novel mechanism(s) of resistance to 5-fluorouracil in...

[CANCER RESEARCH 52. 1855-1864. April 1. 1992]

Novel Mechanism(s) of Resistance to 5-Fluorouracil in Human Colon Cancer(HCT-8) Sublines following Exposure to Two Different ClinicallyRelevant Dose Schedules1

Carlo Aschele,2 Alberto Sobrero, Mary A. Faderan, and Joseph R. Bertino'

Department of Molecular Pharmacology; Memorial Sloan-Kettering Cancer Center, New York, New York 10021 [C. A., M. A. F., J. R. B.J, and Department of MedicalOncology, Istituto Nazionale per la Ricerca sul Cancro, 16132 Genova, Italy [A. S.]

ABSTRACTMechanisms of resistance to 5-fluorouracil (FUra) were compared

between a cell line resistant to a short-term exposure (4 h) to this agent(HCT-8/FU4hR) and a cell line resistant to a prolonged exposure (7days) to the fluoropyrimidine (HCT-8/FU7dR). The two cell lines wereobtained by repeatedly exposing 2 x I(T cells to a constant concentrationof FUra (1000 MMfor 4 h or 15 MMfor 7 days), able to produce 3-4 logsof cell kill. HCT-8/FU4hR cells were still sensitive to FUra given as a7-day exposure, suggesting different mechanisms of resistance. In addition, HCT-8/FU7dR cells were cross-resistant to fluorodeoxyuridine and,to a lesser degree, methotrexate; while HCT-8/FU4hR cells were not.Both HCT-8/FU4hR and HCT-8/FU7dR cells were similar to parentalHCT-8 cells with regard to uptake of FUra as well as the pattern ofFUra-metabolizing and FUra target enzymes. Although neither in situthymidylate synthase (TS) activity nor the degree of its inhibition byFUra showed any evidence of alteration in HCT-8/FU7dR cells, a rapidrecovery of TS activity after drug removal was evident in this cell line.The addition of as much as 100 MMleucovorin did not completely inhibitthe recovery of thymidylate synthesis after FUra exposure. No differences were detected in the kinetic properties (A,,,for 2'-deoxyuridylateand 5,10-methylenetetrahydrofolate, concentration producing 50% inhibition for fluorodeoxyuridylate) or TS from HCT-8/FU7dR cells ascompared to parental HCT-8 TS. Baseline levels of 5,10 methylenete-trahydrofolate were decreased in HCT-8/FU7dR cells, and analysis ofthe chain length distribution of the polyglutamylated form of the folatecofactor showed that in this cell line the defect in 5,10-methylenetetrahydrofolate levels is accompanied by, and possibly due to, a defect in thepolyglutamylation of this cofactor. In contrast, HCT-8/FU4hR cells weresimilar to the parental cell line with regard to both the degree of in situTS inhibition by FUra and duration of inhibition after FUra removal.Labeling studies with |'I !-<>|lIra (4 h exposure, 100 MM)showed that

the incorporation of the fluoropyrimidine into RNA is significantly decreased in HCT-8/FU4hR cells as compared to parental HCT-8 cells.The mechanisms of resistance found in these cell lines indicate that themechanism of cell kill by FUra differs depending on the dose scheduleused: short-term exposure to high concentrations of FUra kills cells byan RNA effect, while prolonged exposure to low doses is cytotoxic viainhibition of thymidylate synthesis and, consequently, DNA synthesis.

INTRODUCTIONFUra4 has been used for 3 decades for the treatment of human

solid tumors (1). Of particular note is its role in the treatment

Received 9/6/91; accepted 1/27/92The costs of publication of this article were defrayed in part by the payment

of page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1Supported by grant 9000140PF70 from Consiglio Nazionale delle Ricerche

(Progetto Biotecnologie), grant 1991 from AIRC, and grant CA 47179 from theUSPHS.

! Fellow of the Istituto Nazionale per la Ricerca sul Cancro, Genova, Italy.

Current address: Department of Medical Oncology, Istituto Nazionale per laRicerca sul Cancro. V.le Benedetto XV, 10 16132 Genova. Italy.

3American Cancer Society Professor of Medicine and Pharmacology. To

whom requests for reprints should be addressed, at Department of MolecularPharmacology, Memorial Sloan-Kettering Cancer Center, 1275 York Avenue,New York, NY 10021.

4The abbreviations used are: FUra, 5-fluorouracil; LV. leucovorin; TS, thymidylate synthase: FdUrd, 5-fluoro-2'-deoxyuridine; CHÂ¡FH4.5,10-methylenetetrahydrofolate: MTX, methotrexate; dUMP. 2'-deoxyuridylate; FdUMP. 5-fluoro-2'-deoxyuridylate; FH4. tetrahydrofolate: TCA, trichloroacetic acid; PBS.phosphate-buffered saline: FUTP. 5-fluorouridine 5'-triphosphate: ICW. concen

tration producing 50% inhibition; ED,0. median effective dose.

of colorectal cancer, in which this drug has been used eitheralone or in combination with modulating noncytotoxic agents,e.g., LV, levamisole, rather than being included in polychem-otherapeutic regimens (2-5). This disease, therefore, providesa unique opportunity to study FUra sensitivity and resistance.

In the last several years, the addition of modulating agentsto FUra have improved the outcome of treatment of coloncancer [both in the advanced stage (4) and in the adjuvantsetting (5)]. However, an objective clinical response is obtainedin <40% of patients with advanced disease, and these responsesare inevitably followed by the development of resistance (4).The outcome of FUra therapy may be improved if the biochemical (and molecular) basis for FUra resistance in tumors isidentified and strategies are designed to overcome the development of resistance. The mechanisms of resistance to FUraoperating in the clinical setting may be complex. FUra metabolism may involve different activation pathways (orotate phos-phoribosyltransferase, thymidine phosphorylase and thymidinekinase, uridine phosphorylase and uridine kinase) and at least3 cellular targets [thymidylate synthetase, RNA, DNA (1)].Indeed, defective drug uptake and alterations in every step ofFUra metabolism as well as mutations or increased levels ofthe target enzyme TS have been described in FUra-resistantexperimental tumors (1,6-15).

Both experimental and clinical data suggest that pulse FUraand continuous infusion FUra might have different mechanismsof cytotoxicity. In the clinic, the maximum tolerated dose anddose-limiting toxicity of FUra depend upon the schedule ofadministration (16). Moreover, the slopes of the dose-responsecurve to FUra, both in experimental systems (17, 18) and inthe clinic (19), are different depending on the schedule of drugadministration. Also, from a theoretical point of view, the S-phase specificity of one mode of FUra cytotoxicity, i.e., TSinhibition, compared to the less tight dependence on cell cyclephase of the other mechanism of FUra action, i.e., FUra incorporation into RNA, is consistent with different mechanisms ofaction (and resistance), depending on the duration of exposureto this drug. In general, this complexity has not been taken intoaccount in the previous studies of FUra resistance. Most investigators have studied resistant clones developed either by increasing the concentration of drug stepwise or by a singleexposure to high doses with or without previous exposure tochemical mutagens. Such approaches usually lead to development of highly resistant clones but involve many disadvantages.The multistep procedure often produces unstable phenotypes,thus requiring the continuous presence of the drug for maintenance of resistance. Moreover, this procedure, as well as theone-step selection method, does not resemble the clinical situation in that patients receive repeated courses of the same doseof chemotherapy and resistance would be expected to be lowgrade (20). Therefore, the mechanisms of resistance operatingunder conditions of low (constant) selection pressure (as in the

1855

Research. on February 16, 2019. © 1992 American Association for Cancercancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

SCHEDULE-DEPENDENT MECHANISM OF RESISTANCE TO FtJra

clinic) may be different from the mechanisms of resistancedescribed previously, and the relevance of some of these observations to the clinical use of FUra is probably limited. Also,while LV is now widely used in combination with FUra for thetreatment of advanced colorectal cancer (4, 21, 22), only recently has its intracellular metabolism and the intracellularpools of its main (to FUra therapy) derivative, CH2FH4, beencorrelated with sensitivity or resistance to fluoropyrimidines(23). Moreover, despite the current controversies about theoptimal schedule of FUra administration (1), no attempts havebeen made to investigate the correlation between the scheduleof FUra administration and the mechanism(s) of resistance.

In this paper, two different sublines resistant to FUra aredescribed. Both were selected from a clone of wild-type HCT-8 cells by repeated exposures to a constant concentration ofFUra either administered for 4 h (high dose) or for 7 days (lowdose) to mimic the dose schedules used in the clinic (pulseversus continuous infusion). The two resistant cell lines obtaineddisplayed different mechanisms of FUra resistance. The pulseFUra selection led to a decreased incorporation of this analogueinto RNA, while continuous exposure to the fluoropyrimidineresulted in a resistant subline that was found to recover fromTS inhibition more rapidly than the parental line. In both celllines, new mechanisms of FUra resistance were found: impairedpolyglutamylation of the thymidylate synthase cofactorCH2FH4 in the 7-day resistant subline (HCT-8/7dR) and decreased incorporation of FUra into RNA in the high-dose pulseresistant subline (HCT8/4hR).

