Indoloquinone EO9: DNA interstrand cross-linking upon reduction by DT-diaphorase or xanthine oxidase.

We report DNA interstrand cross-linking caused by the anti-tumour indoloquinone EO9 following reductive activation with purified rat liver DT-diaphorase or xanthine oxidase. Reduction was a necessary event for cross-linking to occur. DNA cross-link formation by EO9 following DT-diaphorase reduction was completely inhibited by addition 10 microM dicoumarol, whereas only a minor effect of dicoumarol on xanthine oxidase-mediated DNA cross-linking by EO9 was observed. DNA cross-linking was pH dependent, with increasing cross-link formation from pH 5.5 to 7.0 for both DT-diaphorase and xanthine oxidase mediated reactions. Also, conversion of EO9 upon reduction was pH dependent. However, in contrast to DNA cross-linking, conversion rates of EO9 decreased at higher pH. EO9 was shown to be more efficient in DNA cross-linking than mitomycin C under identical conditions, using both DT-diaphorase and xanthine oxidase reductive activation at pH 5.5 and 7.0. This study indicates that the anti-tumour activity of EO9 may be at least partly mediated by interstrand DNA cross-link formation, and that various reducing enzymes may be important for activation of EO9 in vitro and in vivo.

The indoloquinone E09 (see Figure 1) is one of a large group of synthetic quinones based on the clinically used anti-tumour drug mitomycin C (Oostveen and Speckamp, 1987). Because of its activity in solid tumours and lack of bone marrow toxicity it is presently undergoing clinical trials (Hendriks et al., 1993). Like mitomycin C, E09 is expected to be activated bioreductively. Three active centres, i.e. the vinylic group at C-2, the hydroxymethyl group at C-3 and the aziridinyl group at C-5, are possibly activated upon reduction of the quinone group of E09 (Oostveen and Speckamp, 1987). Reduction of E09 can be performed very efficiently with the two-electron reducing enzyme DTdiaphorase [NAD(P)H (quinone acceptor) oxidoreductase, EC 1.6.99.2] (Bailey et al., 1992a). The importance of this enzyme for the anti-tumour activity of E09 is indicated by the correlation that was found between the amount of DTdiaphorase present in a tumour cell and the anti-tumour activity of E09 (Robertson et al., 1992). However, other reducing enzymes may be able to activate E09 as well.
Bioreductive alkylation of DNA has been suggested as the molecular basis of the anti-tumour effect of E09 (Walton et al., 1991(Walton et al., , 1992. Upon reductive activation of E09 by purified DT-diaphorase DNA single-strand breaks were reported. This DNA strand break formation was unaffected by superoxide dismutase, and therefore an alkylating species could be involved in this process (Walton et al., 1991). Moreover, with use of alkaline elution techniques, the presence of DNA cross-links in rat Walker tumour cells was demonstrated, and the involvement of DT-diaphorase was suggested (Bailey et al., 1992b). However, DNA cross-linking by E09 upon reduction has never been demonstrated directly in a cell-free system.
In this paper we report the conversion of E09 upon reduction, using the two-electron reducing enzyme DT-diaphorase as well as xanthine oxidase. Xanthine oxidase, which is normally involved in the oxidation of hypoxanthine to xanthine and further oxidation of xanthine to uric acid, has been reported to be capable of reducing various quinone compounds. In this case, xanthine or NADH can be used as electron donors (Pan et al., 1984;Lusthof et al., 1990). In case NADH is used as co-factor xanthine oxidase mainly functions as a one-electron reducing enzyme (Nakamura and Yamazaki, 1973). Furthermore, we investigated whether activation of E09 with DT-diaphorase or xanthine oxidase could result in DNA interstrand cross-linking, using an ethidium bromide fluorescence assay.
