Abstract
Background: The water-soluble ionic gold(I) complex [Au(PPh2CH2CH2PPh2)2]Cl possesses at least ten times stronger antineoplastic activity than the water-soluble neutral silver(I) carboxylates [AgO2C(CH2OCH2)nH], 1, (n=1), 2 (n=2), or 3 (n=3) even though AuI and AgI are isoelectronic d10 metals lying one above the other in the periodic table of the elements. In this study we determined the cytotoxicity of the stable water-soluble gold(I) carboxylates [(Ph3P)AuO2C(CH2OCH2)nH], 4 (n=1), 5 (n=2) and 6 (n=3) to compare the intrinsic antineoplastic activity of gold(I) and silver(I) under conditions where different complex charges do not influence the result. Materials and Methods: The cytotoxicity of carboxylato gold complexes 4-6 towards the HeLa (human cervix epithelioid) cancer cell line ATCC CCL-2, resting lymphocytes and phytohaemagglutinin(PHA)-stimulated lymphocytes were determined. Cell survival was measured by means of the colorimetric 3-(4,5-dimethylthiazol-2-yl)-diphenyltetrazolium bromide (MTT) assay. Results: The IC50 values of 4-6 from six experiments causing 50% cell growth inhibition were essentially constant; values ranged between 0.50 and 0.82 μmol dm−3. Drug activity was thus independent of the carboxylato ligand chain length. Metal complexes 4-6 were at least one order of magnitude more cytotoxic towards the HeLa cancer cell line than the silver carboxylates 1-3 and were almost as cytotoxic as [Au(PPh2CH2CH2PPh2)2]Cl and cisplatin [(H3N)2PtCl2]. Complexes 4-6 were also 2-5 times more toxic against PHA-stimulated lymphocyte cultures than to HeLa cancer cell. Conclusion: Unlike for the silver complexes 1-3, no meaningful drug activity-structural relationship exists for the gold(I) d10 carboxylates 4-6. Gold(I) complexes in neutral and charged form are intrinsically more cytotoxic than silver(I) against the HeLa cancer cell line.
In the search for new anticancer drugs capable of overcoming the many side-effects associated with them (1, 2), some iron-containing beta-diketones (3) were found to be more active than cisplatin against the COR L23/CPR cancer cell line, a platinum-resistant variety of human large lung cells. Rhodium-cyclooctadiene complexes of these iron-containing beta-diketonates were found to be very active against three different human prostate cancer cell lines (4). A series of ferrocenyl (Fc)-containing alcohols, Fc(CH2)nOH [Fc=FeII(η5-C5H4)(η5-C5H5); n=1, 2, 3 or 4] had an interesting structural relationship in antineoplastic activity in that these compounds became more cytotoxic with increasing values of n (5). To overcome aqueous incompatibility of antineoplastic compounds, use was made of synthetic water-soluble polymeric drug carriers (6-8) and charged molecules (9) to induce water solubility in potentially promising antineoplastic drugs but which cannot be tested in the neat form due to water- and bioincompatibility. However, the water-soluble charged carboxylato complexes of ruthenium (9) were three to four orders of magnitude less active than cisplatin. Despite being ionic, poor solubility in any biocompatible medium prevented determination of the cytotoxicity of ferrocenylcarboxylato complexes of ruthenium, and the multiply charged ruthenium sodium phenyl-sulphonatocarboxylate complexes had low antineoplastic activity. Structurally, ruthenium carboxylates are bridged paddlewheel complexes (10, 11) in which four carboxylato ligands coordinate bi-dentate via both carboxylato oxygen atoms to two adjacent ruthenium ions. Another approach to enhancing water solubility is to build ethylene glycol fragments into the structure of complexes. This was done in the case of silver carboxylates 1-3 (Figure 1) (12). Silver(I) does not bind four carboxylato ligands in a bi-dentate fashion as ruthenium does. Single ligand monodentate coordination via only one oxygen atom may take place but in the solid state, silver carboxylates also show a tendency to oligomerize or polymerize via mono, bi- and bridging bonding motifs (13-15) (Figure 1). In solution, when coordinated to carboxylates bearing ethylene glycol fragments, these ligands can assume a circular chelation pattern via ethylene glycol oxygen atom coordination (14, 15) (Figure 1). Despite being well water-soluble, the silver(I) carboxylato compounds 1-3 were at least one order of magnitude less reactive towards HeLa cells than the ionic gold compound 7. As with ferrocene alcohols (5), a structural dependence of cytotoxicity was observed in that the longer-chain carboxylato silver complex 3 was less cytotoxic than the shorter chain compound 1.
Hydrophylic silver carboxylates 1-3, phosphine gold(I) carboxylato complexes 4-6, [Au(PPh2CH2CH2PPh2)2]Cl, 7, and cisplatin, 8.
