Oral trivalent bismuth ions decrease, and trivalent indium or ruthenium ions increase, intestinal tumor burden in Apc^Δ14/+ mice

Abstract

Immature forms of the peptide hormone gastrin have been implicated in the development of colorectal cancer (CRC). The biological activity of glycine-extended gastrin (Ggly) is dependent on the binding of Fe³⁺ ions in vitro and in vivo. The aim of the present study was to determine the effect of blocking Fe³⁺ ion binding to Ggly, using Bi³⁺, In³⁺ or Ru³⁺ ions, on the development of intestinal tumors in APC^Δ14/+ mice. APC^Δ14/+ mice were treated orally with Bi³⁺, In³⁺ or Ru³⁺ ions for up to 60 days, serum trace metals were analyzed by inductively coupled plasma mass spectrometry, and the incidence and size of intestinal tumors were assessed. Bi³⁺ treatment significantly decreased the number of tumors larger than 3 mm in male mice. In³⁺ or Ru³⁺ treatment significantly increased the tumor burden in all animals and In³⁺ increased the number of tumors larger than 3 mm or 5 mm in male mice alone. The fact that binding of In³⁺ or Ru³⁺ ions to Ggly was orders of magnitude stronger than the binding of Bi³⁺ ions implies that the inhibitory effect of Bi³⁺ ions is not a consequence of a reduction in Ggly activity. However, further testing of higher doses of Bi³⁺ ions for longer periods as an oral treatment for intestinal tumors is warranted.

---

Introduction

Gastrins are a collective term for mature and immature peptides derived from the gastrin gene. Amidated gastrin (Gamide), the fully processed gut hormone, activates the cholecystokinin 2 (CCK2) receptor to stimulate gastric acid secretion and maintain the gastric mucosa.^1,2 Immature forms of gastrin (e.g., progastrin and glycine-extended gastrin (Ggly)) are also biologically active independently of the CCK2 receptor and act primarily in the colon to stimulate cell division,^3–6 migration, and survival.^1,4 Immature gastrins are found in colonic polyps and adenocarcinomas,⁷ and have been implicated in the development of colorectal cancer (CRC). Thus greater concentrations of immature gastrins are found in the circulation of CRC patients,^8,9 and hypergastrinemic patients have a 4-fold risk of developing CRC.^8,10,11 Although several candidates for the Ggly receptor have been identified (e.g. annexin II,¹² F1-ATPase α-subunit¹³), small molecule inhibitors have not as yet been developed.

Most human sporadic CRCs contain an inactivating mutation in the APC (adenomatous polyposis coli) tumor suppressor gene.¹⁴ APC mutant mice provide a model of spontaneous intestinal cancers in vivo without the use of carcinogens.^14,15 Koh et al. crossed APC^min−/+ mice that spontaneously develop intestinal tumors, primarily in the small intestine, with either mice overexpressing glycine-extended gastrin (MTI/Ggly) or gastrin knockout mice (GASKO).¹⁶ The data showed that mice prone to develop intestinal tumors had enhanced tumor size and numbers when Ggly was over-expressed and significantly reduced tumor incidence when all gastrins were absent. The development of APC^Δ14/+ mice, with a deletion of exon 14 engineered into one allele of their Apc tumor suppressor gene, has provided a better model of familial adenomatous polyposis, as the proportion of spontaneous tumors developing in their large intestine is greater than with APC^min−/+ mice.¹⁷

Strong links between excess iron and CRC have also been reported.^18,19 Study of APC^min/+ mice with either iron-deficient, normal or excess iron diets found that high iron drove intestinal tumorigenesis, as tumor number and size correlated with dietary iron status.²⁰ Using GASKO mice and CCK2 receptor knockout mice (which have high circulating gastrin concentrations), Kovac et al. found that gastrins have a direct role in hematopoiesis under iron deficiency.²¹ GASKO mice on a low iron diet had impaired erythropoiesis and developed splenomegaly. Hypoxia studies indicated that gastrins may play a protective role under low oxygen conditions, possibly via catalysis of iron binding to transferrin.^22,23