MATERIALS AND METHODS

Chemicals. FUra and FdUrd were purchased from Sigma ChemicalCo. (St. Louis, MO). MTX (dissolved in saline) and LV were obtainedin clinical form from Lederle (Pearl River, NY). 2'-[5-'H]deoxyuridine,2'-[5-'H]deoxyuridylate, [6-'H]FdUMP, [6-3H]FUra, [6-'4C]FUra, [6-l4C]5-fluorouridine, [6-'4C]FdUrd, and [mif/i>'/--1H]thymidinewere pur

chased from Moravek Biochemicals (Brea, CA). Lactobacillus caseithymidylate synthase (activity, 53.4 ^mol/h/ml; specific activity, 1.4^mol/h/mg) was purchased from Biopure (Boston, MA). Charcoal(activated, neutralized) was supplied by Sigma. Glycine was obtainedfrom Eastman Kodak Co. (Rochester, NY) and Tris (ultrapure) wasfrom Boehringer Mannheim Biochemicals (Indianapolis, IN). Poly-acrylamide and all other reagents for gel electrophoresis were purchasedfrom Bio-Rad Laboratories (Richmond, CA). Pteroyl polyglutamatestandards with 2-7 residues were obtained from Dr. B. Schirks Laboratories (JoÃ±a,Switzerland) and were reduced to the correspondingtetrahydrofolate forms with human recombinant dihydrofolate reducÃaseand NADPH (24). En-'Hance was from New England Nuclear(Boston, MA) and Kodak X-O-mat AR film was purchased fromEastman Kodak. Media and sera for cell culture were obtained fromGrand Island Biological Co.(Grand Island, NY), and plasticware wasobtained from Corning Glass Works (Corning, NY).

Development of Resistant Cell Lines. The human colon adenocarci-noma cell line HCT-8 (25) was grown as a monolayer as previouslydescribed (20). Under these conditions, the doubling time was 18 h andthe cloning efficiency was about 30%. Resistant cell lines were selectedwith a low selection pressure technique as previously described forantifolates (20). Starting from a clone of the parental cell line, weobtained resistance to FUra with either 4-h or 7-day exposure to FUra,repeated 6 times using the same drug concentration from the beginningto the end of the selection period. Cells were passaged (2 x IO5 cells/

flask) and rechallenged with drug as soon as the cultures approached100% of confluence. The initial selecting concentration was the highestthat allowed regrowth of at least a few clones, starting from 2x10"

cells/25-cnr flask (approximately 4 logs of cell kill). A progressivereduction in the time (days) at which confluence of the cultures at each

successive passage was reached allowed the development of resistanceto be monitored. The resulting FUra-resistant sublines were designatedHCT-8/FU4hR, for the one obtained with 4-h exposure to FUra, andHCT-8/FU7dR, for the one obtained with 7-day exposure to FUra,and were maintained as described for the parental cell line. Theseexperiments were repeated 3 times with 4 replicates for each point.

Cytotoxicity Assays. In order to quantitate the degree of resistanceto FUra and cross-resistance of the cell lines to other drugs, a monolayerclonogenic assay was used (26). Cultures were trypsinized for 3 min,and an essentially monocellular suspension was obtained by passing thetrypsinized cells through a 25-gauge needle. The cloning efficiency ofthe resistant sublines was similar to that of the parental sensitive line;thus, 5-10 x 10! cells in 5 ml of medium containing 10% horse serumwere dispensed into sterile 60-mm Petri dishes and incubated at 37"C

and 100% humidity with 5% CO2. Eighteen h later, when the cells wereattached to the bottom of the Petri dish but had not yet divided, 0.1 mlof an appropriate dilution of drug in a 0.9% NaCl solution was addedto each dish. Control dishes received the same volume of saline. Afterincubation as indicated, the medium was decanted, the cells werewashed twice with 5 ml of saline, and 5 ml of fresh medium was added.Clonal growth was determined after staining with orcein. Coloniescontaining >200 cells were scored at xlO magnification using a dissecting microscope. Under these growth conditions, >95% of the colonies from both sensitive and resistant cells contained >200 cells. Thesensitivity of both parental and resistant cells to the antineoplasticagents was determined after 4-h and 7-day exposures to the drugs. Eachexperimental point was determined in triplicate with 4 replicate controls; experiments were repeated at least twice.

FUra Uptake. Exponentially growing cells (IO5 cells/well) were cultured 18-24 h before carrying out the assay in 24-welI plates. Todetermine uptake, a modification of a method previously described forMTX was used (27). Briefly, old medium was discarded, and 2 ml ofprewarmed (37Â°C)complete tissue culture medium containing 100 >Â»M[6-3H]FUra (specific activity, 25 cpm/pmol) was added. The plates wereincubated at 37Â°Cand, at appropriate time points (shown in Fig. 2),triplicate plates were cooled to 0Â°Cand washed three times with 3 ml

of ice-cold PBS. The cells were removed by incubation with 1 ml of 1N NaOH at 37"C for 45 min. A 0.5-ml aliquot was neutralized with an

equal volume of l N HC1, and after 15 ml of Ecolume scintillation fluidwas added, the content of radioactivity was measured in a Beckmanmodel 5801 scintillation counter. Triplicate cultures were collected by10 min trypsinization for all cell counts. All experiments were repeatedat least twice.

Assay of Enzymes of Pyrimidme Metabolism. Enzyme activities wereassayed as previously described using cytosols from logarithmicallygrowing cells (26). Assays were performed at 2 enzyme concentrationsand several time points to ensure that rates were derived from <25%substrate conversion, and the rate was linear for at least 2 time points(10 and 20 min). Activity is expressed as nmol of product formed/mgof protein/h. Thymidine kinase activity was also measured using amodification of the method of Taylor et al. (28). A reaction mixtureconsisting of 5 UMMgCl2, 1.0 mM ATP, 2.0 MM[methyl-3li]tÃ¬iymiame(82.40 Ci/mmol), and 0.2 M Tris-HCl (pH 7.6) in a final volume of100 fi\ was incubated with enzyme at 37Â°Cfor 15 min. At the end of

this period, 50 Â«Ifrom each tube was spotted on DE81 filter discs (24mm; Whatman), washed with 1 mM ammonium formate (4 changes),and rinsed with ethanol before the filters were dried in a microwaveoven at high power for 2 min. Protein concentration was estimated bythe Bio-Rad method (29).

TS Catalytic Assay. Cytosols for TS activity measurements andkinetics studies were prepared according to a different procedure, whichproduces a higher TS yield and enzyme stability. Exponentially growingcells in 225-cnr tissue culture flasks were cooled on ice for 5 min,washed twice with 20 ml of ice-cold PBS, detached by 5- to 10-mintrypsinization, and followed by addition of 15 ml of complete tissueculture medium. Keeping the cells on ice and using ice-cold reagentswas essential for obtaining a higher preservation of enzyme activity.After centrifugation (5 min, 1000 rpm, 4Â°C)the resulting pellet waseither stored at â€”¿�70Â°Cor resuspended in 1-2 ml of extraction buffer

[50 mM Tris-HCl, 50 mM /S-mercaptoethanol (pH 7.4) containing 10%

1856



SCHEDULE-DEPENDENT MECHANISM OF RESISTANCE TO FUra

glycerol and 0.1% Triton X-100, 4 ^1/ml pepstatin, and 1 ii\/m\leupeptin] for further processing.

No detectable loss of TS activity resulted from storage at â€”¿�70Â°Cfor

up to 1 month. Cell disruption was obtained with six 20-s bursts using

a Vibraceli sonicator (Sonics and Materials, Inc., Danbury, CT)equipped with a microtip. No intact cells could be seen by lightmicroscopy. Cell homogenates were then clarified by centrifugation at12,000 x g and 4'C for 30 min, and aliquots of the supernatants were

immediately used as the enzyme source in the following assay. Quan-titation of TS activity was based upon the release of tritium from [5-3H]dUMP as [3H]H2O (30). The assay was performed in a total volumeof 50 ti\ containing 50 MM2'-[5-3H]deoxyuridylate (specific activity, 0.1

Ci/mmol), 250 n\t 5,10-methylenetetrahydrofolate, and enzyme buffered with 50 mM Tris-HCl (pH 7.4), 5 m\i formaldehyde, 50 HIMNaF,10 mM Na ascorbate, and 100 mM /3-mercaptoethanol. After 5 minpreincubation at 37*C, the reaction was started by the addition of 5 or10 n\ of cell extract and carried out at 37Â°Cfor 15 or 30 min before

being stopped by addition of 200 //I of a charcoal suspension (100 mg/ml) in 2% TCA. The samples were vortexed, and the charcoal wasremoved immediately by centrifugation at 10,000 x g for 10 min. AI(10/(Ialiquot of the supernatant was then assayed for 'I I â€¢¿�<)radioactiv

ity by counting in a Beckman model 5801 liquid scintillation spectrometer after adding 5 ml of Ecolume scintillation cocktail. All assays wereperformed 3 times, and each point in the assay was done in triplicate.Enzyme activity is expressed in nmol of tritium released/h/mg ofprotein. All enzyme assays were linear with respect to time and enzymeconcentrations at the substrate concentrations used. In order to measureTS activity in samples of cells previously exposed to FUra, it wasnecessary to remove any FdUMP bound to TS prior to incubation with[5-3H]dUMP (31). This was accomplished by incubating 50-^1 aliquots

of the 12,000 x g supernatants with 50 n\ of dissociation buffercontaining 0.6 M ammonium bicarbonate (pH 8.0), 80 Â¿AIdUMP, 100mM /3-mercaptoethanol, 100 mM NaF, and 15 mM CMP for 4 h at37"C. At the end of the dissociation procedure, dUMP, free FdUMP,

and cellular reduced folates were removed by adding 100 n\ of an ice-cold 10% charcoal suspension, followed by centrifugation at 15,000 xg for 20 min. Twenty-Mi aliquots of the supernatants were then used asthe enzyme source in the assay described above for total TS activitydetermination. Another set of reactions was carried out omitting theternary complex dissociation by immediately adding 50 >i\of dissociation buffer and 100 M!of the 10% charcoal suspension to the 12,000 xÂ£extract. This allowed quantitation of TS activity remaining after FUraexposure (i.e., free TS).