To confirm the hypothesis (12) that 3 is less cytotoxic than 1 due to better encapsulation of the silver ion via semicircular chelation, we report here the results for cytotoxicity of the analogous gold carboxylato compound series 4-6 which cannot assume the open circular chelated structure that silver adopts in solution. Gold(I) complexes are essentially linear in solution and in the solid state (17-19). This study also demonstrates whether water-soluble ionic gold complexes are more reactive than neutral water-soluble ones or not, and if gold d10 complexes are inherently more cytotoxic than silver d10 systems bearing the same carboxylato ligands.
Materials and Methods
Compounds. Gold complexes 4-6 (Figure 1) were synthesised according to published procedures (20).
Sample preparation. Samples were dissolved in phosphate-buffered saline (PBS), giving stock concentrations of 20 mmol dm−3 and diluted in the appropriate growth medium supplemented with foetal calf serum (FCS) to give drug concentrations of 20-2,000 μmol dm−3 prior to the cell experiments.
Cell cultures. The human cervix epithelioid cancer cell line HeLa (ATCC CCL-2) (American Type Culture Collection, Manassas, VA, USA) was grown as monolayer cultures from Eagle's minimum essential medium (MEM). The growth media were maintained at 37°C under 5% CO2 and fortified with 10% FCS and 1% penicillin and streptomycin. Purified mononuclear leukocytes were prepared from whole blood collected from three healthy adult volunteers by density centrifugation on Histopaque-1077 (Sigma-Aldrich) cushions at 400 g for 25 min at ambient temperature. The lymphocyte band was washed and cells resuspended in RPMI-1640 supplemented with 10% FCS. Cells were seeded in 96-well microtiter plates (at 5000 cells/well in the case of cancer cells and 4×105 cells/well in the case of mononuclear leukocytes) in a final volume of 200 μl of growth medium in the presence or absence of different concentrations of experimental drugs. Appropriate solvent control systems were included. To some of the wells a mitogen (phytohaemagglutinin, PHA, Remel Europe Ltd., Dartford, Kent, UK) was added at a concentration of 2.5 μg/ml.
After incubation at 37°C for 7 days in the case of cancer cells and 3 days in the case of the mononuclear leukocytes, cell survival was measured by means of the colourimetric 3-(4,5-dimethylthiazol-2-yl)-diphenyltetrazodium bromide (MTT) assay (17). Plates were read on a spectrophotometer at 570 nm with a reference wavelength of 630 nm. Wells without cells and with cells but without drugs were included as controls. Survival curves were plotted (Figure 2) as a function of drug dose and the drug concentration that caused 50% inhibition of cell growth (IC50) was estimated by extrapolation. Tumour specificity was calculated as the ratio (IC50 of PHA-stimulated lymphocytes)/(IC50 of HeLa cancer cell line).
Results
The cell growth-inhibitory properties of complexes 1-8 expressed as IC50 values are summarized in Table I. The newly determined IC50 values of gold complexes 4-6 were essentially constant ranging between 0.50 and 0.82 μmol dm−3; previously determined (12) IC50 values for silver complexes 1-3 ranged between 2.6 and 6.0 μmol dm−3. A structure–activity relationship was observed in that silver complex 1 with the shortest carboxylate ligand was most reactive, and complex 3, with the longest chain carboxylate ligand, was the least reactive (Figure 3). Gold(I) complexes 4-6 exhibited no significant structure–activity relationship. Tumour specificity for gold complexes was between 2.4 (for 6) and 5.0 (for 5) and compares well with HeLa tumour specificity of 6.9 for cisplatin (3) and silver complexes 1-3. For gold(I) complexes 4-6, IC50 values for resting lymphocytes was 3.3±0.8 (4), 2.1±0.4 (5) and 1.7±0.2 μmol dm−3 (6).
Effect of concentration of 2, 4, and 8 (see Figure 1) on the survival of the indicated cell lines. Data are presented as the mean concentration±standard error of the mean of six HeLa cell experiments or three lymphocyte experiments.
Cytotoxicity of compounds 1-8 expressed as IC50 valuesa after 7 days of incubation with HeLa cancer cell lines or 3 days of incubation with phytohaemagglutinin(PHA)-stimulated human mononuclear leukocytes.
