DNA sequence-specific adenine alkylation by the novel antitumor drug tallimustine (FCE 24517), a benzoyl nitrogen mustard derivative of distamycin

FCE 24517, a novel distamycin derivative possessing potent antitumor activity, is under initial clinical investigation in Europe. In spite of the presence of a benzoyl nitrogen mustard group this compound fails to alkylate the N7 position of guanine, the major site of alkylation by conventional nitrogen mustards. Characterisation of DNA-drug adducts revealed only a very low level of adenine adduct formation. Using a modified Maxam-Gilbert sequencing method the consensus sequence for FCE 24517-adenine adduct formation was found to be 5'-TTTTGA-3'. A single base modification in the hexamer completely abolishes the alkylation of adenine. Using a Taq polymerase stop assay alkylations were confirmed at the A present in the hexamer TTTTGA and, in addition, in one out of three TTTTAA sequences present in the plasmid utilized. The sequence specificity of alkylation by FCE 24517 is therefore the most striking yet observed for an alkylating agent of small molecular weight.

Structure-activity relationship of a series of nitrogen mustard- and pyrrole-containing minor groove-binding agents related to distamycin

Two series of tethered nitrogen mustards based on the minor groove-binding and A/T sequence-specific natural product distamycin have been synthesized and evaluated. The conjugates, which have a modified dimethylamino C-terminus, are comprised of one, two or three pyrrole carboxamide units linked to either benzoic acid mustard (BAM) or chlorambucil (CHL). The DNA binding properties, in vitro cytotoxicities and DNA cross-linking abilities were determined for each of the conjugates. The conjugates were found to bind preferentially to poly(dA.dT) compared to poly(dG.dC) DNA by ethidium displacement and circular dichroism. The di- and tripyrrole conjugates had higher binding affinities than the monopyrrole conjugates. All the conjugates were more cytotoxic than the nitrogen mustards themselves. Cytotoxicity increased with the increase from one to three pyrrole units and the CHL conjugates were more cytotoxic than the corresponding BAM analogues. The CHL conjugates were able to cross-link plasmid DNA at a 10-fold lower dose than CHL itself. The BAM conjugates showed < 10% cross-linking at doses which gave 100% cross-linking with the CHL conjugates. In cells, the CHL conjugates showed significant cross-linking at the IC50 values, while the BAM conjugates showed no evidence of cross-link formation even at 10 times the IC50 value. These results are discussed in reference to a series of previously reported GC-recognizing imidazole analogues possessing the same nitrogen mustard groups.

DNA damage by anticancer agents and its repair: mapping in cells at the subgene level with quantitative polymerase chain reaction

The quantitative polymerase chain reaction (QPCR)-based assay was used to measure DNA damage and repair to a small (523 bp) fragment of the single-copy human N-ras gene in K562 cells. Compared with previous methods DNA preparation from treated cells and the subsequent detection of the radioactive product were considerably simplified. The results demonstrated that QPCR can be used to measure damage in a small gene segment, caused by cisplatin, nitrogen, and quinacrine mustards. Drug-DNA adducts produced by two novel minor groove binding, sequence-specific molecules (AT-486 and DSB-120) could be detected at physiologically relevant concentrations of drug. For both cisplatin and nitrogen mustard the concentrations required to cause damage in cells were higher than those needed to cause equivalent damage in isolated DNA. In contrast both AT-486 and quinacrine mustard caused more damage at equimolar concentrations in cells than in isolated DNA. DSB-120, which is closely related to AT-486, was found to be 15-fold less effective than the latter at causing damage in treated cells despite similar reactivity with isolated DNA. Repair of damage caused by quinacrine mustard to the same small gene fragment was found to proceed at a constant rate over 24 h. The QPCR assay presented here is a simple quantitative method to measure damage and repair in subgene functional units such as promoters, introns, and exons.

Cellular pharmacology of novel C8-linked anthramycin-based sequence-selective DNA minor groove cross-linking agents

The cellular pharmacology of a series of C8-linked pyrrolobenzodiazepine dimers with polymethylene linkers of n = 3-6 (compounds 1-4) has been studied in a range of human tumour cell lines. The four compounds showed the same pattern of relative activity in five ovarian carcinoma cell lines and one cervical carcinoma cell line with the order of IC50 values of 1 < or = 3 < 4 < 2, which correlated with the previously demonstrated DNA interstrand cross-linking ability of the compounds in plasmid DNA. In human leukaemic K562 cells the agents produced a block in the G2/M phase of the cell cycle characteristic of cross-linking drugs, and extensive interstrand cross-linking was observed in cells by alkaline elution with no evidence of single-strand breaks. Cross-links continued to increase up to 24 h following a 1 h exposure to drug, and no repair was evident by 48 h. In a series of ovarian and cervical carcinoma cell lines with acquired resistance to cisplatin no cross-resistance to the most potent compound 1 was observed in two lines whose major mechanism of resistance to cisplatin was reduced platinum transport. Cross-resistance to 1 was observed in a cell line (A2780cisR) possessing elevated glutathione, and depletion of intracellular glutathione using D,L-buthionine-S,R-sulphoximine (BSO) from 10.25 nmol to 2.8 nmol 10(-6) cells reduced the level of resistance from 11-fold to 2-fold compared with sensitive cells. Cross-linking in the resistant cells was restored to 80% of the level in the parent line by BSO pretreatment. There was also a correlation between glutathione levels and sensitivity to 1 measured in several other ovarian cell lines. Compound 1 also showed cross-resistance in the doxorubicin-resistant cell line 41MdoxR and partial cross-resistance in CH1doxR cells. Both these lines possess elevated levels of p170 glycoprotein. Following treatment with 6 microM verapamil, the resistance in these lines decreased almost 2-fold and 8-fold respectively.

DNA damage by anti-cancer agents resolved at the nucleotide level of a single copy gene: evidence for a novel binding site for cisplatin in cells

A new PCR based technique has been developed to investigate the sequence selectivity of adduct formation by DNA damaging agents in a single copy gene in isolated genomic DNA or in drug treated cells. Single-strand ligation PCR (sslig-PCR) demonstrated that cisplatin and nitrogen mustards reacted with guanine in an N-ras fragment with varying sequence specificity similar to that observed previously in plasmid DNA. In cisplatin-treated cells sslig-PCR demonstrated all the adducts found in isolated DNA and with the same sequence selectivity showing a preference for GG and AG sites. However, in cells an additional site of DNA binding of cisplatin was observed at the two occurrences of the sequence 5'-TACT-3' on the transcribed and non-transcribed strands. This sequence is not a recognised target for cisplatin and represents a novel adduct formed in cells.

Molecular and cellular pharmacology of novel photoactive psoralen and coumarin conjugates of pyrrole- and imidazole-containing analogues of netropsin

The molecular and cellular pharmacology of novel sequence-directed photoactive agents, in which either psoralen or coumarin is conjugated to minor groove-binding AT-selective pyrrole-, or more GC-selective imidazole-containing analogues of netropsin, is described. The compounds were relatively non-toxic in the dark and showed marked photoinduced cytotoxicities when irradiated at 366 nm UV. The psoralen-containing pyrrole (1) and imidazole (2) compounds gave the largest photoinduced effect, were more active than 8-methoxypsoralen (8-MP) itself by 333- and 22-fold, respectively, and were more potent than the corresponding coumarin-containing analogues 3 and 4. Following irradiation, 1 and 2 were > 300- and > 10-fold more efficient at producing interstrand cross-links in naked DNA, respectively, than 8-MP. 1 was at least 10-fold more efficient at producing cross-links in cells than 2, reflecting the difference in their IC50 values. No cross-links were observed with the coumarin analogues, but these compounds were more potent than 8-MP. AT and GC sequence recognition was confirmed by DNA footprinting, and sites of covalent modification mapped by a polymerase stop assay. All compounds produced blocks at thymine base sites following irradiation. 1 was more efficient than 8-MP and produced a different pattern of covalent modification, whereas 2 was more selective than either 1 or 8-MP. A 1H-NMR study on a 1:1 complex of 2 with the hexamer (5'-dA1T2G3C4A5T6-3')2 indicated that the imidazole carboxamide moieties of 2 reside in the minor groove of the sequence 5'-GCAT-3' of the hexamer with the C-terminus located on the 3'-TA site, and the psoralen group intercalated between the 5'-A1T2-3' base pairs.

