Bifunctional intercalators: relationship of antitumor activity of diacridines to the cell membrane

Computer-aided molecular modeling and design of DNA-inserting molecules

Intercalators are molecules capable of sliding between base pairs without disturbing the overall stacking pattern. In addition, there may exist molecules capable of inserting into a base pair thereby disrupting the hydrogen bonds and replacing them with new hydrogen bonds. A molecule probably capable of inserting, i.e., an insertor, is the diketopiperazine cyclo-[Gly-Gly] (1). A barbiturate (2), alloxan (3), a pyrimidine derivative (4) and a hydantoin (5) were also studied as possible insertors. Furthermore, molecules such as ethyleneurea (6), succinimide (7), as well as a malonamide derivative (8) and oxamide derivatives (9-11) were studied in order to investigate the arrangement and the number of hydrogen bonds necessary for insertion. Molecules 12-14 were designed and studied for their capacity to act as bisinsertors and/or bisintercalators. These molecules feature two diketopiperazine moieties which are connected via a diphenyl(thio)ether, i.e., 12 and 13, or a bisphenol A spacer, i.e., 14. The latter molecule (14) seems a promising candidate as a bisinsertor.

Subcellular distribution of a nitroxide spin-labeled netropsin-acridine hybrid in living KB cells: Electron Spin Resonance Study

NETGA is an hybrid derivative which possesses an intercalating heterocyclic nucleus related to amsacrine and a minor groove binding squeletton related to netropsin. Cellular uptake of this drug has been studied by Electron Spin Resonance (ESR) spectroscopy using a spin-label derivative of NETGA (SL-NETGA). ESR determination of the kinetics of the drug repartition between the cytoplasm and nucleus showed that NETGA accumulated very rapidly and predominantly in the nucleus. Analysis of the anisotropic ESR spectra recorded in the nuclear compartment are in agreement with a strong binding of the drug to the DNA besides confirmed by a maximum delta Tm of 12 degrees C between the spin-label compound-DNA complex and the DNA alone.

Synthesis and biodistribution of 125I labeled bivalent analogs of practolol as potential myocardial imaging agents

Iodinated bivalent ligands 3 and 4 and a monovalent ligand 5 were prepared from the cardioselective beta-antagonist, practolol. 125I-labeled 3, 4 and 5 were prepared by solid phase isotopic exchange reaction with carrier-free Na125I and examined in rats as potential receptor-site-directed myocardial imaging agents. Biodistribution of these agents in rats indicated that 125I-3 and 125I-4 were localized in the heart similarly to 125I-5 and the [125I]iodobenzoyl (6) that was previously reported. Localization of 125I-3 and 125I-4, was more persistent in the heart than that of 125I-monovalent ligands 5 and 6. Heart-to-blood ratios of 125I-3 and 125I-4 were significantly lower than those of 125I-5 and 125I-6, due mainly to slow blood clearance rates of 125I-3 and 125I-4.

Studies on the fluorescence labeling of human red blood cell membrane ghosts with 4'-(9-acridinylamino)methanesulfon-m-anisidide

4'-(9-Acridinylamino)methanesulfon-m-anisidide (mAMSA) interacts with red cell membranes, resulting in the formation of fluorescent protein adducts. The mAMSA-membrane protein adducts exhibited an emission fluorescence maximum at 445 nm, with two shoulders at approximately 425 and 470 nm. The major labeled proteins were identified as spectrins 1 and 2 and bands 3, 4.1, 4.2 and 5. The fluorescence intensity increased with increasing mAMSA concentrations (0.03 to 1.5 mM), time (15-120 min), and temperature of the reaction. Results from sodium dodecyl sulfate gel electrophoresis show that mAMSA caused no detectable change in the molecular weight of membrane proteins. This indicates that mAMSA is a monofunctional, noncrosslinking agent. Other acridine analogs, 9-aminoacridine and acridine, did not fluorescently label membrane proteins, suggesting that the presence of the acridine nucleus is not sufficient for labeling. Addition of 2-mercaptoethanol to the mAMSA-membrane reaction mixtures reversed the fluorescence labeling. Furthermore, pretreatment of membrane proteins with N-ethylmaleimide or iodoacetamide prevented the formation of fluorescent mAMSA-membrane protein adducts. These data suggest that mAMSA interacts with sulfhydryl groups of the membrane proteins. When the membrane sulfhydryl groups were assayed by labeling with N-[ethyl-2-3H]ethylmaleimide, it was shown that the accessible membrane sulfhydryl groups were reduced after the mAMSA treatment. The above results suggest that mAMSA covalently binds to the sulfhydryl groups in the red cell membrane, with the production of fluorescent mAMSA-protein adducts.

