Novel Opportunities for Thymidylate Metabolism As a Therapeutic Target

3029

Novel opportunities for thymidylate metabolism as a therapeutic target

Peter M. Wilson,1 William Fazzone,1 small-molecule inhibitor to dUTPase represents a viable Melissa J. LaBonte,1 Jinxia Deng,2 strategy to improve the clinical efficacy of these mainstay Nouri Neamati,2 and Robert D. Ladner1 chemotherapeutic agents. [Mol Cancer Ther 2008; 7(9):3029–37] 1Department of Pathology, Norris Comprehensive Cancer Center, Keck School of Medicine, and 2Department of Pharmacology and Pharmaceutical Sciences, School of Pharmacy, University of Southern California, Los Angeles, California Introduction The fluoropyrimidine 5-fluorouracil (5-FU) is widely used in the treatment of a range of cancers, including breast Abstract cancers, and cancers of the aerodigestive and gastrointes- For over 40 years, the fluoropyrimidine 5-fluorouracil tinal tract (1). However, 5-FU has had the greatest effect (5-FU) has remained the central agent in therapeutic and is arguably the most successful drug approved to date regimens employed in the treatment of colorectal cancer for the treatment of colorectal cancer. Throughout 50 years and is frequently combined with the DNA-damaging of clinical development, the response rate of advanced agents oxaliplatin and irinotecan, increasing response colorectal cancer chemotherapy using 5-FU and 5-FU-based rates and improving overall survival. However, many combinations has improved from 10% to 15% to 40% to patients will derive little or no benefit from treatment, 50% primarily due to the introduction of efficacious combi- highlighting the need to identify novel therapeutic targets nation partners such as the topoisomerase I inhibitor to improve the efficacy of current 5-FU-based chemother- irinotecan and the platinum agent oxaliplatin and deter- apeutic strategies. dUTP nucleotidohydrolase (dUTPase) mination of optimal drug scheduling and administration catalyzes the hydrolysis of dUTP to dUMP and PPi, pro- (2–4). Novel biological agents, such as the monoclonal viding substrate for thymidylate synthase (TS) and DNA antibodies cetuximab, which targets the epidermal growth synthesis and repair. Although dUTP is a normal interme- factor receptor and bevacizumab an inhibitor of vascular diate in DNA synthesis, its accumulation and misincor- endothelial growth factor, have recently shown additional poration into DNA as uracil is lethal. Importantly, uracil clinical benefit when included in 5-FU-based regimens in misincorporation represents an important mechanism of metastatic colorectal cancer through suppression of recep- cytotoxicity induced by the TS-targeted class of chemo- tor-mediated tumor processes (5, 6). However, despite therapeutic agents including 5-FU. A growing body of these improvements, approximately one-half of patients evidence suggests that dUTPase is an important mediator treated with 5-FU-based therapies will derive no benefit, of response to TS-targeted agents. In this article, we pre- highlighting the need for the identification of novel sent further evidence showing that elevated expression of therapeutic targets and strategies to overcome the frequent dUTPase can protect breast cancer cells from the expan- occurrence of drug resistance. sion of the intracellular uracil pool, translating to reduced 5-FU continues to remain the mainstay of therapeutic growth inhibition following treatment with 5-FU. We regimens employed in the treatment of colorectal cancer therefore report the implementation of in silico drug and other gastrointestinal malignancies. However, 5-FU development techniques to identify and develop small- continues to prove itself an efficacious agent in additional molecule inhibitors of dUTPase. As 5-FU and the oral 5-FU cancers including breast cancer. Two pivotal clinical trials prodrug capecitabine remain central agents in the treat- have recently reported capecitabine to be an efficacious ment of a variety of malignancies, the clinical utility of a combination partner for novel targeted agents approved for breast cancer (7, 8). In this article, we show that induced dUTP nucleotidohydrolase (dUTPase) expression in a Received 3/26/08; revised 6/16/08; accepted 6/19/08. tetracycline (Tet)-repressible MCF-7 breast cancer cell line Grant support: NIH grant R21 5R21CA104796-3. Wright Foundation, suppresses dUTP pool expansion and increases resistance Whittier Foundation, and Concern Foundation. to 5-FU. Importantly, we also show that dUTPase expres- The costs of publication of this article were defrayed in part by the sion in breast cancer specimens shows marked variation payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to similar to our previous observations in colon cancer. In indicate this fact. summary, dUTPase represents an unexploited therapeutic Requests for reprints: Robert D. Ladner, Department of Pathology, Norris target and the identification of effective inhibitors has the Comprehensive Cancer Center, Keck School of Medicine, University of potential to improve the efficacy of 5-FU-based chemo- Southern California, Room 5322, 1441 Eastlake Avenue, Los Angeles, CA 90089. Phone: 323-865-3116; Fax: 323-865-0522. therapies in a wide variety of cancers. To this end, we have E-mail: [email protected] generated a dUTPase pharmacophore model using in silico Copyright C 2008 American Association for Cancer Research. drug development techniques as a means to identify novel doi:10.1158/1535-7163.MCT-08-0280 small-molecule antagonists to dUTPase.