Materials and methods E09 was kindly provided by the EORTC New Drug Development Office, Amsterdam, The Netherlands. Mitomycin C was from Bristol-Myers. Calf thymus DNA and bovine serum albumin (BSA) were obtained from Boehringer Mannheim (Almere, The Netherlands). Xanthine oxidase (grade III), NADH and NADPH were purchased from Sigma (St Louis, MO, USA). Methanol (HPLC quality) was from Westburg (Leusden, The Netherlands). Acetonitrile (HPLC quality) was obtained from Rathburn (Walkerburn, UK). N,N-Dimethylformamide (DMF) was from Baker (Deventer, The Netherlands). 10 mM stock solutions of E09 and mitomycin C in DMF were used. These stock solutions were stored in the dark at 4°C.
Purification of rat liver DT-diaphorase DT-diaphorase was purified from livers from uninduced male Wistar rats (250 g), using Cibacron blue affinity chromatography, essentially as described by Sharkis and Swenson (1989). DT-diaphorase activity was determined at 25°C using DCPIP as electron acceptor. was measured as the dicoumarol-inhibitable conversion of DCPIP, as measured by the absorbance change at 600 nm. In the presence of 50 ZlM dicoumarol, conversion of DCPIP was inhibited by more than 95%. One unit (U) of DT-diaphorase is defined as the amount converting 1 gmol of DCPIP min' under the conditions mentioned above.

Conversion of E09 by DT-diaphorase and xanthine oxidase
The reactions were performed under nitrogen at 25°C. Reaction mixtures contained 100 tM EO9, 0.15 mg ml'l BSA and 75mUml-' DT-diaphorase or 125mUml-1 xanthine oxidase in 0.1 M phosphate buffer, pH 7.5, 6.5 or 5.5. One unit of xanthine oxidase is defined as the amount converting 1 gLmol xanthine min-' in 0.1 M phosphate buffer, pH 7.4, at 25°C. The reaction mixture was purged for 10 min with nitrogen to remove air before the reaction was started by adding NADH or NADPH (final concentration 500 gsM, total volume 1.0 ml). Samples were taken at several time intervals, and were mixed immediately with an equal amount of acetonitrile. After centrifuging for 1 min at 16 000 r.p.m., the conversion of E09 was monitored by high-performance liquid chromatography (HPLC), using a 4.6 x 200 mm Spherisorb S5-ODS2 Cl8 column with UV detection at 280 nm. EO9 and metabolites were eluted with 49.5:49.5:1 (v/v) methanolwater-0.5 M phosphate buffer, pH 7.4, at a flow rate of 0.5 ml min-'. DNA cross-link formation by E09 DNA cross-linking by EO9 at 25°C was assayed at pH 5.5 and 7.0. A solution of 100 gM EO9, 350 fg ml-I calf thymus DNA, 0.15 mg ml lBSA, and 75 mU ml1 DT-diaphorase or 125 mU ml1 xanthine oxidase was purged with nitrogen for 10 min. Subsequently the reduction was started by adding NADH or NADPH to a final concentration of 500 ItM (total volume 1.0 ml). After 30 min the reaction was stopped and the samples were treated as described elsewhere (Maliepaard et al., 1993). Essentially, the assay makes use of the different fluorescence yields of ethidium bromide intercalated in double-standed DNA or attached to single-stranded DNA. The ethidum bromide fluorescence yields before denaturation and 10 min after denaturation of the DNA in the samples were used to calculate a relative measure of the amount of DNA interstrand cross-links formed.

Results
Conversion of E09 upon reduction Conversion of EO9 was followed with HPLC analysis. Incubation of EO9 at pH 7.5 and pH 6.5, without enzyme or co-factor, did not result in any HPLC-detectable conversion of EO9 within 1 h. However, EO9 appeared to be unstable at pH 5.5, with a half-life of approximately 60 and 150 min in 0.1 M phosphate buffer and 0.15 M Tris-acetate buffer respectively. This instability of EO9 at pH 5.5 has also been reported by others (Phillips et al., 1992). At this pH, one metabolite was formed (see Figures 2a and b), which was attributed to the aziridine ring opened product EOSA, the main acid hydrolysis product of E09 (Phillips et al., 1992).
