Rational Design of an Orally Active Anticancer Fluoropyrimidine, Pencitabine, a Hybrid of Capecitabine and Gemcitabine

Abstract

The structure of the anticancer drug capecitabine was redesigned to prevent metabolic conversion to 5-fluorouracil and its associated potentially fatal toxicities. The resulting cytidine analogue, pencitabine, is a hybrid of capecitabine and gemcitabine, another anticancer drug in clinical use. Preliminary biological evaluation revealed that pencitabine is cytotoxic in vitro in cell culture and orally active in vivo in a human xenograft test system. Pencitabine may mimic the known therapeutically advantageous combination of its parent drugs. Pencitabine is postulated to interfere with DNA synthesis and function by inhibiting multiple nucleotide-metabolizing enzymes and by misincorporation into DNA. Based on detailed mechanistic analyses and literature precedents, the hypothesis is put forward that the significant DNA damage caused by pencitabine may be accounted for by two additional effects not shown by the parent drugs: inhibition of DNA glycosylases involved in base excision repair and of DNA (cytosine-5)-methyltransferase involved in epigenetic regulation of cellular metabolism.

Fluoropyrimidines,¹ an important class of antimetabolites, have been a mainstay in the chemotherapy of cancer, with primary indications in the treatment of colorectal and breast cancer. The prototype of the class is 5-fluorouracil (FU) introduced in 1957 by Heidelberger et al.(2) as an antagonist of uracil (U) that was shown to be utilized more effectively by cancer than normal tissues.

The pharmacological activity of FU requires intracellular metabolism that generates two nucleotide analogues, 5-fluoro-2′-deoxyuridylate (FdUMP) and 5-fluorouridine triphosphate (FUTP), which are the active forms of FU responsible for most of its biological activities. FdUMP inhibits thymidylate synthase (TS, EC 2.1.1.45), preventing formation of dTMP, one of the building blocks of DNA, that leads to inhibition of DNA synthesis and misincorporation of uracil and FU into DNA.² TS inhibition is usually considered mainly responsible for the anticancer activity of all fluoropyrimidines. FUTP, the other main intracellular metabolite of FU, serves as a substrate for RNA polymerase (EC 2.7.7.6), causing misincorporation of FU into RNA in place of uracil.² The disturbances in RNA metabolism and function that result from FU misincorporation into RNA in place of uracil may contribute significantly to most of the toxic side effects of the fluoropyrimidines. Indeed, uridine triacetate, approved by the U.S. Food and Drug Administration (FDA) in 2015, can block FU incorporation into RNA, and it has been used as a life-saving measure in early onset severe FU toxicity, provided it is administered within the first 96 h of therapy.^3,4 Fluoropyrimidine-specific serious toxicity is caused by the hereditary deficiency of the FU-catabolizing enzyme, dihydropyrimidine dehydrogenase,⁵ that also can lead to potentially fatal apparent overdose. Because of the multitude of toxic side effects, there is a very narrow window between the minimum effective dose and the maximum tolerated dose of FU that varies from patient to patient, as well as a limited overall response rate (20–40%). Catabolism of FU leads to toxic products, llike α-F-β-alanine and fluoroacetate, which have been implicated in neurotoxicity⁶ of FU and likely contribute to other toxic side effects, reflected by the low therapeutic index of FU.

The most advanced member of the fluoropyrimidine class of anticancer drugs is capecitabine^7,8 (Cape, Xeloda), one of the five anticancer cytidine analogues approved by the FDA (see Figure 1). Cape is a rationally designed prodrug of FU with good oral bioavailability,⁸ due to the carbamate side chain, eliminating the need for IV administration and associated liabilities. Cape has a wider therapeutic window than FU and has a degree of selectivity due to the overexpression of thymidine phosphorylase (TP, EC 2.4.2.4) in a variety of cancers. TP (and to a lesser extent uridine phosphorylase, EC 2.4.2.3) catalyzes the last, rate-determining step of the metabolic conversion of Cape to FU. Cape shares with FU a variety of serious, dose-limiting toxicities, some potentially fatal.⁴ Drug-induced fatalities have been estimated to reach about 1300 annually.^3,9 In addition to the predominant gastrointestinal toxicity, the less frequent cardiovascular toxicity of fluoropyrimidines is second only to the anthracyclines among cancer chemotherapeutic agents.¹⁰ Cape has a characteristic dose-limiting toxic side effect, known as hand-foot syndrome (HFS) or palmar plantar eryhtrodysesthesia,¹¹ with an incidence of >50%, far exceeding that of FU. After hematologic and gastrointestinal toxicity, HFS is the most frequent undesirable side effect of Cape. Ironically, HFS has been linked to the overexpression in the skin of the palm and feet of the very enzyme, TP, that is claimed to be responsible for the degree of selectivity observed for Cape.¹²