Work to elucidate the mechanism of action of gastrins revealed that ferric ion binding to glutamates 7 and 8/9 in the sequence QGPWLEEEEEAYGWMDF was essential for biological activity of Ggly, but not Gamide.^24–26 Bioactivity of Ggly was diminished or destroyed by deleting these glutamates or by treating with the chelator desferrioxamine (DFO).^25,27 Trivalent Bi³⁺ ions can displace ferric ions from the iron-binding sites on gastrins, with a concomitant reduction in the biological activity of Ggly in vitro.^28,29In vivo, Ggly-induced stimulation of proliferation in the defunctioned rectal mucosa of rats could be abrogated with oral Bi³⁺ treatments, and oral Bi³⁺ also decreased proliferation rates in the colonic mucosa in both hGAS (over-expressing progastrin) and MTI/Ggly mice (over-expressing glycine-extended gastrin).²⁹ In addition, DFO restored the increased colorectal mucosa proliferation rates in mice overexpressing progastrin (hGAS) or Ggly (MTI/Ggly) to the levels observed in wild-type control mice.³⁰

Recently trivalent In³⁺ and Ru³⁺ ions have been shown to form stable, high affinity complexes with gastrins.^26,31 Ru³⁺ and In³⁺ ions were both bound to the iron-binding site of gastrins orders of magnitude more tightly than ferric ions. Because the biological activity of Ggly was dependent on the presence of ferric ions, the possibility arose of using Bi³⁺, In³⁺ or Ru³⁺ ions as competitive inhibitors of the biological activity of Ggly. As inhibitors of the Ggly receptor are not yet available, the aim of the present study was to determine the effect of blocking ferric ion binding to Ggly using Bi³⁺, In³⁺ or Ru³⁺ ions on the incidence and size of intestinal tumors in APC^Δ14/+ mice.

Experimental

Reagents and dosage

All consumables were obtained from Sigma Aldrich (Castle Hill, Australia), unless noted otherwise. Bismuth citrate, indium citrate and ruthenium citrate were all suspended in phosphate-buffered saline (ThermoFisher Scientific, Waltham, MA, USA), and the concentrations described below are based on the metal ion. The dose of Bi³⁺ used was derived from previous work in which 141 mg Bi³⁺ per kg per day was active, but not toxic, in mice treated daily for 4 weeks.²⁹ This dose is 70-fold less concentrated than the single 50% lethal dose (LD50) of 10 000 mg per kg in mice. A single Bi³⁺ dose of 2000 mg per kg (1/5 of LD50), or Bi³⁺ doses of 1000 mg per kg per day (1/10 of its LD50) repeated daily for 28 days in rats, also showed no adverse effects.^32,33 A similar study with oral In³⁺ in rats also demonstrated that a single dose of 2000 mg per kg (1/5 of its LD50 of 10 000 mg per kg in mice), or an In³⁺ dose of 1000 mg per kg per day (1/10 of its LD50) repeated daily for 28 days in rats, had no adverse effects.^33,34 Based on this study, the same concentration was used for In³⁺ as for Bi³⁺. The same logic was used to derive the Ru³⁺ dose of 6.5 mg per kg per day based on a concentration 70-fold less than the reported LD50 of 463 mg per kg in mice.³⁵

Mouse model

APC^Δ14/+ mice (C57Bl/6 background) were kindly provided by Dr Julie Pannequin (National Centre for Scientific Research, Institute of Functional Genomics, Montpellier, France). These mice, which have had a heterozygous deletion of exon 14 engineered into one allele of their Apc tumor suppressor gene, provide a model of familial adenomatous polyposis (autosomal dominant disease), with spontaneous tumors developing in their small and large intestines.^17,36

APC^Δ14/+ mice (8–10 weeks old) were divided into four treatment groups and orally gavaged with 0.1–0.15 mL of PBS (16 female and 17 male mice), bismuth citrate (9 female and 8 male mice, 141 mg Bi³⁺ per kg per day), indium citrate (7 female and 8 male mice, 141 mg In³⁺ per kg per day) or ruthenium citrate (7 female and 8 male mice, 6.5 mg Ru³⁺ per kg per day) once a day, 3 times a week, up to 20 weeks of age. Animals were weighed and sickness scores taken daily. Mice were humanely sacrificed at the end of the experiment by isoflurane inhalation and blood was collected by cardiac puncture. Intestines were collected, inverted on wooden skewers, formalin fixed and prepared as Swiss Rolls for counting of intestinal tumors as described.³⁷ All animal work was approved by the Austin Health Animal Research Ethics Committee (AEC13/4957 and AEC15/5307).