Relationship of biochemical drug effects to their antitumor activity--II. Diacridines and membrane-related reactions

A method is presented that determines the degree of attachment of cancer cells to normal cells. This method may be useful in determining the extent to which treatment of normal cells (or of a tumor-bearing host) with a particular chemotherapeutic agent may affect the degree of attachment of cancer cells to the normal cells. The effects of several diacridines upon this process are described. In addition, we have determined the ability of individual diacridines to alter the permeability of P-388 cells; this effect has been related to their antitumor properties. In general, the most effective antitumor diacridines are those that cause minimal disruption of cell permeability. Conversely, diacridines that disrupt cell permeability tend to have poor antitumor properties. It is considered that the toxicity of these compounds may be a necessary consequence of the assays used for testing anticancer agents, and may not necessarily be related to their antitumor activity.

Photoaffinity labelling of submitochondrial membranes with the 3-azido analogue of 9-amino-3-chloro-7-methoxyacridine

9-Amino-3-azido-7-methoxyacridine has been synthesized and shown to be a suitable photoaffinity probe for the site(s) of interaction of 9-amino-3-chloro-7-methoxyacridine with submitochondrial membranes. Both the excitation and emission spectra of the azido analogue covalently bound to membranes in the energized state display distinctive differences from the spectra of labelled, non-energized membranes (i.e., in the absence of oxidizable substrate, or its presence when uncoupler (FCCP) is also present during photolysis). Enzymatic analyses indicate that the probe interacts with the ATPase and the respiratory chain enzymes; energization appears to afford some protection against inactivation. Electrophoresis of the labelled membranes and isolation of their lipid and protein components indicate that the spectral differences are attributable to differing interactions with the lipid components of energized, relative to non-energized, membranes. Similar results have been obtained with the 3-azido analogue of quinacrine (Mueller, D.M., Hudson, R.A. and Lee, C.P. (1982) Biochemistry 21, 1445-1453), which differs significantly, however, in the extent of its interactions with the enzymes of the respiratory chain and the ATPase. These results indicate that the energy-linked fluorescence responses of 9-aminoacridines with submitochondrial membranes arise from direct interactions with membrane components and may involve redistribution of the probe molecules and/or alteration of their microenvironments upon energization.

1H NMR study of the binding of bis(acridines) to d(AT)5.d(AT)5. 2. Dynamic aspects

Measurements of the 1H NMR spectra and relaxation rates were used to study the dynamic properties of 9-aminoacridine (9AA) and four bis(acridine) complexes with d(AT)5.d(AT)5. The behavior of the 9AA (monointercalator) and that of C8 (bisintercalator containing an eight-carbon atom linker chain) are entirely similar. For both compounds, the lifetime of the drug in a particular binding site is 2-3 ms at approximately 20 degrees C, and neither affects the A.T base pair opening rates. The complex with C10 (bisintercalator containing a 10-carbon atom linker chain) is slightly more stable than the C8 complex since its estimated binding site lifetime is 5-10 ms at 29 degrees C. Base pairs adjacent to the bound C10 are destabilized, relative to free d(AT)5.d(AT)5, but other base pairs in the C10 complex are little affected. Bis(acridine) pyrazole (BAPY) and bis(acridine) spermine (BAS) considerably stabilize those base pairs that are sandwiched between the two acridine chromophores, but in the BAS complex proton exchange from the two flanking base pairs appears to be accelerated, relative to free d(AT)5.d(AT)5. The lifetime of these drugs in specific binding sites is too long (>10 ms) to be manifested in increased line widths, at least up to 41 degrees C. An important conclusion from this study is that certain bisintercalators rapidly migrate along DNA, despite having large binding constants (K>10(6) M-1). For C8 and C10 complexes, migration rates are little different from those deduced for 9AA. The rigid linker chain in BAPY and the charge interactions in BAS retard migration of these two bisintercalators. These results provide new parameters that are useful in understanding the biochemical and biological properties of these and other bisintercalating drugs.

