NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Kufe DW, Pollock RE, Weichselbaum RR, et al., editors. Holland-Frei Cancer Medicine. 6th edition. Hamilton (ON): BC Decker; 2003.

  • By agreement with the publisher, this book is accessible by the search feature, but cannot be browsed.
Cover of Holland-Frei Cancer Medicine

Holland-Frei Cancer Medicine. 6th edition.

Show details

Pyrimidine Analogs

, PhD, PharmD, , MD, and , MD.

Pyrimidine analogs include 5-fluorouracil, cytosine arabinoside, 5-azacytidine, and gemcitabine.


Background and Properties

A major motivation for the development of pyrimidine analogs of uracil was the early observation that preneoplastic rat liver and hepatomas incorporated uracil more actively than did the normal liver.11 Although this may reflect a difference in the relative degradative capacity of these different tissues for uracil, it also provided a focus for the synthetic efforts of Duschinsky and colleagues that led to 5-fluorouracil (5-FU) (Figure 50-1) and a family of related fluorinated pyrimidines.12 This specific site of substitution on the pyrimidine ring was selected because it might inhibit subsequent conversion of a uracil nucleotide to thymine nucleotides. Because insertion of the methyl group occurs on the 5-position, halogen replacement of hydrogen in that position was thought to have a greater chance of inhibiting DNA synthesis and, thus, growth. The selection of fluorine to replace the hydrogen in uracil was based on their similar van der Waals radii (F = 1.47 Å and H = 1.20 Å). Unlike earlier syntheses of halogenated pyrimidines, which involved simple displacement of the hydrogen with the halogens chlorine, bromine, or iodine, 5-FU was originally synthesized from an acyclic precursor. This permitted formation of the corresponding 5-fluoroorotic acid; subsequently, the ribosides and deoxyribosides of 5-FU were prepared (see Figure 50-1). More recently, a direct means of fluorinating 5-FU has been developed that permits positron emission tomography studies with [18F]-5-FU.13

Figure 50-1. 5-Fluorouracil (5-FU) and analog structures.

Figure 50-1

5-Fluorouracil (5-FU) and analog structures.

As anticipated, the pKa of 5-FU (8.1) is more acidic than that of uracil (9.6); thus, under physiologic conditions, 5-FU partially exists as an anionic species. This is undoubtedly important to the metabolic activation to the nucleotide form via the orotidylate pyrophosphorylase reaction. This uridylate analog, 5-fluorouridine monophosphate (5-FUMP), can then substitute for uridine monophosphate (UMP) in a wide spectrum of intermediary reactions. The product of one of these reactions, fluorodeoxyuridine monophosphate (FdUMP), plays a major role by inhibiting displacement of hydrogen from the 5-position of deoxyuridylate and replacing it with a methyl group via a tetrahydrofolate catalyzed reaction (Figure 50-2).14 Many of the properties predicted for 5-FU were seen in early studies of bacterial and model tumor systems, and a remarkably rapid progression to a clinical trial occurred within 2 years of its synthesis.15 These early clinical studies showed enough promise in colon cancer and other solid tumors to sustain 40 subsequent years of further development. A primary focus of this research has been to reduce its very real toxicity to a variety of normal tissues, while retaining its antitumor activity. Today, 5-FU remains an important component in the therapy of several of the most common solid tumors, not only as a single agent but also in combination with other chemotherapy agents.

Figure 50-2. Covalent thymidylate synthase-fluorodeoxyuidylate complex; R = H or CH2FH4 = 5, 10-methylene tetrahydrofolate.

Figure 50-2

Covalent thymidylate synthase-fluorodeoxyuidylate complex; R = H or CH2FH4 = 5, 10-methylene tetrahydrofolate.

Cellular Entry and Efflux Mechanisms

Limited evidence suggests that 5-FU enters cells by a carrier-mediated transport mechanism.16 Early reports suggested that a specific mechanism for the transport of uracil existed in the intestine; however, these studies used methods that made it difficult to distinguish between transport and metabolism. Evidence has been presented for a nonconcentrative transporter in the Novikoff hepatoma that exhibits competitive kinetics between uracil and 5-FU.17 Under conditions in which the 5-FU ring is minimally ionized, enhanced entry of 5-FU occurs if cells are preloaded with uracil, which is consistent with a countertransport mechanism. Using standard analytic techniques, no evidence to date suggests that an alteration of 5-FU entry into cells is responsible for either natural or acquired resistance. However, use of more sophisticated methods has revealed a different picture of 5-FU uptake and retention. Using [19F]-5-FU, a difference in the ability of selected tumors to accumulate free 5-FU was noted to correlate with their response to chemotherapy.18–20 Extension of these studies to four patients with breast and colon carcinoma indicated a half-life of 0.4 to 2.1 h for free 5-FU in the tumor compared with a plasma half-life of less than 10 to 15 min. Independent studies using gas chromatography/mass spectroscopy (GC-MS) documented free 5-FU concentrations in normal and neoplastic tissue that were at least tenfold higher than those in plasma. This study also revealed that after an initial, rapid clearance from plasma, it was possible to detect a second, longer half-life of approximately 3.5 h.21,22 These new observations on the trapping of 5-FU in tumors lend support to the view that 5-FU is transported into the cells by an active transport mechanism16 as well as by a facilitated diffusion mechanism.17,23 Free 5-FU could also be concentrated in the cytoplasm (pH, 7.2) from extracellular spaces of tumors rendered acidic by anaerobic glycolysis (pH, 6.2–7.0) by virtue of ionization trapping of this pyrimidine analog, which has a pKa of 8.1.24 An alternative source would be a slow liberation of free 5-FU from nucleotides and nucleic acids; this liberation sustains an intracellular concentration because of the limited efflux of free 5-FU from the cells. This capacity for trapping free 5-FU may serve as a measure of potential clinical response and deserves further study.

In contrast to 5-FU, the entry of fluorodeoxyuridine (FdUrd) (see Figure 50-1) into most neoplastic cells involves the saturable but nonconcentrative mechanism that is responsible for the facilitated diffusion of a wide spectrum of nucleosides.25 This transporter has been quantified in several cell lines by titration with p-nitrobenzylthioinosine.26 Deletion of this transport mechanism is the basis for resistance to FdUrd27 or purine nucleoside analogs28 in at least two cell lines. Such a deletion makes the cells collaterally sensitive to methotrexate and other inhibitors of thymidylate synthase (TS) because they are unable, or limited in their ability, to salvage thymidine, whether naturally available or administered.29 Fluorouridine and FdUrd released from 5-fluorouridylic acid and 5-fluorodeoxyuridylate by phosphatase action exit the cell via this same facilitated diffusion transporter. Thus, agents that affect this transporter may selectively affect 5-FU cytotoxicity by a differential effect on specific normal or neoplastic cell types. The facilitated diffusion mechanism may play a secondary role in the modulation of 5-FU action in vivo by uridine because this normal nucleoside, but not 5-fluorouridine (5-FUrd) or FdUrd, is actively concentrated by a Na+-dependent system.30 Neoplastic cells appear to be less capable of this transport and are not protected.


