Involvement of Cholecystokinin/Gastrin-Related Peptides and their Receptors in Clinical Gastrointestinal Disorders

Authors

  • Robert T. Jensen

    Corresponding author
    1. Digestive Diseases Branch, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, Maryland, U.S.A.
    Search for more papers by this author

Author for correspondence: Dr. Robert T. Jensen, NIH/NIDDK/DDB, Bldg. 10, Rm. 9C-103, 10 Center Dr MSC 1804, Bethesda, MD 20892-1804, USA (fax +1 301 402 0600, e-mail robertj@bdg10.niddk.nih.gov).

Abstract

Abstract: In this paper the possible roles of cholecystokinin (CCK), gastrin, or gastrin-related peptides and their receptors in human gastrointestinal diseases are reviewed. For CCK/CCKA receptors (CCKA-R), the evidence for their proposed involvement in diseases caused by impaired CCK release or CCKA-R mutations, pancreatic disorders (acute/chronic pancreatitis), gastrointestinal motility disorders (gallbladder disease, irritable bowel syndrome), pancreatic tumor growth and satiety disorders, is briefly reviewed. The evidence that has established the involvement of gastrin/CCKB-R in mediating the action of hypergastrinaemic disorders, mediating hypergastrinaemic effects on the gastric mucosa (ECL hyperplasia, carcinoids, parietal cell mass), and acid-peptic diseases, is reviewed. The evidence for their possible involvement in mediating growth of gastric and pancreatic tumours and possible involvement of gastrin-related peptides in colon cancers, is reviewed briefly.

Since the discovery of cholecystokinin (CCK) in 1928 by Ivy & Oldberg of its action on the gallbladder and structural characterization by Mutt & Jorpes (1968), similar to the closely-related peptide, gastrin, whose acid-releasing activity was described in 1905 by Edkins and structural characterization by Gregory & Tracy in 1964, there have been many studies to attempt to define their roles in both physiological and pathological gastrointestinal processes (Jorpes & Mutt 1973; Lamers et al. 1990; Liddle 1994; Walsh 1994; Baldwin 1995; Guo & Townsend, Jr. 2000; Otsuki 2000; Dockray et al. 2001) . Despite recent insights elucidating the structure of the two G-protein-coupled receptors that mediate their cellular actions (i.e., the CCKA and CCKB receptor) (Wank 1995), characterization of receptor location, peptide and receptor genes, development of receptor antagonists and receptor and peptide knockout animals (Liddle 1994; Jensen 1994; Walsh 1994; Langhans et al. 1997; Wang & Dockray 1999; Hinkle & Samuelson 1999; Lacourse et al. 1999; Miyasaka et al. 1999; deTullio et al. 2000), especially the role of CCK and to a lesser extent, gastrin, in normal gastrointestinal physiology as well as in clinically important gastrointestinal pathologic states, remains largely unclear and in many cases, controversial. In this review the latter aspect will be primarily dealt with, which is their role in clinically important processes rather than their role in normal physiology. By this it is meant their role in disease states which cause clinical symptoms in man.

In this review these disorders will be divided into those involving primarily CCKA and those involving primarily CCKB receptors. Because only sulfated CCK analogues have a high affinity for CCKA (Jensen et al. 1989; Liddle 1994), CCKA-R disorders involve the potential role of CCK in disease states. Gastrin and CCK can interact with CCKB receptors (Jensen et al. 1989; Liddle 1994); however, much more information is available on the role of gastrin causing CCKB receptor-mediated disorders than CCK.

Clinically important gastrointestinal disorders involving CCK or CCKA receptors (table 1)

Table 1.  Gastrointestinal disorders possibly involving CCK/CCKA-R alterations.
1. CCK-deficient states (celiac sprue, bulimia, diabetes mellitus, autoimmune polyglandular syndrome-type 1)

Evidence suggests CCKA receptor activation is important physiologically for gallbladder contraction, pancreatic secretion, effects on gastrointestinal motility (slowing gastric and colonic motility), inhibitory effects on acid secretion, stimulation of pancreatic growth, satiety, and augmentation of meal-related insulin release (Meyer et al. 1989; Schmidt et al. 1991; Liddle 1994; Schmidt et al. 1994; Langhans et al. 1997; Schwizer et al. 1997; Shoji et al. 1997; Hinkle & Samuelson 1999; Lacourse et al. 1999; Miyasaka et al. 1999; Beglinger et al. 2001; Degen et al. 2001; Suzuki et al. 2001; Takiguchi et al. 2002).

General. The review below shows that CCKA receptor's proposed role in various important clinical gastrointestinal disorders has not been unequivocally established (Meyer et al. 1989; Liddle 1994). Evidence exists for a potential gastrointestinal role of CCKA receptors in pancreatic disorders (pancreatitis, pancreatic deficiency states) (Liddle 1994), gastrointestinal motility disorders (gallbladder disorders, irritable bowel syndrome, constipation) (Liddle 1994; Bucceri et al. 2002), satiety (Degen et al. 2001; Moran 2000), and growth of normal and neoplastic tissues (Liddle 1994; Reubi et al. 1997; Guo & Townsend, Jr. 2000). Furthermore, CCK's clinical usefulness is proposed for diagnosis of gallbladder motility disorders causing symptoms (acalculous cholecystitis) (Liddle 1994; Middleton & Williams 2001; Ziessman et al. 1999 & 2001; Ziessman 2001; Ziessman 1999), and as a radiolabeled analogue to image tumors overexpressing CCK receptors (primarily CCKB receptors) (deTullio et al. 2000; Behr & Behe 2002). CCKA receptor's possible role in these disorders will be briefly reviewed below.

CCK secretion or CCKA receptor alterations in clinical gastrointestinal disease. There are no reports where alterations in the CCKA receptor or of CCK hypersecretion have been unequivocally shown to contribute to any disease state. A patient with obesity and gallstones was described (Miller et al. 1995) with a 262 bp deletion in the 3rd exon in the CCKA receptor resulting in a receptor that did not bind CCK. This deletion was present in 93% of the CCKA receptor transcripts but whether this contributes to the disease state remains unclear. A V365I mutation in the CCKA receptor has been reported in obese, diabetic patients (Marchal-Victorion et al. 2002). This mutation demonstrated a decreased expression and efficacy for activating phospholipase C when transfected into COS-7 cells (Marchal-Victorion et al. 2002). Whether this mutation is contributing to either the diabetes mellitus or obesity, remains unclear.

Low levels of CCK, possibly contributing to impaired gallbladder contractility and the development of cholelithiases, are reported in patients with celiac disease (Low-Beer et al. 1975; Calam et al. 1982; Masclee et al. 1991), the short bowel syndrome (Ling et al. 2001) and in diabetics (Bucceri et al. 2002). Recently it has been proposed that diarrhoea and malabsorption in patients with autoimmune polyglandular syndrome-type 1 are due to CCK deficiency secondary to loss of CCK-producing enteroendocrine cells in the proximal small intestine (Hogenauer et al. 2001). However, this proposal has been questioned by others (Creutzfeldt 2001) because CCKA receptor knockout mice do not have malabsorption and/or diarrhoea (Schmitz et al. 2001) and neither suppression of CCK release by somatostatin (Lembcke et al. 1987) nor treatment CCKA receptor antagonists led to a reduction in pancreatic enzyme secretion severe enough to cause severe steatorrhea (Schmidt et al. 1991).

