[CANCER RESEARCH 43, 4039-4044, September 1983]

Metabolic Activation of RS-1-(Tetrahydro-2-furanyl)-5-fluorouracil(Ftorafur) to 5-Fluorouracil by Soluble Enzymes1

Yousry M. El Sayed2 and Wolfgang Sadée3

Departments of Pharmacy and Pharmaceutical Chemistry, School at Pharmacy, University of California, San Francisco, California 94143

ABSTRACT

There are two major fl,S-1-(tetrahydro-2-furanyl)-5-fluorouracil(ftorafur) activation pathways to 5-fluorouracil, one that is mediated by microsomal cytochrome P-450 oxidation at C-5' of the

tetrahydrofuran moiety and one that is mediated by solubleenzymes. This report demonstrates that the soluble enzymepathway proceeds via enzymatic cleavage (possibly hydrolytic)of the N-1—C-2' bond to yield 5-fluorouracil and 4-hydroxybu-

tanal, which is immediately further metabolized to 7-butyrolac-tone or -y-hydroxybutyric acid. The soluble activation pathway

was present in liver, small intestine, and brain. Because of thelimited distribution of cytochrome P-450 in body tissues andbecause of the lack of redistribution of 5-fluorouracil via the

systemic circulation after ftorafur administration, we proposethat the soluble enzyme pathway is at least in part responsiblefor organ toxicity and possibly antitumor effect. Distinction of themicrosomal (C-5') and the soluble enzyme (C-2') activation path

ways can be exploited in the design of more selective prodruganalogues.

INTRODUCTION

The pyrimidine antimetabolite ftorafur4 has shown activity

against several adenocarcinomas with less myelotoxicity butmore central nervous sytem toxicity than those of 5-fluorouracil(29). Ftorafur is considered a prodrug of 5-fluorouracil, and it isslowly metabolized to 5-fluorouracil by several metabolic path

ways (3, 4, 10).Recently, we identified 2 major in vitro pathways of ftorafur

activation to 5-fluorouracil that do not involve the known ftorafur

metabolites (that retain the tetrahydrofuran moiety), one occurring in the 100,000 g supernatant fraction (soluble enzymes) ofliver homogenate and one in the 100,000 g microsomal pellet (3,10). The microsomal pathway involving cytochrome P-450 produces succinaldehyde and 5-fluorouracil via oxidation at the C-5' position of ftorafur (10). In contrast, GBL or its ring open

congener GHB was isolated as a major metabolite in the supernatant soluble enzyme fractions. Chart 1 shows the potentialactivation pathways that can occur at either position C-2' or C-5' of ftorafur with GBL or GHB as the possible common end

product.Precise knowledge of the activation mechanisms of ftorafur

may lead to the development of more selective 5-fluorouracil

prodrugs. In this report, we provide evidence that the solubleenzyme pathway of ftorafur activation proceeds via cleavage of

1This work was supported by National Cancer Institute Grant CA 27866.2 Recipient of a scholarship from the Egyptian Government, in partial support of

this work.3To whom requests for reprints should be addressed.4The abbreviations used are: ftorafur, fl,S-1-<tetrahydro-2-furanyl)-5-fluorouracil;

GBL, -,-butyroiactone: GHB, -y-hydroxybutyric acid; GC, gas chromatography.

Received February 17.1983; accepted May 26.1983.

the N-1—C-2' bond of ftorafur to yield 4-hydroxybutanal as the

intermediate which is further metabolized to GBL or GHB.

MATERIALS AND METHODS

All chemicals and reagents were of spectroquality or analytical grade.Ftorafur was supplied by the Chemical and Drug Procurement Section,Division of Cancer Treatment, National Cancer Institute, Bethesda, Md.NADPH and NADH were purchased from the Sigma Chemical Co. (St.Louis, Mo.). Alcohol dehydrogenase from horse liver (specific activity,about 2.7 units/mg) was purchased from Boehringer Mannheim, Indianapolis, Ind. Because of their chemical instabilities, 4-hydroxybutanal and

succinaldehyde had to be synthesized immediately preceding their use.4-Hydroxybutanal was generated from the precursor 2,3-dihydrofuran

(Aidrich Chemical Co., Milwaukee, Wis.) under acidic conditions (21).Succinaldehyde was generated from 2,5-dimethoxytetrahydrofuran (Aid-

rich Chemical Co.) under acidic conditions (17).Tissue Homogenate Preparations and in Vitro Incubations. Male

Dutch rabbits were stunned and decapitated. The liver, brain, and smallintestine were excised and rinsed in ice-cold isotonic (1.15%) KCI. Eachorgan tissue was homogenized in ice-cold 0.01 M potassiunrsodiumphosphate buffer (pH 7.4):1.15% KCI (1:2, w/v) using a Potter-Elvehjem

Teflon-pestle homogenizer. The homogenates were centrifuged at10,000 g for 20 min at 0-4° to yield the 10,000 g supernatant fraction.

