DNA adducts and DNA damage by antineoplastic and carcinogenic N-nitrosocompounds

Carbamoylating activity associated with the activation of the antitumor agent laromustine inhibits angiogenesis by inducing ASK1-dependent endothelial cell death

The anticancer agent 1,2-bis(methylsulfonyl)-1-(2-chloroethyl)-2-[(methylamino)carbonyl]hydrazine (laromustine), upon decomposition in situ, yields methyl isocyanate and the chloroethylating species 1,2-bis(methylsulfonyl)-1-(2-chloroethyl)hydrazine (90CE). 90CE has been shown to kill tumor cells via a proposed mechanism that involves interstrand DNA cross-linking. However, the role of methyl isocyanate in the antineoplastic function of laromustine has not been delineated. Herein, we show that 1,2-bis(methylsulfonyl)-1-[(methylamino)carbonyl]hydrazine (101MDCE), an analog of laromustine that generates only methyl isocyanate, activates ASK1-JNK/p38 signaling in endothelial cells (EC). We have previously shown that ASK1 forms a complex with reduced thioredoxin (Trx1) in resting EC, and that the Cys residues in ASK1 and Trx1 are critical for their interaction. 101MDCE dissociated ASK1 from Trx1, but not from the phosphoserine-binding inhibitor 14-3-3, in whole cells and in cell lysates, consistent with the known ability of methyl isocyanate to carbamoylate free thiol groups of proteins. 101MDCE had no effect on the kinase activity of purified ASK1, JNK, or the catalytic activity of Trx1. However, 101MDCE, but not 90CE, significantly decreased the activity of Trx reductase-1 (TrxR1). We conclude that methyl isocyanate induces dissociation of ASK1 from Trx1 either directly by carbamoylating the critical Cys groups in the ASK1-Trx1 complex or indirectly by inhibiting TrxR1. Furthermore, 101MDCE (but not 90CE) induced EC death through a non-apoptotic (necroptotic) pathway leading to inhibition of angiogenesis in vitro. Our study has identified methyl isocyanates may contribute to the anticancer activity in part by interfering with tumor angiogenesis.

Thioredoxin reductase is inhibited by the carbamoylating activity of the anticancer sulfonylhydrazine drug laromustine

The thioredoxin system facilitates proliferative processes in cells and is upregulated in many cancers. The activities of both thioredoxin (Trx) and its reductase (TrxR) are mediated by oxidation/reduction reactions among cysteine residues. A common target in preclinical anticancer research, TrxR is reported here to be significantly inhibited by the anticancer agent laromustine. This agent, which has been in clinical trials for acute myelogenous leukemia and glioblastoma multiforme, is understood to be cytotoxic principally via interstrand DNA crosslinking that originates from a 2-chloroethylating species generated upon activation in situ. The spontaneous decomposition of laromustine also yields methyl isocyanate, which readily carbamoylates thiols and primary amines. Purified rat liver TrxR was inhibited by laromustine with a clinically relevant IC(50) value of 4.65 μM. A derivative of laromustine that lacks carbamoylating activity did not appreciably inhibit TrxR while another derivative, lacking only the 2-chloroethylating activity, retained its inhibitory potency. Furthermore, in assays measuring TrxR activity in murine cell lysates, a similar pattern of inhibition among these compounds was observed. These data contrast with previous studies demonstrating that glutathione reductase, another enzyme that relies on cysteine-mediated redox chemistry, was not inhibited by methylcarbamoylating agents when measured in cell lysates. Mass spectrometry of laromustine-treated enzyme revealed significant carbamoylation of TrxR, albeit not on known catalytically active residues. However, there was no evidence of 2-chloroethylation anywhere on the protein. The inhibition of TrxR is likely to contribute to the cytotoxic, anticancer mechanism of action for laromustine.

Impairment of APE1 function enhances cellular sensitivity to clinically relevant alkylators and antimetabolites

