Potentiation of anti-cancer agent cytotoxicity by the potent poly(ADP-ribose) polymerase inhibitors NU1025 and NU1064.

The ability of the potent poly(ADP-ribose) polymerase (PARP) inhibitor, NU1025 (8-hydroxy-2-methyl-quinazolin-4-[3H]one) to potentiate the cytotoxicity of a panel of mechanistically diverse anti-cancer agents was evaluated in L1210 cells. NU1025 enhanced the cytotoxicity of the DNA-methylating agent MTIC, gamma-irradiation and bleomycin 3.5-, 1.4- and 2-fold respectively. The cytotoxicities of the thymidylate synthase inhibitor, nolatrexed, and the cytotoxic nucleoside, gemcitabine, were not increased. Potentiation of MTIC cytotoxicity by a delayed exposure to NU1025 was equally effective as by a simultaneous exposure to NU1025, indicating that the effects of NU1025 were mediated by an inhibition of the cellular recovery. The recovery from potentially lethal gamma-irradiation damage cytotoxicity in plateau-phase cells was also inhibited by NU1025. Investigation of DNA strand breakage and repair in gamma-irradiated cells by alkaline elution demonstrated that NU1025 caused a marked retardation of DNA repair. A structurally different PARP inhibitor, NU1064 (2-methylbenzimidazole-4-carboxamide), also potentiated the cytotoxicity of MTIC, to a similar extent to NU1025. NU1064 potentiated a sublethal concentration of a DNA methylating agent in a concentration-dependent manner. Collectively, these data suggest that the most suitable cytotoxic agents for use in combination with PARP inhibitors are methylating agents, bleomycin and ionizing radiation, but not anti-metabolites.

bound to nicked DNA miaht stabilize histone residues and maintain nucleosomal structure. thereby keeping the txwo DNA ends correctly positioned for subsequent rejoining (Lindahl et al. 1995).
The possible involvement of PARP in DNA repair has stimulated an interest in its role in determining the response to anticancer therapies that damage DNA. The availabilitv of inhibitors of PARP has greatly aided these studies (Shall. 1984). most of which has-e investigated the potentiation of monofunctional alkylatina agents and ionizing, radiation as these therapies are the most potent actixators of PARP. althoug!h some other agents have also been evaluated (reviewed in Griffin et al. 1995).
Studies of the potentiation of DNA-damaging agents by PARP inhibition have primarilx used the benzamide inhibitors. How ever. the benzamides are not very potent and. although their PARP inhibitonr ICconcentrations are in the low micromolar range. millimolar concentrations are invariably required to achieve potentiation in cvtotoxicitv assays. Furthermore. the benzamides also have other effects. most notably on de novo purine biosynthesis (Cleaver. 1984: Hunting et al. 1985: Milam et al. 1986).
Nevertheless. trans-dominant inhibition of PARP. through overexpression of the DBD. also sensitizes cells to y-irradiation and alkylating acgents (Molinete et al. 1993: Kupper et al. 1995. an observation that is consistent with the effects of the benzamides being primarily caused by PARP inhibition. However. given the high concentrations of benzamides required in cy totoxicity studies. more potent and specific inhibitors of PARP are required in order to assess the clinical potential of PARP inhibitors to improve current cancer therapy. Using rational drug desicgn. two novel series of PARP inhibitors have been developed: the benzimidazole4-carboxamides and 'T'his paper is part 4 of the .senies Resistance Modift-ing Agents.: for par . see Griffin et al 1996h Pharn Si 2: 43-7. prepared as previously described (Grnffin et al 1995(Grnffin et al . 1996. were dissolved in DMSO at 100 imM and stored at -20°C. Gemcitabine (a gift from Eli Lilly. Indianapolis. IN. USA) was dissolved in water at 10 mnm and stored at -20°C. MTIC [5-(3-methyltriazen-lyl)imidazole-4-carboxamide: a gift from Dr C Bleasdale.