Mol Cancer Ther 2008;7(9). September 2008

Materials and Methods loading. Horseradish peroxidase signal was detected using Compounds and Reagents HyGlo and Hyblot film (Denville Scientific) and developed 5-FU, fluorodeoxyuridine (FUdR), and paclitaxel were on a Hope Micromax film processor (Hope X-Ray). purchased from Sigma. Growth Inhibition Assay Cell Culture MCF-7 pTet-off cells were transfected in the presence or A The human breast adenocarcinoma MCF-7 pTet-off cell absence of 0.5 g/mL doxycycline and growth inhibition line was obtained from BD Clontech and grown in was measured as previously described (9) using CellTiter DMEM supplemented with 10% Tet-approved fetal 96 AQueous One Solution (Promega). Cells were exposed bovine serum (BD Clontech) with penicillin/streptomycin to increasing concentrations of 5-FU for 72 h. Absorbance and sodium pyruvate (Invitrogen). Cells were maintained was measured using a SpectraMax 190 microplate reader in a humidified Forma incubator (Thermoscientific) at (Molecular Devices) at 490 nm, with drug-treated cells j compared with untreated controls set at 100%. Statistical 37 C with 5% CO2. Overexpression of dUTPase significance was determined using a two-tailed unpaired t MCF-7 pTet-off cells were seeded on 6-cm plates, and Student’s test (GraphPad). Immunohistochemistry 3 h after plating, the cells were washed with PBS and fresh Immunohistochemistry using the DUT415 monoclonal growth medium was added. After 24 h, cells were trans- A fected with 2 Ag pTre-Tight:DUT-N for 6 h and washed in antibody (2 g/mL) was conducted on formalin-fixed, PBS and the appropriate medium was added; to suppress paraffin-embedded breast adenocarcinoma tissue samples the inducible expression of dUTPase, doxycycline was using methods as described previously (12). added to growth medium containing Tet-approved fetal bovine serum at a final concentration of 0.5 Ag/mL. Results and Discussion Twenty-four hours post-transfection, cells were plated for 5-FU Mechanism of Action the appropriate assay and allowed to adhere for 24 h before Following entry into the cell, 5-FU is converted to its medium containing 5-FU, FUdR, or paclitaxel was added. active metabolite, fluoro-dUMP whose primary mechanism Overexpression of dUTPase was confirmed using both of action is inhibition of thymidylate synthase (TS) by Western blotting and enzyme activity assay. formation of a ternary complex with the methyl cofactor dUTPase ActivityAssay 5,10-methylene tetrahydrofolate. This blocks the de novo Cells were harvested and protein was isolated and synthesis of thymidylate resulting in perturbations in A quantified as per Western blotting. Total protein (25 g) nucleotide pools and severe disruption of DNA synthesis A was normalized to a 20 L reaction volume with PBS/ and repair and ultimately leads to lethal DNA damage (1). protease inhibitor. Relative dUTPase activity was deter- Additional mechanisms of action include the incorporation mined as described previously (9) and is expressed as fold of toxic fluoronucleotides into both DNA and RNA and the change compared with an identical transfection in the expansion of the intracellular dUTP pool and subsequent A presence of 0.5 g/mL doxycycline. misincorporation into DNA (1). The 5-FU prodrug capeci- dUTP Accumulation Assay tabine has the convenience of oral administration and has MCF-7 pTet-off cells were treated with specified concen- shown equivalent efficacy as both a single agent and in trations of 5-FU, FUdR, and paclitaxel for indicated times combination with oxaliplatin for the first-line treatment of Â 6 and harvested, and 3 10 cells were analyzed for metastatic colorectal cancer (13, 14). Capecitabine is nucleotide pool content using the assay developed by absorbed intact through the gastrointestinal mucosa where Sherman and Fyfe (10) modified to detect levels of TTP and it undergoes a three-step enzymatic conversion to 5-FU dUTP by preincubating extracts with recombinant dUT- exerting similar mechanisms of action (15). Pase (9, 11). Radioactive incorporation, measured in the Resistance to 5-FU presence of dUTPase, represented the TTP pool, whereas A significant number of studies have correlated intra- untreated extracts represented both dUTP and TTP pools. tumoral TS levels with response to fluoropyrimidine-based dUTP accumulation was determined by subtracting the therapy (16–18). A study by Salonga et al. showed for the results of extracts treated with dUTPase from untreated first time that patients with colorectal cancer who extracts and presented as percent accumulation in histo- responded to 5-FU therapy could be segregated by analysis gram format. Statistical significance was determined using of three genes involved in 5-FU metabolism; patients with a two-tailed unpaired Student’s t test (GraphPad). low expression levels of TS, thymidine phosphorylase, and Antibodies and Western Blotting dipyrimidine dehydrogenase all responded to treatment At specified time points, cells were collected and and those with elevated expression in at least one of the analyzed by Western blot as described previously (9). three genes did not (18). A meta-analysis done by Popat Western blots were probed overnight at 4jC with affinity- et al. analyzed 20 independent studies consisting of over purified anti-dUTPase generated in our laboratory (1:500) 3,000 patients and concluded that those with elevated TS and 2 h with appropriate secondary antibodies (goat anti- expression showed poorer overall survival compared with mouse and goat anti-rabbit horseradish peroxidase). Blots tumors expressing low levels (19). Functional genomic were reprobed for anti-h-actin (Sigma) to control for polymorphisms have been identified within the 5¶-region