CCK and CCKA receptor's role in pancreatic disorders.

General. The exact role of CCKA receptors in clinically important pancreatic disorders such as pancreatitis (acute or chronic) has been difficult to define both because of lack of ideal animal models of these diseases and because of important differences in the occurrences of CCKA and CCKB receptors in human and animal pancreatic acinar cells (table 2). In rat and mouse, the two animals used in most experimental studies of pancreatitis or tumour formation/growth, pancreatic acini possess entirely CCKA receptors, whereas in human pancreatic acini almost entirely CCKB receptor expression is reported (Wank et al. 1994; Monstein et al. 1996; Morisset et al. 2000) (table 1). Furthermore, a recent detailed study (Ji et al. 2001) showed very low expression of both CCKB and CCKA receptors in human pancreatic acini and CCK caused no alteration in human acinar cell function. These observations, combined with reports showing CCK-mediated pancreatic enzyme secretion in humans, is mediated by a cholinergic mechanism (Adler et al. 1991; Soudah et al. 1992)led the authors (Adler et al. 1991; Ji et al. 2001) to conclude that it is unlikely that activation of CCKA receptors on human pancreatic acini caused enzyme secretion or pancreatic growth as reported in various animal studies.

Table 2.  CCK receptors in pancreatic acinar cells in man and different species.
 CCKA-RCCKB-R
  1. Rat, mouse data are from refs. (Wank et al. 1994; Zhou et al. 1995; Monstein et al. 1996; Morisset et al. 2000); guinea pig from ref. (Wank et al. 1994); dog from ref. (Fourmy et al. 1987); calf from refs. (LeMeuth et al. 1993; LeDrean et al. 1999; Saillan-Barreau et al. 1999); pig from refs. (Morisset et al. 1996 & 2000; Philippe et al. 1997; Bourassa et al. 1999; Schweiger et al. 2000); and human from refs. (Wank et al. 1994; Monstein et al. 1996; Nishimori et al. 1999; Morisset et al. 2000; Ji et al. 2001).

Rat, mouse100%  0%
Guinea pig83%17%
Dog84%16%
Calf10%90%
Pig25%75%
Man0–3%97–100%

CCK/CCKA receptor involvement in acute pancreatitis. Evidence from animal studies suggest CCK functioning via CCKA receptors may be involved in both the induction as well as the development of acute experimentally-induced pancreatitis (Niederau et al. 1986; Beglinger 1999; Niederau & Grendell 1999; Otsuki 2000). The evidence for this statement includes the following: first, intravenous administration of large doses of CCK analogues in rats and mice can cause pancreatitis (Adler et al. 1979; Niederau et al. 1985; Otsuki 2000). Second, in the CDE model of pancreatitis in mice (choline-deficient, ethionine-supplemented), CCK worsens the pancreatitis (Niederau et al. 1985). Third, OLETF rats with a selective defect in the CCKA receptor develop less severe pancreatitis induced by taurocholate intraductal infusion, by a closed duodenal loop, or by intraperitoneal L-arginine administration (Tachibana et al. 1997). Fourth, CCKA receptor antagonists (lorglumide, asperlicin, devazepide, loxiglumide, proglumide) administration in various models of acute pancreatitis in rats and mice reduce the severity of the pancreatitis (Niederau et al. 1985; Beglinger 1999; Gorelick & Otani 1999; Niederau & Grendell 1999).

Ethanol is one of the most frequent causes of pancreatitis (Gorelick & Otani 1999; Pandol et al. 1999). In vitro (Katz et al. 1996) and in vivo (Pandol et al. 1999) studies in rats provide evidence that ethanol sensitizes the acinar cells to CCK's ability to cause pancreatitis/alter cell function.

At present there is little information available to assess the applicability of these findings to acute pancreatitis in man. Little is known about the early pathogenesis of acute pancreatitis in man, especially that caused by the most common offending agents: alcohol, biliary tract disorders, drugs, and metabolic abnormalities (Gorelick & Otani 1999; Otsuki 2000). Experimental studies suggest one of the critical initiating steps is the intracellular activation of pancreatic enzymes (Gorelick & Otani 1999; Steer 1999) and the activation of proinflammatory cytokines (especially TNFα, interleukin 1β) (Beglinger 1999; Gorelick & Otani 1999; Steer 1999). Studies in hereditary pancreatitis support the role of premature activation of pancreatic zymogens because a mutation has been found in the cationic trypsinogen gene, which makes it less trypsin resistant (Whitcomb et al. 1996; Whitcomb & Ulrich 1999; Whitcomb 1999; Teich et al. 2000). In models of experimental pancreatitis in animals including that due to large doses of CCK analogues, similar evidence for the importance of intracellular enzyme activation has been provided (Gorelick & Otani 1999; Steer 1999). However, these resemblances in early pathogenesis between experimental animal studies and rare hereditary forms of pancreatitis, do not clearly establish that CCK is important in the pathogenesis of the human disorder as it has been shown in experimental pancreatitis in animals (Beglinger 1999; Gorelick & Otani 1999; Niederau & Grendell 1999; Steer 1999).

Only one study (Ochi et al. 1999) has reported the use of a CCKA receptor antagonist (loxiglumide) in the treatment of acute pancreatitis in a preliminary form. Although this was a double-blind study, it was not placebo-controlled. In this study 189 patients in 104 Japanese centers with acute pancreatitis were treated with one of three doses of loxiglumide (100, 300, 500 mg/day) given intravenously twice a day for 14 days. Of the 178 patients available for analysis, the rate of pain disappearance, changes in nausea or vomiting, changes in physical findings, and changes in serum amylase were similar in the three groups (Ochi et al. 1999). Serum lipase levels remained higher longer in the low-dose group and it was concluded based on all parameters assessed that high-dose loxiglumide (500 mg) showed quicker and better efficacy than the other two doses. Adverse side effects (5%) were low in this study and generally mild. This study suggests CCKA receptor antagonists could be of use to treat acute pancreatitis in man; however, placebo-controlled, double-blind studies are needed to clearly establish this.

CCK/CCKA receptor involvement in chronic pancreatitis. Chronic pancreatitis can not only result in pancreatic insufficiency, it is associated with abdominal pain (Mossner 1999; Otsuki 2000; Toskes 2001). Whether CCK is involved in the pathogenesis of pain in this disease is controversial (Mossner 1999; Toskes 2001). It has been proposed by some (Toskes 2001) that CCK is involved in the pathogenesis of this pain, perhaps by a failure of normal feedback inhibition of CCK release. In rats, chickens and pigs, studies demonstrate the presence of pancreatic enzyme proteases in the duodenum inhibits pancreatic enzyme secretion via a negative feedback, likely mediated by degrading protease-sensitive CCK releasing peptides (Mossner et al. 1992; Otsuki 2000). Whether feedback inhibition of CCK occurs in man and whether it contributes to the pain in chronic pancreatitis is controversial (Mossner et al. 1992; Jin et al. 1994). Some authors (Isaksson & Ihse 1983; Slaff et al. 1984) have reported treatment of patients with chronic pancreatitis with pancreatic enzyme preparations causes pain relief, especially those with mild to moderate disease (abnormal secretin test, no steatorrhea) and minimal to no changes on ERCP (Toskes 2001). In contrast, others have reported no pain reduction in these patients with pancreatic enzyme reduction (Halgreen et al. 1986; Mossner et al. 1992; Malesci et al. 1995). It has been proposed that the therapeutic effect of pancreatic enzyme extracts in some studies is due to the presence of feedback inhibition of CCK in man (Toskes 2001). Of the five placebo-controlled trials of pancreatic extracts in chronic pancreatitis, two studies (Isaksson & Ihse 1983; Slaff et al. 1984) have reported a decrease in abdominal pain, suggesting the proposal of possible feedback inhibition of CCK might be contributing to this effect and three studies (Halgreen et al. 1986; Mossner et al. 1992; Malesci et al. 1995) reported no decrease in abdominal pain. A possible factor contributing to this difference in results in the five studies is whether a conventional or enteric-coated enzyme preparation was used (Mossner 1999; Toskes 2001). The two studies using conventional enzyme preparations both demonstrated pain reduction and the three (Halgreen et al. 1986; Mossner et al. 1992; Malesci et al. 1995) using an enteric preparation did not. It is proposed the enteric preparation may release the enzymes after the duodenum, where they would be ineffective in causing feedback inhibition because the protease-sensitive releasing factors are in the duodenum. Therefore, while a number of the studies reviewed above and other studies (Dlugosz et al. 1988; Jin et al. 1994; Mossner 1999) suggest negative feedback regulation of pancreatic secretion may play a physiological role in humans, whether it is in fact contributing to the pain of chronic pancreatitis, is unproven.

