The mechanism of nitroimidazole damage to DNA: coulometric evidence

New horizons in the treatment of tuberculosis

The development of new chemotherapy for the treatment of tuberculosis has three major objectives: first, the development of faster-acting drugs to shorten the duration of treatment; second, the development of novel antimicrobials to counter the emergence of bacteria resistant to current therapies; and, third, the development of chemotherapeutics that specifically target dormant bacilli to treat the one-third of the world's population latently infected with tubercle bacilli. Strategies based upon optimizing the inhibition of known targets require an extensive knowledge of the detailed mechanism of action of current antimycobacterial agents. For many agents such as isoniazid, ethambutol, rifampin, and pyrazinamide such knowledge is now available. Strategies based upon the identification of novel targets will necessitate the identification of biochemical pathways specific to mycobacteria and related organisms. Many unique metabolic processes occur during the biosynthesis of mycobacterial cell wall components, and some attractive new targets have emerged. The development of targets specific to latency will require a detailed picture of the metabolism and biochemical pathways occurring in dormant bacilli. Recent evidence suggests that anaerobic metabolic pathways may operate in dormant bacilli, and the enzymes involved in such pathways may also provide significant new targets for intervention. The combination of the mycobacterial genome sequence that is anticipated to become available this year with an improved understanding of the unique metabolic processes that define mycobacteria as a genus offers the greatest hope for the elimination of one of mankind's oldest enemies.

Metabolism of a 5-nitroimidazole in susceptible and resistant isogenic strains of Bacteroides fragilis

We investigated the metabolism of dimetridazole (1,2-dimethyl-5-nitroimidazole) (DMZ) by the resting cell method in a susceptible strain of Bacteroides fragilis and in the same strain containing the nimA gene, which conferred resistance to 5-nitroimidazole drugs. In both cases, under strict anaerobic conditions DMZ was metabolized without major ring cleavage or nitrate formation. However, one of two distinct metabolic pathways is involved, depending on the susceptibility of the strain. In the susceptible strain, the classical reduction pathway of nitroaromatic compounds is followed at least as far as the nitroso-radical anion, with further formation of the azo-dimer: 5,5'-azobis-(1,2-dimethylimidazole). In the resistant strain, DMZ is reduced to the amine derivative, namely, 5-amino-1,2-dimethylimidazole, preventing the formation of the toxic form of the drug. The specificity of the six-electron reduction of the nitro group, which is restricted to 4- and 5-nitroimidazole, suggests an enzymatic reaction. We thus conclude that nimA and related genes may encode a 5-nitroimidazole reductase.

Reductive activation of nitroheterocyclic compounds

1. Nitroaromatic compounds are important chemotherapy agents. 2. Their selective toxicity is determined by reduction to the biologically active form in the absence of oxygen. 3. Nitroaromatics are extensively used in the treatment of anaerobic infections and to target hypoxic tumor cells in cancer therapy. 4. Possible mutagenic action is related to the relative ease of nitro group reduction. 5. The mode of action and clinical application of nitroaromatic compounds is summarized.

The interaction of reduced metronidazole with DNA bases and nucleosides

The electrochemical behavior of the 1-electron couple for the bioreductive drug metronidazole has been examined in the presence and absence of the biological target molecules, DNA bases, and nucleosides, including uracil and uridine. Using cyclic voltammetry as the investigation technique, the change in return-to-forward peak current ratio, ipr/ipf, from the control, recorded in the absence of target, was measured as a function of scan rate and biological target concentration. All target molecules, except adenosine and guanine, resulted in interaction with RNO2.-, as measured by the decrease in the ipr/ipf ratio in the following order of increasing reactivity: adenine, guanosine, thymine, uracil, uridine, and thymidine (at a metronidazole:target ratio of 1:1). No decrease in ipr/ipf was observed with cytosine or cytidine until ratios of 1:20 and 1:30, respectively, were attained. An approximately linear relationship was found between the percentage change in the CV response and log[target] allowing us to determine the sensitivity of RNO2.- to the concentration of the target species. The implication for the biological action of metronidazole and other nitro-heterocyclic drugs is discussed.

