[CANCER RESEARCH 46,169-174, January 1986]

Pharmacokinetics of Trimetrexate (NSC 352122) in Monkeys

Frank M. Balis,1 Cynthia M. Lester, and David G. Poplack

Pediatrie Branch, National Cancer Institute, NIH, Bethesda, Maryland 20892

ABSTRACT

The pharmacokinetics of trimetrexate was studied in Rhesusmonkeys following i.v. bolus, continuous i.v. infusion, oral, andsubcutaneous administration. Two methods were used to measure drug concentration in plasma, cerebrospinal fluid (CSF), andurine: the dihydrofolate reducíaseinhibition assay, and a reversephase high-pressure liquid chromatography assay. The phar-macokinetic behavior of trimetrexate was characterized by triex-ponential plasma disappearance, elimination primarily by bio-

transformation, substantial plasma protein binding, poor CSFpenetration, and limited oral bioavailability. Methotrexate. administered in an equimolar dose for comparison, was cleared morerapidly from plasma than was trimetrexate. Trimetrexate concentration remained above 0.1 UM 3-fold longer. In contrast to

methotrexate, which is cleared almost exclusively by renal excretion, renal clearance of trimetrexate accounted for <5% oftotal clearance. A significant discrepancy was observed inplasma and urine trimetrexate concentrations measured by thetwo assay methods. The dihydrofolate reducíaseinhibition assaygave results approximately 2- to 4-fold higher in plasma. Two

metabolites of trimetrexate which inhibit dihydrofolate reductasewere identified in urine (one was also found in plasma) andappear to account for the different results obtained by the twoassays. These metabolites would probably also interfere withthe competitive protein binding assay currently being used tomeasure trimetrexate in ongoing phase I trials.

INTRODUCTION

Trimetrexate (2,4-diamino-5-methyl-6-[(3,4,5-trimethoxyani-

lino) methyl Iquinazoline) is a new folate antagonist which iscurrently undergoing phase I clinical testing. Like methotrexate,the classical antifolate, trimetrexate is a potent inhibitor of theenzyme dihydrofolate reductase (1). However, the mechanism ofcell entry and intracellular metabolism of trimetrexate and methotrexate apparently differ (2-4). Trimetrexate also has a broader

range of antitumor activity in vivo against murine tumors thandoes methotrexate (5). As has been demonstrated previously formethotrexate (6), the antitumor activity of trimetrexate In vivoagainst P388 leukemia is markedly schedule-dependent with

more prolonged exposures to the drug producing a greater effect(5).

The biochemical pharmacology and pharmacokinetics of methotrexate have been studied extensively, and the application ofthis knowledge has resulted in the safer and more efficacioususe of methotrexate. Although some of these principles mayalso be applicable to the clinical use of trimetrexate, differencesin structure, metabolism, mechanisms of resistance, and spectrum of antitumor activity suggest these agents may not be

totally comparable. In the present study the pharmacokinetics oftrimetrexate in the subhuman primate was examined and contrasted to the pharmacokinetics of methotrexate in the samespecies. Significant differences in the route of elimination, rateof clearance, and volume of distribution were observed. In addition, metabolites of trimetrexate capable of inhibiting dihydrofolate reductase were detected.

MATERIALS AND METHODS

Drug. Trimetrexate glucuronate was obtained from the Drug Development Program, Division of Cancer Treatment, National Cancer Institute, Bethesda, MD and was supplied in 10-ml vials, each containing 50

mg. This was reconstituted in 1.9 ml sterile water (25 mg/ml) for administration to the monkeys. Trimetrexate acetate which was used in makingstandards for the HPLC2 and dihydrofolate reductase inhibition assays

was also obtained from the Drug Development Program, National CancerInstitute. 14Cradiolabeled trimetrexate isethionate (specific activity, 26.5ftC'i/mg) was obtained from Warner-Lambert, Ann Arbor, Ml. The radio-

labeled carbon was in the middle methoxy group of the trimethoxyanilinering. Prior to administration of the "C-labeled trimetrexate to two mon

keys, it was diluted with cold trimetrexate glucuronate to a specificactivity of 1.8 ^Ci/mg and filtered through a 0.22-^m filter. The purity ofboth the acetate salt and the 14C-labeled trimetrexate was >98% by

HPLC.Monkeys. Adult male rhesus monkeys (Macaca mulatta) ranging in

weight from 6.7 to 11.7 kg were used in these experiments. The animalswere housed individually and received food and water ad libitum (exceptprior to an oral dose, in which case animals were fasted overnight).Blood samples were drawn from a saphenous or femoral venous catheteropposite from the site of drug administration. CSF samples were obtained from a chronically indwelling Pudenz catheter attached to a s.c.implanted Ommaya reservoir (7).