Role of CCK and CCKA receptor in clinical motility disorders.

CCK/CCKA receptor and motility – General. Recent studies using CCKA receptor agonists and specific antagonists in man (Lembcke et al. 1987; Meyer et al. 1989; Fried et al. 1991; Schmidt et al. 1991 & 1994; Schwizer et al. 1997; Kreiss et al. 1998), as well as similar studies in animals (Dlugosz et al. 1988; Wiley & Owyang 1987) and studies of CCK or CCKA receptor knockout mice (Shoji et al. 1997; Suzuki et al. 2001) show that physiological concentrations of CCK have important effects on gastrointestinal motility. These effects include stimulation of postprandial gallbladder contraction, inhibition of gastric emptying and inhibition of colonic transit. At present, whether CCK or CCKA receptor alterations are contributory to any human motility disorder except those covered under CCK Deficiency is unclear. It has been proposed that either CCKA receptor activation or an alteration in its activation is involved in the generation of dyspeptic symptoms in patients with functional dyspepsia (Feinle et al. 2001), the generation of pain in patients with acalculous cholecystitis (Behar et al. 1989; Behar 1999; Reitter & Aaning 1999; Amaral et al. 2001), the defective gallbladder contraction in patients with cholesterol gallstones and acalculous cholecystitis (Behar et al. 1989; Behar 1999; Xiao et al. 1999 & 2000; Amaral et al. 2001) and the symptoms and altered motility in patients with irritable bowel syndrome (Kellow et al. 1988; Sjolund et al. 1996; Chen et al. 2001). The possible role of CCK and CCKA receptor activation in these disorders will be briefly discussed below.

Role of CCKA receptors in gallbladder disease. Patients with acalculous cholecystitits have impaired gallbladder contraction with infusion of CCK compared to controls or patients with pigmented gallstones (Barr et al. 1997; Amaral et al. 2001). A recent detailed study (Amaral et al. 2001) of the gallbladder muscle from patients with acalculous cholecystitis showed in vitro, the muscle also showed a defective response to CCK. However, this impairment was also seen with general G-protein activation and there was no impairment of binding to the CCKA receptor. Therefore, the contractile abnormality was not specific for CCKA receptor activation and was due to a general defect in the contractile apparatus (Amaral et al. 2001). These results differed from patients with cholesterol cholelithiases, whose gallbladders also have a CCK contractile defect (Behar et al. 1989; Behar 1999; Amaral et al. 2001). In contrast to the case with acalculous cholecystitis gallbladders, gallbladder muscle from patients with cholesterol stones has a defect which resides in the plasma membrane of the muscle (Xiao et al. 1999 & 2000; Amaral et al. 2001). Altered membrane fluidity results in dysfunction of G-protein-coupled transmembrane receptors such as the CCKA receptor (Behar 1999; Xiao et al. 2000; Amaral et al. 2001) and the signal transduction cascade distal to the receptor is normal.

The role of CCK in acalculous gallbladder disease and the disease itself as a cause of chronic abdominal pain remains controversial (Strasberg 1995; Ott 1998; Lillemoe 1997; Reitter & Aaning 1999). In an attempt to identify symptomatic patients with acalculous cholecystitis who might benefit from cholecystectomy, numerous studies have attempted to use either clinical response or alterations in gallbladder emptying with either CCK administration or a fatty meal to release endogenous CCK (Strasberg 1995; Barr et al. 1997; Philippe et al. 1997; Ott 1998; Smythe et al. 1998; Ziessman 1999; Chen et al. 2001; Ziessman 2001; Ziessman et al. 2001). The results have been contradictory (Strasberg 1995; Barr et al. 1997; Ott 1998; Middleton & Williams 2001). One recent study (Barr et al. 1997) reports that by using a slow infusion of CCK and measuring gallbladder volume changes with ultrasound, a sensitivity of 75% and specificity of 100% for acalculous cholecystitis, using pathologic criteria, can be obtained. However, prospective studies need to be performed in a number of centers with sufficient long-term follow-up that give reproducible results to clearly establish, not only the ability of CCK infusion to predict the presence of this disease, but the role of CCK in its pathogenesis and even the existence of acalculous cholecystitis.

Role of CCKA receptors in other human motility disorders. In irritable bowel syndrome and dyspepsia, alterations in CCK release or responses are reported to play a role in their pathogenesis of the symptoms (Sjölund et al. 1996; Schwizer et al. 1997). Numerous studies have reported dysmotility in irritable bowel syndrome (Trotman & Price 1986; Abrahamsson 1987; Kellow et al. 1988; Fisher et al. 1998; Chen et al. 2001; Chey et al. 2001). Some studies have suggested either altered release of CCK (exaggerated) or altered responses to CCK could contribute to the symptoms of irritable bowel syndrome (Snape, Jr. et al. 1977; Sjolund et al. 1996; Chen et al. 2001). At present, there is only a preliminary report on the use of a CCKA receptor antagonist to treat irritable bowel syndrome (Cann et al. 1994). In this pilot multicenter, double-blind parallel design study (Cann et al. 1994), 72 patients with irritable bowel syndrome were treated with placebo, 200, or 400 mg t.i.d. of oral loxiglumide per day for 8 weeks. The 400 mg t.i.d. dose was reported to cause a significant clinical improvement (pain, distension, bowel habit) compared to placebo or the 200 mg t.i.d. dose, and the effect was more evident in constipation-predominant patients. However, until large placebo-controlled double-blind studies are performed, whether CCK contributes to any of the symptoms of patients with irritable bowel syndrome remains unproven.

Many patients with functional dyspepsia have a hypersensitivity to distension of the stomach (Feinle et al. 2001). In one study (Chua et al. 1994) intravenous administration of CCK reproduced symptoms in patients with functional dyspepsia and symptoms were relieved by oral administration of the CCKA receptor antagonist, loxiglumide. In another study (Feinle et al. 2001), the CCKA receptor antagonist, dexloxiglumide, was shown to reduce dyspeptic symptoms due to gastric distension and duodenal lipid infusions in patients with functional dyspepsia and it was concluded that CCKA receptors are involved in the generation of fat-induced dyspeptic symptoms in these patients. Similar to the conclusions with irritable bowel syndrome above, until patients with functional dyspepsia are treated in a placebo-controlled, double-blind trial of CCKA receptor antagonists, it remains unclear the role CCKA receptor activation plays in contributing to their symptoms.

