Pflügers Archiv

, Volume 449, Issue 4, pp 344–355

Gastrin: old hormone, new functions

Authors

    • Physiological LaboratoryUniversity of Liverpool
  • Rod Dimaline
    • Physiological LaboratoryUniversity of Liverpool
  • Andrea Varro
    • Physiological LaboratoryUniversity of Liverpool
Invited Review

DOI: 10.1007/s00424-004-1347-5

Cite this article as:
Dockray, G., Dimaline, R. & Varro, A. Pflugers Arch - Eur J Physiol (2005) 449: 344. doi:10.1007/s00424-004-1347-5

Abstract

It is exactly a century since the gastric hormone gastrin was first described as a blood-borne regulator of gastric acid secretion. The identities of the main active forms of the hormone (the “classical gastrins”) and their cellular and molecular sites of action in regulating acid secretion have all attracted sustained attention. However, recent work on peptides derived from the gastrin precursor that do not stimulate acid secretion (“non-classical gastrins”), together with studies on mice over-expressing the gene, or in which the gastrin gene has been deleted, suggest hitherto unsuspected roles in regulating cell proliferation, migration, and differentiation. Moreover, microarray and proteomic studies have identified previously unsuspected target genes of the classical gastrins. Some of the newer actions have implications for our understanding of the progression to cancer in oesophagus, stomach, pancreas and colon, all of which have recently been linked in one way or another to dysfunctional signalling involving products of the gastrin gene. The present review focuses on recent progress in understanding the biology of both classical and non-classical gastrins.

Keywords

Epidermal growth factorEnterochromaffin-like cellEpithelial proliferationGastrinGastric acid

Introduction

The year 2005 is the centennial anniversary of the discovery of gastrin. Following the description by Bayliss and Starling of the first hormonal reflex (stimulation of exocrine pancreatic secretion by secretin released by intraduodenal acid), it became natural to ask if gastric acid secretion in response to food also might be controlled by a hormone [7]. Edkins provided the first evidence for a gastric acid secretagogue and coined the name gastrin for it [31, 32]. For many years there were doubts as to the relationships between gastrin and histamine, which by the 1920s had also become recognised as a strong acid secretagogue [95]. However, by the middle of the 20th century it was clear that there were at least three separate endogenous gastric secretagogues: gastrin, histamine and acetylcholine, the precise identity of gastrin being clarified with its isolation and sequencing by Gregory and Tracy [39].

In the last few years several lines of evidence have emerged to suggest that there is very much more to the biology of gastrin than indicated by earlier work. In particular, studies of (1) the various peptides generated during gastrin biosynthesis, (2) genetically modified mice that either overexpress the gastrin gene or in which the genes encoding gastrin or its receptor have been deleted, (3) the phenotype of patients with hypergastrinaemia, and (4) new targets indicated by functional genomic methods (gene arrays and proteomics) all indicate that the products of gastrin gene expression are implicated in a wide variety of biological processes, including effects outside the stomach. We propose to review here those aspects of the physiology and pathology of gastrin that have emerged in the last few years. For the most part, we cover only briefly those aspects of the physiology of gastrin that are well defined and have already been reviewed in detail elsewhere [29, 30, 130, 131]. Instead, we focus on new actions of well-recognised forms of gastrin, and the actions of new active forms.

The gastrins

The gastrin family

The pathways of biosynthesis by which the initial product of gastrin gene expression is processed to multiple alternative active peptides have been elucidated by pulse-chase labelling studies [29, 30]. The precursor, preprogastrin, is a peptide of 101 (human) or 104 (rat) amino acid residues, that is co-translationally cleaved (between Ala21 and Ser22, or Arg26 and Ser27) to yield progastrin [23, 99]. In endocrine cells, the latter is cleaved first by subtilisin-like prohormone convertases and then by carboxypeptide E to yield a 35 residue peptide with COOH-terminal glycine (G34-Gly) or an 18-residue peptide corresponding to its COOH-terminal tryptic fragment known as G17-Gly [73, 122, 125, 126] . Collectively we refer to progastrin and the COOH-terminal Gly- gastrins as the “non-classical gastrins” (Fig. 1). This reflects their new-found status as biologically active peptides, rather than their existence per se which has been appreciated for many years [5, 93]. In endocrine cells, G34-Gly and G17-Gly (the Gly-gastrins) are typically converted to the corresponding COOH-terminally amidated peptide by the enzyme peptidyl alpha-amidating mono-oxygenase (PAM), i.e. the peptides G34 and G17 that were originally isolated and characterised by Gregory and Tracy, and that possess the defining biological property of the hormone: the stimulation of acid secretion [38, 39, 40]. We refer to these peptides as the “classical gastrins” (Fig. 1). Endopeptidase cleavage and COOH-terminal amidation occur in secretory vesicles of the regulated secretory pathway [30]. In non-endocrine cells that express the gastrin gene, e.g. some tumour cells, these vesicles are scarce or absent, and subtilisin-like prohormone convertases and PAM are not well expressed, so that the main products of secretion are largely unprocessed peptides (e.g. progastrin).
Fig. 1