Figure 1.

Anticancer cytidine analogues approved by the FDA (approval dates indicated).

It has long been of interest to reduce the dose-limiting toxic side effects of fluoropyrimidines and increase their potency and oral bioavailability. Most previous attempts focused on prodrug approaches and biochemical or pharmacological modulation, but the anticancer activity of fluoropyrimidines remained dependent on the presence or metabolic formation of FU and consequently could not be dissociated from the FU-specific toxic side effects. In this communication, the rational design, synthesis, and biological activity of pencitabine (Pen), a novel, orally active derivative of Cape, which cannot be metabolized to FU and therefore should be devoid of the toxicities characteristically associated with FU, are described. The validity of these predictions awaits future comparative toxicological evaluation of Pen.

It was of interest to discover if the obligatory metabolism of Cape to FU by pyrimidine phosphorylases can be circumvented by structural modifications and thereby eliminate most of the systemic toxicities caused by formation of FU. At the same time, it was important to maintain the beneficial features of Cape, namely, the carbamate prodrug moiety responsible for oral bioavailability and reduced gastrointestinal side effects,^7,8 as well as the 5-fluoropyrimidine moiety responsible for the thymidylate synthase inhibitory activity, which is the cornerstone of the anticancer activity of the fluoropyrimidine class of antimetabolites.

Toward this goal, it was decided to design away substrate activity for TP, which is responsible for the rate limiting formation of FU from Cape. This could be accomplished by considering the mechanism of the enzyme catalyzed reaction that involves a positively charged oxocarbenium ion-like transition state.¹³ It has been established that electron-withdrawing substituents at the 2′-position, such as fluorine, can prevent glycosyl bond cleavage¹⁴ by destabilizing the transition state. It was considered that two fluorines would have an even stronger destabilizing effect (see Chart 1A).

Chart 1. (A) Destabilization of the Oxocarbenium Ionlike Transition State of Nucleoside Phosphorylases by 2′,2′-Difluoro Substitution; (B) Structural Modifications of Cape to Form Pen (the Blue Colored Segment Is Identical to the Sugar Moiety of the Anticancer Drug Gemcitabine).

However, by preventing formation of FU, activation to phosphorylated metabolites would also be prevented. Therefore, the 5′-CH₃ group of Cape had to be replaced by 5′-CH₂OH to allow kinase-mediated phosphorylation. These limited substitutions at the 2′- and 5′-positions proved to be sufficient to convert Cape into a new fluoropyrimidine, that would satisfy the requirements stated above (see Chart 1B), i.e., a loss of substrate activity to TP and a gain of substrate activity to pyrimidine nucleoside kinases. The differential substrate activities predicted may divert Cape and Pen to different metabolic activation pathways (see Figure 2).

Figure 2.

Outline of the predicted metabolic steps required for Cape and Pen to form the corresponding TS inhibitory 5′-monophosphates FdUMP and F₃dUMP, respectively.

Pen can be regarded as a hybrid of two anticancer nucleoside analogues: it is composed of the base of Cape linked to the sugar of gemcitabine (Gem), as shown in Chart 1B. It is a fluorinated derivative of both Cape (2′-deoxy-2′,2′-difluoro-5′-hydroxycapecitabine) as well as that of Gem (N⁴-pentiloxycarbonyl-5-fluorogemcitabine). Even though it was not part of the design strategy that led to Pen, molecular hybridization has been used by medicinal chemists as a useful tool in drug design.¹⁵ Preliminary studies discussed below are consistent with Pen retaining the inhibitory activities of both Cape and Gem against their most characteristic targets, TS and ribonucleotide reductase (RR, EC 1.17.4.1), respectively.

