Gastrins, iron and colorectal cancer

Abstract

This minireview explores the connections between circulating gastrins, iron status and colorectal cancer. The peptide hormone gastrin is a major regulator of acid secretion and a potent mitogen for normal and malignant gastrointestinal cells. Gastrins bind two ferric ions with μM affinity and, in the case of non-amidated forms of the hormone, iron binding is essential for biological activity. The ferric ion ligands have been identified as glutamates 7, 8 and 9 in the 18 amino acidpeptideglycine-extended gastrin. An interaction between gastrin and transferrin was first demonstrated by covalent crosslinking techniques, and has been recently confirmed by surface plasmon resonance. We have therefore proposed that gastrins act as catalysts in the loading of transferrin with iron. Several recent lines of evidence, including the facts that the concentrations of circulating gastrins are increased in mice and humans with the iron overload disease haemochromatosis, and that transferrin saturation positively correlates with circulating gastrin concentrations, suggest that gastrins may be involved in iron homeostasis. In addition the recognition that ferric ions may play an unexpected role in the biological activity of non-amidated gastrins may assist in the development of new therapies for colorectal carcinoma.

---

Graham S. Baldwin

Graham Baldwin is an Associate Professor and NHMRC Senior Research Fellow at the University of Melbourne, Australia. Graham first began investigating the role of the peptide hormone gastrin as a growth factor in gastrointestinal cancer when he joined the Melbourne Branch of the Ludwig Institute for Cancer Research in 1981. Since transferring to the Department of Surgery at Austin Health in 1994, Graham’s research has concentrated on the interactions between gastrins and metal ions, and on ways of utilising those interactions to develop novel therapies for colorectal cancer.

---

The peptide hormone gastrin was originally identified as a stimulant of gastric acid secretion.¹ More recently gastrin has been demonstrated to act as a growth factor in the normal gastrointestinal mucosa and in gastrointestinal cancers. The observation that gastrins bind ferric ions led to the discovery that iron is essential for the biological activity of non-amidated forms of the hormone. This minireview will consider the connections between gastrin and iron homeostasis, and the possibility of targeting the gastrin–iron complex in the development of novel therapies for colorectal cancer.

Gastrins and gastrin receptors

All forms of gastrin are generated from an 80 amino acid precursor, progastrin,² which is cleaved at dibasic sequences by prohormone convertases within G cells in the gastric antrum^3,4 (Fig. 1). Subsequent removal of Arg⁷³ and Arg⁷⁴ by carboxypeptidase E yields glycine-extended gastrin₁₇ (ZGPWLEEEEEAYGWMDFG, Ggly), and oxidative cleavage of the glycine N–Cα bond by peptidylglycine α-amidating monooxygenase results in the C-terminally amidated form gastrin₁₇ (ZGPWLEEEEEAYGWMDFamide, Gamide).³ Gastrins may also be post-translationally modified by sulfation of the sole tyrosine residue.⁵

Fig. 1 Gastrin processing. Preprogastrin (101 amino acids) is converted to progastrin (80 amino acids) by removal of its signal peptide (black box).³ The sequential action of prohormone convertases and carboxypeptidase E in the Golgi network then converts the prohormone to glycine-extended gastrin₃₄ (progastrin_38–72) and glycine-extended gastrin₁₇ (progastrin_55–72, Ggly). The extent of conversion of glycine-extended gastrin₃₄ to amidated forms such as gastrin₃₄ (amidated progastrin_38–71) and gastrin₁₇ (amidated progastrin_55–71, Gamide) by peptidylglycine α-amidating monooxygenase is dependent on the tissue. Both non-amidated and amidated forms are independently active via different receptors. Gastrins may also be post-translationally modified by sulfation of the sole tyrosine residue.⁵