The supernatant fractions were recentrifuged at 100,000 g for 1 hr at 0-4°.The 100,000 g supernatant fraction may be used immediately orfrozen at -20° until use.

Several experiments were carried out with dialyzed 100,000 g liverhomogenates. The dialysis tubes containing the homogenates wereincubated in 200 volumes of 100 mM Tris buffer, pH 7.4, for 5 hr at 4°.

A high concentration of Tris buffer was chosen to accelerate soluteexchange; moreover, the buffer was replaced every hr.

All in vitro incubations were earned out at pH 7.4 at 37°for 1 hr. Final

concentration of all preparations was standardized to 1 g tissue (wetweight) per 2 ml incubation mixture. Chemical reagents and cofactorswere added at the concentrations indicated. The concentrations ofNADPH and NADH were 1 mM each. When incubations were performedfor more than 1 hr, additional amounts of NADPH and NADH were addedevery hr (yielding concentration increments of 1 mw each) in order toprevent cofactor depletion. Higher concentrations of NADPH and NADHfailed to increase the rate of product formation. The rate of GBL or GHBformation from ftorafur was linear over at least 4 hr.

GG Assay for GBL or GHB, 4-Hydroxybutanal, and 1,4-Butanediol.GBL or GHB, 4-hydroxybutanal, and 1,4-butanediol were measured by

the previously published GC assay of GBL or GHB (3) with the followingmodifications. The column, injector, and detector temperatures were145,190, and 210°,respectively. GC retention times for GBL, 4-hydroxy

butanal, and 1,4-butanediol were 3.2, 1.5, and 9.5 min, respectively.There was no detectable conversion of 4-hydroxybutanal and succinal-

dehyde to GBL or GHB under the assay conditions. Sensitivity for GBLwas approximately 0.5 ¿/g/mltissue homogenates. Standard curves werelinear over the range of 0.5 to 200 /<g/ml. There is no measurabledecomposition of ftorafur to GBL at concentrations up to 200 jig/mlduring the assay procedures.

High-Pressure Liquid Chromatographie Assay of 5-Fluorouracil. 5-Fluorouradl was measured in the in vitro incubates with a modified high-

SEPTEMBER 1983 4039

Research. on January 2, 2020. © 1983 American Association for Cancercancerres.aacrjournals.org Downloaded from

Y. M. El Sayed and W. Sadée

C-2' IIIMTMI O°^en tu

4-11

o—1,4-10

O«/* COOH _HO<X\ f00*1

SSI S.ccinic feM

Chart 1. Three possible activation pathways of ftorafur (FT) that lead to 5-fluorouracil (FUra) and the expected products of the tetrahydrofuran moiety. Potential interconversion among these products is also indicated. 4-HB, 4-hydroxybu-tanal; 1,4-BD. 1.4-butanediol: SA. succinaldehyde; SSA. succinic semialdehyde.

pressure liquid Chromatographie assay (30) using prepurifìcationof extracts by column chromatography for sufficient sensitivity (100 ng/ml).The column chromatography afforded separation of 5-fluorouracil from

the endogenous interferences present in the tissue homogenates. To 1ml of the in vitro incubation of ftorafur (2 mw) with the soluble enzymeof the 100,000 g supernatant, 0.1 ml of 0.5 M NaH2PO4 and 5 /¿gof 5-

bromouracil:50 n\ methanol as internal standard were added. Sampleswere then extracted with 2 x 10 ml ethyl acetate, and the ethyl acetatelayers were evaporated to dryness under nitrogen gas at 60°. The

residue was dissolved in a small volume (100 /tl) of methanol andsubjected to column chromatography (14 x 0.9 cm, inside diameter; silicagel, 100 to 200 mesh; Bio-Rad, Richmond, Calif.), eluting with 10 ml

acetonitrile:chloroform:water (40:4:1) (20). The appropriate eluate fraction was evaporated to dryness under nitrogen gas. The residue wasdissolved in 100 n\ of methanol, and 25 to 50 ti\ were injected into theliquid Chromatograph under the following conditions: Ci8-¿iBondapak(30

cm x 4 mm, inside diameter; Waters Associates, Milford, Mass.); mobilephase, 2 (TIMsodium acetate buffer (pH 4.2):acetonitrite (99:1, v/v); flowrate, 2.0 ml/min; detector, UV, 280 nm. The retention times of 5-fluorouracil, 5-bromouracil, and ftorafur were 2.5, 5.5, and 23 min,respectively. Overall analytical recovery of 5-fluorouracil from the homogenates was approximately 50%. 5-Fluorouracil concentrations werequantitated by 5-fluorouracil:5-bromouracil UV absorbance ratios at 280nm. Standard curves for 5-fluourouracil obtained from peak height ratio

measurements were linear over a wide concentration range (0.1 to 200¿ig/mltissue homogenate). The assay was sensitive to 0.1 /<g/ml tissuehomogenste. Ftorafur decomposition to 5-fluorouracil was minimal under

the assay conditions (<0.08%).