Biosynthetic relationships of preprogastrin-derived peptides. Major pathways of processing are indicated by blocked arrows, minor pathways by broken-line arrows. Boxes indicate the “non-classical” and “classical” gastrins (see text). The amidated peptides G34 and G17 are secreted from vesicles of the regulated pathway of exoctyosis which may also contain some unprocessed “non-classical” gastrins; in non-endocrine cells, e.g. colorectal carcinoma, progastrin may be secreted directly from the trans-Golgi network (TGN) to the cell surface by the constitutive route

Control of gastrin gene expression

The gastrin gene is normally expressed in G-cells of the pyloric antral mucosa and, in man, in duodenal G-cells [130]. There is increased expression of the gastrin gene in G-cells in response to food in the stomach, and down-regulation by acid probably secondary to inhibition by the paracrine mediator somatostatin [27, 139]. The cellular signalling pathways are still uncertain, although there is evidence that activation of the canonical MAP kinase pathway, for example by epidermal grown factor (EGF), stimulates gastrin gene expression via phosphorylation of Sp1 (Fig. 2) [19]. Infection with the bacterium Helicobacter pylori leads to inflammation in the antrum and down-regulation of somatostatin which tends to increase gastrin synthesis and release [15]. There appears to be stimulation of NFκB signalling in G-cells of H. pylori-infected patients which may increase gastrin gene expression [123]. It is now clear that the gastrin gene is also expressed in a variety of gastrointestinal tumours, of which colorectal carcinoma is presently the best explored [20, 67, 82, 124]. In cancer cells, activation of Ras and of the TCF-4/β-catenin signalling pathway have been implicated in gastrin gene expression [70, 81].
Fig. 2

Control of G-cell function. Gastrin synthesis and release is stimulated by aminoacids and peptides in the gastric lumen, and inhibited by gastric acid via the paracrine mediator somatostatin released from D-cells. In H. pylori infection there is increased synthesis and release of gastrin which is attributed to inhibition of D-cells by proinflammatory cytokines and stimulation of G-cells (probably mediated by NFκB). In some species the peptide neurotransmitter, gastrin releasing peptide (GRP) stimulates G-cell function, and gastrin synthesis is also stimulated by epidermal growth factor (EGF) possibly via Sp1

Mechanisms of release

At the cellular level, secretion of gastrin from G-cells is via Ca2+-dependent release from secretory vesicles of the regulated pathway [16]. It is well established that gastrin release occurs in response to both gastric luminal amino acids, and neuronal stimulation [130]. There is some recent evidence that the extracellular calcium-sensing receptor may play a role in mediating gastrin release [14]. In some species, e.g. rat and pig, neuronal control of gastrin release is mediated by the peptide neurotransmitter gastrin releasing peptide (GRP) [60, 105], but recent work in vivo using receptor antagonists suggests that while GRP might play a role in regulating acid secretion it does not regulate post-prandial secretion of gastrin from human G-cells [53] (Fig. 2). Physiologically, increased intragastric acid inhibits gastrin release via the paracrine mediator somatostatin [130]. When acid secretion is reduced, for example through treatment with inhibitors of the H+/K+ATPase, or is absent for example in pernicious anaemia, there is increased gastrin release [62].