The two best-characterised receptors for amidated gastrin peptides also bind the related hormone cholecystokinin (CCK) and are therefore called the CCK1 and CCK2 receptors. The CCK1 receptor is found in the pancreas and brain and binds sulfated CCK with 500–1000-fold greater affinity than non-sulfated CCK or Gamide.⁶ The CCK2 receptor was discovered in the brain and stomach, and binds both sulfated and non-sulfated forms of Gamide and CCK with similar affinity.⁶ As the C-terminal sequence WMDFamide is required for full agonist activity at the CCK2 receptor, glycine-extended gastrins and progastrin are not endogenous ligands for this receptor, and the structure of the Ggly receptor is still unknown.⁷

The field of gastrin research was advanced dramatically by the discovery that non-amidated forms of gastrin are also biologically active. For example Ggly stimulates proliferation in the pancreatic cell line AR4-2J⁸ and in other cell lines of gastrointestinal origin.^9–11 In contrast to Gamide, which exerts its biological effects on the gastric mucosa, the predominant in vivo target for non-amidated gastrins is the colorectal mucosa. Transgenic mice which overexpress progastrin (hGAS)¹² or Ggly (MTI/Ggly)¹³ showed two-fold increases in proliferation in the colonic mucosa compared with wild-type controls. The MTI/Ggly mice also had increased colonic mucosal thickness, an increased number of goblet cells per crypt and an expansion of the proliferative zone into the upper third of the colonic crypts.¹³ Conversely gastrin-deficient (Gas^−/−) mice show decreased proliferation in the colonic mucosa,¹⁴ and continuous infusion with Ggly for two weeks resulted in an increase in circulating Ggly concentrations, in colonic mucosal thickness and in proliferation in the colonic mucosa, compared to control Gas^−/− mice that received saline alone.¹⁴ Similarly short-term administration of Ggly to rats after colostomy increased proliferation in the rectal mucosa, and after administration of the carcinogen azoxymethane the Ggly-treated rats had increased numbers of tumour precursors in the colorectal mucosa.¹⁵

Gastrins and iron: structure–function studies

Both Ggly and Gamide bind two ferric ions in aqueous solution.^16,17 A stoichiometry of 2.0 ± 0.3 and an apparent dissociation constant of 0.6 ± 0.2 μM were measured for Ggly at pH 4.0 and 25 °C by absorption and fluorescence spectroscopy, respectively.¹⁶ Fluorescence quenching experiments with peptides derived from the Ggly sequence indicated that one or more of the five glutamic acid residues were necessary for iron binding.¹⁶ The solution structure of Ggly was subsequently determined by NMR spectroscopy (Fig. 2). The iron-binding ligands were identified by the effects of ferric ions on individual resonances in the total correlation spectroscopyspectrum and by comparison of the fluorescence quenching by ferric ions of Ggly and of mutant peptides in which one or more of the five glutamates were replaced by alanines. Glutamate 7 was a ligand at the first ferric ion binding site (and may also contribute to binding of the second ferric ion), and glutamate 8 and glutamate 9 were ligands at the second ferric ion binding site.¹⁸

Fig. 2 Hypothetical model of the gastrin–ferric ion complex. The solution structure of Ggly in 10% Me₂SO, 10% ²H₂O, 80% H₂O, pH 5.3 was determined by NMR spectroscopy.¹⁸ Ggly and Gamide both bound two ferric ions,¹⁶ the first viaglutamate 7 and the second by glutamates 8 and 9.¹⁸ Binding of ferric ions was essential for the biological activity of Ggly,¹⁸ but not Gamide.²² The side chains of the glutamate residues are shown in red, the peptide backbone in blue, and the ferric ions in purple. The positions of the ferric ions in the model, with glutamate 7 acting as a bridging ligand for both ferric ions, and glutamates 8 and 9 each interacting with one of the ferric ions, are hypothetical, but are consistent with the available experimental data. The stability of the progastrin–iron complex (t_1/2 = 117 days at pH 7.6 and 25 °C ¹⁹) suggests that other parts of the progastrin structure may also interact with the ferric ions.

