Search

Duocarmycins are highly cytotoxic natural products that have potential for development into anticancer agents. Herein we describe proposed but previously unidentified NH analogues of the DNA-alkylating subunit and characterise these by solvolysis studies, NMR and computational modelling. These compounds are shown to be the exclusive intermediates in the solvolysis of their seco precursors and to possess very similar structural features to the widely studied O-based analogues, apart from an unusually high basicity. The measured pKa of 10.5 implies that the NH compounds are fully protonated under physiological conditions. Remarkably, their extremely high reactivity (calculated hydrolysis rate 10(8) times higher for protonated NH compared to the neutral O analogue) is still compatible with potent cytotoxicity, provided the active species is formed in the presence of cells. These surprising findings are of relevance to the design of duocarmycin-based tumour-selective therapies., (© 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim.)

Inhibitors of the aldo-keto reductase enzyme AKR1C3 are of interest as potential drugs for leukemia and hormone-related cancers. A series of non-carboxylate morpholino(phenylpiperazin-1-yl)methanones were prepared by palladium-catalysed coupling of substituted phenyl or pyridyl bromides with the known morpholino(piperazin-1-yl)methanone, and shown to be potent (IC50∼100nM) and very isoform-selective inhibitors of AKR1C3. Lipophilic electron-withdrawing substituents on the phenyl ring were positive for activity, as was an H-bond acceptor on the other terminal ring, and the ketone moiety (as a urea) was essential. These structure-activity relationships are consistent with an X-ray structure of a representative compound bound in the AKR1C3 active site, which showed H-bonding between the carbonyl oxygen of the drug and Tyr55 and His117 in the 'oxyanion hole' of the enzyme, with the piperazine bridging unit providing the correct twist to allow the terminal benzene ring to occupy the lipophilic pocket and align with Phe311., (Copyright © 2013 Elsevier Ltd. All rights reserved.)

A high-throughput screen identified 3-(3,4-dihydroisoquinolin-2(1H)-ylsulfonyl)benzoic acid as a novel, highly potent (low nM), and isoform-selective (1500-fold) inhibitor of aldo-keto reductase AKR1C3: a target of interest in both breast and prostate cancer. Crystal structure studies showed that the carboxylate group occupies the oxyanion hole in the enzyme, while the sulfonamide provides the correct twist to allow the dihydroisoquinoline to bind in an adjacent hydrophobic pocket. SAR studies around this lead showed that the positioning of the carboxylate was critical, although it could be substituted by acid isosteres and amides. Small substituents on the dihydroisoquinoline gave improvements in potency. A set of "reverse sulfonamides" showed a 12-fold preference for the R stereoisomer. The compounds showed good cellular potency, as measured by inhibition of AKR1C3 metabolism of a known dinitrobenzamide substrate, with a broad rank order between enzymic and cellular activity, but amide analogues were more effective than predicted by the cellular assay.

Drug lipophilicity is a vital physicochemical parameter that influences drug absorption, distribution, metabolism, excretion and toxicology. A comparative study of a homologous series based on a pharmaceutically active drug represents a powerful approach to the study of the effects of drug lipophilicity. We have developed a rapid and sensitive LC-MS/MS method suitable for such a homologous series and applied it to a series of DNA binding benzonaphthyridine-based antitumour drugs of differing lipophilicity. The method used a gradient elution with a run time of 7 min for simultaneous quantitation of five analogues in a pooled sample. Method validation was carried out in plasma (human and mouse) and mouse tissues (brain, heart, kidney, liver and lung). It had a limit of quantitation of 0.001 μmol/L and was linear (0.001-0.3 μmol/L) in all matrices with acceptable intra- and inter-assay precision and accuracy. This method allowed the pharmacokinetic parameters of these compounds in mice to be related to their lipophilicity as determined by their partition coefficient (LogD). Both the plasma CL (r=0.95; P=2×10⁻⁷) and V(ss) (r=0.95; P=2×10⁻⁷) exhibited a significant positive correlation with LogD values after intravenous bolus administration to mice. Consequently the plasma mean residence time for each of these five analogues decreased with increasing lipophilicity. There was also a significant positive correlation (r=0.91; P=2×10⁻⁷) between LogD values and the brain to plasma AUC ratio indicating the importance of lipophilicity in the distribution of these compounds into the brain tissue., (Copyright © 2012 Elsevier B.V. All rights reserved.)

A series of 3-substituted (5-nitro-2,3-dihydro-1H-benzo[e]indol-1-yl)methyl sulfonate (nitroCBI) prodrugs containing sulfonate leaving groups undergo hypoxia-selective metabolism to form potent DNA minor groove alkylating agents. They were evaluated (along with chloride leaving group analogs for comparison) for their cytotoxicity against cultures of SKOV3 and HT29 human tumor cell lines under both aerobic and hypoxic conditions. Sulfonates with neutral side chains (e.g., 5,6,7-trimethoxyindole; TMI) show consistently higher hypoxic cytotoxicity ratios (HCRs) (34-246) than the corresponding chloro analogs (2.8-3.1) in SKOV3 cells, but these trends do not hold for compounds with cationic or polar neutral side chains., (Copyright © 2011 Elsevier Ltd. All rights reserved.)

Purpose: PR-104, a bioreductive prodrug in clinical trial, is a phosphate ester which is rapidly metabolized to the corresponding alcohol PR-104A. This dinitrobenzamide mustard is activated by reduction to hydroxylamine (PR-104H) and amine (PR-104M) metabolites selectively in hypoxic cells, and also independently of hypoxia by aldo-keto reductase (AKR) 1C3 in some tumors. Here, we evaluate reductive metabolism of PR-104A in mice and its significance for host toxicity., Methods: The pharmacokinetics of PR-104, PR-104A and its reduced metabolites were investigated in plasma and tissues of mice (with and without SiHa or H460 tumor xenografts) and effects of potential oxidoreductase inhibitors were evaluated., Results: Pharmacokinetic studies identified extensive non-tumor reduction of PR-104A to the 5-amine PR-104H (identity of which was confirmed by chemical synthesis), especially in liver. However, high concentrations of PR-104H in tumors that suggested intra-tumor activation is also significant. The tissue distribution of PR-104M/H was broadly consistent with the target organ toxicities of PR-104 (bone marrow, intestines and liver). Surprisingly, hepatic nitroreduction was not enhanced when the liver was made more hypoxic by hepatic artery ligation or breathing of 10% oxygen. A screen of non-steroidal anti-inflammatory drugs identified naproxen as an effective AKR1C3 inhibitor in human tumor cell cultures and xenografts, suggesting its potential use to ameliorate PR-104 toxicity in patients. However, neither naproxen nor the pan-CYP inhibitor 1-aminobenzotriazole inhibited normal tissue reduction of PR-104A in mice., Conclusions: PR-104 is extensively reduced in mouse tissues, apparently via oxygen-independent two-electron reduction, with a tissue distribution that broadly reflects toxicity.

Nitro seco-1,2,9,9a-tetrahydrocyclopropa[c]benz[e]indol-4-ones (nitroCBIs) are a new class of prodrugs for antitumor therapy that undergo hypoxia-selective metabolism to form potent DNA minor groove alkylating agents. Although hindered by poor aqueous solubility, several examples have shown activity against hypoxic tumor cells in vivo. Here we investigate structural properties that influence hypoxic selectivity in vitro, and show that for high hypoxic selectivity nitroCBIs should combine an electron-withdrawing group of H-bond donor capacity on the A-ring, with a basic substituent on the minor groove-binding side chain. Substitution on the A-ring is compatible with the introduction of functionality that can improve water solubility., (Copyright (c) 2010 Elsevier Ltd. All rights reserved.)