Hematological parameters

Full blood examination was performed as previously described²² by the Austin Pathology Laboratory (Heidelberg, VIC, Australia) using an Advia 120 automated hematological analyzer (Bayer, Tarrytown, NY, USA). Serum trace metals were analyzed by inductively coupled plasma mass spectrometry (ICPMS) (Biometals Facility, Florey Institute of Neuroscience and Mental Health, University of Melbourne). Serum from mice was diluted with 1/10 1% (v/v) nitric acid and measured with an Agilent 7700 series ICPMS instrument under routine multi-element operating conditions using a helium gas reaction cell. A certified standard solution containing 200 ppb of Yttrium (Y89) was used as an internal control.

Statistics

Survival statistics were performed using log-rank and cox-regression (SPSS version 22.0, IBM, Armonk, NY, USA). All other statistics were analyzed using either student's t-test or one-way ANOVA (SigmaStat, Jandel Scientific, San Rafael, CA, USA). A P value < 0.05 was considered significant.

Results

Treatment increased metal ion concentrations in the serum

Metal salts were given by oral gavage for up to 60 days at the following doses: Bi³⁺, 141 mg per kg per day; In³⁺, 141 mg per kg per day; Ru³⁺, 6.5 mg per kg per day. At the end of the experiment, concentrations of each metal ion in the serum were measured by ICPMS. All metals were absorbed as the serum concentrations increased significantly in the treated animals (Fig. 1). However, the ratio of Bi³⁺ absorbed to the amount given orally was low (all: 1/980; female: 1/681; male: 1/1,986). The ratios of In³⁺ and Ru³⁺ absorbed to the amounts given orally were closer to unity (In³⁺, all: 1/35; female: 1/92; male: 1/25 and Ru³⁺, all: 1/27; female: 1/64; male: 1/18).

Fig. 1 Serum metal ion analyses. The concentrations of Bi³⁺ (A), In³⁺ (B) or Ru³⁺ (C) in the serum from APC^Δ14/+ mice treated with PBS (control), bismuth citrate, indium citrate or ruthenium citrate were analyzed by inductively coupled plasma mass spectrometry (ICPMS). *P < 0.05; **P < 0.01; ***P < 0.001.

Effect of metal ions on the health of APC^Δ14/+ mice

APC^Δ14/+ mice develop intestinal bleeding and anemia similar to APC^min/+ mice. However, APC^Δ14/+ mice may show earlier signs of decreased health as they develop tumors in both their small and large intestines, while APC^min/+ mice typically develop tumors only in their small intestines.¹⁷ APC^Δ14/+ mice in all treatment groups (PBS, Bi³⁺, In³⁺ or Ru³⁺) had increased sickness scores over time (Fig. S1, ESI†). Male mice appeared to be more sensitive with greater sickness scores (Fig. S1C, ESI†) and greater average weight loss (Fig. S2C, ESI†) than females. Overall, neither Bi³⁺ nor In³⁺ affected mouse survival (Fig. 2A); mice treated with Ru³⁺ died 2.0 times faster than PBS mice (P = 0.079) and the survival of Ru³⁺-treated animals was less than Bi³⁺- or In³⁺-treated animals.

Fig. 2 Survival plots. APC^Δ14/+ mice were treated with PBS (control), bismuth citrate, indium citrate or ruthenium citrate over 60 days. Survival graphs are shown for all (A), female only (B) and male only (C) mice. The difference between PBS- and Ru³⁺-treated mice neared significance as cox-regression analysis indicated Ru³⁺-treated mice died 2.0 times faster than PBS-treated mice (P = 0.079).

APC^Δ14/+ mice have hypoxemia

Anemia is a common outcome of the APC^Δ14/+ mutation due to intestinal bleeding from large polyps. The PBS control and all three treatment groups had significantly decreased hemoglobin concentrations (139 ± 3), hematocrits (0.47 ± 0.01) and red cell counts (9.2 ± 0.2) compared to the values shown in the brackets for untreated C57Bl/6 mice of a similar age (our unpublished data) (Fig. 3A–C). Males, particularly those treated with Ru³⁺, appeared to be more affected. Serum iron concentrations were not significantly different to each other or to the value for untreated wild type animals (26.2 ± 1.5), but In³⁺ and Ru³⁺ appeared to increase or decrease iron levels respectively (Fig. 3E). Treatment with In³⁺ significantly increased WCC compared to the other groups (Fig. 3D), in agreement with previous findings.^34,38 Platelet numbers were significantly reduced in Bi³⁺-treated APC^Δ14/+ mice, to values closer to those observed in untreated wild-type mice (Fig. 3F).