Once inside the cell, 5-FU has several possible routes of activation to the nucleotide form.31 In normal tissues, the predominant mechanism appears to be competition with orotate for condensation with pyrophosphorylribose-5-PO4 (PRPP) via orotidylate pyrophosphorylase to form 5-fluorouridylate.32 In mammalian cells, this protein is a bifunctional enzyme that also catalyzes the decarboxylation of orotidylate to 5′-uridylic acid.32 5-FU can successfully compete with the very low physiologic concentrations of orotate in this reaction because of its acidic pKa (8.1), which generates a significant amount of anionic species.

Alternative activation routes of 5-FU follow the salvage pathways for uracil and thymine, but these are presumed to be less important in most tissues.33–35 The first enzyme in the pathway, uridine phosphorylase, condenses ribose-1-P with uracil or 5-FU in a reaction that energetically favors synthesis; normally this is catabolic in the cell because further reactions such as of PRPP synthesis and phosphatases reduce the concentration of ribose-1-P. The corresponding reaction for thymine uses deoxyribose-1-P, but it is not considered to make a significant contribution to 5-FU activation in current therapeutic regimens. After formation of the nucleoside, phosphorylation by uridine kinase and adenosine triphosphate (ATP) forms 5-fluorouridine monophosphate (5-FUMP) (Figure 50-3). Further phosphorylation of 5-FUMP to 5-fluorouridine diphosphate (5-FUDP) by nucleotide kinase provides a branch point in 5-FU anabolism.36 Additional phosphorylation of a major portion of 5-FUDP to fluorouridine triphosphate (5-FUTP) provides the substrate for RNA polymerases with consequent incorporation into several forms of RNA.37 Alternatively, 5-FUDP can be reduced to 5-fluorodeoxyuridine diphosphate (FdUDP), which is hydrolyzed to the monophosphate FdUMP, the covalent inhibitor of thymidylate synthase.14 Some FdUDP is phosphorylated to fluorodeoxyuridine triphosphate (FdUTP), which is an alternate substrate for deoxythymidine triphosphate (dTTP) in DNA polymerase reactions; however, high deoxyuridine triphosphate (dUTP) pyrophosphatase activity converts most of the FdUTP to FdUMP.38 When 5-FU is incorporated into DNA, uracil N-glycosylase removes it, leaving an apyrimidinic sugar for the process of DNA repair. Errors in this process provide an additional basis for cytotoxicity.39

Figure 50-3. Metabolic activation and targets of fluorinated pyrimidines.

Figure 50-3

Metabolic activation and targets of fluorinated pyrimidines. CO2 = carbon dioxide; dT = thymidine (thymine deoxyriboside); dTMP also called thymidylate; dU = deoxyuridine; dUMP = deoxyuridine monophosphate; FBAL = fluoro-β-alanine; FdU = fluorodeoxyuridine; (more...)

Minor amounts of 5-FUDP sugar derivatives have been detected as anabolic products, but their potential to inhibit cell growth or toxicity has not been documented.40,41 In some of the previously discussed reactions, the analog 5-FU nucleotides are better substrates than the corresponding uracil derivatives.


Consideration of 5-FU pharmacokinetics must focus primarily on the balance between anabolism and catabolism. The conversion to nucleotide derivatives is responsible for most, if not all, of its antineoplastic activity, even though it accounts for a very minor portion of the administered drug. Catabolism via the normal degradation pathway for uracil is the immediate fate of more than 80% of an administered dose of 5-FU.42 Therefore, slight alterations in this pathway can greatly affect the very limited amount that is available for conversion to the nucleotide form.

Because of the appearance of great variability in response among patients and apparent limited bioavailability (10% to 25%) via the oral route,43 there has been a long-held recommendation that 5-FU should be administered intravenously (IV). The basis for this apparent poor bioavailability has not been well understood, particularly since the low molecular weight and pKa of 5-FU should predict excellent absorption. Recent clinical studies in which 5-FU was administered orally together with ethynyluracil—a potent inactivator of the initial enzyme of the pyrimidine pathway, dihydropyrimidine dehydrogenase (DPD)—have demonstrated that 5-FU in fact has excellent absorption and bioavailability44 with the variability from patient to patient being due to the variability of DPD levels in the population.

Dosage used clinically in general depends on the schedule of administration.45 The most common dosage schedules are a monthly course of one dose given on each 5 days as an IV bolus of 400 to 600 mg/m2 or the same dosage given as a single bolus on a weekly basis.46 The limiting toxic effect of these regimens generally is myelosuppression or mucositis. When continuous IV infusion is employed, higher doses are required (1,000 to 2,000 mg/m2/d) to sustain steady-state concentrations of 5-FU (1 to 5 μM) in plasma adequate to achieve therapeutic effects.47 With this route, toxicity is most frequently mucositis, with minimal myelosuppression. Several studies have shown that this regimen is superior to the bolus regimen when 5-FU is given as a single agent.47–49 Optimal treatment was a 48-hour infusion at weekly intervals, which improved both response and survival. Prolonged infusion of 5-FU for up to 12 weeks at 300 mg/m2/d also produced a better response than the bolus regimen.50,51 The most prominent toxicity in this situation was a reversible hand-foot syndrome.51 It was found that during continuous IV infusions, plasma concentrations of 5-FU varied by as much as tenfold, and subsequent studies have demonstrated that variations in dihydropyrimidine dehydrogenase may be responsible for this effect.52 Because 5-FU is most often used in combination with other agents such as leucovorin and methotrexate, it is important to modify the dosage in each case to limit, but not eliminate, host toxicity. It is generally thought that therapeutic benefit requires a dosage intensity that causes significant host toxicity, a result that has been documented in studies of colorectal cancer.53

Administered 5-FU has a volume of distribution of 0.20 to 0.25 L/kg, which suggests distribution into the extracellular space.54,55 Good penetration into the cerebrospinal fluid, lymph, and neoplastic effusions have been documented.56 Since the drug apparently freely permeates cells in culture, it is not clear why the volume of distribution approximates the extracellular space.

The rate of plasma clearance generally is first order with a half-life of 10 to 20 minutes and ranges between 500 and 1,500 mL/min.42 Above a dosage of 800 mg/m2, clearance may decrease rapidly. Because the primary fate of the drug is catabolism, this decreased clearance undoubtedly reflects saturation of these reactions.57 The circulating concentrations of the initial metabolite, dihydro-5-FU, can be much greater than those of 5-FU, and the fate of this metabolite may affect both the pharmacokinetics and response to 5-FU.

Intraarterial infusion of 5-FU has been used with some success in patients with isolated hepatic metastases. As with systemic therapy, extensive single-pass clearance is achieved (19 to 51%), but saturation of catabolism occurs when doses are elevated.58 Nevertheless, hepatic 5-FU concentrations considerably in excess of those tolerated systemically can be achieved. The limiting factor in high-dose regimens is cholestatic jaundice and evidence of chemical hepatitis.