Enzyme Kinetics and Inhibition Studies. Cytosols for kinetic studieswere prepared as described for the TS catalytic assay, without anyfurther purification. Some experiments were performed on desaltedcytosols, but results were similar to those obtained when crude extractswere used as the enzyme source. Apparent Kmvalues for both dUMPand CH2FH4 were determined by varying the concentration of eitherdUMP or CH2FH4 between 1 and 150 MMor 5 and 500 MM,respectively.After correction was made for blank rates (i.e., no enzyme), the datawere Tit by nonlinear least squares regression to Michaelis-Mentenequation by using the Enzyme Fitter program (Elsevier). IC50 valueswere obtained by inhibition assays in which FdUMP was added atvarious concentrations to the standard reaction mixture. In order tocompare the K -,,values of this tight-binding inhibitor between resistantand sensitive cell lines, care was taken to use the same amount ofenzyme activity for each assay. The drug concentration giving 50%inhibition of activity was obtained from plots of the percentage ofremaining activity versus the logarithm of FdUMP concentration.

FdUMP-binding Assay. Cytosols were prepared as described for thecatalytic assay with the exception that the extraction buffer was 200mM Tris-HCl (pH 7.5), 70 mM /3-mercaptoethanol, 100 mM NaF, and15 mM CMP. FdUMP-binding sites were measured as described byFernandes and Cranford (31).

In Situ TS Activity. Inhibition of thymidylate synthase activity byFUra in intact cells was measured by the in situ assay of Yalowich andKaiman (32) as modified by Rodenhuis et al. (33) and Pizzorno et al.(34). The assay is based upon the release of tritium from 2'-[5-3H]

deoxyuridine, because it is methylated by thymidylate synthase after

cellular uptake and phosphorylation to deoxuridylate. The cells werecultured in 24-well plates 24 h before carrying out the assay, at a densityof 50 x IO3cells/ml in complete tissue culture medium. Duplicate wells

for each condition received 100 M!of an appropriate dilution of FUra.Controls received the same volume of PBS. After a 4-h period ofincubation with drug, 1 ml of complete medium containing 2'-[5-'H]deoxyuridine was added to achieve a final concentration of 1 >i('i ml,

and 100-Mlsamples were taken at 0, 30, 60, 90, and 120 min. Thesewere added to 200 n\ of a 10% activated charcoal suspension in 2%TCA, vortexed, and centrifuged for 5 min at 10,000 x g. One hundredM!of the supernatant was added to 5 ml of Ecolume scintillation cocktailand counted in a Beckman model 5801 liquid scintillation counter. Inrecovery experiments, at the end of the 4-h drug exposure, cells werewashed twice with PBS, resuspended in drug-free medium, and incubated for another 8, 24, or 48 h. After this time period, mediumcontaining 2'-[5-'H]deoxyuridine was added, and the assay was carried

out as described for inhibition studies. When effects of LV on stabilityof TS inhibition were investigated, cells were plated in complete medium containing 100 Â»MLV; cultures were provided with the folatecofactor, at this concentration, also during exposure to FUra as well asduring reincubation in drug-free medium following FUra removal. Afterbackground subtraction (/.<â€¢..medium without cells), the counts were

fitted to a straight line by linear regression, and the percentage ofinhibition was calculated by comparing the slope of treated cultureswith that of controls (without drug). Slopes in all assays fitted the linearregression model with a correlation coefficient of at least 0.9.

Assay of Intracellular Folates. A modification of the radioenzymaticbinding assay described by Bunni et al. (35) was used for intracellular</H I 11, and 111, quantitation. This method is based upon the entrapment of endogenous reduced folates, after conversion to CH2FH4, byTS and ['HJFdUMP to form a stable ternary complex (36). The amountof [3H]FdUMP bound to the protein is a function of the available levelsof CH2FH4. Cells were washed twice with ice-cold PBS, detached (5-min trypsinization at 37"C), and centrifuged (5 min, 1000 x g) after

addition of complete tissue culture medium. The resulting pellet wasresuspended in extraction buffer that contained 50 mM Tris-HCl, 50mM sodium ascorbate, 1 mM EDTA, and 0.25 M sucrose (pH 7.4) at adensity of 1.5 x IO7 cells/ml. Cells were boiled for 3 min to achieve

lysis and to prevent enzymatic cycling during the assay. After centrifugation (10,000 x g, 10 min, 4Â°C),aliquots of the supernatant wereincubated with 160 munits L casei TS and 125 nM ['H]FdUMP in a

total volume of 100 M!of the above extraction buffer. The reaction wasallowed to proceed at 37'C for 30 min and stopped by cooling samples

on ice. Uncomplexed radioactivity was removed by addition of a 500-M!slurry of ice-cold 3% acid charcoal (albumin and dextran treated)and centrifugation at 12,000 x g for 20 min. The amount of TS-boundradioactivity was then quantitated in aliquots of the supernatant afteraddition of 10 ml scintillation fluid. A series of standard tubes withknown amounts of ( 11 I 11; was included in each assay. All values arecorrected for binary complex (i.e., TS and ['HjFdUMP without folate

cofactor) radioactivity. The plot of the fraction of bound radioactivityat each concentration of standard CH2FH4 provided a calibration curvefrom which molar levels of cofactor in the test samples were calculated.FH4 levels were quantitated by the same method after conversion toCH2FH4 by the addition of 6.5 mM formaldehyde. The recovery ofknown amounts of CH2FH4 added to cell extracts prior to boiling was94%.

Analysis of the ( 11 111, and FH4 Glutamate Chain Length Distribution. Determination of the predominant glutamate chain length ofCH2FH4 and FH4 was based upon the electrophoretic separation ofternary complexes formed by L. casei TS and |'H]FdUMP reacting

with intracellular CH2FH4 . Folate extraction and ternary complexformation were performed according to the method of Priest and Doig(37). Separation of ternary complexes with different chain lengths ofthe folate cofactor was then obtained by polyacrylamide gel electropho-resis. Samples were electrophoresed for 18 h at 4Â°C,8 mA, on a 0.75-

cm thick polyacrylamide gel (9% in the resolving gel; 4.5% in thestacking gel) using 0.025 M Tris-HCl and 0.192 M glycine (pH 8.5) asthe running buffer. The gels were then sequentially soaked in 5% TCA(30 min, 4Â°C),Enhance (30 min, room temperature), and distilled water

1857




(60 min room temperature) in order to obtain protein Fixation, impregnation with the fluorescent intermediate, and its precipitation in thegel (38). After the gels dried (90 min, 60Â°C),autoradiographs wereobtained by exposure of Kodak X-AR film at -70Â°Cfor 48 h.

Incorporation of FUra into RNA. Cells were cultured 18-24 h in 150-mm dishes (approximately IO7 cells/dish). Old medium was replaced

with complete medium containing a range of concentrations (10-300MM)of FUra labeled in position 6 with 3H (specific activity, 2-20 pdf

i/niol ). At the end of the desired incubation time (4 or 24 h), cell cultureswere washed thrice with PBS and either incubated for 24 h in drug-freemedium or immediately harvested in 1 ml of denaturing solution [4 Mguanidinium thiocyanate, 25 imi sodium citrate (pH 7.0), 0.5% sar-kosyl, and 0.1 M /3-mercaptoethanol). RNA extraction (39) was performed either immediately after cell collection or after storage at -70Â°C

for up to 10 days without any effect on RNA yield. Aliquots of theextracted RNA were placed in scintillation vials and, after addition of5 ml of Ecolume scintillation cocktail, the incorporated radioactivitywas counted in a Beckman model 5801 spectrometer. All results werenormalized for RNA content as determined by measuring UV absorb-ance at 260 nm. Parallel 280 nm readings with determinations of 260/280 ratios excluded any protein contamination.

