Discovery of steroidal lactam conjugates of POPAM-NH2 with potent anticancer activity.

Aim: Steroidal prodrugs of nitrogen mustards such as estramustine and prednimustine have proven effective anticancer agents in clinical use since the 1970s. In this work, we aimed to develop steroidal prodrugs of the novel nitrogen mustard POPAM-NH2. POPAM-NH2 is a melphalan analogue that was coupled with three different steroidal lactams. Methodology: The new conjugates were preclinically tested for anticancer activity against nine human and one rodent cancer experimental models, in vitro and in vivo. Results & conclusion: All the steroidal alkylators showed high antitumor activity, in vitro and in vivo, in the experimental systems tested. Moreover, these hybrid compounds showed by far superior anticancer activity compared with the alkylating agents, melphalan and POPAM-NH2.

Prodrugs for Nitroreductase Based Cancer Therapy- 1: Metabolite Profile, Cell Cytotoxicity and Molecular Modeling Interactions of Nitro Benzamides with Ssap-NtrB.

BACKGROUND: Directed Enzyme Prodrugs Therapy (DEPT) as an alternative method against conventional cancer treatments, in which the non-toxic prodrugs is converted to highly cytotoxic derivative, has attracted an ample attention in recent years for cancer therapy studies. OBJECTIVE: The metabolite profile, cell cytotoxicity and molecular modeling interactions of a series of nitro benzamides with Ssap-NtrB were investigated in this study. METHOD: A series of nitro-substituted benzamide prodrugs (1-4) were synthesized and firstly investigated their enzymatic reduction by Ssap-NtrB (S. saprophyticus Nitroreductase B) using HPLC analysis. Resulting metabolites were analyzed by LC-MS/MS. Molecular docking studies were performed with the aim of investigating the relationship between nitro benzamide structures (prodrugs 1-4) and Ssap-NtrB at the molecular level. Cell viability assay was conducted on two cancer cell lines, hepatoma (Hep3B) and colon (HT-29) cancer models and healthy cell model HUVEC. Upon reduction of benzamide prodrugs by Ssap-NtrB, the corresponding amine effectors were tested in a cell line panel comprising PC-3, Hep3B and HUVEC cells and were compared with the established NTR substrates, CB1954 (an aziridinyl dinitrobenzamide). RESULTS: Cell viability assay resulted in while prodrugs 1, 2 and 3 had no remarkable cytotoxic effects, prodrug 4 showed the differential effect, showing moderate cytotoxicity with Hep3B and HUVEC. The metabolites that obtained from the reduction of nitro benzamide prodrugs (1-4) by Ssap-NtrB, showed differential cytotoxic effects, with none toxic for HUVEC cells, moderate toxic for Hep3B cells, but highly toxic for PC3 cells. CONCLUSION: Amongst all metabolites of prodrugs after Ssap-NtrB reduction, N-(2,4- dinitrophenyl)-4-nitrobenzamide (3) was efficient and toxic in PC3 cells as comparable as CB1954. Kinetic parameters, molecular docking and HPLC results also confirm that prodrug 3 is better for Ssap-NtrB than 1, 2 and 4 or known cancer prodrugs of CB1954 and SN23862, demonstrating that prodrug 3 is an efficient candidate for NTR based cancer therapy.

Chemical genetics analysis of an aniline mustard anticancer agent reveals complex I of the electron transport chain as a target.