The 2′-deoxyriboside of 5-FU, FdUrd, is a much more potent inhibitor of cell growth than 5-FU in cell culture.21 This presumably reflects the ease with which this compound can be activated by thymidine kinase in a single step to FdUMP, the titrating inhibitor of thymidylate synthase, which after further phosphorylation can also be incorporated into DNA. In both humans and animals, IV bolus injection of FdUrd produces a dose response that is essentially that of 5-FU because it is cleaved rapidly to an equivalent amount of 5-FU that subsequently experiences the same metabolic fate as directly injected 5-FU. If, however, FdUrd is given by a 14-day continuous infusion, the maximum tolerated dose is approximately 100-fold less58; however, its therapeutic index is not significantly better than that of 5-FU. Even so, it can be used for isolated hepatic metastases of colon cancer by hepatic artery infusion because approximately 90% of the drug is cleared in a single pass by the liver, thus reducing systemic effects.59 Using this route, major increases in the hepatic concentrations of intact drug are achieved relative to systemic targets of toxicity.

The only other approved preparation of 5-FU is in a 2 or 5% formulation in ethylene glycol or a water-based cream for topical application to treat epithelial dysplasias, particularly actinic keratoses and early basal cell carcinomas.60 Vulvar and vaginal epithelial neoplasms and genital condylomas also respond to this treatment.61,62 Insufficient drug is absorbed from these preparations to cause systemic effects, and reports of local drug kinetics have been limited. It is not clear which, if any, of the biochemical mechanisms detailed earlier are responsible for this therapeutic effect, nor has a reason for the rather selective action on lesions been established, except for their presumably more rapid cell kinetics. Immunologic reactions to 5-FU may play a role.

Although not directly useful in the treatment of cancer, the 4-amino derivative of 5-FU, 5-fluorocytosine (flucytosine) (see Figure 50-1), is a valuable antifungal agent in systemic infections, which are a common complication of antineoplastic therapy.63 5-Fluorocytosine is relatively nontoxic in mammalian systems because it cannot be activated by direct condensation with PRPP and, like uracil, is poorly anabolized by uridine-cytidine phosphorylase. However, pathogenic fungi, including Candida and Cryptococcus species, deaminate 5-fluorocytosine to 5-FU, which is lethal to the organisms, by the same mechanisms as in mammalian cells.64 Although resistant strains rapidly emerge, combination therapy with amphotericin B is valuable in systemic fungal infections. Unfortunately, however, some 5-fluorocytosine appears to be converted to 5-FU in the host, presumably by intestinal organisms, and this causes bone marrow depression, ie, leukopenias and thrombocytopenia.65 Evidence for its relative stability in humans is the observation that approximately 80% of an oral dose is excreted unchanged in the urine, compared to approximately 5% of a comparable dose of 5-FU.

Recently, to improve the specificity of chemotherapy treatment, minimize the systemic toxicity of the drug, and increase the local concentration of the antineoplastic agent at the tumor site, prodrug-activating gene therapy protocols have been developed to activate flucytosine into 5-FU by cytosine deaminase.

Cytosine deaminase is not a mammalian enzyme, but it is present in bacteria and fungi and can be utilized to produce high concentrations of 5-FU in tumors through the enzymatic deamination of flucytosine.66,67 This approach has been initially developed for the treatment of colorectal carcinoma metastatic to the liver, by utilizing a delivery system based on a replication-incompetent adenoviral vector.68 In vivo and in vitro evaluations of this gene therapy system have shown an increased sensitivity of colon carcinoma cells to flucytosine exposure. Even if the adenoviral vector carrying cytosine deaminase is transferred to a limited number of the total tumor cells (10%), a bystander effect is observed, likely due to the local diffusion of 5-FU generated in the virus-infected cells.69–72

Catabolic Reactions

The primary clearance mode of 5-FU is via catabolism along the degradative pathway for uracil.42 Because the products of this pathway do not absorb ultraviolet light, GC-MS or radioisotopic methods must be employed. The initial reaction is reduction by dihydrouracil dehydrogenase. The liver is a major site of 5-FU metabolism, and this is particularly true when the drug is given orally, intraperitoneally, or by intrahepatic arterial infusion. It is now recognized, however, that metabolism in the lung and kidneys may be of equal, or even greater, importance after IV administration.55 These findings have therapeutic relevance because it was previously felt that hepatic metastases might compromise 5-FU clearance and limit dosage.

Marked circadian variations in the metabolism of 5-FU have been detected related to 24-hour cyclic variations in dihydrouracil dehydrogenase activity.52,73 These changes are reflected in the inverse variations of plasma 5-FU concentrations during IV infusions in humans.52,74 Means to employ these differences in the design of clinical protocols have been outlined.75 Preclinical data in murine models have indicated that less toxicity was encountered during a circadian infusion when the maximal concentration of 5-FU was programmed to occur at 4 am.76,77 More recent data indicate that if the maximal concentration is programmed for 9 to 10 pm, even less toxicity is observed than with the previous schedule.78,79 Several clinical protocols comparing a continuous flat infusion with the circadian schedule have been conducted with 5-FU alone and in combination with leucovorin and/or platinum derivatives.80–82 A clinical study using a 14-day continuous infusion of 5-FU and leucovorin suggests that circadian administration with a maximal infusion rate at 4 am increases the maximum tolerated dose (MTD) for both agents: 5-FU, 250 mg/m2/d; leucovorin, 20 mg/m2/d.83 In patients who experienced grade 2 or higher toxicities with this schedule, the peak of their circadian infusion was moved to 9 to 10 pm. Decreased toxicity was observed (mostly diarrhea and stomatitis), and the MTD for 5-FU increased to 300 mg/m2/d, a 50 mg increment over the MTD for a flat continuous infusion.84

Dihydropyrimidine dehydrogenase represents the initial rate-limiting step in the catabolism of the pyrimidines uracil, thymine, and 5-FU. More than 85% of an administered dose of 5-FU is eliminated with rapid formation of dihydrofluorouracil.42 In a small percentage of the population, less than 3%, DPD activity is significantly below the average (below 50% of the control mean). This pharmacogenetic condition, which typically goes undetected until administration of 5-FU, can cause very serious life-threatening toxicity in patients following 5-FU-based chemotherapy; the toxicity is due to increased exposure to and activation of the anticancer agent.85 The variability in DPD activity in normal tissues of the liver and gastrointestinal tract has been recently linked to the erratic oral bioavailability of 5-FU. DPD inhibitors, such as ethynyluracil, have been recently developed in an attempt to increase 5-FU efficacy and improve oral absorption although unfortunately this has not been associated with equivalence to intravenous 5-FU regimens.86–88

The subsequent metabolic step, catalyzed by dihydropyrimidinase, yields β-fluoroureidopropionic acid. A wide variety of tumors apparently express high levels of this activity because they accumulate the subsequent degradation products β-ureidopropionic acid and β-alanine.89

α-Fluoro-β-alanine, the counterpart to the final product of uracil catabolism β-alanine, is the major urinary excretion product of 5-FU.90 In patients with cancer, this has been shown to be conjugated with bile acids and constitutes the primary biliary secretion product of 5-FU.91 It has been suggested that the chenodeoxycholate conjugate may be responsible for the biliary toxicity seen after large-dose, intrahepatic infusion of 5-FU, and cholestasis associated with this conjugate has been demonstrated in isolated, perfused rat livers.91 A summary of 5-FU metabolism is shown in Figure 50-3.