RESULTS

Development of Resistance

The concentrations of FUra used in the selection of resistancewere 1000 MM(4-h exposure) and 15 MM(7-day exposure), i.e.,approximately 3- and 7-fold the ED50 values measured byinhibition of HCT-8 colony growth. These experimental conditions represent the highest drug concentrations that allowedminimal clonogenic survival (5-20 clones), thus reflecting approximately 4 logs of cell kill. The time to reach confluenceafter the first drug treatment illustrates this high degree of cellkill: 33 and 35 days (Fig. 1). In general, no surviving colonieswere observed when slightly higher concentrations of FUrawere used (2 mM for 4-h exposure and 20 MMfor 7-day exposure). Because 2 x 10s cells were exposed to drug, these findings

confirm what we previously reported in this cell line usingMTX and trimetrexate, that it is possible to eradicate 5 logs ofcells in culture with only one treatment cycle (20). At the endof the selection process (6 cycles), a low degree of resistancewas obtained in the HCT-8/FU4hR cells (ED50 for a 4-hexposure, 100 Â±4 versus 30 Â±2MM,mean Â±SE, in the sensitiveparental cell line). In contrast, a higher level of resistance wasachieved in HCT-8/FU7dR cells with the 7-day exposure (ED50,20 Â±1.0 versus 2.0 Â±0.4 MMin the sensitive HCT-8 cells).Also, HCT-8/FU7dR cells were cross-resistant to FdUrd (ED50for a 7-day exposure, 46.0 Â±2.5 versus 1.6 Â±0.3 HM) and,surprisingly, MTX (ED50 for 7-day exposure, 55.0 Â±7.5 versus16.0 Â±3.7 HM), while HCT-8/FU4hR cells showed a normalsensitivity to both the nucleoside and the antifolate. Moreimportantly, HCT-8/FU4hR cells displayed full sensitivity toFUra when challenged with a 7-day exposure of this drug (ED50,2.2 Â±0.5 versus 2.0 Â±0.4 MM).

FUra Uptake

Transport of FUra into sensitive and resistant HCT-8 cells,at an extracellular concentration of 100 MM,was characterizedby an initially rapid uptake followed by a continued slowaccumulation before reaching a plateau at 15 min. In the initialphase of rapid influx, the slopes were similar for both sensitiveand resistant HCT-8 cells. Also, the intracellular concentrationsreached in the plateau phase were approximately equivalent forthe 3 cell lines (42.15 Â±0.65, 39.95 Â±0.96, and 44.81 Â±1.37

(0a-T3

40n

30-

20-

10-

O)

a o

0>

40-1

Â«* SOCI)

F20-

10-

15pM

0 ! 2 3 4 5 6

Treatment cycleFig. 1. Development of resistance to FUra: 4 h (A) or 7 day (B) exposure to a

fixed concentration of drug. HCT-8 cells (2 x 10*)were repetitively exposed to 1

mM FUra for 4 h (A) or 15 JIMFUra for 7 days (B). Shortening of the time toreach confluence at each cycle reflects development of resistance. Point, averageof at least 3 experiments done with 4 duplicates; bar, SE. NO TX, no treatment.

pmol/105 cells in HCT-8, HCT-8/FU7dR, and HCT-8/FU4hR

cells, respectively).

Activities of FUra-metabolizing Enzymes

The activities of thymidine kinase, uridine kinase, thymidinephosphor\ lase, uridine phosphorylase, and orotate phosphori-bosyl transferase, measured using saturating concentrations ofsubstrates (50 MM),were similar in the three cell lines (Table1). Since HCT-8/FU7dR cells were cross-resistant to fluoro-deoxyuridine, thymidine kinase, an enzyme often involved indetermining resistance to this nucleoside, was studied in moredetail. The activity of this enzyme, measured with either thymidine or fluorodeoxyuridine as substrates, was similar in thethree cell lines (Table 1). Also, saturation curves obtained withincreasing concentrations of thymidine were identical, resultingin similar Kmvalues (6.5 Â±1.1, 5.3 Â±2.3, and 5.5 Â±1.1 MMin

Table 1 Activity of enzymes ofpyrimidine metabolismSensitive and resistant cells in the exponential phase of growth were collected

and cytosolic extracts prepared as enzyme source. Reaction conditions are described in "Materials and Methods."

Activity (nmol/mgprotein/h)"Thymidine

kinase*Thymidine kinaser

Uridine kinaseOrotate phosphoribosyl

transferaseThymidine phosphorylaseUridine phosphorylaseHCT-83.3

Â±1.96.6 Â±2.7

31.9 + 9.810.5 Â±3.74.8

Â±0.23.8 Â±2.5HCT-8/FU7dR3.9

Â±1.05.2 Â±1.7

28.6 Â±9.210.1Â±0.83.0

Â±0.44.2 Â±2.6HCT-8/FU4hR3.1

Â±2.45.1 Â±1.925.7

Â±9.312.4Â±2.05.1

Â±2.84.9 Â±2.8

" Mean Â±SD of at least 2 experiments, each done in duplicate.* 5-Fluoro-2'-deoxyuridine was used as substrate.' Thymidine was used as substrate.

1858




HCT-8, HCT-8/FU7dR, and HCT-8/FU4HR cells, respectively).

TS Activity

Two different assays were used to evaluate the catalyticactivity and the level of expression of this enzyme. Neither thecatalytic nor the FdUMP-binding assay demonstrated significant differences in baseline TS levels between the three celllines (Table 2).

Kinetics of Thymidylate Synthase Inhibition by FUra

Although no changes in basal activity of TS and FUra-metabolizing enzymes were detected, alterations in either theactual degree of TS inhibition achieved after FUra treatmentor the stability of this inhibition over time or an increase(induction of this enzyme activity) could still be responsible forthe FUra resistance of these cell lines. These possibilities wereexamined as follows.

Determination of Free and Bound TS. Cells were exposed toFUra for 4 h (30 or 100 MM)and either analyzed immediatelyor resuspended in drug-free medium for 24 or 48 h. TS activityin cell extracts was assayed with or without previous dissociation of the ternary complexes in order to quantitate free, bound,and total TS. After a 4-h exposure to either 30 or 100 MMFUra,free TS levels were similar in sensitive and resistant HCT-8cells (approximately 25 and 18% of total intracellular TS, with30 and 100 MM,respectively) (Table 3). However, after FUraremoval, a marked increase in free TS was observed in HCT-8/FU7dR cells resulting in an almost complete release of theinhibitor FdUMP after 24 h (30 MMFUra) or 48 h (100 MMFUra) (Table 3). No recovery was observed in HCT-8 andHCT-8/FU4hR cells, and at least 65% of the total intracellularTS was still bound by FdUMP 48 h after the end of FUraexposure (Table 3).

In Situ TS Activity. When the in situ assay for thymidylatesynthesis was used to follow the time course of TS inhibition

Table 2 Thymidylate synthase levels in sensitive and resistant HCT-8 cells

Logarithmically growing cells were collected and cytosolic extracts preparedas described in "Materials and Methods." Results shown are means of 3-4

experiments, each done in duplicate Â±SE.

CelllineHCT-8

HCT-8/FU7dRHCT-8/FU4hRFdUMP-binding

assay(pmol/mgprotein)1.87

Â±0.531.17 Â±0.031.27 Â±0.25Catalytic

assay(nmol/mgprotein/h)24.4

Â±6.9017.2 Â±3.0015.5 Â±5.80

Table 3 Free FdUMP-binding sites in sensitive and resistant HCT-8 cells afterFUra treatment

Cells were analyzed both immediately after exposure to either 30 or 100 >AIFUra for 4 h (0 h) and after an additional incubation in drug-free medium (24and 48 h). Data shown are the percentages of free TS compared to total TS valuesobtained from the same samples after ternary complex dissociation, as describedin "Materials and Methods." Results represent the means of 3 experiments, each

done in duplicate Â±SE.

CelllineHCT-8HCT-8/FU7dRHCT-8/FU4hRFUraO.M)30

10030

10030

100Oh19.0

Â±4.525.7 Â±3.226.3

Â±6.127.4 Â±3.226.2

Â±4.134.7Â±7.2%

free TSat:24

h37.0

Â±4.032.3 Â±2.3143.5

Â±15.040.5 Â±3.536.0

Â±4.048.6 Â±2.348

h27.0

Â±3.035.6 Â±4.390.5

Â±7.170.8 Â±4.644.0

Â±5.537.0 Â±6.1

100!