The antitumor agent 11ß (CAS 865070-37-7), consisting of a DNA-damaging aniline mustard linked to an androgen receptor (AR) ligand, is known to form covalent DNA adducts and to induce apoptosis potently in AR-positive prostate cancer cells in vitro; it also strongly prevents growth of LNCaP xenografts in mice. The present study describes the unexpectedly strong activity of 11ß against the AR-negative HeLa cells, both in cell culture and tumor xenografts, and uncovers a new mechanism of action that likely explains this activity. Cellular fractionation experiments indicated that mitochondria are the major intracellular sink for 11ß; flow cytometry studies showed that 11ß exposure rapidly induced oxidative stress, mitochondria being an important source of reactive oxygen species (ROS). Additionally, 11ß inhibited oxygen consumption both in intact HeLa cells and in isolated mitochondria. Specifically, 11ß blocked uncoupled oxygen consumption when mitochondria were incubated with complex I substrates, but it had no effect on oxygen consumption driven by substrates acting downstream of complex I in the mitochondrial electron transport chain. Moreover, 11ß enhanced ROS generation in isolated mitochondria, suggesting that complex I inhibition is responsible for ROS production. At the cellular level, the presence of antioxidants (N-acetylcysteine or vitamin E) significantly reduced the toxicity of 11ß, implicating ROS production as an important contributor to cytotoxicity. Collectively, our findings establish complex I inhibition and ROS generation as a new mechanism of action for 11ß, which supplements conventional DNA adduct formation to promote cancer cell death.

Synthesis and biological activity of stable and potent antitumor agents, aniline nitrogen mustards linked to 9-anilinoacridines via a urea linkage.

To improve the chemical stability and therapeutic efficacy of N-mustard, a series of phenyl N-mustard linked to DNA-affinic 9-anilinoacridines and acridine via a urea linker were synthesized and evaluated for antitumor studies. The new N-mustard derivatives were prepared by the reaction of 4-bis(2-chloroethyl)aminophenyl isocyanate with a variety of 9-anilinoacridines or 9-aminoacridine. The antitumor studies revealed that these agents exhibited potent cytotoxicity in vitro without cross-resistance to taxol or vinblastine and showed potent antitumor therapeutic efficacy in nude mice against human tumor xenografts. It also showed that 24d was capable of inducing marked dose-dependent levels of DNA cross-linking by comet assay and has long half-life in rat plasma.

Significant differences in biological parameters between prodrugs cleavable by carboxypeptidase G2 that generate 3,5-difluoro-phenol and -aniline nitrogen mustards in gene-directed enzyme prodrug therapy systems.

Nine new nitrogen mustard compounds derived from 2,6-difluoro-4-hydroxy- (3a-e) and 2,6-difluoro-4-amino- (4a-d) aniline were synthesized as potential prodrugs. They were designed to be activated to their corresponding 3,5-difluorophenol and -aniline (4)-nitrogen mustards by the enzyme carboxypeptidase G2 (CPG2) in gene-directed enzyme prodrug therapy (GDEPT) models. The compounds were tested for cytotoxicity in the MDA MB-361 breast adenocarcinoma. The cell line was engineered to express stably either CPG2 tethered to the cell surface stCPG2-(Q)3 or beta-galactosidase (beta-Gal) as control. The cytotoxicity differentials were calculated between CPG 2-expressing and -nonexpressing cells and yielded different results for the two series of prodrugs despite their structural similarities. While the phenol compounds are ineffective as prodrugs, their aniline counterparts exhibit outstanding activity in the tumor cell lines expressing CPG2. [3,5-Difluoro-4-[bis(2-chloroethyl)amino]phenyl]carbamoyl-l-glutamic acid gave a differential of >227 in MDA MB361 cells as compared with 19 exhibited by 4-[(2-chloroethyl)(2-mesyloxyethyl)amino]benzoyl-l-glutamic acid, 1a, which has been in clinical trials.

2-Amino metabolites are key mediators of CB 1954 and SN 23862 bystander effects in nitroreductase GDEPT.

