Platinum amine coordination compounds as anti-tumor drugs. Molecular aspects of the mechanism of action

Effects of temperature on the interaction of cisplatin and carboplatin with cellular DNA

Increased levels of cisplatin (cDDP)- and carboplatin (CBDCA)-DNA adducts were detected in cDDP (10 microM)- and CBDCA (6 mM)-treated CC531 cells when the temperature was raised from 37 degrees to 43 degrees. In the case of cDDP, increased DNA adduct formation was already detectable at 38.5 degrees; additional temperature steps led to further increases in DNA modification. Increased CBDCA-DNA adduct formation was observed only at temperatures higher than 40 degrees. In vitro studies on the interaction of CDDP and CBDCA with isolated salmon sperm DNA, however, demonstrated no significant differences in the DNA binding rate between 37 degrees and 43 degrees for cDDP and a minor effect for CBDCA only at 43 degrees, almost totally excluding a direct temperature effect on DNA platination in this temperature range. Furthermore, neither the stability of the formed platinum-DNA adducts nor the rate of adduct loss in CC531 cells was changed at higher temperatures. The observed difference in cellular adduct formation, however, could be related to increased uptake of cDDP and CBDCA into CC531 cells at higher temperatures. In the case of cDDP, a temperature shift from 37 degrees to 38.5 degrees resulted in a significantly higher intracellular platinum concentration (0.03 +/- 0.01 vs 0.071 +/- 0.021 micrograms platinum/10(6) cells, respectively); for CBDCA, temperatures > or = 41.5 degrees were needed to increase the platinum concentration significantly above 37 degree values (0.3 +/- 0.1 vs 0.6 +/- 0.1 micrograms platinum/10(6) cells, respectively). In addition, the increase in DNA adduct formation of cDDP and CBDCA at elevated temperatures was comparable with the increase in cDDP-DNA adducts after a cDDP concentration escalation at 37 degrees, indicating a concentration-dependent increase in cDDP-DNA adducts. It seems that heat affects primarily the cellular uptake of cDDP and CBDCA and not their covalent binding to DNA.

Ligand effects on platinum binding to DNA. A comparison of DNA binding properties for cis- and trans-[PtCl2(amine)2] (amine = NH3, pyridine)

The DNA binding properties of cis- and trans-[PtCl2(pyridine)2] have been examined and compared with their NH3 analogs, cis- and trans-DDP. The presence of a planar ligand reduces the rates of DNA binding but does not greatly affect the overall conformation of CT DNA, as measured by circular dichroism spectroscopy. The sequence specificity of trans-[PtCl2(py)2] includes alternating purine-pyrimidine sequences. The sequence specificity is further different between the two pyridine isomers, and the steric effects of two cis-pyridine groups are demonstrated by the appearance of relatively few binding sites in the 49-bp duplex. The effects of the pyridine ligand are further manifested by a greatly enhanced DNA-DNA interstrand cross-linking efficiency for the trans isomer, with a cross-link per adduct frequency of between 0.14 and 0.23, depending on the rb of the sample. The unwinding of closed circular pUC19 DNA by trans-[PtCl2(pyridine)2] is also more efficient than that by either DDP isomer, with an unwinding angle calculated at phi = 17 degrees (compare cis-DDP with phi = 13 degrees and trans-DDP with phi = 9-10 degrees). In contrast, little unwinding is induced by cis-[PtCl2(pyridine)2], with phi = 4 degrees. These results in particular invert the standard cis/trans structure-activity relationships observed previously for [PtCl2(NH3)2]. The results are discussed with respect to the previously demonstrated effect of activation of the trans-platinum geometry using sterically hindered ligands.