Base excision repair (BER) is the major pathway for removing mutagenic and cytotoxic oxidative and alkylation DNA modifications. Using a catalytically inactive, dominant negative protein form of human APE1, termed ED, which binds with high affinity to substrate DNA and blocks subsequent repair steps, we assessed the role of BER in mediating cellular resistance to clinically relevant alkylating drugs and antimetabolites. Colony formation assays revealed that ED expression enhanced cellular sensitivity to melphalan not at all; to decarbazine, thiotepa, busulfan and carmustine moderately (1.2- to 2.4-fold); and to streptozotocin and temozolomide significantly (2.0- to 5.3-fold). The effectiveness of ED to promote enhanced cytotoxicity generally correlated with the agent's (a) monofunctional nature, (b) capacity to induce N(7)-guanine and N(3)-adenine modifications, and (c) inability to generate O(6)-guanine adducts or DNA cross-links. ED also enhanced the cell killing potency of the antimetabolite troxacitabine, apparently by blocking the processing of DNA strand breaks, yet had no effect on the cytotoxicity of gemcitabine, results that agree well with the known efficiency of APE1 to excise these nucleoside analogues from DNA. Most impressively, ED expression produced an approximately 5- and 25-fold augmentation of the cell killing effect of 5-fluorouracil and 5-fluorodeoxyuridine, respectively, implicating BER in the cellular response to such antimetabolites; the increased 5-fluorouracil sensitivity was associated with an accumulation of abasic sites and active caspase-positive staining. Our data suggest that APE1, and BER more broadly, is a potential target for inactivation in anticancer treatment paradigms that involve select alkylating agents or antimetabolites.

Inhibition of human DNA polymerase beta activity by the anticancer prodrug Cloretazine

The antineoplastic prodrug Cloretazine exerts its cytotoxicity via a synergism between 2-chloroethylating and carbamoylating activities that are cogenerated upon activation in situ. Cloretazine is reported here to inhibit the nucleotidyl-transferase activity of purified human DNA polymerase beta (Pol beta), a principal enzyme of DNA base excision repair (BER). The 2-chloroethylating activity of Cloretazine alkylates DNA at the O(6) position of guanine bases resulting in 2-chloroethoxyguanine monoadducts, which further react to form cytotoxic interstrand DNA crosslinks. Alkylated DNA is often repaired via BER in vivo. Inhibition of the polymerase activity of Pol beta may account for some of the synergism between Cloretazine's two reactive subspecies in cytotoxicity assays. This inhibition was only observed using agents with carbamoylating activity. Furthermore, while therapeutically relevant concentrations of Cloretazine inhibited the polymerase activity of Pol beta, the enzyme's lyase activity, which may also participate in BER, was not significantly inhibited.

Genotoxicity of glycidamide in comparison to 3-N-nitroso-oxazolidin-2-one

Acrylamide (AA) is generated by thermal processing of foods, depending on processing conditions and precursor availability. AA is not genotoxic by itself but becomes activated to its genotoxic metabolite glycidamide (GA) via epoxidation, mediated primarily by cytochrome P450 2E1. In the Comet assay in V79 cells and in human lymphocytes, GA induced DNA damage down to 300 microM concentration (4 h). After post-treatment with the DNA repair enzyme formamidopyrimidine-DNA-glycosylase (FPG), DNA damage became already detectable at 10 microM (4 h). By comparison, the N-nitroso compound 3- N-nitroso-oxazolidin-2-one (NOZ-2) is a much stronger genotoxic agent, significantly inducing DNA damage already at 15 min (3 microM). Post-treatment with FPG in this case did not enhance response. GA induced DNA damage in V79 cells rather slowly, with first response detectable at 4 h. The hPRT forward mutation test encompasses 5 days of expression time during which also repair can take place. GA-induced hPRT mutations only became detectable at concentrations of 800 microM and above. This is 80-fold higher than the lowest significant response to GA in the Comet assay (10 microM with FPG). In contrast, NOZ-2 was as effective in the hPRT test as in the Comet assay (3 microM). These results demonstrate substantial differences in the genotoxic potency of GA and NOZ-2. Whereas NOZ-2 is a pontent genotoxic mutagen, GA in comparison shows only low genotoxic and mutagenic potential, presumably as a result, at least in part, of preferential N7-G alkylation.

Alleviation of 1,N6-ethanoadenine genotoxicity by the Escherichia coli adaptive response protein AlkB

1,N(6)-ethanoadenine (EA) forms through the reaction of adenine in DNA with the antitumor agent 1,3-bis(2-chloroethyl)-1-nitrosourea, a chemotherapeutic used to combat various brain, head, and neck tumors. Previous studies of the toxic and mutagenic properties of the DNA adduct EA have been limited to in vitro experiments using mammalian polymerases and have revealed the lesion to be both miscoding and genotoxic. This work explores lesion bypass and mutagenicity of EA replicated in vivo and demonstrates that EA is neither toxic nor mutagenic in wild-type Escherichia coli. Although the base excision repair glycosylase enzymes of both humans and E. coli possess a weak ability to act on the lesion in vitro, an in vivo repair pathway has not yet been demonstrated. Here we show that an enzyme mechanistically unrelated to DNA glycosylases, the adaptive response protein AlkB, is capable of acting on EA via its canonical mechanism of oxidative dealkylation. The reaction alleviates the unrepaired adduct's potent toxicity through metabolism at the C8 position (attached to N1 of adenine), producing a nontoxic and weakly mutagenic N(6) adduct. AlkB is shown here to be a geno-protective agent that reduces the toxicity of DNA damage by converting the primary adduct to a less toxic secondary product.

