotton rats
[56][57][58], as well as in man with sporadic gastrinoma
[59] or gastrinomas as part of multiple endocrine neoplasia type I
[60], autoimmune gastritis
[20],
H. pylori gastritis
[24], proton pump inhibitor treatment
[39][42], after surgery where antral mucosa is no longer exposed to acidic gastric juice
[61] and in families with an inactivation mutation in the proton pump resulting in anacidity
[43][44]. The neglection of the importance of gastrin in gastric carcinogenesis may have serious consequences, since the incidence of gastric cancer could be markedly reduced by early eradication of
H. pylori before development of oxyntic atrophy, which may be achieved by testing and treating young adults. In those with established oxyntic atrophy either due to
H. pylori or autoimmune gastritis, treatment with the specific gastrin antagonist netazepide
[62][63] could be an option. Moreover, long-term treatment of large proportions of the population with efficient inhibitors of gastric acid secretion like the PPIs and the potassium competitive acid blockers (PCABs) should be avoided, particularly in young individuals. There is an increasing number of papers reporting an association between PPI use and gastric cancer
[64][65]. This should prompt clinicians to reduce gastric acid secretion just sufficient to allow healing of inflammation and relieve symptoms in long-term treatment of acid related gastro-oesophageal reflux. This is particularly important in young individuals
[66]. Treatment of mild to moderate oesophagitis should be started with histamine-2 blockers since their effect is reduced after PPI treatment due to tolerance
[67]. There are other less prevalent causes of gastric cancer such as congenital lack of E-cadherin or infection with Epstein Barr virus
[9], which will not be covered in this review. There are also other possible mechanisms for development of gastric cancer than just hypergastrinemia. Thus, Epidermal growth factor (EGF) was described to play an important role in regeneration of human gastric mucosoid cultures
[68]. It is conceivable that
H. pylori gastritis through inflammation could affect EGF concentration and thus play a role in carcinogenesis. However, as
H. pylori gastritis continues to predispose to gastric cancer development many years after eradication or loss of
H. pylori [69][70] and there is no relationship between
H. pylori infection and expression of EGF receptor 2 in gastric cancer
[71], such a mechanism seems less likely. Similarly, based upon a mouse model of linear tracing, a Lgr5
+ subpopulation of chief cells activated by Wnt signalling functioned as stem cells after injury, and was claimed to play a role in gastric carcinogenesis localised to the corpus
[72]. Since the carcinogenesis continues after loss of
H. pylori in a stomach with some degree of atrophy
[73], this mechanism also seems less plausible. The localisation of gastric carcinomas has been classified into proximal meaning those localised close to the cardia and distal as those localised to the rest of the stomach. Such a classification is very unlogical since the differences between the oxyntic and antral mucosa are ignored. When analysing cancers localised to the antrum and those in the corpus/fundus separately, hypergastrinemia is markedly associated with later development of adenocarcinoma in the corpus/fundus
[74]. Furthermore, when using such a classification, it should be remembered that oxyntic glands may be found in the proximal part of the antrum
[75]. In either way, in order to improve the understanding of gastric carcinogenesis, the tumour localisation should be carried out according to the three mucosae found in the normal stomach (cardia, oxyntic and antral). The data are overwhelming that oxyntic atrophy and gastrin are central in gastric carcinogenesis. The trophic effect of gastrin on the ECL cell predisposes to mutations by increased number of cell divisions each having an inherited risk of mutation. With time mutations with functional effects may lead to changes in malignant direction
[76][77].