PR-104 is the phosphate ester of a 3,5-dinitrobenzamide nitrogen mustard (PR-104A) that is reduced to active hydroxylamine and amine metabolites by reductases in tumors. In this study, we evaluate the excretion of [(3)H]PR-104 in mice and determine its metabolite profile in mice, rats, dogs, and humans after a single intravenous dose. Total radioactivity was rapidly and quantitatively excreted in mice, with cumulative excretion of 46% in urine and 50% in feces. The major urinary metabolites in mice were products from oxidative N-dealkylation and/or glutathione conjugation of the nitrogen mustard moiety, including subsequent mercapturic acid pathway metabolites. A similar metabolite profile was seen in mouse bile, mouse plasma, and rat urine and plasma. Dogs and humans also showed extensive thiol conjugation but little evidence of N-dealkylation. Humans, like rodents, showed appreciable reduced metabolites in plasma, but concentrations of the cytotoxic amine metabolite (PR-104M) were higher in mice than humans. The most conspicuous difference in metabolite profile was the much more extensive O-beta-glucuronidation of PR-104A in dogs and humans than in rodents. The structure of the O-beta-glucuronide (PR-104G) was confirmed by independent synthesis. Its urinary excretion was responsible for 13 +/- 2% of total dose in humans but only 0.8 +/- 0.1% in mice. Based on these metabolite profiles, biotransformation of PR-104 in rodents is markedly different from that in humans, suggesting that rodents may not be appropriate for modeling human biotransformation and toxicology of PR-104.

Nitrochloromethylbenzindolines (nitroCBIs) are a new class of hypoxia-activated prodrugs for antitumor therapy. The recently reported prototypes undergo hypoxia-selective metabolism to form potent DNA minor groove alkylating agents and are selectively toxic to some but not all hypoxic tumor cell lines. Here we report a series of 31 analogues that bear an extra electron-withdrawing substituent that serves to raise the one-electron reduction potential of the nitroCBI. We identify a subset of compounds, those with a basic side chain and sulfonamide or carboxamide substituent, that have consistently high hypoxic selectivity. The best of these, with a 7-sulfonamide substituent, displays hypoxic cytotoxicity ratios of 275 and 330 in Skov3 and HT29 human tumor cell lines, respectively. This compound (28) is efficiently and selectively metabolized to the corresponding aminoCBI, is selectively cytotoxic under hypoxia in all 11 cell lines examined, and demonstrates activity against hypoxic tumor cells in a human tumor xenograft in vivo.

PR-104 is a dinitrobenzamide mustard currently in clinical trial as a hypoxia-activated prodrug. Its major metabolite, PR-104A, is metabolized to the corresponding hydroxylamine (PR-104H) and amine (PR-104M), resulting in activation of the nitrogen mustard moiety. We characterize DNA damage responsible for cytotoxicity of PR-104A by comparing sensitivity of repair-defective hamster Chinese hamster ovary cell lines with their repair-competent counterparts. PR-104H showed a repair profile similar to the reference DNA cross-linking agents chlorambucil and mitomycin C, with marked hypersensitivity of XPF(-/-), ERCC1(-/-), and Rad51D(-/-) cells but not of XPD(-/-) or DNA-PK(CS)(-/-) cells. This pattern confirmed the expected dependence on the ERCC1-XPF endonuclease, implicated in unhooking DNA interstrand cross-links at blocked replication forks, and homologous recombination repair (HRR) in restarting collapsed forks. However, even under anoxia, the hypersensitivity of XPF(-/-), ERCC1(-/-), and Rad51D(-/-) cells to PR-104A itself was lower than for chlorambucil. To test whether this reflects inefficient PR-104A reduction, a soluble form of human NADPH:cytochrome P450 oxidoreductase was stably expressed in Rad51D(-/-) cells and their HRR-restored counterpart. This expression increased hypoxic metabolism of PR-104A to PR-104H and PR-104M as well as hypoxia-selective cytotoxicity of PR-104A and its dependence on HRR. We conclude that PR-104A cytotoxicity is primarily due to DNA interstrand cross-linking by its reduced metabolites, although under conditions of inefficient PR-104A reduction (low reductase expression or aerobic cells), a second mechanism contributes to cell killing. This study shows that hypoxia, reductase activity, and DNA interstrand cross-link repair proficiency are key variables that interact to determine PR-104A sensitivity.

The dinitrobenzamide mustards are a class of bioreductive nitro-aromatic anticancer prodrugs, of which a phosphorylated analog (PR-104) is currently in clinical development. They are bioactivated by tumor reductases to form DNA cross-linking cytotoxins. However, their biotransformation in normal tissues has not been examined. Here we report the aerobic in vitro metabolism of three N-(2 hydroxyethyl)-3,5-dinitrobenzamide 2-mustards and the corresponding nonmustard analog in human, mouse, rat, and dog hepatic S9 preparations. These compounds have a range of mustard structures (-N(CH(2)CH(2)X)(2) where X = H, Cl, Br, or OSO(2)Me). Four metabolic routes were identified: reduction of either nitro group, N-dealkylation of the mustard, plus O-acetylation, and O-glucuronidation of the hydroxyethyl side chain. Reduction of the nitro group ortho to the mustard resulted in intramolecular alkylation and is considered to be an inactivation pathway, whereas reduction of the nitro group para to the mustard generated potential DNA cross-linking cytotoxins. N-Dealkylation inactivated the mustard moiety but may result in the formation of toxic acetaldehyde derivatives. Increasing the size of the nitrogen mustard leaving group abrogated the ortho-nitroreduction and N-dealkylation routes and thereby improved overall metabolic stability but had little effect on aerobic para-nitroreduction. All four compounds underwent O-glucuronidation of the hydroxyethyl side chain and further studies to elucidate the relative importance of this pathway in vivo are in progress.

Purpose: Hypoxia is a characteristic of solid tumors and a potentially important therapeutic target. Here, we characterize the mechanism of action and preclinical antitumor activity of a novel hypoxia-activated prodrug, the 3,5-dinitrobenzamide nitrogen mustard PR-104, which has recently entered clinical trials., Experimental Design: Cytotoxicity in vitro was evaluated using 10 human tumor cell lines. SiHa cells were used to characterize metabolism under hypoxia, by liquid chromatography-mass spectrometry, and DNA damage by comet assay and gammaH2AX formation. Antitumor activity was evaluated in multiple xenograft models (PR-104 +/- radiation or chemotherapy) by clonogenic assay 18 h after treatment or by tumor growth delay., Results: The phosphate ester "pre-prodrug" PR-104 was well tolerated in mice and converted rapidly to the corresponding prodrug PR-104A. The cytotoxicity of PR-104A was increased 10- to 100-fold by hypoxia in vitro. Reduction to the major intracellular metabolite, hydroxylamine PR-104H, resulted in DNA cross-linking selectively under hypoxia. Reaction of PR-104H with chloride ion gave lipophilic cytotoxic metabolites potentially able to provide bystander effects. In tumor excision assays, PR-104 provided greater killing of hypoxic (radioresistant) and aerobic cells in xenografts (HT29, SiHa, and H460) than tirapazamine or conventional mustards at equivalent host toxicity. PR-104 showed single-agent activity in six of eight xenograft models and greater than additive antitumor activity in combination with drugs likely to spare hypoxic cells (gemcitabine with Panc-01 pancreatic tumors and docetaxel with 22RV1 prostate tumors)., Conclusions: PR-104 is a novel hypoxia-activated DNA cross-linking agent with marked activity against human tumor xenografts, both as monotherapy and combined with radiotherapy and chemotherapy.