Mol Cancer Ther 2008;7(9). September 2008

and the 3¶-untranslated region of the TS gene and have been in vitro and in vivo. Overexpression of dUTPase is reported associated with response to fluoropyrimidine treatment to abrogate the expansion of intracellular dUTP following through TS mRNA transcription and stability (20–23). TS-inhibition, providing substrate for TS in the form of Although conflicting results have been reported in the dUMP and preventing DNA damage associated with uracil literature (24), higher TS protein and mRNA expression misincorporation. In addition, dUTPase is reported to in tumors has generally been associated with poor clini- target fluoro-dUTP (FdUTP) for catalysis, thus preventing cal outcome in patients treated with 5-FU-based chemo- the accumulation of DNA damage reported to occur therapy (19). following FdUTP misincorporation into DNA (Fig. 1; Additional enzymes involved in 5-FU metabolism have refs. 29–31). been associated with response to treatment. Dipyrimidine We reported previously that diminished dUTPase dehydrogenase catalyses the rate-limiting step in the expression enhanced dUTP pool expansion following catabolism of 5-FU with >80% of 5-FU degraded by dipyri- treatment with a TS inhibitor, sensitizing yeast cells to the midine dehydrogenase in the liver. Variation in dipyrimi- effects of uracil misincorporation, whereas cells over- dine dehydrogenase expression therefore has direct effect expressing dUTPase were significantly resistant (32). In on 5-FU bioavailability (25). A previously mentioned study colon cancer cells, induced expression of dUTPase was by Salonga et al. showed that colorectal cancer patients who reported to confer resistance to FUdR (33), whereas our responded to 5-FU therapy had low expression levels of laboratory has shown previously that silencing dUTPase TS, thymidine phosphorylase, and dipyrimidine dehydro- expression using small interfering RNA sensitized both genase and nonresponders had elevated expression in at breast and colon cancer cell lines to FUdR through dUTP least one of the three genes (18). pool expansion, uracil misincorporation, and enhanced Role of dUTPase in 5-FU Resistance DNA fragmentation (9). Moreover, we also reported the dUTPase is the sole enzyme responsible for the hydro- results of a small retrospective clinical study negatively lysis of dUTP to dUMP and pyrophosphate, providing correlating elevated nuclear expression of dUTPase with substrate for TS and eliminating dUTP from the DNA response to 5-FU-therapy in colorectal cancer patients, biosynthetic pathway. Although dUTP is a normal inter- bolstering the concept that dUTPase represents an attrac- mediate in DNA synthesis, the extensive expansion of the tive drug target (12). The FOCUS clinical trial analyzed dUTP pool and subsequent uracil misincorporation into dUTPase expression (high versus low) in 846 advanced DNA is lethal in both prokaryotic and eukaryotic organ- colorectal cancer patients randomized to 5-FU alone, 5-FU isms as shown from knockout models (26). Importantly, + oxaliplatin, or 5-FU + irinotecan and showed a negative uracil misincorporation is a significant mechanism of association between elevated dUTPase and response to cytotoxicity induced by TS-inhibiting chemotherapeutic therapy and time to tumor progression (34). Validation in agents including 5-FU and FUdR (1, 27, 28). Expression of an independent 486 patient follow-up analysis showed that dUTPase is therefore reported to be an important deter- dUTPase retained an association with time to tumor minant of cytotoxicity induced by agents that target TS both progression (35).

Figure 1. Mechanism of 5-FU-induced DNA damage. 5-FU is converted to active metabo- lites: fluoro-dUMP and FdUTP. Fluoro-dUMP binds to and inhibits the enzyme TS by formation of a ternary complex with the methyl donor cofactor. Inhibition of TS induces a metabolic blockade, resulting in depletion of thymidylate and the accumulation of dUMP, which can be phosphorylated to dUTP. When dUTP pools expand, dUTPase activity can become saturated resulting in uracil misincorporation into DNA. In tumors with elevated expression of dUTPase, saturation of enzyme activity is unlikely to occur and dUTP misincorporation is eliminated as a mechanism of cytotoxicity. Generation of FdUTP and subsequent incorporation into DNA has also been shown to induce significant cytotoxicity and dUTPase has been shown to possess affinity for FdUTP, catalyzing its hydrolysis and preventing the misincorporation of FdUTP into DNA.

