Use of 5-Fluorodeoxycytidine and Tetrahydrouridine to Exploit High Levels of Deoxycytidylate Deaminase in Tumors to Achieve DNA- and Target-Directed Therapies1

[CANCER RESEARCH 44, 2551-2560, June 1984]

Use of 5-Fluorodeoxycytidine and Tetrahydrouridine to Exploit High Levels of Deoxycytidylate Deaminase in Tumors to Achieve DNA- and Target-directed Therapies1

John A. Mekras. David A. Boothman, Liliana M. Perez, and Sheldon Greet2

Departments of Microbiology and Immunology [J. A. M., D. A. B., L M. P., S. G.] and Oncology and Biochemistry [S. G.J, School of Medicine, University of Miami, Miami, Florida 33101

ABSTRACT We propose that the administration of FdCyd with H4Urd can result in preferential, tumor-directed conversion of a nontoxic In view of the 20- to 80-fold elevation of deoxycytidine-5'- nucleoside analogue to a toxic antimetabolite by an enzyme that phosphate (dCMP) deaminase in many human malignant tumors, is markedly elevated in human tumor tissue. The analogues of we have utilized 5-fluorodeoxycytidine (FdCyd) (Â»administered deoxycytidine are resistant to catabolism and are anabolized by with tetrahydrouridine (H4Urd) as a combination of antitumor a different subset of enzymes than are 5-fluorouracil or 5- agents against two murine solid tumors which possess high fluorodeoxyuridine; therefore, it is a novel approach. Not only levels of dCMP deaminase. This approach is based on our past are there intrinsic selectivity, metabolic stability, and the advan studies in which we demonstrated that FdCyd is an excellent substrate for mammalian 2'-deoxycytidine kinase, and that tages that accrue from prodrug therapy in this strategy, but in addition, the potential for an exclusively DMA-directed effect H4Urd increases the toxicity of FdCyd in the mouse. Cell culture studies utilizing 2'-deoxytetrahydrouridine which inhibits cytidine exists. This is in contrast to approaches with 5-fluorouracil and 5-fluorodeoxyuridine, in which, in addition to DNA effects, parallel deaminase and as 2'-deoxytetrahydrouridine-5'-monophos- effects on RNA metabolism and processing occur. phate inhibits dCMP deaminase, provide indirect evidence for the pathway that we had proposed in the past, INTRODUCTION

FdCyd- FdCMP FdUMP Enzymatic studies in this laboratory first established that dCyd Kinne dCMP deaminase FdCyd3 is an excellent substrate for mammalian dCyd kinase, 2'-Deoxytetrahydrouridine antagonized the toxicity of FdCyd to displaying a Vmaxof 1.8 /Â¿mol/hrand a K, with respect to dCyd a greater extent than did H4Urd and showed marked antagonism phosphorylation of 3 UMunder conditions in which dCyd displays in cytidine deaminase-deficient cells. Cell lines lacking both cyti a Vmaxof 1.7 fÂ¿mol/hrand a Km of 11 ^M [Cooper and Greer dine and 2'-deoxycytidine-5'-monophosphate deaminase were (10)]. We also presented indirect evidence that this analogue markedly resistant to FdCyd. Thymidine and deoxyuridine antag was phosphorylated via dCyd kinase in the mouse [Cooper and onized toxicity in a manner consistent with the proposed path Greer (9)]. This was based on our findings that H4Urd, an inhibitor way of anabolism of FdCyd and consistent with its resulting in of cytidine deaminase, increased the toxicity of FdCyd in C57BL/ the inhibition of thymidylate synthetase. 6 x DBA/2 F, (hereafter called BD2F,) mice, presumably by We have established the efficacy of FdCyd + H4Urd chemo preventing conversion to FdUrd and thereby circumventing ca therapy utilizing adenocarcinoma 755 and Lewis lung carcinoma tabolism by both uridine and thymidine phosphorylases. We in C57BL x DBA/2 F, mice. An example of an optimum schedule concluded that the addition of H4Urd resulted in an increase in versus Lewis lung carcinoma is FdCyd, 10 to 12 mg/kg, plus intracellular levels of FdUMP derived from FdCyd. Our findings H4Urd, 25 mg/kg, coadministered simultaneously, once per day were consistent with previous studies in cell culture (6,11) and in bacteria (4) in which it was shown that dThd kinase-deficient on Days 1 to 7 after tumor implantation. Tumor inhibitions on Days 12, 14, and 16 were 95, 90, and 80%, respectively, with mutants retain sensitivity to FdCyd, whereas they are resistant 8% maximum weight loss. Comparative studies were undertaken to FdUrd. Although we demonstrated that toxicity in mice was only with Lewis lung carcinoma and it was established that increased by (Â»administration of H4Urd with FdCyd, it was FdCyd + H4Urd surpasses the efficacies of 5-fluorouracil and 5- apparent to us that greater insight into the metabolism of FdCyd fluorodeoxyuridine as well as FdCyd when administered without in normal and tumor tissue and characterization of the relevant H4Urd. enzymes would lead to the development of a schedule of FdCyd and H4Urd that would result in efficacy that could complement 1This work was supported by Grant CH-155 from the American Cancer Society (National), the Women's Cancer Association of the University of Miami, a BiomÃ©dical or surpass that obtained with FdUrd or FUra in tumor chemo Research Support Grant from the University of Miami, and gifts from the private therapy. We were guided by the findings of Heidelberger ef al. sector including Readmore Publications and The Weeks Endowment Fund. This is who demonstrated that FdCMP was not an inhibitor of thymi Paper 1 of a series of 2 papers. Paper 2 is entitled, "Use of 5-trifluoromethyldeox- dylate synthetase (20), even though FdCyd was 40-fold more ycy tidine and tetrahydrouridine to circumvent catabolism and exploit high levels of cytidine deaminase in tumors to achieve a DNA- and target-directed therapy," by effective than FdUrd in inhibiting DNA synthesis in HeLa cells J. A. Mekras, D. Boothman, and S. GrÃ©er.This paper is derived, in large part, from (38). the Ph.D. dissertation of John Mekras, University of Miami, September 1982. This paper is dedicated to the memory of Ina GrÃ©er,wife of S. GrÃ©er.She died of cancer 3 The abbreviations used are: FdCyd, 5-fluorodeoxycytidine; ara-C. 1-/3-o-arabi- in July 1980 while these studies were being pursued. It is also dedicated to the nofuranosyteytosine; CHO, Chinese hamster ovary cells; dCyd, 2'-deoxycytidine; memory of Dr. Chartes Heidelberger, a colleague whose studies formed the basis dCMPD, deoxycitidme-5'-phosphate deaminase; dhUUrd, 2'-deoxy tetra of the approach in this and the subsequent paper in the series. hydrouridine: dThd, thymidine; dUrd, 2'-deoxyuridine; FdUrd, 5-fluoro-2'-deoxy 1To whom requests for reprints should be addressed. uridine; FdUMP, 5-fluoro-2'-deoxyuridine-5'-monophosphate; FUra, 5-fluorouracil; Received June 23,1983; accepted March 14,1984. HEp-2, human epidermoid laryngeal carcinoma cells; H.Urd. tetrahydrouridine.