The properties of the progastrin–iron complex were further characterised with recombinant human progastrin_6–80.¹⁹ The stoichiometry (2.5 ± 0.1) and dissociation constant for ferric ion binding (2.2 ± 0.1 μM) for progastrin at pH 4.0 were similar to the values observed for Ggly.¹⁹ The greater size of progastrin_6–80 (M_r 8427) compared to Ggly (M_r 2156) also permitted measurement of the selectivity of metal ion binding and the stability of the ⁵⁹Fe(III)–progastrin complex by equilibrium dialysis. Of the four divalent and seven trivalent non-radioactive metal ions tested, only ferric and ferrous ions competed with ⁵⁹Fe(III) for binding to progastrin at pH 2.8. Furthermore the progastrin–iron complex was extremely stable, with a half life of 117 days in the presence of EDTA at pH 7.6 and 25 °C.

The effect of tyrosinesulfation on iron binding was investigated with the related hormone CCK₈ (DYMGWMDFamide), which has the same C-terminal amidated pentapeptide as Gamide.²⁰ Although no binding of ferric ions to CCK₈ was detected at pH 4.0, the changes in absorbance and fluorescence emission observed on addition of ferric ions at pH 6.5 indicated that tyrosinesulfation of CCK₈ increased the stoichiometry from 1 to 2, without greatly affecting the affinity (0.6–2.8 μM). The small but significant changes observed in the NMR signals from aspartate1 and aspartate7 on addition of ferric ions were consistent with a distant interaction between the carboxylate sidechains and the bound metal ion. By analogy with the results obtained for CCK₈, sulfation of gastrins is not expected to alter substantially their affinity for ferric ions.

The biological role of ferric ion binding in stimulation of both proliferation and migration by Ggly was investigated in the mouse gastric cell line IMGE-5. Binding of ferric ions to Ggly was essential for the peptide’s biological activity as the iron chelator desferrioxamine inhibited the biological action of Ggly and as the GglyE7A mutant was biologically inactive.¹⁸ Further studies of Ggly fragments revealed that the heptapeptides LE₅A and E₅AY each bound two ferric ions, and that iron binding to the heptapeptides was essential for biological activity.²¹ In contrast desferrioxamine had no effect on the binding of Gamide to the CCK2 receptor,²² or of sulfated CCK8 to the CCK1 receptor.²⁰ Furthermore the observation that the biological activity of Gamide was unaffected by desferrioxamine clearly indicated that ferric ions were not essential in this case, and supported the hypothesis that amidated and non-amidated gastrins act through different receptors.²²

Gastrins and transferrin saturation

Serum transferrin plays a key role in iron metabolism by transporting iron to body cells expressing transferrin receptors. Thus reduction of serum transferrin concentration in the hpx mouse²³ results in death from severe anaemia within 14 days of birth.²⁴Transferrin receptor 1 (TfR1) is the major receptor for cellular iron uptake via the endocytic pathway and is ubiquitously expressed on cells which acquire iron.²⁵Transferrin receptor 2 (TfR2), on the other hand, is mainly found on hepatocytes and developing erythroid precursor cells and is upregulated in response to increased transferrin saturation.^26,27 Both receptors bind iron-loaded transferrin with high affinity, but the affinity of TfR2 is lower than TfR1. The observation that mutations in TfR2 cause a rare autosomal recessive form of the iron overload disease hereditary haemochromatosis demonstrates the importance of TfR2 in the regulation of iron homeostasis.²⁸

Both Gamide and Ggly have been shown to interact with the iron transportprotein transferrin. The interaction with Gamide was first detected in extracts of porcine gastric mucosa using covalent cross-linking assays , which demonstrated that the concentration of Gamide required to reduce cross-linking to transferrin by 50% was approximately 100 μM.²⁹ A more detailed ultracentrifugal study subsequently revealed that iron-free (apo-) transferrin bound two molecules of Gamide with a K_d of 6.4 μM at pH 7.4.³⁰ No significant binding of Gamide to diferric-transferrin was detected under the same conditions. Interaction between apo-transferrin and Ggly was first detected by surface plasmon resonance, and the fact that no interaction was observed in the presence of the chelator EDTA suggested that the gastrin–ferric ion complex was the interacting species.³¹