University of Newcastle upon Tyne. UK] and bleomycin (Lundbeck. Milton Keynes. UK) were dissolved in DMSO and used immediately. 3-Aminobenzamide (3AB; Pfaltz and Bauer. Phase Separations. Deeside. UK) was dissolved in tissue culture medium (see below) on the day of use by stirring for 4 h and filter sterilized. Drugs were added to cell cultures so that the final DMSO concentration was always 1% (v/v). All other chemicals were obtained from Sigma (Poole. UK) unless stated otherwise. quinazolin4-[3H]-ones. which exhibit much greater potency than the benzamides (Griffin et al. 1995(Griffin et al. . 1996. The potentiation of DNA methylating agent-induced cytotoxicity and DNA strand breakage by one of these novel PARP inhibitors (NU1025: 8-hydroxy-2-methylquinazolin4-[3H]one. Figure IA) has been described previously (Boulton et al. 1995). The aims of the studies reported here were to investigate the effects of NU1025 (IC,o for PARP inhibition = 0.4 .LM) on the cytotoxicity of a range of anticancer agents, and to compare the effects of quinazolinone (NU 1025) and benzimidazole (NU 1064: 2-methylbenzimiidazole-4-carboxamide: IC;,, for PARP inhibition = 1 Mk. Figure 1B) PARP inhibitors on methylating agent-induced cytotoxicity. The cytotoxic treatment studied were methylating agents (MTIC and temozolomide). y-irradiation. bleomycin and anti-metabolites (nolatrexed and gemcitabine). Studies with topoisomerase I and H inhibitors are described elsewhere  and manuscript in preparation)

MATERIALS AND METHODS Drugs
Nolatrexed (a gift from Agouron Pharmaceuticals. San Diego. CA. USA) and temozolomide (a gift from the Cancer Research Campaign. London. UK) were dissolved in dimethyl sulphoxide (DMSO) at O nmi and stored at -200C. NU1025 and NU1064.
Murine leukaemia L 12 10 cells were used to allow comparison with the PARP inhibitory potency of the compounds that had been determined previously using permeabilized L 1210 cells (Griffin et al 1995(Griffin et al . 1996. Cells were maintained as exponentially growing cultures (<8xlI0 cells ml-') in RPMI-1640 medium supplemented with 10% (v/v) fetal calf serum (Sigma. Poole. UK). at 37°C in an atmosphere of 5% carbon dioxide in air. The cell doubling time was approximately 12 h. For experiments with the thymidylate synthase inhibitor. nolatrexed. cells were adapted to grow in RPMI medium supplemented with 10% (v/v) dialysed serum (to remove thymidine) and maintained in this medium for at least 4 weeks before cytotoxicity assays. when the cell doubling time was also approximately 12 h. Cells were tested to exclude mycoplasma contamination (Chen. 1977) every 4-8 weeks.
Cytotoxicity assays L12 10 cells were diluted to l0 ml-' in medium containing the desired concentration of the cytotoxic drug. with or without the PARP inhibitor, in duplicate wells of a six-well plate. After the selected exposure period. cells were harvested by centrifugation (200 g. room temperature) to remove the drugs. resuspended in fresh medium and counted (Coulter Counter Z 1: Coulter Electronics. Luton. UK). Cell suspensions were then further diluted and dispensed into sterile polyurethane tubes (Falcon. Becton Dickinson. Oxford. UK) in triplicate at a suitable density in 1 ml of culture medium (estimated to give 10-60 colonies). and 5 ml of 0.15% (w/v) agarose (SeaKem ME Agarose. Flowgen. Sittingbourne. UK) in culture medium was added. The tubes were incubated for 1-2 weeks to allow colonies to appear and the contents placed in dishes containing 1 ml 0.5 mg ml-' MTT (3-[4.5-dimethylthiazol-2-yl]-2.5-diphenyltetrazolium bromide) to stain the viable colonies. After 4 h colonies were counted and the plating efficiency. relative to the appropriate control (DMSO alone or PARP inhibitor alone). was calculated. DMSO control incubations gave approximately 100% plating efficiency.