Experiments. The pharmacokinetics of trimetrexate was studied following i.v. bolus and continuous i.v. infusion dosing schedules andfollowing administration by the oral and subcutaneous routes. Threemonkeys received an i.v. bolus dose of 100 mg/sq m dose. Bloodsamples were collected in heparinized tubes prior to administration and5, 15, and 30 min and 1, 2, 3, 4, 6, 8, 12, and 24 h after the dose.Plasma was separated immediately by centrifugation and frozen at -20°

until assayed. Except for the 5-min sample, CSF was collected from the

Ommaya reservoir simultaneously with blood samples and frozen at-20°. The reservoir was pumped four times before and after each

sample to ensure adequate mixing with ventricular CSF.The oral bioavailability of trimetrexate was studied in three animals.

Each received the 30 mg/sq m dose as an i.v. bolus and, in a secondexperiment via a nasogastric tube with the order of administrationrandomized. Blood sampling times were identical to those listed abovefor the 100 mg/sq m dose with additional samples obtained at 45 and90 min following the dose. Two of these animals given the 30 mg/sq mi.v. bolus dose received 14C-labeled trimetrexate. These two animals also

had all urine and stool collected and frozen over 72 h following the dosein 24-h aliquots.

A continuous i.v. infusion was administered to three monkeys at a

Received 5/14/85; revised 8/8/85; accepted 10/2/85.'To whom requests for reprints should be addressed, at Bldg. 10, Room

13N240, Pediatrie Branch, National Cancer Institute, NIH, Bethesda, MD 20892.

2The abbreviations used are: HPLC, high-pressure liquid chromatography;

DHFR INH, dihydrofolate reductase inhibition; AUC, area under the curve; cone.,concentration; CLâ„¢,total body clearance; CSF, cerebrospinal fluid.

CANCER RESEARCH VOL. 46 JANUARY 1986

169

on April 8, 2021. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

PHARMACOKINETICS OF TRIMETREXATE

dose of 10 mg/sq m/h, which was calculated to achieve a steady stateplasma level of 5 MMusing clearance values obtained from the i.v. bolusexperiments. Prior to the 8-h infusion each animal received an i.v. bolus

loading dose of 40 mg/sq m. Blood samples were collected prior to thedose and 5, 15, and 30 min and 1, 2, 4, 6, 8, 8.25, 12, and 24 h afterthe start of the infusion. CSF sampling times were 0,1, 2, 4, 6, and 8 h.The same three animals also received a subcutaneous dose of 125 mg/sq m, which was approximately equal to the total dose given by infusion.Blood sampling times were identical to those for the i.v. infusion.

All animals received leucovorin 24 and 48 h after each dose oftrimetrexate. No significant toxicity was noted in any of the animals withthis regimen.

For comparison, plasma and CSF were monitored in three monkeysfollowing a 34 mg/sq m dose of methotrexate administered as an i.v.bolus. Blood and CSF samples were obtained at 0, 15, and 30 min and1, 2, 3, 4, 6, 8, and 12 h after the dose. The samples were otherwisehandled in the same manner as described for trimetrexate.

Sample Analysis. Trimetrexate in plasma, CSF, and urine was measured by two methods, the DHFR INH assay described previously formethotrexate (8) and a HPLC assay. The DHFR INH assay was alsoused to measure methotrexate in the plasma and CSF of the threemonkeys receiving this drug. The DHFR INH assay did not requiremodification to measure trimetrexate. The lower limit of sensitivity withthe DHFR INH assay is 0.005 pM for trimetrexate and 0.001 ^M formethotrexate. The coefficient of variation for within run and day-to-day

replicates is <10% for both drugs.Plasma samples and standards (1.0 ml) to be measured for trimetrex

ate by HPLC were spiked with [14C]trimetrexate as an internal standard

and applied to a C18 SEP-PAK cartridge (Waters Associates, Milford,MA). After a 3-ml water wash, the trimetrexate was eluted with 2 ml

acetonitrile, and the eluant was evaporated to dryness under a nitrogenstream. The residue was reconstituted in 0.2 ml of mobile phase, resultingin a 5-fold concentration of the original 1.0 ml plasma sample. The

recovery with this sample preparative technique was 71 ±8% (SD). TheHPLC system used consisted of a Waters model 680 automated gradientcontroller with two model 510 pumps and a U6K injector (WatersAssociates). The column was a lO-^m C18Radial-Pak cartridge used ina Z-module radial compression separation system (Waters Associates).