A series of 2,4-dinitrobenzamide mustards were prepared from 5-chloro-2,4-dinitrobenzoic acid or the corresponding 5-dimesylate mustard as potential prodrugs for gene-directed enzyme prodrug therapy (GDEPT) with the E. coli nfsB nitroreductase (NTR). The compounds, including 32 new examples, were evaluated in four pairs of NTR+ve/-ve cell lines for selective cytotoxicity (IC50 and IC50 ratios), in multicellular layer (MCL) cultures for bystander effects, and for in vivo activity against tumors grown from stably NTR transfected EMT6 and WiDr cells in nude mice. Multivariate regression analysis of the IC50 results was undertaken using a partial least-squares projection to latent structures model. In NTR-ve lines, cytotoxicity correlated positively with logP, negatively with hydrogen bond acceptors (HA) and donors (HD) in the amide side chain, and positively with the reactivity of the less-reactive leaving group of the mustard function, likely reflecting toxicity due to DNA monoadducts. Potency and selectivity for NTR+ve lines was increased by logP and HD, decreased by HA, and was positively correlated with the leaving group efficiency of the more-reactive group, likely reflecting DNA crosslinking. NTR selectivity was greatest for asymmetric chloro/mesylate and bromo/mesylate mustards. Bystander effects in the MCL assay also correlated positively with logP and negatively with leaving group reactivity, presumably reflecting the transcellular diffusion/reaction properties of the activated metabolites. A total of 18 of 22 mustards showed equal or greater bystander efficiencies in MCLs than the aziridinylbenzamide CB 1954, which is currently in clinical trial for NTR-GDEPT. The dibromo and bromomesylate mustards were surprisingly well tolerated in mice. High MTD/IC50 (NTR+ve) ratios translated into curative activity of several compounds against NTR+ve tumors. A bromomesylate mustard showed superior activity against WiDr tumors grown from 1:9 mixtures of NTR+ve and NTR-ve cells, indicating a strong bystander effect in vivo.

The 5-aziridinyl-2,4-dinitrobenzamide CB 1954 is a substrate for the oxygen-insensitive nitroreductase (NTR) from E. coli and is in clinical trial in combination with NTR-armed adenoviral vectors in a GDEPT protocol; CB 1954 is also of interest for selective deletion of NTR-marked cells in normal tissues. Since little further drug development has been carried out around this lead, we report here the synthesis of more soluble variants and regioisomers and structure-activity relationship (SAR) studies. The compounds were primarily prepared from the corresponding chloro(di)nitroacids through amide side chain elaboration and subsequent aziridine formation. One-electron reduction potentials [E(1)], determined by pulse radiolysis, were around -400 mV, varying little for aziridinyldinitrobenzamide regioisomers. Cytotoxicity in a panel of NTR-transfected cell lines showed that in the CB 1954 series there was considerable tolerance of substituted CONHR side chains. The isomeric 2-aziridinyl-3,5-dinitrobenzamide was also selective toward NTR+ve lines but was approximately 10-fold less potent than CB 1954. Other regioisomers were too insoluble to evaluate. While CB 1954 gave both 2- and 4-hydroxylamine metabolites in NTR+ve cells, related analogues with substituted carboxamides gave only a single hydroxylamine metabolite possibly because the steric bulk in the side chain constrains binding within the active site. CB 1954 is also a substrate for the two-electron reductase DT-diaphorase, but all of the other aziridines (regioisomers and close analogues) were poorer substrates with resulting improved specificity for NTR. Bystander effects were determined in multicellular layer cocultures and showed that the more hydrophilic side chains resulted in a modest reduction in bystander killing efficiency. A limited number of analogues were tested for in vivo activity, using a single ip dose to CD-1 nude mice bearing WiDr-NTR(neo) tumors. The most active of the CB 1954 analogues was a diol derivative, which showed a substantial median tumor growth delay (59 days compared with >85 days for CB 1954) in WiDr xenografts comprising 50% NTR+ve cells. The diol is much more soluble and can be formulated in saline for administration. The results suggest there may be advantages with carefully selected analogues of CB 1954; the weaker bystander effect of its diol derivative may be an advantage in the selective cell ablation of NTR-tagged cells in normal tissues.

Twelve substituted 4-nitrobenzyl carbamate prodrugs of the 5-aminobenz[e]indoline class of DNA minor groove alkylating agents were prepared and tested as prodrugs for gene-directed enzyme prodrug therapy (GDEPT) using a two-electron nitroreductase (NTR) from E. coli B. The prodrugs and effectors were tested in a cell line panel comprising parental and transfected human (SKOV/Skov-NTR(neo), WiDr/WiDr-NTR(neo)), Chinese hamster (V79(puro)/V79-NTR(puro)), and murine (EMT6/EMT6-NTR(puro)) cell line pairs. In the human cell line pairs, several analogues bearing neutral methoxyethoxy-, 2-hydroxyethoxy-, or 3-hydroxypropoxy-substituted side chains were good substrates for NTR as measured by cytotoxicity ratios, with NTR-ve/NTR+ve ratios similar to the established NTR substrates CB1954 (an aziridinyl dinitrobenzamide) and the analogous bromomustard. Selectivity for NTR decreased with increasing side-chain size or the presence of a basic amine group. Low to modest selectivity was observed in the Chinese hamster-derived cell line pair; however, in the murine EMT6/EMT6-NTR(puro) cell line pair, the above hydroxyalkoxy analogues again showed significant selectivity for NTR. The activity of the 2-hydroxyethoxy analogue was evaluated against NTR-expressing EMT6 tumors comprising ca. 10% NTR+ve cells at the time of tumor treatment. A small decrease in NTR+ve cells was observed after treatment, with a lesser effect against NTR-ve target cells, but these effects were not statistically significant and were much less than for the dinitrobenzamides. These results suggest that useful GDEPT prodrugs based on the 4-nitrobenzyl carbamate and 5-aminobenz[e]indoline motifs may be developed if optimization of pharmacokinetics can be addressed.

5,6-Dimethylxanthenone-4-acetic acid (DMXAA) is a novel anticancer agent with a number of unique activities, and is in clinical trial. The current synthesis of DMXAA involves six steps, beginning with a heterogeneous reaction to form an isonitrosoacetanilide, and gives an overall yield of 11% from 2,3-dimethylaniline. We report an alternative synthesis of the key intermediate 3,4-dimethylanthranilic acid via nitration of 3,4-dimethylbenzoic acid and separation of the key desired isomer by ready crystallisation. This, together with improvements in the rest of the synthesis, provide a shorter and higher-yielding route to DMXAA (22% overall from 3,4-dimethylbenzoic acid).

A series of 5-amino-seco-CBI compounds, designed for use as effectors for prodrugs, were prepared to study structure-activity relationships for the cytotoxicity of side chain analogues. Compounds were prepared by coupling 1-(chloromethyl)-5-nitro-1, 2-dihydro-3H-benz[e]indole to appropriate carboxylic acids, followed by nitro group reduction, or by coupling suitable 5-amino-protected indolines to alpha,beta-unsaturated acids, followed by deblocking. These AT-specific DNA alkylating agents were evaluated for cytotoxicity in a series of tumor cell lines (AA8, UV4, EMT6, SKOV3). For those analogues bearing an indolecarbonyl side chain, the 5'-methoxy derivative was the most cytotoxic (IC(50) 1.3 nM in AA8 cells, 4 h exposure), comparable to that of the parent CBI-TMI (5', 6',7'-trimethoxyindole) derivative (IC(50) 0.46 nM in the above assay). A subset of solubilized derivatives bearing O(CH(2))(2)NMe(2) substituents were about 10-fold less potent. For compounds containing an acryloyl linker in the side chain, the 4'-methoxycinnamoyl derivative proved the most cytotoxic (IC(50) 0. 09 nM in the above assay). A number of these 5-amino-seco-CBI-TMI analogues (including the solubilized compounds) are of interest both as cytotoxins and as components of amine-based prodrugs designed for tumor-specific activation.