Mechanisms of Action

Experimental evidence has suggested numerous sites for the biologic action of 5-FU (see Figure 50-3). The relative importance of each varies widely among different normal tissues and neoplasms. Commonly, the effects are divided into RNA- or DNA-directed toxicity.


The predominant phosphorylated nucleoside of 5-FU, 5-FUTP, is as good a substrate as uridine triphosphate (UTP) for several RNA polymerase reactions. The degree of 5-FUTP incorporation into RNA bears a direct relationship to its concentration relative to that of the normal substrate UTP. In cell lines, greater incorporation is associated with reduced clonogenic survival.92,93 Very substantial amounts of 5-FU replacement of uracil have been reported in each of the RNA species; the highest degree of incorporation generally is seen in the 4S-RNA.37 Some evidence suggests that with a given cell type, the proportion of RNA incorporation in different species depends on the available form of the analog, 5-FU versus FdUrd; this result suggests compartmentalization or channeling of the analog en route to incorporation.94

What is less clear about incorporation into RNA is its contribution to cytotoxicity. Earlier studies indicated effects on transfer RNA acceptor activity, miscoding of protein synthesis, and inhibition of the maturation or processing of ribosomal RNA.95 More recently, attention has focused on the inhibition of processing nuclear RNA to smaller-molecular-weight species.96 Other post-transcriptional effects of 5-FU include inhibiting polyadenylation of messenger RNA (mRNA) and effects on DNA primase. In some model tumors and tumor lines, there is persuasive evidence that these RNA-directed events can be associated with cytotoxicity, particularly when the effects of extended exposure are monitored.97,98

Thymidylate Synthase

The target site that can be defined most clearly is the covalent inactivation of thymidylate synthase by FdUMP.99 This fluorinated deoxyuridylate analog is formed via the reduction of FUDP by ribonucleotide reductase and dephosphorylation.100 Alternatively, it can be formed directly from 5-FdUrd by thymidine kinase101 when this 5-FU deoxynucleoside is regionally infused. The earliest studies by Umeda and Heidelberger102 indicated that in selected cell lines growth inhibition could be prevented by thymidine but not by uridine. Direct inhibition of the enzyme responsible for the 1-carbon transfer confirmed this site of action,103,104 and subsequent research identified specific steps in the reaction in which a methylene group from 5, 10-methylene tetrahydrofolate is transferred to the 5-position of 2′-deoxyuridylate.14 These studies elegantly established the formation of a stable ternary covalent complex among the 5′-fluoro analog of deoxyuridylate, the reduced folate derivative, and thymidylate synthase.105 The obvious consequence of this inhibition is an induced enzyme deficiency, depletion of dTTP, and the accumulation of deoxyuridine monophosphate (dUMP) behind the blockade.99,106,107 More recently, it has been shown that in some tumors or normal tissues the rate-limiting factor in the formation of the abortive ternary complex with FdUMP is availability of the reduced folate derivative.108,109 When this cofactor is limiting, it is possible to enhance inhibition by the administration of leucovorin.110 The consequence of dTTP depletion is generally considered to be unbalanced growth consequent to reduced DNA synthesis. As might be anticipated, this mode of inhibition would be nullified if thymidine were supplied because after phosphorylation by thymidine kinase, it would circumvent the site of inhibition. However, thymidine administration in vivo can actually increase the cytotoxic effects of 5-FU in vivo by inhibiting 5-FU catabolism.111

Recent studies have analyzed a possible correlation between TS expression and therapeutic outcome in colorectal, head and neck, and breast tumors. A retrospective study in patients with rectal carcinoma revealed that the expression of high levels of TS was linked to a significantly reduced disease-free survival, and that the staining intensity of TS monoclonal antibody was stronger in higher-grade, less differentiated tumors.112,113 Another study indicated that patients with advanced gastric and colorectal tumors expressing high TS protein did not respond to 5-FU-leucovorin treatment.114 Association between TS level in tumors and response to chemotherapy in colorectal cancer was also seen when a biochemical assay and a reverse polymerase chain reaction (PCR) method were used.115–117 Such a correlation was not observed in an immunohistochemical study from Findlay and colleagues on primary colorectal tumors.118 A report on head and neck squamous cell carcinomas also failed to establish a relationship between TS level and patient survival or treatment outcome.119 A retrospective immunohistochemical study in breast tumor of patients with early-stage breast cancer indicated that high TS expression was associated with a significantly worse prognosis in node-positive but not in node-negative breast cancer patients.120

A more likely explanation for the discrepancies in these studies is that factors other than TS alone may be involved. Thus more recent clinical studies have shown that tumor response and survival are likely determined by multiple factors including TS; DPD, which controls the amount of 5-FU available for anabolism; and various enzymes in the pyrimidine anabolic pathway which control the interconversions of 5-FU anabolites.121


Initially, the incorporation of 5-FU into DNA was not detected, and it was assumed to be prevented by the active dUTP phosphatases that also dephosphorylate FdUTP as it is formed.39 Subsequently, small quantities of 5-FU were detected in internucleotide linkages within DNA.122,123 Like dUTP, FdUTP, when it is available, is fully active as a substrate for the several DNA polymerases, but a very active glycosylase is present in most cells and excises any 5-FU or uracil that is incorporated in the place of thymine.34,39 Mutants have been found that are relatively deficient in this editing function, and it may be that incorporation per se is not the cytotoxic event, but that the excision and repair involving a pyrimidine endonuclease generates opportunities for error-prone repair that might again re-incorporate 5-FU or uracil instead of thymine nucleotides.38,124,125 Because a considerable accumulation of dUMP occurs behind the blockade of thymidylate synthase, higher concentrations of dUTP are generated. These concentrations and any FdUTP increase the need for an editing function to remove incorporated uracil. Examination of the kinetics of this excision reaction indicates that uracil is removed as much as 30 times more rapidly than 5-FU.

A similar elevation of dUTP concentrations can be achieved by methotrexate therapy via secondary inhibition of thymidylate synthase.126 Under these conditions, uracil incorporation into DNA is also increased, and the potential for error-prone repair is enhanced.

It is not possible to rank the importance of these different potential mechanisms of cytotoxicity: RNA incorporation, dTTP depletion by thymidylate synthase inhibition, DNA incorporation, or damage to DNA consequent to excision of uracil or 5-FU. In fact, the relative importance of each of these sites may vary in different cell types. In some tumor lines, evidence for high sensitivity to RNA-directed effects is seen by the inability of thymidine to overcome growth inhibition, despite the presence of an active thymidine kinase.127 In these same lines, uridine rescue is more successful than in others where thymidine effectively prevents cytotoxicity, presumably by repleting dTTP.