Genotoxicity of glycidamide in comparison to (±)-anti-benzo[a]pyrene-7,8-dihydrodiol-9,10-epoxide and α-acetoxy-N-nitroso-diethanolamine in human blood and in mammalian V79-cells

Genotoxic activity of glycidamide (GA) was investigated in comparison to that of the known carcinogens (+/-)-anti-benzo[a]pyrene-7,8-dihydrodiol-9,10-epoxide ((+/-)-BPDE) and alpha-acetoxy-N-nitroso-diethanolamine (alpha-A-NDELA), using the hypoxanthine-phosphoribosyl-transferase (hPRT) gene mutation assay with V79 mammalian cells and modified alkaline single cell gel electrophoresis (alkaline comet assay with and without treatment of cells with formamido-pyrimidine-DNA-glycosylase (FPG)) in lymphocytes from human whole blood. As shown earlier, GA induced significant DNA damage in lymphocytes from treated whole blood at > or = 300 microM (4 h) (Baum et al., Mutat. Res. 2005, 580, 61-69). In the present study, using the alkaline comet assay with FPG treatment, increased formation of DNA strand breaks was observed in lymphocytes treated with GA (10 microM; 4 h). alpha-A-NDELA and (+/-)-BPDE were genotoxic at 10-30 microM (1 h). Genotoxic activity of these compounds was not enhanced after FPG treatment. FPG treatment thus offers an enhanced sensitivity of DNA damage detection for genotoxic compounds with preference for N(7)- resp. N(3)-purine alkylation. In the hPRT assay with V79 cells, mutagenic activity of (+/-)-BPDE became significant at > or = 3 microM (24 h). For alpha-A-NDELA significant activity was observed at greater, not dbl 10 microM (24 h). As previously observed, GA was considerably less effective, inducing significant mutagenicity roughly at about 80-300-fold higher concentrations (800 microM; 24 h) (Baum et al., Mutat. Res. 2005, 580, 61-69).

Predicting drug response and toxicity based on gene polymorphisms

The sequencing of the human genome has allowed the identification of thousands of gene polymorphisms, most often single nucleotide polymorphims (SNP), which may play an important role in the expression level and activity of the corresponding proteins. When these polymorphisms occur at the level of drug metabolising enzymes or transporters, the disposition of the drug may be altered and, consequently, its efficacy may be compromised or its toxicity enhanced. Polymorphisms can also occur at the level of proteins directly involved in drug action, either when the protein is the target of the drug or when the protein is involved in the repair of drug-induced lesions. There again, these polymorphisms may lead to alterations in drug efficacy and/or toxicity. The identification of functional polymorphisms in patients undergoing chemotherapy may help the clinician prescribe the optimal drug combination or schedule and predict with more accuracy the response to these prescriptions. We have recorded in this review the polymorphisms that have been identified up till now in genes involved in anticancer drug activity. Some of them appear especially important in predicting drug toxicity and should be determined in routine before drug administration; this is the case of the most common variations of thiopurine methyltransferase for 6-mercaptopurine and of dihydropyrimidine dehydrogenase for fluorouracil. Other appear determinant for drug response, such as the common SNPs found in glutathione S-transferase P1 or xereoderma pigmentosum group D enzyme for the activity of oxaliplatin. However, confusion factors may exist between the role of gene polymorphisms in cancer risk or overall prognosis and their role in drug response.

Differential inhibition of cellular glutathione reductase activity by isocyanates generated from the antitumor prodrugs Cloretazine™ and BCNU