A series of acridine-substituted bis(acridine-4-carboxamides) linked by a (CH2)3N(Me)(CH2)3 chain have been prepared by reaction of the isolated imidazolides of the substituted acridine-4-carboxylic acids with N,N-bis(3-aminopropyl)methylamine. These dimeric analogues of the mixed topoisomerase I/II inhibitor N-[2-(dimethylamino)ethyl]acridine-4-carboxamide (DACA), currently in clinical trial, show superior potencies to the corresponding monomeric DACA analogues in a panel of cell lines, including wild-type (JLC) and mutant (JLA and JLD) forms of human Jurkat leukemia. The latter mutant lines are resistant to topoisomerase II targeted agents because of lower levels of the enzyme. Analogues with small substituents (e.g., Me, Cl) at the acridine 5-position were clearly superior, with IC50's as low as 2 nM against the Lewis lung carcinoma and 11 nM against JLC. Larger substituents at any position caused a steady decrease in potency, likely due to lowering of DNA binding affinity. A small series of analogues of the most potent bis(5-methylDACA) compound, with second substituents (Me and Cl) in the 1- or 8- position had broadly similar potencies to the 5-Me compound, indicating that, while the 1- and 8-substituents are acceptable, they add little to the enhancing effect of the 5-methyl group. All of the compounds were at least equitoxic (some up to 4-fold more cytotoxic) against the mutant Jurkat lines than in the wild-type, consistent with a relatively greater effect on topoisomerase I compared with topoisomerase II. The bis(5-methylDACA) compound was found to inhibit the action of purified topoisomerase I in a cell-free assay. Compounds were on average 10-fold less cytotoxic in an MCF7 breast cancer line overexpressing P-glycoprotein than in the wild-type line and showed some selectivity for colon tumor lines in the NCI human tumor cell line panel. Several analogues produced significant growth delays in the relatively refractory subcutaneous colon 38 tumor model in vivo at substantially lower doses than DACA. The bis(acridine-4-carboxamides) represent a new and interesting class of potent topoisomerase inhibitors.

A series of tricyclic aromatic carboxamides, and their corresponding dimeric analogues, were prepared and their growth-inhibitory properties were evaluated in a series of cell lines. The dimeric compounds were prepared by reaction of the appropriate acids with carbonyl-1,1'-diimidazole, isolating the resulting imidazolides, and reacting these with a stoichiometric amount of the diamine. The monomeric carboxamides containing a (CH2)2NMe2 side chain had widely differing inhibitory potencies, with the known nitronaphthalimide (mitonafide) and acridine-4-carboxamide (DACA) being the most potent. The corresponding bis analogues, linked by a (CH2)3NMe(CH2)3 chain, were generally more potent, with the largest increases (dimer/monomer ratio 20- to 30-fold) seen for the nitronaphthalimides and the phenazines. Based on the intrinsic cytotoxicity of the monomers and the highest degree of increase in cytotoxicity on dimerization, the most interesting chromophores appear to be the acridine-4-carboxamide and phenazine-1-carboxamide. Both of these compounds showed significant growth delays (approximately 6 days) in an in vivo colon 38 tumour model in mice.

A series of racemic 6-amino-seco-cyclopropylindole (seco-CI) compounds was prepared by coupling 1-(tert-butyloxycarbonyl)-3-(chloromethyl)-6-nitroindoline with appropriate acids, followed by nitro group reduction, and evaluated for cytotoxicity in AA8, UV4, EMT6, and SKOV3 cell lines. These compounds are of interest due to their close structural relationship to known AT-specific alkylating agents and cytotoxins and also for the possible construction of stable amine-based prodrugs designed for tumor-specific release. Variations included indole or furan side chains with different substituents, sulfonamide or carboxamide linkers, extension of the minor groove binding side chain to two subunits, and the use of a pyrroylacryloyl unit previously reported to give extremely potent analogues. The parent compound, with a trimethoxyindole side chain, was a moderately potent cytotoxin (IC50 = 0.34 microM in AA8 cells, 4 h exposure). A single 5-methoxy group on the indole minor groove binding unit was sufficient to maintain potency, and a series of dimethylaminoethoxy-substituted analogues retained the cytotoxicity of the parent compound, while providing increased aqueous solubility.

A series of N-dinitrophenylamino acid amides [(4-CONHZ-2, 6-diNO2Ph)N(R)C(X,Y)CONHPhOMe] were prepared as potential bioreductive prodrugs and reduced radiolytically to study their rates of subsequent intramolecular cyclization. Compounds bearing a free NH group (R = H) underwent rapid cyclization in neutral aqueous buffers (t1/2 < 1 min) following 4-electron reduction, with the generation of a N-hydroxydihydroquinoxalinone and concomitant release of 4-methoxyaniline. Amine release from analogous N-methyl analogues (R = Me) was relatively slow. These results are consistent with intramolecular cyclization of a monohydroxylamine intermediate. The high rates of cyclization/extrusion by these very electron-deficient hydroxylamines suggest that the process is greatly accelerated by the presence of an H-bonding "conformational lock" between the anilino NH group and the adjacent o-nitro group (Kirk and Cohen, 1972). Changes in the phenylcarboxamide side chain or in C-methylation in the linking chain had little effect on the rate of cyclization. The model compounds had 1-electron reduction potentials in the range appropriate for cellular reduction (-373 mV for a measured example) and appeared suitable for development as prodrugs that release amine-based effectors following enzymic or radiolytic reduction. Prodrug examples containing 4-aminoaniline mustard and 5-amino-1-(chloromethyl)benz[e]indoline alkylating units were evaluated but were not activated efficiently by cellular nitroreductases. However, cell killing by the radiation-induced reduction of the latter prodrug was demonstrated.

The action of the anticancer drug amsacrine appears to involve molecular interactions with both DNA and topoisomerase II. It has been shown previously that DNA intercalators can inhibit the action of amsacrine and several other topoisomerase II poisons, presumably as a result of interference with the DNA binding sites for the enzyme. We show here that drug molecules such as N-phenylmethanesulfonamide, which mimic the anilino side chain of amsacrine, inhibit the cytotoxicity against cultured Lewis lung murine carcinoma of amsacrine, amsacrine analogues including asulacrine and DACA (N-[2-(dimethylamino)-ethyl]acridine-4-carboxamide dihydrochloride), and etoposide. In contrast, the cytotoxicity of doxorubicin was slightly increased by co-incubation with N-phenylmethanesulfonamide. The cytotoxicity of amsacrine was also modulated in human Jurkat leukemia, HCT-8 colon, and HT-29 colon cell lines. Because o-AMSA, an amsacrine analogue containing a methoxy group in the ortho rather than in the meta position, is known to be inactive as an antitumor drug, the abilities of the ortho and meta methoxy-substituted derivatives of methyl-N-phenylcarbamate to reverse the cytotoxicity of amsacrine, asulacrine, and DACA were compared. The ortho substitution decreased activity while meta substitution slightly increased it, suggesting that the side chains were binding to a similar site to that occupied by amsacrine. To determine whether the side chain variants actively inhibited the formation of DNA-topoisomerase II covalent complexes, cultured cells were treated with amsacrine or asulacrine, harvested, and lysed directly on acrylamide gels before electrophoresis and Western blotting to identify non-DNA-bound topoisomerase II. Extractable topoisomerase II was depleted in cells incubated with amsacrine but partially restored by coculture with methyl-N-phenylcarbamate. The findings are consistent with the hypothesis that low molecular weight molecules can modulate the effects of topoisomerase II poisons by directly interacting with the enzyme.