In normal volunteers, the CCKA receptor antagonist, loxiglumide, increased colonic transit (Meyer et al. 1989), suggesting CCK had an inhibitory effect on normal colonic transit. These results raised the possibility that CCKA receptor antagonists could be pro-kinetic agents that could be useful to treat constipation (Meyer-Wyss et al. 1991; Hildebrand & Beglinger 1994). The effect of loxiglumide on chronic constipation was investigated in a randomized, double-blind trial involving 21 chronically constipated elderly patients (mean age, 83 years) (Meier et al. 1993). After 3 weeks of treatment with placebo or loxiglumide (800 mg 3 times daily), colonic transit was assessed. Treatment with loxiglumide significantly (P<0.005) accelerated colonic transit time from 113±6 hr to 81±10 hr, and the stool frequency per week increased significantly from 3.9±0.5 to 4.9±0.5 (P<0.006). Furthermore, the number of aenemas administered in a 3-week period decreased from 2.7±0.6 to 1.3±0.4 with loxiglumide treatment (P<0.005) (Meier et al. 1993). These data demonstrated that blockage of CCKA receptors can significantly improve chronic constipation in geriatric patients.

CCK and CCKA receptors in tumour growth. CCK can stimulate growth of normal and neoplastic tissues by interacting with CCKA or CCKB receptors (Lamers et al. 1990; Shulkes & Baldwin 1997; Baldwin & Shulkes 1998b; Guo & Townsend, Jr. 2000; Rozengurt & Walsh 2001) . In this section only the role of CCKA receptors will be briefly discussed. The possible role of gastrin or CCK activation of CCKB receptors or possibly another not yet identified receptor subtype will be discussed later. CCK administration has been shown to cause pancreatic growth in animals (Solomon et al. 1978; Zucker et al. 1989; Neiderau et al. 1994; Ohlsson et al. 1995, 1998 & 1999b). Endogenous CCKA receptor activation assessed by the use of CCKA receptor antagonists results in stimulation of pancreatic weight in animals (rats, hamsters, mice) in some studies (Ohlsson et al. 1995 & 1998; Neiderau et al. 1994), but not others (Zucker et al. 1989). Furthermore, mice that are CCK-deficient by target disruption of the CCK gene have similar pancreatic weights to wild-type controls fed a similar diet (Lacourse et al. 1999), as do mice with a CCKA receptor knockout (Takiguchi et al. 2002). These results suggest endogenous CCK or CCKA receptor activation is not essential to maintain pancreatic growth and perhaps compensation mechanisms exist to replace the loss of CCK (Lacourse et al. 1999).

In one study (Ohlsson et al. 1999b) in hamsters, CCK infusion increased liver weight; however, 4 weeks of treatment with the CCKA receptor antagonist, devazepide, had no effect on liver weight. In rats, neither hypergastrinaemia (induced by omeprazole or fundectomy) nor hyper-CCKaemia (induced by pancreaticobiliary diversion) stimulated growth of the intact or regenerating liver (Chen et al. 1994).

Numerous experimental studies in animals or in vitro studies of tumour cells have provided evidence that CCKA receptor activation may be important in both the induction and growth of cancers, especially of the pancreas (Guo & Townsend, Jr. 2000). At present there is no study that unequivocally establishes that CCKA receptor activation is clinically important in human tumours. The role of CCKA receptor activation in pancreatic tumour growth will be briefly discussed below.

Pancreatic cancers and CCKA receptors. Prolonged elevation of CCK levels alone induces the formation of pre-neoplastic lesions in rat pancreas; CCK promotes the development of chemical carcinogen-induced pancreatic cancers in rats and hamsters, and this effect is blocked by CCKA receptor antagonists; and CCK stimulates growth of some pancreatic cancers and inhibited the growth of others (Guo & Townsend, Jr. 2000). In two studies, 90–100% of pancreatic adenocarcinomas expressed CCKA receptors (Weinberg et al. 1997; Moonka et al. 1999) and 100% of pancreatic adenocarcinoma cell lines (n=8) (Mandair et al. 1998). However, even though CCKA receptors were present, CCK did not alter tumour cell function (growth, p125FAK phosphorylation) in one study (Mandair et al. 1998). In another study (Monstein et al. 2001) both CCK and CCKA receptor expression were not present in any pancreatic adenocarcinoma cell line studied, nor was CCK detected in the culture medium from any cell, leading the authors to suggest CCKA receptor activation was unlikely to be involved in their growth.

In one study (Militello et al. 1997) the safety and efficacy of loxiglumide, a CCKA receptor antagonist, was evaluated in a randomized, placebo-controlled, double-blind study in 64 patients with advanced pancreatic cancer. Loxiglumide did not alter survival in this study even when patients were stratified by gender and stage. In a second pilot study (Abbruzzese et al. 1992) the effect of the CCKA receptor antagonist, MK-329, was studied in 18 patients with advanced pancreatic cancer. MK-329 had no effect on tumour progression, pain or nutrition.

CCKA receptors and non-pancreatic tumour growth. In a more recent survey of 406 human tumours, CCKA receptors were expressed in 38% of gastroenteropancreatic tumours, 19% of neuroblastomas, 30% of meningiomas, 0% of colon cancers, and 0–3% of other tumours (Reubi et al. 1997). In general, CCKA receptor expression in human tumours was much less frequent than CCKB receptor expression (Reubi et al. 1997). In one study (Okada 1996) using RT-PCR, 4/14 (29%) of gastric cancers were found to express CCK mRNA and CCKA receptor expression was found in 36% (5/14). Whether CCK or CCKA receptor activation effects growth of gastric cancer, is unknown.

CCK and CCKA receptors' role in satiety. Extensive studies in animals (Reidelberger 1994; Moran 2000) support the conclusion that CCK functioning through gastrointestinal CCKA receptors functions as a satiety factor. Most (Ballinger & Clark 1994; Lieverse et al. 1994 & 1995a; Ballinger et al. 1995; Matzinger et al. 1999; Moran 2000; Rayner et al. 2000; Beglinger et al. 2001; Degen et al. 2001), but not all (French et al. 1994; Lieverse et al. 1995b) studies in man using either infusions of CCK (Lieverse et al. 1994; Ballinger et al. 1995; Rayner et al. 2000), intraduodenal nutrient infusions to increase endogenous CCK (Ballinger & Clark 1994; Lieverse et al. 1994; Lieverse et al. 1995a; Matzinger et al. 1999) or CCKA receptor antagonists (French et al. 1994; Lieverse et al. 1995b; Beglinger et al. 2001) support a role of CCK as a satiety factor function through gastrointestinal CCKA receptors. Recent studies show CCK and leptin have a number of interactions (Lewin & Bado 2001; Attoub et al. 1999; Matson et al. 1997 & 2000). CCK and leptin have an additive effect on satiety, CCK can stimulate gastric leptin release, CCKA receptors are involved in leptin release, and CCK and leptin have an additive effect on the NTS in the CNS. At present the exact role of CCK in human obesity remains unclear as does the possible therapeutic potential of CCK analogues to treat obesity (Moran 2000; Degen et al. 2001).