An important feature of gene-directed enzyme-prodrug therapy is that prodrug activation can provide diffusible cytotoxic metabolites capable of generating a local bystander effect in tumours. Activation of the aziridinyl dinitrobenzamide CB 1954 by E. coli nitroreductase (NTR) provides a bystander effect assumed to be due to the potently cytotoxic 4-hydroxylamine metabolite. We show that there are four cytotoxic extracellular metabolites of CB 1954 in cultures of NTR-expressing tumour cells (the 2- and 4-hydroxylamines and their corresponding amines). The 4-hydroxylamine is the most cytotoxic in DNA crosslink repair defective cells, but the 2-amino derivative (CB 10-236) is of similar potency to the 4-hydroxylamine in human tumour cell lines. Importantly, CB 10-236 has much superior diffusion properties to the 4-hydroxylamine in multicellular layers grown from the SiHa human cervical carcinoma cell line. These results suggest that the 2-amine, not the 4-hydroxylamine, is the major bystander metabolite when CB 1954 is activated by NTR in tumours. The corresponding dinitrobenzamide nitrogen mustard SN 23862 is reduced by NTR to form a single extracellular metabolite (also the 2-amine), which has superior cytotoxic potency and diffusion properties to the CB 1954 metabolites. These results are consistent with the reported high bystander efficiency of SN 23862 as an NTR prodrug in multicellular layers and tumour xenografts.

Quantitation of bystander effects in nitroreductase suicide gene therapy using three-dimensional cell cultures.

The efficacy of cancer gene therapy depends critically on "bystander effects" by which genetic modification of tumor cells results in killing of unmodified cells in the local microenvironment. In gene-dependent enzyme-prodrug therapy, expression of a prodrug-activating suicide gene is used to generate a cytotoxic metabolite that diffuses to nontransduced cells. The objective of this study was to develop a physiologically relevant tissue culture model for quantifying bystander effects and to validate the model using as an example the activation of dinitrobenzamide prodrugs (e.g., CB 1954) by Escherichia coli aerobic nitroreductase (NTR). Bystander effects were measured in three-dimensional multilayer cocultures of NTR+ and NTR- cells by determining clonogenic survival curves for both cell types using V79, Skov3, or WiDr as parental cell lines. Bystander killing by CB 1954 was much more efficient in multilayers than monolayers at equivalent cell:medium ratios, whereas the chloromustard analogue of CB 1954 showed even greater efficiency. For a series of dinitrobenzamides, bystander killing in multilayers showed a positive correlation with prodrug lipophilicity and also correlated with the bystander effect in mixed tumor xenografts grown from the same NTR+ and NTR- WiDr cell lines (r(2) = 0.84; P < 0.001). The multilayer model identified a bromomustard prodrug (SN 24927) with superior therapeutic activity to CB 1954 that provided curative activity against WiDr tumors comprising 1:1 mixtures of NTR+ and NTR- cells. This study demonstrates the utility of the multilayer tissue culture model for quantifying and optimizing bystander effects in tumors and identifies a new lead prodrug for NTR gene-dependent enzyme-prodrug therapy.

A rationally designed genotoxin that selectively destroys estrogen receptor-positive breast cancer cells.

We describe a novel strategy to increase the selective toxicity of genotoxic compounds. The strategy involves the synthesis of bifunctional molecules capable of forming DNA adducts that have high affinity for specific proteins in target cells. It is proposed that the association of such proteins with damaged sites in DNA can compromise protein function and/or DNA repair resulting in increased toxicity. We describe the synthesis of a bifunctional compound consisting of an aniline mustard linked to the 7alpha position of estradiol. This novel compound can form covalent DNA adducts that have high affinity for the estrogen receptor. Breast cancer cells that express high levels of the estrogen receptor showed increased sensitivity to the cytotoxic effects of the new compound.

Characterization of an antineoplastic glucuronide prodrug.