Electrochemical studies and DNA damaging effects of the benzotriazine-N-oxides

The electrochemical behaviour of eight benzotriazine 1,4 di-N-oxides has been examined and compared with the mono- and zero-N-oxides. The di-N-oxides all show two reduction steps, an irreversible followed by a quasi-reversible response assigned to the 4 electron reduction of both N-oxide groups, followed by the 2 electron reduction of the benzotriazine ring. Mono- and zero-N-oxides show only a single, quasi-reversible reduction step, similar in character to the second reduction of the di-N-oxides. This has been assigned to reduction of the benzotriazine ring, with the available, redox-active, N-oxide group of the mono-N-oxide complex being reduced at less negative potentials, but only after ring reduction, hence only a single electrode response. The importance of reductive activation of the N-oxide group has been examined using a phi X174 double transfection technique which assays biologically relevant DNA damage. For the di-N-oxides, no effect on DNA was recorded under oxic conditions, however, DNA damage was marked under anoxic reduction conditions. The extent of DNA damage was found to increase with the acidity of the medium, suggesting the protonated form of the reduction product as being responsible for the cytotoxic action. The mono-N-oxide was shown to be biologically inactive under all conditions.

Electrochemical characteristics of nitroheterocyclic compounds of biological interest. VI. The misonidazole radical anion

The addition of four aprotic solvents to misonidazole in an aqueous buffer system has been examined electrochemically. Qualitatively they all result in separation of the initial irreversible 4 electron reduction step into two stages, the RNO2/RNO2- and RNO2-/RNHOH couples respectively. Despite some difficulties in achieving measurements for the discrete RNO2/RNO2- without interference from the following reduction step, it was clear that the various aprotic solvents influenced the lifetime of the RNO2- species to different degrees. Resolution of the two processes was best achieved using a water-acetone system and this has been employed to study the lifetimes of the misonidazole radical anion as a function of acetone content and drug concentration. Analysis of the cyclic voltammetric response showed a second order decay pathway, in line with the metronidazole system studied under similar conditions. This has been compared with results from pulse radiolysis work, which suggested a first order reaction of unknown pathway for 2-nitroimidazole radical anions.

Electrochemical characteristics of nitroheterocyclic compounds of biological interest. V. Measurement and comparison of nitro radical lifetimes

Using mixed aqueous/dimethylformamide solvents we have generated nitro radical anions by electrochemical reduction of nitroaromatic compounds. Six drugs have been examined: metronidazole, nitrofurazone, nifuroxime, chloramphenicol, M&B 4998 and 4(5)-nitroimidazole, chosen to represent a variety of ring structures and a range of reduction potentials. Analysis of the cyclic voltammetric response as a function of scan rate and dimethylformamide content yields information on the reactivity of RNO2.-. A kinetic analysis of the return-to-forward peak current ratio based on a theoretical treatment was employed. Second-order kinetics for the decay of RNO2.- for all six drugs examined was established. By extrapolation, first half-lives in purely aqueous media were found to increase in the order: nitrofurazone, nifuroxime, chloramphenicol, metronidazole and M&B 4998 (from 8.9 x 10(-2) seconds for nitrofurazone to 98s for M&B 4998 at a radical anion concentration of 1 x 10(-6) mol/dm3). Comparison with reduction potentials showed that as the lifetime of RNO2.- increased, the drug became progressively less electron-affinic (reduced at more negative potentials). The reactivity of RNO2.- was also examined in relation to the DNA damaging capability following electrochemical reduction of these nitroaromatic drugs.

Kinetic studies of four types of nitroheterocyclic radicals by pulse radiolysis. Correlation of pharmacological properties to decay rates

The chemical properties of the nitro radical of four types of nitroheterocyclic compounds, nitrofurans, 2-nitroimidazoles, 4(5)-nitroimidazoles, 5-nitroimidazoles, having radiosensitizing and cytotoxic properties, have been studied by pulse radiolysis. The acid-base equilibria involving the nitro radical, the imidazole ring and some residues on the heterocycle have been determined. The pH-dependence of the rate of the disproportionation reaction of the nitro radical have been extensively studied. While the nitro radical derived from nitrofurans, 4- and 5-nitroimidazoles had a second-order decay, those of the 2-nitroimidazoles were found to decay through simultaneous first-order and second-order processes. Intrinsic second-order rate constants of the decay of the radical species in its various acidic and basic forms, could be determined. The intrinsic rate constants that determine the overall decay rates in the physiologically important 6 to 7.5 pH-range could be related to the one-electron redox potential E7(1). The implication of such chemical properties to enzyme-catalyzed reduction processes and to the mechanisms of radiosensitization and cytotoxicity of nitroheterocyclic compounds are briefly discussed. Pharmacological properties such as in vitro radiosensitization efficiency or metabolic reduction rates could be related to two of the nitro radical intrinsic disproportionation rates.