In vitro photoinduced cytotoxicity and DNA binding properties of psoralen and coumarin conjugates of netropsin analogues: DNA sequence-directed alkylation and cross-link formation

he synthesis, DNA binding and in vitro photoinduced cytotoxic properties of a number of minor groove and sequence-directed psoralen and coumarin conjugates of pyrrole- and imidazole-containing distamycin analogues 2-5 are described. Results from an ethidium displacement assay on calf thymus and T4 DNA suggest that like distamycin these agents bind strongly to the minor groove of DNA. The data show that these conjugates exhibit a lower AT preference than distamycin and the decrease is significantly greater for the imidazole-containing compounds. All of the compounds along with 8-methoxypsoralen, 1, were relatively noncytotoxic in the dark with only the imidazole-psoralen compound 3 giving an IC50 value below 100 microM. Following UV activation, all compounds showed an increased potency with photoinduced dose modifications in the human chronic myeloid leukemia K562 cells of > 333, 12, > 1.3, and > 2.5 for compounds 2-5, respectively, under the UV irradiation conditions employed. The psoralen-pyrrole analogue 2 was over 300 times more active following UV activation than agent 1, 250 times more potent than the corresponding coumarin conjugate 4, and 15-fold more potent than its imidazole analogue 3. Data from CD dilution (with DMF) studies show that upon irradiation with light at 366 nm, compounds 2-5 bind irreversibly to DNA. Furthermore, upon irradiation compound 2 produced interstrand cross-linked DNA in quantitative yield, with isolated DNA, at > 300- and > 3000-fold lower drug concentrations than the imidazole analogue 3 and 8-methoxypsoralen, respectively. As expected coumarin conjugates 4 and 5 did not produce any cross-linked DNA under any conditions. Since the psoralen conjugates are more phototoxic than their coumarin analogues, these results suggest that DNA interstrand cross-link formation may be an important mechanism by which they exert their biological activity in cells. In addition, the enhanced photocytotoxic potency of conjugate 2 over 3 may be related to its larger binding constant, more efficient DNA cross-linking ability, and possibly to its preference for AT-rich sequences.

Electrophoretic and chromatographic separation methods used to reveal interstrand crosslinking of nucleic acids

Several electrophoretic and chromatographic techniques, many of which have only been developed recently, provide sensitive methods for the detection and separation of DNA containing interstrand crosslinks such as those produced by many cancer chemotherapeutic drugs and photoactive psoralen derivatives. Most of the methods rely on the fact that the presence of such crosslinks prevent the complete denaturation of the two complimentary DNA strands by heat or alkali. A simple and highly sensitive neutral agarose gel electrophoresis method is particularly applicable to detailed time-course experiments of both total crosslink formation, and the "second-arm" of the crosslink reaction. This method separates denatured single-stranded from double-stranded DNA which has reannealed as a result of an interstrand crosslink. Polyacrylamide gel-based assays using denaturing gels are more suited to the separation of smaller crosslinked DNA fragments and, in particular, small oligonucleotides on high-percentage gels. In addition, they provide methods for the determination of the exact base position and sequence selectivity of crosslink formation. Sephadex chromatography and high-performance liquid chromatography can separate small crosslinked oligonucleotides from non-crosslinked duplexes, and the hydroxyapatite column chromatographic separation of single- and double-stranded cellular DNA can be used to quantitate the level of interstrand crosslinking present in the bulk of the genome. Finally, the analysis of damage by crosslinking agents, and its repair, at the level of specific genes can be achieved by hybridization with specific probes following membrane transfer from neutral agarose gels used to fractionate restricted and fully denatured genomic DNA from drug-treated cells.

A quantitative assay to measure the relative DNA-binding affinity of pyrrolo[2,1-c] [1,4]benzodiazepine (PBD) antitumour antibiotics based on the inhibition of restriction endonuclease BamHI

An assay has been developed (restriction endonuclease digestion assay--RED100) based on inhibition of the restriction endonuclease BamHI that is capable of quantitative evaluation of the relative DNA-binding affinity of pyrrolo[2,1-c] [1,4]benzodiazepine (PBD) antitumour antibiotics. This method provides comparable results to those obtained from thermal denaturation and ethidium bromide displacement assays but is much more sensitive, discriminating between molecules of similar structure such as DC-81, iso-DC-81 and neothramycin. The results reveal a trend between relative DNA-binding affinity and in vitro cytotoxicity for the PBDs in two tumour cell lines studied.

bcl-2 modulation of apoptosis induced by anticancer drugs: resistance to thymidylate stress is independent of classical resistance pathways

The hypothesis was tested that expression of bcl-2 could provide protection against apoptosis induced by cytotoxic drugs via a mechanism which was different from the classical determinants of drug resistance. Sensitivity and resistance to inhibitors of thymidylate synthase (EC 2.1.1. 45) were chosen for study since these drugs have a well-defined and quantifiable locus of action with similarly well defined biochemical sequelae resulting from enzyme inhibition. Human lymphoma cells transfected with the vector alone readily underwent apoptosis after a 36-h exposure to various drugs. For example, 5-fluorodeoxyuridine (0.1 microM) induced 67% apoptosis in vector control cells 24 h after removal of the drug. In contrast, cells treated under identical conditions, but which expressed the bcl-2 protein, showed only basal levels of apoptosis (8%), with no significant fall in viability. Similar results were obtained using two quinazoline-based inhibitors of thymidylate synthase, N10-propargyl-5,8-dideazafolic acid (CB3717) and ICI M247496. Determinants of resistance to these three drugs were investigated. Analysis of the cell cycle, thymidylate synthase levels, and activity showed these to be unchanged by expression of bcl-2. Addition of the drugs brought about equivalent inhibition of proliferation in the presence or absence of bcl-2 expression. 5-Fluorodeoxyuridine treatment reduced TTP synthesis, induced strand breaks in nascent DNA, measured by alkaline elution, and increased the synthesis of thymidylate synthase; these changes preceded the onset of apoptosis and were identical in the vector controls and bcl-2 transfectants. Resistance to thymidylate stress in bcl-2-expressing cells therefore occurred by a mechanism different from those which classically define resistance to this type of cytotoxic drug.

Design, synthesis, and biological evaluation of DNA sequence and minor groove selective alkylating agents

The syntheses of oligoimidazolecarboxamido analogues of distamycin wherein the N-terminus contains either a benzyl-mustard 8 or chlorambucil moiety 9-11 are reported. Based on data from an ethidium displacement assay and CD studies, these compounds, along with the N-benzoyl mustards 6 and 7, were shown to have increased acceptance of GC-rich sequences over distamycin. Compounds 8-11 which contain an electron-donating group (sigma < 0) para to the bischloroethylamino moiety, were found to have significantly enhanced reactivity with DNA compared to the benzoyl mustards 6 and 7. Through dialysis experiments, the benzyl and chlorambucil mustards were shown to alkylate calf thymus DNA more readily than the benzoyl mustards, presumably due to destabilization of the aziridinium intermediate by the electron-withdrawing (sigma > 0) carboxamido group of the benzoyl compounds. Compounds 8-11 were found to alkylate guanine-N7 in the major groove, while compounds 6 and 7 did not, suggesting that they may have different modes of DNA interaction. Mustards 8-11 were also more efficient than 6 and 7 at producing DNA interstrand cross-links in isolated DNA. In general, for these compounds, the cytotoxicity against human chronic myeloid leukemia K562 cells and the panel of human tumor cell lines of the National Cancer Institute increased with the number of imidazole moieties. The IC50 values of compounds 7 and 8 were similar, even though the latter compound was at least 100-fold more efficient at forming DNA cross-links in isolated DNA. Similarly, compounds 9-11 were less cytotoxic than 6 and 7, although they were more efficient cross-linkers in isolated DNA. A direct comparison of the three imidazole-containing benzoyl mustard 7 with the corresponding chlorambucil-containing 11 for their ability to form interstrand cross-links in cells revealed that the former compound showed no cross-linking even at doses in excess of the IC50, whereas the latter produced extensive cross-linking. This further suggests that these compounds exert their biological activity through different mechanisms. It is proposed that the aromatic moiety of compounds 8-11, which bind to the minor groove, may be able to intercalate between GC base pairs and the protruding bischloroethylamino group would be positioned to alkylate and cross-link at guanine-N7 sites in the major groove. However, the benzoyl mustards, which have a rigid amino linkage between the imidazole and aromatic-mustard moieties, do not have the flexibility to intercalate into the DNA.(ABSTRACT TRUNCATED AT 400 WORDS)