As with most drugs, partial or complete responses of human cancer to 5-FU generally are followed by the eventual regrowth of tumor despite sustained, or even increased, dosages. Understanding some of the factors that contribute to natural or acquired resistance has stimulated several approaches to modulating 5-FU therapy. The most prominent mechanism seen in experimental tumors is reduced anabolism of the analog to nucleotide form.128,129 This may reflect altered condensation with PRPP or activation via the two-stage salvage pathway involving ribose-1-phosphate or deoxyribose-1-phosphate and the appropriate nucleoside phosphorylase, with subsequent phosphorylation of the resultant nucleoside by uridine or thymidine kinase. Alternatively, lack of sensitivity has been correlated with an increased disappearance rate of 5-FU nucleotides, which were documented in one case to reflect enhanced nucleotide phosphatase activity.130 Alterations in the catabolism of 5-FU appear to affect sensitivity and predict responsiveness to the drug. DPD, the rate-limiting enzyme in the catabolism of pyrimidines, regulates the amount of 5-FU available for the activation to nucleotide forms. In hepatocellular carcinomas inherently resistant to fluoropyrimidine-based chemotherapy, DPD activity was found elevated compared to that of normal tissue.131 DPD activity was also found to predict response to 5-FU in head and neck tumors, and DPD mRNA levels predicted resistance to the drug in colorectal cancer patients.132,133 Other well-documented mechanisms of resistance reflect changes in the thymidylate synthase, with reduced affinity for FdUMP,134 or increases in the rate of synthesis and activity of the enzyme, possibly associated with gene amplification or altered enzyme turnover rates.135 The mode of exposure to the drug can result in the selection of tumor cells with different mechanisms of resistance.136–138 Finally, effective deletion of the facilitated diffusion transport of FdUrd has been shown to confer resistance to this 5-FU derivative, but not to 5-FU in a human colon cancer cell line.27

Modulation of Therapy: Leucovorin

To improve the limited response rate to therapy with 5-FU, a rate of 10% to 25% in the most responsive cancers, various biochemical strategies have been investigated.139 The degree of 5-FU activation by orotidylate pyrophosphorylase is affected by the available concentrations of PRPP. Because alterations of traffic along both the purine and pyrimidine nucleotide biosynthesis pathways affect the available concentrations of PRPP, several drug or metabolite combinations have been shown to modify the activation of 5-FU, presumably by altering the concentration of this ribose-5′-phosphate donor.140–142 Others have explored depletion of pyrimidine nucleotides by inhibitors of the de novo synthesis of pyrimidines. A major focus in this area has been enhancing the efficiency with which the covalent complex of FdUMP with the folate cofactor and thymidylate synthase is formed by supplementation with the reduced folate cofactor.143

Formation of the ternary complex of FdUMP, thymidylate synthase, and folate coenzymes may be limited by the availability of reduced folates in some cell lines and tumors.105,144 To optimize formation of the covalent complex, large doses of leucovorin, or D,L-N-5-formyl tetrahydrofolate, have been employed to saturate target enzymes with l-5-10-methylene tetrahydrofolate via conversion of the l-isomer of leucovorin to 5-methyl tetrahydrofolate.145

Sound experimental evidence supports the logic of this approach to modulation. Early studies have demonstrated that optimal 5-FU cytotoxicity in cell lines was achieved only when the cells were supplemented with folates to achieve concentrations much greater than those required for optimal growth.97,143 These effects directly related to the quantity of the ternary complex formed within the cells. The importance of sustaining the folate levels to stabilize the ternary complex could be seen in xenografts of human tumors, in which only transient inhibition of thymidylate synthase with 5-FU would be expected unless supplemental reduced folates were present.143,146 The importance of polyglutamylation to enhance binding to thymidylate synthase in retaining folates within cells also has been documented, using cells that were defective in polyglutamate synthase.147

If modulation by leucovorin in human disease is to be successful, the enhancement of ternary complex formation must be selective for tumor tissue. In a murine tumor model, leucovorin expanded the reduced folate pools in the tumor but not in bone marrow.148 This result was consistent with the antitumor effect seen without increased host toxicity. In other model systems, however, a consistent improvement in the therapeutic index is not seen. Because of the enhanced inhibition of thymidylate synthase when prior supplementation with leucovorin is employed, the dose of 5-FU must be reduced by approximately 20%.46 Under these conditions, diarrhea and mucositis remain the dose-limiting toxicities.

A wide range of clinical studies have generally confirmed the increased rate of response to 5-FU therapy in colorectal cancer when supplemented by leucovorin.110,149,150 Evidence for increased survival in these trials is limited, however.46,149 In breast and stomach cancers, the response rate in patients who are not previously treated with 5-FU appears to be increased by the addition of leucovorin; data for other diseases are insufficient to draw conclusions. The generally favorable results obtained in these studies have led to a rather universal addition of leucovorin to 5-FU trials of combination with other drugs. Particularly promising are three studies combining 5-FU-leucovorin with cisplatin in head and neck cancer.110 Despite these positive results, carefully controlled studies are needed to ensure the validity of this mode of modulation, particularly as other new drugs and modulators are combined with 5-FU-leucovorin regimens.

Oral Prodrug of 5-Fluorouracil: Capecitabine

A recent metaanalysis of infusional versus bolus 5-FU has concluded that protracted low-dose infusion of 5-FU has resulted in a higher response rate, 22% versus 14%, with improvement in survival.48 However, the long-term delivery requires a surgically implanted venous access and the use of an infusion pump. The administration of oral 5-FU could reduce the cost of treatment and be more convenient to the patient. Its oral use has been hampered by an incomplete and variable bioavailability. Over the past several years several oral fluoropyrimidines have been evaluated clinically. Although several had potentially desirable pharmacologic attributes, only capecitabine has received Food and Drug Administration (FDA) approval in the United States.

Capecitabine is an orally administered fluoropyrimidine carbamate prodrug (Figure 50-4) that is activated to 5-FU by three sequential enzymatic steps. First, hepatic carboxyesterase hydrolyses the N-pentyl carbamate chain to form 5′-deoxy-5-fluorocytidine, which is then deaminated to 5′-deoxy-5-fluorouridine 5′-d5-FUR by cytidine deaminase; then thymidine and uridine phosphorylases hydrolyze 5′-d5-FUR to produce 5-FU. The higher phosphorolytic activity expressed in human tumor tissue compared to that of the surrounding normal tissue has been suggested to result in selective activation and an improved therapeutic index. A higher concentration of 5-FU (2.9-fold) has been demonstrated in colorectal tumor specimens when compared to adjacent normal tissue of patients who received oral capecitabine 5 to 7 days prior to surgical removal of the tumor.151 Capecitabine is typically administered bid at a total daily dose of 2,000–2,500 mg/m2/d over 14 days. This dose generates plasma peak levels 2 hours after administration, comparable to the ones achieved with a continuous intravenous infusion of 300 mg/m2/d of 5-FU. Toxicities were also similar to a continuous 5-FU infusion, with diarrhea, mucositis, and hand-foot syndrome.152 The toxicities are for the most part tolerable, in particular at the 2,000 mg/m2/d dose, and reversible after a short interval off therapy.

Figure 50-4. Structure and metabolic activation of capecitabine.

Figure 50-4

Structure and metabolic activation of capecitabine. Adapted from Xeloda product information, Roche USA.