CCK-2 receptors and signal transduction

There are two well-characterised receptors for peptides of the gastrin family. These are the CCK-1 (also known as the CCK-A) receptor which has low affinity for gastrin, but high affinity for the related hormone cholecystokinin (CCK) and the CCK-2 (also known as the CCK-B or gastrin-CCKB) receptor which has high affinity for both gastrin and CCK [83]. The CCK-2 receptor is expressed by parietal cells, enterochromaffin-like (ECL) cells, some smooth muscle cells, neurons of the central and peripheral nervous systems and (depending on the species) pancreatic acinar cells. The primary endogenous ligand for CCK-2 receptors in the CNS is thought to be CCK (which is abundantly produced by CNS neurons, whereas gastrin is not). However, in the periphery where receptors are exposed to both circulating hormones, the main endogenous ligand is likely to be gastrin since its concentrations in plasma are about five to ten times higher than those of CCK. Neither progastrin nor the Gly-gastrins are considered to be endogenous ligands of the CCK-2 receptor, but both are reported to have their own characteristic pattern of biological activities, and these are thought to be mediated by non-CCK-1/non-CCK-2 receptors [1, 112]. The term “CCK-C receptor” is sometimes used to describe non-CCK-1/non-CCK-2 receptors responding to progastrin-derived peptides. However, some authors have used the term to describe a splice variant of the CCK-2 receptor [115]. As discussed in more detail below, on present evidence there may be multiple other receptors responding to progastrin or Gly-gastrin. For these reasons we prefer to avoid the term “CCK-C receptor”.

The sequence WMDFamide is required for full agonist activity at CCK-2 receptors. This explains why peptides derived from preprogastrin (such as progastrin and the Gly-gastrins), which lack the COOH-terminal amide moiety have low or no affinity for CCK-2 receptors. A number of CCK-2 receptor antagonists have been reported, these include L-365,260, L-740,093, YM022, YF476, spiroglumide, and gastrozole [10, 74, 83]. These are useful for experimental studies, and have for example provided evidence supporting a role for gastrin in post-prandial increases in gastric acid secretion [8]. There is on going work to determine whether these and related compounds might be useful in the clinic.

The CCK-2 receptor belongs to the seven-transmembrane domain, G-protein coupled, receptor superfamily. It is normally coupled to Gαq/11 and activation leads to an increase in intracellular Ca2+ and PKC. The mechanisms have been studied in parietal cells and ECL cells [65, 100]. In addition, there is a considerable volume of work on signalling via CCK-2 receptors in cancer cell lines. The data indicate activation of the MAPkinase and PI3kinase/Akt signalling pathways and have been linked to a variety of outcomes including migration [84], proliferation [120], tubulogenesis [90] and inhibition of apoptosis [121]. One of the emerging themes in recent research is the activation of paracrine pathways downstream of the CCK-2 receptor; as yet relatively few studies have discriminated between intracellular signalling pathways that are directly activated by the CCK-2 receptor compared with those indirectly activated as a consequence of autocrine or paracrine activation of other receptors (see below).

A CCK-2 receptor variant in which intron 4 is retained (CCK2Ri4sv) has an extra 69 amino acid residues in the third intracellular loop. This variant appears to be expressed in colorectal tumours but not adjacent normal colon [48], and in pancreatic tumours [114]. It is also expressed in the condition of Barrett’s metaplasia of the oesophagus, which is known to be a precursor lesion of oesophageal cancer, while it seems not to be expressed in normal oesophageal mucosa [45]. The variant receptor has increased constitutive activity compared with the wild type receptor and there is evidence for slightly higher affinity of Gly-gastrin, although still relatively low (<0.001) compared with G17 [48]. When transfected into HEK293 cells, which have low Src activity, CCK2Ri4sv is reported to be associated with an agonist-independent increase in Src and it has been suggested that expression of this form of the CCK-2 receptor could contribute to the increased Src activity in some tumours [88].

Control of expression of the CCK-2 receptor

In addition to the primary sites of expression of the CCK-2 receptor mentioned earlier, it is now clear that in some experimental or clinical circumstances there is increased expression in other cell types. In a transgenic mouse model of hypergastrinaemia in which there is a gastric hyperproliferative condition, there is reported to be induction of the CCK-2 receptor on proliferating cells [80]. Moreover, there is now evidence for rapid, specific, increases in expression at the margin of cryoulcers induced in rat stomach [104]. There is also threefold higher expression in the epithelium of Barrett’s oesophagus compared with oesophageal epithelium from unaffected individuals [42]. These data suggest that the expression of the CCK-2 receptor gene may be upregulated in hyperproliferative conditions including the responses to injury or inflammation of gastrointestinal epithelia. The mechanisms involved in transcriptional regulation of the gastrin gene remain largely unexplored, although in human colon cancer cells over-expression of oncogenic Ras is reported to increase transcription via a MEK-dependent mechanism [61]. Various clinical conditions, e.g. schizophrenia, panic disorder, Parkinson’s disease and alcoholism [44, 46, 86, 133] have been reported to be associated with polymorphisms in the promoter region of the human CCK-2; it is possible these influence receptor expression but again the underlying mechanisms are not well understood.