The antitumor, DNA-alkylating agent 1,3-bis[2-chloroethyl]-2-nitrosourea (BCNU; Carmustine), which generates 2-chloroethyl isocyanate upon decomposition in situ, inhibits cellular glutathione reductase (GR; EC 1.8.1.7) activity by up to 90% at pharmacological doses. GR is susceptible to attack from exogenous electrophiles, particularly carbamoylation from alkyl isocyanates, rendering the enzyme unable to catalyze the reduction of oxidized glutathione. Evidence implicates inhibition of GR as a cause of the pulmonary toxicity often seen in high-dose BCNU-treated animals and human cancer patients. Herein we demonstrate that the prodrug Cloretazine (1,2-bis[methylsulfonyl]-1-[2-chloroethyl]-2-[(methylamino)carbonyl]hydrazine; VNP40101M), which yields methyl isocyanate and chloroethylating species upon activation, did not produce similar inhibition of cellular GR activity, despite BCNU and Cloretazine being equally potent inhibitors of purified human GR (IC(50) values of 55.5 microM and 54.6 microM, respectively). Human erythrocytes, following exposure to 50 microM BCNU for 1h at 37 degrees C, had an 84% decrease in GR activity, whereas 50 microM Cloretazine caused less than 1% inhibition under the same conditions. Similar results were found using L1210 murine leukemia cells. The disparity between these compounds remained when cells were lysed prior to drug exposure and were partially recapitulated using purified enzyme when 1mM reduced glutathione was included during the drug exposure. The superior antineoplastic potential of Cloretazine compared to BCNU in animal models could be attributed in part to the contribution of the methyl isocyanate, which is synergistic with the co-generated cytotoxic alkylating species, while at the same time unable to significantly inhibit cellular GR.

Reduction of nilutamide by NO synthases: implications for the adverse effects of this nitroaromatic antiandrogen drug

Nitric oxide synthases (NOSs) are flavohemeproteins that catalyze the oxidation of l-arginine to l-citrulline with formation of the widespread signal molecule NO. Beside their fundamental role in NO biosynthesis, these enzymes are also involved in the formation of reactive oxygen species and in the interactions with some xenobiotic compounds. Nilutamide is a nonsteroidal antiandrogen that behaves as a competitive antagonist of the androgen receptors and is proposed in the treatment of metastatic prostatic carcinoma. However, therapeutic effects of nilutamide are overshadowed by the occurrence of several adverse reactions mediated by toxic mechanism(s), which remain(s) poorly investigated. Here, we studied the interaction of NOSs with nilutamide. Our results show that the purified recombinant neuronal NOS reduced the nitroaromatic nilutamide to the corresponding hydroxylamine. The reduction of nilutamide catalyzed by neuronal NOS proceeded with intermediate formation of a nitro anion free radical easily observed by EPR, was insensitive to the addition of the usual heme ligands and l-arginine analogues, but strongly inhibited by O(2) and a flavin/NADPH binding inhibitor. Involvement of the reductase domain of nNOS in the reduction of nilutamide was confirmed by (i) the ability of the isolated reductase domain of nNOS to catalyze the reaction and (ii) the stimulating effect of Ca(2+)/calmodulin on the accumulation of hydroxylamine and nitro anion radical. In a similar manner, the recombinant inducible and endothelial NOS isoforms also displayed nitroreductase activity, albeit with lower yields. The selective reduction of nilutamide to its hydroxylamino derivative by the NOSs could explain some of the toxic effects of this drug.

Involvement of human polynucleotide kinase in double-strand break repair by non-homologous end joining

The efficient repair of double-strand breaks (DSBs) in DNA is critical for the maintenance of genome stability. In mammalian cells, repair can occur by homologous recombination or by non-homologous end joining (NHEJ). DNA breaks caused by reactive oxygen or ionizing radiation often contain non- conventional end groups that must be processed to restore the ligatable 3'-OH and 5'-phosphate moieties which are necessary for efficient repair by NHEJ. Here, using cell-free extracts that efficiently catalyse NHEJ in vitro, we show that human polynucleotide kinase (PNK) promotes phosphate replacement at damaged termini, but only within the context of the NHEJ apparatus. Phosphorylation of terminal 5'-OH groups by PNK was blocked by depletion of the NHEJ factor XRCC4, or by an inactivating mutation in DNA-PK(cs), indicating that the DNA kinase activity in the extract is coupled with active NHEJ processes. Moreover, we find that end-joining activity can be restored to PNK-depleted extracts by addition of human PNK, but not bacteriophage T4 PNK. This work provides the first demonstration of a direct, specific role for human PNK in DSB repair.