Nonclassical platinum antitumor agents: perspectives for design and development of new drugs complementary to cisplatin

Studies over the last few years have shown that the range of platinum complexes with useful cytotoxicity and antitumor activity is not strictly limited to structural analogs of cisplatin. In general, we can expect that cells will process structurally different species in different manners. The metabolic chemistry and DNA binding will be altered in in comparison to the cis-[PtX2 (amine)2] class. This point is of particular importance because any altered pattern of antitumor activity of structural analogs of cisplatin is likely to be due to unpredictable pharmacokinetic, rather than truly mechanistic, factors. The fact that discrete cisplatin-DNA adducts vary in their biological activity further supports the hypothesis that complexes structurally dissimilar to cisplatin may produce biological activity complementary to the parent drugs. The mechanism of action of nonclassical complexes is different from that of cisplatin and its analogs. Their pattern of antitumor activity is also altered with respect to cisplatin--thus, not all platinum-containing drugs need necessarily be similar in their clinical profile to cisplatin. Note that both the dinuclear bis(platinum) complexes and the trans complexes give their own distinct patterns of tumor specificity--different from cisplatin and each other (see Tables 1 and 3). New cytotoxic mechanisms for platinum complexes may also be placed in context with cisplatin resistance. Modes of overcoming cisplatin resistance may reside at the various levels of uptake, interaction with "detoxifying" intracellular thiols, and DNA repair. Likewise complexes with novel mechanisms of action may circumvent resistance by more than one unique route. Indeed, the three major routes to resistance are all affected to varying degrees by the complexes outlined above. From the discovery of cisplatin, the development of analogs has essentially been an empirical exercise. Because of their similar mechanism of action, much comparison has been made between platinum complexes and the classic alkylating agents. Yet the alkylating agents represent a good example where a number of structurally distinct drugs with different anticancer activities are clinically available. This desirable feature may be achieved for platinum complexes by emphasis on complexes structurally dissimilar to the presently used agents. The dinuclear bis(platinum) complexes and mononuclear complexes in the trans geometry are of special interest. Comparison of common features and differences between different classes may point to guidelines for the rational design of complexes with a different spectrum of clinical antitumor activity to cisplatin and activity against cisplatin-resistant tumors.

Antibodies against cisplatin-modified DNA and cisplatin-modified (di)nucleotides

Cytotoxic effects of cis-diamminedichloroplatinum-(II) (cis-DDP) are thought to be mediated by binding to DNA. Studies on binding of cis-DDP to cellular DNA rely heavily on the availability of specific antibodies. We therefore raised and characterized four rabbit antisera: one against cis-DDP-modified DNA (antiserum NKI-A59) and three others against the cis-DDP-modified (di)nucleotides cis-Pt(NH3)2d(pApG) (NKI-A68), cis-Pt(NH3)2d(GMP)2 (NKI-A10), and Pt(NH3)3dGMP (NKI-A39). Reactivities to platinum compounds were determined in an enzyme-linked immunosorbent assay (ELISA) and in a quantitative immunocytochemical assay. In the ELISA, NKI-A59 showed a high affinity for DNA heavily substituted with either cis-DDP or CBDCA [cis-diammine(1,1-cyclobutanedicarboxylato)platinum(II)]; amounts of platinum per well giving 50% inhibition (IA50) were as low as 15 and 76 fmol, respectively. NKI-A59 also showed affinity to cis-DDP-modified poly[d(G-C)].poly[d(G-C)], poly(dC), and poly(dG). No affinity was found for trans-DDP [trans-diamminedichloro-platinum(II)]-modified DNA, enzymatically digested cis-DDP-DNA, or cis-DDP-DNA, or cis-DDP-modified poly(dA).poly(dT), oligo(dA)15.oligo(dT)15, oligo(dG)21, oligo(dG)42, or oligo(dAAAG)10. The efficiency of binding to cis-DDP-DNA decreased with decreasing DNA modification levels. Although other cis-DDP-DNA- and cis-DDP-(di)nucleotide-specific antisera have been identified, NKI-A59 is the first antiserum described that is suitable for the in situ detection of cis-DDP-DNA adducts at clinically relevant platinum levels. Adduct-specific immunostaining signals in cultured RIF-1 cells or rat liver paralleled platinum-DNA binding as measured by atomic absorption spectroscopy. The antisera NKI-A68, NKI-A10, and NKI-A39 showed high affinity for their corresponding haptens and varying affinity for non-hapten cis-DDP-DNA adducts. Their affinity for digested cis-DDP-modified DNA was up to 30 times that for intact cis-DDP-DNA. Neither NKI-A68 nor NKI-A10 resulted in specific immunocytochemical staining of cis-DDP-DNA adducts. We conclude that NKI-A68, NKI-A10, and NKI-A39 are suitable for platinum-DNA adduct analysis of digested DNA in ELISA and that NKI-A59 is suitable for platinum-DNA adduct detection at the single-cell level using immunocytochemical methods.