Paracrine cascades downstream of the CCK-2 receptor

The role of histamine as a paracrine mediator of the action of gastrin on acid secretion is well recognised [11, 30, 43]. However, recent work suggests gastrin also activates many other paracrine signalling pathways. These include release of somatostatin [107], activation of COX-2 [41, 72, 113], shedding or induction of members of the EGF family [77, 128], FGF-1 [84], Reg [35, 52] and the chemokine IL-8 [55] (Fig. 3). The EGF family members and Reg have attracted particular attention because they may mediate at least some of the effects of gastrin on cell proliferation.

Stimulation of a number of GPCRs coupled to Gαq/11 leads to shedding of growth factors of the EGF family [96]. It seems that both an increase in intracellular calcium and activation of PKC lead to activation of an extracellular metalloproteinase that acts on the membrane bound precursors of EGF-receptor ligands including HB-EGF, amphiregulin and TGF-α. Several proteases have been implicated in this process including TACE/ADAM17 [54, 102]. In a model system designed to allow studies of paracrine signalling pathways there is evidence that gastrin acts via shedding of HB-EGF to stimulate proliferation [78, 128]. Gastrin may also stimulate HB-EGF gene expression [108, 134]. Autocrine and paracrine pathways involving EGF family members, and perhaps also FGF-1, account for activation of p42/44 MAPkinase by gastrin in a gastric cancer cell line (AGS) [84]. Interestingly, there may also be interactions downstream of EGF and CCK-2 receptor stimulation [87], since activation of the MAPkinase pathway is reported to modulate the amplitude of gastrin-dependent intracellular calcium release. Interactions between CCK-2 and EGF receptors may also influence COX-2, since gastrin increases transcription of COX-2 mRNA, and EGF enhances its stability [113].

Expression of the Reg family of growth factors, which includes the pancreatitis associated proteins, is markedly increased in response to damage and inflammation in a wide range of tissues including gastrointestinal tract, pancreas and nervous system [3, 24, 75]. In both experimental animal models of hypergastrinaemia (omeprazole-treated rats), and in patients with hypergastrinaemia (pernicious anaemia) there is increased expression of gastric Reg 1(rat) and Reg 1α (human) [35, 52]. In both rat and human, Reg is expressed in ECL cells, but in human stomach there is also expression in chief cells. The cellular mechanisms by which gastrin stimulates Reg expression include activation of PKC and stimulation of the small GTPase RhoA, and they involve a CG-rich cis-element in the proximal region of the Reg1 promoter [4]. Reg increases proliferation of gastric epithelial cells and mutations of Reg1α are also associated with ECL cell carcinoid tumours [35, 52]. Moreover, in transgenic mice that over-express Reg there is increased differentiation of parietal and chief cells [77], which is reminiscent of the phenotype of mice over-expressing gastrin at least up to about 6-months [134].

Control of acid secretion

The evidence that gastrin regulates gastric acid secretion after a meal is based on the observations that similar concentrations of endogenous or exogenous gastrin in plasma are associated with comparable rates of acid secretion, and that the post-prandial acid response is inhibited by either gastrin immunoneutralisation or by CCK-2 receptor antagonists [130]. Quantitative pharmacological studies using histamine H2 receptor antagonists support the idea that gastrin acts mainly via release of histamine from ECL cells, which then functions as a paracrine mediator in the gastric mucosa to stimulate parietal cells. The relevant issues have been well reviewed [11, 30, 49]. In addition, gastrin also regulates ECL cell numbers, and the expression of genes in ECL cells that direct histamine synthesis and storage. Nevertheless, parietal cells also express the CCK-2 receptor and the significance of this is attracting renewed attention. One view has been that direct effects of gastrin on parietal cells may contribute to acid secretory responses when plasma concentrations are high. However, recent data from studies on mice in which the genes encoding gastrin or the CCK-2 receptor have been deleted suggest roles in parietal cell maturation (see below)

Control of parietal cell number, maturation and position

In the absence of the gastrin gene there is a failure of parietal cells to complete the full sequence of maturation events. Morphological studies indicate that parietal cells are present in Gas-KO mice, but they appear immature; there is reduced basal acid secretion and insensitivity to stimulation by gastrin, histamine or carbachol [34, 68]. These effects can be reversed by administration of gastrin for >24 h, so for example in Gas-KO mice implanted with osmotic minipumps G17 induced the capacity for stimulation of acid secretion by gastrin, histamine and carbachol [18, 34]. Interestingly, a potentiating interaction between G17 and G17-Gly has been described. Thus, in Gas-KO mice receiving G17, acid secretion increased initially and then faded over a period of 2 weeks, whereas in animals receiving G17 and G17-Gly there was maintained secretion over many weeks [18].

