Protein metalation by metal-based drugs: X-ray crystallography and mass spectrometry studies.

The combined use of X-ray crystallography and mass spectrometry represents a valuable strategy to investigate and characterize protein metalation induced by anticancer metal-based drugs. Here, we summarize a series of significant results recently obtained in our laboratories upon the examination of the structures of several adducts of proteins with representative metallodrugs (mostly containing ruthenium, gold and platinum). The general mechanisms of protein metalation that emerge from a careful comparative analysis of these structures are illustrated and their mechanistic implications are discussed. Possible directions for future work in the field are delineated.

PtI2(DACH), the iodido analogue of oxaliplatin as a candidate for colorectal cancer treatment: chemical and biological features.

Colorectal cancer (CRC) is a global health problem being the fourth most common cause of death due to cancer worldwide. Oxaliplatin plays a key role in current CRC treatment but shows serious drawbacks, such as a high systemic toxicity and the frequent insurgence of Pt resistance. In search of novel and more efficacious Pt-based drugs for CRC treatment, we synthesized and characterised PtI2(DACH), an oxaliplatin analogue. PtI2(DACH) was obtained through the replacement of bidentate oxalate with two iodides. PtI2(DACH) turned out to be more lipophilic than oxaliplatin, a fact that led to an enhancement of its cellular uptake. In contrast to oxaliplatin, PtI2(DACH) showed a scarce reactivity towards model proteins, while maintaining affinity for a standard DNA oligo. Notably, PtI2(DACH) induced cytotoxicities roughly comparable to those of oxaliplatin in three representative CRC cell lines. Moreover, it was able to trigger cell apoptosis, to an extent even better than cisplatin and oxaliplatin. Overall, a rather promising picture emerges for this novel Pt drug that merits, in our opinion, a deeper and more extensive preclinical evaluation.

Water-soluble Ru(II)- and Ru(III)-halide-PTA complexes (PTA=1,3,5-triaza-7-phosphaadamantane): Chemical and biological properties.

Four structurally related Ru(II)-halide-PTA complexes, of general formula trans- or cis-[Ru(PTA)4X2] (PTA=1,3,5-triaza-7-phosphaadamantane, X=Cl (1, 2), Br (3, 4), were prepared and characterized. Whereas compounds 1 and 2 are known, the corresponding bromo derivatives 3 and 4 are new. The Ru(III)-PTA compound trans-[RuCl4(PTAH)2]Cl (5, PTAH=PTA protonated at one N atom), structurally similar to the well-known Ru(III) anticancer drug candidates (Na)trans-[RuCl4(ind)2] (NKP-1339, ind=indazole) and (Him)trans-[RuCl4(dmso-S)(im)] (NAMI-A, im=imidazole), was also prepared and similarly investigated. Notably, the presence of PTA confers to all complexes an appreciable solubility in aqueous solutions at physiological pH. The chemical behavior of compounds 1-5 in water and in physiological buffer, their interactions with two model proteins - cytochrome c and ribonuclease A - as well as with a single strand oligonucleotide (5'-CGCGCG-3'), and their in vitro cytotoxicity against a human colon cancer cell line (HCT-116) and a myeloid leukemia (FLG 29.1) were investigated. Upon dissolution in the buffer, sequential halide replacement by water molecules was observed for complexes 1-4, with relatively slow kinetics, whereas the Ru(III) complex 5 is more inert. All tested compounds manifested moderate antiproliferative properties, the cis compounds 2 and 4 being slightly more active than the trans ones (1 and 3). Mass spectrometry experiments evidenced that all complexes exhibit a far higher reactivity towards the reference oligonucleotide than towards model proteins. The chemical and biological profiles of compounds 1-5 are compared to those of established ruthenium drug candidates in clinical development.

Cytotoxic activity and protein binding through an unusual oxidative mechanism by an iridium(I)-NHC complex.

A new NHC iridium(I) complex (1) showing significant antiproliferative properties in vitro is described here. Its crystal structure, solution behaviour and interactions with the model proteins cytochrome c (cyt c) and lysozyme were investigated. High resolution ESI-MS measurements suggest that this iridium(i) complex acts as a prodrug and binds cyt c tightly through an unusual "oxidative" mechanism. Eventually, an iridium(III)-NHC fragment is found associated to the protein.