Fig. 3 Hematological parameters. Blood samples from APC^Δ14/+ mice treated with PBS (control), bismuth citrate, indium citrate or ruthenium citrate were analyzed for hemoglobin (A), red blood cell count (B), hematocrit (C), white blood cell count (D), and platelet count (F) by the Austin Pathology Laboratory (Heidelberg, VIC, Australia). Serum iron levels (E) were analyzed by ICPMS. Boxed numbers are the corresponding mean ± SEM values from untreated wild-type C57Bl/6 mice of similar ages. *P < 0.05; **P < 0.01.

Bi³⁺ treatment decreased, and In³⁺ or Ru³⁺ treatment increased, tumor burden in the small intestine

Intestinal tumor size and numbers were determined in the small and large intestines from APC^Δ14/+ mice treated with PBS, Bi³⁺, In³⁺ or Ru³⁺. Bi³⁺ treatment significantly decreased the number of tumors larger than 3 mm in male mice (Fig. 4C). In³⁺ or Ru³⁺ treatment significantly increased the tumor burden in all animals and in male mice alone (Fig. 4A and C). No significant differences were observed in the colon (Fig. 5).

Fig. 4 Tumor burden in small intestine. APC^Δ14/+ mice were treated with PBS, Bi³⁺, In³⁺ or Ru³⁺, and the total number and number of tumors over 3 mm or 5 mm in the small intestine were counted for all mice (A), female only (B) and male only (C). Significance was determined compared to the PBS-treated mice. *P < 0.05; ***P < 0.001.

Fig. 5 Tumor burden in colon. APC^Δ14/+ mice were treated with PBS, Bi³⁺, In³⁺ or Ru³⁺, and the total number and number of tumors over 3 mm or 5 mm in the colon were counted for all mice (A), female only (B) and male only (C). No significance difference was detected.

Discussion

Previous work has revealed that APC^min−/+ mice, which are prone to develop intestinal tumors, had enhanced tumor size and numbers when Ggly was over-expressed and significantly reduced tumor incidence when all gastrins were absent.¹⁶ The observation that Bi³⁺ ions blocked the binding of Fe³⁺ ions to gastrins, and thereby decreased the bioactivity of Ggly in vivo and in vitro,^28,29 suggested that Bi³⁺ treatment might also decrease tumor burden in APC mice. Further work revealed that the binding of In³⁺ or Ru³⁺ ions to gastrins was orders of magnitude stronger than the binding of Bi³⁺ ions.²⁶ The effect of blocking ferric ion binding to Ggly using Bi³⁺, In³⁺ or Ru³⁺ ions on the incidence and size of intestinal tumors in APC^Δ14/+ mice was therefore investigated.

Bi³⁺ treatment significantly decreased the number of small intestinal tumors larger than 3 mm in male mice (Fig. 4C). No reduction in tumor numbers was observed in the colon, or on tumors larger than 5 mm in either small intestine or colon, possibly because of the smaller numbers involved. The mechanism is likely to involve competition by Bi³⁺ for Fe³⁺ ion uptake and/or inhibition of intracellular processes involving iron. The effectiveness of Bi³⁺ treatment might possibly be improved by increasing the dosage or route of administration, as bismuth citrate was poorly absorbed and/or retained in the circulation (Fig. 1), in agreement with previous reports on its poor absorption.^39,40 The observation that Bi³⁺ did not affect circulating iron concentrations or most blood parameters at the dosage used (Fig. 3) suggests that greater dosages might be tolerated, although the significant reduction in platelet counts following Bi³⁺ treatment suggests that a cautious approach would be appropriate.