JUNE 1984 2551

With this background, we sought to undertake biochemical Mammalian Cells and Culture Techniques. HEp-2 cells as well as and tumor inhibition studies and proposed that the advantages L1210; L5178Y, a leukemic cell line; C1498, a myelogenous leukemic of our approach were that: (a) FdCyd would serve as a meta- cell line; and Ehrlich ascites cells were obtained from the American Type bolically stable prodrug of FdUMP since dCyd and its analogues Culture Collection. CHO cells were obtained from Dr. H. Gratzner, Department of Medicine, University of Miami. CCL39 and GM1 Chinese are not substrates of undine and thymidine phosphorylases (17); hamster fibroblasts were kindly supplied by Dr. G. Buttin, Institut Pasteur, (b) a different subset of enzymes is involved in the anabolism of Paris, France. These 2 cell lines were described by Robert de Saint FdCyd than is involved in FUra and FdUrd anabolism, thereby Vincent ef al. (44). Briefly, CCL39 is cytidine deaminase deficient but increasing the number of tumors potentially responsive to ther does possess dCMPD. GM1 is a mutant derived from CCL39 and is apy with fluorinated pyrimidines; e.g., a tumor with low levels of both cytidine deaminase and dCMPD deficient; it possesses a TTP pool thymidine kinase or undine phosphoribosyltransferase could that is 25% that of the wild type, CCL39. GM1 is resistant to 8- nonetheless be responsive to FdCyd due to the fact that dCyd azaguanine, lacking guanine-hypoxanthine phosphoribosyl transferase; kinase activity has been shown to be elevated in many human it is not resistant to ara-C, and it possesses a dCTP pool that is only and murine tumors (25, 30); (c) because FdUrd is rapidly cata- slightly expanded above that of its parent, CCL39. For colony-forming assays, cells were grown in a 95% air-5% COa bolized to form FUra, which can readily be incorporated into RNA, the combination of FdCyd and H4Urd is, in contrast, a humidified atmosphere; medium was supplemented with amphotericin B DMA-directed approach [a growing body of evidence suggests (0.25 Mg/ml) penicillin (100 units/ml), streptomycin (100 units/ml), and sodium bicarbonate was increased to 0.1%. that the incorporation of FUra into RNA contributes more to Media, sera, glutamina, trypsin, and antibiotics were purchased from general toxicity than to tumor inhibition (41); this issue is not Grand Island Biological Company (Grand Island, NY), Flow Laboratories settled]. Furthermore, FdUrd has been shown by Chadwick (8) (McLean, VA), and K-C Biologicals (Lenexa, KS). to be a better source of RNA-FUra than is FUra itself; and (d) in Cells were passed at confluency (about 3 days) and subcultured at a view of the marked elevated levels of dCMPD in many human 1:5 or 1:6 dilution by detaching the monolayers with trypsin. HEp-2 cell malignant tumors (12, 18, 30), FdCyd + H4Urd is a combination monolayers were detached for passage by exposure to trypsin-EDTA [trypsin (0.25%), EDTA (0.2 g/liter, in Puck's Saline A] for 2 min. All 3 of agents in which a nontoxic metabolite (FdCyd or FdCMP) could be preferentially converted at the tumor site to a toxic Chinese hamster lines were exposed for only 30 sec, and detached cells antimetabolite, FdUMP, a potent inhibitor of thymidylate synthe- were pelleted by centrifugaron and resuspended in fresh medium to wash them free of trypsin before subculturing. tase. It is the inhibition of this enzyme that is thought to be Cytotoxicity Assays. Cells (5x105) were plated into 60-mm tissue primarily responsible for tumor inhibition by fluorinated pyrimi culture dishes in 4 ml of Eagle's minimal essential growth medium for dines rather than by effects on RNA metabolism, maturation, growth in a humidified atmosphere. These plates were then incubated and processing (45, 46). Indeed, use of an exclusively DNA- for 12 to 18 hr to allow for attachment and log-phase growth to begin. directed approach would help settle the controversy regarding At the end of this period, the growth medium was aspirated off and the relative importance of the DNA and RNA components in replaced with 4 ml of medium containing drugs. Four times concentrated solutions of drugs were prepared in Eagle's minimal essential growth tumor therapy with fluorinated pyrimidines. Recently, Newman and Santi (42) have recognized the DNA- medium, and 1 ml was added to the plates to a total volume of 4 ml. directed aspects of FdCyd + H4Urd therapy, which they state is The plates were then incubated for 48 hr, and the cells were harvested based on our past findings (9). They did not refer to aspects of by trypsinization. The cells were then serially diluted in growth medium, and appropriate dilutions were replated into 60-mm dishes and incubated selectivity or emphasize that FdCyd could act as a metabolically for 6 to 9 days depending on the cell lines. After this 6- to 9-day incubation stable prodrug but suggested that it was an approach worthy of period, the colonies were washed and stained, and colonies consisting animal tumor studies. This has, indeed, been a long-standing of sou cells were counted. All conditions were done in duplicate. interest of our laboratory, and the present studies establish the Enzyme Assays. Enzyme activities and protein concentration were efficacy of FdCyd + H4Urd as well as its superiority to FdUrd determined as described by Dobersen and Greer (13). To determine and FUra as a combination of antitumor agents utilizing 2 solid- cytidine deaminase activity, [3H]dCyd and cellulose thin-layer chromatog tumor model systems, mammary adenocarcinoma 755 and raphy were utilized to separate dCyd from dUrd and to quantitate the Lewis lung carcinoma. A preliminary account of these studies extent of dUrd formed from dCyd. Enzyme activity is expressed as nmol converted/g protein/hr at 37Â°.Reactions remained linear for 30 min, and has been presented (37). 15-min assay points were used in quantitating enzyme activity. In the dCMPD assay, [5-3H]dCMP was utilized; the reaction mixture included MATERIALS AND METHODS dCTP, the activator of dCMPD. Reactions did not normally remain linear beyond 15 min; consequently, 5-min points were used in quantitating Chemicals. Pyrimidine bases, nucleosides, and nucleotides were pur enzyme activity. Enzyme activity was expressed as ^mol converted/g chased from Sigma Chemical Co., St. Louis, MO, and P-L Biochemicals, protein/10 min at 37Â°. Milwaukee, Wl, with the following exceptions. dH4Urd was supplied and In the dCyd kinase assay, H4Urd was added to the reaction mixture, FdCyd was custom synthesized by Calbiochem-Behring Corp., La and enzyme activity was expressed as nmol converted/g protein/hr at Jolla. CA. H4Urd was a gift of Dr. Leonard Kedda, National Cancer 37Â°. Institute, Bethesda, MD. Animal Care and Maintenance of Tumors. BD2F, and C57/BLK mice Radiochemicals. [5-3H]dCyd was purchased from New England Nu were obtained from The Jackson Laboratory, Bar Harbor, ME. Animals clear (Boston, MA) and ICN (Irvine, CA). [5-3H]dCMP was purchased were provided Purina Laboratory (St. Louis, MO) feed and water ad from New England Nuclear. All labeled materials were greater than 99% libitum. pure. Solid tumors were maintained as follows. Animals were sacrificed by Chromatography Supplies. DE81 chromatography paper for use in cervical dislocation. Tumors were removed aseptically and placed into dCyd kinase enzyme assays was obtained from Whatman (Bedford, MA). Petri dishes containing ice-cold phosphate-buffered saline. Tumors were Flexible cellulose thin-layer chromatography sheets (Baker-flex) were debrided, necrotic tissue was removed, and viable tissue was cut into 2- supplied by J. T. Baker (Atlanta, GA). to 3-mm fragments. These fragments were loaded into 13-gauge trocars

2552 CANCER RESEARCH VOL. 44

Downloaded from cancerres.aacrjournals.org on September 27, 2021. © 1984 American Association for Cancer Research. Target- and DNA-directed Therapy with FdCyd and subsequently transplanted s.c. into the axillary region via inguinal reduction of the toxicity of FdCyd in these cells. In cytidine puncture. Tumors were passed every 7 to 10 days depending on their deaminase-deficient cells that are, in addition, impaired in growth kinetics. dCMPD activity (GM1 cells),dH4Urddisplays a less striking effect Tumor Inhibition Studies. All antitumor studies were done with female than in cells possessing dCMPD. Consistent with the pathway BD2F! mice, unless otherwise noted, using the transplantation proce dures already described. Tumor-inhibitory activity was assessed by in Chart 1 is the finding that cells that are impaired in dCMPD activity are more resistant to FdCyd with or without coadminis- measuring tumors in 3 dimensions with vernier calipers and computing volumes. The groups were coded after treatment so that the measure tered deaminase inhibitors than are the dCMPD-containing cells. ments were not biased. A T/C value (the ratio of the volumes of the This is exemplified by CHO cells which are totally lacking cytidine treated to that of the control) was then computed, and this value was deaminase and dCMPD activity and are strikingly resistant to used to calculate the extent of tumor inhibition (1 - T/C x 100) to FdCyd. As expected, H4Urd and dH4Urd have only marginal compare different drug regimens. Weight loss data were also obtained effects on the toxicity of FdCyd in CHO cells; for example, in an and used as a measurement of toxicity along with mortality data. experiment not shown in Table 1 with 20 UMFdCyd, 28,42, and All drug solutions were sterilized by filtration (0.45-pm Millipore filters; 47% survival was obtained with no addition, 4 mM H4Urd,and 4 Millipore, Bedford, MA) before being injected i.p. The regimens used are mMdH4Urd,respectively. indicated in the tables summarizing the tumor inhibition studies. Statistical Evaluation. Student's f test was used for statistical eval The marked resistance of CHO cells to FdCyd is due to their uation of differences in mean tumor size between groups of treated mice. lack of converting enzymes and not to the possession of an Differences between groups with a statistical probability of 0.05 or less intrinsically resistant thymidylate synthetase or ineffective folate were considered significant; p values relating to tumor inhibition studies levels as evidenced by the fact that FdUrd at concentrations of reflect comparisons between treated groups on Days 10 to 16. Standard 0.05 and 0.5 MMresulted in 40 and 1.5% survival, respectively. deviations in treated and untreated groups were similar and, in most Although the dCyd kinase activity of CHO cells was not deter- instances, were within 15% of mean tumor volumes. FdUrdinDNA RESULTS t Â¡FUraIIil lemoiedby FdUTP (Â¡paircascai)!thai Itads Cell Culture Studies. The cytotoxicityof FdCyd with and ti siifli situi kinks] without HjUrd or dH4Urdin several cell lines is shown in Table 1. In HEp-2 cells, H4Urdprevents toxicity to a great extent, and FdCDP FdUDP the addition of dH4Urd results in even greater survival. These A ICMP results are consistent with the pathway shown in Chart 1. IEANIMSE t UHIITIM Js. TITIIIITL1TE Without HÂ«Urd,FdCyd may be anabolized to FdUMP via the FOP- HUMP ">^ JWIETKE dCyd kinase or dThd kinase pathway. dH4Urd is an inhibitor of cytidine deaminase and is an inhibitor of dCMPD, as the nucleotide, dH4UMP (33). With dH4Urd, all routes of conversion of TICK FdCyd to FdUMPare inhibited. Another lineof evidence indicating CI the importance of dCMP deaminasefor the conversion of FdCMP to FdUMP and the achievement of a toxic effect is displayed in H4U Chinese hamster fibroblasta (CCL39 cells) which lack cytidine deaminase (44). In accord with the absence of cytidine deami Chart! Proposed pathway of mtracellular metabolism of FdCyd in the presence of H,Ur d (H4U).The size and intensity of the arrows are meant to reflect the relative nase in CCL39 cells, H4Urdhas no effect on the cytotoxicity of rate and extent of anabolism or conversion in tumor tissue. dCK. dCyd kinase; CO, FdCyd, whereas the addition of dH4Urd results in a dramatic cytidine deaminase; TK, dThd kinase.