Capecitabine was initially approved for the treatment of metastatic breast cancer resistant to chemotherapy containing both paclitaxel and antracyclines. In this patient population, an 18.5% response rate was observed.153

Subsequently, capecitabine was approved for use in advanced colorectal cancer based on demonstrated equivalence in Phase III studies to intravenously administered 5-FU-leucovorin (Mayo regimen).154,155 It is of interest that coadministration of leucovorin was not required to obtain a comparable effect.

Most recently capecitabine has been approved for combined use with Taxotere in advanced breast cancer.156 The desirable effect of this combination may be due to upregulation of thymidine phosphorylase by Taxotere. Several other chemotherapy agents in addition to radiation are now recognized to upregulate thymidine phosphorylase, thereby increasing selective intratumoral activation of capecitabine.157

Cytosine Arabinoside


Cytosine arabinoside, or cytarabine or ara-C, is a nucleoside analog of deoxycytidine that was first synthesized in 1950 and introduced into clinical medicine in 1963.158 One of the most important drugs in the treatment of acute myeloid leukemia, it is also active against acute lymphocytic leukemia and, to a lesser extent, is useful in chronic myelocytic leukemia and non-Hodgkin lymphoma.159 It has not proven to be particularly useful in the treatment of nonhematologic neoplasms. Myelosuppression and gastrointestinal epithelial injury are the primary toxic effects of ara-C. Using high-dose ara-C regimens, additional toxic effects such as intrahepatic cholestasis and central nervous system (CNS) toxicity are frequently observed.160 This toxicity could be due to its impact on mitochondrial DNA synthesis in nonproliferating tissues.


Cytosine arabinoside is rapidly deaminated by cytidine deaminase to a much less active compound, arabinosyluracil (ara-U).161–163 Ara-C enters cells through a carrier-mediated process or by simple diffusion.26,164 At low concentrations of ara-C (< 2 μM), the carrier-mediated process predominates. The efficiency of this transport process depends on the binding affinity of ara-C for the carrier, the number of carrier molecules in the membrane, and the presence of competing nucleosides sharing the same system. After entering the cells, it is metabolized primarily by the enzymes that normally metabolize deoxycytidine or, in some instances, cytidine (Figure 50-5).

Figure 50-5. Structure and metabolism of arabinosylcytosine (ara-C).

Figure 50-5

Structure and metabolism of arabinosylcytosine (ara-C). MP, DP, TP = mono-, di-, and triphosphate; araCDP= cytosine arabinoside diphosphate; araCMP = cytosine arabinoside monophosphate; araCTP = cytosine arabinoside triphosphate araU = arabinoxyluracil; (more...)

The enzyme that is responsible for cytosine arabinoside monophosphate (ara-CMP) synthesis is cytoplasmic deoxycytidine kinase. Mitochondrial deoxypyrimidine nucleoside kinase, which can phosphorylate deoxycytidine and thymidine, does not efficiently phosphorylate ara-C.165 The activity of the cytoplasmic deoxycytidine kinase is higher in the S phase of the cell cycle. The amount of ara-CMP formed depends on the relative activity of cytoplasmic deoxycytidine kinase and cytidine deaminase. Tetrahydrouridine is a potent inhibitor of cytidine deaminase, with a Ki value of 10-8 M.161,166 Potentiation of the cytotoxic effect of low ara-C concentrations by tetrahydrouridine underscores the role of cytidine deaminase in ara-C metabolism. The enzyme responsible for conversion of ara-CMP to cytosine arabinoside diphosphate (ara-CDP) is cytidylate-uridylate-deoxycytidylate (CMP-UMP-dCMP) kinase. There are two forms of this enzyme, and both are capable of phosphorylating ara-CMP. It has been suggested that ara-CMP could be deaminated to uracil arabinoside monophosphate (ara-UMP) by dCMP deaminase.167 Whether this pathway is functional in cells is questionable, however, because ara-CMP is a very poor substrate for dCMP deaminase compared to dCMP. Several mammalian cell lines are partially resistant to ara-C because of a decreased activity of dCMP deaminase.168,169 Enzymes responsible for the phosphorylation of ara-CDP to cytosine arabinoside triphosphate (ara-CTP) are nucleoside diphosphate (NDP) kinases. There are multiple species of NDP kinase activities in human cells; many of them belong to the nm23 gene family.170 Whether a preference exists for one isozyme over another in the phosphorylation of ara-CDP is unclear, but the formation of ara-CDP choline in human cells incubated with ara-C has been reported.171,172 The enzyme that catalyzes this reversible process is phosphorylcholine cytidyltransferase. Both cytidine diphosphate (CDP) choline and deoxy-cytidine diphosphate (dCDP) choline serve as donors of the phosphorylcholine moiety in phosphatidylcholine synthesis; how ara-CDP choline participates in or interferes with this reaction is not clear.

Major attention has also been focused on the incorporation of ara-CTP into DNA in competition with deoxy-cytidine triphosphate (dCTP).173–175 Elongation of DNA by polymerase α is considerably retarded by the incorporation of ara-CMP, whereas no significant impact on elongation by DNA polymerase β could be seen after incorporation of a single ara-C nucleoside residue. However, neither polymerase alone could appreciably elongate the DNA if two consecutive ara-CMP residues were incorporated. Thus, the behavior of ara-CTP on DNA polymerase is not only polymerase-dependent but also sequence-dependent.176,177

Mechanism of Action

The primary action of ara-C is inhibition of nuclear DNA synthesis.178,179 Mitochondrial DNA synthesis is not affected by ara-C, even at concentrations 10 times greater than that required to inhibit cell growth by 50%. The possibility remains, however, that the functional nature of mitochondrial DNA may be compromised through incorporation of ara-C internally.162

Three mechanisms have been suggested to account for the inhibition of nuclear DNA synthesis by ara-C. The relative importance of each mechanism may depend on the intracellular concentration of ara-CTP. The first mechanism is inhibition of the initiation of new replication units in chromosomes consequent to the incorporation of ara-C into the replicon-initiation primer.180 The second mechanism is the retardation of DNA-chain elongation because of the incorporation of ara-C into DNA.173,174 This effect is DNA polymerase- and sequence-dependent, as discussed earlier. Reactions catalyzed by DNA polymerase α, and perhaps DNA polymerase Δ, are more susceptible than other DNA polymerase activities. The third mechanism, which may become important only when a high-dose ara-C protocol is used, is the inhibition of DNA primase.181 Ara-CTP can inhibit the formation of the RNA oligomer required for the initiation of DNA synthesis with Ki values of 25 to 125 μM (depending on the template being used). Although there is no evidence that ara-CMP can be incorporated into an RNA oligomer in vitro, it has been found that some of the ara-C that is associated with DNA is alkaline labile.182 This indicates the possibility that ara-C is incorporated into the RNA primer of DNA and requires further investigation.