Discussion
In this study, hydrophylic ethylene glycol fragments as part of carboxylate ligands −O2C(CH2OCH2)nH with n=1, 2 or 3 were used to solubilize otherwise water-insoluble gold(I) complexes 4-6. This resulted in gold complexes [(Ph3P)AuO2C(CH2OCH2)nH] which are much more comparable to the analogous silver complexes 1-3, [AgO2C(CH2OCH2)nH] we reported on earlier (12) than the ionic species [Au(PPh2CH2CH2PPh2)2]Cl, 7; this is despite the gold(I) complexes being linear and monomeric in the solid state, while the silver complexes aggregate into oligomeric and polymeric species. The similarity in structure of silver and gold molecules 1-6 allows meaningful comparisons of the intrinsic relative cytotoxicity of the two d10 metallic systems (Au and Ag) lying one above the other in group 11 of the periodic table of the elements. Despite the structural and electronic similarities of 1-6, the gold(I) complexes 4-6 were about one order of magnitude more cytotoxic than the silver complexes 1-3. The ionic gold(I) complex 7 with an IC50 of 0.14 μmol dm−3 was only slightly more cytotoxic than 4-6. We conclude that irrespective of whether gold(I) complexes are ionic (i.e. charged) or neutral, they will always be much more reactive than the equivalent silver compounds. The reactivity of hydrophylic gold carboxylates 4-6 is also in the same order of magnitude as that of cisplatin (IC50=0.19 μmol dm−3 under identical conditions)
For the silver complexes 1-3 (Figure 1), a structural relationship was observed in that compounds become less cytotoxic with increasing value of n. For the gold complexes 4-6 (Figure 1), IC50 values were for all practical purposes independent of n. Compound numbers are noted next to each data point.
To understand why gold complexes of type [(Ph3P)AuO2C(CH2OCH2)nH] have no structure–activity dependence but silver complexes of the type [AgO2C(CH2OCH2)nH] do, one has to consider the structure in solution of these compounds (14, 15). While silver complexes 1-3 exist in coordinated polymer form in the solid state and have a tendency towards semicircular chelation in solution (14, 15) (Figure 1), gold carboxylates 4-6 can only exist as linear monomeric molecules, both in the solid state and in solution (16-19). This implies the gold centres of 4-6 are all equally unprotected and available for interaction with biological molecules despite the increase in ligand chain length. In contrast, the silver centres of 1-3 will become progressively more hindered in gaining access to the reactive site of any target biological molecule upon increasing ethylene glycol chain length. To help visualize this, the silver ion of 1, being coordinated by the shortest carboxylato ligand, coordinates only two oxygen atoms. Due to the short carboxylato chain length, one side of this molecule is left completely open for substrate interactions. However, the silver(I) ion of 3 is coordinated by four oxygen atoms and the carboxylato ligand is long enough to stretch completely around the central silver ion, thereby encapsulating it (14, 15). This will seriously impair the capability of 3 to interact with biologically relevant target molecules, such as DNA, and explains the lower cytotoxicity of 3. The observed decrease in antineoplastic activity of silver complexes [AgO2C(CH2OCH2)nH] with increasing value of n completely contrasts with the behaviour of iron complexes of the type [Fc(CH2)nOH] (Fc=FeII(η5-C5H5)(η5-C5H4); n=1-4) which exhibit an increase in antineoplastic activity with increasing value of n (5). The increase in cytotoxicity of [Fc(CH2)nOH] is related to the lowering of the oxidation potential of the iron centre (5) of these compounds with increasing n values.
Conclusion
The antineoplastic activity of hydrophylic gold(I) d10 carboxylates [(Ph3P)AuO2C(CH2OCH2)nH] (n=1, 2, 3) are approximately one order of magnitude higher than the equivalent hydrophobic silver(I) d10 carboxylato complexes [AgO2C(CH2OCH2)nH]. The cytotoxicity of ionic [Au(PPh2CH2CH2PPh2)2]Cl is only slightly higher than that of neutral [(Ph3P)AuO2C(CH2OCH2)nH] implying that the charge is not an important factor in determining the cytotoxicity of water-soluble gold(I) complexes. For the corresponding silver(I) complexes, a structure–reactivity relationship was observed in that compounds became less cytotoxic with increasing value of n. However, no such relationship exists for the gold complex series. The silver structure–reactivity relationship is ascribed to the propensity of [AgO2C(CH2OCH2)nH] to adopt a chelated structure with n+1 oxygen atoms coordinated to the central silver ion. The silver centres of complexes 1-3 therefore become more protected and less available for interaction with biological molecules with an increase in ligand chain length with larger n values. Gold complexes cannot undergo an equivalent geometrical transformation; they remain linear in solution. This implies the gold centres of complexes 4-6 are all equally poorly protected (or equally available) and reactive towards biological molecules despite the increase in ligand chain length.
Acknowledgements
The National Research Foundation of South Africa, the Central Research Fund of the University of the Free State, the Deutsche Forschungsgemeinschaft (DFG) and the Fonds der Chemischen Industrie (FCI) are acknowledged for their financial support.
- Received May 26, 2012.
- Revision received June 15, 2012.
- Accepted June 15, 2012.
- Copyright© 2012 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.