Tabtel Cytotoxicity of FdCyd with and without HJJrd or dHJJrd in several cell lines and effect of pyrimidine deoxyribonucleosides on the toxicity of FdCyd Cytidine deaminaseor dCMPD activity (% of cells surviving) cells (cytidine deaminase-deficient, GM1 cells (cytidine deaminase-defi- cells' (cyti cells (cyti dCMPD-deficient)NodCMPD-containing) dent. dine deaminase- dine dearnmase- deficient, FdCyd containing,dCMPD-containmg)101 dThd + dUrd + dThd dCMPD-defi-cient)87.5 (MM)0000.50.50.50.50.51.01.01.020H,Urd(mM)4142dH,Urd(mM)4142HEp-2 addition99.5 (0.1 mM) (0.1 mM)mM)98.2 - dThd (0.1 Â±0.02*89.3 Â±1.992.1 Â±3.194.1 Â±3.40.16 Â±2.40.35 2.7101 Â± Â±1.667.2Â±1.5 94.5 Â±0.0214.0 Â±0.080.24 Â±0.8216.5 Â±1.248.0 Â±0.029.4 Â±2.848.2 Â±3.40.12.8 Â±0.400.0080.042

Â±0.71 1.26 Â±0.02 0.066Â±5.10.28 Â±0.0007 62.0 Â±8.828.1 Â±0.1426.0 Â±7.533.5 Â±0.014 56.2 Â±1.8CCL39 Â±0.01+ Â±7.7 78.0 Â±1.6 0.29 Â±0.010 66.0 Â±9.1CHO 6.53 Â±

â€¢ i-.-lâ€” -J iâ€žâ€ž , ..,.Â»K M.M-UW AÂ«JI linn - UVMtl*AetÂ¡Â«*n Mut Â«WAS* n-11-ic-t-.n^l-i A< t~~LJf~\ f~r\\lf 4Â« CÃ-tr*Â«lH 6 Mean Â±S.E.

JUNE 1984 2553

Downloaded from cancerres.aacrjournals.org on September 27, 2021. © 1984 American Association for Cancer Research. J. A. Mekras et al. mined, the fact that they display sensitivity to ara-C is in accord Table 2 with their possession of an intact dCyd kinase. Survivals of 1.9 Cytidine deaminase, dCPMD. and dCyd kinase activities of various neoplastic and normal tissues of BD2F, mice, as well as cytidine deaminase and dCMPD and 1.5% were obtained with 5 and 10 U.Mara-C, respectively. activities of human and hamster cell cultures A possible explanation of why H.,Urd has an effect in GM1 Cytidine deami- dCMPO dCyd kinase cells is that, in lacking cytidine deaminase activity, all the H4Urd nase Gimol/g pro- (nmol/g pro- is available to be anabolized to H4UMP, which Maley and Maley Tissue (nmol/g protein/hr) tein/10 min) tein/hr) (34) have shown to be a weak but nonetheless discernable Tumor Lewis lung carcinoma Â±1,410* Â±6.11 Â±52.0 inhibitor of dCMPD, an activity that is very low and therefore 755NormalAdenocarcinoma 2,185 Â±2574 80.39 Â±2.6213.66 217.3 Â±47.2323.0 readily inhibited in this cell line. It should be noted that GM1 cells differ from CCL39 cells not only with respect to their levels of Spleen Â± 23 Â±9.36 Â± 1.0 dCMPD; in addition, GM1 cells lack guanine-hypoxanthine phos- Liver3.9" 1,380 Â± 130 7.18 Â±4.25 11.6 Â± phoribosyl transferase, and their pool of TTP is almost one-fourth Intestinal mucosa 7,827 Â±5737,300 4.18 Â±3.55 that of CCL39 cells (44). This latter feature may explain, in part, BonemarrowCultured 7.51 Â±0.554.3 why GM1 cells are much more sensitive to FdCyd than are CHO cells cells, which are also cytidine deaminase and dCMPD deficient. HEp-2 Â±7.1 CHO12,010 1030.66 1.0272.6 Although GM1 cells were shown to possess less than 0.5 unit Mean Â±S.E. (nmol dUMP formed/15 min/mg protein) of dCMPD activity com pared to 37 units in the parental cell line (44), the fact that these cells are elevated only to a slight extent (27%) in their dCTP Jackson (26) indicate that approximately 60 to 80% of dUMP is pools (44) indicates that the block in dCMPD activity is not derived from dCMP in rat hepatomas. The levels of dCyd kinase absolute. These characteristics of GM1 cells may also contribute in Lewis lung carcinoma, mammary adenocarcinoma 755, and to their relative sensitivity to FdCyd and to the partial response normal liver were 273 Â±52, 217 Â±47, and 11.6 Â±3.9 nmol/g to dHÂ«Urdthat was obtained with these cells. protein/hr, respectively. This is consistent with the findings of Studies utilizing metabolites to antagonize the inhibition of elevated levels of dCyd kinase in tumors by others (19, 25, 30) FdCyd are in accord with the proposed pathway in Chart 1 and and indicates that these tumors should be responsive to FdCyd also provide indirect evidence regarding the contribution of in + H4Urd therapy. corporation of FdCyd, as such into DNA, to its cytotoxic effect. The addition of H4Urd or dH4Urd to log-phase HEp-2 cells dThd markedly antagonized the toxicity of FdCyd, consistent inhibited the appropriate target enzymes. After 3 hr, cytidine with the anabolism of FdCyd to FdUMP via the dCyd kinase â€”Â»deaminase activity was inhibited 93 and 92% with 4 mw H4Urd dCMPD pathway. In GM1 cells, there is no indication that the and dH4Urd, respectively. After a 4-hr exposure to 100 /Ã•M presumed incorporation of FdCyd as such into DNA contributes concentrations of the inhibitors, 67 and 59% inhibition of dCMPD substantially to toxicity; however, the slightly higher levels of was obtained with dH4Urd and H4Urd, respectively. A 4-hr ex dCTP in these cells (44) may compete with FdCyd for incorpo posure to 20 U.Minhibitor concentrations resulted in 53 and 39% ration into DNA. The results in Table 1 with CCL39 cells are inhibition of dCMPD with dH4Urd and H4Urd, respectively. It is consistent with a slight effect of FdCyd incorporation into DNA likely that the use of lower concentrations of dH4Urd and H4Urd on toxicity in view of the fact that 67% survival was obtained would have resulted in a greater difference in their ability to with the combination of FdCyd and dThd and only 34% survival inhibit dCMPD, with the inhibition by dH4Urd persisting and that was obtained upon the addition of dH4Urd to this combination. of H4Urd diminishing. The addition of 340 U.Mconcentrations of However, when dUrd at a concentration of 0.1 mw was added each inhibitor to a crude extract of HEp-2 cells resulted in 95 along with dH4Urd and FdCyd, the fraction of surviving cells and 96% inhibition of cytidine deaminase with H4Urd and dH4Urd, increased to 78%. respectively. The addition of 34 UM concentrations of each dCyd antagonized the cytotoxicity of FdCyd in the absence inhibitor to a crude extract of HEp-2 cells resulted in 66 and 6% and in the presence of H4Urd in HEp-2 cells. In an experiment inhibition of dCMPD with dH4Urd and H4Urd, respectively. Maley not shown in Table 1, FdCyd (1 UM) resulted in 0.33 and 35% and Maley (33) found that dH4UMP was 100-fold more effective survival with no addition and with 0.1 mu dCyd, respectively; than H4UMP in inhibiting purified dCMPD obtained from chick the addition of H4Urd (1 mw) and H4Urd with dCyd resulted in embryos. 3.3 and 67% survival, respectively. Tumor Inhibition Studies. The effect of FdCyd coadministered Enzymatic Studies. The cytidine deaminase, dCMPD, and with H4Urd on mammary adenocarcinoma 755 as a solid tumor dCyd kinase activities of Lewis lung carcinoma, mammary ade- in BD2F! mice is shown in Table 3. The range of concentrations nocarcinoma 755, and several murine normal tissues are shown utilized in our initial experiments was based, in part, on our in Table 2. The HEp-2 cell line, with which extensive studies previous studies with C3H mice (9). Various schedules were were undertaken, has a high level of dCMPD that is represent utilized that resulted in effective tumor inhibition without encoun ative of the tumor lines. All murine tumors that were examined tering toxic deaths. Extensive and sustained tumor inhibition was possessed high levels of dCMPD. The comparable values for obtained in Experiment 5, although this was achieved with Sarcoma 180, L1210 leukemic cells, Ehrlich ascites, L5178Y extensive weight loss. These results are comparable to those cells and C1498 cells were 51 Â±17 (S.E.), 76 Â±10, 103 Â±57, obtained in Heidelbergers laboratory with FUra (24) and FdUrd 53 Â±18, and 30 Â±3.2 (/Â¿mol/gprotein/10 min at 37Â°),respec (22) with adenocarcinoma 755. tively. Levels of dCMPD have been shown to be markedly Various schedules of FdCyd and H4Urd were examined against elevated in rodent hepatomas by Maley and Maley (32) as well Lewis lung carcinoma in its solid form; some representative as by Cummins and Balinsky (12). Furthermore, studies by schedules are shown in Table 4. As indicated by Cooper and