1H NMR study of the binding of Bis(acridines) to d(AT)5.d(AT)5. 1. Mode of binding

1H NMR has been used to investigate the mode of binding to d(AT)5.d(AT)5 of a series of bis(acridine) derivatives connected by different types of linker chains. The length and character (ionic, aliphatic, rigid, and flexible) of the linker chains are found to have a profound effect on the binding of these derivatives to the DNA. Bis(acridine) derivatives with linker chains shorter than 9 A monointercalate under the conditions used in the NMR study, whereas those bis(acridines) with chains of 9.8 A or longer bisintercalate. We find no evidence for the violation of the so-called neighbor exclusion principle. Although all of the bis(acridines) contain the same chromophores, their NMR spectra clearly demonstrate that they form complexes with d(AT)5.d(AT)5 which have different structures. This emphasizes the important effect that the linker chain has on the structure of the intercalation complex.

Bleomycin control of transplasma membrane redox activity and proton movement in HeLa cells

Bleomycin, tallysomycin A, tallysomycin S10b and copper-bleomycin have been tested for their capacity to inhibit the transplasma membrane electron transport and associated proton release by HeLa cells. Transplasma membrane redox activity is measured using reduction of external ferricyanide by the cells. At 75 micrograms/ml bleomycin, tallysomycin A and tallysomycin S10b gave a maximum of 65% inhibition of the ferricyanide reduction rate; half-maximum inhibition was observed at 30 micrograms/ml. The copper-bleomycin complex was slightly more effective as an inhibitor with half-maximum inhibition at 20 micrograms/ml. Survival of cells after 1 hr of drug treatment was 50% at 25 micrograms/ml for bleomycin and copper-bleomycin and at 75 micrograms/ml for tallysomycin A. Tallysomycin A and tallysomycin S10b gave 75 to 83% inhibition of ferricyanide-induced proton extrusion, respectively at 50 micrograms/ml, whereas bleomycin and copper-bleomycin appeared to be slightly less effective with 50 to 60% inhibition, respectively, at 50 micrograms/ml. In all aspects studied, which included transplasma membrane ferricyanide reduction, ferricyanide-induced proton release, and cell survival, there were significant effects by these compounds on HeLa cells in the range of 25-50 micrograms/ml.

Amsacrine (m-AMSA): a new antineoplastic agent. Pharmacology, clinical activity and toxicity

The synthetic aminoacridine derivative amsacrine (m-AMSA) is capable of preventing DNA from serving as a template in replication and DNA synthesis. This mechanism of action is similar to that of anthracyclines, but clinical evidence suggests the lack of cross-resistance. The recommended dosage in patients with solid tumors is 90-120 mg/m2 intravenously every 3-4 weeks. Despite the initial encouraging reports from experimental models, m-AMSA has shown no real impact in the treatment of patients with a wide variety of solid tumors. In relapsed acute nonlymphocytic leukemia, 20-30% of patients will achieve complete remission. An increased remission rate is obtained when m-AMSA is combined with other agents, especially with high-dose cytosine arabinoside, with a complete remission rate of 50-60% in relapsed patients. Currently, several phase III trials are evaluating m-AMSA combinations against daunorubicin-containing regimens in patients with previously untreated acute leukemia. The potential role of these regimens in this disease remains to be defined.