The eluant was monitored with a Beckman 165 dual channel, variablewavelength detector (Beckman Instruments, Inc., Berkeley, CA) at awavelength of 240 nm. The mobile phase was a 0.01 M KH2PO4 buffer,pH 7.5, with 40% acetonitrile with a flow rate of 2.5 ml/min. The retentiontime for trimetrexate under these conditions was 12.5 min. With the 5-

fold concentration of the samples, the lower limit of sensitivity is 0.05MM.The coefficient of variation for replicates is <10%. Plasma standardsfor the HPLC assay were measured with the DHFR INH assay to ensurecomparability of the two methods.

The plasma samples from the monkeys given [14C]trimetrexate had tobe assayed under different conditions, since the 14Calready present inplasma would interfere with using [14C]trimetrexate as an internal stand

ard. Instead, 0.2 nmol of trimethoprim (Sigma Chemical Co., St. Louis,MO) was added as an internal standard to each sample prior to extractionon the SEP-PAK. In addition, a 5-^m Nova-Pak Ci8 Radial-Pak cartridge

(Waters Associates) with a mobile phase of 0.02 M KH2PO4 buffer, pH4.5, with 23% acetonitrile at a flow rate of 2 ml/min replaced the 10-jim

column and mobile phase described above. The retention times fortrimethoprim and trimetrexate are 3 and 5.8 min. This is a more rapidand convenient technique for future studies. Plasma was screened forthe presence of metabolites using the 5-^m Nova-pak column and a

mobile phase of 0.01 M KH2PO4 buffer, pH 7.5, with 12% acetonitrile at2 ml/min.

Urine from the two monkeys receiving radiolabeled drug was centri-fuged and injected directly onto the lO-^m Ci8 Radial-Pak column. The

mobile phase was 0.01 M KH2PO4 buffer, pH 5.0, at 2 ml/min with alinear gradient of acetonitrile from 0 to 60% over 25 min. Fractions werecollected and measured with the DHFR INH assay and counted in a

Beckman LS 8100 scintillation counter (Beckman Instruments). A fractioneluting before trimetrexate which inhibited DHFR was collected, lyophi-

lized, reconstituted in a small amount of mobile phase and réinjectée)onto the 10-^m C18column with a mobile phase of 0.01 M KH2PO4 buffer,

pH 7.5, with 12% acetonitrile at 2 ml/min. The eluant was monitoredwith the UV detector at 240 nm and fractions were again collected forthe DHFR INH assay and counting for 14C.

The 72-h stool collection was homogenized in 3 volumes of water. Analiquot was centrifuged, and the supernatant was counted for 14C.

Counts per minute for all radiolabel studies were corrected for quenchusing the H# technique (9).

Plasma Protein Binding. Fresh plasma from a normal monkey wasspiked with trimetrexate glucuronate to a final concentration of 5 /¿Mandplaced in one chamber of a dialysis cell (Technilab Instruments, Pequan-nock, NJ). An equal volume of phosphate-buffered saline, pH 7.2 (Bio-

fluids, Rockville, MD), with 0.01% sodium azide was added to the otherchamber. The two chambers were separated by a regenerated cellulosedialysis membrane (Technilab Instruments). Dialysis was performed at37°with gentle shaking for 16 h. At the end of the incubation an aliquot

from each chamber was assayed using the DHFR INH assay. Thepercentage of drug bound was equal to:

Plasma concentration - buffer concentration

Plasma concentrationx 100

Pharmacokinetic Analysis. Plasma concentration-time data from thei.v. bolus experiments were fitted to both biexponential (n = 2) andtriexponential (n = 3) equations,

C(0 = îAfi-x¡-i

using MLAB, a nonlinear curve fitting program (10). Akaike's information

criterion (11 ) was used to determine which equation best fitted the data.The half-life for each phase of elimination was calculated by dividing

0.693 by the rate constant (A,) for that phase. Other pharmacokineticparameters were calculated using model-independent methods. The AUC

was derived by the linear trapezoidal method (12), and extrapolated toinfinity by adding the quotient of the final plasma concentration dividedby the terminal rate constant (\„).CL™was the dose divided by the AUC,

and renal clearance equaled the amount excreted in urine over 72 hdivided by the AUC. In the monkeys receiving trimetrexate by continuousinfusion, CLTBwas estimated by dividing the infusion rate by the steady

state concentration in plasma. The volume of distribution at steady statewas calculated from the area under the moment curve (13).