In contrast to Bi³⁺, both In³⁺ and Ru³⁺ treatments significantly increased tumor numbers in the small intestine of APC^Δ14/+ mice. Again no effect was observed on tumors in the colon, possibly because of the smaller numbers involved. The fact that binding of In³⁺ or Ru³⁺ ions to gastrins was orders of magnitude stronger than the binding of Bi³⁺ ions,²⁶ therefore implies that the inhibitory effect of Bi³⁺ ions is not a consequence of a reduction in Ggly activity. The failure to observe an effect does not appear to be caused by poor absorption as the absorption ratios (μmol L⁻¹ in the serum: amount given orally) for In³⁺ and Ru³⁺ were greater than for Bi³⁺. Furthermore both In³⁺ and Ru³⁺ treatments were effective in other respects. Thus In³⁺ treatment significantly increased white cell counts, in agreement with previous reports that In³⁺ can stimulate the immune system,^34,38 and Ru³⁺ treatment decreased survival. The latter effect may be caused by displacement by Ru³⁺ of circulating iron from transferrin and hemoglobin,^41,42 as ruthenium is in the same column of the periodic table as iron.

Conclusions

Bi³⁺ treatment significantly decreased the number of tumors larger than 3 mm in male APC^Δ14/+ mice. Comparison of the effects of Bi³⁺, In³⁺ and Ru³⁺ treatment on tumor numbers suggests that the inhibitory effect of Bi³⁺ ions is unlikely to be gastrin-dependent. However, further testing of higher doses of Bi³⁺ ions for longer periods as an oral treatment for intestinal tumors is warranted. The use of In³⁺ or Ru³⁺ ions for preventative reduction of intestinal tumor burden appears inadvisable based on the present findings, and their use in other tumor treatments should be approached with caution, especially if the patient is anemic.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

We thank Jennifer Hunt for preparation of the metal ion solutions and Dannel Yeo for invaluable assistance with the survival statistics. This work was supported by the National Health and Medical Research Council of Australia [grant number 1020983 to GSB].