Mol Cancer Ther 2008;7(9). September 2008

Figure 2. Variation in dUTPase expression in caner cell lines. Western blot analysis of dUTPase expression in a panel of colon (SW620, HCT-8, HCT116, HT29, and LoVo) and gastric (AGS) cancer cell lines. Mitochondrial and nuclear isoforms are indicated. h- Actin was used to control for loading. dUTPase relative enzyme activity assay from corresponding Western blot cell lysates. Mean F SE of corresponding Western lysates analyzed in duplicate.

Dysregulation and variation in expression of dUTPase is Elevated dUTPase Protects Breast Cancer Cells from observed in many cancer cell lines (Fig. 2) and tumor 5-FU specimens including tumor types frequently treated with Our laboratory has shown previously in the MCF-7 breast agents that target TS; thus, validation of dUTPase expres- cancer model that diminishing dUTPase expression sion as a marker of resistance to TS-directed chemotherapy, resulted in a significant increase in dUTP pools and such as 5-FU, is of clinical interest (12, 28, 36). enhanced cytotoxicity induced by the fluoropyrimidine

Figure 3. Elevated dUTPase protects breast cancer cells from 5-FU. A, Western blot and analysis of dUTPase expression and corresponding dUTPase enzyme activity assay following 72 h transfection with pTre-Tight:DUT-N in the presence and absence of 0.5 Ag/mL doxycycline. Mean F SE of corresponding Western lysates analyzed in duplicate. B, analysis of dUTP accumulation in pTre-Tight:DUT-N-transfected MCF-7 pTet-off cells with basal (+dox) and induced (Àdox) dUTPase expression and treated with 10 Amol/L 5-FU, 1 Amol/L FUdR, and 5 nmol/L paclitaxel for 24 h. % dUTP was calculated as described in Materials and Methods. Mean F SE of two independent treatments. c, dUTP not detected. C, growth inhibition measured by MTS assay in pTre-Tight:DUT-N-transfected MCF-7 pTet-off cells following treatment with increasing concentrations (in Amol/L) of 5-FU for 72 h. Mean F SE of three independent treatments. *, P < 0.05; **, P < 0.01.

Mol Cancer Ther 2008;7(9). September 2008

Figure 4. Immunohistochemical analysis of dUTPase in human breast adenocarcinoma. Formalin-fixed, paraffin-embedded breast cancer tumor specimens were routinely processed and stained using the DUT415 monoclonal antibody. A, low dUTPase expression. B and C, elevated nuclear and cytoplasmic expression. Magnifica- tion, Â20 (A and B) and Â40 (C).

FUdR. As an extension to this, we sought to determine parental cells, where depletion of dUTPase correlated with if overexpression of dUTPase could significantly protect increased dUTP pool expansion, greater DNA double-strand MCF-7 cells from the expansion of dUTP pools and reduce break formation, and enhanced chemosensitivity to the growth-inhibitory effects of 5-FU. We used a Tet- FUdR. These data further validate the significance of the repressible dUTPase expression construct and overex- uracil-DNA pathway as a potential contributory mechanism pressed dUTPase with an approximately 4-fold increase of cytotaxicity to 5-FU-based therapies in breast cancer. in expression and enzymatic activity in transfected Tet- Novel Fluoropyrimidine-Based Strategies in Breast inducible MCF-7 pTet-off cells (Fig. 3A). We show that Cancer cells that overexpress dUTPase were significantly more 5-FU has proven itself as the mainstay agent in resistant to 5-FU-induced dUTP pool expansion with a 37% therapeutic combinations used to treat colorectal cancer. reduction in dUTP detected compared with control trans- More recently, the 5-FU prodrug capecitabine has shown fected cells with basal dUTPase. We also used FUdR, a efficacious combinations with novel agents approved for a more TS-directed fluoropyrimidine and observed a similar variety of different cancers and has recently been of 38% reduction in dUTP pool expansion in the presence of particular interest in the treatment of chemorefractory increased dUTPase expression when compared with basal breast cancer. A recent clinical trial showed improved expression. As a control, we used paclitaxel, which inhibits efficacy when capecitabine was used in combination with cell growth by binding to dynamic microtubules (37) and the novel microtubule-stabilizing agent ixabepilone when show that treatment with a non-TS-directed agent resulted compared to capecitabine alone in heavily pre-treated in no detection of dUTP in the presence of basal or induced metastatic breast cancer patients (8). In a HER2-expressing dUTPase expression (Fig. 3B). patient population, capecitabine, in combination with the We subsequently investigated the effects of dUTPase dual epidermal growth factor receptor/HER2 tyrosine overexpression on growth inhibition following exposure kinase inhibitor lapatinib, showed improved progression of MCF-7 pTet-off cells to increasing concentrations of free survival in heavily pretreated metastatic breast cancer 5-FU. Despite the inherent resistance of MCF-7 cells to 5-FU patients (7). These clinical trials show that patients who A (>10 mol/L IC50), overexpression of dUTPase resulted in a previously failed multiple lines of therapy, which included statistically significant decrease in growth inhibition anthracyclines and taxanes, still derive benefit from a compared with cells expressing basal dUTPase at all con- fluoropyrimidine-based therapy. Although only prelimi- centrations of 5-FU tested (Fig. 3C). These preliminary data nary evidence exists regarding resistance to capecitabine, confirm and extend our previous observations in MCF-7 the distinct overlap in mechanism of action with 5-FU