The fraction of drug penetrating into the CSF was derived from theratio of the AUCs in CSF and plasma for the i.v. bolus dose and the ratioof steady state concentration in CSF and plasma for the continuous i.v.infusion. Bioavailability of the oral dose was the ratio of the AUC followingthe oral dose and the AUC following the same dose given by i.v. bolus.

RESULTS

Pharmacokinetics. The plasma disappearance of trimetrexatefollowing an i.v. bolus dose (Chart 1) was best fitted by atriexponential equation. The half-lives and other pharmacokinetic

parameters calculated from the i.v. bolus doses are listed inTable 1. Plasma concentrations measured by the more specificHPLC assay were lower than those obtained with the DHFR INHassay after the initial time points, which suggests the presenceof one or more circulating metabolites of trimetrexate capable ofinhibiting DHFR (Chart 1). The values for pharmacokinetic parameters calculated from plasma concentrations measured by thetwo techniques were also significantly different (Table 1). Plasma

CANCER RESEARCH VOL. 46 JANUARY 1986

170

on April 8, 2021. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

PHARMACOKINETICS OF TRIMETREXATE

and CSF trimetrexate conœntrations measured by the DHFRINH assay and reported as UM actually represent trimetrexateplus unknown concentrations of DHFR inhibiting metabolites.

Chart 2 and Table 1 show the comparison between trimetrexate and methotrexate administered in approximately equimolardoses. Methotrexate is cleared more rapidly from plasma andhas a smaller volume of distribution at steady state. The plasmaconcentration of trimetrexate remains above 0.1 MM[the concentration at which methotrexate and trimetrexate produced >90%inhibition of cell growth in vitro against a human lymphoblasticleukemia cell line (3)] 3 times longer than does that of methotrexate. The route of elimination also appears to be different.Renal clearance of trimetrexate accounts for only a small fractionof CLTB,which indicates that trimetrexate is cleared primarily by

biotransformation, in contrast to methotrexate, which is knownto be cleared almost exclusively by renal excretion in humans(14). Despite its greater lipophilicity, trimetrexate does not penetrate into the CSF to any greater extent than does methotrexate(Chart 3 and Table 1).

When administered as a 10 mg/sq m/h continuous i.v. infusion,trimetrexate achieved a steady state concentration in plasma of3.32 ± 0.51 MM by HPLC (5.75 ± 0.48 MM by DHFR INH).

100 r

i 10e

1—I—I—I—I—I—ii i i i i i

12 16 20 24

TIME [hr]

Chart 1. Plasmatrimetrexate concentration following an i.v. bolus dose of 100mg/sq m measured by the HPLC assay (•)and the DHFR INH assay (O).Measurements made by the DHFR INH assay represent trimetrexate plus anunknown concentration of metabolites producing an equivalent degree of DHFRinhibition. All points are the geometric mean from three monkeys.Bars, SD.

Estimates of CLTB calculated from this plasma steady state

concentration was 138 ±22 ml/min/sq m. The ratio of steadystate concentration in CSF to plasma was 0.029 ±0.029 (DHFRINH assay). These values are in close agreement to thoseobtained from the i.v. bolus doses (Table 1).

Chart 4 is a comparison between plasma levels following the8-h i.v. infusion and the s.c. dose. The s.c. dose resulted in a

slow release of trimetrexate that simulated the infusion andprovided a more prolonged exposure to the drug than an i.v.bolus dose.

Results of the oral bioavailability study are shown in Chart 5.The fraction of parent drug absorbed unchanged into the systemic circulation was only 0.23 ±0.05 (HPLC method). Therewas a greater discrepancy in plasma concentration and AUCbetween assay methods with the oral dose than the i.v. dose(Table 2). Following the oral dose trimetrexate (unchanged), asmeasured by the HPLC assay, accounts for only one-fourth toone-fifth of the DHFR inhibiting activity in plasma, while followingan i.v. dose it accounts for one-half of the DHFR inhibiting activity

in plasma.Protein Binding Studies. Trimetrexate was 90 ±1% bound

to plasma proteins at a concentration of 5 MM,as measured bythe equilibrium dialysis method. Ninety-three % of the drug

100 r

o

p .1-

.001 -

12

TIME [hr]

16 20 24

Chart 2. Plasma methotrexate concentration (D) measured by the DHFR INHassay following a 34 mg/sq m i.v. bolus dose and trimetrexate concentrationmeasuredby the HPLC assay (•)following a 30 mg/sq m i.v. bolus dose. Allpointsare the geometric mean from three monkeys.Bars, SD.