RESULTS

Metabolism of Succinaldehyde, 4-Hydroxybutanal, and

GBL or GHB. Under a variety of homogenates conditions tested(10,000 g and 100,000 g supernatants), GBL or GHB wasessentially stable over the incubation period. In contrast, both 4-

hydroxybutanal and succinaldehyde were rapidly metabolized,with GBL or GHB as a major product. In order to understand theintermediate role that either one of these 2 tetrahydrofuranmetabolites may play in the conversion of ftorafur to GBL orGHB, it was necessary to study their metabolism in more detail.1,4-Butanediol was not measurable in detectable quantities in

any of the incubations.

Chart 2 shows the percentage of conversion of succinaldehyde(500 MM)to GBL or GHB in rabbit liver homogenates (100,000 gsupernatant fraction). The addition of NADPH or NADH greatlyenhanced the conversion of succinaldehyde to GBL or GHB,while incubation without any cofactor added to the solubleenzyme preparation generated only 12% GBL or GHB. In addition, incubations without enzyme and only with phosphate bufferdid not generate any detectable GBL or GHB. Acid treatment ofsuccinaldehyde which is part of the GC assay procedure did notproduce any detectable amounts of GBL:GHB, thereby rulingout any procedural artifacts. Incubation of lower concentrationsof succinaldehyde (50 Õ/M)gave similar yields of GBL or GHB.

Incubations of succinaldehyde (500 A/M)with the 100,000 gsupernatants in the presence of NADPH and hydrazine (2 mw)produced an 80% inhibition of GBL or GHB product formation(Chart 2). Hydrazine was chosen because it was expected toform a cyclic bis(hydrazone) with succinaldehyde (17). In contrast, pyrazole (2 HIM)(an alcohol dehydrogenase inhibitor) (22),A/-ethylmaleimide (1 mw) (an inactivator of sulfhydryl groups) (8),and semicarbazide (2 mw) (an aldehyde-trapping reagent), did

not produce a significant change in the yield of GBL or GHB.Chart 3 shows the percentage of conversion of 4-hydroxybu-

tanal (50 ^M) to GBL or GHB under different liver 100,000 ghomogenate incubation conditions. Most strikingly, 4-hydroxy-

butanal is very rapidly and quantitatively converted to GBL orGHB without any cofactor added. NADPH slightly reduced theyield of GBL or GHB, while hydrazine had very little effect on thereaction. However, the conversion to GBL or GHB was inhibitedby 80% with NADH plus horse liver alcohol dehydrogenaseadded to the 100,000 g supernatant fractions. NADH and horseliver alcohol dehydrogenase were selected for this experiment,since they were expected to reduce 4-hydroxybutanal to 1,4-

butanediol, thereby inhibiting the conversion to GBL or GHB.Incubations in the presence of A/-ethylmaleimide (1 mw), sem-

Chart 2. Conversion of succinaldehyde (500 »M)to GBL or GHB by the 100,000g supernatants of rabbit liver homogenates under different incubation conditions.Incubation time was 1 hr. Control samples contained supernatant alone. Theinhibition effect of hydrazine was tested in the presence of NADPH, which affordsan efficient conversion of succinaldehyde to GBL or GHB. Reagent concentrationswere 1 mw NADPH, 2 HIM NADH, and 2 mM hydrazine. The reducing cofactorsNADPH and NADH were added, since the conversion of succinaldehyde to GBL orGHB includes a reduction of one of the aldehydic functions to an alcohol.

4040 CANCER RESEARCH VOL. 43

Research. on January 2, 2020. © 1983 American Association for Cancercancerres.aacrjournals.org Downloaded from

Ftorafur Activation by Soluble Enzymes

Charts. Conversion of 4-hydroxybutanal to GBL or GHB by the 100,000 g

supernatant of rabbit liver homogenates under different incubation conditions. Theincubation time was 1 hr. Hydrazine was tested in the absence of reducingcofactors, since they are not required but indeed inhibit the conversion of 4-hydroxybutanal to GBL or GHB. Reagent concentrations were 1 mM NADPH, 2HIM NAOH, 2 mM hydrazlne, and horse liver alcohol dehydrogenase (ADH) (4 mg/

icarbazide (2 mM),and pyrazole (2 mM)again had no significanteffect on the yield of GBL or GHB, while incubations in thepresence of EDTA (5 mw) produced only a slight inhibition(~20%). Preheating of the 100,000 g supernatant liver homogenates at 100°for 30 min produced 85 to 90% inhibition of GBL

or GHB yield, which suggested that a small portion of the reactionmay occur nonenzymatically in the homogenate incubations.Omission of the liver homogenates from the incubation buffereliminated GBL or GHB yield from 4-hydroxybutanal.