In order to investigate if the potentiation of a cytotoxic agent by NU1025 was dependent on the simultaneous presence of the inhibitor. or if NU1025 would be equally effective when administered after the cytotoxic agent had been removed. the following protocol was employed: cells were exposed to varying concentrations of MTIC for 20 min. duplicate samples were then harvested by centrifugation and resuspended in drug-free medium or medium containing 200 gm NU1025 for a further 16 h. In addition. cells were exposed to vary ing concentrations of MTIC in the A a NH CH3 OH 0 Cancer Research Campaign 1996 were exposed to a`Cs source (Gammacell 1000 Elite. Nordian International. Canada) then incubated at 37C. in the presence or absence of 200 im Ni 1025 for 2 h before determining colony formation as described above. To insvestigate the repair of potentially lethal damage. duplicate cultures of cells were held at plateau phase by maintaining, them at a density of 106 cells ml-' in conditioned medium from plateau-phase cells (which ensured a complete cessation of cell division without a decrease in cell viability durinc the exposure period: data not shown) and exposed to 8 Gy of y-irradiation. Cells were kept on ice immediately before and after irradiation. then incubated at 37C in the presence or absence of 200 .iM NUT1025 for the time period indicated in the results. counted and seeded for colony formation in 0.15%c (w/s) agarose as above.
The degree of potentiation [enhancement factor (EF)] produced bv the PARP inhibitor was calculated by comparing the IC9, or ID90 [the concentration (C) of druc or dose (D) of radiation causin, 90% cell death. i.e. a 10% relative plating efficiency] of the crtotoxic agent alone with the ICg) or ID, in the presence of PARP inhibitor.
EFI,, = IC., or ID., controlfIC,, or ID., + PARP inhibitor Figure 2 Effect of NU1025 on the cytotoxicity of MTIC. Cells were exposed to varying concentrations of MTIC for 20 min and resuspended in fresh medium for 16 h (-) or medium containing 200 JM NU1025 for 16 h (=) before seeding for colony formabon. In additon, cells were exposed to varying concentrations of MT1C in the presence of 200 JM NU1025 for 20 min and resuspended in fresh medium containing 200 gm NU1025 for 16 h (:) before seeding for colony formation. Data are mean + s.d. of triplicate colony counts from each of the cell populations exposed in duplicate. from a single representative experiment with pooled data from three experiments given in Table 1 presence of 200 gm NU 1025 for 20 rmn. harvested and resuspended in fresh medium containing 200 im NU 1025 for a further 16 h. After the 16-h incubation cells were counted and seeded for colony formation as described above.
To study the effects of PARP inhibition on y-irradiated cells. cultures were dispensed into sterile plastic bijoux bottles (Bibby Sterilin. Aldershot. UK) and duplicate samples (cooled to 0-4C) Alkaline elution DNA strand break formation in druc-treated or irradiated cells was measured by alkaline elution as previously described (Kohn et al. 1981 ). This technique employs filters that mechanically impede the passage of DNA. such that the rate of elution through the filter is inversely proportional to the length of the DNA strand. Treated cells are co-eluted with intemal standard cells that hase been irradiated immediately before elution in alkaline buffer. Measurement of the proportion of total DNA from the treated cells remaining on the filter compared with the percentage of the total DNA from the intemal standards in each fraction collected gives the standardized elution rate of the experimentally treated cells. Treated cells (40 ml. 2.5 x I0W ml-' ) were labelled wvith 14.8 kBq ml' [2-'4C]thNmidine (1 96 GBq mmol: Amersham International. Amersham. TK) for 24 h. Following a 4 h chase period in fresh medium. duplicate aliquots of intemal standard cells were also added to all filters and both treated and internal standard cells were allowed to settle on to the filters by gravity in ice-cold PBS. A solution of 2% (w/v) sodium dodecyl sulphate (SDS) in 25 mm EDTA pH 10 was added to lyse the cells and the filters were then exposed to 0.5 mg ml' proteinase K in lysis solution for 1 h. The filters were washed three times with 20 mm EDTA pH 10 and eluted with 20 mm EDTA (acid form) + 1% (w/v) SDS adjusted to pH 12.1 with tetrapropylammonium hydroxide (Aldrich. Gillingham. UK). The elution rate was 2 ml h-' and eight 90-min fractions were collected into scintillation vials containing 15 ml of scintillant (Optiphase Hisafe 2, Fisons. Loughborough. UK). The filters were transferred to scintillation vials. baked at 60°C in 0.4 ml of 1 M hydrochloric acid for 1 h then neutralized with 2.5 ml of 0.4 M sodium hydroxide at room temperature for 15 min before adding 15 ml of scintillant and counting with the eluted samples. The total radioactivity collected in the fractions and hence that remaining on the filters was determined. and the proportion of the total retained on the filter for each of the 90 min intervals was calculated. The proportion of the total '4C retained on the filter at each time point was plotted against the proportion of the total 3H retained to standardize the assay against inter-and intra-assay variations in pump efficiency.