Mol Cancer Ther 2008;7(9). September 2008

would imply that similar mechanisms of resistance should of this important class of TS-targeted chemotherapeutic exist for both agents; as a result, dUTPase expression agents. In addition to the pathway driven rationale for should be investigated as a marker of response. dUTPase as a therapeutic target, several other practical Dysregulation of dUTPase Expression in Breast advantages make this enzyme amenable to drug develop- Adenocarcinoma ment: (a) dUTPase has a well-characterized, nonredundant Previously, we did immunohistochemistry analyzing cellular function, (b) crystallographic data of enzyme- dUTPase expression and localization in colon cancer substrate complexes are available at high resolution (38), specimens from patients treated with 5-FU-based therapy. and (c) the biochemical consequences of dUTPase inhibi- This study negatively correlated elevated expression of tion have measurable outcomes that can be used to validate intratumoral dUTPase with response to 5-FU therapy (12). the mechanism of drug action. We therefore did immunohistochemistry on breast adeno- The in silico approach to inhibitor design is a useful carcinoma tissue using the DUT415 monoclonal antibody to technology that uses existing structural data from crystal- investigate both dUTPase expression and intracellular lographic and site-directed mutagenesis studies to identify localization. Interestingly, similar to our previous observa- lead inhibitory compounds with optimum potency, selec- tions in colon cancer specimens, we report that dUTPase tivity, and/or pharmacokinetic properties. Pharmacophore shows significant variation in total expression and locali- models are used as queries to cull structurally diverse zation. Three representative photomicrographs were molecules that may potentially show requisite inhibitory obtained following immunohistochemistry analysis and activity from existing chemical databases. Such an ap- show breast adenocarcinomas with low expression (Fig. A) proach allows for rational design of inhibitory molecules to and elevated nuclear and cytoplasmic expression of a therapeutic target by enabling establishment of quantita- dUTPase (Fig. 4B and C). This study shows that the pre- tive-structure activity relationships between potential viously reported variation in dUTPase expression in colon molecules and their inhibitory ability (39). cancer specimens extends to other neoplastic tissues To identify potential inhibitors to human dUTPase, the including breast cancer, in which fluoropyrimidine-based structure-based focusing module equipped in the Cerius2 therapies are routinely implemented. software package (Accelrys) was used to efficiently gene- In silico Identification of dUTPase Small-Molecule rate a pharmacophore model by mapping the functional Antagonists features of the dUTPase active site from the crystal Evidence suggests that the accumulation of excessive structure (1Q5H) of recombinant human dUTPase enzyme dUTP pools and its iterative misincorporation into DNA bound to a dUDP substrate (Dud778). We defined the may be a key mediator of cytotoxicity subsequent to binding site centered at the average position of Dud778, a inhibition of de novo thymidylate biosynthesis. As the sole ligand cocrystallized with human dUTPase in the X-ray regulator of dUTP pools, dUTPase represents a novel structure, and with a radius of 15 A˚ covering the ligand- intervention point that could improve the clinical efficacy binding cavity. Catalyst software (Accelrys) package was

Figure 5. Shape-merged pharmacophore model derived from Dud778. A, Dud778 mapping onto the feature model. Green sphere, H-bond acceptor; magenta sphere, H-bond donor. B, shape-merged feature model of Dud778 mapping onto the shape query. Gray area, shape constraint generated from the crystallized conformation. C, Dud778 mapping onto the shape-merged feature model as shown in B.