A number of studies (Geracioti, Jr. & Liddle 1988; Geracioti et al. 1989; Devlin et al. 1997) suggest CCK could be involved in another disorder involving altered human weight control, bulimia nervosa. Patients with bulimia have recurrent episodes of binge eating and vomiting (Geracioti, Jr. & Liddle 1988). These patients have reduced meal-stimulated CCK release and this defect is resolved with tricyclic antidepressant-induced amelioration of the bulimic behavior (Geracioti, Jr. & Liddle 1988). A study (Devlin et al. 1997) showed that the decreased CCK release after a meal is associated with decreased gastric emptying. It was proposed (Devlin et al. 1997) that increased gastric capacity in these patients, perhaps resulting from repeated binge eating, gives rise to delayed gastric emptying and decreased postmeal CCK release which leads to an impaired satiety response, which may perpetuate the illness.

Clinically important disorders involving gastrin or CCKB receptors (table 4)

Table 4.  Gastrointestinal disorders involving or possibly involving gastrin-related peptides and their receptors.
I. Proven:

General. In contrast to CCKA receptors, the role of gastrin CCKB receptor activation is well defined in a number of gastrointestinal disorders. Gastrin is an important mediator of gastric mucosal growth (hyperplasia of enterochromaffin-like cells (ECL cells) which can progress to the formation of gastric carcinoids (Håkanson et al. 1986; Solcia et al. 1993; Bordi et al. 1995 & 1999; Nagata et al. 1996), stimulation of gastric mucosal growth, increased parietal cells), and is an important mediator of gastric acid secretion in various acid-peptic diseases and in various acid hypersecretory studies including Zollinger-Ellison syndrome (Walsh 1994; Smith et al. 1995; Jensen 1996; DeWeerth et al. 1999; Ohlsson et al. 1999a; Goetze et al. 2000; Szabo et al. 2000). Numerous studies suggest gastrin may have an important growth effect on various gastrointestinal malignancies because they frequently overexpress or ectopically express CCKB receptors (Baldwin 1995; Rehfeld 1995; Guo & Townsend, Jr. 2000; Smith & Watson 2000b; Rozengurt & Walsh 2001). Furthermore, gastrin itself or gastrin-related peptides, functioning through either the CCKB receptor or an unknown receptor, may have an important growth effect in some tumors, especially colon cancer (Baldwin 1995; Rehfeld 1995; Singh et al. 1996; Baldwin & Shulkes 1998a; McWilliams et al. 1998; Guo & Townsend, Jr. 2000; Smith & Watson 2000a & b; Rozengurt & Walsh 2001). Each of these areas will be reviewed briefly below.

Gastrin or CCKB receptor alterations in clinical gastrointestinal disease.

Hypergastrinaemic/hypogastrinaemic states. There are no clinically important diseases specifically due to hyposecretion of gastrin. There are two main categories of human hypergastrinaemic states: physiological responses due to acid hyposecretion/achlorhydria in such disorders as pernicious anemia/atrophic gastritis (table 3), and hypergastrinaemia associated with hypersecretion of gastric acid (Jensen 1996) (table 3). Chronic hypergastrinaemia in both of these categories can be of clinical significance because it can result in gastric ECL cell hyperplasia and the development of gastric carcinoid tumours (Bordi et al. 1974, 1986 & 1995; Solcia et al. 1993; Håkanson et al. 1994; Peghini et al. 2002). This will be discussed. Hypergastrinaemia with acid hypersecretion can be caused by a number of conditions (table 3). Of these the most frequent is H. pylori infections, which in a proportion of infected patients causes hypergastrinaemia and hyperchlorhydria (Metz et al. 1995; el-Omar et al. 1995). The most virulent hypersecretory disease occurs in the Zollinger-Ellison syndrome due to the presence of a neuroendocrine tumour ectopically secreting gastrin (i.e., a gastrinoma) (Jensen & Gardner 1993; Jensen 1997).

Table 3.  Composition of the test diets.
Causes of chronic hypergastrinaemia.

Strictly speaking a gastrinoma includes any tumour that produces gastrin and thus the term would not be synonymous with Zollinger-Ellison syndrome. Zollinger-Ellison syndrome is the clinical syndrome due to ectopic hormone release by a gastrinoma, which results in hypergastrinaemia causing gastric acid hypersecretion (Jensen 1997). Gastrin mRNA or various forms of gastrin peptides have been detected in bronchogenic carcinoma, acoustic neuromas, pheochromocytomas, ovarian carcinomas, colorectal carcinomas, and other pancreatic endocrine tumour syndromes than Zollinger-Ellison syndrome (Kochman et al. 1992; Rehfeld & van Solinge 1994; Walsh 1994; Dockray et al. 2001). Normal and tumourous tissue give rise to identical cDNA clones (Rehfeld & van Solinge 1994), which suggests that the overexpression of the gastrin gene may be due to an alteration of the regulatory regions of the gene, such as in the 5′ untranslated region, resulting in an altered transcription rate or altered stability of the gene-specific mRNA. These results in ovarian carcinomas are of particular interest. In a recent study (van Solinge et al. 1993) either amidated gastrin, glycine-extended gastrin, or progastrin was detected in 12 ovarian serous cystadenocarcinomas, three ovarian non-differentiated carcinomas, and five ovarian serous cystadenomas, mucinous cystadenomas, or follicular cysts. In 50% of the malignant ovarian tumours, significant concentrations of amidated gastrin were found. These results demonstrate that, in contrast to bronchogenic carcinomas, acoustic neuromas, and most colon cancers, ovarian tumours can fully process the progastrin to the biologically active amidated form (van Solinge et al. 1993; Rehfeld & van Solinge 1994). This result is particularly interesting because there are numerous case reports of Zollinger-Ellison syndrome occurring in patients with ovarian tumours (Maton et al. 1989; Jensen & Gardner 1993), but hypergastrinemia does not generally occur in bronchogenic carcinoma, colorectal carcinoma, or acoustic neuromas (Rehfeld & van Solinge 1994).

Because tumours can produce gastrin without producing hypergastrinaemia and a clinically important syndrome (i.e., Zollinger-Ellison syndrome), clinically the term ‘gastrinoma’ is used synonymously with Zollinger-Ellison syndrome in most papers (Jensen 1997). Clinically significant gastrinomas producing the Zollinger-Ellison syndrome until recently were thought to be entirely intra-abdominal in location (duodenum >pancreas >lymph nodes >liver >other abdominal sites) (Jensen & Gardner 1993; Jensen 1997; Jensen 2001). However, studies show they can occur in the heart (Gibril et al. 1997; Noda et al. 1999) and, due to ectopic gastrin release, by a non-small cell lung cancer (Abou-Saif et al. 2001).

The study of Zollinger-Ellison syndrome has provided important insights into the effects of chronic hypergastrinemia in man (Jensen 1993). These studies demonstrate hypergastrinaemia can cause marked gastric acid hypersecretion which can result in severe refractory peptic ulcer disease, malabsorption and gastroesophageal reflux disease (Ellison & Wilson 1964; Jensen 1993; Mignon et al. 1995; Roy et al. 2000 & 2001); can result in increased gastric mucosal thickness and increased parietal cell mass 4 to 6 times with no increase in peptic cells (Polacek & Ellison 1966; Rosenlund 1967; Sum & Perey 1969; Neuburger et al. 1972); and result in a mean 2-fold increase in mucosal argyrophil cells (Bardram et al. 1986; Helander 1986; Solcia et al. 1986; Lehy et al. 1989; D'Adda et al. 1990; Maton et al. 1990). Of the seven types of gastric endocrine cells the increase in gastric argyrophil cells with hypergastrinaemia is due to an increase only in the ECL cells (D'Adda et al. 1990), which is similar to results in animal studies (Havu 1986; Hakanson & Sundler 1986; Larsson et al. 1986; Creutzfeldt 1988). No other clinical effects of chronic hypergastrinaemia have been clearly demonstrated from studies of patients with Zollinger-Ellison syndrome, especially in regard to increased tumour growth or frequency of tumours in other sites. These points will be discussed in more detail in the following sections.