Statistical analysis
Comparison of data sets was made by paired or unpaired Student's two-tailed t-test analyses as appropriate using Graphad Instat software (GraphPad Software. San Diego. CA. USA).

Monofunctional alkylating agents
Two monofunctional alkylating agents were used in these experiments: MTIC. a very reactive DNA-methylating agent with a halflife in aqueous medium of 8 min (Shealy and Krauth, 1966). and the clinically used agent, temozolomide, which spontaneously decomposes to yield MTIC with a half-life of 1.24 h in phosphate buffer and 0.42 h in human serum (Stevens et al, 1987). NU1025 has previously been shown to potentiate the cytotoxicity of temozolomide when cells were exposed to both drugs simultaneously (Boulton et al. 1995). In order to investigate the mechanism of NU1025 potentiation. experiments were performed with MTIC as its rapid half-life allows a pulse exposure to the methylating species. Exposure to MTIC and concomitant or sequential exposure to NU1025 was used to determine if potentiation required both drugs to be present at the same time. Cells were exposed for 20 min to MTIC followed by a 16-h recovery period in fresh mean ± s.d. of triplicate coony counts from each of the ceU populabons exposed in duplicate, from a single representative experiment with pooled data from three experiments gven in Table 1 medium: 200 gM NU1025 was either omitted from both incubations. added only to the recovery medium or present during both the exposure and recovery periods. MTIC caused a concentrationdependent decrease in cell survival and the cytotoxicity was increased by NU1025 both when it was co-administered with the MTIC or added after the MTIC had been removed (Figure 2).
Comparison of the IC, data ( also caused a threefold enhancement of cytotoxicity (Table 1). There was no significant reduction in cell survival following exposure to 200 gm NU1025 alone but 200 gM NU1064 reduced survival by approximately 50%; this was accounted for in the calculation of relative plating efficiency (see Methods). The growth inhibition caused by NU1025 alone and the synergistic effect of NU1025 on growth inhibition caused by a fixed concentration of temozolomide has been described previously (Boulton et al, 1995). In the current study the cytotoxicity of increasing concentrations of the benzimidazole PARP inhibitor, NU1064. alone and in combination with the same fixed concentration of temozolomide (100 gM, a concentration causing 10% reduction in cell survival), was investigated. The results were normalized to controls without NUI1064, i.e. 1% DMSO or 100 gIM temozolomide alone. NU1064 alone was slightly cytotoxic. resulting in a 45% reduction in cell survival at 200 gs, the highest achievable concentration (Figure 4). There was a very marked Brinsh Joumral of Cancer (1998) 78(10) PotentiatKon of the cytotoxicity of 100 uM temozolomide by NU1064. Cells were exposed to increasing concentrations of NU1064 in the presence (7) or absence (V) of 100 gM temozokomide for 16 h before seeding for colony formation. (Data norrnalized to DMSO or temozolomide alone controls: exposure to 100 gm temozolomide alone results in approximatety 900% cell survival.) Data are means + s.d. of triplicate colony counts from each of the cell populations exposed in duplicate, from a single representative effect of increasing NU 1064 concentrations on the cytotoxicitv of 100 JIM temozolomide. resulting in a 98% reduction in cell survival at 200 -iM NU1064 (Figure 4). the IC-o for NLT1064 + 100 JIM temozolomide being 46 gtm.