GC base sequence recognition by oligo(imidazolecarboxamide) and C-terminus-modified analogues of distamycin deduced from circular dichroism, proton nuclear magnetic resonance, and methidiumpropylethylenediaminetetraacetate-iron(II) footprinting studies

The DNA binding properties of a series of imidazole-containing and C-terminus-modified analogues 4-7 of distamycin are described. These analogues contain one to four imidazole units, respectively. Data from the ethidium displacement assay showed that these compounds bind in the minor groove of DNA, with the relative order of binding constants of 6 (Im3) > 7 (Im4) > 5 (Im2) > 4 (Im1). The reduced binding constants of these compounds for poly(dA-dT) relative to distamycin, while they still interact strongly with poly(dG-dC), provided evidence of GC sequence acceptance. The preferences for GC-rich sequences by these compounds were established from a combination of circular dichroism (CD) titration, proton nuclear magnetic resonance (1H-NMR), and methidiumpropylethylenediaminetetraacetate-iron(II) [MPE.Fe-(II)] footprinting studies. In the CD studies, these compounds produced significantly larger DNA-induced ligand bands with poly(dG-dC) than poly(dA-dT) at comparable ligand concentrations. 1H-NMR studies of the binding of 5 to d-[CATGGCCATG]2 provided further evidence of the recognition of GC sequences by these compounds, and suggested that the ligand was located on the underlined sequence in the minor groove with the C-terminus oriented over the T residue. MPE footprinting studies on a GC-rich BamHI/SalI fragment of pBR322 provided unambiguous evidence for the GC sequence selectivity for some of these compounds. Compounds 4 and 7 produced poor footprints on the gels; however, analogues 5 and 6 gave strong footprints.(ABSTRACT TRUNCATED AT 250 WORDS)

Two structurally related diaziridinylbenzoquinones preferentially cross-link DNA at different sites upon reduction with DT-diaphorase

The nucleotide sequence preferences for the formation of interstrand cross-links induced in DNA by 2,5-diaziridinyl-1,4-benzoquinone (DZQ) and 3,6-dimethyl-2,5-diaziridinyl-1,4-benzoquinone (MeDZQ) were studied using synthetic duplex oligonucleotides and denaturing polyacrylamide gel electrophoresis (PAGE). Reaction of these bifunctional alkylating agents with a DNA duplex containing several potential cross-linking sites resulted in the formation of cross-linked DNAs with different electrophoretic mobilities. Analysis of the principal cross-linked products by piperidine fragmentation revealed that the preferential site of cross-linking was altered from a 5'-GNC to a 5'-GC sequence upon reduction of DZQ to the hydroquinone form by the enzyme DT-diaphorase. In contrast, the reduced form of MeDZQ was found to preferentially cross-link at 5'-GNC sites within the same sequence. These preferences were confirmed in duplex oligonucleotides containing single potential cross-linking sites. Additional minor cross-linked products were characterized and revealed that DZQ and MeDZQ are both capable of cross-linking across four base pairs in a 5'-GNNC sequence.

In vitro cytotoxicity of GC sequence directed alkylating agents related to distamycin

Imidazole containing analogues 7, 10, and 17 of distamycin wherein the C-terminus contain a dimethylamino moiety have been shown to selectively bind to the minor groove of GC-rich sequences. Accordingly, these agents were employed as vectors for the delivery of a variety of alkylating agents to GC-rich sequences. The alkylating agents are attached to the N-terminus of these vectors thus providing the benzoyl N-mustards (8, 15, and 18 that contain one, two, and three imidazole units, respectively) and substituted acetamides 11-14. Results from the ethidium displacement assay for the formamides 7, 10, and 17 and mustards 15 and 18 showed that these agents bind to calf thymus DNA, poly(dA.dT), poly(dG.dC), and also to coliphage T4 DNA, thus confirming their binding in the minor groove. The reduced binding constants of these compounds for poly(dA.dT) while still binding as strongly, or more strongly, to poly(dG.dC) than distamycin provided evidence for their acceptance of GC sequences. Selectivity for GC-rich sequences was also indicated by CD titration studies. Titration of 10, 15, 17, and 18 to poly(dA.dT) produced weak drug-induced CD bands at approximately 330 nm; however, interaction of these agents to poly(dG.dC) in equimolar drug concentrations gave strong bands in this region. Results from dialysis and cross-link gel experiments provided evidence of alkylation and cross-linking of DNA by the mustards which could explain their enhanced cytotoxicity over the formamido analogues. The bifunctional N-mustard-containing analogues 15 and 18 are significantly more cytotoxic than the monoalkylating acetamides 11-14. The mustards also exhibited significant activity against cell lines derived from solid tumors such as melanomas, ovarian cancers, CNS cancers, and small cell lung cancer.

Photosensitization by anticancer agents. 11. Mechanisms of photosensitization of human leukemic cells by diaminoanthraquinones: singlet oxygen and radical reactions

The synthesis of several aminoanthraquinone derivatives (AAQs), designed to suppress the dark toxicity and to promote more efficient cancer cell photosensitization for potential use in photodynamic therapy (PDT), is described. The following AAQs were synthesized: 1-NH2-4,5-(MeO)2-AQ (1), 1,5-(NH2)2-4,8-(MeO)2-AQ (2), 1,8-(NH2)2-4,5-(MeO)2-AQ (3), and 1,5-(NHPhMe)2-4,8-(MeO)2-AQ (8). The agents exhibit strong absorption in the region 480-620 nm. Possible mechanisms of photosensitization were studied by measuring 1O2 phosphorescence at 1270 nm, detecting superoxide radicals employing an electron paramagnetic resonance (EPR)-spin trapping technique, and measuring oxygen consumption during the photo-oxidation of a representative biological electron donor, NADH. Strong phosphorescence from 1O2 was observed upon illumination of 2 and 3 in C6H6 (quantum yield of 0.25 and 0.5 respectively), and in EtOH (quantum yield of 0.23 and 0.34). The 1-amino-AQ (1) was the weakest 1O2 sensitizer, with quantum yield of 0.13 in benzene. No phosphorescence was observed in EtOH. A superoxide radical was detected as a spin adduct of 5,5-dimethyl-1-pyrroline-N-oxide (DMPO) in irradiated benzene solutions of 1, 2 or 3 and DMPO. AAQs 2 and 3 sensitized photo-oxidation of NADH in H2O/EtOH mixture with the intermediacy of singlet oxygen as judged by the effect of sodium azide on the photostimulated oxygen consumption. Evolution of O2 upon addition of catalase to the illuminated solution confirmed the ultimate formation of hydrogen peroxide. These findings suggested that the (di)amino-dimethoxyanthraquinones might exert photosensitization via both Type I and Type II mechanisms. The AAQs were tested for their ability to photosensitize K562 human chronic myeloid leukemic cells in culture. Viability was measured using the 3,4,5-diethylthiazol-2,5-diphenyl tetrazolium blue assay, and DNA and possible membrane damage were assessed. The results from illuminating cells with light > 475 nm show that for the 1,5-compounds, the presence of methoxy substituents at 4,8 positions reduces the dark toxicity from ID50 of 23 to 250 microM and for the 1,8-compounds correspondingly from ID50 of 53 to > 300 microM. In the 1,5-series this decrease of the dark toxicity is accompanied by an increase in light-induced dose modification from 8.85 to 14.4. Differences exist in the mechanisms of cytotoxicity between the prototype phenolic AAQs and their methoxy counterparts. It appears that the cytotoxic action of the latter causes cell damage by the formation of a high proportion of alkali labile sites in addition to frank strand breaks.(ABSTRACT TRUNCATED AT 400 WORDS)