Interactions of selected gold(III) complexes with DNA G quadruplexes.

The interactions of three representative gold(III) complexes with human telomeric DNA sequences were analysed using a variety of biophysical methods, including DNA melting, circular dichroism, SPR and ESI MS; remarkable interactions were highlighted for all tested complexes, although they were associated to significantly different binding profiles. The most interesting compound was Auoxo6, a dinuclear gold(III) complex, which beyond manifesting a conspicuous binding affinity for the G-quadruplex conformation, turned out to be very effective in inducing a non-canonical secondary structure. These findings may clear the way for novel biological and pharmacological applications of this class of metal compounds.

X-ray absorption spectroscopy studies of the adducts formed between cytotoxic gold compounds and two major serum proteins.

Gold metallodrugs form a class of promising antiproliferative agents showing a high propensity to react with proteins. We exploit here X-ray absorption spectroscopy (XAS) methods [both X-ray absorption near-edge spectroscopy (XANES) and extended X-ray absorption fine structure (EXAFS)] to gain insight into the nature of the adducts formed between three representative gold(I, III) metallodrugs (i.e., auranofin, [Au(2,2'-bipyridine)(OH)(2)](PF(6)), Aubipy, and dinuclear [Au(2)(6,6'-dimethyl-2,2'-bipyridine)(2)(µ-O)(2)](PF(6))(2), Auoxo6) and two major plasma proteins, namely, bovine serum albumin (BSA) and human serum apotransferrin (apoTf). The following metallodrug-protein systems were investigated in depth: auranofin/apoTf, Aubipy/BSA, and Auoxo6/apoTf. XANES spectra revealed that auranofin, upon protein binding, conserves its gold(I) oxidation state. Protein binding most probably takes place through release of the thiosugar ligand and its subsequent replacement by a thiol (or a thioether) from the protein. This hypothesis is independently supported by EXAFS results. In contrast, the reactions of Aubipy with serum albumin and of Auoxo6 with serum apoTf invariantly result in gold(III) to gold(I) reduction. Gold(III) reduction, clearly documented by XANES, is accompanied, in both cases, by release of the bipyridyl ligands; for Auoxo6 cleavage of the gold-gold dioxo bridge is also observed. Gold(III) reduction leads to formation of protein-bound gold(I) species, with deeply modified metal coordination environments, as evidenced by EXAFS. In these adducts, the gold(I) centers are probably anchored to the protein through nitrogen donors. In general, these two XAS methods, i.e., XANES and EXAFS, used here jointly, allowed us to gain independent structural information on metallodrug/protein systems; detailed insight into the gold oxidation state and the local environment of protein-bound metal atoms was achieved in the various cases.

Copper and zinc dismetabolism in the mouse brain upon chronic cuprizone treatment.

Recent reports describe successful treatment using copper chelation therapy in neurodegenerative animal models. However, the success claimed for chelation therapy in neurodegenerative diseases is still rather controversial. To acquire new information on copper metabolism/homeostasis, we utilized cuprizone, a very sensitive and selective copper-chelating agent with well-known neurotoxic properties, as a relevant chemical model in mice. Upon cuprizone treatment, mice developed a pronounced astrocytosis, with brain oedema and spongiosis characterised by vacuolisations of the neuropil predominantly in the white matter. In addition, cuprizone treatment severely altered copper and zinc homeostasis in the central nervous system (CNS) as well as in all other tissues examined, with increasing metal ion concentrations particularly in the CNS. Concomitant with this increase in the Cu and Zn concentration in the brain, metallothionein-I and -II were also highly immunoreactive in astrocyte, consistent with the astrocytosis and demyelination observed in our and other laboratories.

Structural effects of titanium citrate on the human erythrocyte membrane.