2554 CANCER RESEARCH VOL. 44

Table 3 Effect of FdCyd coadmimstered with HJJrd versus mammary adenocarcinoma 755 as a solid tumor in BD2F, mice activityExperimentlalb23a3b45FdCyd(mg/kg)7.57.511.512.512.58.2511.25H,Urdmg/kg25252525012.512.5Timings"-30Antitumor of tumorinhibition)'Day of toxic deaths frp-quency1/day;srheriiite treated%Maximum among 6 1081808497629098(%1276728291548993Day1468677582546889Day1651496369215480No.wt lossmice71141198170000000 Days1/day; minSSSsTreatment Days1/day; Days1/day; Days1/day; Days2/day; Days1-71-71-71-71,3,5,7,91,3,5,7,9Day ^ volume treated ' S, simultaneously\volume control/ coadministered with nucieoside analogue.

Table 4 Effect of FdCyd coadmimstered with and without HJJrd or dHJJrd versus Lewis lung carcinoma in BD2F, mice activityH,UrdExperi Antitumor inhibition)Day8828378838388887193959389Day10738668816344846390929289Day126752565851815983909477Day1458565480858676Day161565736350Maximum%of tumor of toxicdeathsamong sched-mg/kg 6treatedmice060300000203 ment1a1b2a2b3a3b4Sa5b5c5d5eFdCyd(mg/kg)55555371010101010mg/kg1050252510251025100Timing-30Timing ulefrequency2/day; wtloss1222122233699101514No. min-30 1-52/day;Days min-30min5aSSSSSfSHJMTreatment 1-52/day;Days

1-52/day;Days

1-41/day;Days

1-725 Days

S100 S(% S, simultaneously coadministered with nucieoside analogue.

Greer (9), there is a concentration range in which the coadmin- implantation was more effective. istration of H4Urd with FdCyd can lead to toxic deaths. This kind The effect of FdCyd (with and without H4Urd or dH4Urd) against of interaction between FdCyd and H4Urd is not seen in cell Lewis lung carcinoma in BD2F, mice is shown in Table 4. It can culture, presumably because catabolism of FdUrd in cell culture be seen that FdCyd, without H4Urd, did not inhibit tumor growth is very slight compared to systemic catabolism encountered in as effectively as did FdCyd coadministered with an inhibitor of the mouse. For example, Laskin et al. (29) found very low levels deamination (p < 0.01). To explain the effective tumor inhibition of thymidine phosphorylase in HEp-2 cells compared to the levels obtained with FdCyd and dH4Urd (25 mg/kg), we speculate that found in human normal and tumor tissue. Of primary importance the low concentration of dH4Urd may have resulted in modulation in this regard is the extent of catabolism that occurs in liver and of dCyd deaminase and dCMPD to provide a sustained source serum. In the animal system, high concentrations of H4Urd pre of FdUMP, available for tumor cells as they are recruited into S sumably increase the plasma half-life of FdCyd so that more phase. High doses of H4Urd or dH4Urd (100 mg/kg) coadminis FdCMP is formed in all tissues. At very low concentrations of tered with FdCyd resulted in toxic deaths. We had hoped to H4Urd or in the absence of H4Urd, the competing pathways of demonstrate that, at high doses, dH4Urd would prevent tumor conversion to FdUrd and catabolism to FUra and further catab inhibition by FdCyd to obtain in vivo evidence for tumor inhibition olism by dihydrouracil dehydrogenase may be favored over by the pathway shown in Chart 1. However, only a slight and anabolism of FdCyd to FdCMP by dCyd kinase and subsequent insignificant decrease in tumor inhibition was observed, and conversion to FdUMP. Clearly, the concentration and timing of death occurred in 50% of the animals. In an experiment not both agents is a critical factor in obtaining optimal efficacy. The shown in Table 4, 50 mg of dH4Urd/kg coadministered with 12 use of a 30-min pretreatment schedule with H4Urd followed the mg FdCyd/kg (one injection/day on Days 1 to 7) resulted in 19% tradition of past in vivo studies (7). In Table 4, Experiment 2, it weight loss and no toxic deaths; again, only a slight and insig can be seen that the simultaneous administration of H4Urd nificant reduction in tumor inhibition occurred compared to that resulted in greater toxicity than when H4Urd was coadministered obtained with 25 mg H4Urd/kg coadministered with the same 30 min prior to FdCyd. Results of other experiments (not shown) concentration of FdCyd (81 and 88% tumor inhibition, respec indicate that there may be a simple dose increase effect with tively, on Day 16). simultaneous addition of FdCyd and H4Urd. In initial experiments, A rational basis for timing and scheduling of FdCyd and H4Urd the rapid growth of Lewis lung carcinoma suggested that inten or dH4Urd must await appropriate pharmacokinetic, enzymatic, sive treatment (2 injections/day) over a 5-day period would be and incorporation studies, which are planned. These experiments appropriate; however, further experimentation indicated that a will also clarify our understanding of the basis of enhanced dose schedule of one injection/day on Days 1 to 7 following morbidity and mortality obtained when high concentrations of

JUNE 1984 2555

TabteS Effect of FOCydcoadministered with and without HJJrd compared to higher concentrations of FdUrd and to FUraagainst Lewis lung carcinoma in BD2F, mice Antitumor activity - â€¢â€¢" (% of tumor inhibition) No. of toxic deaths Experi simulta schedule 6 treated ment67878897777898deoxy-ribonucteosideFdUrdFdUrdFdUrdFdUrdFUraFUraFdCydFdCydFdCydFdCydFdCydFdCydFdCydFdCydmg/kg10408010025301240101111121214(mg/kg,neously)252525252525Treatmentfrequency1/day; 87092839878946895989498959795Day106475819678955792979697949795Day124167759573925288959396939695Day144057728570864273888889919287Day162543687564763865777879858279Day181531625744712440626058707577Maximum% wttoss10511159215109968816amongmice00010300000002 1-71/day;Days 1-71/day;Days 1-71/day;Days 1-71/day;Days

1-71/day;Days 1-71/day;Days

1-71/day;Days 1-72/day;Days 3,5,7,91/day;Days 1, 1-71/day;Days 1-71/day;Days Days 1-7Day