Chlorambucil-monoglutathionyl conjugate is sequestered by human alpha class glutathione S-transferases

The spontaneous reaction of 110 microM chlorambucil (4-[p-[bis(2-chloroethyl)amino]phenyl]-butanoic acid; CHB) with 5 mM GSH at 37 degrees C in physiological phosphate-buffered saline for 35 min gave primarily the monoglutathionyl derivative, 4-[p-[N-2-chloroethyl,N-2-S-glutathionylethyl]amino]phenyl]-butano ic acid; CHBSG) and the diglutathionyl derivative, 4-[p-[bis(2-S-glutathionylethyl]amino]phenyl]-butanoic acid (CHBSG2) with small amounts of the hydroxy-derivatives: 4-[p-[N-2-chloroethyl,N-2-hydroxy-ethyl]amino] phenyl-butanoic acid (CHBOH) and 4-[p-[N-2-S-glutathionylethyl-2-hydroxyethyl]amino]phenyl]-butanoi c acid (CHBSGOH). The inclusion of approximately physiological amounts of human glutathione S-transferases (GSTs) A1-1, A2-2, P1-1, M1a-1a M3-3 or P1-1 (for nomenclature see Mannervik et al., 1992, Biochem. J., 282, 305) had little or no catalytic effect on these reactions as determined by loss of CHB. However, GTSs A1-1 and A2-2 were associated with a significant increase of CHBSG at the expense of CHBSG2 + CHBSGOH suggesting that these GTs sequestered CHBSG at the active site. This interpretation was supported by inhibition studies which showed that CHBSG was a pure competitive inhibitor of the activity of GSTs A1-1 and A2-2 towards 1-chloro-2,4-dinitrobenzene with Ki's of 1.3 and 1.2 microM respectively. GSH transferases P1-1 and M1a-1a were inhibited by CHBSG above 10 microM. Incubation of 2 microM CHB, a concentration which may be of more significance for chemotherapy, in the presence or absence of GST A1-2 (20-50 microM) showed catalysis of GSH monoconjugation equivalent to 18% of the spontaneous rate. However, the dominant effect again was the sequestration of CHBSG which reached 74.3 +/- 1.5 (SEM)% of the total reactants at 60 min compared to 28.9 +/- 0.3(SEM)% in controls. CHBSG, although possessing a potential electrophilic centre, showed no detectable alkylation of plasmid DNA but indirect evidence was obtained that it alkylated other cellular macromolecules. It is concluded that the contribution of GSTs to catalysis of CHB detoxication will depend on factors not previously considered, namely the relative molarities of CHB, CHBSG and GSTs, and the cellular capacity to excrete CHBSG to relieve product inhibition.

Relationship between DT-diaphorase-mediated metabolism of a series of aziridinylbenzoquinones and DNA damage and cytotoxicity

A series of 2,5-bis-substituted 3,6-diaziridinyl-1,4-benzoquinones have been tested for their ability to be reduced by the two-electron NAD(P)H:(quinone acceptor) oxidoreductase [DT-diaphorase (DTD); EC 1.6.99.2]. Symmetrically alkyl-substituted carbamoyl ester analogs of 2,5-ethyl(carboethoxyamino)3,6-diaziridinyl-1,4- benzoquinone [AZQ], 3,6-diaziridinyl-1,4-benzoquinone (DZQ), and its 2,5-dimethyl derivative (MeDZQ) were tested. The rate of reduction by DTD was DZQ greater than MeDZQ greater than n-butyl- (D5) greater than sec-butyl- (D7) greater than n-propyl- (D3) greater than methyl- (D1) greater than ethyl- (AZQ) greater than i-butyl- (D6) greater than i-propyl- (D4) substituted derivatives. The hydroxyethylamino analog (BZQ) was not a substrate for DTD. The order of toxicity to HT-29 human colon carcinoma cells (at 1-log cell kill) was MeDZQ greater than DZQ greater than BZQ greater than D1 greater than D5 greater than AZQ greater than D7 greater than D3 greater than D6 greater than D4. Dicumarol, a known inhibitor of DTD, was capable of inhibiting the cytotoxicity of DZQ, MeDZQ, AZQ, D3, D4, D5, D6, and D7, with little inhibition of D1 cytotoxicity. Alkaline elution assays suggested that DZQ induced DNA strand breaks, whereas MeDZQ induced DNA interstrand crosslinks in HT-29 cells. The formation of both classes of lesions was inhibited by dicumarol. DZQ and MeDZQ were 5-6-fold less cytotoxic to the DTD-deficient BE cell line, whereas BZQ was more cytotoxic to this cell line than the HT-29 cell line. BZQ was capable of inducing dicumarol-insensitive DNA interstrand crosslinks in both cell lines. In summary, these data show a trend between the rate of reduction by DTD of an analog and its ability to induce cytotoxicity in HT-29 cells, and they support a role for DTD in the bioreductive activation of AZQ and its analogs.

Targeting of tumor cells and DNA by a chlorambucil-spermidine conjugate

Many tumor cells, including murine ADJ/PC6 plasmacytoma cells, possess an active energy dependent polyamine uptake system which selectively accumulates endogenous polyamines and structurally related compounds. We have attempted to target the cytotoxic drug chlorambucil to a tumor possessing this uptake system by conjugating it to the polyamine spermidine. Furthermore, since polyamines have a high affinity for DNA, the attachment of spermidine to chlorambucil should also facilitate its targeting to DNA. This was supported by the observation that the chlorambucil-spermidine conjugate was approximately 10,000-fold more active than chlorambucil at forming interstrand crosslinks with naked DNA. In vitro cytotoxicity and in vivo antitumor studies were carried out using the ADJ/PC6 plasmacytoma. In vitro, using [3H]thymidine incorporation to assess cell viability following a 1-h exposure to control and polyamine depleted ADJ/PC6 cells, chlorambucil-spermidine was 35- and 225-fold, respectively, more toxic than chlorambucil. The increased toxicity of the conjugate compared to chlorambucil was possibly due to enhanced DNA binding and/or facilitated uptake via the polyamine uptake system. The enhanced toxicity of the conjugate but not chlorambucil by prior polyamine depletion with difluoromethylornithine, together with the observation that the conjugate but not chlorambucil competitively inhibited spermidine uptake into tumor cells, supported the suggestion that the conjugate utilized the polyamine uptake system. In vivo following a single i.p. dose, the conjugate was 4-fold more potent than chlorambucil in its ability to inhibit ADJ/PC6 tumor growth in BALB/c mice. However, the therapeutic index was not increased. Our results support the hypothesis that polyamines linked to cytotoxics facilitate their entry into tumor cells possessing a polyamine uptake system and increase their selectivity to DNA. This may have therapeutic application in the delivery of cytotoxic agents linked to polyamines to certain tumors.

Quinone-induced DNA single strand breaks in rat hepatocytes and human chronic myelogenous leukaemic K562 cells

In rat hepatocytes exposed to the quinones menadione and 2,3-dimethoxy-1,4-naphthoquinone (2,3-diOMe-1,4-NQ) a decrease in NAD+ is observed. DNA damage and activation of poly(ADP-ribose)polymerase are often associated with a decrease in NAD+. Using rat hepatocytes and human myeloid leukaemic cells (K562), we examined the extent of DNA damage induced by these quinones at non-toxic concentrations, i.e. at concentrations at which the cells completely exclude the dye trypan blue. Both quinones caused significant DNA damage at very low concentrations (5-100 microM). With 2,3-diOME-1,4-NQ (15 microM) or menadione (15 microM) single strand breaks (SSB) were observed at very early time points (less than 5 min), reaching a maximum between 20 and 30 min. Most SSB were repaired within 45 min of the removal of the quinones. Whilst extensive repair was observed within 4 hr of the removal of 2,3-diOMe-1,4-NQ (15 microM), only partial repair was observed following exposure to menadione (15 microM). SSB induced by 2,3-diOMe-1,4-NQ (15 microM) were completely inhibited by the iron chelator 1,10-phenanthroline (25 microM), whereas in cells exposed to menadione (15 microM) they were only partially inhibited. Finally, although the membrane integrity of K562 cells was unaffected by exposure to high concentrations of both quinones (less than or equal to 400 microM), cytostasis was observed at much lower concentrations (50 microM). Our results demonstrate that at very low concentrations these quinones induce extensive DNA damage possibly caused by hydroxyl radicals. The DNA damage was accompanied by an early cytostasis but no loss of membrane integrity.