Parietal cells are generated from epithelial progenitor cells that lie in the isthmus region of the gland. Some parietal cells are found in this region, and may migrate upwards; however, it appears that most parietal cells migrate along the gland to its base. In Gas-KO mice, the rate of migration is depressed compared with wild type mice, indicating that gastrin stimulates cell migration in vivo [66].

In mice over-expressing the gastrin gene with elevated plasma amidated gastrin, there is an initial (up to about 3 months of age) response characterised by increased acid secretion and increased parietal cell numbers. However, with age there is a progressive loss of parietal cells and expansion of the mucus neck cell population [134]. This condition resembles the human condition of atrophic gastritis, which is characterised by loss of gastric glands, shortening of the glands, loss of parietal cells and increased plasma gastrin. The condition is found in association with H. pylori infection and is recognised to be premalignant. Interestingly, in transgenic mice with elevated plasma amidated gastrin, infection with H. felis, also accelerates the progression to atrophy and to gastric cancer [134].

It seems therefore that while gastrin is required for the normal regulation of gastric epithelial organisation and function, it can also lead to disruption of epithelial organisation, particularly when accompanied by inflammation. The mechanisms are likely to be important because it is well recognised that increased plasma gastrin often accompanies H. pylori infection in man [15]. Even so, the possibility that gastrin might exacerbate the effects of H. pylori have been largely neglected. Elevated plasma gastrin alone does not predispose to gastric atrophy in humans (unlike mice), for example there is elevated plasma gastrin in the human condition of sporadic gastrinoma but not gastric atrophy [62]. However, in the presence of inflammation it seems that increased plasma gastrin is associated with expression of a number of genes that predispose towards loss of parietal cells, loss of epithelial organisation and hyperproliferation (see below).

Control of gene expression in ECL cells (HDC, VMAT2 and CGA)

The key events in the production and storage of histamine by the ECL cell in response to gastrin are synthesis of the amine by the cytosolic enzyme histidine decarboxylase (HDC, EC 4.1.1.22.), and the subsequent sequestration of histamine into secretory granules by vesicular monoamine transporter type 2 (VMAT2). Expression of the mRNA encoding HDC is regulated over the physiological range of circulating gastrin concentrations, and may be increased within the time taken to digest a single meal [28]. In response to hypergastrinaemia in humans and rodents, mRNAs encoding both HDC and VMAT2 are markedly upregulated, together with that encoding a third protein, chromogranin A (CGA) that may have a role in the stabilisation of amine-containing secretory granules [17, 25, 26, 51, 101, 118]. Transcriptional activation by gastrin of these three genes has been investigated primarily in human gastric epithelial cell lines and involves transcription factor binding to disparate GC rich regions within the promoter sequences. In the case of HDC, three as yet uncharacterised nuclear proteins bind to response elements just downstream of the transcriptional start site to stimulate transcription in response to gastrin [97, 140]. Interestingly, these gastrin-response elements can also direct inhibition of HDC transcription, when bound by the gut-enriched transcription factor Kruppel-like factor 4 (KLF4); KLF4 can also inhibit HDC gene transcription through an upstream SP1-binding GC-rich region [2]. Transcriptional stimulation by gastrin of the VMAT2 gene also depends in part on a yet-to-be characterised protein that binds to GC-rich, overlapping SP1/AP2 response elements. In this case, gastrin responsiveness is also dependent on binding of phosphorylated CREB to a canonical CRE site; binding of CREB and the uncharacterised factor may be co-ordinated by p300/CBP [135]. The importance of the CRE site in gastrin-stimulated transcription of VMAT2 in non gastric tissue (rat pheochromocytoma cells) has also been reported [36]. Gastrin-stimulated transcription of CGA can be accounted for through activation of the transcription factors EGR-1, SP1 and CREB [98]. The cis-regulatory elements that mediate the effects of gastrin on these three genes lie within 200 base pairs of the start of transcription; studies in transgenic mice indicate that in the case of CGA, approximately 4.5 kb of promoter seems sufficient to direct expression to ECL cells [56].