Due to its hybrid nature, Pen may be considered a “dual antagonist”, a concept introduced by Bardos¹⁶ to simplify the administration of combination chemotherapies. The main advantage of incorporating two drugs into one molecule is that in their hybrid form they have a single set of ADMET properties; they colocalize and can act simultaneously. Numerous clinical trials of the combination of oral capecitabine and IV gemcitabine for the treatment of pancreatic cancer have established their potential advantage over monotherapy.^17,18 Elimination of the requirement for IV administration of the “gemcitabine component” while maintaining its therapeutic contribution is another important advantage. It is conceivable that Pen may mimic the beneficial effects of the therapeutic combinations of Cape and Gem.^17,18

A comparison of the metabolic activation pathways available for Cape and Pen reveals that, after removal of their carbamate prodrug moiety by carboxylesterases CES1 and CES2 (EC 3.1.1.1.),¹⁹ Cape requires five more steps than Pen to form the respective TS inhibitory metabolites FdUMP and F₃dUMP (see Figure 2). This includes the step catalyzed by TP, which is not part of the pathway available for Pen. Indeed, F₃dUrd (see preparation in the Supporting Iinformation (SI)) could not serve as a substrate for human TP, while FdUrd and 5′-doxy-FdUrd were completely cleaved to FU (see Enzyme studies in the SI) as predicted by the design strategy.

Scheme 1 outlines the synthesis of Pen (7), which starts with persilylation of cytosine 1 to yield 2 using HMDS in the presence of (NH₄)₂SO₄. Condensation of 2 with mesyl dibenzoyl-α-ribofuranoside 3 was attempted under a variety of conditions (see the SI).

Scheme 1. Outline of the Synthesis of Pen (7).

Conditions favoring S_N2-substitution to give the β-nucleoside were unsatisfactory; therefore, Vorbrueggen’s S_N1-type glycosylation method using a Lewis acid promoter was applied. The best results were obtained using 3.0 equiv of trifluoromethanesulfonate (TMSOTf) in dichloroethane that yielded 4 as a 55:45 α/β mixture of anomers in 100% yield. The preparation of the l-enantiomer of 4 was reported to involve an equimolar amount of TMSOTf for the condensation reaction with a yield of 57%.²⁰ Separation of the anomers by silica gel column chromatography to obtain the pure β-anomer 6 was unsuccessful. After debenzoylation with methanolic ammonia, the anomeric mixture was converted to the HCl salt which crystallized out as an enriched mixture (∼80% beta), which was basified by adding NaOH until pH 8.5, at which point the pure beta free base 6 crystallized out in 30% yield over three steps. To a solution of 6 in a mixture of CH₃CN, pyridine, and HMDS, 1.1 equiv of amyl chloroformate was added to yield the desired product 7, contaminated with the 5′-carbonate 8, which was hydrolyzed back to 7 to give an overall yield of 66% 7. A doubling of the molar equivalent of amyl chloroformate may increase the yield by more complete N-acylation.

A preliminary evaluation of the biological activity of Pen was undertaken. It was of interest to evaluate the substrate activity for TP of F₃dUrd (see preparation in the SI), the metabolite of Pen designed to resist cleavage to FU (see Figure 2). Silica gel thin layer chromatography revealed that F₃dUrd did not show any conversion to FU. In contrast, 5′-dFUrd and FdUrd were completely cleaved to FU by TP. Spectrophotometric studies confirmed these results, verifying the prediction of the design strategy. Initial velocities obtained for 0.5 mM FdUrd, 5′-dFUrd, and F₃dUrd were 0.126 ± 0.003, 0.068 ± 0.004, and 0.000 ΔOD₂₉₀/min, respectively (see Enzyme studies in the SI). When 0.5 mM F₃dUrd was added to the cuvette with 5′-dFUrd, the rate decreased to 0.032 ± 0.001, indicating competition between the two, showing that F₃dUrd must be able to bind to TP, but cannot be cleaved.