H4Urd and dH4Urd are coadministered with FdCyd. These high As seen in Table 5, Experiment 8, when the concentration of concentrations of deaminase inhibitors may channel incorpora FUra was increased to 30 mg/kg, comparable tumor inhibition tion of FdCyd into DNA. occurred; however, toxic deaths occurred in 50% of the animals. Studies comparing FUra, FdUrd, and FdCyd alone with FdCyd Additional data not shown in Table 5 but derived from Exper + H4Urd against Lewis lung carcinoma are shown in Table 5. iment 9 is of interest. The administration of H4Urd or dH4Urd FdCyd at a concentration of 12 mg/kg (coadministered with alone at a dose of 50 or 100 mg/kg, one injection/day for 7 days H4Urd) surpassed the efficacy of FdUrd at a concentration of 10, resulted in no weight loss. No significant tumor inhibition was 40, and 80 mg/kg (one injection/day for 7 days) (p < 0.01 ). When displayed by these deaminase inhibitors except at high concen the concentration of FdUrd was increased to 100 mg/kg, tumor trations of dH4Urd (100 mg/kg, one injection/day, for 7 days). inhibition comparable to that obtained with FdCyd + H4Urd was This dose resulted in 28, 26, and 15% tumor inhibition on Days seen; however, there was a substantial increase in weight loss 10, 12, and 14, respectively, with weight gain equal to that of and one toxic death among 6 animals. FdCyd (40 mg/kg) when untreated animals. It is conceivable that this slight but significant administered alone was more effective than an equal dose of (p < 0.05) inhibition of tumor growth may reflect the importance FdUrd (p < 0.01), but it was not as effective as were lower of dCMPD activity to the metabolism of the tumor. FdCyd (12 concentrations (10 mg/kg) of FdCyd coadministered with H4Urd mg/kg), when administered alone, resulted in 15 and 0% tumor (p < 0.05). In an experiment repeating, combining, and extending inhibition on Days 20 and 22, respectively, and an increase in life the conditions of Experiments 7 and 8 (Experiment 10, not shown span ratio of 1.15; whereas the same concentration of FdCyd, in Table 5), FdUrd and FdCyd (40 mg/kg, one injection/day on coadministered with H4Urd (25 mg/kg), resulted in 66 and 53% Days 1 to 7) again resulted in relatively poor tumor inhibition (67 tumor inhibition on Days 20 and 22, respectively, with an increase and 61%, respectively, on Day 14 and 42 and 48% on Day 16). in life span ratio of 1.72. Although this is not an impressive FdUrd, at concentrations of 80 and 100 mg/kg, resulted in 1 and increase in life span when compared to other agents in other 3 deaths among 6 animals, respectively. FdCyd (without H4Urd) mouse tumor models, the increase in life span surpassed that at a concentration of 80 and 100 mg/kg resulted in 5 deaths obtained with FUra and FdUrd with Lewis lung carcinoma, which, among 6 animals in both groups, whereas FdCyd (12 mg/kg), it should be noted, is a highly metastatic tumor model. For coadministered with H4Urd (25 mg/kg), resulted in no toxic example, FdUrd at a dose of 40 mg/kg (one injection/day for 7 deaths, and tumor inhibition and weight loss were identical to days) resulted in 20% tumor inhibition on Day 20 and an increase those shown in Table 5. Lowering the dose of H4Urd to 10 mg/ in life span ratio of only 1.30. FdUrd at a dose of 80 mg/kg kg, coadministered with 12 mg FdCyd/kg, resulted in much less resulted in 41% tumor inhibition on Day 20 with an increase in tumor inhibition than when 25 mg H4Urd/kg were coadministered life span ratio of 1.46 (omitting 1 of 6 toxic deaths). Experiment (48% compared to 88% tumor inhibition, respectively, on Day 9 (Table 5) was the only experiment in which male rather than 16). female mice were used. There is no significant difference in FdCyd (12 mg/kg + H4Urd, 25 mg/kg, one injection/day for 7 tumor inhibition or weight loss compared to an equivalent exper days) surpassed the efficacy of FUra, when the latter is utilized iment with females (Experiment 8). at its optimum and most generally used concentration of 25 mg/ FdCyd at a dose of 10 mg/kg for 7 days is essentially equal kg (p < 0.0025). This is consistent with studies which demon to 12 mg/kg for 7 days [coadministered with H4Urd (25 mg/kg)] strated that Lewis lung carcinoma possesses very low levels of in efficacy; no toxic deaths were encountered in 9 separate uracil phosphoribosyltransferase; e.g., the levels of this enzyme experiments involving 70 animals utilizing a range of 10 to 12 in this tumor are more than 12 times lower than in Ehrlich ascites mg of FdCyd/kg. Increasing the concentration of FdCyd to 14 carcinoma and 50 times lower than in Novikoff hepatoma (23). mg/kg resulted in toxic deaths (33%). Although there was only

2556 CANCER RESEARCH VOL. 44

Downloaded from cancerres.aacrjournals.org on September 27, 2021. © 1984 American Association for Cancer Research. Target- and DNA-directed Therapy with FdCyd a slight difference in weight loss with a change in the schedule be an important factor in the effectiveness of fluoropyrmidine and total dose seen when FdCyd + H4Urd was given twice per therapy (40, 45). day on Days 1, 3, 5, 7, and 9 (Table 5, Experiment 7), further The evidence supporting the route of anabolism of FdCyd changes in the scheduling of FdCyd and (Â»administration of shown in Chart 1 has been derived from cell culture studies in H4Urd could lead to even greater efficacy with this combination which we have shown that the cytotoxicity of FdCyd is markedly of agents. reduced when dCMPD is inhibited by the addition of dH4Urd or The results of these tumor inhibition studies demonstrate that by utilizing mutants deficient in both cytidine deaminase and substantial efficacy has been obtained with FdCyd + H4Urd in 2 dCMPD. Supporting evidence for the pathway as depicted in murine solid tumor model systems. Furthermore, with Lewis lung Chart 1 derives from the findings that FdCyd displays substantial carcinoma, which is not usually responsive to nucleic acid con toxicity in the presence of H4Urd although that toxicity is not as stituent analogues, FdCyd + H4Urd has been shown to surpass great as when H4Urd is omitted, presumably because 2 pathways the efficacies of FdUrd and FUra. are available for the conversion of FdCyd to FdUMP in the absence of H4Urd. In addition, in the absence of an inhibitor of DISCUSSION cytidine deaminase, FdUrd may be catabolized to FUra and result in RNA-directed toxicity. Balzarini and DeClercq (3) recently The utilization of FdCyd with H4Urd (and 5-trifluoromethyl- confirmed our past enzyme kinetic studies with crude extracts deoxycytidine + H4Urd, to be described in Paper 2 in this series) of human and murine cells (10) in which it may be inferred that possesses the potential to approach some of the major goals in FdCyd is a good substrate for the dCyd kinase from L1210 cells, cancer therapy: selective action against the tumor; metabolic displaying a K/Kâ„¢value of 0.24, which is similar to the value of stability; and retention of effective anabolites of the drug in tumor 0.27 that we obtained for the enzyme from P388 cells in our tissue. Once the pyrimidine nucleosides are phosphorylated in earlier studies. Balzarini and DeClercq demonstrated that dThd tumor tissue, they will be retained to inhibit their target enzymes kinase-deficient mutants of L1210 cells are 17-fold less sensitive after they are deaminated. These 2 approaches are DNA-directed to FdCyd than are wild-type L1210 cells and that dCyd kinase- approaches; therefore, there exists an even greater potential for deficient cells are 25-fold less sensitive than are wild type. selectivity since there is mounting evidence that tumor inhibition Newman and Santi (42), utilizing mutants deficient in 3 of the by fluorinated pyrimidines Â¡sa result of their inhibition of thymi- enzymes of the pathway, found that no single mutant deficient dylate synthetase or their transient incorporation and removal in dCyd kinase, dThd kinase, or dCMPD was resistant to FdCyd from DMA and the repair which follows rather than a result of without H4Urd; however, with H4Urd, the dCyd kinase-deficient incorporation into RNA, which appears to account more for cells were resistant to FdCyd. In the absence of H4Urd, dCyd general toxicity. kinase-deficient, dThd kinase-deficient double mutants and Cytidine and its analogues are not catabolized in mammalian dCMPD-deficient/dThd kinase-double mutants were resistant to cells; therefore, in contrast to FdUrd, which is catabolized by FdCyd. These studies support the pathway shown in Chart 1, uridine and thymidine phosphorylase, FdCyd is resistant to ca- indicating that there are 2 possible routes of anabolism of FdCyd tabolism when its deamination is prevented. Consequently, to FdUMP. FdCyd does not present the problems encountered with FdUrd, It is of interest that in CCL39 cells, in the absence of inhibitors since there are no effective in vivo inhibitors of thymidine phos of deamination, thymidine antagonized FdCyd cytotoxicity to a phorylase. much greater extent than did deoxyuridine: 67% compared to The studies of Chadwick ef a/. (8) indicate that FdUrd and 1.3% survival with dThd and dUrd, respectively (Table 1). This FUra are equivalent as a source of FUra. Furthermore, these may be due to several factors, including the relative Â«â€žsofdThd investigators found that the levels of FdUMP derived from FdUrd and dUrd with respect to dThd kinase. dUrd has a 100-fold higher were lower than those attained after FUra administration. More Km with respect to dThd kinase than dThd (5); furthermore, in over, a greater percentage of FUra in RNA was obtained when the absence of dH4Urd, the levels of dUMP are most probably FdUrd was administered than when FUra was given. Therefore, elevated in cells in which thymidylate synthetase is inhibited by in view of catabolism, FdUrd may be viewed not as a DNA- FdUMP (40). Therefore, the addition of dUrd, the direct precursor directed deoxyribonucleoside analogue but as a storage form of of the substrate of the inhibited enzyme, may be of less conse FUra. Clinical trials comparing FUra to FdUrd have failed to quence than adding dThd, the direct precursor of the product of demonstrate much difference either in terms of toxicity or ther the inhibited reaction. Yoshida ef al. (48) have also observed apeutic efficacy (39). more effective antagonism of FdCyd cytotoxicity by dThd than H4Urd, an intermediate-state analogue and very effective inhib by dUrd in L5178Y cells that are deficient in cytidine deaminase. itor of cytidine deaminase, does not display toxicity in animals Once deamination is blocked and the levels of dUMP are depleted and humans. At modulating concentrations of H4Urd or dH4Urd, in cells treated with dH4Urd, then the addition of dUrd has some of the benefits that accrue from prolonged infusions of profound effects on reversing the toxicity of FdCyd as seen in fluoropyrimidines may be gained in our approach. Prolonged Table 1. The ability of the 2 pyrimidine deoxyribonucleosides, administration of FUra has resulted in less incorporation of FUra dUrd and dThd, to effectively antagonize FdCyd toxicity indicates into bone marrow nucleic acid and less myelotoxicity (16), similar that thymidylate synthetase is the major target of FdCyd when to that obtained with ftorafur. As cells are recruited into S phase, it is coadministered with H4Urd. modulating concentrations of H4Urd or dH4Urd coadministered Our past studies demonstrated that FdCyd is an excellent with FdCyd may, in effect, result in FdCyd serving as a prodrug substrate for dCyd kinase. Other studies, in several laboratories, of FdUMP and result in greater retention of FdUMP in target demonstrated that mutants which lack dThd kinase retain their tissue. Retention of FdUMP in tumor tissue has been shown to sensitivity to FdCyd. When the findings of these and other recent