A mechanism has been proposed for the connection observed between changes in circulating Gamide concentrations and changes in transferrin saturation.³² The mechanism is based on the fact that efficient loading of apo-transferrin requires the presence of an anion such as bicarbonate or an anionic chelator such as nitrilotriacetate.³³ Our working hypothesis envisages that, after export of ferrous ions from the enterocyte by ferroportin and their oxidation to ferric ions by hephaestin, circulating gastrins may act as chaperones for the uptake of ferric ions by apo-transferrin. The failure to detect significant binding of gastrin to diferric-transferrin³⁰ suggests that gastrin dissociates after iron transfer has occurred, and hence plays a catalytic role consistent with the difference in the circulating concentrations of gastrin and transferrin. Changes in the circulating concentration of diferric-transferrin may in turn cause significant alterations in iron traffic throughout the body by altering the production by the liver of the regulatory peptide hepcidin.³⁴

Gastrins and iron: biological connections

In addition to the strong evidence that a gastrin–ferric ion complex exists in solution and can interact with apo-transferrin, accumulating data supports the existence of a physiological connection between iron status and gastrins. Mutations in the HFE gene are the most common cause of the iron-overload disease hereditary haemochromatosis, and Gamide and Ggly concentrations were increased in the gastric mucosa and plasma of Hfe^−/− mice, and in the sera of patients with HFE-related haemochromatosis.³⁵ The observed changes in gastrin in the Hfe^−/− mice were not due to structural changes in the gastric mucosa, which contained normal numbers of parietal cells, or to reduced gastric acid production, as the pH of the luminal contents was lower in Hfe^−/− mice than in wild-type animals.

Conversely, modulation of circulating gastrin concentrations was also shown to alter iron homeostasis. In juvenile Gas^−/− mice intestinal iron absorption measured by ⁵⁹Fe uptake after oral gavage was increased sixfold, and concentrations of the mRNA encoding the divalent metal transporter-1 (DMT-1) were increased fourfold, compared with age-matched wild-type mice.³² Although there was no increase in iron absorption in hypergastrinaemic Cck2r^−/− mice, DMT-1 mRNA was 5.4-fold higher than in age-matched wild-type mice. Importantly transferrin saturation was reduced 0.8-fold in Gas^−/−mice, and increased 1.5-fold in Cck2r^−/−mice, compared with age-matched wild-type mice. Similarly, in humans hypergastrinaemic because of multiple endocrine neoplasia type I, transferrin saturation correlated positively with circulating Gamide concentrations.³² This correlation is consistent with the proposal outlined above that gastrins catalyse the binding of iron by transferrin.

Relevance to colorectal cancer

There is now abundant evidence that non-amidated gastrins accelerate the development of colorectal carcinoma (CRC).^7,36 CRC patients have increased circulating concentrations of Gamide (3.9-fold) and total gastrins (5.2-fold),³⁷ and the concentration of progastrin in CRC (1.2 ± 0.5 pmol g⁻¹) is higher than in surrounding normal mucosa (0.1 ± 0.1 pmol g⁻¹).³⁸ Thorburn and coworkers have provided additional evidence that increased circulating Gamide concentrations contribute to the increased risk of colon cancer.³⁹ Furthermore Ggly is active in mouse models of colon cancer. In the APC Min^+/−mouse model of the early stages of CRC the number of intestinal polyps is increased either by infusion of Ggly,⁴⁰ or by crossing Min^+/− mice with MTI/Ggly mice that over-express Ggly.¹³ Progastrin has also been designated as a co-carcinogen since transgenic mice that overexpress progastrin develop more tumour precursors⁴¹ and colon cancers⁴² than wild-type mice in response to the mutagen azoxymethane. In vitro, Ggly stimulates proliferation and migration of the mouse colon cell line YAMC,⁹ and invasion and migration of the human colorectal cell line LoVo.¹¹ In other cell lines Ggly also stimulates proliferation¹⁰ and the stimulation can be blocked by expression of gastrin antisense constructs⁴³ or by immunization with a gastrin-diphtheria toxin conjugate (gastrimmune).⁴⁴