f-Irradiation
Exponentially grow-ing cells exposed to y-irradiation. followed by a 2-h recovery period. display a dose-related reduction in survival (Figrure 5A). If the recovery was in the presence of 200 j.si NU1025 the survival (IC G) was significantly (P<0.05) reduced by a factor of 1.4 ( Figure 5A. Table 1). For companrson, the effect of 10 m-M 3-aminobenzamide (3AB) was also investigated. a concentration chosen to be equipotent with that of NU 1025 on the basis of the inhibition of L 1210 PARP activitv in a permeabilized cell assay (Griffin et al. 1995). The potentiation of y-irradiation by 3AB was less than with NU1025 (EF 9 = 1.2). and only marginally significant (P < 0.1). Furthermore. 10 mm 3AB alone caused an 18 ± 13% reduction in cell survival (compared with 3 ± 5% reduction for NU1025). possibly a reflection of the effect of 3AB on enzymes other than PARP.
To investigate the effect of NUJ1025 and 3AB on the repair of potentially lethal damage. cells were held at plateau phase and exposed to 8 Gy of y-irradiation. Cells were then seeded for colony formation. either without recovery or allowed to recover in the presence or absence of the PARP inhibitor for 2 or 4 h after irradiation. As shown in Figure SB. there was a slight recovery of the cells with time in the absence of PARP inhibitor: however. in the presence of 200 ptM NV 1025 or 10 Xm1 3AB recovery was inhibited. A Exponentially growing cells were exposed to varying doses of -tLirradiation followed by a 2-h recovery period in control medium (0), 200 gM NU1025 or 10 mm 3AB (7) before seeding for colony formation. B Recovery from potentially lethal k-irradiation damage in the presence or absence of NU1025 or 3AB. Plateau-phase cells were exposed to 8 Gy of y-irradiation and allowed to recover for up to 4 h in control medium (0) or medium containing 200 gm NU1025 (.), or 10 mm 3AB (7) before seeding for colony formation.
Data are mean ± s.d. of triplicate colony counts from each of the cell populations exposed in duplicate, from a single representative experiment with pooled data from three expenments of the type shown in A given in Table 1 The effects of NU 1025 on the repair of DNA damaged by y-irradiation was measured by alkaline elution. The elution profiles of cells exposed to y-rays and eluted immediatelv. or after recovery for 2 h in the presence or absence of NU1025. are shown in Figure 6. The elution profiles of cells allowed to recover for 2 h after inradiation British Joumal of Cancer (1998) 78(10) [3H]Thymidine retained Figure 6 Effect of NU1025 on the repair of DNA damage after potentialty lethal y-irradiation damage. Cells were exposed to 8 Gy of y-irradiaton and eluted immediatety (V) or allowed to recover for up to 2 h in control medium (A) or medium containing 200 gM NU1025 ( ) compared with unirradiated control cells eluted with (-) or wrthout (U) the 2-h recovery incubation and unirradiated cells exposed to NU1025 for 2 h (_ ) before determination of DNA strand breakage by alkaline elution. Data are from representative elubon profiles of duplicate samples from a single representative experiment were indistinguishable from profiles of control cells. demonstratin, that complete repair had occurred in the recovery period-The elution profiles of cells exposed to NU1025 dungn the recovery period were intermediate between the profiles of the control cells and cells irradiated and eluted without recovery. indicating that repair A-as hindered by NU1025 but not completely blocked a result consistent with the cvtotoxicitx data. Bleomycin A concentration-dependent increase in cytotoxicity A as observed following a 16-h exposure to the radiomimetic drug, bleomvcin. and cytotoxicio-swas enhanced twofold by co-exposure to NU 1025 ( Figure 7. Table 1): however. the potentiation was only marginallv significant (P < 0.1 ).