JUNE 1984 2557

Downloaded from cancerres.aacrjournals.org on September 27, 2021. © 1984 American Association for Cancer Research. J. A. Me/eras ef al. studies (48) are integrated, the pathway shown in Chart 1 may result not only in toxicity but also in a self-defeating inhibition appears to be the pathway of anabolism of FdCyd when H4Urd of the key converting enzyme. This has been considered in our is (Â»administered; direct confirmation awaits pharmacokinetic tumor therapy protocols. On the other hand, this phenomenon studies and evidence that counts derived from FdCyd are not may be responsible for adding selectivity to our strategy. Normal incorporated into RNA. tissue, with its low levels of dCMPD may shut off the conversion Further evidence that FdCyd will be channeled towards of FdCMP to FdUMP; in contrast, the high levels of dCMPD in FdUMP, especially in tumors, is derived from the studies in which tumors may remain sufficiently active with this moderate level of it was shown that dCyd kinase is elevated in many tumors and inhibition. Another important consideration is that FdUMP will transformed tissues (19,25, 30). dCyd kinase activity correlates bind much more avidly to thymidylate synthetase, the target enzyme, than it will to dCMPD, (KÂ«,for FdUMP, 10'12 M) (21). with the rate of tumor growth (19) and is not elevated in the small intestine (25). Giusti er al. (18), in extensive studies, showed The ultimate consequence of these effector interactions could that dCMPD is markedly elevated in human malignant tumors. well be that they result in a sustained continuous conversion of These studies utilized many different tumors with regard to tissue FdCMP to FdUMP. Understanding the nature and importance of involved, multiple independent sources of tumor, and equivalent these regulatory factors is essential to this approach. It is impor normal tissue. They found a 20- to 80-fold elevation in dCMPD tant to determine the kinetic parameters in human tissue; for activity above that of normal or benign tumors. Normal liver was example, to what extent dCMPD from human tumors and normal an exception and also displayed elevated levels of dCMPD. tissue is inhibited by FdUMP. Mancini and Cheng (35), utilizing Machara ef al. (30) obtained less dramatic, but nonetheless human dCMPD from blast cells of individuals with acute myeloge- substantial, increases in dCMPD activity in human tumor tissue nous leukemia, found FdUMP to be a very weak inhibitor of that (3- to 6-fokJ). These authors did not detect any significant ele enzyme. Indeed, it was the weakest inhibitor in the halogenated vation of the enzyme in normal human liver. Cummins and series of dUMP analogues. Balinsky (12) also described a 2- to 6-fold increase in dCMPD in Deoxycytidylate kinase, the enzyme with the capacity to phos- human hepatomas over that of normal liver and commented on phorylate dCMP, has been shown to possess a very high Km (6 the fact that normal human liver possesses higher levels of dCMP â€¢104 M) with respect to its substrate and is not usually elevated than does rat liver. Our enzymatic studies in the mouse, and in tumors (1, 26). If the addition of a fluorine atom to position 5 those of Maley and Maley (32) and Cummins and Balinsky (12) does not affect binding to dCMP kinase, then there is further in the rat, indicate a marked (10- to 75-fold) elevation of dCMPD assurance that FdCMP will be channeled to FdUMP rather than in rodent tumor tissue over that of normal tissue. If the findings to FdCDP. The findings of Arima ef al. (1) also suggest that the of Giusti ef al. are more representative of the dCMPD of human dCMPD pathway is the predominant pathway in malignant tis liver, then it may be argued that the mouse is not a suitable or sues. Their studies indicated that a dUMP-dTMP kinase, rather stringent model for toxicity studies. However, the human liver than a dCMP kinase, is the more prevalent monophosphate possesses low levels of dCyd kinase (25, 30); in addition, the kinase in malignant tumors of a rat sarcoma. human liver may not be dependent for its function on DMA If FdCMP was not deaminated by dCMPD and was further replication as are bone marrow and intestine, which have low phosphorylated to the triphosphate, it would probably be incor levels of dCMPD. Cummins and Balinsky (12) have shown that porated into DNA. Studies by Tanaka ef al. (47) have shown that both thymidylate synthetase and DNA polymerase activities were 5-FdCTP is utilized by a and .>'DNA polymerases of calf thymus low and, in some cases, undetectable in normal human liver, with K,,,s comparable to that of natural substrates. Jones and indicating that little or no DNA synthesis occurs in that organ. In Taylor (27) have shown that cultured mouse embryo cells, when addition, ftorafur is converted to a proximate or intermediate treated with FdCyd, were stimulated to differentiate into muscle antimetabolite in the liver, yet it is heralded as a drug less toxic cells and that the methylation of DNA was inhibited, although to than FUra. Cytoxan is another drug which is activated in the liver a lower extent than that obtained with 5-azacytidine or 5-aza- 2'-deoxycytidine. This is indirect evidence that FdCyd is incor that is successful in cancer therapy. Due to the profound effector control of dCMPD exerted by porated into DNA as such, in view of the fact that FdUrd did not TTP (K) 10~7 M) (34), the lower TTP pools that result from the have similar effects. Recent studies from our laboratory (15) with inhibition of thymidylate synthetase by FdUMP should lead to cells transformed by a restriction fragment containing the dThd greater activity of dCMPD and to further conversion of FdCMP kinase gene of herpes simplex virus type 1, but containing host to FdUMP. Another consideration regarding effector control is cell DNA polymerases, indicate that FdCyd would very probably that tumor cells with high levels of dCMPD may possess low be incorporated into DNA as such in vivo, without the necessity levels of dCTP so that dCyd kinase would not be end product for prior deamination in view of the fact that we demonstrated inhibited, thereby favoring the anabolism of FdCyd to FdCMP. (32) that 5-iododeoxycytidine and 5-bromodeoxycytidine are in Fortunately, dCMPD is only very weakly inhibited by its product, corporated as such into the DNA of these transformed cells dUMP (K, 1.1 x 10"3 M) (26) which would accumulate in a cell in which are unique in that they are capable of phosphorylating the which thymidylate synthetase is inhibited by FdUMP. 5-halo, 4-aminopyrimidine deoxyribonucleosides. A complexity that one may encounter in this strategy is the A complexity that must be considered in strategies utilizing moderate competitive inhibition of dCMPD by FdUMP that was FdCyd and 5-trifluoromethyldeoxycytidine is that, upon incorpo reported in one study utilizing a human lymphoblastic leukemia ration into DNA, these analogues may affect the maintenance cell line as source of the enzyme (Ki for FdUMP of 2.3 x 10~5 M methylation patterns of DNA and thereby affect gene expression. compared to a Kâ€žfor dCMP of 1.2 x 10~4 M) (14). Utilizing a FdCyd has been shown to affect differentiated functions (27), dCMPD preparation from chick embryos, Maley (31) obtained a presumably by activating silent genes. Even in the absence of Kn, for FdCMP of 1.9 x 10~3 M. High concentrations of FdCyd dH4Urd or H4Urd, there may be significant incorporation of FdCyd