CCKB receptor alterations in diseases. Recently CCKB receptor mutations were reported in pancreatic cancer (Ding et al. 2002), colon cancer (Hellmich et al. 2000; Laghi et al. 2002), and gastric cancers (Laghi et al. 2002). A misspliced form of the CCKB receptor in which intron 4 is retained is reported in pancreatic (Ding et al. 2002) and colon cancer (Hellmich et al. 2000). In pancreatic cancer (Ding et al. 2002) the expression of this misspliced receptor form was linked to decreased levels of the U2 small nuclear ribonucleoprotein particle auxiliary splicing donor (U2HF35) (Ding et al. 2002). The abnormal spliced receptor was shown to have constitutive activity and to have a trophic effect on cells expressing it (Hellmich et al. 2000), suggesting its presence can have important growth-promoting effects. In 43 gastrointestinal tumours with a high degree of microsatellite instability due to defects in the DNA mismatch repair genes, frameshift mutations were found in the CCKB receptor gene in 19% (Laghi et al. 2002). Frameshift mutations occurred in 23% of the gastric cancers studied, 13% of sporadic and 20% of hereditary colorectal carriers, and all tumours also had frameshift mutations in other genes (Laghi et al. 2002). The gastrin-sensitive LoVo colorectal cancer cell line also showed a similar frameshift mutation in the CCKB receptor gene (Laghi et al. 2002). The results in this study led the authors to propose that the human CCKB receptor gene is a new candidate target gene possibly playing a role in the tumourigenesis of a fraction of MSI tumours. In a study of obese, diabetic patients, 2 of 18 type-2 diabetes mellitus families were found to have a V125I mutation in the CCKB receptor (Marchal-Victorion et al. 2002). This receptor, when expressed in COS-7 cells, had an increased affinity for CCK and enhanced potency for stimulated phospholipase C activity assessed by measuring formation of inositol phosphates (Marchal-Victorion et al. 2002). Co-segregation studies showed the mutation was not associated with diabetes or early age at diagnosis of the disease. At present, the role of this mutation in the pathogenesis of either the obesity or diabetes mellitus in the families remains unclear.

Gastrin and CCKB receptors in gastric ECL cell hyperplasia/carcinoids. umerous studies in animals demonstrate that manipulations that result in chronic hypergastrinaemia (antisecretory drug treatment, gastrin infusions, surgical procedures) can lead to gastric ECL proliferation and in some cases (rat, mouse, mastomys) the development of gastric carcinoids (Poynter et al. 1985 & 1986; Havu 1986; Håkanson et al. 1986; Sundler et al. 1986; Creutzfeldt 1988; Ryberg et al. 1990; Mattsson et al. 1991), which can occasionally be malignant (Wangberg et al. 1995). It is proposed these carcinoids arise from the ECL cell through a progression of ECL cell proliferative changes from increasing hyperplasia (linear, micronodular, adenomatoid) to dysplasia and carcinoid formation (Creutzfeldt 1988; Solcia et al. 1988 & 2000; Bordi et al. 1995). A similar sequence of events may occur in man because various human conditions with chronic hypergastrinaemia (especially atrophic gastritis/pernicious anemia and Zollinger-Ellison syndrome (table 3)) as well as in patients with chronic acid suppressive treatment, ECL cell proliferative changes occur (Borch et al. 1985; Creutzfeldt 1988; Lamberts et al. 1988 & 1993; Solcia et al. 1988; Maton et al. 1990; Lehy et al. 1989 & 1992; Cadiot et al. 1993; Bordi et al. 1995; Delle Fave et al. 1998; Peghini et al. 2002). Furthermore, in patients with atrophic gastritis (Borch et al. 1985; Solcia et al. 1991; Rindi et al. 1996; Delle Fave et al. 1998) and with Zollinger-Ellison syndrome with multiple endocrine neoplasia-type 1 (MEN1) (Rindi et al. 1996) and rarely without MEN1 (Jensen 1993; Feurle 1994; Cadiot et al. 1995; Rindi et al. 1996), gastric carcinoid tumors occur.

Increasing attention is being given to the effect of chronic hypergastrinaemia on ECL cell proliferation for two reasons. First, increasing numbers of patients with idiopathic gastroesophageal reflux disease/PUD are being treated long-term with potent acid suppressants such as PPIs and in 80–100% of patients (Jansen et al. 1990; Lamberts et al. 1993) hypergastrinaemia develops and in 20–30% of patients the fasting gastrin levels can reach levels frequently seen in Zollinger-Ellison syndrome (Jansen et al. 1990; Jensen 1993; Lamberts et al. 1993; Ligumsky et al. 2001). Second, in hypeergastrinaemic states a subset (4–30%) of gastric carcinoids in hypergastrinaemic states are malignant (Rindi et al. 1996 & 1999).

Whereas the above studies establish that chronic hypergastrinaemia can have a proliferative effect on ECL cells, which can contribute to the development of gastric carcinoids, the exact role of hypergastrinaemia per se in human gastric carcinoid development has remained unclear until recently (Peghini et al. 2002). This has occurred because studies of patients with the most frequently studied chronic hypergastrinaemic states (atrophic gastritis, chronic PPI use) demonstrate that both the hypergastrinaemia, grade of corporal gastritis and/or atrophy are closely correlated with the degree of ECL hyperplasia (Borch et al. 1985; Solcia et al. 1991; Lamberts et al. 1993). In addition, in animal studies gender and age of the animal are important predisposing factors to the development of gastric carcinoids (Ekman et al. 1985; Havu 1986; Solcia et al. 2000). The contribution of these different factors in humans has been difficult to define (Peghini et al. 2002).

A recent study (Peghini et al. 2002) was able to address a number of these points because it assessed gastric ECL changes in 106 patients with the sporadic form of Zollinger-Ellison syndrome. These patients infrequently have gastritis or atrophy and therefore are a good model to study the effects of hypergastrinaemia alone in man. Furthermore, at least one-half of these patients have fasting gastrin levels in the range seen with chronic PPI treatment (Lamberts et al. 1988; Jansen et al. 1990; Jensen 1993; Peghini et al. 2002). In this study (Peghini et al. 2002) only 1% of the patients with active disease had a normal ECL cell pattern and 52% had at least linear hyperplasia or a more advanced qualitative ECL change, with 7% having dysplasia and 0% carcinoid tumours. Fasting gastrin levels correlated closely with the magnitude of the ECL change (P<0.0001) and no threshold effect of gastrin on ECL cell hyperplasia was found, as proposed by others (Coupe et al. 1990; Bordi et al. 1995; Modlin & Tang 1996). In contrast to animal studies, gender (Ekman et al. 1985; Poynter et al. 1985; Cui et al. 2000), age, or vagal tone Håkanson et al. 1986; Axelson et al. 1988; Cao et al. 1991) did not affect the degree of ECL change. These results show the risk of developing gastric carcinoids with chronic hypergastrinaemia alone in man is low; however, even mild hypergastrinaemia can cause ECL hyperplastic changes without a threshold effect.