The specificity of tumor therapy may be improved by preferentially activating antineoplastic prodrugs at tumor cells pretargeted with antibody-enzyme conjugates. In this study, the conditions required for the efficient activation of p-hydroxyaniline mustard glucuronide (BHAMG) to p-hydroxyaniline mustard (pHAM) were investigated. pHAM induced cross-links in linearized double-stranded DNA at about 180-fold lower concentrations than BHAMG, indicating that the nucleophilicity of pHAM was decreased by the presence of a glucuronide group. The partition coefficient of BHAMG was about 1890 times lower than pHAM in an octanol-water two-phase system, suggesting that the reduced toxicity of BHAMG was due to both hindered diffusion across the lipid bilayer of cells and decreased reaction with nuclear DNA. BHAMG was significantly less toxic to BHK cells that expressed cytosolic Escherichia coli-derived beta-glucuronidase (betaG) compared with cells that were engineered to secrete betaG, demonstrating that extracellular localization of betaG was required for optimal activation of BHAMG. The extended retention of mAb RH1 on the surface of AS-30D cells was also consistent with extracellular activation of BHAMG. Taken together, our results indicate that the low toxicity of BHAMG was due to hindered cellular uptake and low alkylating activity. BHAMG must be enzymatically activated outside of tumor cells for maximum cytotoxicity, and non-internalizing antibodies are preferred for human tumor therapy by targeted antibody-enzyme activation of BHAMG.

Novel inhibitors of carboxypeptidase G2 (CPG2): potential use in antibody-directed enzyme prodrug therapy.

The design and synthesis of potent thiocarbamate inhibitors for carboxypeptidase G2 are described. The best thiocarbamate inhibitor N-(p-methoxybenzenethiocarbonyl)amino-L-glutamic acid 6d, chosen for preliminary investigations of in vitro antibody-directed enzyme prodrug therapy (ADEPT), abrogated the cytotoxicity of a combination of A5B7-carboxypeptidase G2 conjugate and prodrug PGP (N-p-{N,N-bis (2-chloroethyl)amino}phenoxycarbonyl-L-glutamate) toward LS174T cells. This is the first report of a small-molecule enzyme inhibitor proposed for use in conjunction with the ADEPT approach.

Macromolecular derivatives of N,N-di-(2-chloroethyl)-4-phenylene diamine mustard. 2. In vitro cytotoxicity and in vivo anticancer efficacy.

Prodrugs of N,N-di-(2-chloroethyl)-4-phenylene diamine (PDM) based on soluble poly[N5-(2-hydroxyethyl)-l-glutamine] (PHEG) have been evaluated as tumour-targeted drugs. These materials are designed to exploit the enhanced permeability of tumour vasculature, combining a passive tumour tropism with systemic liberation of free PDM. Modification of PDM by coupling via oligopeptide spacers onto a polymeric carrier significantly reduced its cytotoxicity towards different cell types in vitro. On the other hand, incubation of the cells with the PHEG-Gly-Phe-Ala-Leu-PDM conjugate in the presence of collagenase IV led to the release of lethal amounts of free drug, resulting in higher cytotoxicity for this derivative. The PHEG-Gly-Phe-Ala-Leu-PDM conjugate, which is rapidly degraded by lysosomal and tumour-associated enzymes also showed a decreased systemic toxicity in vivo and could be administered at a dose of 8 mg PDM/kg body weight intravenously, compared with just 2 mg/kg for free PDM. Furthermore, this derivative also showed better antitumour activity against a C26 colorectal carcinoma tumour model, compared with no activity for the free drug. The results indicate that the PHEG-Gly-Phe-Ala-Leu-PDM conjugate is a promising candidate for cancer treatment.

Synthesis and hypoxia-selective cytotoxicity of a 2-nitroimidazole mustard.

A four-step synthesis of 5-[N,N-bis(2-chloroethyl)amino]-1-methyl-2-nitroimidazole from 1-methyl-2-nitroimidazole is described. This compound showed similar hypoxia-selective cytotoxicity to the dinitrobenzamide mustard SN 23,862 in UV4 cells (ca. 40-fold), and superior selectivity (> 7-fold) in repair-competent AA8 cells.

Retinoic acid conjugates as potential antitumor agents: synthesis and biological activity of conjugates with Ara-A, Ara-C, 3(2H)-furanone, and aniline mustard moieties.