Control of gastric epithelial cell proliferation

Patients with elevated plasma gastrin due to a gastrin-secreting tumour, and animals treated with gastrin for prolonged periods, exhibit increased thickness of the gastric mucosa and increased parietal cell and ECL-cell mass [62]. In fasted rats, rates of gastric epithelial cell proliferation measured by BrdU incorporation are low, but feeding rapidly increases proliferation which is due in part to increased gastrin secretion [85]. The gastric epithelial progenitor cell population is thought not to express CCK-2 receptors and instead gastrin is considered to act indirectly via release of growth factors (see above) [30], which presumably originate from parietal or ECL cells, since these are the main cell types expressing the CCK-2 receptor (Fig. 3). Parietal cells are terminally differentiated and the increase in their number in hypergastrinaemia is likely secondary to stimulation of proliferation and increased commitment to the parietal cell lineage. In a range of different types of genetically modified mice characterised by loss of parietal cell function, the reflex increase in plasma gastrin concentrations (due to loss of acid inhibitory feedback) leads to increased gastric proliferation [30, 33].
Fig. 3

Paracrine cascades downstream of CCK-2 receptor stimulation. Gastrin acts on ECL cells to release histamine which stimulates acid secretion from parietal cells, and D-cells to release somatostatin (Som) which inhibits acid secretion. Gastrin also stimulates proliferation, differentiation and cell migration; since proliferating cells are thought not to express the CCK-2 receptor these actions are indirect and possible mediators include EGF family members (HB-EGF, TGFα, amphiregulin) and Reg family members. Other mediators include COX-2 products and IL-8

Control of ECL cell number

Both patients with hypergastrinaemia, and animals with experimentally induced hypergastrinaemia (either exogenous or endogenous in origin) typically exhibit ECL cell hyperplasia [62]. There is evidence that these cells have the capacity for proliferation so that ECL cell hyperplasia may reflect a direct mitogenic action of gastrin [76]. Present interest in the trophic effects of gastrin on ECL cells stems from two facts: patients treated with proton pump inhibitors (PPIs) have elevated plasma gastrin, and in both patients and animals ECL cell hyperplasia can lead to dysplasia and ECL cell carcinoid tumours.

The first clear evidence of a link between gastrin and ECL cell carcinoid tumours came from studies in rats on long-term treatment to suppress acid secretion [9, 47]. The role of gastrin was demonstrated by showing that the administration of PPIs to animals in which the pyloric antrum was removed did not cause ECL cell hyperplasia [117]. In other animal models too there is evidence that CCK-2 receptor stimulation is linked to ECL cell proliferation. For example, the sub-Saharan rodent Mastomys spontaneously exhibits ECL cell carcinoid tumours possibly due to amino acid substitutions in the CCK-2 receptor in this species that increase constitutive activity [79, 103].

The combination of hypergastrinaemia and ECL cell carcinoid tumours in humans occurs in two conditions: gastrin-secreting tumours (gastrinomas), which may be sporadic or on a background of multiple endocrine neoplasia type-1 (MEN-1), and atrophic gastritis or pernicious anaemia in which there is loss of parietal cells, and hence acid secretion, relieving the normal inhibition of the G-cell. It is reported that approximately 30% of patients with gastrinoma and MEN-1 and 5% of patients with atrophic gastritis/pernicious anaemia develop ECL cell carcinoid tumours [12, 37]. Importantly, however, ECL cell carcinoids occur in less than 1% of patients with sporadic gastrinoma although the range of plasma gastrin concentrations in this condition is similar to that in gastrinoma on a background on MEN-1 [94]. It appears therefore that mutation of the gene encoding the tumour suppressor menin, or an inflammatory condition in the stomach, together with hypergastrinaemia, predispose to ECL cell carcinoid tumours. Almost certainly other factors are also involved, e.g. mutation of Reg [52].

New actions at the CCK-2 receptor

Over the last few years it has become clear that in different cellular and animal models activation of the CCK-2 receptor is associated with a far wider range of biological responses than previously supposed. The effects include stimulation of migration [84], invasion [137], tubulogenesis [90], and inhibition of apoptosis [121]. These, together with the proliferative response to CCK-2 receptor stimulation, are likely to account for the oncogenic effects of the amidated gastrins. There is a clear example in pancreas where targeted expression of the receptor and locally high concentrations of amidated gastrin lead to cancer [21]. The situation in colon, stomach, lung and oesophageal cancer may be more complex, with evidence for the involvement of aberrantly spliced CCK-2 receptors, aberrant expression of CCK-2 receptors in the premalignant phase, or roles for non-classical gastrins (see below).