The structural effects of titanium citrate on the human erythrocyte membrane were studied through its interaction with intact erythrocytes and isolated unsealed human erythrocyte membranes (IUM). The studies were carried out by scanning electron microscopy and fluorescence spectroscopy, respectively. Titanium citrate induced shape changes in erythrocytes, which were damaged and ruptured leaving empty and retracted membranes. Fluorescence spectroscopy measurements in IUM indicated a disordering effect at both the polar head group and the acyl chain packing arrangements of the membrane phospholipid bilayer. Titanium citrate also interacted with molecular models of the erythrocyte membrane consisting in bilayers of dimyristoylphosphatidylcholine (DMPC) and dimyristoylphosphatidylethanolamine (DMPE), representing classes of phospholipids located in the outer and inner monolayers of the erythrocyte membrane, respectively. X-ray diffraction indicated that titanium citrate induced structural perturbation of the polar head group and of the hydrophobic acyl regions of DMPC, while the effects on DMPE bilayers were negligible. This conclusion is supported by fluorescence spectroscopy measurements on DMPC large unilamellar vesicles. All these findings indicate that the structural perturbations induced by titanium to human erythrocytes can be extended to other cells, thereby affecting their functions.

A comparative study of adduct formation between the anticancer ruthenium(III) compound HInd trans-[RuCl4(Ind)2] and serum proteins.

Formation of adducts between the antitumor ruthenium(III) complex [HInd]trans-[RuCl(4)(Ind)(2)] (KP1019) and the plasma proteins serum albumin and serum transferrin was investigated by UV-vis spectroscopy, for metal-to-protein ratios ranging from 1:1 to 5:1. In both cases, formation of tight metal-protein conjugates was observed. Similar spectroscopic features were observed for both albumin and transferrin derivatives implying a similar binding mode of the ruthenium species to these proteins. Surface histidines are the probable anchoring sites for the bound ruthenium(III) ions in line with previous crystallographic results. In order to assess the stability of the KP1019-protein adducts the influence of pH, reducing agents and chelators was analysed by UV-vis spectroscopy. Notably, there was no effect of addition of EDTA on the UV-vis spectra of the conjugates. The pH-stability was high in the pH range 5-8. Experiments with sodium ascorbate showed that there was just some alteration of selected bands. The implications of the present results are discussed in relation to the pharmacological behavior of this novel class of antitumor compounds.

Investigation of the effects of copper ions on protein aggregation using a model system.

Protein aggregation is a notable feature of various human disorders, including Parkinson's disease, Alzheimer's disease and many others systemic amyloidoses. An increasing number of observations in vitro suggest that transition metals are able to accelerate the aggregation process of several proteins found in pathological deposits, e.g. alpha-synuclein, amyloid beta (Abeta) peptide, beta(2)-microglobulin and fragments of the prion protein. Here we report the effects of metal ions on the aggregation rate of human muscle acylphosphatase, a suitable model system for aggregation studies in vitro. Among the different species tested, Cu(2+) produced the most remarkable acceleration of aggregation, the rate of the process being 2.5-fold higher in the presence of 0.1 mM metal concentration. Data reported in the literature suggest the possible role played by histidine residues or negatively charged clusters present in the amino acid sequence in Cu(2+)-mediated aggregation of pathological proteins. Acylphosphatase does not contain histidine residues and is a basic protein. A number of histidine-containing mutational variants of acylphosphatase were produced to evaluate the importance of histidine in the aggregation process. The Cu(2+)-induced acceleration of aggregation was not significantly altered in the protein variants. The different aggregation rates shown by each variant were entirely explained by the changes of hydrophobicity or propensity to form a beta structure introduced by the point mutation. The effect of Cu(2+) on acylphosphatase aggregation cannot therefore be attributed to the specific factors usually invoked in the aggregation of pathological proteins. The effect, rather, seems to be a general related to the chemistry of the polypeptide backbone and could represent an additional deleterious factor resulting from the alteration of the homeostasis of metal ions in cells.

Biological role of adduct formation of the ruthenium(III) complex NAMI-A with serum albumin and serum transferrin.

NAMI-A is an innovative ruthenium(III) complex with a very encouraging preclinical profile of metastasis inhibition, which is undergoing initial phases of clinical trials. To assess the pharmacological relevance of the drug fraction associated to plasma proteins, adducts of NAMI-A with either serum albumin or serum transferrin were prepared and their biological effects tested in vitro and in vivo. Specifically, adducts of NAMI-A with either serum albumin or serum transferrin, prepared and characterized at a ruthenium-to-protein molar ratio of 4:1, were evaluated in vitro on the KB human tumor cell line and in vivo on the MCa mammary carcinoma tumor. The effects of NAMI-A/protein adducts on cell viability and on cell cycle progression were found to be far smaller than those produced by free NAMI-A. GFAAS measurements point out that the amount of ruthenium that gets into cells is drastically reduced when NAMI-A is presented in its protein-bound form. In vivo use of NAMI-A adducts with albumin and transferrin resulted markedly less effective on lung metastasis reduction, than free NAMI-A. Overall, the present results suggest that binding to plasma proteins causes a drastic decrease of NAMI-A bioavailability and a subsequent reduction of its biological activity, implying that association to plasma proteins essentially represents a mechanism of drug inactivation.