The cytotoxicity, DNA crosslinking ability and DNA sequence selectivity of the aniline mustards melphalan, chlorambucil and 4-[bis(2-chloroethyl)amino] benzoic acid

Three aniline derivatives melphalan (L-PAM), chlorambucil (CHL) and 4-[bis(2-chloroethyl)amino] benzoic acid (BAM) have been compared on the basis of their in vitro cytotoxicities, DNA interstrand crosslinking ability and DNA sequence selectivity. Cytotoxicity was assessed in the human colonic adenocarcinoma LS174T and leukaemic K562 cell lines using the sulpho-rhodamine B and tetrazolium dye reduction assays. The order of cytotoxicities was L-PAM greater than CHL greater than BAM in both cell lines with K562 being less sensitive than LS174T. This was different from the order CHL greater than L-PAM greater than BAM which would be predicted from simple chemical reactivity or rate of hydrolysis, parameters which have been used previously as indicators of biological potency for aromatic nitrogen mustards. DNA interstrand crosslinking in cells as determined by alkaline elution showed a correlation with IC50 values. The ranking order of activity was further predicted by the ability of the agents to produce interstrand crosslinks in isolated DNA. The extent of guanine N-7 alkylation, assessed using a modified DNA sequencing technique, mirrored cytotoxicity and crosslinking ability, but at equivalent levels of alkylation there was no significant difference in DNA sequence selectivity. These data demonstrates that simple chemical reactivity or hydrolysis rate is not a good indicator of DNA reactivity or cytotoxicity for a number of aniline mustards, whereas DNA interstrand crosslinking ability either measured directly in cells or in isolated DNA, gives a good indication of biological activity.

DNA sequence selectivity of guanine-N7 alkylation by nitrogen mustards is preserved in intact cells

Nitrogen mustard alkylating agents react with isolated DNA in a sequence selective manner, and the substituent attached to the drug reactive group can impose a distinct sequence preference. It is not clear however to what extent the observed DNA sequence preferences are preserved in intact cells. The highly reiterated sequence of human alpha DNA has been used to determine the sites of guanine-N7 alkylation following treatment of cells with three nitrogen mustards, mechlorethamine, uracil mustard and quinacrine mustard, known to react in isolated DNA with distinctly different sequence preferences. Alpha DNA from drug treated cells was extracted, purified, end-labeled, and a 296 base pair, singly end-labelled, fragment isolated. Following the quantitative conversion of alkylation sites to strand breaks the fragments were separated on DNA sequencing gels. Clear differences were observed between the alkylation patterns of the three compounds, and the selectivities were qualitatively similar to those predicted and observed in the same sequence alkylated in vitro. In particular the unique preferences of uracil and quinacrine mustards for 5'-PyGC-3' and 5'-GT/GPu-3' sequences, respectively, were preserved in intact cells suggesting that the pattern of sequence dependent reactivity is not grossly affected by the nuclear milieu.

Potentiation of etoposide-induced cytotoxicity and DNA damage in CCRF-CEM cells by pretreatment with non-cytotoxic concentrations of arabinosyl cytosine

Pretreatment of the human lymphoblastoid cell line CCRF-CEM with 0.02 microM arabinosyl cytosine (ara C) enhances both the cytotoxic and the DNA-damaging effects of etoposide. This concentration of ara C is itself non-cytotoxic and results in no detectable DNA damage as measured by alkaline elution. Ara C pretreatment results in the synchronisation of cells, a 24-h pretreatment resulting in the accumulation of cells in the early S phase. The sensitivity of cells to etoposide-induced cytotoxicity was increased 2.5 times and DNA damage was enhanced 1.66 times by this pretreatment. Maximal potentiation of etoposide-induced DNA damage (2.06-fold increase) was observed after 48 h continuous treatment with ara C, but no further enhancement of cytotoxicity occurred. Cell-cycle analysis demonstrated that 48 h ara C treatment resulted in the accumulation of cells in the late S/G2M phase. Cells returned to a normal cell-cycle distribution within 24 h of the removal of ara C, and the potentiation of etoposide activity was then reduced to a 1.3- to 1.4-fold level. DNA damage induced by etoposide following ara C pretreatment was qualitatively identical to that produced by etoposide alone, suggesting a mechanism involving topoisomerase II. To investigate this possibility, we measured topoisomerase II protein levels by immunoblotting. Measurement of topoisomerase II levels in whole-cell lysates of ara C-pretreated cells showed a 3- to 5-fold increase in topoisomerase levels relative to total protein content. This suggests that elevated enzyme levels may be responsible for the increased sensitivity of ara C-pretreated cells to etoposide.

DNA cross-linking and sequence selectivity of aziridinylbenzoquinones: a unique reaction at 5'-GC-3' sequences with 2,5-diaziridinyl-1,4-benzoquinone upon reduction

Several bifunctional alkylating agents of the aziridinylbenzoquinone class have been evaluated as potential antitumor agents. 3,6-Bis[(2-hydroxyethyl)amino]-2,5- diaziridinyl-1,4-benzoquinone (BZQ), 2,5-diaziridinyl-1,4-benzoquinone (DZQ), 3,6-bis(carboxyamino)-2,5-diaziridinyl- 1,4-benzoquinone (AZQ), and six analogues of AZQ have been studied for their ability to induce DNA interstrand cross-linking, as measured by an agarose gel technique, and to determine whether they react with DNA in a sequence-selective manner, as determined by a modified DNA sequencing technique. At an equimolar concentration (10 microM), only DZQ and BZQ showed any detectable cross-linking at pH 7 without reduction. Cross-linking was enhanced in both cases at low pH (4). Reduction by ascorbic acid at both pH's increased the cross-linking, which was particularly striking in the case of DZQ. In contrast, AZQ and its analogues only produced a significant level of cross-linking under both low-pH and reducing conditions, the extent of cross-linking decreasing as the size of the alkyl end group increased. The compounds reacted with all guanine-N7 positions in DNA with a sequence selectivity similar to other chemotherapeutic alkylating agents, such as the nitrogen mustards, although some small differences were observed with BZQ. Nonreduced DZQ showed a qualitatively similar pattern of reactivity to the other compounds, but on reduction (at pH 4 or 7) was found to react almost exclusively with 5'-GC-3' sequences, and in particular, at 5'-TGC-3' sites. A model to explain this unique reaction is proposed.

Relationship between topoisomerase II level and chemosensitivity in human tumor cell lines

Patients with metastatic testis tumors are generally curable using chemotherapy, whereas those with disseminated bladder carcinomas are not. We have compared levels of the nuclear enzyme topoisomerase II in three testis (SuSa, 833K, and GH) and three bladder (RT4, RT112, and HT1376) cancer cell lines which differ in their sensitivity to chemotherapeutic agents. The testis cell lines were more sensitive than the bladder lines to three drugs whose cytotoxicity is mediated in part by inhibiting topoisomerase II: amsacrine; Adriamycin; and etoposide (VP16). The frequency of DNA strand breaks induced by amsacrine was higher (1.5- to 13-fold) in the testis cells than in the bladder cells. The level of topoisomerase II-mediated DNA strand breakage in vitro, measured by filter trapping of amsacrine-induced protein:DNA cross-links, was similarly higher in nuclear extracts from the testis than the bladder cells. Western blot analysis showed a generally higher level of topoisomerase II protein in testis than in bladder cell nuclear extracts. Topoisomerase II protein expression broadly correlated with drug-induced strand breakage in both protein extracts and whole cells, but not with population doubling time. However, despite a 2- to 20-fold increased sensitivity to the different topoisomerase II inhibitors, the testis line 833K had a less than 2-fold higher level of topoisomerase II protein than that of the bladder line RT4. These results indicate that the level of expression of topoisomerase II is an important determinant of the relative chemosensitivity of testis and bladder tumor cell lines, but that additional factors must contribute to the extreme chemosensitivity of testis cells.