Electrochemical characteristics of nitro-heterocyclic compounds of biological interest. III. Nitroso derivative formation

Upon electrolytic reduction of a range of nitro-aromatic complexes (including imidazoles, benzenoids, furans and pyrazoles) an associated oxidation-reduction process is observed at more positive potentials with respect to nitro group reduction when using repeat scan cyclic voltammetry. This new couple has been identified as the reversible first reduction of the nitroso derivative for chloramphenicol, by the addition of a genuine sample of nitrosochloramphenicol to the electrochemical cell. We have failed to observe formation of the new redox-active species for five 5-nitroimidazoles examined. Possible reaction schemes for nitroso formation under electrolytic reduction conditions and the importance of the nitroso redox couple with respect to the cytotoxic action of the parent drug are discussed. The applicability of nitrosochloramphenicol as a model for the behaviour of nitro-heterocycles in general is shown.

Electrochemical characteristics of nitro-heterocyclic compounds of biological interest. II. Nitrosochloramphenicol

The electrochemical characteristics of nitrosochloramphenicol have been studied in aqueous buffer systems (pH 7.1) using direct current (d.c.) and differential pulse polarography, cyclic voltammetry and coulometric techniques. Up to 4 charge-transfer steps can be identified. The first reduction step is reversible both chemically and electrochemically, the charge-transfer product showing no tendency to undergo further reaction on the electrochemical time-scale. In contrast, the second reduction step is irreversible, with the product undergoing a fast following reaction to yield a redox-active species which was detected by cyclic voltammetry. From the data and by comparison with related systems, two reduction mechanisms are possible and are discussed.

Role of hydrogenase 1 of Clostridium pasteurianum in the reduction of metronidazole

Competition studies between the phosphoroclastic reaction and the metronidazole reduction reaction using dialyzed crude cell-free extracts of Clostridium pasteurianum which were essentially devoid of Hydrogenase 1 activity demonstrated that this enzyme plays an important role in the reduction of metronidazole. To determine further the exact function for Hydrogenase 1 in the reduction of the drug, this enzyme was highly purified from C. pasteurianum. Metronidazole reduction activity copurified with Hydrogenase 1 specific activity throughout the purification procedure. Drug reduction required the presence of an electron carrier and could not be accomplished by the enzyme alone. Ferredoxin, and also the low potential electron carrier dyes, methyl and benzyl viologen, and the flavin coenzymes, FAD and flavin mononucleotide (FMN), could couple the reduction of metronidazole. Hydrogenase 1 activity and its metronidazole reduction activity were inactivated irreversibly in the presence of oxygen. Metronidazole could be reduced only by an electron carrier-Hydrogenase 1 mechanism or directly by sodium dithionite.

DNA damage induced by reduced nitroimidazole drugs

Five nitroimidazole drugs were reduced electrolytically and by gamma-radiolysis at fast (300 mumoles or 100% per hr) and slow (3-9 mumoles or 1-3% per hr) reduction rates in the presence of Escherichia coli DNA and single stranded or double stranded DNA from the bacteriophage phi X174. The degree of DNA damage depends upon the rate of drug reduction, where slow reduction produces more damage than fast reduction. The efficiency of damage produced is in the order metronidazole greater than ornidazole greater than azomycin greater than misonidazole greater than benznidazole which reveals a linear correlation between the one-electron reduction potential (E17) and the negative logarithm of the concentration of reduced drug at which 37% of the original DNA activity remains. Damage is not influenced by the presence of O2 at least between about 1-100 ppm. We suggest the protonated one-electron nitro radical anion as a possible candidate for the active damaging species and explain the basis of the relative cytotoxicity of these drugs under conditions of hypoxia.

Effect of radiation-induced reduction of nitroimidazoles on biologically active DNA

Radiation-chemical reductions have been carried out with several nitroimidazoles. Reduction of these drugs in the presence of single-stranded phi chi 174 DNA causes extensive lethal damage. However, relatively stable (end) products, do not contribute to the damage, although glyoxal is potentially toxic. This demonstrates that a short-lived intermediate in the reduction process is responsible. Further, the quantity of damage in the DNA depends on both dose (reduction)-rate and also the nature of the drug.