Table 1Comparisonof pharmacokineticparametersof trimetrexateadministered by i.v. bolus in two different doses and measuredby two methodsand comparisonof

equimolar doses of trimetrexateand methotrexateTrimetrexate concentrations measured by DHFR INH actually represent trimetrexate plus DHFR-inhibitingmetabolites.

DrugTrimetrexateTrimetrexate

MethotrexateDose

(mg/m2)1003034No.33

3AssayH

DHDAUCpa

(¿iM-min)2262

±468"3773 ±468"

596 ±173385 ±39AUC«*(/i

u-nun)142

±151

8.4 ±3.9CSF:P0.034

±0.029

0.021 ±0.008CI™(ml/min/m2)123

±2473 ±8

144 + 39194 ±21CI"

(ml/min/m2)5.0

±3.4°Half-lives

(min)(liters/m2)23.7

±8.417.8 ±1.725.1 ±6.89.5 ±2.4a3.3

3.93.76.3ß39

452626T199

248162140

" P, plasma; CLR,renal clearance; Vd„,volume of distribution at steady state; H, HPLC assay; D, DHFR INH assay.6 Mean ±SD.cn = 2.

CANCER RESEARCH VOL. 46 JANUARY 1986

171

on April 8, 2021. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

PHARMACOKINETICS OF TRIMETREXATE

100 r

10 r

.01

.001 -1 1 i I

4 8 12 16

TIME [hr]

20 24

Chart 3. Plasma (O) and CSF (A) trimetrexate concentration measured by theDHFR INH assay following an i.v. bolus dose of 100 mg/sq m. Measurementsmade by the DHFR INH assay represent trimetrexate plus an unknown concentration of metabolites producing an equivalent degree of DHFR inhibition. All pointsare the geometric mean from three monkeys. Bars, SO.

initially present in the plasma was recovered from the plasmaand buffer at the end of the incubation.

Metabolism Studies. In an attempt to evaluate trimetrexatemetabolism, two monkeys received 14C-labeled trimetrexate.

Over the 72 h following drug administration, only 16% of thetotal dose of 14Cadministered was recovered in urine, and only1% was found in the stool. Plasma clearance of total 14C was

biexponential, with an initial half-life of 3.7 min, which is similarto the initial half-life of the parent compound. However, theterminal 14C half-life of 1030 min was much longer than the

terminal half-life of trimetrexate.Chart 6 shows the chromatograms of urine from the two

monkeys receiving 14C-labeled trimetrexate. Six to seven peaks

of radioactivity are present, one representing trimetrexate at 40min. Two peaks of DHFR inhibiting activity were also present inurine, one corresponding to the parent drug (40 min) and a largerpeak at 18 min, apparently representing a more polar metabo-

lite(s) of trimetrexate. This peak was not present in urine collected before or 6 days after the dose. The metabolite peak doesnot coelute with a peak of radioactivity, which suggests that themiddle methoxy group containing the 14C has been demethyl-

ated.This peak containing DHFR inhibiting material (fraction 18) was

collected, lyophilized to dryness, reconstituted in a small amountof mobile phase, and reinjected under conditions described in"Materials and Methods," yielding the chromatogram shown in

Chart 7. Two peaks can be detected by DHFR INH assay ofeluant fractions and by monitoring UV absorption at 240 nm thatare not present in baseline urine samples. There was no radioactivity associated with either peak. These peaks apparentlyrepresent metabolites of trimetrexate that are capable of inhibiting DHFR (the metabolites inhibit both Lactobacillus case/' and

bovine DHFR).Plasma was also screened for the presence of metabolites.

Chart 8 shows chromatograms of extracted plasma prior to the

100 r

10

.1 -i I I 1

4 8 12 16

TIME [hr]

20 24

Chart 4. Plasma trimetrexate concentration measured by the DHFR INH assayfollowing a subcutaneous dose of 125 mg/sq m (D) and a continuous i.v. infusionfor 8 h at a dose rate of 10 mg/sq m/h (O). Monkeys receiving the i.v. infusionwere administered an initial loading dose of 40 mg/sq m. Measurements made bythe DHFR INH assay represent trimetrexate plus an unknown concentration ofmetabolites producing an equivalent degree of DHFR inhibition. All points are thegeometric mean from three monkeys. Bars, SD.