In cell culture, Pen was found to be cytotoxic to both solid tumors and leukemias. After 72 h incubation, the growth of HCT-116 colorectal carcinoma and KG-1 acute myelogenous leukemia cells was inhibited by 50% at 0.37 ± 0.13 and 0.13 ± 0.011 μM, respectively. In comparison, the two components of the hybrid, Gem and Cape, under the same conditions showed differential activities. Cape was less active with IC₅₀ values of 34.2 ± 1.5 and 8.9 ± 1.5 μM, respectively, whereas Gem was more active with IC₅₀ values of 0.05 ± 0.0087 and 0.018 μM, respectively (see representative dose–response curves in Figure 3). It can be concluded that modification of the structure of Cape increased its potency more than 1 order of magnitude.

Figure 3.

Inhibition of the growth of HCT-116 human colorectal carcinoma cells (A) and KG-1 human myelogenous leukemia cells (B) by Gem, Pen, and Cape.

As expected, based on the transport properties of Gem, it was found dipyridamole, an inhibitor of nucleoside transport, was able to decrease the cytotoxicity of F₃dCyd (6), the free nucleoside form of Pen. In one experiment using BxPC3 pancreatic ductal adenocarcinoma cells, 10 mM dipyridamole caused a 150-fold increase in the IC₅₀ value of F₃dCyd (0.5 to 75 μM), suggesting the involvement of the equilibrative nucleoside transport (ENT1) system in its cellular uptake (G. J. Peters, personal communication).

Since one of the prime advantages of Cape as a prodrug of FU is its oral bioavailability, it was of interest to obtain evidence that, after structural modification, Pen retained the oral activity of Cape. To this end, a HCT-116 subcutaneous human colorectal carcinoma xenograft tumor model in athymic female nude mice was used with 100 μM capecitabine as the positive control. As shown by the results in Figure 4, Pen inhibited tumor growth by about 40% in vivo at a 25 mg/kg oral dosing during a 26 day period. It was found that Pen had about a 10-fold lower systemic toxicity to mice than its free nucleoside parent F₃dCyd, as measured by weight loss and time of death (see Table S3), likely due to the more favorable ADMET properties of the prodrug, justifying the decision to preserve the carbamate side chain of Cape. Future work should establish the full ADMET profile of Pen in a complete preclinical pharmacology/toxicology study to facilitate further development.

Figure 4.

Antitumor activity of Pen and Cape by oral dosing in the HCT-116 human xenograft model in mice (see the SI for treatment of animals). Error bars: ± SD (N = 3).

Based on the analogy of the well-established metabolic pathways of cytidine and its analogues, it can be predicted that, after carbamate hydrolysis and cellular uptake assisted by nucleoside transporters, the free nucleosides, F₃dCyd, and its deaminated product F₃dUrd, will undergo two parallel pathways of intracellular metabolism, leading to two sets of enzyme inhibitory mono-, di-, and triphosphate metabolites, and incorporation into DNA, as outlined in Figure 5 (red, cytosine-based nucleotides; blue, uracil-based nucleotides). In contrast, due to its obligatory deamination by cytidine deaminase (CDA) during its metabolic activation (see Figure 2), Cape cannot yield any cytosine-based nucleotides for enzyme inhibition or incorporation into DNA.

Figure 5.

Outline of the predicted metabolism of Pen.

Detailed studies of the mechanisms of action of Pen have not been carried out, but reasonable predictions of likely targets can be made based on the extensive work in the past on both Cape⁸ and Gem²¹ as well as other nucleosid e analogues.²²

It has been established that TS and RR are the most characteristic enzyme targets of Cape⁸ and Gem,²² respectively. It was of interest to gain some preliminary insight into the likelihood of both being targeted by F₃dCyd, the free nucleoside metabolite of Pen in the cell. To this end, deoxynucleoside triphosphate pools were examined as described²³ after exposure of HeLa cells to 50 μM F₃dCyd for 16 h in parallel with 50 μM Gem. Both compounds decreased the dATP pool size equally to almost undetectable levels (∼3%), a hallmark of RR inhibition even by non-nucleotide RR inhibitors, like hydroxyurea.²⁴ In contrast, the dTTP and dUTP pool sizes showed a differential response to the two compounds. For Gem, the results showed a ratio of dTTP/dUTP = 14.7, whereas for F₃dCyd results showed a ratio of dTTP/dUTP = 0.97, a 15-fold differential. These results are consistent with equally strong inhibition of RR by the two drugs but a stronger inhibition of TS by F₃dCyd than by Gem, which is known to generate F₂dUMP, a weak competitive inhibitor of TS.²⁵