A series of bis(hydroxymethyl)-substituted imidazoles, thioimidazoles, and pyrrolizines and related bis(carbamates), linked to either 9-anilinoacridine (intercalating) or 4-(4-quinolinylamino)benzamide (minor groove binding) carriers, were synthesized and evaluated for sequence-specific DNA alkylation and cytotoxicity. The imidazole and thioimidazole analogues were prepared by initial synthesis of [(4-aminophenyl)alkyl]imidazole-, thioimidazole-, or pyrrolizine dicarboxylates, coupling of these with the desired carrier, and reduction to give the required bis(hydroxymethyl) alkylating moiety. The pyrrolizines were the most reactive alkylators, followed by the thioimidazoles, while the imidazoles were unreactive. The pyrrolizines and some of the thioimidazoles cross-linked DNA, as measured by agarose gel electrophoresis. Strand cleavage assays showed that none of the compounds reacted at purine N7 or N3 sites in the gpt region of the plasmid gpt2Eco, but the polymerase stop assay showed patterns of G-alkylation in C-rich regions. The corresponding thioimidazole bis(carbamates) were more selective than the bis(hydroxymethyl) pyrrolizines, with high-intensity bands at 5'-NCCN, 5'-NGCN and 5'-NCGN sequences in the PCR stopping assay ( indicates block sites). The data suggest that these targeted compounds, like the known thioimidazole bis(carbamate) carmethizole, alkylate exclusively at guanine residues via the 2-amino group, with little or no alkylation at N3 and N7 guanine or adenine sites. The cytotoxicities of the compounds correlated broadly with their reactivities, with the bis(hydroxymethyl)imidazoles being the least cytotoxic (IC50s >1 microM; P388 leukemia) and with the intercalator-linked analogues being more cytotoxic than the corresponding minor-groove-targeted ones. This was true also for the more reactive thioimidazole bis(carbamates) (IC50s 0.8 and 11 microM, respectively), but both were more active than the analogous "untargeted" carmethizole (IC50 20 microM). The bis(hydroxymethyl)pyrrolizine analogues were the most cytotoxic, with IC50s as low as 0.03 microM.

The mixed topoisomerase I/II inhibitor N-[2-(dimethylamino)ethyl]acridine-4-carboxamide (DACA) is currently in clinical trial as an anticancer drug. A series of acridine-substituted analogues were prepared, using a new synthetic route to substituted acridine-4-carboxylic acids (conversion of substituted diphenylamine diacid monoesters to the corresponding aldehydes and mild acid-catalyzed ring closure to form the acridines directly). The analogues were evaluated in a panel of cell lines which included wild-type (JLC) and mutant (JLA and JLD) forms of the human Jurkat leukemia line. The latter mutant lines are resistant to topoisomerase II targeted agents due to lower levels of the enzyme. Structure-activity studies suggest that the electronic properties of the substituents do not markedly affect cytotoxicity, but steric bulk is important, with larger groups leading to loss of activity. The compounds fell broadly into two categories. The majority had cytotoxicities similar to (or lower than) that of DACA itself and were equitoxic in all the Jurkat lines, suggesting a relatively greater effect on topoisomerase I compared with topoisomerase II. Most of the 5-substituted derivatives and the 7-Ph compound were more cytotoxic than DACA, but were less effective against JLA and JLD cell lines than in the wild-type JLC, suggesting a mode of cytotoxicity largely mediated by effects on topoisomerase II. Both DACA and selected acridine-substituted analogues were active in the relatively refractory subcutaneous colon 38 tumor model in vivo.

Studies have shown that 4-(alkylamino)-5-nitroquinolines possess high selectivity (20-60-fold) for hypoxic tumor cells in vitro, but are not active as hypoxia-selective cytotoxins (HSCs) in vivo. The compounds show inadequate rates of extravascular diffusion, likely due both to sequestration of the bisbasic compounds into lysosomes and rapid nitroreduction. A further series of analogues, designed to counteract these limitations, has been synthesized and evaluated. Analogues bearing one to three electron-donating substituents on the quinoline have one-electron reduction potentials up to 100 mV lower than that of the unsubstituted compound (5), but do not have improved biological activity. The relationship between hypoxic selectivity and rates of metabolic reduction suggests at least two mechanisms of cytotoxicity for this series of 5-nitroquinolines. Compounds with high rates of reduction are toxic via oxygen-sensitive net bioreduction, while compounds which are poor substrates for nitroreduction are toxic through an oxygen-insensitive non-bioreductive mechanism. As rates of metabolic reduction are lowered, the non-bioreductive mechanism of toxicity becomes dominant and hypoxic selectivity is lost. A small series of analogues bearing hydrophilic but neutral side chains were also prepared. Compounds with a dihydroxypropyl side chain retained cytotoxic potency and hypoxic cell selectivity in cell culture assays, and had lowered uptake into lysosomes, but none of three analogues evaluated against KHT tumors in mice showed activity as an HSC in vivo.

2,4-Dinitrobenzamide mustards, exemplified by the parent compound SN 23862 (2) are activated under aerobic conditions by an Escherichia coli nitroreductase enzyme (NR2) via selective reduction of the 2-nitro group, and are thus of interest as prodrugs for antibody-directed enzyme-prodrug therapy (ADEPT). A series of related compounds 12a-12d, where the 4-nitro group of 2 was replaced by other substituents of varying electronic properties, were prepared and evaluated as potential ADEPT prodrugs. One-electron reduction potentials of the compounds correlated well with the substituent sigma m values, with the exception of the unsubstituted (4-H) analogue 13, which had a much lower value than expected on electronic grounds, due to a coplanar conformation of the mustard. The cytotoxicities of the compounds towards aerobic UV4 cells correlated positively with the electron-donating ability of the 4-substituent (measured by sigma p values), indicating that the cytotoxicities of the compounds in the absence of the NR2 enzyme are due substantially to the parent (unreduced) compounds. A positive, although less strong, correlation was seen between the electronic properties of the 4-substituent and their cytotoxicities in the presence of the NR2 enzyme, suggesting that, in this closely related series, the degree of activation by the enzyme is significantly dependent on the reduction potential of the 2-nitro group. While the 4-SO2Me derivative 12d was the next most preferred substrate after the parent 2, it was considerably less so (degree of activation as measured by IC50 ratio of 26 compared with 145), despite the similar electronic properties of the two 4-substituents.

A series of polybenzamide DNA minor groove binding ligands bearing either one or two monofunctional mustards have been synthesised, and their cytotoxicities and interactions with DNA have been studied. Analogues with two alkylating functions (e.g. compounds 7 and 14) are the most cytotoxic, with 7 being 1000-fold more potent than the clinical mustard chlorambucil against P388 leukemia in culture, as well as being more potent in vivo. Monofunctional analogues were also significantly more cytotoxic than chlorambucil, despite bearing much less reactive mustard species. These results support the concept that targeting nitrogen mustard alkylating agents to DNA by attachment to DNA-affinic carriers can greatly enhance cytotoxicity due to alkylation, and that even for such DNA-targeted mustards, crosslinking is a more toxic event than monoalkylation. Close analogues of 7 differing only in their radius of curvature, appear to alkylate and crosslink DNA in similar fashion, yet have widely differing cytotoxicities. The most cytotoxic compound (7) possesses a geometry most complementary to that of duplex DNA, suggesting that the most toxic lesions are those which result in least DNA distortion, thus being less easily recognised by DNA repair systems.