Dietary iron has been identified as a risk factor for CRC.^45,46 A population-based, case-control study demonstrated that mutations in the HFE gene that are responsible for body iron overload are associated with increased risk of CRC.⁴⁷In vitro studies have shown that progression to CRC is associated with increased expression of iron import proteins and decreased expression of iron export proteins, possibly because of greater oxidative damage from the resultant increase in intracellular iron.⁴⁸ Alternatively, since progastrin and Ggly act as colonic growth factors, the increased circulating Ggly concentrations in the plasma of haemochromatotic patients and Hfe^−/− mice may contribute to CRC development.³⁵ Although the mechanism underlying the involvement of gastrins in iron homeostasis is still unknown, non-amidated gastrins and iron may synergistically promote the development of CRC.

Inhibition of gastrin–iron complex formation

The recognition that the complexes between iron and non-amidated gastrins act as growth factors in CRC suggests at least two new therapeutic options for this all-too-common disease. One approach to disruption of the gastrin–iron complex is treatment with chelating agents.¹⁸ The iron-selective chelatordesferrioxamine inhibited the ability of Ggly to stimulate both cell migration and proliferation in vitro,¹⁸ but had no effect on the biological activity of Gamide.¹⁷ Unpublished data from our laboratory indicates that daily injection of rats with desferrioxamine also blocks the stimulatory effect of Ggly on cell proliferation in the normal colorectal mucosa.⁴⁹

A second approach utilises the trivalent bismuth ion, which is an analogue of the trivalent ferric ion and binds to both ferric ion-binding sites of transferrin.^50,51 Bismuth has the added advantage that its salts have been used for the treatment of gastrointestinal disorders such as ulcers and diarrhoea for many years.⁵² Binding of Bi³⁺ ions to Ggly or Gamide was detected by fluorescence quenching²² and, as with Fe³⁺ ions, NMR studies indicated that the side chains of glutamates 7, 8 and 9 were acting as metal ligands.²² Importantly, Bi³⁺ ions competitively inhibited the ability of Ggly to stimulate both cell migration and proliferation in vitro.²²

The immediate challenge is to test the effectiveness of either chelating agents or bismuth ions in animal models representative of the different stages of CRC. The three principal models available are polyp development in Min^± mice, representative of the early stages of CRC,⁴⁰ growth of CRC cell lines as xenografts in nude mice, representative of the growth of an established cancer,⁵³ and the splenic injection model, representative of liver metastasis.⁵⁴ Treatment of thalassemic patients with desferrioxamine or with newer oral iron chelators,⁵⁵ or of gastrointestinal disorders with oral bismuth,⁵² does not adversely affect iron homeostasis in the short term. Hence, blockade of the stimulatory effects of Ggly in any of these models by prevention of iron binding would open the way for future trials in human CRC.

Abbreviations

CCK	cholecystokinin
CCK2R	cholecystokinin2 receptor
CRC	colorectal cancer
DMT-1	divalent metal transporter-1
Ggly	glycine-extended gastrin17
Gamide	amidated gastrin17
TfR1	transferrin receptor 1
TfR2	transferrin receptor 2

Acknowledgements

This work was supported in part by grants from the National Health and Medical Research Council of Australia (400062, 454322) and the National Institutes of Health (5RO1GM065926-07). We thank Dr Kevin Barnham for many helpful discussions and for the preparation of Fig. 2.