W 40

^ 20

100-

Â¿-80'l 60nCO 40

5s 20

100-

Â¿-80'I 60-toco 40-

35 20

B

10 30

[FUra], pM

100

Fig. 2. Inhibition of in situ TS activity by FUra (4-h exposure; 10, 30, and 100MM)in sensitive and resistant HCT-8 cells. TS activity was assayed either immediately after drug exposure (.-I) or after an additional 24- (B) or 48-h (C)incubation in drug-free medium. M, HCT-8 cells; D, HCT-8/FU4hR cells; G,HCT-8/FU7dR cells. Column, mean percentage of tritium released as comparedto untreated controls in 3-4 experiments, each in duplicate; bar, SE.

after FUra exposure in intact cells, only slight differences inthe degree of thymidylate synthesis inhibition between sensitiveand resistant cells were shown at the end of a 4-h exposure to

either 10 or 30 MM FUra (Fig. 2A). However, at 100 MM,complete inhibition (>95%) of thymidylate synthesis wasachieved only in HCT-8 and HCT-8/FU4hR cells, while HCT-8/FU7dR cells still showed at least 10% of activity. A markedrecovery of thymidylate synthesis activity occurred in HCT-8/FU7dR cells when FUra was removed and the cells were incubated in drug-free medium for 24 and 48 h (Fig. 2, B and C).In particular, at 48 h both HCT-8 and HCT-8/FU4hR cellspreviously exposed to 100 MMFUra still showed >90% inhibition of thymidylate synthesis activity, while under the sameconditions thymidylate synthesis in HCT-8/FU7dR cells recovered to 80% of untreated controls (Fig. 2C). Exposure to100 MMLV prior to, during, and after exposure to the fluoro-pyrimidine enhanced the degree of thymidylate synthesis inhibition achieved with a 4-h exposure to FUra (10-100 MM)inparental sensitive and pulse FUra-resistant cells, but this enhancement was not observed in HCT-8/FU7dR cells (Fig. 3/1).While no differences were observed for the stability of thymidylate synthesis inhibition after drug removal in HCT-8 andHCT-8/FU4hR cells, the addition of LV resulted in a lessprofound recovery in HCT-8/FU7dR cells (Fig. 3, B and C).However, in this cell line, residual thymidylate synthesis activityafter FUra removal was still significantly higher than in bothsensitive and pulse FUra-resistant HCT-8 cells (26.0 Â±5.0%versus 3.0 Â±2.0% versus 3.5 Â±1.5%, at 100 MM48 h after drugremoval).

1859




100

f-(oW 40

55 20

100-,

â€¢¿�f80

C/5 40

^ 20

100-,

B

-i-i

'â€¢g60to

W40-^

20EÃ•TX 110

30100[FUra],

LJM

Fig. 3. Effects of 100 MMLV on in situ TS activity inhibition by FUra (4-hexposure; 10, 30. and 100 MM)in sensitive and resistant HCT-8 cells. TS activitywas assayed either immediately after FUra exposure (A) or after an additional24- (II) or 48-h (C) incubation in drug-free medium containing 100 MMLV. D.HCT-8 cells; D, HCT-8/FU4hR cells; D, HCT-8/FU7dR cells. Column, meanpercentage of tritium released as compared to untreated controls of 3-4 experiments, each in duplicate; bar, SE.

Table 4 Kinetic parameters ofthymidylate synthase from sensitive and resistantHCT-8 cells

Reaction conditions and analysis are described in "Materials and Methods."

1CÂ«,FdUMP (HM)Km CHjFH/ (MM)Km dUMP* (MM)HCT-890.5

Â±25.449.6 Â±6.3

8.2 Â±0.7FU7dR74.2

Â±13.856.3 Â±12.110.7 Â±1.8FU4HR70.1

Â±9.359.1 Â±10.9

6.4 Â±0.8Â°Apparent Kâ€žvalues at 125 MMdUMP." Apparent Kmvalues at 250 MMCH3FH4.

Intracellular Folate Pools

Baseline (i.e., without previous exposure to LV) levels of FH4were measured in the three cell lines (Table 5). FH4 levels weredecreased by about one third in both resistant cell lines ascompared to the parenteral line (Table 5). CH2FH4 levels werefound to be decreased in HCT-8/FU7dR cells as compared toboth sensitive and pulse FUra-resistant cells (Table 5). Themagnitude of this reduction (approximately 60% decrease) ismuch greater than the intrasample variability due to the partialinstability of CH2FH4 to heat treatment in the absence of excessHCHO (<20%).

Analysis of the Distribution of OII11, and FH4 PolyglutamateChain Length in Sensitive and Resistant HCT-8 Cells

A deficit in the formation of long-chain polyglutamates ofCH2FH4 could result in a less stable ternary complex, leadingto a less prolonged TS inhibition by FUra. This would explainboth the dramatic recovery of TS activity observed in HCT-8/FU7dR cells after FUra removal as well as the limited capabilityof LV to inhibit this recovery. In basal conditions, the parentalcell line mainly contained heptaglutamates of CH2FH4 with asmall amount of hexa- and possibly octaglutamates. In contrast,HCT-8/FU7dR cells contained only short-chain glutamates(Fig. 4). The same pattern was observed after a 4-h exposure to100 firn LV, penta- and hexaglutamates predominated in HCT-8 cells, while no folates with >3 glutamic residuals could bedetected in HCT-8/FU7dR cells (Fig. 4). In HCT-8/FU4hRcells, a significant amount of monoglutamate and diglutamateswere found prior to treatment, and short-chain polyglutamatespredominated after LV exposure. However, in contrast to HCT-8/FU7dR cells, this was only a quantitative difference in thatthis cell line also formed some penta- and hexaglutamates (datanot shown).

Incorporation of FUra into RNA

In parental sensitive HCT-8 cells FUra incorporation intoRNA increased linearly with increasing concentrations of thefluoropyrimidine (Fig. 5/4). The amount of drug incorporatedinto RNA after a 4-h exposure to 300 UM was approximately6-fold greater than that obtained with 30 ^M (Fig. 5A). Incubation of cell cultures in drug-free media for 24 h after the end

Table 5 CHÂ¡FH4and FH4 pool size in sensitive and FUra-resistant HCT-8 cells

Logarithmically growing cells were collected and resuspended at a density of15 x 10" cells/ml in extraction buffer. Folate extraction and measurement wereperformed as described in "Materials and Methods." Data represent the means

Â±SE of 3 experiments, each done in triplicate.

Cell line CH,FH4 FH4(pmol/10' cells)

HCT-8HCT-8/RU7-dRHCT-8/FU4-hR

18.1 Â±0.87.4 Â±1.8

16.1 Â±3.7

32.1 Â±5.519.3 Â±2.924.3 Â±2.7

1 234567

TS Kinetics and Inhibition Studies

Since a mutated enzyme would be expected to be less sensitiveto inhibition by the FUra metabolite FdUMP and/or to resultin a less stable ternary complex (and consequently less prolonged TS inhibition), Kmvalues for both dUMP and CH2FH4were determined. As shown in Table 4, only minimal differencesbetween the 3 cell lines were detected. Furthermore, the IC50values for in vitro inhibition of TS with FdUMP were similarfor the 3 cell lines (Table 4).

Fig. 4. Analysis of folylpolyglutamates extracted from sensitive and FUra-resistant HCT-8 cells by electrophoretic separation as ternary complexes with /-.casei TS and ['HjFdUMP on a polyacrylamide gel in nondenaturing conditions.

Lane I, Glu 1 and Glu 4 standards; lane 2, Glu 3 and Glu 7 standards; lanes 3and 5, parental HCT-8 cells, before and after a 3 h exposure to 100 MMLV,respectively; lanes 4 and 6, HCT-8/FU7dR cells, before and after a 3 h exposureto 100 MMLV, respectively; lane 7, Glu 2 and Glu 5 standards. Polyglutamatechain length increases from the top to the bottom of the gel. Samples wereprepared and gels were processed as described in "Materials and Methods."

1860




15.0

12.5

10.0

7.5

5.0

25

50.0

i 40.0I 10.0"

Ã¬ 7.5oI 5.0

| 2.5LL_

500

7.5

5.0

2.5

B

10 30 100 300

[FUra], \

Fig. 5. Incorporation of FUra into RNA in sensitive and resistant HCT-8 cells.Cell cultures were exposed for either 4 (A and B) or 24 h (C) to differentconcentrations of FUra: A. 30, 100, and 300 UM;B, 30 and 100 Â»IM;C, 10 and 30JJM.Specific activity ranged from 2-20 pCi/nmoi. Cells were analyzed either atthe end of FUra exposure (A and C) or after an additional 24-h incubation indrug-free medium. Â®,HCT-8 cells; G, HCT-8/FU4hR cells; D, HCT-8/FU7dRcells. Column, average of two different experiments; bar, Â±SE.

of drug exposure (Fig. 5B) led to increased incorporation ofFUra into RNA at both 30 MM(3-fold) and 100 MM(7-fold)concentrations. Increasing the duration of drug exposure with30 MMFUra up to 24 h did not result in a higher incorporationof FUra into RNA (Fig. 1C). The same relationship betweenthe concentration of FUra and its incorporation into RNA wasobserved in the two resistant cell lines. However, in HCT-8/FU4hR cells, the amount of FUra incorporated into RNA aftera 4-h exposure to this drug was significantly lower when compared to parental sensitive or HCT-8/FU7dR cells, independently of the concentration of FUra used (30-300 MM)(Fig. 5/1).Incubation of cells in drug-free medium after FUra removalalso resulted in an increased drug accumulation into RNA inthis cell line. However, the total amount of FUra incorporatedinto RNA was 3-fold lower than both the parental sensitive andthe continuous infusion resistant cell line (Fig. 5.8).