Relevance of DNA repair to carcinogenesis and cancer therapy

DNA-reactive carcinogens and anticancer drugs induce many structurally distinct cytotoxic and potentially mutagenic DNA lesions. The capability of normal and malignant cells to recognize and repair different DNA lesions is an important variable influencing the risk of mutation and cancer as well as therapy resistance. Using monoclonal antibody-based immunoanalytical assays, very low amounts of defined carcinogen-DNA adducts can be quantified in bulk genomic DNA, individual genes, and in the nuclear DNA of single cells. The kinetics of DNA repair can thus be measured in a lesion-, gene-, and cell type-specific manner, and the DNA repair profiles of malignant cells can be monitored in individual patients. Even structurally very similar DNa lesions may be repaired with extremely different efficiency. The miscoding DNA alkylation products O6-methylguanine (O6-MeGua) and O6-ethylguanine (O6-EtGua), for example, differ only by one CH2 group. These lesions are formed in DNA upon exposure to N-methyl-N-nitrosourea (MeNU) or N-ethyl-N-nitrosourea (EtNU), both of which induce mammary adenocarcinomas in female rats at high yield. Unrepaired O6-alkylguanines cause transition mutations via mispairing during DNA replication. O6-MeGua is repaired at a similar slow rate in transcribed (H-ras, beta-actin) and inactive genes (IgE heavy chain; bulk DNA) of the target mammary epithelia (which express the repair protein O6-alkylguanine-DNA alkyltransferase at a very low level). O6-EtGua, however, via an alkyltransferase-independent mechanism, is excised approximately 20 times faster than O6-MeGua from the transcribed genes selectively. Correspondingly, G:C-->A:T transitions arising from unrepaired O6-MeGua at the second nucleotide of codon 12 (GGA) of the H-ras gene are frequently found in MeNU-induced mammary tumors, but are absent in their EtNU-induced counterparts.

Protective effect of O6-methylguanine-DNA methyltransferase (MGMT) on the cytotoxic and recombinogenic activity of different antineoplastic drugs

The DNA repair protein O6-methylguanine-DNA methyltransferase (MGMT) removes alkyl groups from the O6 position of guanine in DNA and thus may protect cells against genotoxic effects of agents inducing this lesion. To analyze quantitatively the level of protection mediated by MGMT against antineoplastic drugs, we determined the cytotoxic and recombinogenic (sister-chromatid exchange inducing) effects of various chemotherapeutic agents in a pair of isogenic Chinese hamster cell lines deficient and proficient for MGMT, generated upon transfection with human MGMT cDNA. Furthermore, we compared the responses of the human cell lines HeLa MR (MGMT deficient) and HeLa S3 (MGMT proficient) to the various agents. It is shown that: (1) MGMT proficient cells are resistant in cell killing to the methylating drug streptozotocin and all the chloroethylating nitrosoureas tested. There was a marked agent specificity in protection. The level of resistance provoked by MGMT increased in the order BCNU < CCNU < ACNU < HeCNU < streptozotocin. (2) MGMT did not protect cells against killing induced by chlorambucil, cisplatin, melphalan, activated cyclophosphamide (malosfamide) and activated ifosfamide (4-hydroperoxy-ifosfamide). (3) MGMT caused protection against the recombinogenic effect of all nitrosoureas tested. The lowest level of protection was again observed for BCNU, followed by CCNU, ACNU < HeCNU < streptozotocin. (4) MGMT proficient cells did not exhibit resistance in SCE induction towards cyclophosphamide (activated by microsomes), 4-hydroperoxy-ifosfamide, mafosfamide, chlorambucil and melphalan. Some protection was afforded, however, against cisplatin (and transplatin). This effect was abolished by pretreatment of cells with O6-benzylguanine, which depletes MGMT, indicating that some lesion(s) induced by cisplatin giving rise to SCEs can be repaired by MGMT. Taken together, these results indicate that streptozotocin, HeCNU and ACNU are more selective than CCNU and BCNU in killing MGMT deficient cells, and that in the cases of cyclophosphamide, ifosfamide, chlorambucil, cisplatin and melphalan MGMT is not involved in mediating cytotoxic drug resistance.

In the search for new anticancer drugs. 28. Synthesis and evaluation of highly active aminoxyl labeled amino acid derivatives containing the [N'-(2-chloroethyl)-N'-nitrosoamino]carbonyl group

The aminoxyl (nitroxyl) labeled (2-chloroethyl)nitrosocarbamoyl (CNC) derivatives of amino acids, i.e., N-[[N'-(2-chloroethyl)-N'-nitrosoamino]carbonyl]-A-(1-oxy-2,2,6,6- tetramethylpiperidin-4-yl)amides, A = glycyl (10a), A = L-alanyl (10b), A = L-valyl (10c), A = L-phenylalanyl (10d), were synthesized and evaluated in vitro for their anticancer activities against the murine lymphocytic leukemia P388. Compounds 10a-d possessed activities ranging from 242 to 456% increase in life span (%ILS). All CDF1 male mice treated with the highly active compounds 10b and 10c at 12 mg/kg/day for 9 days were alive after 30 days. Compounds 10a-d were then tested in vivo against the murine lymphoid leukemia L1210. Compounds 10a-d exhibited, on day 60, a %ILS of 496, 663, 663, and 581, respectively. All CDF1 male mice treated with the highly active compounds 10b and 10c at 12 mg/kg/day for 9 days were alive after 60 days. The lipophilicities of compounds 10a-d were determined using the UV method. The %ILS parameters obtained against the P388 and L1210 tumor lines were correlated with the corresponding lipophilicities, and a trend was generally observed toward an increase in cytotoxicity with a concomitant decrease in hydrophobicity.