N2,N4,N6-tri(hydroxymethyl)-N2,N4,N6-trimethylmelamine (trimelamol) is an efficient DNA cross-linking agent in vitro

An investigation of the mechanism of action of the antitumour agent trimelamol has established that it is an efficient interstrand DNA cross-linker in vitro, comparable to nitrogen mustards such as melphalan. Studies have shown that the cross-linking reaction is acid-catalysed but, unlike the nitrogen mustards, only partially reversible after treatment with piperidine. The bisalkylation (cross-linking) reaction appears to be concerted, and no "second arm" reaction has been detected. The results of thermal denaturation studies are consistent with general DNA binding, and suggest a preference for GC-rich sites. The acid-catalysed reaction of trimelamol with a model nucleophile (thiophenol) has also been investigated and an adduct resulting from displacement of the three carbinolamine functions has been isolated and characterized.

The effect of tranexamic acid on bleeding time and haemostasis

The effect of tranexamic acid (TXA) on platelet function was examined in 37 patients with ruptured intracranial aneurysms. TXA was administered in a double-blind, placebo controlled manner to 37 patients with ruptured intracranial aneurysms. TXA was given to 17 patients in a dose of 9 g/24 hours, to another 6 patients in a dose of 6 g/24 hours; 14 patients were given a placebo. The effect of TXA on certain haemostatic function tests was evaluated between the treatment groups. We conclude that patients receiving TXA do not appear to be at risk of recurrent haemorrhage due to abnormal haemostasis.

Measurement of the sequence specificity of covalent DNA modification by antineoplastic agents using Taq DNA polymerase

A polymerase stop assay has been developed to determine the DNA nucleotide sequence specificity of covalent modification by antineoplastic agents using the thermostable DNA polymerase from Thermus aquaticus and synthetic labelled primers. The products of linear amplification are run on sequencing gels to reveal the sites of covalent drug binding. The method has been studied in detail for a number of agents including nitrogen mustards, platinum analogues and mitomycin C, and the sequence specificities obtained accord with those obtained by other procedures. The assay is advantageous in that it is not limited to a single type of DNA lesion (as in the piperidine cleavage assay for guanine-N7 alkylation), does not require a strand breakage step, and is more sensitive than other primer extension procedures which have only one cycle of polymerization. In particular the method has considerable potential for examining the sequence selectivity of damage and repair in single copy gene sequences in genomic DNA from cells.

DNA interstrand crosslinking and sequence selectivity of dimethanesulphonates

Members of the homologous series of alkanediol dimethanesulphonates of general formula H3C.SO2O.(CH2)n.O.SO2.CH3 have been tested for their ability to produce DNA interstrand crosslinking and DNA sequence selectivity of guanine-N7 alkylation. In a sensitive crosslinking gel assay the efficiency of DNA interstrand crosslink formation, dependent on the ability of the alkylating moiety to span critical nucleophilic distances within the DNA, was found at 6 h to be 1,6-hexanediol dimethanesulphonate (Hexa-DMS) (n = 6) greater than methylene dimethanesulphonate (MDMS) (n = 1) greater than 1,8-octanediol dimethanesulphonate (Octa-DMS) (n = 8) greater than Busulphan (n = 4). The DNA interstrand crosslinking produced by MDMS was not due to either of its hydrolysis products, formaldehyde or methanesulphonic acid (MSA). In contrast the extent of monoalkylation at guanine-N7 as determined by a modified DNA sequencing technique was found to be Busulphan much greater than Hexa-DMS = Octa-DMS, with a sequence selectivity somewhat less than that of other chemotherapeutic alkylating agents such as nitrogen mustards. MDMS at high levels induced a non-specific depurination as a result of the reduction in pH resulting from MSA release. More strikingly MDMS (and MSA) produced a single strong site of guanine reaction (depurination) in a guanine-rich 276 base pair fragment of pBR322 DNA in the sequence of 5'-ATGGTGG-3'. This was observed when non-specific depurination was negligible and was not seen with formic acid. Thus structurally similar alkylating agents can differ in their type and extent of DNA monoalkylation and interstrand crosslinking, and in some cases (e.g. MDMS/MSA) produce reactions with a high degree of selectivity.

An agarose gel method for the determination of DNA interstrand crosslinking applicable to the measurement of the rate of total and "second-arm" crosslink reactions

A simple agarose gel electrophoresis method for the determination of DNA interstrand crosslinks is described. Following complete denaturation of 32P-end-labeled DNA the presence of an interstrand crosslink results in renaturation to double-stranded DNA. The single- and double-stranded bands separated on an agarose gel can be accurately quantitated by densitometry of the autoradiograph produced from the dried gel. The technique is particularly applicable to detailed time-course experiments of both total crosslink formation and, following removal of free drug, the "second-arm" of the crosslink reaction. The method is illustrated for a number of nitrogen mustard antitumor agents, showing how the moiety attached to a bifunctional reactive group can influence the extent and rate of crosslink formation and, in particular, the conversion of monoadducts to crosslinks. It is sensitive enough to follow the formation of crosslinks by slow and inefficient cross-linking agents such as busulfan which have not previously been measured by physical procedures.

Photosensitization by selected anticancer agents

Novel anticancer anthrapyrazoles and anthracenediones are available as alternatives to the cardiotoxic clinical agents, doxorubicin and daunorubicin. Certain representatives of these new classes of compounds possess photosensitizing properties. The structural features influencing the photophysical parameters of these agents are discussed. Photosensitizing reactions involving singlet oxygen production, free radical formation, decomposition of hydrogen peroxide and organic hydroperoxides, oxidation of certain biochemical electron donors, DNA damage and killing of human leukemic cells in vitro in the presence of photoactive anthrapyrazoles, anthracenediones and anthracyclines are described.

Effect of ionic strength and cationic DNA affinity binders on the DNA sequence selective alkylation of guanine N7-positions by nitrogen mustards

Large variations in alkylation intensities exist among guanines in a DNA sequence following treatment with chemotherapeutic alkylating agents such as nitrogen mustards, and the substituent attached to the reactive group can impose a distinct sequence preference for reaction. In order to understand further the structural and electrostatic factors which determine the sequence selectivity of alkylation reactions, the effect of increased ionic strength, the intercalator ethidium bromide, AT-specific minor groove binders distamycin A and netropsin, and the polyamine spermine on guanine N7-alkylation by L-phenylalanine mustard (L-Pam), uracil mustard (UM), and quinacrine mustard (QM) was investigated with a modification of the guanine-specific chemical cleavage technique for DNA sequencing. For L-Pam and UM, increased ionic strength and the cationic DNA affinity binders dose dependently inhibited the alkylation. QM alkylation was less inhibited by salt (100 mM NaCl), ethidium (10 microM), and spermine (10 microM). Distamycin A and netropsin (100 microM) gave an enhancement of overall QM alkylation. More interestingly, the pattern of guanine N7-alkylation was qualitatively altered by ethidium bromide, distamycin A, and netropsin. The result differed with both the nitrogen mustard (L-Pam less than UM less than QM) and the cationic agent used. The effect, which resulted in both enhancement and suppression of alkylation sites, was most striking in the case of netropsin and distamycin A, which differed from each other. DNA footprinting indicated that selective binding to AT sequences in the minor groove of DNA can have long-range effects on the alkylation pattern of DNA in the major groove.