It is conceivable that antagonism to the cytotoxicity of Pen, as a fluoropyrimidine, may result from its own RR-inhibitory activity. DNA fragmentation, a hallmark of thymine-less death,²⁶ contributes significantly to the cytotoxicity of fluoropyrimidines, caused by uracil incorporation into DNA in place of thymine, due to the expansion of intracellular dUTP pools during inhibition of TS. When RR is inhibited, the pools of deoxynucleotides, including dUTP, are depleted and uracil incorporation into DNA is reduced or prevented. Under normal conditions, when the nucleotide pools are not perturbed, it is the efficiency of dUTPase (EC 3.6.1.23) that safeguards the integrity of DNA by rapidly hydrolyzing dUTP to dUMP.

To find out if Pen is capable of damaging DNA despite its inhibition of RR, alkaline comet assays²⁷ of F₃dCyd, the intracellular free nucleoside form of Pen, were carried out using etoposide as positive control. The comet assay has been used to measure DNA damage and BER activities.²⁸

Table 1 shows the results expressed as % DNA in tail showing that F₃dCyd causes dose-dependent DNA damage at micromolar concentrations, less potent than etoposide, but in the same order of magnitude. The corresponding dose-dependent increase of comet numbers displayed as fluorescent images are shown in Figure S1. As expected, the effects of etoposide were more proportional to the concentration increase than those of F₃dCyd; a doubling of etoposide concentration increased tail DNA 1.6-fold, whereas a 10-fold increase in F₃dCyd concentration increased tail DNA only 1.3-fold. This is consistent with etoposide binding directly to the DNA-topoisomerase II complex while F₃dCyd must become phosphorylated and incorporated into DNA before it can exert its effect.

Table 1. Percent DNA in Tail (Comet Assay Data Points).

sample	Na	mean	mean SEb	80% CIc
control	246	9.15	0.973	7.90–10.40
1.0 μM F3dCyd	2718	20.70	0.359	20.2–21.17
3.0 μM F3dCyd	1956	24.28	0.442	23.7–24.84
10 μM F3dCyd	2454	26.73	0.432	26.1–27.28
0.5 μM etoposide	178	15.87	1.534	13.8–17.84
1.0 μM etoposide	184	24.98	2.222	22.1–27.83
2.0 μM etoposide	151	35.17	3.022	31.2–39.06

^aNumber of data points.

^bStandard error of the mean.

^cConfidence interval.

Based on detailed mechanistic analyses discussed below, combined with literature precedents cited in the references, the following hypothesis is put forward to account for the substantial DNA damage caused by Pen. It is proposed that, in addition to other possible effects on DNA structure and function, Pen may cause inhibition of two enzymes of DNA metabolism:

• 1.DNA (cytosine-5)-methyltransferase (MTase, EC 2.1.1.37) and
• 2.DNA glycosylases (EC 3.2.2.) involved in DNA base-excision repair.

Neither of the two components of the hybrid, Cape or Gem alone were reported to inhibit these enzymes.

Validation of the above hypothesis must await future experimentation guided by the ideas discussed in the present communication.

Based on the analogy to the behavior of all anticancer cytidine analogues, it is expected that the triphosphates formed from Pen, F₃dCTP and F₃dUTP, can serve as substrates for DNA polymerases (EC 2.7.7.7.), leading to incorporation of the analogues into DNA. Accordingly, both triphosphates must compete with the natural triphosphates for binding to the DNA polymerases thereby inhibiting DNA synthesis. Just like in the case of Gem, inhibition of DNA polymerases²⁹ may be one of the determinants of the cytotoxicity of Pen.