Characterization of cisplatin adducts of oligonucleotides by fast atom bombardment mass spectrometry

The products of the reaction of the antitumor drug cisplatin (cis-diamminedichloroplatinum(II)) with four oligonucleotide tetramers, d(GpCpGpC), d(GpGpCpC), d(TpGpApT), and d(TpGpCpT), were separated by gel permeation chromatography and characterized by negative- and positive-ion fast atom bombardment (FAB) mass spectrometry. Fragment ions indicating the oligonucleotide sequence and the position of cisplatin binding were observed in MS/MS spectra following collisional activation and B/E-linked scanning. Positive-ion FAB MS/MS spectra were characterized by platinum-containing product ions. Nonplatinated sequence ions and internal fragment ions were present primarily in the negative-ion spectra. The most prominent fragment ions containing platinum were [HB2.Pt.B3H]+ and [HB1.Pt.B2H]+, where B1, B2, and B3 were bases in the oligonucleotide tetramer, one of which was usually guanine. Both singly and doubly charged platinum complexes were observed, probably indicating reduction of Pt(II) during the FAB ionization process. The location of the platinum complex bound to each oligonucleotide sequence could be determined, and the binding sites observed by mass spectrometry were similar to those previously determined by other methods. FAB ionization with collisional activation and MS/MS analysis could serve as a new method for structural analysis of platinated oligonucleotides.

Regeneration experiments of the platinated enzyme fumarase, using sodium diethyldithiocarbamate, thiourea, and sodium thiosulfate

The enzyme fumarase is inhibited by [cis-Pt(NH3)2(H2O)2] (NO3)2. The Pt compound most likely binds at a S-methionine site. Sodium diethyldithiocarbamate (Naddtc) appears to be a powerful regenerator of enzymatic activity. Thiourea is less active, while sodium thiosulfate (STS) is almost inactive in restoring the activity of the enzyme. The regeneration phenomena are based on the dissociation of the Pt-S bonds of the methionine type, and formation of species like [Pt(ddtc)2]. In the model adduct [Pt(dien)GS-Me]2+ Naddtc, thiourea and STS easily break the Pt-S bond of the methionine type. It is concluded that the model system for Naddtc and thiourea does resemble fumarase quite well. S-donor ligands, which may be used as rescue agents in Pt antitumor therapy, are known to suppress nephrotoxicity caused by [cis-PtCl2(NH3)2]. A parallel is drawn between the enzyme reactivation, modeled by fumarase, and the [cis-PtCl2(NH3)2] nephrotoxicity suppression by rescue agents. It is proposed that a Pt-methionine type binding is broken by the rescue agents Naddtc and thiourea, but that the rescue agent STS only inhibits the nephrotoxicity by inactivating unbound Pt species in the cell.

Biophysical studies of the modification of poly(rG) . poly(rC) by cisplatin. Relations to the biological activity of the complex

The integrity of the double-stranded complex polyriboguanylic.polyribocytidylic acid [poly(rG).poly(rC)] modified by antitumour cis-diamminedichloroplatinum(II)(cis-DDP) was studied with the aid of differential pulse polarography and terbium fluorescence measurement. The modification was made to level corresponding to rb = 0.05 (rb is defined as the number of platinum atoms covalently bound per one nucleotide residue). Two modes of the modification of the polynucleotide complex were employed: The action of cis-DDP on poly(G) before formation of the complex with poly(C) and on the complex already formed from non-modified polynucleotides. It was shown that in the latter case modification disordered the integrity of the complex only negligibly. while in the former case the modification resulted in a noticeably more extensive disturbance of the double-stranded polynucleotide complex. Moreover, the modification of the complex (after its formation) at rb = 0.02 led to improved interferon-inducing and antiviral activity of poly(rG).poly(rC) tested on mice infected by influenza virus. It was suggested that the combined effects of interferon-inducing and antiviral activities of poly(rG).poly(rC) and antiviral activity of cis-DDP may result in an increased effect over and above what may be expected from the actions of the two modalities separately.