Metabolic Conversion of Ftorafur to GBL or GHB and 5-Fluorouracil.Chart 4 shows the conversionof ftorafurto GBLor GHB (percentage of control) under different 100,000 g liverhomogenate incubation conditions. Potential inhibitors selectedfor this experiment were those active as inhibitors of either 4-hydroxybutanal or succinaldehyde conversion to GBL or GHB.Ftorafur is converted to GBL or GHB in highest yield withoutcofactor addition. Hydrazine did not reduce GBL or GHB yield,while addition of NADPH and, more strongly, NADH (±horseliver alcohol dehydrogenase) inhibited GBL:GHB formation fromftorafur to a similar extent as that from 4-hydroxybutanal. Incubation of ftorafur with phosphate buffer (pH 7.4) at 37°for 1 hrresulted in no detectable 4-hydroxybutanal or GBL or GHB. Inaddition, preheating of the liver homogenates (100°for 30 min)

eliminated the formation of GBLGHB from ftorafur, which suggests that the reaction is enzymatic.

In order to address the question of whether or not thymidineor undine phosphorylase is responsible for metabolic activationof ftorafur to 5-fluorouracil, we carried out in vitro incubationswith 10 mw thymidine and with 2'-deoxy-5-trifluoromethyluridine

at a concentration (1.6 HIM)that was shown to inhibit nucleosidephosphorylase (7, 14). Under these conditions, ftorafur conver-

Chart4. Conversion of ftorafur (2 mM) to GBL or GHB by the 100,000 gsupernatant of rabbit liver homogenates under different incubation conditions.Incubation time was 1 hr. Control samples contained supernatant alone. Reagentconcentrations were 1 mM NAOPH, 2 mM NADH, 2 mM hydrazine, and horse liveralcohol dehydrogenase {ADH) (4 mg/ml).

sion to GBL or GHB was affected only to a small extent (Table1). Furthermore, the enzyme preparation was dialyzed as described ("Materials and Methods") in order to remove phosphate

which is required for the phosphorylation reaction. Subsequentincubation of ftorafur with the dialyzed enzyme preparation withor without phosphate buffer (10 HIM)yielded similar amounts ofGBL or GHB. However, dialysis resulted in partial loss (~70%)of enzymatic activity, which prevents any firm conclusions thatmay have been derived from this experiment.

Activationof Ftorafur by Different Tissue Homogenates. Invitro experiments were carried out using various organ tissuehomogenates. Table 2 represents the results of in vitro incubations of ftorafur with rabbit liver, brain, and small intestine100,000 g supernatant enzyme fractions. The yield of 5-fluorouracil from ftorafur is lower than that of GBL or GHB. Therefore,the rate of disappearance of 5-fluorouracil (77 /¿M)was measuredin 100,000 g liver supernatants (without the addition of anycofactor). Within the first 20 min of incubation, 5-fluorouracil wasrapidly metabolized by more than 80%, which was possibly aresult of the presence of residual endogenous cofactor (e.g.,NADPH, ribose 1-phosphate, 5-phosphoribosyl 1-pyrophos-phate). After this initial rapid phase, 5-fluorouracil metabolismwas much slower (f1/2~ 150 min). Although exact quantitativecalculations of the expected 5-fluorouracil yield from ftorafurwere not readily possible under these conditions, the data areroughly compatible with the hypothesis that 5-fluorouracil andGBL or GHB were generated in stoichiometric amounts. Liverhomogenates showed highest levels of ftorafur activation, followed by small intestines and brain. Although 5-fluorouracil concentrations were rather small in the brain homogenates, they areclearly above the nonenzymatic background (<0.08% conversion). In order to estimate the Michaelis-Menten constant («„,)for the overall reaction of ftorafur to GBL or GHB, we carried out

SEPTEMBER 1983 4041

Research. on January 2, 2020. © 1983 American Association for Cancercancerres.aacrjournals.org Downloaded from