The presence of 5-fluorocytosine (FC) in place of cytosine in DNA has been shown to inactivate MTase to form a covalent binary complex³⁰ by a mechanism reminiscent of the mechanism-based inactivation of TS by FdUMP.^31,32 It is proposed that an analogous covalent complex may form between F₃dCyd-containing DNA and MTase (see Figure 6), resulting from the inability of the enzyme to cleave the C–F bond. Incorporation of the MTase-inhibitory F₃dCyd residues into DNA would not only result in direct inhibition of 5-C-methylation and depletion of the relatively small MTase pool, but its covalent adduct would also be expected to block DNA replication.³³ Such inhibitory activity resulting in partial “demethylation” of DNA could potentially interfere with epigenetic regulation of cellular metabolism. MTase inhibitory activity would be a novel function for Pen since it has not been reported for either Cape or Gem and could not be anticipated based solely on the combination of the two components in a hybrid. It should be noted that MTase inhibition happens to be the primary mechanism of cytotoxicity of two other anticancer cytidine analogues, azacytidine and decitabine (see Figure 1).

Figure 6.

Proposed covalent ternary complex between F₃dUMP, CH₂–H₄folate, and TS (A) and the analogous binary covalent complexes between F₃dCyd-containing DNA and MTase (B), and their similarity to the covalent interaction of 5-aza-C-containing DNA and MTase (C).

The presence of the two fluorines at the 2′-position of the incorporated analogues derived from Pen are predicted to have at least two distinct effects. First, the sugar pucker should be different from that in DNA, having a 3′-endo conformation. Such a conformational alteration may affect DNA structural dynamics and interactions with DNA binding proteins. This effect is shared by Gem having the identical 2′,2′-difluoro sugar component. Second, the two fluorines are predicted to interfere with the most efficient DNA repair mechanisms dealing with misincorporated base analogues: base excision repair (BER) and mismatch repair (MMR), which operate primarily via a large family of enzymes, the DNA glycosylases.³⁴ This effect is not shared by Gem, because it does not contain a 5-fluorinated base analogue, that is recognized by the BER system to signal for excision by DNA glycosylases. The catalytic mechanisms of DNA glycosylases involve positively charged oxocarbenium ion-like transition states³⁵ that are expected to be destabilized by the two electronegative fluorines at the 2′-position of the sugar (see Figure 7), analogous to the case of TP¹⁴ discussed above.

Figure 7.

Destabilization of the oxocarbenium ion-like like transition state of DNA glycosylases. (A) Monofunctional glycosylase and (B) bifunctional glycosylase.

Accordingly, DNA glycosylases involved in BER may not be able to excise the bases of incorporated F₃dCMP and F₃dUMP and leave the damaged DNA unrepaired. This prediction is supported by the finding that both 2′-mono- and difluoro substitutions of deoxyuridine in oligonucleotides prevented glycosylase mediated base excision.³⁶ Interestingly, some bifunctional glycosylase activities capable of removing the base of a 2′-F-substituted analogue have been detected,³⁷ but these enzymes could not carry out the required cleavage of the 2′-C–F bond in the arabino-configuration (which would be one of the two fluorines during the subsequent β-elimination step (see Figure S2)), stalling the enzyme at the site of the damage.³⁷ This is reminiscent of the mechanisms of inactivation of TS and MTase by fluoropyrimidines³¹⁻³³ discussed above, which results from the lack of cleavage of the C–F bond in the 5-position of the respective substrates (see Figure 6).

Considering the above, structure–activity relationships related to the specific locations of the three fluorines in the molecule are proposed as follows:

• 1.5-F: inactivation of TS by FdUMP; inactivation of MTase by FC in DNA; FC and FU in DNA as unnatural bases are recognized for removal by BER;
• 2.2′, 2′-F: inhibition by F₃dUrd of glycosyl bond cleavage by TP; inactivation of RR by F₃dCDP; inhibition by F₃dCyd and F₃dUrd in DNA of glycosyl bond cleavage by DNA glycosylases;
• 3.2′-F in thearabino-configuration: inhibition of the β-elimination step in bifunctional BER (see Figure S2).