Gastrin and CCKB receptors in peptic ulcer disease. One of the most important clinical effects of gastrin is its ability to stimulate gastric acid secretion (Walsh 1994). Although studies in various species show the parietal cell possesses CCKB receptors, evidence suggests the principal pathway of stimulation of acid secretion by gastrin is by stimulating release of histamine from ECL cells (Walsh 1994; Chen et al. 2000; Lindström et al. 2001). Gastrin is the major hormonal mediator of the gastric phase of acid secretion (Walsh 1994) because gastrin immunoneutralization of circulating gastrin completely inhibits acid secretion stimulated by peptone or glucose-induced gastric distension (Kovacs et al. 1989). Gastrin also plays a variable role in the cephalic and intestinal phases in different species (Walsh 1994). CCKB receptor antagonists in rats inhibit gastrin-stimulated acid secretion and histamine release from ECL cells (Håkanson et al. 1999; Chen et al. 2000; Björkqvist et al. 2001).

In man spiroglumide, a CCKB receptor antagonist, inhibited acid secretion stimulated by a meal as well as sham feeding stimulated secretion (Beltinger et al. 1999). These results support a physiological role for gastrin in regulating acid secretion in man. The central role of the CCKB receptor mediating the action of gastrin on acid secretion is also supported by the results of CCKB receptor and gastrin knockout studies in mice (Langhans et al. 1997; Friis-Hansen et al. 1998; Hinkle & Samuelson 1999). In both cases the mice have an elevated gastric pH and no not secrete acid in response to gastrin (Friis-Hansen et al. 1998; Hinkle & Samuelson 1999).

Gastrin also has a trophic effect on parietal cells and chronic hypergastrinaemia results in an increased parietal cell mass (Eissele et al. 1992; Jensen & Gardner 1993; Koh & Chen 2000). The parietal cell mass correlates with the maximal acid output. In various chronic hypergastrinaemic states in man, such as Zollinger-Ellison syndrome, both increased parietal cell mass and an increased maximal acid output are frequently found (Polacek & Ellison 1966; Rosenlund 1967; Sum & Perey 1969; Neuburger et al. 1972; Jensen & Gardner 1993; Roy et al. 2001).

It is now known that Helicobactor pylori (H. pylori) infections are the principal cause of duodenal ulcer disease; however, the exact mechanisms are still unclear (Soll 1998; Calam 1999; Dore & Graham 2000; McColl et al. 2000). The exact role of abnormalities in gastrin secretion or responses in H. pylori-mediated duodenal ulcer disease remain unclear (Soll 1998; Calam 1999; Dore & Graham 2000; McColl et al. 2000). It is known that a proportion of patients with duodenal ulcers have an increased parietal cell mass resulting in an increased MAO, an exaggerated acid and gastrin release with meals or gastrin-releasing peptide administration, an altered sensitivity to gastrin, impairment in inhibiting responses mediating secretion of acid and gastrin, in addition to an increased basal acid output in 30% of these patients (Soll 1998).

Duodenal ulcer patients with H. pylori characteristically have an antrum-dominant, body-sparing, non-atrophic gastritis (Calam 1999; McColl et al. 2000)which results in increased acid secretion and gastrin release. The increased gastrin release is mediated primarily by an impairment of the acid-mediated inhibitory control of gastrin release, which is mediated by somatostatin release from antral D cells (Moss et al. 1992; Calam 1999; McColl et al. 2000). Lower levels of somatostatin are found in the antral mucosa of H. pylori infected subjects and they increase post-eradication of the H. pylori (Moss et al. 1992; Calam 1999; McColl et al. 2000). Functional studies suggest that alterations in the ability of CCK functioning through CCKA receptors may be involved in the impairment of acid-mediated inhibitory control of gastrin release in H. pylori infected subjects (Konturek et al. 1995a & b). CCK acting via CCKA receptors on antral D cells stimulate somatostatin release, which has an inhibitory effect on gastrin secretion from G cells (Eissele et al. 1991; Schmidt et al. 1994; Calam 1999). The CCKA receptor antagonist, loxiglumide, increased acid output with a meal in healthy controls, but not in duodenal ulcer patients (Konturek et al. 1995b). Eradicating H. pylori resulted in correction of the abnormal response to CCKA receptor blockage in duodenal ulcer patients to that seen in healthy controls (Konturek et al. 1995a & b. The mechanism by which H. pylori infection or accompanying gastritis alters the acid-inhibitory control of gastrin release and somatostatin levels is not completely clear (Calam 1999; McColl et al. 2000). Possible mechanisms include secondary to increased cytokine production, and alterations induced by ammonia production by H. pylori (Levi et al. 1989; Beltinger et al. 1999; Calam 1999).

It has been proposed in some studies (McColl et al. 2000), but not others (Dore & Graham 2000), that alterations in gastrin regulation can explain most of the acid secretory abnormalities seen in patients with H. pylori infection. This includes the proposal that 1) the increased acid output is, in large part, due to the altered gastrin release, and this results in increased duodenal acid load which progressively damages the duodenal mucosa leading to gastric metaplasia and eventually to duodenal ulcers; and 2) the increased gastrin release results in an increased BAO and the trophic effects of gastrin cause the increased MAO (McColl et al. 2000).

Gastrin, gastrin-related peptides on normal and tumour growth (non-ECL cell growth).

In addition to their growth effects on the gastric mucosa and gastric ECL cells (discussed previously), gastrin and gastrin-related peptides have been reported to have numerous growth effects in the gastrointestinal tract both on normal tissues and on cancer growth (Lamers et al. 1990; Baldwin 1995; Nagata et al. 1996; Shulkes & Baldwin 1997; Baldwin & Shulkes 1998a; Guo & Townsend, Jr. 2000; Koh & Chen 2000; Rozengurt & Walsh 2001). Particularly controversial is the role of gastrin, gastrin-related peptides and their possible receptors in colon cancer (Baldwin 1995; Rehfeld 1995; Baldwin & Shulkes 1998a & b; Koh & Chen 2000; Smith & Watson 2000a & b; Dockray et al. 2001; Rozengurt & Walsh 2001). Therefore, the possible clinical importance of the effects of gastrin and related peptides in colon cancer and in growth of other gastrointestinal neoplasms will be briefly considered separately below.

General. Almost all of the evidence of the growth effects of gastrin and/or gastrin-related peptides in gastrointestinal malignancies is from studies in animals (tumour growth studies, transgene expression in animals, etc.) and studies of isolated human cell preparations. At present no studies have clearly established in man that therapies directed at gastrin, gastrin analogues or their receptors will result in effective tumour treatment in man.

Possible importance of gastrin and gastrin-related peptides in human colonic neoplasia. The possible involvement of gastrin or gastrin-related peptides in the growth and/or development of colon cancer has received considerable attention over the last 15 years since their growth effects on these tumours and their presence in these tumours with their receptors were first described in colorectal cancer cells (Singh et al. 1985; Townsend Jr. et al. 1989; Hoosein et al. 1988 & 1990; Smith & Solomon 1988; Rehfeld 1995; Guo & Townsend, Jr. 2000; Smith & Watson 2000b). Since then a number of studies have shown that gastrin can stimulate the growth of colonic cancers, isolated colon cancer cells, and that CCKB receptor inhibitors can inhibit their growth (Townsend, Jr. et al. 1989; Baldwin 1995; Chu et al. 1995; Rehfeld 1995; Orbuch et al. 1996; Smith & Watson 2000b).