Y. M. El Sayed and W. Sadée

Tabte!In vitro metabolism of ftorafur (2 mu) to GBL or GHB by rabbit liver 100,000g

supernatant enzyme fraction

Condition % of conversion to GBL or GHB/hr100,000 g supernatant atone 1.4 ±0.2° (5)"100,000 g supernatant + 2'-deoxy-5'-trifluo- 1.1"

romethyluridine (1.6 mM)100,000 g supernatant + thymidine (10 RIM) 1.0 ±0.1 (2)100,000 g dialyzed supernatant 0.4 ±0.03 (4)"

100,000 g dialyzed supernatant + phosphate 0.5 ±0.04 (2)buffer. pH 7.4 (10 mm)* Mean ±S.D., unless 2 determinations; then, value is the mean ±one-half of

the difference between the 2 observations.'' Numbers in parentheses, number of determinations.c Single observation." The dialyzed enzyme preparations were prepared as described under "Mate

rials and Methods" using 10 mM Tris buffer for homogenization and 100 mM Tris

buffer for accelerated dialysis.

Table2In vitro metabolism of ftorafur (2 mu) to GBL or GHB and 5-fluorouracil by

100,000 g supernatant fractions of rabbit liver, small intestine, and brain

% of converston/hr

Liver Small intestine Brain

GBLGHB5-Fluorouracil

1.4 ±0.2*(5)"

0.6 ±0.1 (2)0.9 ±0.2 (2)0.7 ±0.1 (2)

0.43 ±0.04 (2)0.2 ±0.03(2)

* Mean ±S.D., unless 2 determinations; then, value is the mean ±one-half

range.6 Numbers in parentheses, number of determinations.

in vitro incubations with increasing ftorafur concentrations (2, 5,and 10 HIM)and determined the rate of formation of GBL or GHBper hr. Reaction rates were linear over at least 4 hr. The hourlyyields of GBL or GHB formed were 33.8, 51.4, and 64.8 fiM(percentage of conversion, 1.7, 1.03, and 0.65, respectively).Using the double-reciprocal plot (Lineweaver-Burk) of these data,

the apparent Kmof the enzyme is approximately 3 ITIM.Since thereaction rate of 4-hydroxybutanal to GBL or GHB was muchgreater at rather high 4-hydroxybutanal levels (10 to 100 UM), it

is likely that this Kmvalue is associated with the direct activationof ftorafur.

DISCUSSION

The formation of GBL or GHB and 5-fluorouracil from ftorafur

in the soluble enzyme preparation of 100,000 g supernatantfractions may be mediated by either of the 3 pathways describedin Chart 1. These include oxidation at C-2' and C-5' positions

which leads to chemically labile intermediates that spontaneouslycleave to 5-fluorouracil and to GBL or GHB (C-2' oxidation) orto succinaldehyde (C-5' oxidation). Another potential pathway offtorafur activation involves C-2' hydrolytic cleavage which pro

duces 5-fluorouracil and 4-hydroxybutanal. Furthermore, succinaldehyde and 4-hydroxybutanal can undergo enzymatic conver

sion to GBL or GHB. In fact, GBL or GHB was isolated as amajor metabolite of ftorafur in the soluble enzyme fraction of100,000 g supernatants (3).

Since GBL or GHB is a major product of several possiblemetabolic pathways of ftorafur activation, it is necessary to firststudy the metabolic activation of the postulated intermediates,4-hydroxybutanal and succinaldehyde. The specific inhibition of

their conversion to GBL to GHB can then be applied to test GBLor GHB formation pathways from ftorafur.

The formation of GBL or GHB from succinaldehyde in thesoluble enzyme preparation could occur by either one of the 2reaction sequences described in Chart 1: (a) reversible reduction

to 4-hydroxybutanal followed by oxidation; or (b) oxidation to

succinic semialdehyde followed by reversible reduction to GBLor GHB. It has been previously reported that succinic semialdehyde (an intermediate metabolite in 7-aminobutyric acid metab

olism) is reversibly metabolized to GBH by soluble enzymefractions of brain and other tissue homogenates (11). The conversion of succinaldehyde to GBL or GHB requires NADPH as acofactor, and this conversion was strongly inhibited by hydrazine(an aldehyde-trapping reagent which can undergo bifunctional

condensation with succinaldehyde to form a heterocyclic ringstructure) (Chart 2). Kaufman and Nelson (18) had previouslyused hydrazine to trap succinic semialdehyde produced fromoxidation of GHB in vitro. In contrast, hydrazine had very littleeffect on the conversion of both 4-hydroxybutanal and ftorafur

to GBL or GHB (Charts 3 and 4). This result rules out succinaldehyde as an intermediate of the conversion of ftorafur to GBLor GHB and thus rules out C-5' oxidation by the soluble enzymes

as the major activation pathway.There are also several pathways of 4-hydroxybutanal metab

olism by the soluble enzymes of 100,000 g supernatants (Chart1), 4-Hydroxybutanal could be directly oxidized to GBL or GHB,