References

1. J. S. Edkins, Proc. Roy. Soc. B, 1905, 76, 376.
2. O. J. Yoo, C. T. Powell and K. L. Agarwal, Proc. Natl. Acad. Sci. U. S. A., 1982, 79, 1049–1053.
3. G. J. Dockray, A. Varro, R. Dimaline and T. Wang, Annu. Rev. Physiol., 2001, 63, 119–139.
4. S. Kovac, A. Shulkes and G. S. Baldwin, Curr. Opin. Endocrinol. Diabetes Obes., 2009, 16, 79–85.
5. S. J. Brand, J. Klarlund, T. W. Schwartz and J. F. Rehfeld, J. Biol. Chem., 1984, 259, 13246–13252.
6. A. Shulkes and G. S. Baldwin, Clin. Exp. Pharmacol. Physiol., 1997, 24, 209–216.
7. A. Aly, A. Shulkes and G. S. Baldwin, Biochim. Biophys. Acta, 2004, 1704, 1–10.
8. C. Seva, C. J. Dickinson and T. Yamada, Science, 1994, 265, 410–412.
9. F. Hollande, A. Imdahl, T. Mantamadiotis, G. D. Ciccotosto, A. Shulkes and G. S. Baldwin, Gastroenterology, 1997, 113, 1576–1588.
10. V. M. Stepan, D. F. Krametter, M. Matsushima, A. Todisco, J. Delvalle and C. J. Dickinson, Am. J. Physiol.: Cell Physiol., 1999, 277, R572–581.
11. S. Kermorgant and T. Lehy, Biochem. Biophys. Res. Commun., 2001, 285, 136–141.
12. T. C. Wang, T. J. Koh, A. Varro, R. J. Cahill, C. A. Dangler, J. G. Fox and G. J. Dockray, J. Clin. Invest., 1996, 98, 1918–1929.
13. T. J. Koh, G. J. Dockray, A. Varro, R. J. Cahill, C. A. Dangler, J. G. Fox and T. C. Wang, J. Clin. Invest., 1999, 103, 1119–1126.
14. T. J. Koh, J. R. Goldenring, S. Ito, H. Mashimo, A. S. Kopin, A. Varro, G. J. Dockray and T. C. Wang, Gastroenterology, 1997, 113, 1015–1025.
15. A. Aly, A. Shulkes and G. S. Baldwin, Int. J. Cancer, 2001, 94, 307–313.
16. G. S. Baldwin, C. C. Curtain and W. H. Sawyer, Biochemistry, 2001, 40, 10741–10746.
17. J. Pannequin, J. P. Tantiongco, S. Kovac, A. Shulkes and G. S. Baldwin, J. Endocrinol., 2004, 181, 315–325.
18. J. Pannequin, K. J. Barnham, F. Hollande, A. Shulkes, R. S. Norton and G. S. Baldwin, J. Biol. Chem., 2002, 277, 48602–48609.
19. G. S. Baldwin, Protein J., 2004, 23, 65–70.
20. G. S. Baldwin, M. F. Bailey, B. P. Shehan, I. Sims and R. S. Norton, Biochem. J., 2008, 416, 77–84.
21. H. He, B. P. Shehan, K. J. Barnham, R. S. Norton, A. Shulkes and G. S. Baldwin, Biochemistry, 2004, 43, 11853–11861.
22. J. Pannequin, S. Kovac, J. P. Tantiongco, R. S. Norton, A. Shulkes, K. J. Barnham and G. S. Baldwin, J. Biol. Chem., 2004, 279, 2453–2460.
23. J. I. Huggenvik, C. M. Craven, R. L. Idzerda, S. Bernstein, J. Kaplan and G. S. McKnight, Blood, 1989, 74, 482–486.
24. N. C. Andrews, Nat. Rev. Genet., 2000, 1, 208–217.
25. M. Wessling-Resnick, Annu. Rev. Nutr., 2000, 20, 129–151.
26. A. Robb and M. Wessling-Resnick, Blood, 2004, 104, 4294–4299.
27. M. B. Johnson and C. A. Enns, Blood, 2004, 104, 4287–4293.
28. C. Camaschella, A. Roetto, A. Cali, M. De Gobbi, G. Garozzo, M. Carella, N. Majorano, A. Totaro and P. Gasparini, Nat. Genet., 2000, 25, 14–15.
29. G. S. Baldwin, R. Chandler and J. Weinstock, FEBS Lett., 1986, 205, 147–149.