Nolatrexed L1210 cells were adapted to growth in dialysed serum as the thvmidine concentration of undialysed serum was sufficient to prevent the cytotoxicity of nolatrexed during a 16-h exposure period (data not shown). Nolatrexed wvas selected for these studies as. unlike classical antifolate thymidylate synthase inhibitors. it does not require active uptak-e or polyglutamation (Webber et al. 1996). In dialysed serum there was a decrease in cell survival followinc a 16-h exposure to increasing concentrations of nolatrexed. However. there was no enhancement of nolatrexed cytotoxicity by co-exposure to NU1025 ( Figure 8A}: the ICSo for nolatrexed alone was 1.57 ± 0.66 Jm. and in the presence of 200JM NU1025 was 2.42±1.42 -i (P>0.1). NU1025 also failed to potentiate the classical antifolate thymidylate synthase inhibitor. CB3717 (data not shown).
[Bleomycin](IU ml') Fgure 7 Effect of NU1025 on the cytotoxicity of bleomycin. Cells were exposed to varying concentrations of bleomycin in the presence (X) or absence (e) of 200 gm NUl 025 for 16 h before seeding for colony formation.
Data are means ± s.d. of triplicate colony counts from each of te cell populations exposed in duplicate, from a single representative experiment Gemcitabine The novel anti-cancer nucleoside analogue gremcitabine WdFdCH is phosphorylated intracellularlv to dFdCDP. vAhich inhibits nrbonucleotide reductase. and dFdCTPC which is incorporated into DNA . L 1 210 cells were exposed to gremcitabine for 16 h. resulting, in concentration-related cvtotoxicity that wvas not potentiated by NUJ1025 ( Figure 8B): the IC, for gaemcitabine alone was 7.9 n-Nw and in the presence of 200 JIm NU 1025 was 8.4 nmt.

DISCUSSION
Investigations of the effects of PARP inhibition on the cvtotoxicitV of DNA-damaging anti-cancer agents have been assisted by the recent development of inhibitors that are markedly more potent (> tenfold). and possibly more specific. than the classical benzamide inhibitors. The primary aim of this study as to evaluate the effects of two representative examples of the recently developed PARP inhibitors on the cv-totoxicitv of a range of different anticancer aaents in order to select suitable agents for preclinical in vivo and. ultimately. clinical studies. In addition. studies were performed to investigate whether these PARP inhibitors functioned by inhibition of repair/recovery. or by a mechanism that requires the cytotoxic agyent and the PARP inhibitor to be present at the same time.
The majonrtx of studies on the potentiation of cvtotoxicity by PARP inhibition have used alky lating (particularly monofunctional alkylating) agents and in the studies described here the greatest potentiation was observed with MTIC (Table 1). MTIC is not clinically useful as it is too reactive. but temozolomide. which breaks down chemically to produce MTIC. is currently undergoing clinical trials (Newlands et al. 1997). Previous studies. also in L12 10 cells (Boulton et al. 1995). have  [Gemcitabine] (nm) Figure 8 Effect of NU1025 on the cytotoxicity of the anti-metabolites nolatrexed and gemcitabine. A Cells were exposed to varying concentrations of nolatrexed in the presence ( ) or absence (-) of 200 gm NU1025 for 16 h before seeding for colony formation. Data are means + s.d. of tripicate colony counts from each of the cell populations exposed in duplicate. from a single representative experiment. B Cells were exposed to varying concentratons of gemcitabine in the presence ( ) or absence (0) of 200 gm NU1025 for 16 h before seeding for colony formation. Data are mean ± s.d. of triplicate colony counts from each of the cell populations exposed in duplicate. from a single experiment increases growth inhibition induced by a low fixed ( 100 JIM) concentration of temozolomide in a concentration-dependent manner. The enhancement of the cytotoxicitx (clonogenic assay) of 100 Im temozolomide by NU1064 (Figure 4) also illustrated PARP inhibitor concentration dependence. The results suggest that the enhancement of temozolomide cvtotoxicitv is related to the degree of PARP inhibition. and that the concentration of the tw-o PARP inhibitors required to reduce cell growth and survival. in the presence of 100 gm temozolomide. by 50%c in the two studies were remarkablx similar (41 and 46 nm for NU1025 and NU1064 respectivelx). Although NU1064 is 2.5-fold less potent than NU1025 in permeabilized L1210 cell PARP inhibition assavs (Gnrffin et al. 1995(Gnrffin et al. . 1996. NU1064 caused a similar potentiation of MTIC cytotoxicitv (threefold) compared with NU,1025 (3.5-fold potentiation). The concentration of NUT1025 mav has-e been in excess of that needed for maximum cy totoxic potentiation as 50 msi N-1025 has been demonstrated to be sufficient to give maximum potentiation of temozolomide growth inhibition (Boulton et al. 1995).