In a dual targeting approach, to explore the ability of tretinoin (all-trans-retinoic acid) to behave as a covalent carrier for cytotoxic entities, conjugates of retinoic acid with a few representative molecules, being important examples of antitumor pharmacophores (i.e., nucleoside analogues and alkylating agents), have been synthesized and tested for their cytostatic and differentiating activity. All compounds were stable to in vitro hydrolysis in human plasma and more lipophilic than the parent compounds, thus consenting enhanced uptake into the cells. Among the nucleoside analogues the Ara-C derivatives 3 and 6 and the Ara-A derivative 7 proved the most cytostatic (IC50 < 0.32 microgram/mL) resulting from 25- to > 144-fold more active (Ara-A derivatives) or at least as equally active (Ara-C derivatives) as compared to the parent nucleosides. Compound 3, endowed with a highly lipophilic silyl moiety at the 3' and 5' positions, showed the highest differentiating activity (54% and 44% differentiated HL-60 cells at 0.2 and 0.05 microgram/mL respectively). With regard to the retinoic acid conjugates of alkylating agents, compound 10 was the most cytostatic agent (IC50 < 0.32 microgram/mL) and the most potent differentiating agent (33-34% at 0.32 and 0.08 microgram/mL). These structures may also be regarded as analogs of either retinoic acid or the cytotoxic compound.

Aniline mustard analogues of the DNA-intercalating agent amsacrine: DNA interaction and biological activity.

Two series of analogues of the clinical antileukemic drug and DNA-intercalating ligand amsacrine have been prepared, containing aniline mustard sidechains of varying reactivity, linked either at the 4-position of the intercalating acridine chromophore (type A) or at the 1'-position of the 9-anilino group (type B). DNase I footprinting assays showed that compounds of type B had stronger reversible binding to DNA than did compounds of type A. Compounds of each type showed similar patterns of alkylation-induced cleavage of DNA, and alkylate at the N7 of guanines in runs of guanines (similar to the pattern for untargeted mustards) as well as some adenines. Both classes of compounds crosslinked DNA, although those bearing relatively inactive mustards did so only at high drug/base pair ratios. However, while the patterns of DNA alkylation were broadly similar, the compounds were considerably more cytotoxic than analogous untargeted mustards. Comparison of their cytotoxicities in wild-type and DNA repair-deficient lines indicated this toxicity was due to DNA crosslinks (except for the least reactive SO2-linked mustards). The 4-linked analogues showed slightly higher in vivo antileukemic activity than the corresponding 1'-linked analogues.

Hypoxia-selective antitumor agents. 14. Synthesis and hypoxic cell cytotoxicity of regioisomers of the hypoxia-selective cytotoxin 5-[N,N-bis(2-chloroethyl)amino]-2,4-dinitrobenzamide.

A series of regioisomers of the novel hypoxia-selective cytotoxin (HSC) 5-[N,N-bis(2-chloroethyl)-amino]-2,4-dinitrobenzamide (2a) have been prepared by displacement of the chloro group from methyl chlorodinitrobenzoates or the corresponding carboxamides with diethanolamine, followed by dimesylation and mesylate displacement with LiCl. The compounds fall into two classes, where the two nitro groups have either a meta or an ortho (or para) disposition to each other. The four meta derivatives had one-electron reduction potentials in the range -340 to -375 mV, similar to that of the known isomer 2a, while the other isomers had much higher values (-262 to -285 mV). The meta derivatives were much less cytotoxic to AA8 cells under aerobic conditions (IC50s from 75 to 470 microM) than were the other compounds (IC50s from 1.6 to 20 microM). However, the ratios of IC50s of the compounds in repair-proficient (AA8) and repair-deficient (UV4) cell lines varied, indicating differing contributions of DNA alkylation to aerobic toxicity between the isomers, with no clear relationship between this and nitro group disposition. The hypoxic selectivities of the (dimethylamino)ethylcarboxamide analogues for each isomer were determined by clonogenic assay against both AA8 and UV4 cells. With one exception, the meta derivatives showed excellent hypoxic selectivities (ca. 45-115-fold) against UV4 cells, while the ortho or para isomers had little selectivity (ca. 2-7-fold). A possible reason may be that the latter compounds, with higher reduction potentials, undergo rapid bioreduction even under aerobic conditions. None showed hypoxic selectivities greater than 2-3-fold against AA8 cells. The 3-[N,N-bis(2-chloroethyl)amino]-2,6-dinitrobenzamide isomer (5b), which showed the highest hypoxic selectivity for UV4 cells in this series, was active against both hypoxic and aerobic cells in KHT tumors in mice at well-tolerated doses, and showed superior in vivo activity to the previously studied 2,4-dinitro isomer 2b.