30. S. C. Longano, J. Knesel, G. J. Howlett and G. S. Baldwin, Arch. Biochem. Biophys., 1988, 263, 410–417.
31. S. Kovac, A. Ferrand, J.-P. Estève, A. Mason and G. S. Baldwin, FEBS J., 2009 DOI:10.1111/j.1742-4658.2009.07186.x , in press.
32. S. Kovac, K. Smith, G. J. Anderson, J. R. Burgess, A. Shulkes and G. S. Baldwin, Am. J. Physiol.: Gastrointest. Liver Physiol., 2008, 295, G855–861.
33. G. W. Bates and M. R. Schlabach, J. Biol. Chem., 1973, 248, 3228–3232.
34. D. M. Frazer and G. J. Anderson, Blood Cell. Mol. Dis., 2003, 30, 288–297.
35. K. A. Smith, S. Kovac, G. J. Anderson, A. Shulkes and G. S. Baldwin, FEBS Lett., 2006, 580, 6195–6198.
36. S. A. Watson, A. M. Grabowska, M. El-Zaatari and A. Takhar, Nat. Rev. Cancer, 2006, 6, 936–946.
37. G. D. Ciccotosto, A. McLeish, K. J. Hardy and A. Shulkes, Gastroenterology, 1995, 109, 1142–1153.
38. W. W. Van Solinge, F. C. Nielsen, L. Friis-Hansen, U. G. Falkmer and J. F. Rehfeld, Gastroenterology, 1993, 104, 1099–1107.
39. C. M. Thorburn, G. D. Friedman, C. J. Dickinson, J. H. Vogelman, N. Orentreich and J. Parsonnet, Gastroenterology, 1998, 115, 275–280.
40. T. J. Koh, C. J. Bulitta, J. V. Fleming, G. J. Dockray, A. Varro and T. C. Wang, J. Clin. Invest., 2000, 106, 533–539.
41. P. Singh, M. Velasco, R. Given, M. Wargovich, A. Varro and T. C. Wang, Am. J. Physiol.: Gastrointest. Liver Physiol., 2000, 278, G390–399.
42. P. Singh, M. Velasco, R. Given, A. Varro and T. C. Wang, Gastroenterology, 2000, 119, 162–171.
43. P. Singh, A. Owlia, A. Varro, B. Dai, S. Rajaraman and T. Wood, Cancer Res., 1996, 56, 4111–4115.
44. S. A. Watson, D. Michaeli, S. Grimes, T. M. Morris, G. Robinson, A. Varro, T. A. Justin and J. D. Hardcastle, Cancer Res., 1996, 56, 880–885.
45. R. G. Stevens, D. Y. Jones, M. S. Micozzi and P. R. Taylor, New Engl. J. Med., 1988, 319, 1047–1052.
46. P. Knekt, A. Reunanen, H. Takkunen, A. Aromaa, M. Heliovaara and T. Hakulinen, Int. J. Cancer, 1994, 56, 379–382.
47. N. J. Shaheen, L. M. Silverman, T. Keku, L. B. Lawrence, E. M. Rohlfs, C. F. Martin, J. Galanko and R. S. Sandler, J. Natl. Cancer Inst., 2003, 95, 154–159.
48. M. J. Brookes, S. Hughes, F. E. Turner, G. Reynolds, N. Sharma, T. Ismail, G. Berx, A. T. McKie, N. Hotchin, G. J. Anderson, T. Iqbal and C. Tselepis, Gut, 2006, 55, 1449–1460.
49. S. Lachal, PhD thesis, University of Melbourne, 2005.
50. H. Li, P. J. Sadler and H. Sun, J. Biol. Chem., 1996, 271, 9483–9489.
51. H. Sun, H. Li, A. B. Mason, R. C. Woodworth and P. J. Sadler, J. Biol. Chem., 2001, 276, 8829–8835.
52. S. L. Gorbach, Gastroenterology, 1990, 99, 863–875.
53. A. Ferrand, M. S. Sandrin, A. Shulkes and G. S. Baldwin, Biochim. Biophys. Acta, 2009, 1793, 477–488.
54. J. H. Neo, C. Malcontenti-Wilson, V. Muralidharan and C. Christophi, J. Gastroenterol. Hepatol., 2007, 22, 577–584.
55. E. J. Neufeld, Blood, 2006, 107, 3436–3441.

---