Comparative cardiac oxygen radical metabolism by anthracycline antibiotics, mitoxantrone, bisantrene, 4'-(9-acridinylamino)-methanesulfon-m-anisidide, and neocarzinostatin

This study examined the effects of various anthracycline antibiotics and mitoxantrone, bisantrene, 4'-(9-acridinylamino)-methanesulfon-m-anisidide (m-AMSA), and neocarzinostatin on oxygen radical formation by cardiac sarcoplasmic reticulum and submitochondrial particles. Doxorubicin, daunorubicin, rubidazone, and aclacinomycin A stimulated superoxide production by both heart fractions in a dose-dependent fashion that appeared to follow saturation kinetics. The anthracycline drugs also significantly increased hydrogen peroxide production by heart sarcosomes and submitochondrial particles. On the other hand, mitoxantrone, bisantrene, m-AMSA, and neocarzinostatin did not significantly enhance cardiac reactive oxygen metabolism. Thus, it is unlikely that the mechanism of the cardiac toxicity produced by mitoxantrone and m-AMSA in patients previously treated with anthracycline drugs can be directly related to oxidation-reduction cycling catalyzed by cardiac flavin dehydrogenases.

Subcellular distribution of nitroxide spin-labelled 9-aminoacridine in living KB cells

A spin-labelled derivative of 9-aminoacridine, the AATEMPO, was studied with respect to its localization in KB cells in vivo. It was found that both nuclear and mitochondrial DNAs were targets for this intercalating dye. The observed speed of intercalation and the absence of a blocked signal in cellular membranes suggested that no receptor -or carrier- proteins were implied in the penetration process. No changes in cell membrane fluidity were observed following the administration of m-AMSA, the unlabelled structural analog, to the living cells, which seemed to exclude the plasma membrane as a possible site of action of 9-aminoacridines. Side effects of phenol/chloroform mixtures and high-salt concentrations on the intercalation phenomenon were also described.

Tranquillizers can block mitogenesis in 3T3 cells and induce differentiation in Friend cells

Compounds of diverse structure induce murine Friend erythroleukaemia (MEL) cell differentiation. Seeking a common property, Bernstein reported that the relationship between activity and octanol/water partition coefficients for inducers resembled that reported for anaesthetics; moreover, anaesthetics inhibited induction. Since anaesthetics were known to increase membrane fluidity, they suggested that inducers might decrease it. Reporting evidence against unitary theories of anaesthesia, Richards et al. suggested that lipophilic drugs competed, according to their individual structure, with membrane lipids for hydrophobic regions on membrane proteins. The antagonism between pairs of anaesthetics, anaesthetics and inducers and pairs of inducers, might thus be explained economically by Richards' anaesthesia model, inducers and lipophilic drugs acting by similar rather than contrary mechanisms. Lipophilic drugs should, therefore, induce differentiation. We report here that some tranquillizers do so. As reported elsewhere for classical inducers, they also block non-differentiating 3T3 cells in pre-S. These findings may have application in cancer chemotherapy.

Cell differentiation and flagellar elongation in Naegleria gruberi. Dependence on transcription and translation

This paper presents evidence that the phenotypic transformation of Naegleria gruberi from amebae to flagellates that occurs when cells are placed in a nutrient-free aqueous environment is dependent on transcription and translation. RNA and protein are synthesized during the hour-long differentiation. Actinomycin D and daunomycin selectively inhibit RNA synthesis, and cycloheximide selectively inhibits protein synthesis, throughout the time required for differentiation. These inhibitors prevent differentiation if added soon after the cells are transferred to nonnutrient buffer but cease to block specific differentiation events at subsequent, reproducible times, the transition points. After each transition point, morphogenesis can occur in the presence of the inhibitor and in the virtual absence of transcription or translation. A map of the transition points indicates that RNA synthesis is required until halfway through the temporal process from initiation to flagellum assembly, and that protein synthesis is required until three-fourths of the way through. Even when flagellum outgrowth can occur in the presence of cycloheximide, the length of the flagella formed is determined by the extent of synthesis of an unknown "limiting precursor." The transition points for formation of flagella and for formation of the streamlined flagellate body shape are temporally separate. These results indicate that differentiation in Naegleria involves a redirection of cell metabolism to produce new RNA and protein molecules that are essential for morphogenesis.