The 4-alkylamino-5-nitroquinolines (5NQs) are a new series of bioreductive drugs that exhibit varying degrees of selective toxicity (up to 60-fold) under hypoxic conditions in cell culture. This study tested the hypothesis that differences in hypoxia-selective cytotoxicity in this series reflect differences in the efficiency with which oxygen inhibits metabolic reduction. The products of reduction of six 5NQs were characterized and rates of reduction compared in aerobic and hypoxic AA8 cells. The major stable products of both radiolytic and metabolic reduction under anoxic conditions were the corresponding amines, which were not responsible for the toxicity of the parent nitro compounds. Metabolism of each compound was inhibited completely in aerobic cells, indicating that differences in hypoxia-selective toxicity in this series are not due to variations in efficiency as substrates for oxygen-insensitive nitro reduction. Rates of hypoxic metabolism correlated broadly with hypoxia-selective cytotoxicity; the 5NQ derivatives with high rates of hypoxic metabolism had good hypoxia-selective cytotoxicity, whereas the compounds with low rates of reduction (the 3,6-dimethyl and 8-methylamino compounds; 3,6diMe-5NQ and 8NHMe-5NQ) were non-selective. Low rates of drug-induced oxygen consumption by 3,6-diMe-5NQ and 8NHMe-5NQ in respiration-inhibited cells confirmed that these compounds are poor substrates for enzymatic nitro reduction. While there was an overall correlation between one-electron reduction potential at pH 7 (E1(7)) and rate of metabolic reduction, the relatively high E1(7) of 3,6diMe-5NQ (-367 mV) indicates that rates of reduction, and hypoxic selectivity of cytotoxicity, cannot be predicted from reduction potential alone. 3,6diMe-5NQ and 8NHMe-5NQ are cytotoxic through a non-bioreductive mechanism, the variable contribution of which may underlie the differences in hypoxia-selective cytotoxicity within this series of bioreductive drugs.

A series of analogues of the novel hypoxia-selective cytotoxin 5-[N,N-bis(2-chloroethyl)amino]-2,4-dinitrobenzamide (6) have been prepared and evaluated, in a search for compounds which retain high hypoxic selectivity but have increased potency and/or aqueous solubility. Several analogues with ionizable or dipolar carboxamide side chains showed improved solubility but generally had reduced cytotoxic potency and hypoxic selectivity. Modification of the mustard leaving groups or replacement of the carboxamide moiety provided some compounds with superior potency, but only the mixed chloro/mesylate mustard 20 provided a gain in potency relative to solubility while retaining the hypoxic selectivity of 6. These nitrogen mustards did not show the remarkable activity demonstrated by the related aziridine 7 [CB 1954, 5-(N-aziridinyl)- 2,4-dinitrobenzamide] in Walker 256 adenocarcinoma cells and are not efficient substrates for the DT-diaphorase which activates the latter compound by aerobic nitroreduction in Walker cells. Variations in hypoxic selectivity within the dinitrobenzamide mustards appear not to be due to differences in sensitivity to activation by this enzyme. Walker cells showed intermediate sensitivity to the mono(2-chloroethyl) analogue 26 but not to the related half-mustard 27, suggesting that the inhibition of DT-diaphorase activity is due to steric effects in the 5-position. The preferred compound overall with respect to solubility, potency, and in vitro hypoxic cell selectivity was the (dimethylamino)-ethyl derivative 11. DNA elution studies and comparison of the sensitivity of AA8 and UV4 cells to this compound indicated reductive activation to form a DNA cross-linking agent under hypoxia. Radiobiological studies indicated 11 to be equally active against both aerobic and hypoxic cells in KHT tumors. It is not clear whether this reflects efficient killing of aerobic cells as a result of diffusion of reduced metabolites from hypoxic regions or whether cytotoxicity in tumors is independent of hypoxia.

2-Nitroaryl amides of general structure I are proposed as bioreducible prodrugs, capable of releasing cytotoxic aminoaniline mustards V on bioactivation by spontaneous cyclization of the resulting 2-aminoarylamides II via a tetrahedral intermediate, III. This concept allows separate optimization of the substituent effects influencing nitro-group reduction and mustard reactivity. A series of model 2-aminoaryl amides has been synthesized, and their rates of cyclization have been studied; these varied by a factor of more than 50,000-fold (kobs from 0.00040 to 21 min-1) at pH 2.4. For three compounds studied in detail, the rates were linearly dependent of pH, indicating that no change in the mechanism of the rate-determining step occurs over the pH range studied. The nucleophilicity of the amino group had a modest influence on the kinetics of cyclization, with electron-withdrawing groups slowing the rate. The geometry of the compound was also important, with structure-activity relationships indicating that the rate of cyclization is greatly enhanced by the preorganization of the molecule. In contrast, 4-substitution on the leaving aniline by a variety of groups had little effect on the cyclization reaction. These results are consistent with the rate-determining step being formation of the tetrahedral intermediate. These model studies suggest that the phenyldimethylacetamide system could be developed as a prodrug system for the bioreductively-triggered release of amines. Further substantial rate enhancements appear possible by alterations in the geometry of the system, whereas substitution of electron-withdrawing groups (required to raise the nitro-group reduction potential into the appropriate range) has only relatively modest retardation effects on rates of cyclization. More rigid systems may also be useful; a nitronaphthaleneacetamide analogue cyclized spontaneously during nitro-group reduction, suggesting a very short half-life for the reduced intermediate (amine or hydroxylamine).

A series of isomeric 4-[[3-(dimethylamino)propyl]amino]nitroquinolines has been synthesized and evaluated as hypoxia-selective cytotoxins and as radiosensitizers of hypoxic cells. The compounds showed widely-differing hypersensitivity factors (ratios of cytotoxicity against wild-type and repair-deficient mammalian cells). Many compounds showed oxygen-sensitive bioreduction resulting in DNA alkylation, while others show oxygen-insensitive modes of action. Of the nitro isomers studied, the 5-nitro showed the greatest hypoxic selectivity. A series of ring-substituted analogues were then prepared, in an effort to lower its reduction potential of -286 mV. Structure-activity studies showed that the effects of substitution on reduction potential were complex, being mediated by electronic and steric effects on the nitro group, as well as by effects on quinoline pKa. Two compounds of lower reduction potential, the 3- and 8-methyl analogues, showed improved selectivity (47- and 60-fold in a clonogenic assay). These two compounds also showed the highest "in vitro therapeutic indices" of the series as hypoxic cell radiosensitizers. Despite these favorable in vitro properties, neither compound had activity against hypoxic cells in SCCVII tumors when administered at 60% of the MTD.

A series of 9-oxo-9,10-dihydroacridine-4-acetic acids (acridone-4-acetic acids) were prepared by Jourdan-Ullmann condensation of 2-halobenzoic acids with 2-aminophenylacetic acids, followed by H2SO4-induced cyclodehydration of the resulting 2-[2-(carboxymethyl)phenylamino]benzoic acids. These were evaluated for their ability to induce haemorrhagic necrosis in transplanted colon 38 tumours in mice, using a short-term histology assay. The results broadly paralleled those seen previously for xanthenone-4-acetic acids, with 1-, 2- and 7-substitution being dystherapeutic, and substitution at the 5- and 6-positions by lipophilic groups increasing activity. While some analogues were as active as xanthenone-4-acetic acids in the histology assay and gave significant growth delays against colon 38 tumours in vivo, as a class the 9-oxo-9,10-dihydroacridine-4-acetic acids were generally less potent than the xanthenone-4-acetic acids.