100

10

A

HPLC

B

DHFR Inh.

F - .51

12 16 20 24 0 4

TIME [hr]

16 20 24

Chart 5. Plasma trimetrexate concentration following a 30 mg/sq m dose givenby i.v. bolus (•,O) and via a nasogastric tube (•,11).A shows the levels measuredby the HPLC assay, and B shows levels measured by the DHFR INH assay.Measurements made by the DHFR INH assay represent trimetrexate plus anunknown concentration of metabolites producing an equivalent degree of DHFRinhibition. F is the fraction of the oral dose absorbed derived from the ratio of theAUC's following oral and i.v. doses. All points represent the geometric mean from

three monkeys. Bars, SD.

dose and at serial time points after the dose in a monkeyreceiving an i.v. bolus dose. Two peaks which only appearedafter trimetrexate was administered can be identified. One elutedat 4.5 min, corresponding to one of the DHFR inhibiting metabolites seen in urine. The second peak at 5.9 min had no detectable DHFR inhibiting activity associated with it and may representan inactive metabolite. A very small peak is seen in the chromatogram at the time (7.3 min) the second DHFR inhibiting metab-

CANCER RESEARCH VOL. 46 JANUARY 1986

172

on April 8, 2021. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

PHARMACOKINETICS OF TRIMETREXATE

Table 2Comparisonof area under the plasma concentration time curve following i.v. and

oral dosing and peak plasma levels following oral dosing, using both assaymethods to measureplasma concentration

Trimetrexate concentrations measured by the DHFR INH assay actually represent trimetrexate plus DHFR-inhibitingmetabolites.

.01 r

AUC (nM-

Route of Doseadministration (mg/sq m)

Peak plasma concentration (fiM)

DHFR INH HPLC DHFRINH HPLC

i.v.P.O.

3030

999 +103s 596 ±173499 + 120 128±13 1.31 ±0.15 0.36 ±0.10

8 Mean ±SD.

10 20 30 40 50 10

FRACTION[min]

Charte. Two chromatograms of the urines collected over 24 h from twomonkeys given a 30 mg/sq m dose of 14C-labeledtrimetrexate by i.v. bolus. One-

min fractions of eluant were collected and measured by DHFR INH assay ( )and for UC ( ). The peak at 40 min represents unchanged trimetrexate. Thepeak in fraction 18 was collected and reinjected under different conditions (Chart7).

.01 r

O

io;O

OI

CD 'A<o>

O 5 10 15

TIME [min]

Chart 7. Chromatogramof material in fraction 18 shown in Chart 6. Eluant wasmonitoredwith a UV detector at a wavelengthof 240 nm ( ) and 1-minfractionsof eluant were collected and measured by the DHFR INH assay ( ). Peatelabeled 1 and 2 are believedto be metabolites of trimetrexate capable of inhibitingDHFR.Eluant from underpeate 1 and 2 did not contain 14C.

olite would be expected, but it is present in too low a concentration to confirm its presence by the DHFR INH assay.

360 m

TIME[min]

Chart 8. Chromatogramsof serial plasmafrom a monkey receivingan i.v. bolusdose of 100 mg/sq m of trimetrexate. Assay conditions are described in "Materialsand Methods." Eluant was monitored with a UV detector at a wavelength of 240

nm. Peak 1 corresponds to peak 1 in the Chromatogramof urinary metabolites(Chart 7). Eluant from under peak 1 inhibited DHFR. The small peak labeled 2corresponds io peak 2 in Chart 7 but is present in too low a concentration to inhibitDHFR, if it does represent the metabolite. The retention times of peaks 1 and 2differ from those in Chart 7 because a different column was used (see "Materialsand Methods" section).Eluant from underpea* 3, which may represent an inactive

metabolite of trimetrexate, did not inhibit DHFR. Under these conditions trimetrexate élûtesat >20 min.

DISCUSSION

The purpose of this study was to define the pharmacokineticbehavior of trimetrexate in a subhuman primate model whichhas previously been predictive of drug kinetics in humans (15).Trimetrexate pharmacokinetics in this model were characterizedby triexponential plasma disappearance, clearance primarily bybiotransformation, poor CMS penetration, substantial plasmaprotein binding, and limited oral bioavailability.