New mustard prodrugs for antibody-directed enzyme prodrug therapy: alternatives to the amide link.

Antibody-directed enzyme prodrug therapy (ADEPT) is a two-step approach for the treatment of cancer which seeks to generate a potent cytotoxic agent selectively at a tumor site. In this work described the cytotoxic agent is generated by the action of an enzyme CPG2 on a relatively nontoxic prodrug. The prodrug 1 currently on clinical trial is a benzamide and is cleaved by CPG2 to a benzoic acid mustard drug 1a. We have synthesized a series of new prodrugs 3-8 where the benzamide link has been replaced by, for example, carbamate or ureido. Some of these alternative links have been shown to be good substrates for CPG2 and therefore new candidates for ADEPT. The active drugs 3a and 4a derived from the best of these prodrugs are potent cytotoxic agents (1-2 microM) some 100 times more than 1a. The prodrugs 3 and 4 are some 100-200-fold less cytotoxic, in a proliferating cell assay, than their corresponding active drugs 3a and 4a.

Toward antibody-directed "abzyme" prodrug therapy, ADAPT: carbamate prodrug activation by a catalytic antibody and its in vitro application to human tumor cell killing.

Antibody-directed enzyme prodrug therapy, ADEPT, is a recent approach to targeted cancer chemotherapy intended to diminish the nonspecific toxicity associated with many commonly used chemotherapeutic agents. Most ADEPT systems incorporate a bacterial enzyme, and thus their potential is reduced because of the immunogenicity of that component of the conjugate. This limitation can be circumvented by the use of a catalytic antibody, which can be "humanized," in place of the bacterial enzyme catalyst. We have explored the scope of such antibody-directed "abzyme" prodrug therapy, ADAPT, to evaluate the potential for a repeatable targeted cancer chemotherapy. We report the production of a catalytic antibody that can hydrolyze the carbamate prodrug 4-[N,N-bis(2-chloroethyl)]aminophenyl-N-[(1S)-(1,3- dicarboxy)propyl]carbamate (1) to generate the corresponding cytotoxic nitrogen mustard (Km = 201 microM, kcat = 1.88 min-1). In vitro studies with this abzyme, EA11-D7, and prodrug 1 lead to a marked reduction in viability of cultured human colonic carcinoma (LoVo) cells relative to appropriate controls. In addition, we have found a good correlation between antibody catalysis as determined by this cytotoxicity assay in vitro and competitive binding studies of candidate abzymes to the truncated transition-state analogue ethyl 4-nitrophenylmethylphosphonate. This cell-kill assay heralds a general approach to direct and rapid screening of antibody libraries for catalysts.

Bioactivation of dinitrobenzamide mustards by an E. coli B nitroreductase.