In the search for new anticancer drugs. XXIII: Exploration of a predictive design for anticancer drugs of carbohydrates containing N-nitrosochloroethylamino, N-nitrosomethyl, and N-nitrosoaminoxyl components

The spin-labeled glucose nitrosoureas 13-16, streptozotocin (18), chlorozotocin (31), streptozotocin analogues of galactosyl 24 and mannosyl 28, and their tetra-O-acetyl derivatives 25 and 29, MCNU (Cymerin, 34), and glucamine (21) were synthesized and evaluated in vivo for their anticancer activities against the murine lymphocytic leukemia P388. Compounds 13-16, 18, 24, 28, 31, and 34 possessed activities ranging from 33 to 603% increase in life span (%ILS), whereas 21, 25, and 29 were inactive (%ILS = 9 to 10). All CD2F1 male mice treated with the most active compounds (13, 14, and 34) at 20 mg/kg were alive after 30 days, whereas all mice treated with the clinical drug streptozotocin (18) and clinically tested chlorozotocin (31) succumbed. Compounds 13-16, 18, 31, and 34 were further evaluated for their antineoplastic activity against lymphoid leukemia L1210. Compounds 13 and 34 on day 60 exhibited %ILS values of 557 and 713, respectively, as compared with %ILS values of 646 and 713 for CCNU (1) and the spin-labeled SLCNU (3), respectively. The lipophilicities of 13-16, 18, 21, 24, 25, 28, 29, 31, and 34 were determined using EPR and/or UV methods. A predictive design pattern was observed, with the most active drug (34) possessing some hydrophobic property (log P = 1.24), followed by 13 (log P = 1.87) and 14 (log P = 1.81) as the more active drugs with higher hydrophobicity than 34. The clinical drugs streptozotocin (18) and chlorozotocin (31) were distinctly hydrophilic and less active. Finally, it was concluded that various scattered results of anticancer activity in the literature can be explained by a linear correlation of activities with lipophilicities.

Nucleophilic selectivity as a determinant of carcinogenic potency (TD50) in rodents: a comparison of mono- and bi-functional alkylating agents and vinyl chloride metabolites

Using published data, the carcinogenic potency (TD50) in rodents of a series of monofunctional alkylating agents, bifunctional antitumor drugs and the vinyl chloride (VC) metabolites chloroethylene oxide (CEO) and chloroacetaldehyde (CAA) was compared to their nucleophilic selectivity (Swain and Scott's constant s or initial ratio of 7-/O6-alkylguanine in DNA). A positive correlation between the log of TD50 estimates and the s values for a series of 14, mostly monofunctional, alkylating agents was observed. This linear relationship also included 2 bifunctional chloroethylnitrosoureas, although their carcinogenic potency was compared to their initial 7-/O6-alkylguanine ratio rather than their s values (n = 16, r = 0.91, p less than 0.005). In addition, the carcinogenic potency of 2 alkyl sulfates, which is not yet known accurately, may correlate with their nucleophilic selectivity through the same relationship. By contrast, 2 methyl halides and 5 bifunctional antitumor drugs (nitrogen mustards and azyridinyl derivatives) did not follow this linear relationship: at similar nucleophilic selectivity, they were more potent carcinogens than the above 18 alkylating agents; this may hold true for CEO and CAA too, although further carcinogenicity experiments are needed to calculate their precise TD50 values. The possible molecular mechanisms involved in tumor induction by these agents are discussed on the basis of these findings. Comparison of the estimated TD50 for CEO, CAA and VC in rodents confirms that CEO is the ultimate carcinogenic metabolite of VC and suggests that only a very small proportion of metabolically generated CEO is available for DNA alkylation in vivo.