At present it is still controversial, with considerable differences in results in the literature about the type of gastrin-related peptide that is present in human colorectal cancers (Baldwin 1995; Dickinson 1995; Rehfeld 1995; Baldwin & Shulkes 1998b; Smith & Watson 2000b). Whereas most studies demonstrate the presence of progastrin (87–100%), the percentage of tumours possessing amidated gastrin varies from 0–100% (mean – 52% (n=6)) and glycine gastrin from 44–100% (mean – 63% (n=3)). The importance of these precursor forms is receiving increased attention because of recent reports of their growth-promoting effects (Baldwin & Shulkes 1998b; Koh & Chen 2000; Dockray et al. 2001). Glycine gastrins have been shown to have growth effects in a number of different cells (Seva et al. 1994; Baldwin & Shulkes 1998b; Koh & Chen 2000; Dockray et al. 2001) and mice overexpressing either glycine gastrin (Koh et al. 1999) or progastrin (Wang et al. 1996) develop hyperplastic colonic mucosa.

Whether hypergastrinaemia occurs or not in patients with colorectal cancers is also controversial, with divergent results in different studies (Suzuki et al. 1988; Smith et al. 1989; Yapp et al. 1992; Penman et al. 1994; Ciccotosto et al. 1995; Rehfeld 1995; Thorburn et al. 1998; Smith & Watson 2000b). Some of this difference was due to a failure to control for the presence or absence of H. pylori in some studies. In one large recent study (Thorburn et al. 1998) which was a nested case control study among 128,992 patients in the Kaiser Health Maintenance Program, a significantly higher proportion of colon cancer patients had serum gastrin >90 pg/ml, probably due to an H. pylori infection. One study assessed serum levels of both gastrin and non-amidated gastrins and found both H. pylori positive and negative patients with colorectal cancers had higher plasma non-amidated gastrins, but not amidated gastrin levels (Ciccotosto et al. 1995).

Controversy exists about whether colorectal cancers or colorectal cancer cell lines express CCKB receptors, some other non-CCKA/CCKB receptor, or none of these (Baldwin 1995; Rehfeld 1995; Baldwin & Shulkes 1998b; Koh & Chen 2000; Smith & Watson 2000b). Furthermore, a truncated form of the CCKB receptor has been described in colon cancer cell lines as well as a splice variant of the CCKB receptor that has constitutive activity (McWilliams et al. 1998; Hellmich et al. 2000). Although there are a limited number of binding studies to colorectal tumours or cell lines, in general these studies show that CCKA and CCKB receptors are likely present on only a small percentage of these tumours (Baldwin 1995; Hellmich et al. 2000).

The role of gastrin or gastrin-related peptides as an autocrine growth factor in colorectal cancer is also controversial (Baldwin 1995; Rehfeld 1995; Baldwin & Shulkes 1998a & b; Smith & Watson 2000b). In patients with hypergastrinemia due to either atrophic gastritis (Brinton et al. 1989; Creutzfeldt & Lamberts 1991; Maton 1995) or Zollinger-Ellison syndrome (Sobhani et al. 1993; Orbuch et al. 1996; Smith & Watson 2000b) in most studies, there is no increased occurrence of either colonic adenomas or cancers. However, in patients with Zollinger-Ellison syndrome there was increased rectal cell proliferation found in one study (Renga et al. 1997) and colonic mucosal cell proliferation in a second study (Sobhani et al. 1993). In vitro studies generally do not support a role for an autocrine loop involving amidated gastrin in colorectal cancers (Baldwin & Shulkes 1998b). However, the possibility of an autocrine loop involving non-amidated gastrins is still an open question and is supported by some studies using gastrin antisense constructs (Baldwin & Shulkes 1998b; Singh et al. 1996).

At present, there are no published results of large prospective trials in patients with colorectal cancer attempting to assess the possible therapeutic effect of inhibiting the action of gastrin. Recent studies show that antibodies directed against the amino terminus of gastrin, which inhibit gastrin-17 and glycine-extended forms of gastrin-17, decrease the growth or development of colorectal tumours in a number of experimental models (Smith & Watson 2000b; Watson et al. 2000; Watson & Gilliam 2001; Watson & Smith 2001). This is accomplished in vivo by administering gastrimmune, a conjugate of diphtheria toxoid and gastrin-17 (1–9) (Smith et al. 2000; Smith & Watson 2000b; Watson et al. 2000; Watson & Gilliam 2001; Watson & Smith 2001). Recently, results of a phase II multicenter, sequential group, open label study involving 50 patients with advanced colorectal cancer was reported (Smith et al. 2000). The study involved a primary and two booster intramuscular injections of gastrimmune. Eighty percent of the patients produced a measurable antibody response and the treatment was well tolerated systematically; however, 46% had myalgias at the injection site and 14% developed a sterile abscess at the injection site. Two of 14 patients assessed demonstrated disease stabilization (Smith et al. 2000). At present, additional studies are needed to determine whether this approach alone or in an adjuvant setting has a role in the treatment of colonic cancer or other cancers (gastric, pancreatic) which in experimental studies gastrin-related peptides affect their growth. In one study (Morris et al. 1990) the effect of the non-selective CCKB receptor antagonist, proglumide, was assessed in 41 patients with advanced colorectal cancer. Proglumide did not alter survival or increased tumour regression.

Possibility of gastrin and CCKB receptors in growth of other tumours (non-colon). Numerous studies in animals as well as studies of tumour cells have shown growth effects of gastrin and/or CCKB receptor activation on cancers of the pancreas (Baldwin 1995; Sjölund et al. 1996; Guo & Townsend, Jr. 2000), stomach (Baldwin 1995; Iwase et al. 1997; Szabo et al. 2000), and liver (Caplin et al. 1999 & 2001). Similarly, studies using CCKB receptor antagonists have inhibited growth of tumour and/or cultured tumor cells from the pancreas (Baldwin 1995; Smith et al. 1995; Ohlsson et al. 1999a) and stomach (Baldwin 1995). Furthermore, gastrin and/or the CCKB receptor have been shown to be expressed by cancers of the stomach (Baldwin 1995; Clerc et al. 1997; McWilliams et al. 1998; Henwood et al. 2001), pancreas (Smith et al. 1995; Mandair et al. 1998; DeWeerth et al. 1999; Goetze et al. 2000; Monstein et al. 2001), and liver (Caplin et al. 1999 & 2001). At present the clinical importance of CCKB receptors on these tumours remains unclear. There are no controlled studies using highly selective CCKB antagonists alone or in combination with other antitumour agents in these tumours that demonstrate their importance as a growth factor in vivo. In one study the effectiveness of the relatively non-selective CCKA/CCKB receptor antagonist, proglumide, was studied in 110 patients with gastric cancer (Harrison et al. 1990). Proglumide did not increase survival or cause increased tumour regression.

Results from studies in gastric and pancreatic cancer using gastrimmine are reported in preliminary form only (Brett et al. 2000; Watson & Gilliam 2001). It is reported in these preliminary reports that gastrimmune increased the duration of survival in patients with advanced pancreatic carcinoma (Brett et al. 2000; Watson & Gilliam 2001).

Ancillary