The metabolism of trimetrexate appears to be complex, withboth inactive and DHFR inhibiting metabolites identified in plasmaand urine. Although the exact structure of these metabolites hasyet to be determined, it appears that one step ¡nthe metabolicpathway is the demethylation of the middle methoxy group onthe trimethoxyaniline ring. The loss of this single carbon fragmentfrom the trimetrexate molecule and its subsequent incorporationinto other molecules or excretion via the lungs probably accountsfor the loss of over 80% of the radiolabel and the more prolongedterminal half-life of 14Cin plasma. Demethylation appears to be

a major step in the degradative metabolism of trimetrexate. Sincethe 14C was cleaved from trimetrexate, the multiple peaks of

radioactivity found on chromatography of urine probably do notall represent metabolites of trimetrexate but rather incorporationof the 14Cinto other compounds. The metabolites identified were

more polar than trimetrexate based on their shorter retention onthe reversed-phase columns.

Even though the two metabolites inhibit DHFR, it cannot beconcluded that they are active metabolites until cytotoxicitystudies are performed. Since they appear to be more polar thantrimetrexate, they may not gain entry into cells as readily. Furtherpurification and characterization of these metabolites are ongoing.

The metabolites identified in urine and plasma, which arecapable of inhibiting DHFR, interfere with the DHFR INH assayand presumably would also interfere with the competitive proteinbinding assay (16), which uses DHFR as the binding protein. Asa result, plasma and urine levels are falsely high, and pharmacokinetic parameters calculated from these levels are incorrect.The HPLC method appears to be more specific. When ¡nterpret-

CANCER RESEARCH VOL. 46 JANUARY 1986

173

on April 8, 2021. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

PHARMACOKINETICS OF TRIMETREXATE

¡ngor comparing pharmacokinetic data from ongoing phase Itrials, the method of measuring trimetrexate must be considered.

The disposition kinetics of trimetrexate and methotrexate differsignificantly in the monkey. Trimetrexate is cleared more slowlyand, with an equivalent dose, may therefore provide a moreprolonged drug exposure. The primary route of elimination of thetwo drugs is also different. While methotrexate is excretedalmost exclusively by the kidneys (14), the renal clearance oftrimetrexate accounts for <5% of the total clearance. If trimetrexate is found to have significant antitumor activity, it may beuseful in patients with renal dysfunction in place of methotrexate.

As in the present study in monkeys, Weir ef al. (17) havereported relatively low CSF drug concentrations following systemic administration in dogs. Trimetrexate was originally synthesized to be a lipid soluble antifol, and thus it was anticipated thattrimetrexate would penetrate into the CSF to a greater extentthan methotrexate [the log of the partition coefficient for trimetrexate in octanol versus water is 0.88 compared to -1.85 for

methotrexate (18)]. One explanation for its limited central nervous system penetration in monkeys may be the high percentageof circulating trimetrexate bound to plasma protein. In contrast,protein binding of trimetrexate in dogs was reported to benegligible when measured by a size exclusion column technique(17). This marked difference in results is more likely to be due todifferences in techniques used to meausre protein binding thanspecies differences. Our study used the classical equilibriumdialysis method, which is not affected by nonspecific drug bindingwithin the system.

Using the DHFR INH assay, Weir ef al. (17) reported a highbioavailability of 80% following an oral dose in dogs. In our studywith the more specific HPLC assay, the bioavailability in monkeyswas considerably lower. This difference is in large part due tothe more extensive metabolism of trimetrexate to DHFR-inhibit-

ing metabolites following an oral dose. This phenomenon isprobably a result of presystemic or first pass metabolism oftrimetrexate in the liver or intestinal mucosa.

The slow-release effect following subcutaneous administrationof trimetrexate simulated the levels achieved with an 8-h contin

uous i.v. infusion. Since the duration of exposure to antifols is asignificant factor in producing cytotoxic effect, this convenientroute of administration could provide an alternative to prolongedintravenous infusions.

REFERENCES

1. Berlino, J. R., Sawicki, W. L, Moronson, B. A., Cashmore,A. R., and Elslager,E. F. 2,4-Diamino-5-methyl-6-[(3,4,5-trimethoxyanilino)methyl]quinazoline(TMQ), a potent non-classicalfolate antagonist inhibitor. I—Effectof dihydro-folate reducíaseand growth of rodent tumors in vitro and in vivo. Biochem.Pharmacol.,28: 1983-1987, 1979.