In general, the inhibition of cell growth correlates well with the degree of the incorporation of ara-C into cellular DNA. The majority of incorporated ara-CMP is in internucleotide linkage in DNA. The relative ratio of ara-C in internucleotide, compared with chain-terminal positions, depends on the concentration of ara-C; the higher the concentration of ara-C to which the cells are exposed, the lower the relative amount of internucleotide ara-C residues. This could result from the higher probability of consecutive ara-CMPs being incorporated into DNA, which stops further DNA-chain elongation catalyzed by DNA polymerase α and DNA polymerase β. The amount of ara-CMP that is incorporated into DNA also depends on the relative ratio of ara-CTP to dCTP. Decreases in the intracellular pool of dCTP can increase the amount of ara-CMP that is incorporated. Exonucleases such as the recently identified TREX 1 or 2 and even p53 could remove ara-C incorporated in terminal positions to limit the cytotoxic effects.

Among other potential targets, ara-CTP is not a potent inhibitor of ribonucleotide reductase, a key enzyme early in the course of dCTP formation.183 Ara-CTP can act in lieu of dCTP to activate dCMP deaminase for the deamination of dCMP to dUMP, the substrate for dTMP synthesis. Because ara-CMP is a poor substrate for dCMP deaminase, the accumulation of ara-CTP enhances the deamination of dCMP and subsequently decreases the intracellular pool of dCTP.167 This could “self-potentiate” the incorporation of ara-CTP into DNA. This hypothesis is based on enzyme studies in vitro, but it is substantiated by the observation that cells become resistant to ara-C because of decreased dCMP deaminase activity.168

The mechanism of action for ara-C may be dosage-dependent. At noncytotoxic concentrations, ara-C can cause human promyeloblast HL-60 cell lines to differentiate. It has been suggested that the success of low-dose ara-C therapy in patients with myelodysplastic syndrome may result from the differentiation effects of ara-C.178 When given to patients with leukemia, high doses of ara-C cause rapid tumor-cell lysis.171 Whether additional mechanisms of ara-C also play important roles in this protocol is unclear. In patients who receive high doses, the concentration of ara-U, the deamination product of ara-C, can exceed 100 μM in plasma.184 The high concentrations of ara-U may act in concert with ara-C, and it also may affect cell growth by mechanisms that have not yet been established.185

Mechanism of Resistance

Cells could become resistant to ara-C because of (1) decreased activities of the carrier for ara-C transport and for cytoplasmic deoxycytidine kinase, (2) increased catabolism of ara-C through the action of cytidine deaminase, (3) increased formation of dCTP by ribonucleotide reductase and NDP kinase, or (4) decreased activity of dCMP deaminase, which could lead to increased competition between dCTP and ara-CTP for incorporation into DNA. An increased activity of 3′ to 5′ exonuclease, which could remove ara-CMP from the DNA-chain terminus, has also been suggested.186



5-Azacytidine (5-AC) was first synthesized in 1963, and it was later isolated as a natural product from fungal cultures.187,188 The clinical utility of this cytidine analog is primarily in the treatment of acute myelocytic leukemia and myelodysplastic syndrome where in low dose, it is able to cause partial or complete differentiation in hematopoiesis of the majority of patients; occasionally, clinical response has been observed in patients with solid tumors. This compound can promote the expression of genes that are suppressed by hypermethylation.189 This activity suggested use of 5-AC in genetic diseases, such as sickle cell anemia and thalassemia, but its usefulness in treating these diseases has been limited by its bone marrow toxicity and concerns over its carcinogenic potential. The major toxicity of 5-AC is leukopenia and, to a lesser degree, thrombocytopenia. Hepatotoxicity has also been reported, particularly in patients with pre-existing hepatic dysfunction.159


The replacement of carbon by nitrogen in position 5 of the cytidine heterocyclic ring results in a marked chemical instability. The product of the ring opening, N-formylamidinoribofuranosyl guanylurea, may recycle to form the parent compound, but it is also susceptible to further decomposition. This tendency to decompose not only may play a role in its mechanism of action but also is troublesome in its clinical use.190 Although 5-AC can be deaminated by cytidine deaminase to 5-azauridine (5-AU), a less toxic compound, the efficiency of this deamination by cytidine deaminase is less than that of cytidine. Nevertheless, inhibition of the deamination by tetrahydrouridine can enhance 5-AC toxicity. 5-AC enters mammalian cells by a facilitated nucleoside transport mechanism that is shared with other nucleosides.191 The initial step in its activation is the conversion to 5-azacytidine monophosphate (5-ACMP) by uridine-cytidine kinase.192 5-ACMP is further phosphorylated to 5-AC di- and triphosphate by CMP-UMP-dCMP kinases and nucleoside diphosphate kinases, respectively. 5-AC triphosphate, which for several hours is the predominant metabolite in cells treated with 5-AC, can be incorporated into RNA, but its pathway for incorporation into DNA is not well defined. 5-azacytidine diphosphate (5-ACDP) likely is reduced by ribonucleotide reductase to the corresponding deoxynucleotide diphosphate. This diphosphate is phosphorylated by nucleoside diphosphate kinases to 5-azadeoxycytidine diphosphate (5-AdCTP), which can be efficiently incorporated into DNA by DNA polymerases α and β. The incorporated 5-azadeoxycytidine monophosphate (5-AdCMP) at the 3′ terminus of DNA has less effect on subsequent DNA-chain elongation than the incorporated ara-CMP at the 3′ terminus of DNA. 5-azadeoxycytidine (5-AdC) also is stabilized against hydrolytic degradation by incorporation into DNA, which could result, in part, from hydrophobic shielding of the triazine ring from water and other polar nucleophiles within the DNA double helix.177,193

A summary of 5-AC metabolism is shown in Figure 50-6. 5-AC is most cytotoxic to cells in the DNA-synthetic phase of the cell cycle, but the exact mechanism of its cytotoxic action has not been well established. It could inhibit both DNA and RNA synthesis. Incorporation into RNA can inhibit the processing of ribosomal RNA from higher-molecular-weight species, disassembly of polyribosomes, and markedly inhibit protein synthesis. Incorporation into DNA also could inhibit DNA synthesis.194–197 One important, welldocumented effect is the inhibition of DNA methylation because of stoichiometric binding with DNA-methyltransferase after incorporation. The methylation of cytosine residues in DNA is responsible for the inactivation of specific genes; thus, treatment of cells with 5-AC leads to reduced levels of cytosine methylation and enhanced expression of selected genes that are normally suppressed. At minimally cytotoxic concentrations, 5-AC stimulates the differentiation of some tumor cell lines in culture, and it has been suggested for the treatment of genetic diseases that are associated with hypermethylation189 in myelodysplastic syndrome,188a as well as for virus-associated cancers in combination with antiviral compounds.

Figure 50-6. Structure and metabolism of 5-azacytidine (5-AC).

Figure 50-6

Structure and metabolism of 5-azacytidine (5-AC). 5-AdC = 5-azadeoxycytidine; 5-AU = 5-azauridine; MP, DP, TP = mono-, di-, and triphosphate.