A d(GpG)-platinated decanucleotide duplex is kinked. An extended NMR and molecular mechanics study

A conformational study of the double-stranded decanucleotide d(GCCG*G*ATCGC).d(GCGATCCGGC), with the G* guanines chelating a cis-Pt(NH3)2 moiety, has been accomplished using 1H and 31P NMR, and molecular mechanics. Correlation of the NMR data with molecular models has disclosed an equilibrium between several kinked conformations and has ruled out an unkinked structure. The deformation is localized at the CG*G*.CCG trinucleotide where the helix is kinked by approximately 60 degrees towards the major groove and unwound by 12-19 degrees. The models revealed an unexpected mobility of the cytosine complementary to the 5'-G*. This cytosine can stack on either branch of the kinked complementary strand. The energy barrier between the two positions has been calculated to be less than or equal to 12 kJ/mol. The NMR data are in support of rapid flip-flopping of this cytosine. An explanation for the strong downfield shift observed in the 31P resonance of the G*pG* phosphate is given.

The new antitumor compound, cis-[Pt(NH3)2(4-methylpyridine)Cl]Cl, does not form N7,N7-d(GpG) chelates with DNA. An unexpected preference for platinum binding at the 5'G in d(GpG)

The reaction of the antitumor active agent cis-[Pt(NH3)2(4-mepy)Cl]Cl (4-mepy stands for 4-methylpyridine) with d(GpG) has been investigated by 1H magnetic resonance spectroscopy. Initially, two mononuclear complexes cis-Pt(NH3)2(4-mepy)[d(GpG)-N7(1)] 1 and cis-Pt(NH3)2(4-mepy)[d(GpG)-N7(2)] 2 are formed in an unexpected ratio 65:35, as determined by 1H NMR and enzymatic digestion techniques. Both products react further with a second equivalent of cis-[Pt(NH3)2(4-mepy)Cl]Cl forming the dinuclear platinum complex [cis-Pt(NH3)2(4-mepy)]2[mu-d(GpG)- N7(1),N7(2)] 3. With [Pt(dien)Cl]Cl and [Pt(NH3)3Cl]Cl similar complexes are formed. No evidence was found for the formation of chelates cis-Pt(NH3)(4-mepy) [d(GpG)-N7(1),N7(2)], which would be formed upon ammonia release from the mononuclear complexes 1 and 2. Even addition of strong nucleophiles, like sodium diethyldithiocarbamate, thiourea, cysteine, or methionine, before or after reaction, do not induce the formation of a chelate. Under all conditions the N-donor ligands remain coordinated to Pt in 1,2 and 3. In addition, the results of bacterial survival and mutagenesis experiments with E. coli strains show that the in vivo formation of bifunctional adducts in DNA, comparable to those induced by cis-Pt(NH3)2Cl2, by treatment of cells with cis-[Pt(NH3)2(4-mepy)Cl]Cl is unlikely. Also, a mechanism of binding and intercalation is not supported by experimental data. All experiments suggest that the mechanism of action of this new class of antitumor agents must be different from that of cis-Pt(NH3)2Cl2.