It is concluded that the rational design of a modified structure of capecitabine aimed at blocking its metabolic conversion to 5-fluorouracil and its associated toxicities yielded a new orally active anticancer fluoropyrimidine, Pen, a hybrid of two anticancer drugs, Cape and Gem. It is suggested that, in addition to the characteristic inhibitory effects produced by the individual components, misincorporation of the trifluorinated analogues into DNA derived from Pen may result in two addition functions: inhibition of DNA glycosylases involved in base excision and mismatch repair of DNA damage and inhibition of DNA (cytosine-5)-methyltransferase, involved in epigenetic regulation of cellular metabolism. Both may contribute significantly to the cytotoxicity of Pen.

A comparison of the attributes of Cape, Gem, and Pen with respect to the variety of their targets (shown in Table S2) highlights the multiple potential targets of action of Pen, which may reduce the likelihood of the emergence of drug resistance, one of the aims of the combination chemotherapy of cancer. It is also likely that, as a hybrid drug, Pen may be able to mimic the beneficial effects of the therapeutic combinations^17,18 of its components. Based on its DNA-targeted effects, combinations of Pen with DNA damaging agents, such as radiation, platinum-based drugs, and/or topoisomerase inhibitors may be expected to enhance its efficacy.

Glossary

Abbreviations Used

ADMET

absorption, distribution, metabolism, excretion and toxicity

BER

base excision repair

Cape

capecitabine

CDA

cytidine deaminase (EC 3.5.4.5)

DCDT

deoxycytidylate deaminase (EC 3.5.4.12)

dCK

deoxycytidine kinase (EC 2.7.1.74)

dCMPK

deoxycytidylate kinase (EC 2.7.4.35)

5′-dFUrd

5′-deoxy-5-fluorouridine

dR-1-P

2′-deoxyribose-1-phosphate

dTMPK

dTMP kinase (EC 2.7.4.9)

dUMP

deoxyuridine 5′-monophosphate

dUTP

deoxyuridine 5′-triphosphate

ENT

equilibrative nucleoside transport

FC

5-fluorocytosine

FdUMP

5-fluoro-2′-deoxyuridine 5′-monophosphate

FdUrd

5-fluoro-2′-deoxyuridine

F₂dUMP

2′,2′-fluoro-2′-deoxyuridine 5′-monophosphate

FU

5-fluorouracil

FUTP

5-fluorouridine 5′-triphosphate

F₃dCyd

5,2′,2′-trifluoro-dCyd

F₃dCMP

5,2′,2′-trifluoro-dCMP

F₃dUMP

5,2′,2′-trifluoro-dUMP

F₃dUrd

5,2′,2′-trifluoro-dUrd

Gem

gemcitabine

gem

gemcitabine

HFS

hand-foot syndrome

HMDS

hexamethyldisilazane

MMR

mismatch repair

MTase

DNA (cytosine-5)-methyltransferase (EC 2.1.1.37)

NDPase

nucleoside diphosphatase (EC 3.6.1.6)

NDPK

nucleoside diphosphokinase (EC 3.6.1.6)

NTPase

nucleoside triphosphatase (EC 3.6.1.15)

Pen

pencitabine

RR

ribonucleotide reductase (EC 1.17.4.1)

Pol α

DNA polymerase α

TMSOTf,

trifluoromethane-sulfonate

TP

thymidine phosphorylase (EC 2.4.2.4)

TS

thymidylate synthase (EC 2.1.1.45)

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.1c00565.

• Experimental section; condensation conditions; nucleotide pool determinations; preparation of F₃dUrd (9); enzyme studies; fluorescent images of the comet assay results; beta-elimination step in bifunctional glycosylases blocked by 2′-F (arabino); care and treatment of animals; systemic toxicity; cell culture conditions; comet assay procedure; references; SMILES; HPLC purity check of compound 7; MS spectrum of compound 7; ¹H NMR, ¹³C NMR, and ¹⁹F-NMR spectra of compound 7 (PDF)

Author Present Address

^† 955 Pine Tree Court, East Amherst, NY 14051

A preliminary account of this work was presented at the 2021 Spring Meeting of the American Chemical Society, Abstract No. 3551334.

The author declares the following competing financial interest(s): Kalman, T.I. Multitargeted Nucleoside Derivatives. US Patent, 2020, US10,751,358 B2.

Dedication

This paper is dedicated to the memory of Prof. Thomas J. Bardos—mentor, colleague and friend, whose inspiration played a major role in the development of the ideas described herein.