Antiangiogenic properties of selected ruthenium(III) complexes that are nitric oxide scavengers.

The nitric oxide synthase (NOS) pathway has been clearly demonstrated to regulate angiogenesis. Increased levels of NO correlate with tumour growth and spreading in different experimental and human cancers. Drugs interfering with the NOS pathway may be useful in angiogenesis-dependent tumours. The aim of this study was to pharmacologically characterise certain ruthenium-based compounds, namely NAMI-A, KP1339, and RuEDTA, as potential NO scavengers to be used as antiangiogenic/antitumour agents. NAMI-A, KP1339 and RuEDTA were able to bind tightly and inactivate free NO in solution. Formation of ruthenium-NO adducts was documented by electronic absorption, FT-IR spectroscopy and (1)H-NMR. Pretreatment of rabbit aorta rings with NAMI-A, KP1339 or RuEDTA reduced endothelium-dependent vasorelaxation elicited by acetylcholine. This effect was reversed by 8-Br-cGMP. The key steps of angiogenesis, endothelial cell proliferation and migration stimulated by vascular endothelial growth factor (VEGF) or NO donor drugs, were blocked by NAMI-A, KP1339 and RuEDTA, these compounds being devoid of any cytotoxic activity. When tested in vivo, NAMI-A inhibited angiogenesis induced by VEGF. It is likely that the antitumour properties previously observed for ruthenium-based NO scavengers, such as NAMI-A, are related to their NO-related antiangiogenic properties.

Interactions of two cytotoxic organotin(IV) compounds with calf thymus DNA.

The reactions with DNA of two antitumor active organotin(IV) compounds, the dimer of bis[(di-n-butyl 3,6-dioxaheptanoato)tin] (C(52)H(108)Sn(4)O(1) x 2H(2)O), compound 1, and tri-n-butyltin 3,6,9-trioxodecanoate (C(19)H(40)SnO(5) x 1/2H(2)O), compound 2, were analysed by circular dichroism, DNA melting experiments and gel mobility shift assays. It is found that both complexes modify only slightly the B-type circular dichroism spectroscopy (CD) spectrum of calf thymus DNA. On the other hand, both complexes were found to affect significantly the parameters of the thermally induced helix-to-coil transition. Addition of 1 or 2 to calf thymus DNA samples does not favor DNA renaturation after melting ruling out formation of interstrand crosslinks. Moreover, the effects of both compounds on plasmid DNA gel mobility were investigated. From the analysis of the present results it is inferred that both organotin(IV) compounds do interact with DNA, probably at the level of the phosphate groups.

Solution chemistry and DNA binding properties of MEN 10755, a novel disaccharide analogue of doxorubicin.

The behavior under physiological conditions of MEN 10755, a novel disaccharide analogue of doxorubicin, was investigated in detail by a variety of spectroscopic techniques including spectrophotometry, fluorescence, and (1)H NMR. The pH dependent properties of MEN 10755 were also analysed by spectrophotometry and potentiometry within the pH range 5--11. It is found that MEN 10755 behaves very similarly to doxorubicin and reproduces closely its pH dependent pattern. Like doxorubicin, MEN 10755 undergoes dimerization with a significantly smaller association constant. The interaction of MEN 10755 with calf thymus DNA was studied in detail. Spectrophotometric and fluorescence titrations of MEN 10755 with calf thymus DNA show spectral patterns almost identical to those obtained with doxorubicin implying that the binding mechanism and the stability of the resulting adducts are very similar. An apparent affinity constant of 1.2 x 10(6) was determined for the interaction of MEN 10755 with calf thymus DNA to be compared with the value of 3.3 x 10(6) measured for doxorubicin, under the same conditions. The effects of both anthracyclines on the thermal denaturation profiles of calf thymus DNA were also analyzed; both compounds turned out to stabilize to a similar extent the DNA double helix and to give rise to a characteristic two-step melting profile. The implications of the present results for the pharmacological activity and the mechanism of action of this novel and promising antitumor compound are discussed.