Inhibition of the hepatic O6-alkylguanine-DNA alkyltransferase in vivo by pretreatment with antineoplastic agents

The mammalian DNA repair enzyme O6-alkylguanine-DNA alkyltransferase (AT) is inactivated during the repair process and its activity can only be restored by de novo synthesis. We have made use of this property to determine whether and to what extent various chemotherapeutic agents alkylate DNA in the O6-position of guanine, ie. produce lesions susceptible to AT repair. Adult female Fischer rats received a single i.p. injection of a high dose (LD50) of the respective agent and, 5 hr later, a chasing dose of N-nitroso-[14C]dimethylamine (0.2 mg/kg; 4 hr survival). The amount of 7-[14C]methylguanine formed was approximately 95 mumol/mol guanine and not significantly altered by pretreatment with any of the drugs. The ratio of O6-[14C]methylguanine/7-[14C]methylguanine was 0.019 for control animals, indicating that during the observation period of 4 hr, 83% of the O6-[14C]methylguanine produced had been removed by the hepatic AT. Little or no effect was found in rats that received spirohydantoin mustard, hexamethylmelamine, cis-platinum or mitomycin C. A significant increase in the O6-/7-[14C]methylguanine ratio was found after pretreatment with AZQ (0.026) and cyclophosphamide (0.028), agents for which lesions involving the O6-position of guanine have not yet been identified. N-(2-Hydroxyethyl)-N-nitrosourea and the cytostatic haloethylinitrosoureas, 1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU), 1-(2-chloroethyl)-3-(2,6-dioxo-3-piperidyl)-1-nitrosourea (PCNU), and N-chloroethyl-N-hydroxyethylnitrosourea (HECNU) inhibited the hepatic AT, producing a ratio of 0.025-0.035. Considerably higher ratios of 0.059 and 0.101 were observed after administration of the methylating agents procarbazine and 5-(3,3-dimethyl-1-triazeno)imidazole-4-carboxamide (DTIC), respectively. Complete saturation of the repair system (O6-/7-[14C]methylguanine ratio, 0.11) was only achieved with N-methyl-N-nitrosourea.

Nitrosoureas. Modes of action and perspectives in the use of hormone receptor affinity carrier molecules

Mechanisms of DNA adduct formation by antineoplastic 2-chloroethyl-N-nitrosoureas (CNUs) and of DNA damage induced by these compounds are discussed. CNUs are alkylating agents that form DNA-DNA cross-links as well as 2-chloroethylated and 2-hydroxyethylated adducts, the N-7-position of guanine being the predominantly alkylated site. A close correlation exists between the potential of a given compound to induce DNA-DNA cross-links and its antineoplastic effectiveness. However, levels of DNA-DNA cross-linking in bone marrow and extent of myelosuppression as measured in rodents are also closely correlated. The design of new cross-linking analogues capable of directing the antineoplastically relevant activity predominantly to the target tumour appears therefore to be of great promise. Cross-linking agents have been attached to a variety of steroid hormone carrier molecules and the conjugates have been tested in structure-activity studies using hormone-receptor containing animal tumours. These studies have revealed that some hormone-linked antineoplastic agents are highly effective in receptor positive experimental tumours and are superior to mixtures of unlinked alkylating agents with hormones. Indications for a relative enrichment of DNA damaging effects in the tumour tissue and for reduced myelotoxicity have been obtained with specific hormone conjugates.

Cytogenetic effects of N-nitrosodiethanolamine (NDELA) and NDELA-monoacetate in human lymphocytes

The mutagenicity of N-nitrosodiethanolamine (NDELA) and NDELA monoacetate was tested in vitro on lymphocytes of two healthy probands by determining the frequencies of chromosome aberrations, micronuclei and sister chromatid exchanges (SCE). A dose-dependent increase was found in all three test systems for NDELA as well as its monoacetate. The SCE test proved to be most sensitive for the genotoxic effect of NDELA because the differences to the control cultures had already become significant at 250-625 mumol/culture (26.6-65.4 mM). However, NDELA monoacetate showed a higher reactivity in the micronuclei and chromosome aberration test: significantly increased values were found even at 12.5 mumol (1.3 mM), whereas in the SCE test the differences became significant at the 25-mumol (2.7 mM) level. NDELA caused significantly increased rates of micronuclei and chromosome aberrations only at the highest test levels (625-1250 mumol; 65.4-127.6 mM). The results indicate important differences in the genotoxic effects of the two compounds, which might be explained by different lipophilicity and/or special activation processes.