Because increased cytotoxicity of EO9 at lower pH has been reported (Phillips et al., 1992), conversion of E09 following reduction was measured at various pH values. At all pH values used, reduction of EO9 by DT-diaphorase as well as by xanthine oxidase resulted in conversion of EO9. Chromatograms demonstrating the conversion of E09 upon DTdiaphorase reduction at pH 7.5 and pH 5.5 are shown in Figures 2c and d respectively. Xanthine oxidase-mediated reduction of EO9 yielded identical metabolite patterns, as did reactions that were performed in 0.15 M Tris-acetate buffer (data not shown). EO9 conversion rates in phosphate buffer are shown in Table I E09. Conversion of E09 upon DT-diaphorase-mediated reduction was accelerated at lower pH. However, for xanthine oxidase-mediated reduction of E09, after an initial increase in conversion rate observed upon lowering the pH from 7.5 to 6.5, a decrease was noted upon further lowering the pH to 5.5.
DT-diaphorase-mediated conversion of E09 was inhibited completely by omitting NAD(P)H, or by adding 10 liM of the frequently used DT-diaphorase inhibitor dicoumarol (Table  I). E09 conversion by xanthine oxidase was inhibited by omitting NADH. However, adding 10 liM dicoumarol to this reaction mixture resulted in an increased conversion rate of E09 (Table I). This potentiation of xanthine oxidasemediated metabolism by dicoumarol has also been reported for mitomycin C (Gustafson and Pritsos, 1992). Interestingly, upon reduction of E09 at pH 7.5, only small amounts of the aziridinyl ring-opened product EOSA are observed (see Figure 2c). Moreover, the observed amount of EO5A, formed at pH 5.5 (see Figure 2d), can be completely accounted for by non-enzymatic ring opening. Once formed,   Figure 2). These metabolites were not further characterised.
DNA interstrand cross-link formation Before denaturation, the absolute fluorescence yields of control DNA samples and E09-treated DNA samples were essentially equal. Only after denaturation (and 10 min renaturation) did differences in fluorescence yields between control samples and E09-treated samples become apparent. Therefore, it is clear that reduction of E09 by DT-diaphorase or xanthine oxidase in the presence of calf thymus DNA results in interstrand cross-link formation (Table II), whereas loss of DNA owing to E09-induced DNA strand breaks was not observed. For DT-diaphorase as well as for xanthine oxidase-mediated DNA cross-linking by E09, a distinct pH dependency was noticed, with a higher amount of DNA interstrand cross-links formed at higher pH. Using the co-factor NADPH instead of NADH in DT-diaphorase mediated reactions resulted in an equal amount of DNA cross-links formed by E09. No DNA cross-linking by E09 was detected without NADH or NADPH or without enzyme (data not shown). Furthermore, DT-diaphorase-mediated DNA cross-linking by E09 was prevented by adding 1O AM dicoumarol. However, xanthine oxidase-mediated DNA cross-linking was inhibited only for approximately 24% in the presence of dicoumarol (Table II). The acid hydrolysis product EOSA appeared to be a very poor cross-linking agent under the conditions as described in this paper (Table  II). DNA cross-linking by E09 was compared with that by mitomycin C. Mitomycin C is known to inhibit DT-diaphor-ase at pH 7.0 owing to covalent binding to the enzyme (Ross et al., 1993). The relatively small number of DNA cross-links detected at pH 7.0 using DT-diaphorase reduction of mitomycin C (Table II) can be explained by this enzyme inhibition. At pH 5.5, increased DNA cross-linking by mitomycin C was observed. This increased DNA cross-link formation by mitomycin C following DT-diaphorase reduction at lower pH was also noted by others (Siegel et al., 1992;Ross et al., 1993). However, following DT-diaphorase reductive activation at pH 5.5, E09 is clearly a more potent DNA crosslinker than mitomycin C (Table II). Notably, also, xanthine oxidase-mediated reductive activation of E09 yielded more DNA interstrand cross-links than mitomycin C under identical conditions (Table II). However, this difference in DNA cross-linking efficiency was less pronounced than noted following DT-diaphorase reduction.