reversibly oxidized to succinaldehyde or reversibly reduced to1,4-butanediol. Each of these reactions could involve a different

enzymatic mechanism, requiring different cofactors. The metabolic conversion of 1,4-butanediol to GBL or GHB as the major

product has been demonstrated previously (12, 24). We havefound that the conversion of 4-hydroxybutanal to GBL or GHB

in 100,000 g homogenates is extremely rapid. Therefore, it wasnot possible to directly demonstrate the presence of 4-hydroxybutanal in ftorafur incubations. Conversion of 4-hydroxybutanalto GBL or GHB did not require any external cofactor but it wasslightly reduced by NADPH and strongly inhibited (80%) by NADHplus added horse liver alcohol dehydrogenase. Similarly, NADPHand, more strongly, NADH (±horse liver alcohol dehydrogenase)inhibited GBL or GHB formation from ftorafur to a similar extentas that from 4-hydroxybutanal. These results provide strongevidence in support of the hypothesis that 4-hydroxybutanal is

an essential intermediate in the conversion of ftorafur to GBL orGHB. Moreover, the lack of any cofactor requirements arguesagainst an oxidation of ftorafur at C-2' to give 5-fluorouracil and

GBL or GHB directly.It has been previously reported that horse liver thymidine

phosphorylase and rat liver undine phosphorylase were inactiveagainst ftorafur (5, 30). Recently, Kono ef al. (19) suggested thatthe cleavage of ftorafur to 5-fluorouracil was catalyzed by a

thymidine phosphorylase activity, which was enhanced in humantumor tissues. Moreover, ftorafur activation was suppressed inthe presence of a 7-fold excess amount of thymidine (19). Lack

of a rigorous quantitative analysis of the data makes it difficultto assess the significance of their results. In contrast, in ourexperiments, addition of thymidine had little effect on the yieldof GBL or GHB, and the presence of the nucleoside phosphorylase inhibitor (2'-deoxy-5-trifluoromethyluridine) did not inhibit

the metabolic activation of ftorafur to GBL or GHB. These resultsargue against the hypothesis that activation occurs via phospho-rolysis of the N-1—C-2' bond of ftorafur, yielding 5-fluorouracil

and 2-hydroxytetrahydrofuran 2-phosphate, which is then converted to 4-hydroxybutanal by phosphatases. Rather, we pro

pose that a hydrolytic enzyme mechanism is responsible foractivation of ftorafur in the 100,000 g supernatants.

Although ftorafur is thought to act predominantly via 5-fluoro-

4042 CANCER RESEARCH VOL. 43

Research. on January 2, 2020. © 1983 American Association for Cancercancerres.aacrjournals.org Downloaded from

Ftorafur Activation by Soluble Enzymes

uracil formation, plasma concentrations of 5-fluorouracil gener

ated in vivo were shown to be negligible (5); this finding isconsistent with the hypothesis that intracellulariy formed 5-

fluorouracil is further metabolized without subsequent redistribution via the systemic circulation (2, 5). Therefore, the mechanism of metabolic activation of ftorafur to 5-fluorouracil might be

a determinant of its selective tissue toxicity towards the gastrointestinal tract and the central nervous system as well as of itsantitumor effects. We demonstrate here that these tissues arecapable of enzymatically activating ftorafur to 5-fluorouracil by

soluble enzyme catalysis. Although the extent of in vitro activation is rather small, it must be considered that ftorafur is presentin rather large concentrations after therapeutic doses. Moreover,because of the long elimination half-life (6 to 17 hr) of ftorafur (5,6,16), tissue exposure to 5-fluorouracil as a result of the soluble

ftorafur activation pathway could be sufficient to account for theobserved organ toxicities of the drug. On the other hand, wecannot rule out the possibility that metabolites other than 5-

fluorouracil contribute to organ toxicity (1, 13). For example,GHB occurs physiologically in brain and cerebrospinal fluids (9,23, 25, 28) and has been used as an anesthetic adjuvant (15).Its effects on the central nervous system lead to behavioraldepression and an induction of an epileptic-like stupor (lethargy)

(6, 27). However, the plasma concentrations of GBL or GHBrequired for such effects are far above those generated fromftorafur.