References

1. A. Aly, A. Shulkes and G. S. Baldwin, Biochim. Biophys. Acta, 2004, 1704, 1–10.
2. G. J. Dockray, A. Varro, R. Dimaline and T. Wang, The gastrins: their production and biological activities, Annu. Rev. Physiol., 2001, 63, 119–139.
3. F. Hollande, A. Imdahl, T. Mantamadiotis, G. D. Ciccotosto, A. Shulkes and G. S. Baldwin, Gastroenterology, 1997, 113, 1576–1588.
4. 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.
5. 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.
6. A. Ferrand and T. C. Wang, Cancer Lett., 2006, 238, 15–29.
7. C. Do, C. Bertrand, J. Palasse, M. B. Delisle, E. Cohen-Jonathan-Moyal and C. Seva, BMC Cancer, 2013, 13, 531.
8. C. Do, C. Bertrand, J. Palasse, M. B. Delisle, A. Shulkes, E. Cohen-Jonathan-Moyal, A. Ferrand and C. Seva, Cancer Prev. Res., 2012, 5, 675–684.
9. A. C. Paterson, F. A. Macrae, C. Pizzey, G. S. Baldwin and A. Shulkes, J. Gastroenterol. Hepatol., 2014, 29, 480–486.
10. C. M. Thorburn, G. D. Friedman, C. J. Dickinson, J. H. Vogelman, N. Orentreich and J. Parsonnet, Gastroenterology, 1998, 115, 275–280.
11. D. A. Westwood, O. Patel, C. Christophi, A. Shulkes and G. S. Baldwin, Int. J. Colorectal Dis., 2017, 32, 1061–1064.
12. P. Singh, Cancer Lett., 2007, 252, 19–35.
13. A. Kowalski-Chauvel, S. Najib, I. G. Tikhonova, L. Huc, F. Lopez, L. O. Martinez, E. Cohen-Jonathan-Moyal, A. Ferrand and C. Seva, J. Biol. Chem., 2012, 287, 41458–41468.
14. L. N. Kwong and W. F. Dove, Adv. Exp. Med. Biol., 2009, 656, 85–106.
15. A. E. McCart, N. K. Vickaryous and A. Silver, Pathol., Res. Pract., 2008, 204, 479–490.
16. 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.
17. S. Colnot, M. Niwa-Kawakita, G. Hamard, C. Godard, S. Le Plenier, C. Houbron, B. Romagnolo, D. Berrebi, M. Giovannini and C. Perret, Lab. Invest., 2004, 84, 1619–1630.
18. H. Padmanabhan, M. J. Brookes and T. Iqbal, Nutr. Rev., 2015, 73, 308–317.
19. A. C. Chua, B. Klopcic, I. C. Lawrance, J. K. Olynyk and D. Trinder, World J. Gastroenterol., 2010, 16, 663–672.
20. S. Radulescu, M. J. Brookes, P. Salgueiro, R. A. Ridgway, E. McGhee, K. Anderson, S. J. Ford, D. H. Stones, T. H. Iqbal, C. Tselepis and O. J. Sansom, Cell Rep., 2012, 2, 270–282.
21. S. Kovac, G. J. Anderson, W. S. Alexander, A. Shulkes and G. S. Baldwin, Endocrinology, 2011, 152, 3062–3073.
22. M. Laval, G. S. Baldwin, A. Shulkes and K. M. Marshall, Increased gastrin gene expression provides a physiological advantage to mice under hypoxic conditions, Am. J. Physiol.: Gastrointest. Liver Physiol., 2015, 308, G76–84.
23. S. C. Longano, J. Knesel, G. J. Howlett and G. S. Baldwin, Interaction of gastrin with transferrin: effects of ferric ions, Arch. Biochem. Biophys., 1988, 263, 410–417.
24. G. S. Baldwin, C. C. Curtain and W. H. Sawyer, Biochemistry, 2001, 40, 10741–10746.
25. J. Pannequin, K. J. Barnham, F. Hollande, A. Shulkes, R. S. Norton and G. S. Baldwin, J. Biol. Chem., 2002, 277, 48602–48609.
26. G. S. Baldwin, G. N. George and M. J. Pushie, PLoS One, 2015, 10, e0140126.
27. J. Pannequin, J. P. Tantiongco, S. Kovac, A. Shulkes and G. S. Baldwin, J. Endocrinol., 2004, 181, 315–325.
28. 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.
29. S. Kovac, S. W. Loh, S. Lachal, A. Shulkes and G. S. Baldwin, Biochem. Pharmacol., 2012, 83, 524–530.
30. A. Ferrand, S. Lachal, G. Bramante, S. Kovac, A. Shulkes and G. S. Baldwin, Am. J. Physiol.: Gastrointest. Liver Physiol., 2010, 299, G220–227.
31. M. Laval, K. M. Marshall, J. Sachinidis, A. Scott, M. Eutick and G. S. Baldwin, J. Biol. Inorg. Chem., 2017, 22, 999–1006.
32. Y. Sano, H. Satoh, M. Chiba, M. Okamoto, K. Serizawa, H. Nakashima and K. Omae, J. Occup. Health, 2005, 47, 293–298.
33. F. P. Castronovo, Jr. and H. N. Wagner, Jr., J. Nucl. Med., 1973, 14, 677–682.
34. K. Asakura, H. Satoh, M. Chiba, M. Okamoto, K. Serizawa, M. Nakano and K. Omae, J. Occup. Health, 2008, 50, 471–479.
35. European Medicines Agency, Guidelines on the specification limits for residues of metal catalysts, 2007, http://www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2009/09/WC500003586.pdf.
36. N. Huynh, K. Wang, M. Yim, C. J. Dumesny, M. S. Sandrin, G. S. Baldwin, M. Nikfarjam and H. He, BMC Cancer, 2017, 17, 431.
37. K. H. Liu, N. Huynh, O. Patel, A. Shulkes, G. Baldwin and H. He, Cancer Lett., 2013, 340, 22–29.
38. M. E. Blazka, D. Dixon, E. Haskins and G. J. Rosenthal, Fundam. Appl. Toxicol., 1994, 22, 231–239.
39. A. J. Wagstaff, P. Benfield and J. P. Monk, Drugs, 1988, 36, 132–157.
40. S. L. Gorbach, Gastroenterology, 1990, 99, 863–875.
41. F. Kratz, M. Hartmann, B. Keppler and L. Messori, J. Biol. Chem., 1994, 269, 2581–2588.
42. D. R. Frasca, L. E. Gehrig and M. J. Clarke, J. Inorg. Biochem., 2001, 83, 139–149.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c7mt00272f

---