A nitroreductase isolated and purified from Escherichia coli B has been demonstrated to have potential applications in ADEPT (antibody-directed enzyme prodrug therapy) by its ability in vitro to reduce dinitrobenzamides (e.g. 5-aziridinyl 2,4-dinitrobenzamide, CB 1954 and its bischloroethylamino analogue, SN 23862) to form cytotoxic derivatives. In contrast to CB 1954, in which either nitro group is reducible to the corresponding hydroxylamine, SN 23862 is reduced by the nitroreductase to form only the 2-hydroxylamine. This hydroxylamine can react with S-acetylthiocholine to form a species capable of producing interstrand crosslinks in naked DNA. In terms of ADEPT, SN 23862 has a potential advantage over CB 1954 in that it is not reduced by mammalian DT diaphorases. Therefore, a series of compounds related to SN 23862 has been synthesized, and evaluated as potential prodrugs both by determination of kinetic parameters and by ratio of IC50 against UV4 cells when incubated in the presence of prodrug, with and without the E. coli enzyme and cofactor (NADH). Results from the two studies were generally in good agreement in that compounds showing no increase in cytotoxicity in presence of enzyme and cofactor were not substrates for the enzyme. None of the analogues were activated by DT diaphorase isolated from Walker 256 carcinoma cells. For those compounds which were substrates for the E. coli nitroreductase, there was a positive correlation between kcat and IC50 ratio. Two compounds showed advantageous properties: SN 25261 (with a dihydroxypropylcarboxamide ring substituent) which has a more than 10-fold greater aqueous solubility than SN 23862 whilst retaining similar kinetic characteristics and cytotoxic potency; and SN 25084, where a change in the position of the carboxamide group relative to the mustard resulted in an increased cytotoxicity ratio and kcat compared with SN 23862 (IC50 ratios 214 and 135; kcat values of 75 and 26.4 sec-1, respectively). An analogue (SN 25507) incorporating both these structural changes had an enhanced kcat of 576 sec-1. This study elucidates some of the structural requirements of the enzyme and aids identification of further directions in the search for suitable prodrugs for an ADEPT nitroreductase system.

Reductive chemistry of the novel hypoxia-selective cytotoxin 5-[N,N-bis(2-chloroethyl)amino]-2,4-dinitrobenzamide.

5-[N,N-Bis(2-chloroethyl)amino]-2,4-dinitrobenzamide (1; SN 23862) is a novel bioreductive drug whose selective toxicity for hypoxic cells appears due to oxygen-inhibited enzymatic reduction of one of the nitro groups to the corresponding amine or hydroxylamine. Radiolytic reduction of 1 using up to four reducing equivalents in 1 N sodium formate was shown to proceed via electron addition to the 4-nitro group, thereby identifying this substituent as the most electron-affinic site in the molecule. The initially-formed 4-hydroxylamine and its N-hydroxytetrahydroquinoxaline half-mustard cyclization product (formed by intramolecular reaction with one arm of the adjacent mustard group) are reduced to the corresponding 4-amines upon further addition of electrons, although reduction of the 2-nitro group leading to 2,4-diamino products begins after addition of only six electron equivalents. Radiolytic reduction of the structurally similar 5-(aziridin-1-yl)-2,4-dinitrobenzamide (2; CB 1954) with six electron equivalents also occurs at the 4-nitro group to give the 4-hydroxylamine and 4-amine. The product mixture from reduction of 2 is less complex, largely because the corresponding 4-hydroxylamine and 4-amine are stable. The major reduction products of 1 were chemically synthesized by unequivocal routes to provide authentic samples for identification of the products of radiolytic reduction and to allow determination of their cytotoxicities. The 2- and 4-amino derivatives of 1 are significantly more cytotoxic than the parent drug, although the toxicity of the 4-amine is moderated by its facile conversion to the corresponding less toxic tetrahydroquinoxaline half-mustard. Although the 2- and 4-hydroxylamino derivatives were prepared by chemical reduction of 1, their toxicity could not be evaluated because of their instability. The 4-hydroxylamine reacts intramolecularly with the 5-mustard group somewhat more rapidly than does the 4-amine, while the 2-hydroxylamine is converted into a 2,2'-azoxy dimer following aerial oxidation to the 2-nitroso derivative. The fully reduced 2,4-diamino derivative of 1 is 10-fold more cytotoxic again than the 2-amine and, surprisingly, does not undergo spontaneous intramolecular alkylation. This elucidation of the reduction chemistry of 1 will facilitate further investigations of the toxic products generated from this compound both by hypoxic tumor cells and by ADEPT enzymes.