DISCUSSION

A large number of studies of the mechanisms of FUra resistance have been conducted in the past three decades (1). Themain limitation to the clinical relevance of most of these studieshas been the procedures used to select resistant cell lines. Themethods used previously were aimed at obtaining highly resistant clones and usually involved a high selection pressure withincreasing drug concentration. The resistant cell lines weredeveloped in the present study to investigate the mechanisms

of resistance to FUra operating under experimental conditionsthat closely mimic the clinical situation. The concentration ofFUra used for the selection of resistance was kept constant forthe six repeated treatment cycles, resulting in rapid development of low-levelresistance, analogous to the clinical situation(20).This model also allows study of the possible relationship(s)between the development of resistance and the schedule ofFUra administration. This issue is of crucial importance in viewof the uncertainties regarding the optimal method of deliveringfluoropyrimidines (1). The hypothesis that different schedulesof FUra administration may lead to different mechanisms ofresistance was mainly based on clinical data that pulse andcontinuous infusion FUra have different dose-limiting toxicitiesand maximum tolerated doses (16). Evidence presented in thisstudy that supports this hypothesis derives from cytotoxicityassays that showed that cells resistant to pulse FUra stillretained full sensitivity to the fluoropyrimidine given as a 7-dayexposure, as well as a different pattern of cross-resistance toother drugs (FdUrd, MTX) for the two resistant cell lines. Ofinterest is that the cell line resistant to 7-day FUra was cross-resistant to 4-h FUra, possibly indicating that both TS inhibition and incorporation of FUra into RNA are required forcytotoxicity (data not shown). TS inhibition has long beenthought to be the primary mechanism of FUra cytotoxicity (40),and either increased levels or mutations of this enzyme havebeen the explanation for fluoropyrimidine resistance in themajority of previous studies (1). Amplification of the TS generesulting in higher enzyme activity is the only mechanism ofresistance to fluoropyrimidine detected in human samples untilnow (41). In the present study, both FdUMP-binding sites andTS catalytic activity were found to be lower in the two resistantcell lines as compared to the parenteral cell line. However, thisslight decrease is probably not significant with regard to themechanism of resistance since one would expect to find anelevation, rather than a decrease, of this enzyme in resistantcells.

Although no changes in basal activity of TS and FUra-activating enzymes weredetected, alterations in either the actualdegree of TS inhibition achieved after FUra exposure or thestability of this inhibition over time after FUra removal couldstill be responsible for the FUra resistance of these two celllines. After a 4-h exposure to FUra, TS activity was inhibitedby 75-80% in the 3 cell lines, regardless of the drug concentration used (either 30 or 100 MM)(Table 3) and despite the acuteinduction of this enzyme observed in the 3 cell lines (approximately 5-fold increase in total intracellular TS as compared topretreatment values, data not shown). In the parental cell line,a 4-h exposure to 100 MMFUra produces a >95% cell kill; thus,the residual 20% free TS at this concentration (Table 3) was asurprising finding possibly related to TS induction followingFUra treatment, suggesting that other mechanisms, in additionto TS inhibition, are involved in FUra cytotoxicity upon exposure to high doses for a short time.

Prolongation of TS inhibition after drug removal is an important determinant of fluoropyrimidine cytotoxicity (42). TSlevels (43) and dUMP pools (44) have been shown to increaseafter FUra treatment, and either or both increases may lead toa less prolonged inhibition of thymidylate synthesis. A dramaticincrease in free TS was observed in HCT-8/FU7dR but not inparental sensitive and pulse FUra-resistant cells when cultureswere incubated in drug-free medium after FUra exposure (Table3). However, this assay (31), although widely used, involves ahigh degree of manipulation of cell extracts, and different

1861




stabilities of TS from different cell lines could affect the results.Moreover, incorporation of TS into ternary complexes couldresult in a lower rate of degradation of this enzyme as comparedto the enzyme not bound, leading to an overestimation of totalTS as compared to free TS (43). The use of an in situ assay forTS activity overcomes these limitations and also allows a studyof the kinetics of TS inhibition after FUra treatment in intactcells. This assay showed a higher sensitivity for detecting TSinhibition, making it possible to obtain a dose-response curveafter a 4-h exposure to concentrations of FUra ranging from10-100 fi\i. The dose-response curves for TS inhibition weresimilar for the 3 cell lines (Fig. 14). However, in the HCT-8/FU7dR subline when FUra was withdrawn and cells were keptin drug-free medium, thymidylate synthesis recovered, approaching 80% (as compared to untreated controls) at 24 and48 h (Fig. 2). These results could be due to either an altered TSwith reduced affinity for FdUMP or a reduced availability ofCH2FH4 polyglutamates required for stable ternary complexformation (45,46). The former possibility was unlikely becausethe degree of inhibition of TS activity achieved after a 4-hexposure to FUra was similar in this cell line and in the parentalcells (Fig. 2A). Furthermore, kinetic studies of TS showed thatits affinity for both the substrates dUMP and CH2FH4 and theinhibitor FdUMP is approximately the same in the 3 cell lines,excluding an alteration of this enzyme as a possible reason forthe less persistent TS inhibition observed in HCT-8/FU7dRcells. A further possible explanation for the recovery of TSactivity observed in HCT-8/FU7dR cells was an increasedsynthesis of new TS protein. This possibility was excluded inthat an induction of this enzyme did occur after FUra treatment,but the magnitude of this increase was similar in sensitive andFU7dR-resistant HCT-8 cells Moreover, this phenomenon wasobserved at the end of FUra exposure (4 h), but 24 h (and 48h) after drug removal total intracellular TS was similar intreated and untreated cells. The levels of CH2FH4 were foundto be decreased in HCT-8/FU7dR cells, providing a possibleexplanation for the previous findings. The addition of 100 ÃŸMLV resulted in a higher degree of inhibition of TS activity at10 and 30 UM FUra in both sensitive and pulse resistant cells(Fig. 3A), but the same potentiation was not observed for HCT-8/FU7dR cells (Fig. 3/1). Also, although in this cell line theaddition of the reduced folate resulted in a more prolongedinhibition of TS activity after FUra removal, the recovery at 24and 48 h was still high compared to both wild-type and HCT-8/FU4hR cells. The capability of LV to only partially inhibitthe recovery of TS activity might be explained by either animpaired uptake of this compound or an altered intracellularmetabolism of 5-formyltetrahydrofolate. It is known that thepotentiating effects of LV on fluoropyrimidine cytotoxicity aredue to its conversion to CH2FH4 and FH4 after cellular uptake,and a 6-fold increase in the pool size of these reduced folateshas been described in LI210 cells exposed to LV (42, 47). Ithas also been shown that the polyglutamylation of this cofactorresults in a decreased dissociation rate of the ternary complex(48) and consequently in more prolonged TS inhibition (43). Adefective uptake of LV was an unlikely explanation because asimilar increase (approximately 3-fold, data not shown) inCH2FH4 and FH4 pools occurred in the three cell lines after a3-h LV exposure. These data provide indirect evidence thatneither 5-formyltetrahydrofolate transport nor intracellularconversion to other reduced folates was altered in this cell line.The analysis of the intracellular polyglutamates of CH2FH4 andII1 ! clearly shows a lack of long-chain polyglutamates in the

HCT-8/FU7dR cells, both in basal conditions and after a 3-hexposure to 100 //M LV. This defect is specific for this cell linein that, although HCT-8/FU4hR cells probably have a differentdistribution as compared to the wild type, they still show theability to synthesize long-chain polyglutamates. The defectivepolyglutamylation of CH2FH4 may explain both the less prolonged inhibition of TS in this cell line and the partial lack ofeffectiveness of leucovorin. Since short-chain folates are retained less effectively inside the cell this might also explain theslight reduction in FH4 and CH2FH4 pools. It is of interest thatYin et al. (49) described a Hep 2 cell line naturally resistant toFUra that contained a decreased total pool size of folates anda decreased proportion of higher polyglutamate forms.

In HCT-8/FU4hR cells, TS inhibition was achieved andmaintained at a fairly stable level even after FUra removal;therefore, blockade of thymidylate synthesis does not seem toplay a role in the resistance that develops after high-dose short-term FUra administration. This hypothesis is consistent withthe lack of cross-resistance to FdUrd and is supported by thediscrepancy between the percentage of TS inhibition and percentage of cell kill achieved with a 4-h exposure to 100 AIMFUra (Table 3). When compared to the parental and the HCT-8/7dR line, the amount of FUra incorporated into RNA wasless in the HCT-8/FU4hR cell line, using a range of FUraconcentrations and with 3 different conditions of exposure to[6-'H]FUra. Although FUTP levels were not determined, the

normal values obtained for uridine kinase, as well as orotatephosphoribosyltransferase activity, make it unlikely that a decreased synthesis of this nucleotide is the reason for the defectobserved in this cell line. However, either a reduction in pyrim-idine monophosphate kinase or an increased phosphatase activity might lead to lower FUTP levels. The former enzyme isusually found in very high concentrations as compared tonucleoside kinases in most tumor cells (50); therefore, it isunlikely that it becomes rate limiting in FUra activation. Furthermore, the increase in FUra accumulation into RNA observed in this cell line after FUra removal would exclude anincreased degradation rate or a reduced synthesis of FUTP asbeing the lesion responsible for this defect. Either a reducedaffinity of RNA polymerase for FUTP or an increased removalof FUTP from RNA [in analogy to what has been reported forFdUTP in a cell line with a decreased accumulation of FUrainto DNA (51)] might explain these findings. These possibilitiesare under investigation. Ardalan et al. (10) described a leukemiccell line that was resistant to FUra that had both a decreasedrate of incorporation of FUra into RNA, as well as an "accelerated excretion of FdUMP." The incorporation of FUra into