Recent studies using gene arrays have started to identify a number of previously unsuspected target genes downstream of the CCK-2 receptor, which suggest novel functions for the amidated gastrins. Initial array studies of the gastric cancer cell line, AGS, expressing the CCK-2 gene, identified plasminogen activator inhibitor (PAI)-2 and matrix metalloproteinases (MMP)-7 and -9 as previously unsuspected gastrin-regulated genes [127, 137]. The relevance of these observations was confirmed by the demonstration of increased abundance of the gene products in the stomach of patients with hypergastrinaemia, and by studies in cell lines showing increased transcription using promoter-luciferase reporter vectors. The results are interesting because they raise the novel possibility that gastrin regulates cell invasion and the deposition of extracellular matrix (Fig. 4).
Fig. 4

Control by gastrin of expression of plasminogen activator inhibitor-(PAI)-2. Gastrin stimulates the small GTPase RhoA, and the MAPkinase pathway increasing PAI-2 expression via AP-1, CREB and NFκB; the transcriptional repressor menin inhibits expression. In part, the actions of gastrin are mediated by IL-8 and the products (PGs) of COX-2 induction acting by autocrine/paracrine mechanisms. PAI-2 secreted by cells inhibits the urokinase plasminogen activator decreasing fibrinolysis and inhibiting cell invasion and remodelling of the extracellular matrix (ECM). Some PAI-2 is retained within cells and is thought to inhibit apoptosis. PAI-2 is an example of a previously unsuspected downstream target of the CCK-2 which is a potential mediator of gastrin in regulating tissue organisation

In parallel studies, Khan et al. [64] applied differential mRNA display to the gastric mucosa of wild type and gastrin-knockout mice. This led to the identification of trefoil factor-1 (TFF-1) as a putative gastrin-regulated gene. Importantly, TFF-1 is expressed by gastric epithelial cells that are not thought to express the CCK-2 receptor. Moreover, PAI-2 is expressed both in cells that express the CCK-2 receptor (e.g. ECL cells) and those that do not (e.g. mucus cells). It seems possible, therefore, that gastrin activates paracrine pathways leading to the regulated expression of genes such as TFF-1 and PAI-2 (Fig. 4). In the case of PAI-2 the evidence from studies in AGS cells suggests a role for both COX-2 and IL-8 as paracrine mediators of gastrin-stimulated gene expression [126, 129].

Responses to non-classical gastrins

Several studies have dealt with the cellular signalling mechanisms activated by Gly-gastrin [58, 116, 119]. The proliferative effects of Gly-gastrin have been reported to depend on both MAPkinase-dependent and independent pathways [58, 116]. In an immortalised mouse gastric cell line, in which Gly-gastrin and G17 activate different receptors, Gly-gastrin produced a sustained increase in p42/44 MAP kinase phosphorylation, whereas G17 stimulated a transient increase, and both peptides increased PI-3-kinase activity [58]. Both peptides also stimulated proliferation, but a sustained increased in MAPkinase was required for the cell dissociation and migration that was seen with Gly-gastrin but not G17 [58]. Gly-gastrin also activates Src tyrosine kinase and this has been linked to dissociation of tight junctions [59].

Actions of Gly-gastrin

Seva et al first demonstrated that Gly-gastrin stimulated the proliferation of AR4–2J cells [106] and at about the same time Singh et al. demonstrated that non-classical gastrins were mitogenic to some cells lines [109]. The proliferative effects of Gly-gastrins have now been reported in both transformed and non-transformed cells [57, 136]. In addition transgenic mice over-expressing Gly-gastrin have been shown to have increased proliferation in the colon [69]. These animals are also prone to the development of lung adenocarcinomas; interestingly, the expression of Gly-gastrin (but not progastrin or amidated gastrin) in lung adenocarcinoma in patients is associated with decreased survival suggesting a specific oncogenic role for Gly-gastrin [71]. In addition to increased proliferation, Gly-gastrin stimulates cell migration [58] and cell invasion at least in part by induction of MMP-1 and MMP-3 [6, 63] which taken together are likely to contribute to the oncogenic actions of the peptide.