Interactions of selected gold(III) complexes with calf thymus DNA.

DNA represents the primary target for platinum antitumor metal complexes and is the probable target for newly developed cytotoxic gold(III) complexes. To test this hypothesis the reactions with calf thymus DNA of five representative gold(III) complexes--namely [Au(en)(2)]Cl(3), [Au(dien)Cl]Cl(2), [Au(cyclam)](ClO(4))(2)Cl, [Au(terpy)Cl]Cl(2) and [Au(phen)Cl(2)]Cl--were analyzed in vitro through various physicochemical techniques including circular dichroism, absorption spectroscopy, DNA melting, and ultradialysis. It is shown that all tested complexes interact with DNA and modify significantly its solution behavior. The solution conformation of DNA is affected to variable extents by the individual complexes as shown by CD titration experiments. Notably, in all cases, the gold(III) chromophore is not largely perturbed by addition of calf thymus DNA ruling out occurrence of gold(III) reduction. Ultradialysis experiments point out that the binding affinity of the various complexes for the DNA double helix is relatively low; in most cases the gold(III)/DNA interaction is electrostatic in nature and reversible. The implications of these findings for the mechanism of action of antitumor gold(III) complexes are discussed.

Gold(III) complexes as potential antitumor agents: solution chemistry and cytotoxic properties of some selected gold(III) compounds.

Gold(III) complexes generally exhibit interesting cytotoxic and antitumor properties, but until now, their development has been heavily hampered by their poor stability under physiological conditions. To enhance the stability of the gold(III) center, we prepared a number of gold(III) complexes with multidentate ligands - namely [Au(en)(2)]Cl(3), [Au(dien)Cl]Cl(2), [Au(cyclam)](ClO(4))(2)Cl, [Au(terpy)Cl]Cl(2), and [Au(phen)Cl(2)]Cl - and analyzed their behavior in solution. The solution properties of these complexes were monitored by visible absorption spectroscopy, mass spectrometry, and chloride-selective potentiometric measurements; the electrochemical properties were also studied by cyclic voltammetry and coulometry. Since all the investigated compounds exhibited sufficient stability under physiological conditions, their cytotoxic properties were tested in vitro, via the sulforhodamine B assay, on the representative human ovarian tumor cell line A2780, either sensitive or resistant to cisplatin. In most cases the investigated compounds showed relevant cell-killing properties with IC(50) values falling in the 0.2-10 microM range; noticeably most investigated gold(III) complexes were able to overcome, to a large extent, resistance to cisplatin when tested on the corresponding cisplatin-resistant cell line. The cytotoxic properties of the free ligands were also determined under the same solution conditions. Ethylenediamine, diethylenetriamine, and cyclam were virtually nontoxic (IC(50) values > 100 microM) so that the relevant cytotoxic effects observed for [Au(en)(2)]Cl(3) and [Au(dien)Cl]Cl(2) could be quite unambiguously ascribed to the presence of the gold(III) center. In contrast the phenanthroline and terpyridine ligands turned out to be even more cytotoxic than the corresponding gold(III) complexes rendering the interpretation of the cytotoxicity profiles of the latter complexes less straightforward. The implications of the present findings for the development of novel gold(III) complexes as possible cytotoxic and antitumor drugs are discussed.

DNA as a possible target for antitumor ruthenium(III) complexes.

The interaction of two experimental ruthenium(III)-containing antitumor complexes-Na[trans-RuCl(4)(DMSO)(Im)] (NAMI) and dichloro(1,2-propylendiaminetetraacetate)ruthenium(III) (RAP)-with DNA was investigated through a number of spectroscopic and molecular biology techniques, including spectrophotometry, circular dichroism, gel shift analysis, and restriction enzyme inhibition. It was found that both complexes slightly alter DNA conformation, modify its electrophoretic mobility, and inhibit DNA recognition and cleavage by some restriction enzymes, though they were less effective than cisplatin in producing such effects. Notably, the effects produced by NAMI on DNA were much larger than those induced by RAP. Implications of these results for the mechanism of action of ruthenium(III) antitumor complexes are discussed.