RNA was ascribed to a low rate of enzymic conversion of 5-fluorouridine 5'-monophosphate into 5-fluorouridine 5'-di-

phosphate, thus decreasing the amount of FUTP available forRNA synthesis. It is also evident from the data in Fig. 5 thatFUra accumulation into RNA increases almost linearly withthe concentration of drug. Increasing the duration of exposuredid not result in a similar increase. This finding is in contrastwith some previous studies that report an increased accumulation of FUra into RNA over time (52, 53). The present studydiffers from the previous investigations in that 3 differentconditions of FUra exposure (i.e., 4, 24, and 4 h followed by24 h in drug-free medium) were compared over a range of FUraconcentrations from 10-300 pM. Also, in some of the previousstudies, noncytotoxic concentrations of FUra were used andthe time span over which FUra accumulation into RNA wasfollowed was often too short. Furthermore, the technique used

1862




for RNA isolation in our study results in a purer product thanobtained by the methods used in other reports. In addition tothese considerations, some of this discrepancy might be explained by the normalization of data with RNA content in thesestudies instead of with cell number. It is possible that when thetime of exposure to FUra is prolonged TS blockade may slowdown the incorporation of FUra into RNA because of anincrease in the pool size of UTP due to the lack of negativefeedback inhibition on pyrimidine synthetic pathway by dTTP.Alternatively, the arrest in S phase caused by inhibition ofthymidylate formation might protect cells from RNA-directedcytotoxicity that has been reported to be more pronounced inthe G i phase (54).

These studies show that different schedules of FUra administration may lead to resistance via different mechanisms andalso provides evidence supporting the contention that the FUramechanism of action depends on the schedule of its administration. The clinically relevant schedules of FUra exposure usedin this study resulted in two novel mechanisms of drug resistance, decreased incorporation into RNA and decreasedCH2FH4 polyglutamate formation. A recent report from thislaboratory described a MTX-resistant cell line which was cross-resistant to the combination FdUrd-LV and displayed a lessstable TS inhibition associated with a lack of long-chain poly-glutamates of CH2FH4 (24). This study is the first in whichacquired resistance to FUra has been correlated with decreasedlevels of CH;FH4 accompanied by and probably due to a decreased polyglutamylation of this cofactor.

These results have important clinical implications. The addition of LV may not enhance the cytotoxicity of pulse FUra,at least when optimal cytotoxic concentrations of this agent areused, because this effect is mediated via thymidylate synthaseinhibition. Channeling FUra into RNA using other modulatingagents (MTX, PALA) may improve results when high-doseshort-term administration is used. On the other hand, enhancement of FUra cytotoxicity with LV may be greater when thefluoropyrimidine is administered as a continuous infusion. Experimental support for this suggestion has recently appeared(55). Finally, these data support the concept that pulse FUraand continuous infusion FUra have different mechanisms ofaction. It might be possible to take advantage of the lack ofcross-resistance between these schedules to treat tumors resistant to pulse FUra administration with continuous infusionFUra. Limited clinical data support this concept in that continuous infusion FUra Â±LV may be effective in patients previouslytreated with pulse FUra (56).

REFERENCES

1. Grem. L. G. Fluorinated pyrimidines. In: B. A. Chabner and J. M. Collins(eds.). Cancer Chemotherapy: Principles and Practice, pp. 180-224. Philadelphia, PA: J. B. Lippincott Company, 1990.

2. Buroker, T. R., Moertel. C. G., Fleming, T. R., Everson, L. K., Cullinan. S.A., Krook, J. E., Mailliard. J. A., Marschke, R. F., Klaassen, D. J., Laurie,J. A., and Moon, M. D. A controlled evaluation of recent approaches tobiochemical modulation or enhancement of 5-fluorouracil therapy in colo-rectal carcinoma. J Clin. Oncol.. 3: 1624-1631. 1985.

3. Poon, M. A., O'Connell, M. J.. Moertel, C. G.. Wieand, H. S.. Cullinan. S.

A., Everson, L. K., Krook, J. E., Mailliard, J. A., Laurie, J. A., Tschetter, L.K., and Wiesenfeld, M. Biochemical modulation of fluorouracil: evidence ofsignificant improvement of survival and quality of life in patients withadvanced colorectal carcinoma. J. Clin. Oncol., 7: 1407-1417, 1989.

4. Einhorn, L. H. Improvements in fluorouracil chemotherapy? J. Clin Oncol.,7; 1377-1379, 1989.

5. Moertel. C. G., Fleming. T. R.. Macdonald, J. S., Haller. D. G.. Laurie, J.A., Goodman. P. J., Ungerleider, J. S., Emerson, W. A., Tormey, D. C.,Glick, J. H., Vivila. M. H., and Mailliard, J. A. Levamisole and fluorouracilfor adjuvant therapy of resected colon carcinoma. N. Engl. J. Med., 322:352-358, 1990.

6. Fukushima, M., Ikenaka. K., Shirasaka, T., and Fujii, S. Metabolism of 5-fluorouracil in sensitive and resistant tumor cells. Gann. 70: 47-53, 1979.

7. Goldberg, A. R., Machledt. J. H.. Jr.. and Pardee. A. B. On the action offluorouracil on leukemia cells. Cancer Res., 26: 1611-1615, 1966.

8. Reichard, P., Skold, O., Klein, G., Revesz, L., and Magnusson, P-H. Studieson resistance against 5-fluorouracil. I. Enzymes of the uracil pathway duringthe development of resistance. Cancer Res., 22: 235-243, 1962.

9. Mulkins, M. A., and Heidelberger, C. Biochemical characterization of fluo-ropyrimidine-resistant murine leukemic cell lines. Cancer Res., 42:965-973,1982.

10. Ardalan. B., Cooney, D. A., Jayaram, H. N.. Carrico. C. K.. Glazer, R. L.,Macdonald, J.. and Schein, P. S. Mechanisms of sensitivity and resistanceof murine tumors to 5-fluorouracil. Cancer Res., 40: 1431-1437, 1980.

11. Fernandes. D. J.. and Cranford. S. K. Resistance of CCRF-CEM clonedsublines to 5-fluorodeoxyuridine associated with enhanced phosphatase activities. Biochem. Pharmacol., 24: 125-132, 1985.

12. Umeda, M., and Heidelberger, C. Comparative studies of fluorinated pyrimidines with various cell lines. Cancer Res., 28: 2529-2538, 1968.

13. Bresnick, E., and Thompson, U. B. Properties of deoxythymidine kinasepartially purified from animal tumors. J. Biol. Chem., 240: 3967-3974, 1965

14. Berger, S. H., Jenh, C. H.. Johnson. L. F.. and Berger, F. Thymidylatesynthase overproduction and gene amplification in fluorodeoxyuridine-re-sistant human cells. Mol. Pharmacol., 28: 461-467, 1985.

15. Bapat, A. R., Zarov, C., and Danenberg. P. V. Human leukemic cells resistantto 5-fluoro-2'-deoxyuridinc contain a thymidylate synthetase with loweraffinity for nucleotides. J. Biol. Chem., 25Â«:4130-4136, 1983.

16. Myers, C. E. The pharmacology of fluoropyrimidines. Pharmacol. Rev., 33:1-15, 1981.

17. Calabro-Jones, P. M.. Byfield, J. E., Ward, J. F., and Sharp. J. R. Time-doserelationships for 5-fluorouracil cytotoxicity against human epithelial cancercells in vitro. Cancer Res.. 42: 4413-4420, 1982.

18. Santelli. G.. and Valeriote. F. Schedule-dependent cytotoxicily of 5-fluorouracil in mice. J. Nati. Cancer Inst., 76: 159-164. 1986.

19. Hryniuk, W. M., Figueredo, A., and Goodyear, M. Applications of doseintensity to problems in chemotherapy of breast and colorectal cancer. Semin.Oncol.. 14:)-\l. 1987.

20. Sobrero, A., Aschele, C., Rosso, R., Nicolin, A., and Berlino, J. R. Rapiddevelopment of resistance to antifolates in vitro: possible clinical implication.J. Nati. Cancer Inst., 83: 24-28, 1991.

21. Petrelli. N., Douglass. H. O.. Jr., Herrera. L.. Russell. D., Stablein, D. M.,Bruckner, H. W., Mayer, R. J., Schinella. R., Green, M. D., Muggia, F. M.,Megibow, A., Greenwald, E. S., Bukowski, R. M., Harris, J., Levin, B.,Gaynor, E., Loutfi, A., Kaiser, M. H.. Barkin. J. S.. Benedetto, P., Woolley,P. W., Nauta, R.. Weaver. D. W., and Leichman, L. P. The modulation offluorouracil with leucovorin in metastatic colorectal carcinoma: a prospectiverandomized phase III trial. J. Clin. Oncol., 7: 1419-1426, 1989.