NMR studies of an oligonucleotide with an unusual structure induced by platinum anti-cancer drugs

The 31P NMR spectra of Pt(en)[d(T1A2T3G4G5G6T7A8C9C10C11A12T13A14)] (14-mer) and Pt(en)[d(A2T3G4G5G6T7A8C9C10C11A12T13)] (12-mer) (en = ethylenediamine) each contain two signals far downfield (ca. -2.9 and -2.6 ppm from trimethylphosphate standard), two signals slightly downfield, and at least one signal slightly upfield of the normal range (ca -4.0 to -4.4 ppm). This pattern suggested a distorted structure. The unusual 31P signals of the 12-mer were assigned by analogy to signals of the 14-mer previously assigned by 17O-labeling methods. A combination of heteronuclear multiple-quantum coherence, one-dimensional and two-dimensional nuclear Overhauser effect (1D and 2D-NOE) and homonuclear shift correlation spectroscopy (COSY) experiments assigned all aromatic 1H signals of the 12-mer except H8 of G5 or G6. One of these H8 signals is missing from the spectrum and the nucleotide is labeled Gm. The other H8 is the most downfield signal and has a strong NOE to its H1'. Since this strong NOE indicates that this nucleoside exists in a syn conformation, it is labeled Gs. A strong NOE was observed between the Gs and A8 H8 signals. Several lines of evidence suggested a hairpin-like structure with a loop region (G6T7A8C9) and a stem region involving A2T3G4G5 and C10C11A12T13. The 31P signals for the stem region are within or slightly outside the normal range. 3JH3'-P values (3-6 Hz), measured by a 2D-J experiment, of stem nucleotides were characteristic for a DNA duplex. Imino signals for base pairs A2T13, T3A12, G4C11, and probably G5C10, and the observation of internucleotide NOE connectivities for these nucleotides (e.g. between an H8 signal and the H1' signal of the 5' nucleotide) suggested a right-handed helical structure. For the loop region, a distorted sugar-phosphate backbone is indicated by far downfield positions of the G5pG6 and A8pC9 31P signals, the 3JH3'-P values for C9p (8.0 Hz) and A8p (6.8 Hz), and the absence of H3'-P coupling for G5p. In the loop region, no imino signals or internucleotide NOEs characteristic of a right-handed duplex were observed. However, A8H8, C9H6, and C10H6 each exhibited unusual internucleotide NOEs to the H4' signal of the 5' residue. NOE crosspeaks between T7 1H signals and signals attributed to sugars of the Gs and Gm suggested that the T7 moiety is located within the space encircled by the loop. The few NOE crosspeaks, pH dependence, and Cu2+ broadening of C9 1H signals indicate an isolated location accessible to solvent.(ABSTRACT TRUNCATED AT 400 WORDS)

Reaction of cis- and trans-[PtCl2(NH3)2] with reduced glutathione inside human red blood cells, studied by 1H and 15N-[1H] DEPT NMR

Reactions of cis- and trans-[PtCl2(NH3)2] with glutathione (GSH) inside intact red blood cells have been studied by 1H spin-echo nuclear magnetic resonance (NMR). Upon addition of trans-[PtCl2(NH3)2] to a suspension of red cells, there was a gradual decrease in the intensity of the resonances for free GSH, and new peaks were observed that were assignable to coordinated GSH protons in trans-[Pt(SG)Cl(NH3)2], trans-[Pt(SG)2(NH3)2], and possibly the S-bridged complex trans-[[NH3)2PtCl)2SG]+. Formation of trans-[Pt(SG)2(NH3)2] inside the cell was confirmed from the 1H NMR spectrum of hemolyzed cells, which were ultrafiltered to remove large protein molecules; the ABM multiplet of the coordinated GSH cys-beta CH2 protons was resolved using selective-decoupling experiments. Seventy percent of the total intracellular GSH was retained by the ultrafiltration membrane, suggesting that the mixed complex trans-[Pt(SG)(S-hemoglobin)(NH3)2] also is a major metabolite of trans-[PtCl2(NH3)2] inside red cells. The reaction of cis-[PtCl2(NH3)2] with intracellular GSH was slower; only 35% of the GSH had been complexed after a 4-hr incubation compared to 70% for the trans isomer. There was a gradual decrease in the intensity of the GSH 1H spin-echo NMR resonances, but no new peaks were resolved. This was interpreted as formation of high-molecular weight Pt:GSH and mixed GS-Pt-S(hemoglobin) polymers. By using a 15N-[1H] DEPT pulse sequence, we were able to study the reaction of cis-[PtCl2(15NH3)2] with red cells at concentrations as low as 1 mM. 15NH3 ligands were released, and no resonances assignable to Pt-15NH3 species were observed after a 12-hr incubation.