NU 1025 also increased the cytotoxicity of y-irradiation to exponentially arowing L 1210 cells. affecting both the shoulder and the slope of the survival curve ( Figure 5A). Similar results have been observed in exponentially growing V79 cells with 500 Jim PD 128763 (Arundel-Suto et al. 1991). In the studies reported here. a 2-h recovery period ± NU 1025 was used. as preliminary experiments had shown that no greater potentiation Awas achieved bv extending the period to 8 h (data not shown). This obserxation is in agreement with previously reported results showing that 2 h is long, enough to vield near maximum radiosensitization with other PARP inhibitors (Ben-Hur et al. 1985: Arundel-Suto et al. 1991. Alkaline elution studies ( Figure 6) confirm that DNA repair is Xvirtuallx complete 2 h after irradiation. Howuever. in studies on the repair of potentially lethal irradiation damage. the sunriving fraction increased not only during the first 2 h of the recovery period but also continued to increase for a further 2 h (Fiaure Sb). In these studies cells were held in plateau phase (i.e. growth arrested) to allow DNA repair before being 'fixed' by DNA replication.
The cvtotoxicity of bleom cin w as increased tw ofold by NU 1025. which confirms data obtained w-ith the benzamide PARP inhibitors: For example. 1 mrm 3AB has been shown to cause a 2.4fold reduction in the IC; of bleomycin in growth inhibition assays in L1210 cells (Kato et al. 1988). and 3-methoxNbenzamide (2.5 mxt) and 3AB (5 mm) both caused a threefold potentiation of the cytotoxicity of bleomycmn (10 pzg ml') in Chinese hamster oxarv (CHO) cells (Huet and Laval. 1985).
There w as no potentiation of the cvtotoxicity of the thymidy late svynthase (TS) inhibitor nolatrexed. If anvthing. there w-as a modest protection. although this was not significant by statistical analysis.
A lack of potentiation would not be predicted from the suggested role of PARP in base excision repair. and the proposed mechanism of cvtotoxicity of TS inhibitors. TS inhibitor cvtotoxicitx is thought to involve base excision repair as a result of dUTP pool elexations. which lead to extensive uracil misincorporation into DNA. Excision of misincorporated uracil by the base excision repair enzyme. uracil glvcosvlase. leads ultimately to DNA strand breakaae and cell death. the elevation of dUTP pools correlatinc well with DNA strand breakaae in TS-inhibited cells (Curtin et al. 1991). However. consistent with the current results are those of Prise et al (1986). who showed that in HeLa cells the survixing fraction follow ing a 24-h exposure to methotrexate alone w as 0.31 ± 0.02. whereas co-incubation with 5 mM\ 3AB increased this to 0.41 ± 0.10. Similarlv. in a studx in CHO cells. 3 mM\ 3AB had a protective effect on the cytotoxicitv of fluorodeoxvuridine (FUTdR: another TS inhibitor) even though DNA strand break levels were increased (Willmore and Durkacz. 1993). These latter authors proposed that the NAD+ depletion induced by FLTdR may contribute to the cvtotoxicitv of TS inhibitors, and hence the prevention of this depletion by 3AB may exert a protective effect.