Novel acriflavin resistance genes, acrC and acrD, in Escherichia coli K-12

Acriflavine-resistant mutants were isolated from an acriflavine-sensitive (acrA) strain of Escherichia coli K-12 and then tested for temperature sensitivity of cell division. Genetic analysis characterized two new genetic loci, acrC and acrD. The former was mapped between tonA and proA, and the latter between the origin of genetic transfer of HfrH and serB. acrC and acrD mutants could divide but did not initiate a new round of deoxyribonucleic acid (DNA) replication at 43 degrees C. DNA synthesis of the acrC mutant cells ceased after a period of time following temperature shift-up, and thereafter DNA degradation occurred. However, cell mass continued to increase for a long time at the nonpermissive temperature. On the other hand, DNA synthesis of the acrD mutant cells ceased soon after the shift-up, and the cell mass did not appreciably increase during the prolonged incubation.

An electron microscopic study of the photochemical cross-linking of DNA in guinea pig epidermis by psoralen derivatives

Albino guinea pigs were treated with psoralen derivatives plus 320--400 nm ultraviolet radiation, and DNA was extracted from their epidermis. The DNA was assayed for the presence of interstrand cross-links by standard denaturation-renaturation assays and by a new technique, electron microscopy of the DNA under totally denaturing conditions. The latter method allows individual cross-links to be directly observed and counted. When either 4,5',8-trimethylpsoralen or 8-methoxypsoralen was applied topically to the skin (8--20 microgram/cm2) or administered orally (10--12 mg/kg body weight), followed by exposure to 320--400 nm ultraviolet radiation, most of the epidermal DNA was found to contain a high frequency of cross-links. For example, oral or topical trimethylpsoralen treatment gave an average of one cross-link per 250 nucleotide pairs or about 3 . 10(5) cross-links per guinea pig chromosome. When the dose of either drug was decreased 20-fold to the level used in the clinical treatment of psoriasis, however, no cross-links coulld be detected in the epidermal DNA. The electron microscopic assay is sensitive enough that we can put an upper limit of 1 cross-link per 10(6) nucleotide pairs (80 cross-links per chromosome) for the low dose studies. The significance of these findings to the understanding of the effectiveness of psoralens in psoriasis therapy is discussed.

Electrostatic complexes of mitomycin C with nucleic acids and polyanions

Reductively activated mitomycin C exhibits strong, non-covalent electrostatic binding to polyanions such as polyvinylsulfate and polyphosphate. The protonated C-2 amino group generated by the reduction is most likely responsible for this type of interaction. At moderate drug and salt concentrations only covalent binding to nucleic acids is observable. This is shown to be guanine-specific in DNA for the first time, as well as in synthetic polyribo- and polydeoxyribonucleotides at 10--20 times higher binding levels than previously tested. At higher mitomycin C concentration, however, strong non-covalent electrostatic binding to nucleic acids also occurs, resulting in a binding ratio up to 1 mol drug bound per mol mononucleotide, although this non-specific binding is relatively inhibited compared to polyvinylsulfate. Salts also have an inhibitory effect on the non-specific binding to nucleic acids. A series of mitomycin derivatives were compared for their binding and cross-linking abilities using DNA as substrate, with the following results: (a) the presence of a basic nitrogen . funtion at C-2 promotes binding, both covalent and electrostatic, presumably by kinetically facilitating the approach between positively charged nitrogen and DNA. (b) The aziridine ring is the major covalent binding site, indispensable for crosslinking and determines the guanine-specificity of the binding.