A series of 16 analogues of the solid tumor active compound 9-oxo-9H-xanthene-4-acetic acid (XAA), with variations in the acetic acid side chain, have been prepared and evaluated for their ability to cause early haemorrhagic necrosis of colon 38 tumors in mice. The results extend the previous SAR for this class and confirm the necessity for a carboxylic acid group in a fixed disposition with respect to the xanthenone chromophore. None of the compounds showed superior potency to XAA itself, with virtually all alterations in the nature of the anionic center or its geometry with respect to the chromophore greatly reducing or abolishing activity. However, alpha-methylation of the side chain was permissible, and the two enantiomers of 5-methyl-alpha-methyl-XAA were separated and tested. Both were active, but the S-(+) enantiomer was much more dose-potent than the R-(-) enantiomer, in both the in vivo tumor necrosis assay and an in vitro assay measuring the stimulation of nitric oxide production by macrophages. This suggests that the enantiomers have different intrinsic activities, rather than differing in their vivo metabolism.

A series of tricyclic analogues of 9-oxo-9H-xanthene-4-acetic acid have been prepared and evaluated for their ability to cause hemorrhagic necrosis in subcutaneously implanted colon 38 tumors in mice, in an effort to extend the structure-activity relationships for this series. As was found previously with analogues of flavone-8-acetic acid (FAA) (Atwell et al. Anti-Cancer Drug Des. 1989, 4, 161), all electronic modifications of the XAA nucleus led to severe decreases or complete abolition of activity, suggesting narrow structure-activity relationships. Dipole moments for many of the compounds were computed, and the degree to which the molecular dipole moment lay out of the plane of the aromatic part of these molecules was found to be determined largely by the contributions from the acetic acid moiety relative to that from the tricyclic ring system. There did not appear to be any general relationship between the magnitude of the dipole moment and activity. However, for compounds containing the 9-carbonyl functionality, the orientation of the dipole vector may be of significance. In all compounds possessing an ether group peri to the acetic acid side chain, there was a close approach (ca. 2.4 A) between this and the side chain OH.

Analogues of 9-oxo-9H-xanthene-4-acetic acid (XAA) bearing small, lipophilic 5-substituents are among the most dose-potent compounds yet reported with the capability of causing hemorrhagic necrosis of implanted colon 38 tumors in mice. To further extend structure-activity relationships among this class of compound, a series of XAA derivatives bearing two small lipophilic groups at various positions have been prepared and evaluated. The 5,6-disubstituted compounds in particular show consistently high levels of both dose potency and activity, suggesting this is the optimal configuration among substituted 9-oxo-9H-xanthene-4-acetic acids. The 5,6- dimethyl and 5-methyl-6-methoxy are the most effective analogues, showing in vivo colon 38 activity comparable to that of FAA at 10-15-fold lower doses and superior activity to FAA at the respective optimal doses, and the former has been selected for detailed evaluation.

Flavone-8-acetic acid (FAA) induces haemorrhagic necrosis and tumour regression in experimental tumours and induces natural killer (NK) activity. Xanthenone-4-acetic acid (XAA) forms the basis of a series of analogues of FAA which vary in antitumour potency. FAA, XAA and 15 XAA derivatives were tested for their ability to induce either NK activity in mouse spleens or haemorrhagic necrosis in mouse colon 38 tumours. Some derivatives were active in both assays (one at a dose 8-fold lower than that of FAA). When both assays were quantitated, a significant correlation (r = 0.85; P less than 0.001) was found. NK assays could be useful in screening compounds such as FAA and XAA analogues which appear to mediate their antitumour activity by biological response modification. Since tumour necrosis may not be mediated directly by NK cells, FAA and active XAA derivatives may exert pleiotropic effects that include NK induction and tumour necrosis by acting on host cells to release cytokines.

A total of 25 different tricyclic DNA-intercalating chromophores bearing a common -CONH(CH2)2N(CH3)2 solubilizing sidechain have been compared with the 'classical' frameshift mutagen 9-aminoacridine for their ability to induce revertants in Salmonella typhimurium strain TA1537 (sensitive to frameshift mutation by acridine mutagens). The compounds showed varying levels of activity in this strain. For the fused linear and fused angular tricyclics, activity varied from zero to similar levels to 9-aminoacridine, but with no discernable relationship between activity and either structure or the measured physico-chemical properties. However, the '2-1' tricyclic compounds had essentially no mutagenic activity. Since several of these compounds have high in vivo antitumor activity, this is useful knowledge.

9-Oxoxanthene-4-acetic acids are a class of antitumor agents effective against the mouse colon adenocarcinoma 38 in vivo. Within this class, 5-substituents on the xanthenone are known to enhance potency. To extend structure-activity relationships for the class, a series of derivatives bearing a wide variety of substituents at the 5-position have been prepared and evaluated. The results suggest that activity correlates better with the lipophilic properties of substituents rather than with their electronic properties. Generally, lipophilic substituents result in more active compounds, but there may be a size limitation on such substituents. The 5-methyl derivative is the most dose-potent of the analogues studied.

The nitroacridine derivative nitracrine is a potent hypoxia-selective cytotoxin for mammalian cells in culture. In an attempt to modulate the degree of hypoxia selectivity among this class of compounds, we have studied a series of side-chain analogues of nitracrine. Both the electronic and steric properties of the side chain are shown to be important in determining the hypoxia selectivity of the compounds, by controlling the degree of aminoacridine/iminoacridan tautomerism. Studies with the repair-defective Chinese hamster cell line UV4 indicate that the cytotoxicity of all the compounds is due to nitro group reduction and subsequent macromolecular adduct formation. However, compounds such as the 9-amino derivative, which exist totally as the aminoacridine tautomer, form much less lethal lesions than the 9-alkylamino derivatives, which exist to varying degrees in the iminoacridan conformation. For the whole set of compounds, the degree of hypoxia-selective cytotoxicity correlates well with the proportion of iminoacridan tautomer present.

Replacement of the 1'-methanesulfonamide group of the 9-anilinoacridine class of antitumor agents with the 1'-(dimethyl phosphoramidate) group provides compounds that are generally more lipophilic and bind more tightly to DNA. On the average, the dimethyl phosphoramidates are twice as dose potent as the corresponding methanesulfonamide (AMSA) compounds against P388 leukemia in vivo, but also show about twice the acute toxicity and no resultant improvement in tumor cell selectivity (ILSmax values) is seen. A pairwise comparison of a range of acridine-substituted compounds shows that structure-activity relationships within each series are similar and dominated by the acridine substitution pattern.

The DNA-intercalating agent amsacrine is an effective drug for the treatment of human leukemias and lymphomas but has minimal solid tumor activity. As a first step in identifying analogues with a wider spectrum of activity, a comparison was made of the in vivo antileukemic (P-388) activity of amsacrine analogues given by oral (po) and intraperitoneal (ip) routes. A series of 4-substituted and 4,5-disubstituted derivatives all showed high activity when administered ip against ip-implanted P-388, but activity varied widely when the compounds were given orally. 4-Methoxy and 4-carbamoyl derivatives proved essentially inactive, whereas 4-methyl and 4-methylcarbamoyl derivatives retained activity. Exceptional oral activity was shown by the 4-methyl-5-methylcarbamoyl derivative, making this amsacrine derivative worthy of further testing.