Mechanism of Resistance

Cells can become resistant to 5-AC by the reduction or elimination of uridine-cytidine kinase. Decreased nucleoside transport by the facilitated diffusion mechanism also can decrease sensitivity to 5-AC, and cytosine deaminase may play an important role in cell sensitivity as well. In animal models, tumor cells that are resistant to ara-C because of the deletion of cytoplasmic deoxycytidine kinase activity— a frequent mechanism of cellular resistance to ara-C— are more susceptible to 5-AC than is the parent tumor line. Sequential treatment with ara-C and then 5-AC deserves further study, particularly in patients who become refractory to ara-C.



2′, 2′-difluoro-2′-deoxycytidine (dFdC) is a deoxycytidine analog with two fluorine atoms in the 2′ position of the sugar moiety (Figure 50-7).198 First synthesized in 1986, this molecule was initially developed as an antiviral agent because of its potent inhibitory activity against both DNA and RNA viruses.199 Subsequently, its broad spectrum of activity in murine tumors and human tumor xenografts200,201 led to evaluating this antineoplastic activity in clinical trials.

Figure 50-7. Structure, metabolism, and actions of 2′, 2′-difluoro-2′-deoxycytidine (dFdC) and its nucleotides.

Figure 50-7

Structure, metabolism, and actions of 2′, 2′-difluoro-2′-deoxycytidine (dFdC) and its nucleotides. Dashed lines indicate inhibitory actions. Modified from Heinemann et al. dU = deoxyuridine; MP, DP, TP = mono-, di-, and triphosphate; (more...)

dFdC was approved by the FDA in 1996 as a first-line treatment for patients with locally advanced or metastatic adenocarcinoma of the pancreas.202,203 For the first time, approval was granted on the basis of clinical benefit response as the main clinical end point for assessing the drug's effect. Subsequently, dFdC has received FDA approval for the treatment of inoperable, locally advanced or metastatic non-small cell lung cancer, in combination with cisplatin.204–210

dFdC has been found active as a single agent in the first-line treatment for breast cancer and has been combined with cisplatin in patients pretreated with anthracycline and taxane. 211,212 Activity has also been shown in bladder cancer213–219 and in ovarian cancer.220–224 dFdC has demonstrated a promising effect in non-Hodgkin lymphoma,225,226 in Hodgkin disease,227 and in patients with relapsed or refractory cutaneous T-cell lymphoma.228

The dose-limiting toxicity of dFdC in both single-agent and combination studies has been mild to moderate myelosuppression. The nonhematologic toxicity was mild, with nausea, vomiting, occasional skin rash, alopecia, and pneumonitis. Rare occurrences of hemolytic-uremic syndrome have been reported.229–231


2′, 2′-difluoro-2′-deoxycytidine requires phosphorylation by deoxycytidine kinase to exert its cytotoxic activity (see Figure 50-7). The major intracellular metabolite is 2′, 2′-difluoro-2′-deoxycytidine triphosphate (dFdCTP); lesser amounts of the monophosphate (dFdCMP) and the diphosphate (dFdCDP) are also present.232 The cellular elimination of dFdCTP was investigated in several human cell lines: CCRF-CEM, K562, and A2780.233,234 Elimination of dFdCTP follows a biphasic course, with a short initial half-life followed by a second, slower phase of degradation. The biphasic elimination of dFdCTP differs from the linear monophasic kinetic that is exhibited by the triphosphate of ara-C,235 arabinosyladenine,236 and arabinosyl-2-fluoroadenine.237

Deoxycytidine deaminase inactivates dFdC to 2′, 2′-difluoro-2′-deoxyuridine (dFdU), which has no antitumor activity.232 The monophosphate of dFdC also can be deaminated to the uracil derivative 2′difluoro-deoxyridine monophosphate (dFdUMP) by deoxycytidylate deaminase.238

Pharmacokinetic studies during phase I clinical trials have shown a very rapid half-life (8 minutes) for dFdC because of deamination over a wide range of dosages.239 The deamination product, dFdU, which is the only metabolite present in the urine, exhibits a biphasic elimination from plasma, with a long terminal phase of 14 hours. The concentration of dFdCTP in mononuclear cells increases in proportion to the dose of dFdC infused, up to 250 mg/m2. Above this dose, the process shows saturation in accumulation of the triphosphate derivative.

Mechanism of Action

2′, 2′-difluoro-2′-deoxycytidine exerts its inhibitory activity on DNA synthesis through several distinct mechanisms. The accumulation of dFdCTP causes a reduction in the deoxyribonucleotide pools in both CCRF-CEM and HT-29 human tumor cells.240,241 This reflects a direct inhibition of ribonucleotide reductase, caused mainly by dFdCDP; however, dFdCTP was not as inhibitory of the partially purified enzyme.240 Another important mechanism is the incorporation of dFdCTP into DNA; dFdCTP competes with dCTP for incorporation into the C sites of DNA as catalyzed by DNA polymerases α and ε. The primer extension pauses one deoxynucleotide after dFdCMP incorporation.242 Moreover, the exonuclease activity of polymerase ε was unable to excise nucleotides from DNA containing dFdCMP at either the 3′ end or at an internal position.242 The cytotoxic activity of dFdC strongly correlates with the amount of monophosphate that is incorporated into cellular DNA.

Incorporation of dFdC into RNA has been detected in murine colon 26-10 cells as well as human A2780 and CCRF-CEM cells.243 Although the extent of this incorporation was two- to tenfold less than that into DNA, it may play a role in cytotoxicity.

Inhibition of ribonucleotide reductase could have a self-potentiation effect on the inhibitory activity of this drug. The activity of deoxycytidine kinase, which is required for the phosphorylation of dFdC, is regulated by dCTP levels; therefore, a decrease in dCTP pools likely will lead to increased dFdC activation.232 dCTP also is required as an activator of dCMP deaminase, an enzyme that is critical for the catabolism of dFdC nucleotides; thus, a reduction in dCTP could slow the deamination process and prolong the half-life of dFdC nucleotides.238 Finally, dCTP competes with dFdCTP for incorporation into DNA by polymerases α and ε, and lower dCTP levels also could enhance dFdC incorporation into DNA as well as increase its inhibitory effect on cell proliferation.242

To date, two examples of resistance to dFdC have been reported.244,245 Human ovarian carcinoma A2780 cells that were exposed to increasing concentrations of dFdC became highly resistant to the drug, cross-resistant to ara-C and 2-chlorodeoxyadenosine, and modestly resistant to doxorubicin, vincristine, and cis-platinum. Resistant cells did not possess deoxycytidine kinase activity; therefore, they were not able to phosphorylate dFdC as well as the other two nucleoside analogs. Western blot analyses of the cell extract using a polyclonal, anti-deoxycytidine kinase antibody could not detect this protein in the resistant subline. Another mechanism of resistance has been recently reported. Human KB tumor cells could become resistant to dFdC as the result of increased expression of the M2 unit of ribonucleotide reductase. Resistance leads to elevated activity of the same enzyme, as well as an augmented intracellular dCTP pool, which could prevent the phosphorylation of dFdC by deoxycytidine kinase.245

By agreement with the publisher, this book is accessible by the search feature, but cannot be browsed.

Copyright © 2003, BC Decker Inc.
Bookshelf ID: NBK13287


  • Cite this Page

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...