How does cisplatin alter DNA structure? A molecular mechanics study on double-stranded oligonucleotides

Molecular models for two double-stranded decanucleotides, d(GCCG*G*ATCGC)-d(GCGATCCGGC) (1) and d(GCTG*G*ATCGC)-d(GCGATCCAGC) (2), with the G* guanines cross-linked by a cis-Pt(NH3)2 moiety, were calculated using molecular mechanics. Nine models for 1 and eight models for 2 are reported; in all of them, the double helix is kinked by approx. 60 degrees towards the major groove and slightly unwound. The model building has been guided by comparison with the NMR data available for duplex 1. The influence of the base at the 5'-side of the coordinated G*G* dinucleotide is discussed.

Formation and structure of reaction products of cis-PtCl2(NH3)2 with d(ApG) and/or d(GpA) in di-, tri- and penta-nucleotides. Preference for GpA chelation over ApG chelation

The reaction products of cis-PtCl2(NH)3)2 with several deoxyribonucleotides containing d(ApG) and/or d(GpA) have been studied. The various reaction products were separated by high-performance liquid chromatography and characterized by means of absorbance at 254 nm in combination with atomic absorption spectroscopy and 300-MHz 1H-NMR (pH dependence of the non-exchangeable base-protons, T1 relaxation time determinations). For the larger fragments the results from these techniques were confirmed by enzymatic degradation studies of the platinated fragments. The smallest of the investigated nucleotides, d(ApG) and d(GpA), both formed a variety of different platinum chelates. In the reaction with d(ApG) 15% cis-Pt(NH3)2-[d(ApG)N1(1),N7(2)] and 78% cis-Pt(NH3)2[d(ApG)N7(1),N7(2)] were found, 4% of the reacted material consisted of a 1 mol Pt/2 mol dinucleotide product, and 3% of an unidentified 1:1 product. From the main product two rotamers were found to occur: at room temperature, 81% anti,anti and 19% anti,syn product is present. With d(GpA) about equal amounts of N1,N7 and N7,N7 products were found; for both products the anti,anti and anti,syn conformations were found, respectively. Upon reaction of cis-PtCl2(NH3)2 with d(pApG) and d(pGpA) only the N7,N7 products were found; at room temperature and pH greater than 1.5 these products were present in anti,anti conformation. However, for the d(pApG)-platinum chelate at -20 degrees C a small amount (less than 5%) of a second product could be observed in NMR. For the d(pGpA)-platinum chelate a second N7,N7-coordinated product was observed when the pH of the NMR sample was lowered to 1.1 (at this pH the free 5'-phosphate group is protonated). With the larger fragments d(ApGpA), d(pApGpA) and d(TpApGpApT) the intra-molecular competition between the formation of the d(ApG) or the d(GpA) chelates could be studied. Using these nucleotides no N1-coordinated products or rotamers were observed. In the case of d(ApGpA) and d(TpApGpApT) the d(GpA) chelate (67% and 75% respectively) was favoured over the d(ApG) chelate, while with d(pApGpA) about equal amounts of both chelates were formed.

Alterations in the d(CpGpT) structure in solution as a result of [PtCl(diethylenetriamine)]+ binding

The trinucleotide d(CpGpT) reacts with [PtCl(dien)]Cl (dien = diethylenetriamine) to yield as a single adduct Pt(dien)[d(CpGpT)-N7(2)]. The structure of this adduct in solution has been analysed with the aid of NMR spectroscopy and compared with that of the unmodified trinucleotide. A change in the population of the S conformer of the guanosine deoxyribose ring and a syn preference of the guanine residue are the most important changes occurring upon platination. As a result the dC-dG stack disappears, whereas the dG-dT stack is hardly affected. The CD spectra of both platinated and free d(CpGpT) confirm the different nature of the two molecules.