Discussion
Both reduction with the obligate two-electron reducing enzyme DT-diaphorase and the mainly one-electron reducing enzyme xanthine oxidase results in conversion of E09 and its activation to a DNA cross-linking agent. Comparing DNA cross-linking data with conversion data, the pH dependence of conversion of E09 and DNA cross-linking by it appears to be reversed: lowering pH results in accelerated conversion, but less DNA cross-linking by E09. This holds for both the DT-diaphorase and xanthine oxidase-mediated reactions. Differences between buffers used in conversion and DNA cross-linking experiments do not appear to be crucial, as the metabolite patterns upon reduction and pH dependency of the conversion in 0.15 M Tris-acetate buffer were similar to those obtained in 0.1 M phosphate buffer. Furthermore, the smaller number of cross-links formed at lower pH can only partly be explained by acid hydrolysis of E09, and resulting formnation of the less active EO5A, at this low pH. From the above-mentioned half-life of E09 in 0.15 M Tris-acetate buffer at pH 5.5, it is concluded that in the DNA crosslinking experiment at this pH more than 60% of E09 is still present (or already reduced) after 40 min of incubation. A reversed effect of pH on E09 conversion and DNA crosslinking was also evident in experiments using dicoumarol in xanthine oxidase-mediated reactions: by adding 10 JLM dicoumarol, conversion of E09 in the absence of DNA was enhanced by a factor 3.5 (Table I), whereas DNA crosslinking by E09 diminished by 24% (Table II). Although the exact mechanism of action of E09 is presently unknown, reasons for the deviating pH dependence of conversion and DNA cross-linking can be hypothesised. Firstly, the chemical mechanism responsible for the formation of E09 metabolites may be different in the absence or in the presence of nucleophiles (e.g. DNA). If so, such behaviour can result in different pH profiles for electrophilic and nucleophilic reactions. Such a phenomenon has been described for mitomycin C (Schiltz and Kohn, 1992). Another explanation for the lower amount of DNA cross-links formed at pH 5.5 may be a decreased lifetime of the alkylating intermediates of E09 at lower pH, which would diminish DNA adduct formation. More research is needed to clarify this matter.
Cytotoxicity of E09 has been shown to increase at lower extracellular pH in a human adenocarcinoma DLD-1 and breast carcinoma MCF-7 cell line (Phillips et al., 1992). In contrast to this, our results show diminished DNA cross-link formation at lower pH. Although sensitivity of cells towards DNA cross-linking damage may increase at lower extracellular pH, this could suggest that other mechanisms of action also play a role in the cytotoxicity of E09. A candidate mechanism is induction of single-strand DNA breaks, which have been demonstrated in pBR 322 plasmid DNA following DT-diaphorase reduction of E09 (Walton et al., 1991). The effect of lower pH on induction of DNA single-strand breaks by E09 is not clear at this moment. Notably, DNA interstrand cross-linking is regarded as an important mechanism for anti-tumour activity of bioreductive quinones (Workman, 1992). Therefore, it is interesting that under our conditions E09 is a more potent DNA cross-linking agent than mitomycin C, using DT-diaphorase as well as xanthine oxidase reductive activation. DNA cross-linking by E09 can therefore be regarded as potentially important for anti-tumour activity of this compound. DNA cross-linking by E09 upon reduction by DT-diaphorase is interesting in view of the correlation between DT-diaphorase content and the activity of E09 in tumour cells (Robertson et al., 1992;Walton et al., 1992). This supports the conclusions of these authors that DT-diaphorase reduction may be important for the anti-tumour effect of E09 in vitro and in vivo. Moreover, in some tumours the content of DT-diaphorase is increased several fold, com-pared with non-malignant cells (Cresteil and Jaiswal, 1991). However, our results with xanthine oxidase indicate that other reducing enzymes, the amount of which can also be elevated in certain tumour cells (Nemeikaite and Cenas, 1993), may potentially be activators of E09 to a DNA alkylating species as well. Therefore these enzymes should not be excluded in studying the activation mechanisms of E09.