22. Valone, F. H., Friedman, M. A., Wittlinger, P. S., Drakes, T., Eisenberg. P.D., Malee, M., Hanningan, J. F., and Brown, B. W.. Jr. Treatment of patientswith advanced colorectal carcinomas with fluorouracil alone, high-dose leucovorin plus fluorouracil, or sequential methotrexate, fluorouracil, and leucovorin: a randomized trial of the Northern California Oncology Group. J.Clin. Oncol., 7: 1427-1436, 1989.

23. Romanini, A., Lin, J. T., Niedzwiecki, D., Bunni. M., Priest, D. G., andBenino. J. R. Role of folylpolyglutamates in biochemical modulation offluoropyrimidines by leucovorin. Cancer Res., 51: 789-793, 1991.

24. Priest, D. G., and Mangum, M. Relative affinity of 5.10-methylenetetrahy-drofolylpolyglutamates for the Laclobacillus casei thymidylate synthase-5-fluorodeoxyuridylate binary complex. Arch. Biochem. Biophys., 210: 118-123, 1981.

25. Tompkins, W. A. F., Watrach, A. M., Schmale, J. D., Sanili/. R., andHarris, J. Cultural and antigcnic properties of newly established cell strainsderived from adenocarcinomas of the human colon and rectum. J. Nati.Cancer Inst., 52; 1101-1110, 1974

26. Sobrero, A. F., Moir, R. D.. Berlino. J. R.. and Handschumacher, R. E.Defective facilitated diffusion of nucleosides, a primary mechanism of resistance to 5-fluoro-2'-deoxyuridine in the HCT-8 human carcinoma line. Cancer Res., 45: 3155-3160. 1985.

27. Galivan, J. Transport and metabolism of methotrexate in normal and resistant cultured rat hepatoma cells. Cancer Res., 39: 735-743, 1979.

28. Taylor, A. T., Stafford, M. A., and Jones, O. W. Properties of thymidinekinase partially purified from human fetal and adult tissue. J. Biol. Chem.,247: 1930-1935. 1972.

29. Bradford, M. A rapid and sensitive method for the quantitation of microgramquantities of protein utilizing the principle of protein-dye binding. Anal.Biochem., 72: 248-254, 1976.

30. Roberts, D. An isotopie assay for thymidylate synthase. Biochemistry. 5:3546-3548. 1966.

31. Fernandes. D. J., and Cranford. S. K. A method for the determination oftotal, free and 5-fluorodeoxyuridylate-bound thymidylate synthase in cellextracts. Anal. Biochem., 142: 378-385, 1984.

32. Yalowich, J. C.. and Kaiman. T. I. Rapid determination of thymidylatesynthase activity and its inhibition in intact LI210 leukemic cells in vitro.Biochem. Pharmacol.. 34: 2319-2324. 1985.

33. Rodenhuis. S., McGuire. J. J., Narayanan. R., and Berlino, J. R. Development of an assay system for the detection and classification of methotrexateresistance in fresh human leukemic cells. Cancer Res., 46:6513-6519. 1986.

34. Pizzorno. G., Chang, Y. M., McGuire. J. J.. and Berlino, J. R. Inherent

1863




resistance of human squamous carcinoma cell lines to methotrexate as aresult of decreased polyglutamylation of this drug. Cancer Res., 49: 5275-5280, 1989.

35. Bunni, M., Doig, M. T., Donato, H., Kesavan. V., and Priest, D. G. Role ofmethylenetetrahydrofolate depletion in methotrexate-mediated intracellularthymidylate synthesis inhibition in cultured LI210 cells. Cancer Res., 48:3398-3404, 1988.

36. Santi, D. V., McHenry, C. S., and Perriard, E. R. A filter assay for thymidylate synthetase using 5-flouro-2'deoxyuridylate as an active site titrant.Biochemistry, 13:467-470, 1974.

37. Priest, D. G., and Doig, M. T. Tissue folate polyglutamate chain lengthdetermination by electrophoresis as thymidylate synthase-fluorodeoxyuridy-late ternary complexes. Methods Enzymol., 122: 313-319, 1986.

38. Houghton, J. A., Williams, L. G., Radparvar, S., and Houghton, P. J.Characterization of the pools of 5,10-methylenetetrahydrofolates and tetra-hydrofolates in xenografts of human colon adenocarcinoma. Cancer Res.,48: 3062-3069, 1988.

39. Chomczynski, P., and Sacchi, N. Single-step method of RNA isolationby acid guanidinium thiocyanate-phenol-chloroform extraction. Anal.Biochem., 762: 156-159, 1987.

40. Heidelberger, C. Fluorinated pyrimidines and their nucleosides. In: A. C.Sartorelli and D. G. Johns (eds.), Handbook of Experimental Pharmacology,Vol. 38, Part 2. pp. 193-231. New York: Springer-Verlag, 1975.

41. Clark, J. L., Berger, S. H.. Mittelman, A., and Berger, F. G. Thymidylatesynthase gene amplification in a colon tumor resistant to fluoropyrimidinechemotherapy. Cancer Treat. Rep., 71: 261-265, 1987.

42. Spears, C. P., Shahinian. A. H., Moran, R. G., Heidelberger, C., and CorbeÂ«,T. H., In vivo kinetics of thymidylate synthetase inhibition in 5-fluorouracil-sensitive and -rÃ©sistantmurine colon adenocarcinomas. Cancer Res., 42:450-456, 1982.

43. Keyomarsi, K., and Moran, R. G. Mechanism of the cytotoxic synergism offluoropyrimidines and folinic acid in mouse leukemic cells. J. Biol. Chem.,263: 14402-14409, 1988.

44. Myers, C. E., Young, R. C., and Chabner, B. A. Biochemical determinantsof 5-fluorouracil response in vivo: the role of deoxyuridylate pool expansion.J. Clin. Invest., 56: 1231-1238, 1975.

45. Danenberg, P. V., and Danenberg, K. D. Effect of 5,10-methylenetetrahydro-folate on the dissociation of 5-fluoro-2'-deoxyuridylate from thymidylatesynthetase: evidence for an ordered mechanism. Biochemistry, 17: 4018-4024, 1978.

46. Lockshin, A., and Danemberg, P. V. Biochemical factors affecting the tightness of 5-fluorodeoxyuridylate binding to human thymidylate synthetase.Biochem. Pharmacol., 30: 247-257, 1981.

47. Keyomarsi, K., and Moran, R. G. Folinic acid augmentation of the effects offluoropyrimidines on murine and human leukemic cells. Cancer Res., 46:5229-5235, 1986.

48. Radparvar, S., Houghton, P. J., and Houghton, J. A. Effect of polyglutamylation of 5,10-methylenetetrahydrofolate on the binding of 5-floro-2'-

deoxyuridylate to thymidylate synthase purified from a human colon adenocarcinoma xenograft. Biochem. Pharmacol.. 38: 335-342, 1989.

49. Yin, M. B., Zakrezewski, S. F., and Hakala, M. T. Relationship of cellularfolate cofactor pools to the activity of 5-fluorouracil. Mol. Pharmacol., 23:190-197, 1982.

50. Hande, K. R., and Chabner, B. A. Pyrimidine nucleoside monophosphatekinase from human leukemic blast cells. Cancer Res., 38: 579-585, 1978.

51. Chu, E., Lai, G. M., Zinn, S., and Allegra, C. J. Resistance of a humanovarian cancer line to 5-fluorouracil associated with decreased levels of 5-fluorouracil in DNA. Mol. Pharmacol., 38: 410-417, 1990.

52. Schuetz, J. D., Wallace, H. J.. and Diasio, R. B. 5-fluorouracil incorporationinto DNA of CF-1 mouse bone marrow cells as a possible mechanism oftoxicity. Cancer Res., 44: 1358-1363, 1984.

53. Benz, C., and Cadman, E. Modulation of 5-fluorouracil metabolism andcytotoxicity by antimetabolite pretreatment in human colorectal adenocarcinoma HCT-8. Cancer Res., 41:994-999, 1981.

54. Maybaum, J., oilman, B., Mandel, H. G., et al. Regulation of RNA- andDNA-directed actions of 5-fluoropyrimidines in mouse T-lymphoma (S-49)cells. Cancer Res., 40: 4209-4215, 1980.

55. Moran, R. G., and Scanlon, K. L. Schedule dependent enhancement of thecytotoxicity of fluoropyrimidines to human carcinoma cells in the presenceof folinic acid. Cancer Res., 51: 4618-4623, 1991.

56. Hartmann, H., Kessinger, A., and Lemon, H. Five days continuous infusionof 5-fluorouracil for advanced colorectal, gastric and pancreatic adenocarcinoma. J. Surg. Oncol., //: 227-238, 1979.

1864



1992;52:1855-1864. Cancer Res Carlo Aschele, Alberto Sobrero, Mary A. Faderan, et al. Different Clinically Relevant Dose SchedulesColon Cancer (HCT-8) Sublines following Exposure to Two Novel Mechanism(s) of Resistance to 5-Fluorouracil in Human

Updated version

http://cancerres.aacrjournals.org/content/52/7/1855

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

[email protected] 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/52/7/1855To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



novel mechanism(s) of resistance to 5-fluorouracil in...

Documents