Photosensitization of human leukemic cells by anthracenedione antitumor agents

1,4-Diamino-substituted anthraquinone antitumor agents (mitoxantrone and ametantrone) and structurally related 1,5- and 1,8-diamino-substituted compounds (AM1 and AM2) were tested for their ability to photosensitize human leukemic cells in culture. Viability was measured using the 3,4,5-dimethylthiazol-2,5-diphenyl tetrazolium bromide assay, and DNA and membrane damage were assessed. Following a 1-h exposure to AM2, a dose of drug required to give 50% loss of cell viability (53 microM) was obtained in the dark, which was reduced to approximately 2.4 microM following illumination for 2 min (lambda greater than 475 nm), a dose of light that was completely nontoxic to the cells in the absence of drug. A shift in the cell viability curve was also observed for AM1 but, under identical conditions, the dose modification was only 8.9. In contrast, neither ametantrone nor mitoxantrone gave a decreased viability upon illumination. DNA single-strand breaks as measured by alkaline elution correlated with cell viability. Frank DNA single-strand breaks were produced by AM2 and light, suggesting the production of free radicals. The strand breaks produced by AM2 in the dark and by mitoxantrone (with or without illumination) were protein concealed. No evidence of photo-induced membrane damage, as determined by transport of the model amino acid cycloleucine, could be observed even at supralethal doses.

Liblomycin-mediated DNA cleavage in human head and neck squamous carcinoma cells and purified DNA

Liblomycin (LBM), a novel bleomycin analogue, and bleomycin A2 (BLM A2) were compared with respect to their relative potential to inhibit growth in a human head and neck squamous carcinoma cell line and to produce DNA damage within cellular DNA and nuclei DNA and against isolated naked DNA. Against the BLM-sensitive cell line 183A, the concentration of LBM that inhibits cell growth by 50% was 1.1 microM for a 30-min drug exposure, while it was 23 microM for BLM A2. Drug-mediated DNA double-strand cleavage within cells was compared with the relative ability of these drugs to produce DNA cleavage in isolated 183A cell nuclei. Though 30-min exposures of cells to equimolar concentrations of both drugs resulted in 4-fold greater cellular DNA damage by LBM than BLM A2, the two drugs were nearly equipotent in producing DNA injury within isolated nuclei. Against Simian virus 40 DNA, however, LBM was 10-fold less effective than BLM A2 in producing Forms II and III DNA from Form I DNA. Radioactivity from either [3H]BLM A2 or 125I-LBM found associated with cells after a 30-min incubation period was also assessed in the 183A cell line. The exposure of cells to radiolabeled drug (1 microM) resulted in a 71-fold greater amount of cell-associated radioactivity for LBM than for BLM A2. The relative abilities of the 183A cell line to partially reseal LBM- or BLM A2-mediated DNA double-strand breaks were also assessed. No preferential repair of overall drug-mediated DNA injury, however, was observed. Finally, drug-mediated specific cleavage sites on pBR322 DNA were determined. At doses that gave the same extent of DNA cleavage, both BLM A2 and LBM gave similar patterns of strand scission, although minor differences were observed. Taken together, these data demonstrate that the greater efficacy of LBM against the BLM-sensitive head and neck squamous cell line is due mainly to LBM's greater association with cells over a defined time period, even though the DNA cleaving ability of LBM is relatively lower than that of BLM A2.

DNA damage and sequence specificity of DNA binding of the new anti-cancer agent 1,4-bis(2'-chloroethyl)-1,4-diazabicyclo-[2.2.1] heptane dimaleate (Dabis maleate)

The DNA damage and the sequence specificity of guanine-N7 alkylation produced by the novel, positively charged, antineoplastic agent 1,4-bis(2'-chloroethyl)-1,4-diazabicyclo-[2.2.1] heptane dimaleate (Dabis maleate) and its uncharged tertiary amine analogue 1,4-bis(2'-chloroethyl)-1,4-diazacyclohexane (Dabis analogue) were investigated in L1210 cells and isolated DNA. Both compounds are cytotoxic in vitro causing an arrest of L1210 cells in G2/M phase of the cell cycle. In isolated DNA, Dabis maleate alkylates guanine at the N7-position with some differences in specificity compared to other alkylating agents (e.g. nitrogen mustard). Significant differences are also evident between Dabis maleate and Dabis analogue, suggesting that Dabis analogue is not the sole alkylating species of Dabis maleate. Using the alkaline elution technique a moderate number of DNA interstrand cross-links were detected in L1210 cells treated with both compounds, which were completely repaired within 24 h. Dabis maleate and Dabis analogue do not cause DNA single strand breaks or DNA protein cross-links at the doses at which DNA interstrand cross-links were detected.

Interaction of the peroxidase-derived metabolite of mitoxantrone with nucleic acids. Evidence for covalent binding of 14C-labeled drug

The antitumor agent mitoxantrone undergoes horseradish peroxidase-catalyzed oxidation by hydrogen peroxide to an identifiable cyclic metabolite which is a substituted hexahydronaphtho[2,3-f]-quinoxaline-7,12-dione. Binding of mitoxantrone to DNA inhibited enzymatic oxidation of the drug. The metabolite of mitoxantrone, derived from the action of the HRP/H2O2 system on the drug, bound non-covalently to DNA oligomers. Spectrophotometric analyses of such complexes showed formation of a new, blue-shifted, metachromatic absorption band which was observed when the DNA base pair to drug ratio was close to 1. Measurements of DNA unwinding angles suggest that the metabolite, in contrast to mitoxantrone, did not intercalate but rather bound externally to DNA. Experiments with 14C-labeled mitoxantrone confirmed that peroxidase-activated drug binds covalently to DNA.

Molecular recognition between oligopeptides and nucleic acids. Sequence specific binding of (4S)-(+)- and (4R)-(-)-dihydrokikumycin B to DNA deduced from 1H NMR, footprinting studies and thermodynamic data

The sequence specific binding of the antibiotic (4S)-(+)-dihydrokikumycin B and its (4R)-(-) enantiomer, [(S)-1 and (R)-1, respectively] to DNA were characterized by DNase I and MPE footprinting, calorimetry, UV spectroscopy, circular dichroism, and 1H NMR studies. Footprinting analyses showed that both enantiomers [(S)-1 and (R)-1] bind to AT-rich regions of DNA. 1H NMR studies (ligand induced chemical shift changes and NOE differences) of the dihydrkikumycins with d-[CGCAATTGCG]2 show unambiguously that the N to C termini of the ligands are bound to 5'-A5T6T7-3' reading from left to right. From quantitative 1D-NOE studies, the AH2(5)-ligand H7 distance of complex A [(S)-1 plus decamer (which is bound more strongly)] and complex B [(R)-1 and decamer] are estimated to be 3.8 +/- 0.3 A and 4.9 +/- 0.4 A, respectively. This difference in binding properties is reflected in the thermodynamic profiles of the two enantiomeric ligands determined by a combination of spectroscopic and calorimetric techniques. The binding free energies (delta G degrees) of (S)-1 and (R)-1 to poly d(AT).poly d(AT) at 25 degrees C are -31.8 and -29.3 kJ mol-1, respectively while the corresponding binding enthalpies (delta H degrees) are -11.3 and -0.8 kJ mol-1. These data permit the construction of models for the binding of the enantiomeric dihydrokikumycins to DNA and account for the more efficient binding of the natural (S) isomer to DNA.

Interaction of bisantrene anti-cancer agents with DNA: footprinting, structural requirements for DNA unwinding, kinetics and mechanism of binding and correlation of structural and kinetic parameters with anti-cancer activity

A topoisomerase I assay in conjunction with covalently closed circular DNA afforded unwinding angles for a series of eight analogs of the anti-cancer agent bisantrene. Values ranging from 33 degrees to zero comprised three classes wherein the extent of DNA interaction correlates with anti-cancer activity. The binding of bisantrene compounds to T4 DNA suggests at least partial interaction via the minor groove. MPE footprinting evaluation of bisantrene showed no evidence for selective protection of a Hind III-EcoRI fragment of pBR322 suggesting a sequence neutral binding mode. Stopped flow spectrophotometry studies indicate an association rate constant for bisantrene with DNA of greater than 1 x 10(7) M-1 S-1 at 20 degrees at ionic strength of 0.1 and is considerably slower than ethidium or propidium at all ionic strengths. At least two exponential components are required to match the kinetic curves indicating a biphasic mechanism and estimates of k1 and k2 were obtained for seven drugs. The ionic strength dependence of the surfactant driven dissociation of bisantrene is in accord with a competitive mechanism for intercalation of the drugs at alternative sites. The residence time and t1/2 (app) for bisantrene are slightly larger for poly d(G.C)2 than poly d(A.T)2 and calf thymus DNA is at an intermediate value. Considerations of structural factors in the drug dissociation kinetics permit a correlation with observed anti-cancer properties of these bisantrene analogs.