In the studies described here. a modest protection from gemcitabine-induced cytotoxicity by NU 1025 wvas also observed. Sphase cells that most actively incorporate the drug are the most sensitive to gemcitabine . and protection against a v-ariety of S-phase acting drugs (hydroxyurea. fluorodeoxvuridine and thioguanine) by 3AB has been reported and attributed to the cytostatic effects of the benzamides (Moses et al. 1988). The growth-inhibitory effects of NU1025 alone in L1210 cells has been described previously: IC for a 48 h expo sure is 410 Im (Boulton et al. 1995). It is possible that the mildly cvtostatic effect of NUL1025 at 200 -iM may hinder entry into Sphase and hence protect cells from agents acting in this phase. such as the anti-metabolites gemcitabine and nolatrexed.
PARP inhibitors are considered to potentiate cytotoxic arents bv inhibiting DNA repair and hence the recovery from genotoxic insult. In order to verify that NU 1025 was acting by repair inhibition. the survival of cells exposed to an alkylating agent or ionizing radiation. with a recovery period in the presence or absence of NU1025. were compared. The highly reactive methylating, agent MTIC was selected in order to achieve rapid induction of DNA damage for a limited period. NUT1025 caused a similar potentiation of MTIC cytotoxicity w-hether it was present from the beginningr of the experiment or added only after the MTIC had been removed. This result indicates that NU 1025 does not interact with MTIC to influence the initial level of DNA damage. but rather that it suppresses recovery. In the case of y-iradiation. both 200 g.m NJ 1025 and 10 mM 3AB inhibited recovery from potentially lethal damage. DNA repair following y-irradiation by NU 1025 was investigated by alkaline elution. and 2 h in drur-free medium at 37°C was sufficient to pernit complete repair of DNA strand breaks. NU1025 significantly retarded the repair of DNA strand breaks but did not block it completely. Taken together these data indicate that NU 1025 is indeed inhibiting the repair of alkylating agent-and y-irradiation-induced DNA damage. and that this effect underlies the potentiation produced by PARP inhibition.
The normal growth and development of PARP knockout mice (Wang et al. 1997). which exhibit normal DNA repair capacity after MNNG exposure. suggests that PARP activity is not necessary for cell survival. However. the use of PARP inhibitors to retard DNA repair and hence increase the cytotoxicitv of certain classes of anti-cancer agrents remains a viable option as the effects of inhibited PARP are arguably different from those of PARP deficiency. This difference has been most clearly illustrated bv the work of Satoh and Lindahl (1992 (Molinete et al. 1993). On the basis of this model it may be no coincidence that during apoptosis PARP is cleaved into two fragments. one comprising the DBD and the other the automodification and catalytic domains (Kaufmann et al. 1993). Cleavage would presumably be analogous in effect to the overproduction of the DBD and the use of PARP inhibitors. i.e. an inhibition of DNA repair. which is presumably counter-productiVe during programmed cell death.
In recent years considerable effort has been devoted to the investigation of the role of PARP and its inhibition in the cvtotoxicitv of anti-cancer acents using a wide variety of molecular biological approaches. as w-ell as inhibitor studies. The precise role of the enzyme still remains to be fully elucidated. but the use of the new generation of potent and structurally different PARP inhibitors mav facilitate the probing of its function. The data presented here. using two novel potent inhibitors that are structurallv different from the classical benzamides. supports many of the studies w-ith the benzamides and conclusions drawn from them. Furthermore. these data suggest that for the preclinical evaluation of PARP inhibitors it would be appropriate to use combinations with alkvlating agents. bleomvcin and ionizinc radiation. but not antimetabolites.