In conclusion, this report demonstrates that ftorafur is activated to 5-fluorouracil by the soluble enzymes of target tissuesvia cleavage of the N-1—C-2' bond (Chart 5). The intermediate

product, 4-hydroxybutanal, is rapidly converted to GBL or GHB

which has been observed previously as a major ftorafur metabolite in circulating plasma. The other major metabolic activationpathway, i.e., ftorafur oxidation at C-5' by a cytochrome P-450-

mediated mechanism (10), leads to 5-fluorouracil and succinal-

dehyde (Chart 5), the latter being also converted, at least partially, to GBL or GHB. Since cytochrome P-450 is mainly localized

in the liver, with lower levels in the gastrointestinal tract andmuch lower levels in the brain, and since 5-fluorouracil is not

redistributed via the circulation (2, 5), it is possible that the

?//* "W**/> V«o +

Chart 5. The 2 major activation pathways of ftorafur (FT) to 5-fluorouracil (FUra).Note that the 2 pathways attack at different sites of the ftorafur moiety. While onlya fraction of the generated succinaldehyde (SA) is converted to GBL or GHB. 4-hydroxybutanal (4-HB) quantitatively yields GBL or GHB.

soluble enzyme pathway is responsible for part of the organtoxicity as well as for the antitumor effects of ftorafur. Otherpreviously reported activation pathways are unlikely to contributesignificantly to the ftorafur effects. Further development of pro-drugs related to ftorafur should therefore be directed towardsseparating the microsomal (C-5') and soluble enzyme (C-2')

activation pathways of the tetrahydrofuran ring moiety. Moreover, assay of the soluble activation pathway of ftorafur mayserve as a predictor of the antitumor efficacy of the drug.

REFERENCES

1. Ansfieid. F., Kallas, G., and Singson, J. Clinical results with Tegafur (ftorafur)in cotorectal and breast cancer. Proc. Am. Assoc. Cancer Res., 22: 450,1981.

2. Au, J. L, and Sadée,W. 5-Fluorouracil concentrations in human plasmafollowing fl,S-1-(tetrahydro-2-furanyl)-5-fluorouracil (ftorafur) administration.Cancer Res.. 39; 4289-4290,1979.

3. Au, J. L., and Sadée,W. Activation of ftorafur [fl,S-1-(tetrahydro-2-furanyl)-5-fluorouracil] to 5-fluorouracil and -y-butyrolactone. Cancer Res., 40: 2814-2819, 1980.

4. Au, J. L., and Sadée,W. The pharmacology of ftorafur (fl,S-1-(tetrahydro-2-furanyO-5-fluorouracil). Recent Results Cancer Res., 76:100-114,1981.

5. Au, J. L., Wu, A. T., Friedman. M. A., and Sadée,W. Pharmacokinetics andmetabolism of ftorafur in man. Cancer Treat. Rep., 63: 343-350,1979.

6. Benvenuto, J. A., Lu, K., Hall, S. W., Benjamin, R. S., and Loo, T. L. Dispositionand metabolism of 1-(tetrahydro-2-furanyl)-5-fluorouracil (ftorafur) in humans.Cancer Res., 38: 3867-3870,1978.

7. Cooper, G. M., and Gréer,S. Irreversible inhibition of dehalogenation of 5-¡odouracilby 5-diazouraoil and réversibleinhibition by 5-cyanouracil. CancerRes., 30: 2937-2941,1970.

8. Culp, H. W., and McMahon, R. E. Reducíase for aromatic aldehydes andketones. J. Biol. Chem., 243: 848-852,1958.

9. Doherty, J. D., Hattox, S. E., Snead, O. C., and Roth, R. H. Identification ofendogenous •y-hydroxybutyratein human and bovine brain and its regional

distribution in human, guinea pig and rhesus monkey brain. J. Pharmacol. Exp.Ther., 207:130-139,1978.

10. El Sayed, Y. M., and Sadée,W. Metabolic activation of ftorafur [fl,S-1-(tetrahydrc-2-furanyl)-5-fluorouracil]: the microsomal oxidative pathwayBrachem. Pharmacol., 37: 3006-3008,1982.

11. Fishbein, W. N., and Bessman, S. P. -x-Hydroxybutyrate in mammalian brain.Reversible oxidation by lactic dehydrogenase. J. Bio). Chem.. 239: 357-361,1964.

12. Gessa, G. L., Spano, P. F., Vargin, L., Crabai, F., Tagliamonte, A., and Mameli.L. Effect of 1,4-butanediol and other butyric acid congeners on brain cate-cholamines. Life Sci., 17: 289-298,1968.

13. Harrison, S. D., Jr., Donine, E. P., and Giles, H. D. Evidence for toxicologieactivity of ftorafur independent of conversion to 5-fluorouracil. Cancer Treat.Rep., 63: 1389-1391,1979.

14. Heidelberger, C., Bimie, G. D., Boohar, J., and Wentland, D. Fluorinatedpyrimidines XX. Inhibition of the nucleoside phosphorylase cleavage of 5-fluoro-2'-deoxyuridine by 5-trifluoromethyl-2'-deoxyuridine. Btochim. Biophys.