2. Kamen, B. A., Eibl, B., Cashmore, A., and Benino, J. Uptake and efficacyof trimetrexate (TMQ, 2,4-diamino-5-methyl-6-[(3,4,5-trimethoxyanilino>-methylIqumazolme),a nonclassicalantifolate in methotrexate-resistant leukemia cells in vitro. Biochem. Pharmacol.,33:1697-1699,1984.

3. Ohnoshi,T., Ohnuma,T., Takahashi, l., Scanion,K., Kamen,B. A., and Holland,J. F. Establishment of methotrexate-resistant human acute lymphoblasticleukemia cells in culture and effects of folate antagonists. Cancer Res.,42:1655-1660, 1982.

4. Diddens,H., Niethammer,D., and Jackson, R. C. Patterns of cross-resistanceto the antifolate drugs trimetrexate, metoprine, homofolate, and CB3717 inhuman lymphoma and osteosarcoma cells resistant to methotrexate. CancerRes., 43: 5286-5292,1983.

5. Clinical Brochure on Trimetrexate Glucuronate (TMTX), NSC 352122. Be-thesda, MD: Divisionof Cancer Treatment, NationalCancer Institute, 1983.

6. Gangji, D., Ducore, J., Poplack, D., Levine, A., Kohn, K., and Glaubiger, D.Concentration and time dependence of methotrexate cytotoxicity in mouseL1210 leukemia cells and human Burkitt's lymphoma cells (Abstract). Clin.

Res., 28: 525A, 1980.7. Wood, J. H., Poplack,D. G., Bleyer, W. A., and Ommaya,A. K. Primate model

for the chronic study of intraventricularly or intrathecally administered drugsand intracranial pressure. Science(Wash. DC), 195: 499-500,1977.

8. Falk, L. C., Clark, D. R., Kaiman, S.M., and Long, T. F. Enzymatic assay formethotrexate in serum and cerebrospinal fluid. Clin. Chem., 22: 785-788,1976.9. Long, E. C. Applications of quench monitoring by Compton edge: the "H#."Publication no. 1096-NUC-77-2T.Fullerton, CA: Beckman Instruments, Inc.,1977.

10. Knott, G. D. MLAB: a mathematicalmodelingtool. Comput. ProgramsBiomed.,ÕO.261-280,1979.

11. Yamaoka, K., Nakagawa, T., and Uno, T. Application of Akaike's informationcriterion (AIC) in the evaluation of linear pharmacokinetic equations. J. Phar-macokinet. Biopharm., 6:165-175, 1978.

12. Gibaldi, M. and Perrier,D. Pharmacokinetics,ed. 2. New York: Marcel Dekker,Inc., 1982.

13. Perrier, D. and Mayershon, M. Noncompartmental determination of steadystate volumeof distribution for any mode of administration.J. Pharm.Sci., 77:372-373,1982.

14. Chabner, B. A. Methotrexate. In: B. Chabner (éd.),PharmacologiePrinciplesof CancerTreatment, pp. 229-255. Philadelphia:W. B. Saunders Co., 1982.

15. Poplack, D. G., Bleyer, W. A., Wood, J. H., Kostolich, M., Savitch, J. L., andOmmaya, A. K. A primate model for study of methotrexate pharmacokineticsin the central nervous system. Cancer Res., 37: 1982-1985,1977.

16. Drake,J. C., Allegra,C. J., Curt, G. A., and Chabner,B. A. Competitive protein-binding assay for trimetrexate. CancerTreat. Rep., 69: 641-644,1985.

17. Weir, E. C., Cashmore,A. R., Dreyer,R. N.,Graham,M. L., Hsiao,N., Moroson,B. A., Sawicki, W. L., and Benino, J. R. Pharmacologyand toxicity of a potent"nonclassical" 2,4-diamino quinazolinefolate antagonist, trimetrexate, in normal dogs. Cancer Res., 42: 1696-1702,1982.

18. Hamrell.M. R. Inhibitionof dihydrofolate reductase and cell growth by antifolsin a methotrexate-resistant cell line. Oncology (Basel),47: 343-348,1984.

CANCER RESEARCH VOL. 46 JANUARY 1986

174

on April 8, 2021. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1986;46:169-174. Cancer Res   Frank M. Balis, Cynthia M. Lester and David G. Poplack  Pharmacokinetics of Trimetrexate (NSC 352122) in Monkeys

  Updated version

  http://cancerres.aacrjournals.org/content/46/1/169

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/46/1/169To request permission to re-use all or part of this article, use this link

on April 8, 2021. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from