From structure--anti-L1210 relationships developed earlier for the 4'-(9-acridinylamino)alkanesulfonanilides it was predicted that congeners bearing both lipophilic 3-acridine substituents and compensatory hydrophilic function(s), together providing an overall molecular lipophilic--hydrophilic balance close to optimum, should have augmented antitumor properties. The acceptability of a variety of hydrophilic functions, and optimum positioning of these, has now been investigated. A variety of sterically demanding, hydrophilic functions may be acceptably appended to the acridine 4(5) position suggesting considerable site bulk tolerance. A variant with both a lipophilic 3-acridine substituent (3-iodo) and a hydrophilic 5-(2,3-dihydroxypropoxy) function is markedly more active than previous examples in the early treated, intraperitoneally (ip) dosed, ip implanted L1210 system, the assay system employed in the structure--activity analyses. However, this latter compound, on ip administration, failed to significantly inhibit subcutaneously implanted L1210 whereas earlier variants, under the same conditions, provided significant tumor inhibition. In this drug series the observed order of relative drug effectiveness alters with changing site of tumor implantation.

A series of amide-linked triacridines of varying interchromophore separation were synthesized as potential DNA trisintercalating agents. The corresponding diamide diacridines (lacking the central chromophore) were also prepared, and the DNA-binding and biological activities of both series of compounds were evaluated. Although one of the triacridines shows evidence of a trisintercalative binding mode, most of the triacridines (and all the diacridines) bound by bisintercalation. The diacridines showed great cytotoxicity and higher DNA association constants than the corresponding 9-[[3-(dimethylamino)propyl]amino]acridine monomer, but addition of a third chromophore did not provide corresponding increases in either DNA affinity, inhibition of RNA synthesis, or cytotoxicity. Some members of both series show minimal in vivo antileukemic activity. The results suggest that further development of trimeric molecules as potential antitumor agents should focus on smaller chromophores with lower capacity for nonspecific binding and/or the employment of rigid linker chains to provide true molecular "staples" for DNA.

A series of isomeric phenylquinoline-8-carboxamides have been synthesized and evaluated as antitumor agents. This configuration is close to the minimum chromophore required for intercalative binding, since the binding mode of the compounds is dependent on the presence and position of the phenyl ring. If the ring is appended at the 4- or 5-position, it cannot lie within the DNA-intercalation site, and the compounds do not intercalate as shown by both unwinding and helix extension assays. In contrast, the 2-, 3-, and 6-phenyl isomers (where the phenyl ring lies coplanar with the quinoline and in the intercalation site) bind by intercalation. Only those isomers that intercalate show in vivo antitumor activity, with the 2-phenyl derivative in particular possessing broad-spectrum activity in both leukemia and solid-tumor assays.

Quantitative relationships between physicochemical drug properties and antileukemic (L1210) efficacy have been examined for a series of bisquaternary ammonium heterocycles employing multiple variable regression analysis. Three measures of biologic response were examined: ILSmax, the percentage increase in mean life span of leukemic animals at the LD10 dose; D40, the drug dose necessary to provide 40% increase in life span; and CI (=LD 10/D40), the chemotherapeutic index. A cross correlation matrix between these three measures and the LD10 values demonstrates ILSmax and CI to be independent of toxicity. D40 is highly inversely correlated with LD10 and positively correlated with ILSmax, suggesting that this parameter measures a composite of both drug selectivity and toxicity. Superior regression equations resulted at all stages employing ILSmax as a measure of antitumor selectivity. Acceptable equations modeling LD10 could not be obtained. There was a parabolic relationship between agent lipophilic-hydrophilic balance, measured as chromatographic Rm values, and ILSmax. To reduce residual variance in the L1210 screening data, not accepted by this parabolic equation, measures of agent-DNA interaction were investigated as possible indices of site fit. Relative levels of drug-DNA interaction were obtained by spectrofluorimetric quantitation of drug displacement of DNA-bound ethidium. Addition to regression equations of agent C50 values for calf thymus DNA, those micromolar drug concentrations necessary to displace 50% of the ethidium bound to that DNA, provided a significant reduction in the screening data variance. C50 values for drug interactions with poly[d(A-T)] and poly[d(G-C)] were also investigated as possible indicators of drug selectivity towards different DNA sites. Marked differences were observed in the C50 values for the two synthetic nucleic acids, with those for calf thymus DNA and poly[d(G-C)] proving highly covariant. A regression equation containing a positive term in C50[poly[d(G-C)]] and a negative term in C50-[poly[d(A-T)]] provided the greatest acceptance of the variation in the biologic data. The term in C50[poly[d(A-T)]] is the single most important equation term, alone accounting for 35% of the variance in the data. This best equation provides both an adequate summary of the L1210 screening data for the 174 active compounds considered and a physical explanation for the antitumor selectivity for these agents. The major factor influencing antitumor selectivity in these drugs is their ability to distinguish alternating adenine-thymine sequences in DNA.

The 4-(N-methylcarboxamido)-5-methyl derivative of amsacrine (NSC 249 992) has been synthesized as part of a program aimed at optimizing solid tumor activity in this series. Physicochemical studies of this analogue (N-5-dimethyl-9-[(2-methoxy-4-methylsulfonylamino) phenylamino]-4-acridinecarboxamide; NSC 343 499) indicate a slightly increased lipophilicity (estimated log p = 1.10), a decreased acridine base strength (pKa 6.40), and a 16-fold-higher association constant for double-stranded calf thymus DNA (Ka 2.1 X 10(6) M-1 at 0.01 ionic strength). Like amsacrine, the drug binds to DNA by intercalation. Inhibition of cell growth has been monitored by continuous drug exposure assays with a variety of rodent and human cell lines. The concentration for 50% inhibition varied from 6.7 nM (T-47D, a human breast carcinoma line) to 800 nM (P388/ADR, a murine cell line resistant to Adriamycin). N-5-Dimethyl-9-[(2-methoxy-4-methylsulfonylamino) phenylamino]-4-acridinecarboxamide was cytotoxic at growth-inhibitory concentrations and also induced cell cycle arrest in the G2 phase. It was active against P388 leukemia following administration by p.o., i.v., or i.p. routes, and it was superior to amsacrine, daunorubicin, and Adriamycin. It was curative towards i.v.-injected Lewis lung tumor in a proportion of animals when treatment was started on Day 1 or Day 5 after tumor inoculation. It also produced highly significant life extensions against advanced tumors (treatment starting Day 9 after i.v. inoculation or on Day 8 after s.c. inoculation) and was comparable to cyclophosphamide in its effectiveness. It is a candidate drug for clinical trial.

Factors influencing dose potency of 4'-(9-acridinylamino)methanesulfon-m-anisidide (m-AMSA) analogues in L1210 assays have been investigated by multiple regression analysis. The dependent variable was D40, the dose to provide 40% life extension in L1210 tests. Independent variables examined were chromatographic Rm values, as a measure of agent lipophilic-hydrophilic balance; Rm2; log K, where K is the agent-DNA association constant for poly[d(A-T)]; log T1/2, the half-life for congener thiolytic cleavage; and agent pKa values. A regression equation containing terms in Rm2 and log K was derived with the latter term accepting the greater proportion of the biological variance. DNA binding, of acridine substituted m-AMSA variants, is the most important factor influencing dose potency. Modeling of log K for 3-substituted derivatives provided an equation in substituent R constants and molar refractivities (MR).