Glutathione reductase inhibitors as potential antimalarial drugs. Effects of nitrosoureas on Plasmodium falciparum in vitro

Malarial parasites are believed to be more susceptible to oxidative stress than their hosts. BCNU(1,3-bis(2-chloroethyl)-1-nitrosourea) and HeCNU(1-(2-chloroethyl)-3-(2-hydroxythyl)-1-nitrosourea), inhibitors of the antioxidant enzyme glutathione reductase, were found to prevent the growth of Plasmodium falciparum in all intraerythrocytic stages. When exposing infected red blood cells to 38 microM BCNU or 62 microM HeCNU for one life cycle of synchronously growing parasites, the parasitemia decreased by 90%. During the formation of new ring forms, the parasites are even more susceptible to these drugs. The treatment with BCNU or HeCNU produced a rapid depletion of GSH in the parasites and their host cells; in addition, protection against lipid peroxidation was impaired in these cells. Possible mechanisms for the antimalarial action of the inhibitors are discussed. Our results suggest that erythrocyte glutathione reductase, an enzyme of known structure, might be considered as a target for the design of antimalarial drugs.

Glutathione reductase-deficient erythrocytes as host cells of malarial parasites

BCNU [1,3-bis(2-chloroethyl)-1-nitrosourea] and its less toxic derivative HeCNU [1-(2-chloroethyl)-3-(2-hydroxyethyl)-1-nitrosourea] are clinically-used antitumour drugs. In erythrocytes BCNU is a highly specific inhibitor of the enzyme glutathione reductase [H. Frischer and T. Ahmad, J. Lab. clin. Med. 89, 1080 (1977)]. When treating erythrocytes in vitro, 50% enzyme inhibition was obtained with 1 microM BCNU or 3 microM HeCNU within 2 hr. The two drugs were used for preparing red cell populations with various levels of glutathione reductase activity; complete inhibition (greater than or equal to 98%) was only achieved when the medium contained glucose as a source of reducing equivalents. The erythrocytes were then tested in drug-free media as host cells for the malaria parasite Plasmodium falciparum. In the range of 0-300 mU/ml cells, there was a correlation between glutathione reductase activity and parasite growth; erythrocytes with an activity of less than 20 mU/ml did not serve as host cells for P. falciparum at all although these erythrocytes were viable. When the culture medium was supplemented with 20 mM glutathione (GSH), parasite growth was normal irrespective of the glutathione reductase level in the erythrocytes. This is consistent with the finding that poisoning glutathione reductase led to a 10-fold decrease of the cytosolic GSH level. Our results corroborate the concept that intraerythrocytic inhibition of glutathione reductase mimicks the biochemistry of drug-sensitive glucose-6-phosphate dehydrogenase deficiency (favism), an inherited condition which confers protection from malaria.

Inhibition of human glutathione reductase by the nitrosourea drugs 1,3-bis(2-chloroethyl)-1-nitrosourea and 1-(2-chloroethyl)-3-(2-hydroxyethyl)-1-nitrosourea. A crystallographic analysis

Glutathione reductase from human erythrocytes was inhibited by incubation with the drugs 1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU) and 1-(2-chloroethyl)-3-(2-hydroxyethyl)-1-nitrosourea (HeCNU) under quasi-physiological conditions. For reference purposes, iodoacetamide was used for inactivating alkylation of the enzyme. In each case the modified glutathione reductase was crystallized and its structure determined. These analyses showed that in all experiments the enzyme had reacted at the distal sulfur, that is at the thiol of Cys-58, and virtually nowhere else in the visible structure. The electron density of the HeCNU derivative at 0.3 nm resolution is consistent with a 2-hydroxyethyl group. This alkyl moiety has recently been identified by chemical analysis [Schirmer, R. H., Schöllhammer, T., Eisenbrand, G. and Krauth-Siegel, R. L. (1987) Free Radical Res. Commun. 3, 3-12]. The 0.2 nm resolution electron-density map of the BCNU-derivatized enzyme cannot be explained by a 2-hydroxyethyl group. Instead the modification appears as a carbamoyl moiety containing at least five non-hydrogen atoms. In this derivative the distal cysteine is forced into an unusual conformation.

Rapid oxidative stress induced by N-nitrosamines

We have investigated the generation of prooxidant state shortly after administration of N-nitrosamines (NA) to rats. N-Nitrosodimethylamine (NDMA) was found to increase ethane exhalation (EE) rapidly in a dose-related manner. EE remained elevated for several days after single doses of NDMA. Similarly, lipid peroxidation (LP) in the liver (measured by four methods) increased rapidly showing a peak 20 min after NDMA dose. The increase of LP was preceded by a decrease in retinol concentration in the liver. N-Nitrosodiethanolamine, too, increased EE and LP in the liver, whereas N-nitrosomethylbenzylamine had no effect. Thus, hepatocarcinogenic NA induced LP in their target tissue, and the LP enhancing effects of NA were not related to their acute toxic effects.