Cytotoxicity and DNA binding properties of a chloro glycylhistidinate gold(III) complex (GHAu).

The chloro glycylhistidinate gold(III) complex (GHAu) is shown to be fairly cytotoxic towards the established A2780 ovarian carcinoma human cell line either sensitive or resistant to cisplatin. Remarkably, GHAu is far more cytotoxic than the corresponding zinc(II), palladium(II), platinum(II) and cobalt(II) complexes implying that cytotoxicity is essentially to be ascribed to the presence of a gold(III) center. Circular dichroism (CD) spectra, atomic absorption measurements and DNA melting profiles suggest that GHAu in vitro is able to bind DNA, the presumed target for several antitumor metal complexes, and to modify its conformation, even if the observed changes are generally small. Implications of these findings for the mechanism of action of cytotoxic gold(III) complexes are discussed.

A spectroscopic study of the reaction of NAMI, a novel ruthenium(III)anti-neoplastic complex, with bovine serum albumin.

The reaction of Na[transRuCl4Me2SO(Im)] (NAMI; where Im is imidazole), a novel anti-neoplastic ruthenium(III) complex, with BSA, was studied in detail by various physico-chemical techniques. It is shown that NAMI, following chloride hydrolysis, binds bovine serum albumin tightly; spectrophotometric and atomic absorption data point out that up to five ruthenium ions are bound per albumin molecule when BSA is incubated for 24 h with an eightfold excess of NAMI. CD and electronic absorption results show that the various ruthenium centers bound to albumin exhibit well distinct spectroscopic features. The first ruthenium equivalent produces a characteristic positive CD band at 415 nm whereas the following NAMI equivalents produce less specific and less marked spectral effects. At high NAMI/BSA molar ratios a broad negative CD band develops at 590 nm. Evidence is provided that the bound ruthenium centers remain in the oxidation state +3. By analogy with the case of transferrins it is proposed that the BSA-bound ruthenium ions are ligated to surface histidines of the protein; results from chemical modification experiments with diethylpyrocarbonate seem to favor this view. Spectral patterns similar to those shown by NAMI are observed when BSA is reacted with two strictly related ruthenium(III) complexes Na[transRuCl4(Me2SO)2] and H(Im)[transRuCl4(Im)2] (ICR), implying a similar mechanism of interaction in all cases. It is suggested that the described NAMI-BSA adducts may form in vivo and may be relevant for the biological properties of this complex; alternatively NAMI/BSA adducts may be tested as specific carriers of the ruthenium complex to cancer cells. Implications of these findings for the mechanism of action of NAMI and of related ruthenium(III) complexes are discussed.

Cytotoxicity, DNA damage, and cell cycle perturbations induced by two representative gold(III) complexes in human leukemic cells with different cisplatin sensitivity.

The gold(III) complexes [Au(phen)Cl2]Cl and [Au(dien)Cl]Cl2 were recently shown to exert important cytotoxic effects in vitro on human tumor cell lines. To elucidate the biochemical mechanisms leading to cell death, the effects produced by these gold(III) complexes on the leukemic CCRF-CEM cell line--either sensitive (CCRF-CEM) or resistant to cisplatin (CCRF-CEM/CDDP)--were analyzed in detail by various techniques. For comparison purposes the effects produced by equitoxic concentrations of cisplatin were also analyzed. First, the dependence of the IC50 values of either complex on the incubation time was investigated. Cytotoxicity experiments confirmed that both gold(III) compounds retain their efficacy against the cisplatin-resistant line: only minimal cross-resistance with cisplatin was detected. Notably, [Au(phen)Cl2]Cl is more cytotoxic than [Au(dien)Cl]Cl2, with IC50 values of 7.4 and 6.0 M at 24 and 72 h, respectively, on the resistant line. Results of the COMET assay point out that both gold(III) complexes directly damage nuclear DNA. Remarkably, DNA damage inferred by either gold(III) complex in the two cell lines is larger than that produced by equitoxic cisplatin concentrations. Finally, the effects that either gold(III) complex produces on the cell cycle were investigated by flow cytometry. It was found that both complexes cause only moderate and transient cell cycle perturbations. Larger cell cycle perturbations are induced by equitoxic concentrations of cisplatin. The implications of the present results for the mechanism of action of cytotoxic gold(III) complexes are discussed.