2558 CANCER RESEARCH VOL. 44

Downloaded from cancerres.aacrjournals.org on September 27, 2021. © 1984 American Association for Cancer Research. Target- and DNA-directed Therapy with FdCyd into DNA; e.g., Jones and Taylor (27) did not (Â»administer egies are more selective than that involved in the use of ftorafur, inhibitors of deamination in their studies of the effect of FdCyd for that prodrug is converted to its active form by liver micro- on differentiated functions in cell culture. The problems of somes rather than by an enzyme that is elevated in tumors. changes in maintenance methylation which should be investi If a tumor possesses low levels of uracil phosphoribosyltrans- gated may not be unique to therapy with 5-halodeoxycytidine ferase, pyrimidine phosphorylases, and dThd kinase, so that analogs, for when any agent, including FdUrd, invites repair of FUra and FdUrd therapy would not be appropriate, then the DNA, the possibility exists that the cytosines incorporated into approach described in this paper offers an alternate course of the repaired patch of DNA will not be methylated, and thereby action if dCyd kinase and dCMPD levels are high, as they are in the patterns of maintenance methylation will be lost (28). many tumors. We propose that the use of FdCyd and 5-trifluo- Although one may readily assume that the high levels of romethyldeoxycytidine with H4Urd, an inhibitor of cytidine de dCMPD in tumors serve primarily to provide a secondary source aminase, will result in a novel strategy of exclusive DNA-directed of dUMP for accelerated DNA synthesis, another function for tumor therapy by fluoropyrimidines that are resistant to catabo- elevated dCMPD activity is that it could provide a growth advan lism. FdCyd when (Â»administered with H4Urd will be preferen tage to tumor tissue by preventing the reutilization of 5-methyl- tially anabolized by dCyd kinase to the nucleoside FdCMP and dCyd that is available as a result of degradation of DNA in the will be preferentially converted in target tissue by dCMP deami necrotic tissue of an enlarging tumor mass. It should be noted nase, an enzyme that is markedly elevated in human tumors. that 5 to 9% of the cytosine residues in DNA are methylated. The antimetabolites generated will inhibit thymidylate synthetase Because 5-methyl-dCyd displays a relatively low Â«â€ž,withrespect and invite repair in DNA; this results in tumor inhibition. We to dCyd kinase (15), exclusion of 5-methyl-dCyd from inappro propose that the anabolism of these analogues is via a different priate sites in DNA may reside in the heightened activity of subset of enzymes than that utilized by fluoropyrimidines in dCMPD in tumors. current use and that our approach with analogues of deoxycyti- An additional consideration regarding the strategy of utilizing dine involves greater selectivity than do current approaches with FdCyd and H4Urd concerns possible mutagenic effects due to FUra or FdUrd. The efficacy that we have established with FdCyd nucleotide pool imbalances caused by FdUMP. In view of the and H4Urd warrants further tumor inhibition studies and associ higher dCTP/TTP ratios in FdUMP-inhibited cells, Peterson ef al. ated pharmacokinetic, biochemical, enzymatic, and incorporation (43) have suggested that mutagenesis, which occurs because studies so that the concepts and strategies that we have devel of this imbalance, may be prevented by the utilization of an oped can be utilized in future clinical trials. inhibitor of dCyd kinase. The use of FdCyd would result in some inhibition of the phosphorylation of dCyd, although it is merely a REFERENCES competitive inhibitor. Peterson ef a/, have indicated that muta genesis by this mechanism could lead to resistance to fluoropy- 1. Arima, B., Akiyoshi, H., and Fujii, S. Characterization of pyrimidine nucleoside monophosphokinasein normal and malignanttissues. Cancer Res., 37:1593- rimidine therapy. Although it appears unlikely that the addition of 1597, 1977. H4Urd would result in an increase in dCTP pools, this must be 2. Armstrong, R., and Diasio. R. Selective activation of 5'-deoxy-5-fluorouridine by tumor cells as a basis for an improved therapeutic index. Cancer Res., 41: investigated in cells that are grown in the presence of FdCyd. 4891-4894,1981. Tumors responsive to FdCyd + H4Urd therapy, having high 3. Balzarmi,j.. and DeClercq.E. D. Role of deoxycytidine kinase in the inhibitory activity of 5-substituted 2'-deoxycytidmes and cytosine arabmosideson tumor levels of dCyd kinase and dCMPD, have an enzymatic profile cell growth. Mol. Pharmacol.23:175-181,1982. that would also result in a response to ara-C (not (Â»administered 4. Block, A., and Hutchinson, D. J. A mechanism of resistance to fluoropyrimi with FdCyd), since elevated dCMPD would result in low dCTP dines. Cancer Res., 24: 433-439,1964. pools. Because ara-UTP that is formed via the ara-CMP -Â»ara- 5. Bresnick, E., and Thompson, U. Properties of deoxythymidine kinase partially UMP pathway may inhibit dUTPase,4 this inhibition may lead to purified from animal tumors. J. Bid. Chem., 240: 3967-3974,1965. 6. Brockman, R. Mechanisms of resistance to anticancer agents. Adv. Cancer increased incorporation of FdUrd and dUrd into the DNA of tumor Res., 7: 129-234,1963. cells, thereby inviting repair and the enhanced formation of single- 7. Camiener, G. W. Studies of the enzymatic deamination of ara-cytidine-V. Inhibitionin vitro ana in vivo by tetrahydrouridineand other reduced pyrimidine strand breaks. In addition, the inhibition may block a possibly nudeosides. Biochem. Pharmacd., 17:1981-1991,1968. 8. Chadwick. M., and Chang, C. Comparative pharmacologicaldispositions of 5- important source of dUMP, thereby making the inhibition of f)uoro-2'-deoxyuridine and 5-fluorouracilin mice bearing solid L1210 lympho- thymidylate synthetase by FdUMP more effective. Regarding the cytic leukemia. Cancer Treat. Rep., 60: 845-855,1976. use of FdCyd and H4Urd in combination chemotherapy, in view 9. Cooper, G., and GrÃ©er,S.The effect of inhibition of cytidine deaminase by tetrahydrouridineon the utilization of deoxycytidineand 5-bromodeoxycytidine of the studies of Maybaum ef al. (36), the utilization of an entirely for DNA synthesis. Md. Pharmacd., 9: 698-703,1973. DNA-directed approach will mean that cells will not be arrested 10. Cooper, G., and GrÃ©er,S.Phosphorylation of 5-hatogenated deoxycytidine in GÃ¬,a result of RNA-directed events. Therefore, new consid analogs by deoxycytidine kinase. Md. Pharmacd., 9: 704-710,1973. 11. Cramer,J., Prusoff, W., and Welch, A. 5-Bromo-2'-deoxycytidine (BCDR)â€”II. erations regarding combination chemotherapy may be necessary Studies with murine neoplastic cells in culture and in vitro. Biochem. Pharma when FdCyd and H4Urd are utilized. cd., 8: 331-335.1961. Our strategy with FdCyd and H4Urd is somewhat analogous, 12. Cummins, R., and Balinsky, D. Activities of some enzymes of pyrimidine and DNA synthesis in a rat transplantablc hepatoma and primary hepatomas, in although not as qualitatively selective, to the chemotherapeutic cell linesderived from these tissues, and in human fetal liver. Cancer Res., 40: approach used against Candida infections with 5-fluorocytosine, 1235-1239,1980. 13. Dooersen, M., and GrÃ©er,S.HSV type 2 induced pyrmidinenucleosidekinase: an analogue that is converted to FUra by cytosine deaminase, enzymatic basis for the selective antiherpetic effect of 5-halogenatedanalogs an enzyme absent in humans but present in yeast. The approach of deoxycytidine. Biochemistry, 17: 920-928,1978. is also similar to some extent to the utilization of 5'-deoxy-5- 14. Drake, J., Hande, K., Fuller, R., and Chabner, B. Cytidine and deoxycytidylate deaminationinhibition by uridine analogs. Biochem. Pharmacd., 29: 807-811, fluorouridine which appears to be converted to FUra in tumors 1980. possessing high levels of uridine phosphorylase (2). These strat- 15. Fox, L, Dobersen, M. J., and GrÃ©er,S.Incorporation of 5-substituted analogs of deoxycytidine into DNA of herpes simplex virus-infected or -transformed 4T. A. Bolton, Universityof South Florida,Tampa, FL, personal communication. cells without deaminationto the thymidine analog. Antimicrob. Agents Chem-