GC-rich regions in genomes as targets for DNA alkylation

For many DNA-damaging agents, the extent of damage at any given base site is influenced by the DNA sequence surrounding that site. Most agents that alkylate the guanine N7 position, including mechlorethamine (nitrogen mustard) and benzo[a]pyrene diol epoxide, alkylate oligo-guanine sequences preferentially. Since these data suggest that guanine-cytosine(GC)-rich regions in genes could be preferred sites of damage by these agents, GenBank was searched for genes containing 30 bp sequences of greater than 90% GC (GC runs). While primate, rodent, other mammalian, vertebrate and animal virus genes constituted 57% of the annotated entries, they included 90% of the entries with the GC runs. In addition, the percentage of oncogenes in the group of the entries with GC runs was higher than that in the overall database. One gene of interest containing GC runs was the human c-Ha-ras oncogene. All seven GC runs in the c-Ha-ras gene are in the 5'-flanking region, rather than in the coding sequences. In fact, some of the GC runs are contained in Sp1-binding enhancer sequences. Gel analysis of the alkylation of cloned c-Ha-ras DNA by several carcinogenic alkylating agents strongly suggest that in this gene GC runs can be preferred sites of damage. These observations suggest mechanisms by which DNA damage at sites other than oncogene coding sequences may play a role in carcinogenesis and/or chemotherapy.

DNA sequence specificity of guanine N7-alkylations for a series of structurally related triazenes

The base sequence selectivity for reaction at the guanine-N7 position was examined for a series of structurally related triazenes by a modification of a standard DNA sequencing method. The monomethyl and monochloroethyl triazenes alkylate guanines extensively at the N7 position with a general preference for runs of contiguous guanines, similar to, but not as striking as that observed previously for the chloroethylnitrosoureas. In contrast to the nitrosoureas, the triazenes had patterns of base sequence selectivity that differed somewhat from agent to agent, with the monochloro-ethylphenyltriazene having the pattern most different from the others in the series. Thus, the nature of the nonalkylating portion of the molecule can influence the ultimate alkylation preference. The monoethylating analogues alkylated weakly with little sequence preference, and the dimethyl analogues were essentially unreactive in this system.

Characteristics of the interaction of anthrapyrazole anticancer agents with deoxyribonucleic acids: structural requirements for DNA binding, intercalation, and photosensitization

The binding constants for interaction of several novel anthra[1,9-cd]pyrazol-6(2H)-ones (anthrapyrazoles) with DNA have been determined by an ethidium displacement method. The apparent binding constants range from less than 2 X 10(6) to 2.7 X 10(8) M-1. The binding is influenced not only by the nature of the side chains but also by the number and position of hydroxyl groups on the chromophore. Unwinding angles, determined by a topoisomerase I assay, ranged from 0 degrees to 29.2 degrees. The deshydroxy compound 1 gave the highest unwinding angle, and both substitution of hydroxyl groups in the chromophore and alterations in the side chains decrease the unwinding angle, consistent with a decreased or partial intercalation. Representative anthrapyrazoles cause an increase in sonicated DNA viscosity as expected for intercalators. Spectrophotometric examination of the binding of compound 1 to DNAs of different base composition show that the apparent binding to GC is approximately 3 times that of AT, a result which was paralleled by thermal denaturation studies. Certain of the anthrapyrazoles exhibit marked visible light photosensitization and induce DNA single-strand breakage upon illumination in the presence of NADH. The essential structural requirement for photosensitizing properties with these agents was the absence of hydroxyl groups in the chromophore. By employing 32P-labeled DNA of known sequence, it was possible to examine the anthrapyrazole 1-photosensitized cleavage of DNA at the individual base level employing denaturing polyacrylamide sequencing gels. Smooth sequence neutral photosensitized cleavage of DNA is observed analogous to hydroxyl radical "footprinting."

Sequence specific molecular recognition by a monocationic lexitropsin of the decadeoxyribonucleotide d-[CATGGCCATG]2: structural and dynamic aspects deduced from high field 1H-NMR studies

All 1H-NMR resonances of d-[CATGGCCATG]2 and the 1:1 complex of lexitropsin 1 and the DNA were assigned by the NOE difference, COSY and NOESY methods. Addition of 1 causes the base and imino protons for the sequence 5'-CCAT to undergo the most marked drug-induced chemical shift changes, thereby indicating that 1 is located in this base pair sequence. NOEs confirmed the location and orientation of the drug in the 1:1 complex, with the amino terminus oriented to C(6). The van der Waals interaction between H12a,b of 1 and AH2(8) may be responsible for reading of the 3' A.T base pair in the 5'-CCAT sequence. Exchange NMR effects allow an estimate of approximately equal to 62 s-1 for the intramolecular "slide-swing" exchange of the lexitropsin between two equivalent binding sites with delta G = 58 +/- 5 kJ mol-1 at 301 degrees K.

Structural and dynamic aspects of binding of a prototype lexitropsin to the decadeoxyribonucleotide d(CGCAATTGCG)2 deduced from high-resolution 1H NMR Studies

Structural and dynamic properties of the self-complementary decadeoxyribonucleotide d(CGCAATTGCG)2 and the interaction between a prototype lexitropsin, or information-reading oligopeptide, and the decadeoxyribonucleotide are deduced by using high-resolution 1H NMR techniques. The nonexchangeable and imino proton resonances of d(CGCAATTGCG)2 have been completely assigned by two-dimensional NMR studies. The decadeoxyribonucleotide exists as a right-handed B-DNA. In the 1H NMR spectrum of the 1:1 complex, the selective chemical shifts and removal of degeneracy of AH2(4), AH2(5), T-CH3(6), and T-CH3(7) due to the anisotropy effects of the heterocyclic moieties of the ligand, and with lesser effects at the flanking base sites C(3) and G(8), locate the drug centrally in the decadeoxyribonucleotide. This conclusion is supported by plots of individual chemical shift changes across the decadeoxyribonucleotide. Similarly, imino protons IV and V experience larger shifts and II and III smaller shifts in accord with this conclusion while drug complexation permits the detection of imino proton I. Strong nuclear Overhauser effects (NOEs) between pyrrole H5 and AH2(5), and weaker NOEs to AH1'(5), TH3'(6), and AH2'(5), firmly locate the ligand in the minor groove. Intraligand NOEs between the adjacent heterocyclic moieties indicate that the lexitropsin is subject to propeller twisting about the N6-C9 bond in both the bound and free forms. Nuclear Overhauser effect spectroscopy (NOESY) and correlated spectroscopy (COSY) experiments also indicate that the removal of degeneracy of the C16 methylene protons upon complexation may arise from restricted rotation about the C15-N9, C15-C16, and C16-C17 bonds. Specific hydrogen bonds between amide NH groups on the concave face of the ligand (N4H, N6H, N9H) and adenine N3 or thymine O2 on the floor of the minor groove are in accord with displacement of the hydration shell by the drug. NOE measurements on the decadeoxyribonucleotide in the 1:1 complex confirm it exists as a right-handed helix and belongs to the B family. Exchange NMR effects permit an estimate of a rate of approximately equal to 44 s-1 for the two-site exchange of the lexitropsin between two equivalent sites on the decamer with delta G++ approximately equal to 70 +/- 5 kJ mol-1 at 294 K. Alternative mechanisms for this exchange process are considered.