The role of Gly-gastrin in normal gastric physiology is still rather uncertain. However, an early report suggested synergistic interactions between amidated and Gly-gastrin for stimulation of acid secretion in the rat [50]. More recently, it has been found that administration of the two peptides in combination enhances and maintains parietal cell function in mice in which the gastrin gene has been deleted [18]. It seems that parietal cell responses to amidated gastrin may therefore be modulated by Gly-gastrin although the mechanisms are unknown.

Recently, Pannequin et al. have reported that bismuth and ferric ions bind to and influence the biological actions of Gly-gastrin. Using NMR spectroscopy, ferric ions were shown to bind Glu-7 and to be required for the action of Gly-gastrin on cell proliferation and migration [91]. Conversely, bismuth ions which also bind Glu-7, inhibited the action of Gly-gastrin on cell proliferation and migration, but had little effect on responses to amidated gastrin mediated by the CCK-2 receptor [92]. Whether these observations are important in gastric physiology, or in the progression to cancer, requires clarification.

Actions of progastrin

In parallel with the emerging evidence that Gly-gastrin might have its own properties it has also become clear that progastrin is biologically active. The evidence from studies of the phenotype in transgenic mice or cell lines over-expressing progastrin, and from studies of the action of recombinant progastrin suggests roles in stimulating proliferation [132], inhibition of apoptosis [138], and stimulation of cell migration [59]. The receptor at which recombinant progastrin acts is distinct from the CCK-2 receptor and is reported to be linked to activation of Src [13, 112].

In transgenic mice over-expressing progastrin, there is increased epithelial cell proliferation in the colon, and increased thickness of the colon mucosa [132]. Importantly, these mice might also exhibit increased susceptibility to the damaging effects of the carcinogen azoxymethane [110, 111]. These effects have now been described in two different lines of transgenic mouse: hGas mice in which expression is directed to the liver, and Fabp-PG mice in which there is targeted expression to the intestine [22, 111, 132]. Moreover, there is also increased susceptibility to the carcinogenic effects of azoxymethane in mice expressing a mutant form of progastrin in which the dibasic amino acid residue cleavage sites have been deleted [22], indicating that the carcinogenic effects of progastrin are not dependent on processing to Gly-gastrin or amidated gastrin.

The mechanisms by which progastrin stimulates proliferation and predisposes to carcinogenesis are still not well understood. Normally, in mice exposed to 8G γ-irradiation there is complete inhibition of mitosis in the colon. In contrast, in irradiated mice that over-express progastrin, colonic cells remain in the cell cycle. It seems therefore that progastrin allows cells to stay in the cell cycle even after DNA damage, which presumable contributes to its apparent effects as a co-carcinogen [89]. In this irradiation model there is no difference in rates of colonic epithelial cell apoptosis between control and progastrin-over expressing mice, although in IEC6 cells it has been reported that progastrin has inhibitory effects on apoptosis [138].

Overview

The data reviewed here indicate that in addition to the acute control of acid secretion in the post-prandial period, the COOH-terminally amidated (or classical) gastrins also regulate parietal cell maturation, ECL cell numbers and the capacity for histamine production. Together, these actions can be considered aggressive since they directly or indirectly lead to increased acid in the gastric lumen which in turn enhances the potential for acid-peptic damage to upper gastrointestinal mucosa. We suggest that recent data indicate that classical gastrins can also activate defence mechanisms that protect against acid-peptic damage and promote healing responses. The protective mechanisms include induction of the CCK-2 receptor in cells that do not normally express it, stimulation of cellular proliferation and migration, induction of COX-2, and of various genes contributing to tissue repair, e.g. Trefoil peptides, PAI-2 and perhaps MMPs. The non-classical gastrins such as progastrin and Gly-gastrin are also now emerging as active entities—they have effects on proliferation and migration that might enhance or augment the protective effects of the amidated gastrins. However, some of the actions of non-classical gastrins may well be subverted in cancer or in preneoplastic conditions where their increased production leads to enhanced cellular proliferation even after DNA damage.

A century after the discovery of gastrin, it is clear that there are still many open questions regarding the biology of this hormone. We argue that future progress will depend on an appreciation of actions that go beyond the post-prandial control of acid secretion. The issue is important not least because a more complete understanding may provide novel therapeutic targets relevant to gastrointestinal disease.

Copyright information

© Springer-Verlag  2004