Acta, 76:315-318,1963.15. Helrich, M., McAslan, R. C.. and Skolnik, S. Correlation of blood levels of 4-

hydroxybutyrate with stages of consciousness. Anesthesiology. 23:771-775,1964.

16. Hills. E. B.. Godefroi, V. G., O'Leary, I. A., Burke, M., Amdrezejewski, D.,

Brukwinski, W., and Horwitz, J. P. GLC determination for ftorafur in biologicalfluids. J. Pharm. Sci., 66:1497-1499,1977.

17. Hjeds, H., and Larsen. I. K. A reinvestigation of the structure of the condensation products of phenylhydrazine and succinaldehyde. Acta Chem. Scand.,25:2394-2402,1971.

18. Kaufman, E. E., and Nelson, T. Kinetics of coupled -x-hydroxybutyrate oxidationand D-glucuronate reduction by an NADP*-dependent oxidoreductase. J. Bid.Chem., 256: 6890-6894,1981.

19. Kono, K., Hará,Y., and Matsushima, Y. Enzymatic formation of 5-fluorouracilfrom 1-{tetrahydro-2-furanyl)-5-fluorouracil (Tegafur) in human tumor tissues.Chem. Pharm. Bull.. 29: 1486-1488,1981.

20. Marunaka, T., Minami, Y., Umeno, Y., Yasudia, A., Sato, T., and Fujii, S.Identification of metabolites of 1-(tetrahydro-2-furanyl)-5-fluorouracil (FT-207)formed in vitro by rat liver microsomes. Chem. Pharm. Bull., 28: 1795-1803,1980.

21. Paul, R., Fluchaire. M., and Collardeau, G. Transposition des dihydro-2.5furannes en dihydro-2.3 furannes. Application a la préparationde l'hydroxy-4

butanal. Bull. Soc. Chim. Fr., 668-671,1950.22. Raskin, N. H., and Sokoloff, L. Alcohol dehydrogenase activity in rat brain and

liver. J. Neurochem., 17: 1677-1687,1970.23. Roth, R. H., and Giarman, N. J. -y-Butyrolactone and yhydroxybutyric. i.

Distribution and metabolism. Brachem. Pharmacol., 75:1333-1348,1966.24. Roth, R. H., and Giarman, N. J. Evidence that central nervous system depres-

SEPTEMBER 1983 4043

Research. on January 2, 2020. © 1983 American Association for Cancercancerres.aacrjournals.org Downloaded from

Y. M. El Sayed and W. Sadée

sten by 1,4-butaned¡olis mediated through a metabolite, gamma-hydroxybu- graphic and behavioraleffects. Neurology, 26: 51-56,1976.tyrate. Biochem. Pharmacol.. 17:735-739,1968. 28. Tabakoff, B., and Radulavacki,M. -,-Hydroxybutyrate in CSF during steepand

25. Roth, R. H., and Giarman, N. J. Natural occurrence of f-hydroxybutyrate in wakefulness. Res. Commun. Chem. Pathol. Pharmacol., 14: 587-590,1976.mammalianbrain. Biochem. Pharmacol., 19:1087-1093,1970. 29. Valdivieso, M., Bodey, G. P., Gottlieb, J. A., and Freireich, E. J. Clinical

26. Snead. O. C. Gamma-hydroxybutyrate in the monkey. I. Electroencephalo- evaluation of ftorafur (pynmidine-deoxyribose W-1̂ '-íuranidyl-S-fluorouracil).

graphic, behavioral and pharmacokmetic studies. Neurology, 28: 636-648, Cancer Res., 36:1821-1824,1976.1978. 30. Wu, A. T., Au, J. L., and Sadée,W. Hydroxylated metabolites of fl.S-1-

27. Snead, O. C., Yu, R. K., and Hutttentocher, P. R. Gamma-hydroxybutyrate, (tetrahydro-2-furanyl)-5-fluorouracil(ftorafur) in rats and rabbits. Cancer Res.,correlation of serum and cerebrospinal fluid levels with electroencephalo- 38:210-214,1978.

4044 CANCER RESEARCH VOL. 43

Research. on January 2, 2020. © 1983 American Association for Cancercancerres.aacrjournals.org Downloaded from

1983;43:4039-4044. Cancer Res   Yousry M. El Sayed and Wolfgang Sadée  5-Fluorouracil by Soluble Enzymes-1-(Tetrahydro-2-furanyl)-5-fluorouracil (Ftorafur) to

R, SMetabolic Activation of

  Updated version

  http://cancerres.aacrjournals.org/content/43/9/4039

Access the most recent version of this article at:

   

   

   

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

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/43/9/4039To request permission to re-use all or part of this article, use this link

Research. on January 2, 2020. © 1983 American Association for Cancercancerres.aacrjournals.org Downloaded from