JUNE 1984 2559

Other., 23: 465-476,1983. 33. Maley, F., and Maley, G. The regulatory influence of altosteric effectors on 16. Fraile. R., Baker, L, Buroker, T., Horwitz, J., and Vaitkevicius, V. Pharmaco- deoxycytidylate deaminases.Curr. Top. Cell. Regul., 5:177-228,1972. kinetics of 5-fluorouracil administered orally by rapid intravenous and by slow 34. Maley, G., and Maley, F. Nucleotide interconversions. X. Deoxyribo- and infusion. Cancer Res., 40:2223-2228,1980. nbonucleoside 5'-phosphate synthesis via a phosphotransferase reaction in 17. Fnedkin, M,. and Roberts, D. The enzymatic synthesis of nucleosides II. chick embryo extracts. Arch. Biochem. Biophys., 707: 342-349,1963. Thymidine and related pyrimidine nucleosides. J. Bid. Chem., 207: 257-266, 35. Mancini,W. R., and Cheng,Y. C. Humandeoxycytidylate deaminase:substrate 1954. and regulator specificities and their chemotherapeutic implications. Mol. Phar- 18. Giusti, G., Mangoni, C., De Petrocellis. B . and Scarano, E. Deoxycytidme macol., 23:159-164,1982. deaminase in normal and neoplastic human tissues. Enzymoi. Biol Clin., 11: 36. Maybaum,J., Ullman.B., Mandel, H. G., Day, J. L, and Sadee, W. Regulation 375-383,1970. of RNA-and DNA-directedactionsof 5-tluoropynmidinesin mouse T-lymphoma 19. Harkrader, R. J., Jackson, R. C., Ross, D. A., and Weber, G. Increase in liver (S-49)cells. Cancer Res., 40:4209-4215,1980. and kidney deoxycytidme kinase activity linked to neoplastia transformation. 37. Mekras, J., and Greer, S. Use of 5-fiuorodeoxycytidme and tetrahydrouridine Biochem. Biophys Res. Commun., 96: 1633-1639,1980. to exploit high levels of deoxycytidylate deaminase in tumors: target-directed 20. Hartmann, K. U., and Heidelberger, C. Studies on fluorinated pyrimidines. J. conversion of a nontoxic analog to a toxic antimetabohte.Fed. Proc . 42:360, Biol. Chem., 236: 3006-3013,1961. 1983. 21. Heidelberger,C. Fluorinated pyrimidines and their nucleosides. In: Handbook 38. Mukherjee, K., and Heidelberger,C. Inhibition of the incorporation of formate- of Experimental Pharmacology,Vol. 38, pp. 193-231. Berlin, Heidelberg,and "C into DNA thymine of Ehrlich ascites carcinoma cells by 5-fluoro-2-deox- New York: Springer-Verlag,1975. yuridine-5'-monophosphate and related compounds. Cancer Res., 22: 815- 22. Heidelberger, C., and Anderson, S. Fluorinated pyrimidines XXI. The tumor- 822,1962. inhibitory activity of 5-trifluoromethyl-2'-deoxyuridine. Cancer Res., 24:1979- 39. Myers, C. The pharmacology of the fluoropyrimidines.Pharmacol. Rev., 3:1- 1985.1964. 15,1981. 23. Heidelberger,C., Danenberg,P. V., and Moran, R. G. Fluorinatedpyrimidines 40. Myers, C., Young, R., and Chabner, B. Biochemicaldeterminants of 5-fluoro and their nucleosides.Adv. Enzymoi., 54:57-120,1983. uracil response in vivo. The role of deoxyuridylate pool expansion. J. Clin. 24. Heidelberger, C., Griesbach.L., Montag, B., Mooren, D., Cruz, 0., Schnitzer, Invest., 56:1231-1238,1975. R., and Grunberg, E. Studies on fluorinated pyrimidines. II. Effects on trans 41. Myers, C., Young, R., and Chabner, B. Kinetic alterations induced by 5- planted tumors. Cancer Res., 78: 305-317,1958. fluorouracil in bone marrow, intestinal mucosa, and tumor. Cancer Res., 36: 25. Ho, D. Distribution of kinase and deaminaseof 1-.i-o-arabmofuranosylcytosme 1653-1658,1976. in tissues of man and mouse. Cancer Res., 33:2816-2820,1973. 42. Newman, E.. and Santi, D. Metabolism and mechanism of action of 5-fluoro- 26. Jackson, R. C. The regulationof thymidylate biosynthesisin Novikoffhepatoma deoxycytidme. Proc Nati. Acad. Sci USA, 79: 6419-6423,1982. cellsand the effects of amethopterin,5-fluorodeoxyuridine,and 3-deazaurÂ¡dine. 43. Peterson, A. R., Peterson, H., and Danenberg,P. V. Induction of mutations by J. Bnl. Chem., 253: 7440-7446,1978. 5-fluorodeoxyuridine: a mechanism of self-potentiated drug resistance? 27. Jones, P., and Taylor, S. Cellular differentiation, cytidme analogs and ONA Biochem. Biophys. Res. Commun., 770: 573-577,1983. methylation CeÂ«,20:85-93,1980. 44. Robert de Saint Vincent, B., Dechamps, M., and Buttin, G. The modulation of 28. Kastan. M., Gowans, B . and Lieberman. M. Methylation of deoxycytidine the thymidmetriphosphate pool of Chinesehamster cete by dCMP deaminase incorporated by excision-repairsynthesis of DNA. Cell, 30: 509-516,1982. and UDP reducÃase.J.Biol. Chem., 255: 162-167,1980. 29. Laskin, J. D., Evans, R. M., Slocum. H. K., Burke, D., and Hakala.M. T. Basis 45. Rustum, Y., Danhauser, L., and Wang, G. Selectivity of action of 5-FU: for natural variationin sensitivity to 5-fluorouracil in mouse and human cells in biochemicalbasis. Bull. Cancer (Paris),66:43-47,1979. culture. Cancer Res., 39: 383-390,1979. 46. Spears, C., Shahinian,A., Moran, R., Heidelberger,C., and CorbeÂ«,T.In vivo 30. Maehara,Y., Nakamura, H., Nakane. Y , Kawai, K., Okamoto. M., Nagayama, kinetics of thymidylate synthetase inhibition in 5-fiuorouracil-sensitive and S., Shirasaka. T., and Fujii, S. Activities of various enzymes of pyrimidine -resistant murine colon adenocaroinoma.Cancer Res., 42:450-456,1982. nucleotide and DNA synthesis in normal and neoplastiahuman tissues. Gann, 47. Tanaka. M., Yoshida, S., Saneyoshi. M., and Yamaguchi. T. Utilization of 5- 73: 289-298,1982. fluoro-2'-deoxycytidine triphosphate in DNA synthesis by DNA polymerase 31. Maley, F. Deoxycyt.dylatedeaminase.Methods Enzymoi. 72:170-182,1967. and from calf thymus. Cancer Res., 41:4132-4135,1987. 32. Maley, F., and Maley. G. Nucleotide interconversions. IV. Activities of deoxy 48. Yoshida, M., Hoshi, A., Kuretani, K., and Saneyoshi, M. Mode of action of 5- cytidylate deaminase and thymidylate synthetase in normal rat liver and fluorocytidine and 5-fluoro-2'-deoxycytidine in L5178Y cells in vitro. Chem. hepatomas. Cancer Res., 27:1421-1426,1961. Pharm. Bull. (Tokyo), 30:1018-1023,1982.

2560 CANCER RESEARCH VOL. 44

Downloaded from cancerres.aacrjournals.org on September 27, 2021. © 1984 American Association for Cancer Research. Use of 5-Fluorodeoxycytidine and Tetrahydrouridine to Exploit High Levels of Deoxycytidylate Deaminase in Tumors to Achieve DNA- and Target-directed Therapies

John A. Mekras, David A. Boothman, Liliana M. Perez, et al.

Cancer Res 1984;44:2551-2560.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/44/6/2551

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://cancerres.aacrjournals.org/content/44/6/2551. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.