Mol Cancer Ther 2008;7(9). September 2008

Figure 6. Docking validation of Dud778-dUTPase cocrystal structure. A, Dud778 as observed in the X-ray structure (1Q5H; green). The atom-type colored conformation of Dud778 is predicted by Genetic Optimization for Ligand Docking. B, superimposition of a representative compound, colored by atom type, with X-ray determined Dud778 conformation (green). The compound was selected from database screening that favorably interacts with the ligand binding domain by efficiently filling the deep cavity.

employed to map the functional features (H-bond donor, ranked the candidate antagonists according to the pre- H-bond acceptor, hydrophobic feature, or aromatic ring) dicted docking scores. As a result, those candidate onto the ligand (Dud778) according to the active confirma- molecules displaying inefficient contact with the dUTPase tion observed in the X-ray structures. To develop the active site were not tested in the enzyme assay. The most feature model, geometrical constraints were assigned to favorable docking conformation was nearly identical to the each feature, and the selected features were merged to a X-ray structure, indicating that surface docking is an single model (Fig. 5). The resulting pharmacophore model efficient approach to sampling the ligand-protein interac- was then queried against our in-house compound database tions in the dUTPase system (Fig. 6). The binding pocket is for potential dUTPase antagonists. Docking simulation characterized by a deep cavity localized by the uracil base approaches have proven useful in accurately predicting motif and a channel filled by phosphate fragment. As a small-molecule binding interactions with a receptor. To final selection step, ADMETPredictor (Simulations Plus) further refine selection of compounds for testing in our software was used to calculate the ADMET (absorption, in vitro enzymatic assay, we used Genetic Optimization for distribution, metabolism, excretion, and toxicity) properties Ligand Docking (Cambridge Crystallographic Data Centre, of candidate molecules. version 2.1) software to perform docking studies on our 24- The pharmacophore-based search of a database of processor Silicon Graphics Onyx workstation (40) and approximately 360,000 small molecules yielded an initial

Table 1. Selected compounds identified using the dUTPase pharmacophore model

Compound ID Genetic Optimization for Ligand Docking score* Molecular weight % dUTPase inhibitionc

DU7 70.4 443.52 55.72 DU6 64.5 523.99 46.24 BB37 66.4 418.5 33.98 DU44 78.9 680.86 28.46 DU202 56.6 494.54 Inactive DU53 56.5 535.68 Inactive

*Docking score calculated using Genetic Optimization for Ligand Docking software docking simulations. cPercentage inhibition of dUTPase enzymatic catalysis by respective compound at a fixed concentration of 100 Amol/L. Activity assay was done as described in Materials and Methods.