DNA damage induced by reductively activated nitroimidazoles--pH effects

The effect of pH on E. coli DNA damage measured viscometrically and induced by electrolytically reduced metronidazole and misonidazole has been studied, together with the effect on the statistical average number of electrons required for reduction, measured by high-resolution coulometry, and nitrite production measured colorimetrically. In general, nitroimidazole-induced DNA damage is greatest at acid pH and decreased at alkaline pH, but whereas metronidazole exhibits a linear relationship between DNA damage and increased pH, misonidazole shows a plateau between pH 6 and 8. The electron requirements for complete reduction (n) vary with pH. For misonidazole n increases with an increase in pH both in the absence and presence of DNA with a shallow plateau between pH 6 and 8. In contrast, for metronidazole, n decreases with increased pH and exhibits breakpoints between pH 6 and 8. Nitrite (NO2-) production is linear with increased pH for misonidazole but for metronidazole (NO2-) production shows a sudden increase at 7.5 yielding ca. 35% on a molar basis. The results may reflect differences in the relative stability and reactivity of the nitro radical anion.

Electrolytic reduction of nitroheterocyclic drugs leads to biologically important damage in DNA

The effect of electrolytic reduction of nitroimidazole drugs on biologically active DNA was studied. The results show that reduction of the drugs in the presence of DNA affects inactivation for both double-stranded (RF) and single-stranded phi X174 DNA. However, stable reduction products did not make a significant contribution to the lethal damage in DNA. This suggests that probably a short-lived intermediate of reduction of nitro-compounds is responsible for damage to DNA.

Induction of DNA strand breaks by RSU-1069, a nitroimidazole-aziridine radiosensitizer. Role of binding of both unreduced and radiation-reduced forms to DNA, in vitro

[2-14C]-RSU-1069 [1-(2-nitro-1-imidazolyl)-3-(1-aziridino)-2-propanol], either as a parent (unreduced) or following radiation reduction, binds to calf thymus DNA in vitro. Radiation-reduced RSU-1069 binds to a greater extent and more rapidly than the parent compound. RSU-1137, a nonaziridino analogue of RSU-1069, binds following radiation reduction. Radiation-reduced misonidazole (1-(2-nitro-1-imidazolyl)-3-methoxy-2-propanol) exhibits binding ratios a thousand-fold less than those of reduced RSU-1069. There is no evidence for binding of parent misonidazole. Both parent and reduced RSU-1069 cause single strand breaks (ssbs) in pSV2 gpt plasmid DNA with the reduced compound causing a greater number of breaks. Parent and reduced RSU-1137 and misonidazole do not cause ssbs. It is inferred that the aziridine moiety present in both parent and reduced RSU-1069 is required for ssb production. RSU-1069 reacts with inorganic phosphate probably via nucleophilic ring-opening of the aziridine fragment. Incubation of plasmid DNA with reduced RSU-1069 in the presence of either phosphate or deoxyribose-5-phosphate at concentrations greater than 0.35 mol dm-3 prevents strand breakage, whereas 1.2 mol dm-3 deoxyribose does not protect against strand breakage formation. From these findings it is proposed that the observed binding to DNA occurs via the aziridine and the reduced nitro group of RSU-1069 and that these two have different target sites. Binding to DNA via the reduced nitro group may serve to increase aziridine attack due to localization at or near its target.

DNA damage in normal and neoplastic mouse tissues after treatment with misonidazole in vivo

Alkaline elution has been used to examine the integrity of DNA isolated from various tissues from mice treated with misonidazole (MISO). High doses (1-3 mg/g) of MISO caused extensive DNA strand breakage in cells isolated from two fibrosarcoma tumors that were known to contain hypoxic cells, and also in cells from certain normal tissues (liver and kidney in particular). The incidence of strand breaks gives further support to the suggestion that MISO can be metabolically nitroreduced beyond the singly reduced nitro radical-anion in some normal tissues as well as in hypoxic tumor cells, generating DNA-reactive species. Nitroreductases must therefore be able to compete successfully with molecular oxygen for the MISO nitro radical-anion in such tissues.