Mol Cancer Ther 2008;7(9). September 2008

set of 400 structurally diverse compounds possessing irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med 2004;350:2335 – 42. drug-like physicochemical properties. These molecules 6. Van Cutsem E, Nowacki MP, Lang I, et al. Randomized phase III study were tested in our dUTPase activity assay at an initial of irinotecan and 5-FU/FA with or without cetuximab in the first-line 100 Amol/L in vitro to determine their percentage inhibition treatment of patients with metastatic colorectal cancer (mCRC): the CRYSTAL trial. J Clin Oncol 2007; ASCO Annu Meet Proc Part I 2007;25 of dUTPase catalysis relative to uninhibited control. Table 1 [abstract 4000]. illustrates the docking fitness scores of the top four active 7. Geyer CE, Forster J, Lindquist D, et al. Lapatinib plus capecitabine compounds and their ability to inhibit dUTPase activity. for HER2-positive advanced breast cancer. N Engl J Med 2006;355: Ranked according to inhibitory activity, two representative 2733 – 43. inactive molecules are also shown with their corresponding 8. Thomas ES, Gomez HL, Li RK, et al. Ixabepilone plus capecitabine for metastatic breast cancer progressing after anthracycline and taxane docking scores, indicating a reasonable correlation between treatment. J Clin Oncol 2007;25:5210 – 7. predicted docking fitness values and the dUTPase inhibi- 9. Koehler SE, Ladner RD. Small interfering RNA-mediated suppression of tory potential of these small molecules. Lead compounds dUTPase sensitizes cancer cell lines to thymidylate synthase inhibition. with promising inhibitory activity are currently undergoing Mol Pharmacol 2004;66:620 – 6. further optimization in structure-activity studies in an 10. Sherman PA, Fyfe JA. Enzymatic assay for deoxyribonucleoside in vitro triphosphates using synthetic oligonucleotides as template primers. Anal effort to develop more potent analogues that may Biochem 1989;180:222 – 6. also show clinical utility as dUTPase inhibitors in cellular 11. Horowitz RW, Zhang H, Schwartz EL, Ladner RD, Wadler S. cancer disease models. Measurement of deoxyuridine triphosphate and thymidine triphosphate Future of 5-FU in the extracts of thymidylate synthase-inhibited cells using a modified DNA polymerase assay. Biochem Pharmacol 1997;54:635 – 8. Fifty-one years since Charles Heidelberger first described 12. Ladner RD, Lynch FJ, Groshen S, et al. dUTP nucleotidohydrolase the antitumor activity of the fluoropyrimidines, many tens isoform expression in normal and neoplastic tissues: association with of thousands of cancer patients worldwide have benefited survival and response to 5-fluorouracil in colorectal cancer. Cancer Res from chemotherapy stemming from those original obser- 2000;60:3493 – 503. vations (41). 5-FU and other fluoropyrimidine derivatives 13. Van Cutsem E, Twelves C, Cassidy J, et al. Oral capecitabine compared with intravenous fluorouracil plus leucovorin in patients with continue to show clinical efficacy in a wide variety of metastatic colorectal cancer: results of a large phase III study. J Clin Oncol cancers and remain central agents for the treatment of 2001;19:4097 – 106. colorectal cancer and an effective agent in refractory breast 14. Tyagi P, Grothey A. Commentary on a phase III trial of bevacizumab plus XELOX or FOLFOX4 for first-line treatment of metastatic cancer. However, a large number of patients do not colorectal cancer: the NO16966 trial. Clin Colorectal Cancer 2006;6: respond to 5-FU-based treatment strategies and progress 261 – 4. remains hindered by the lack of predictive markers to these 15. Miwa M, Ura M, Nishida M, et al. Design of a novel oral therapies. Therefore, identification and exploitation of fluoropyrimidine carbamate, capecitabine, which generates 5-fluorouracil selectively in tumours by enzymes concentrated in human liver and cancer novel alternative targets within this pathway of proven tissue. Eur J Cancer 1998;34:1274 – 81. therapeutic utility is of considerable importance. As cancer 16. Johnston PG, Lenz HJ, Leichman CG, et al. Thymidylate synthase treatment undergoes a paradigm shift toward individual- gene and protein expression correlate and are associated with response to ized treatment strategies, the identification and exploitation 5-fluorouracil in human colorectal and gastric tumors. Cancer Res 1995; 55:1407 – 12. of novel targets such as dUTPase introduces the possibility 17. Lenz HJ, Hayashi K, Salonga D, et al. p53 point mutations and of delivering a more effective 5-FU-based chemotherapy, thymidylate synthase messenger RNA levels in disseminated colorectal increasing response rates in advanced and metastatic cancer: an analysis of response and survival. Clin Cancer Res 1998;4: disease and enabling more patients to undergo complete 1243 – 50. surgical resection with curative intent. 18. Salonga D, Danenberg KD, Johnson M, et al. Colorectal tumors responding to 5-fluorouracil have low gene expression levels of dihydro- pyrimidine dehydrogenase, thymidylate synthase, and thymidine phosphorylase. Clin Cancer Res 2000;6:1322 – 7. Disclosure of Potential Conflicts of Interest 19. Popat S, Matakidou A, Houlston RS. Thymidylate synthase expres- R.D. Ladner: consultant, Eisai; grant recipient, GlaxoSmithKline. No other sion and prognosis in colorectal cancer: a systematic review and meta- potential conflicts of interest were disclosed. analysis. J Clin Oncol 2004;22:529 – 36. 20. Horie N, Takeishi K. Functional structure of the promoter region of the human thymidylate synthase gene and nuclear factors that regulate the expression of the gene. Nucleic Acids Symp Ser 1995;34:77 – 8. References 21. Pullarkat ST, Stoehlmacher J, Ghaderi V, et al. Thymidylate synthase 1. Longley DB, Harkin DP, Johnston PG. 5-fluorouracil: mechanisms of gene polymorphism determines response and toxicity of 5-FU chemother- action and clinical strategies. Nat Rev Cancer 2003;3:330 – 8. apy. Pharmacogenomics J 2001;1:65 – 70. 2. Douillard JY, Cunningham D, Roth AD, et al. Irinotecan combined with 22. Mandola MV, Stoehlmacher J, Muller-Weeks S, et al. A novel single fluorouracil compared with fluorouracil alone as first-line treatment for nucleotide polymorphism within the 5¶ tandem repeat polymorphism of the metastatic colorectal cancer: a multicentre randomised trial. Lancet 2000; thymidylate synthase gene abolishes USF-1 binding and alters transcrip- 355:1041 – 7. tional activity. Cancer Res 2003;63:2898 – 904. 3. Giacchetti S, Perpoint B, Zidani R, et al. Phase III multicenter 23. Mandola MV, Stoehlmacher J, Zhang W, et al. A 6 bp polymorphism randomized trial of oxaliplatin added to chronomodulated fluorouracil- in the thymidylate synthase gene causes message instability and is leucovorin as first-line treatment of metastatic colorectal cancer. J Clin associated with decreased intratumoral TS mRNA levels. Pharmacoge- Oncol 2000;18:136 – 47. netics 2004;14:319 – 27. 4. Tournigand C, Andre T, Achille E, et al. FOLFIRI followed by FOLFOX6 24. Edler D, Glimelius B, Hallstrom M, et al. Thymidylate synthase or the reverse sequence in advanced colorectal cancer: a randomized expression in colorectal cancer: a prognostic and predictive marker of GERCOR study. J Clin Oncol 2004;22:229 – 37. benefit from adjuvant fluorouracil-based chemotherapy. J Clin Oncol 5. Hurwitz H, Fehrenbacher L, Novotny W, et al. Bevacizumab plus 2002;20:1721 – 8.

Mol Cancer Ther 2008;7(9). September 2008

25. Diasio RB, Harris BE. Clinical pharmacology of 5-fluorouracil. Clin expression of Escherichia coli deoxyuridinetriphosphatase. Cancer Res Pharmacokinet 1989;16:215 – 37. 1994;54:2296 – 8. 26. el-Hajj HH, Zhang H, Weiss B. Lethality of a dut (deoxyuridine 34. Adlard JW, Richman S, Royston P, et al. Assessment of multiple triphosphatase) mutation in Escherichia coli. J Bacteriol 1988;170: markers for association with response rate (RR) and failure-free survival 1069 – 75. (FFS) in patients with advanced colorectal cancer (CRC) treated with 27. Curtin NJ, Harris AL, Aherne GW. Mechanism of cell death following chemotherapy in the MRC CR08 (FOCUS) randomized trial. J Clin Oncol thymidylate synthase inhibition: 2¶-deoxyuridine-5¶-triphosphate accumu- 2004; ASCO Annu Meet Proc 2006;22 [abstract 9506]. lation, DNA damage, and growth inhibition following exposure to CB3717 35. Richman S, Braun MS, Adlard JW, et al. Prognostic value of and dipyridamole. Cancer Res 1991;51:2346 – 52. thymidylate synthase (TS) expression on failure-free survival of fluoroura- 28. Webley SD, Hardcastle A, Ladner RD, Jackman AL, Aherne GW. cil-treated metastatic colorectal cancer patients. J Clin Oncol 2006; ASCO Deoxyuridine triphosphatase (dUTPase) expression and sensitivity to Annu Meet 2006;24 [abstract 10011]. the thymidylate synthase (TS) inhibitor ZD9331. Br J Cancer 2000;83: 36. Fleischmann J, Kremmer E, Muller S, et al. Expression of deoxyuridine 792 – 9. triphosphatase (dUTPase) in colorectal tumours. Int J Cancer 1999;84: 29. Ingraham HA, Tseng BY, Goulian M. Nucleotide levels and incor- 614 – 7. poration of 5-fluorouracil and uracil into DNA of cells treated with 37. Dumontet C, Sikic BI. Mechanisms of action of and resistance to 5-fluorodeoxyuridine. Mol Pharmacol 1982;21:211 – 6. antitubulin agents: microtubule dynamics, drug transport, and cell death. 30. An Q, Robins P, Lindahl T, Barnes DE. 5-Fluorouracil incorporated into J Clin Oncol 1999;17:1061 – 70. DNA is excised by the Smug1 DNA glycosylase to reduce drug 38. Mol CD, Harris JM, McIntosh EM, Tainer JA. Human dUTP cytotoxicity. Cancer Res 2007;67:940 – 5. pyrophosphatase: uracil recognition by a h hairpin and active sites formed 31. Caradonna SJ, Cheng YC. The role of deoxyuridine triphosphate by three separate subunits. Structure 1996;4:1077 – 92. nucleotidohydrolase, uracil-DNA glycosylase, and DNA polymerase a in 39. Neamati N, Barchi JJ, Jr. New paradigms in drug design and the metabolism of FUdR in human tumor cells. Mol Pharmacol 1980;18: discovery. Curr Top Med Chem 2002;2:211 – 27. 513 – 20. 40. Jones G, Willett P, Glen RC, Leach AR, Taylor R. Development and 32. Tinkelenberg BA, Hansbury MJ, Ladner RD. dUTPase and uracil-DNA validation of a genetic algorithm for flexible docking. J Mol Biol 1997;267: glycosylase are central modulators of antifolate toxicity in Saccharomyces 727 – 48. cerevisiae. Cancer Res 2002;62:4909 – 15. 41. Heidelberger C, Chaudhuri NK, Danneberg P, et al. Fluorinated 33. Canman CE, Radany EH, Parsels LA, et al. Induction of resistance to pyrimidines, a new class of tumour-inhibitory compounds. Nature 1957; fluorodeoxyuridine cytotoxicity and DNA damage in human tumor cells by 179:663 – 6.

Mol Cancer Ther 2008;7(9). September 2008

Peter M. Wilson, William Fazzone, Melissa J. LaBonte, et al.

Mol Cancer Ther 2008;7:3029-3037.

Updated version Access the most recent version of this article at: http://mct.aacrjournals.org/content/7/9/3029

Cited articles This article cites 37 articles, 20 of which you can access for free at: http://mct.aacrjournals.org/content/7/9/3029.full#ref-list-1

Citing articles This article has been cited by 10 HighWire-hosted articles. Access the articles at: http://mct.aacrjournals.org/content/7/9/3029.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mct.aacrjournals.org/content/7/9/3029. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.