Introduction

Abstract Introduction Methods Results Discussion References

Irinotecan (CPT-11,² Camptosar, Pharmacia & Upjohn, Kalamazoo, MI) is a carboxylesterase-labile, carbamate prodrug of the antineoplastic topoisomerase I inhibitor SN-38 (1-3). CPT-11 and SN-38 exist in a pH- and protein-dependent equilibrium between active lactone and inactive hydroxy acid anion forms (4). Bioactivation of CPT-11 by microsomal carboxylesterases (E.C. 3.1.1.1) occurs primarily in the liver (5, 6) and has been studied previously using purified human carboxylesterase (7, 8). In this study, an in vitro model for the assessment of CPT-11 bioactivation by carboxylesterases was developed using human liver microsomes. Michaelis-Menten kinetic constants were determined and a new high affinity isoform was discovered.

Since CPT-11 will be used clinically with a variety of other drugs, an in vitro screen for potential drug interactions involving CPT-11 bioactivation was developed.

Some chemicals were chosen based on possible clinical use and have no known carboxylesterase-inhibiting ability (e.g. 5-FU, ethanol (EtOH), caffeine, acetaminophen, coumarin, dexamethasone, cisplatin, and probenecid). Others were chosen based on previously described effects on carboxylesterases (e.g. ASA (9), fluoride (10), HMB, CMB (6), physostigmine (6, 11), bis-nitrophenylphosphate (BNPP) (7), metoclopramide (12), loperamide (8), and prochlorperazine (13)). To avoid misleading conclusions about the possible clinical significance of in vitro experiments conducted at high inhibitor concentrations, statistically significant results were discussed in the context of clinically relevant concentrations of inhibitor.

As a carbamate, CPT-11 is a relatively poor substrate for carboxylesterases. This was proposed to be a result of slow decarbamylation of the serine esteratic site, inferring that CPT-11 is a slowly reversible competitive inhibitor, possibly also with an allosteric inhibitory effect on other substrates at a modulator site of the enzyme (8). Therefore, CPT-11 could act as an inhibitor of endogenous or xenobiotic high turnover substrates of carboxylesterases. To investigate this, CPT-11 was tested as an inhibitor of microsomal carboxylesterase-mediated hydrolysis of the model substrate, -naphthyl acetate (-NA) (14,15).

	Methods

Abstract Introduction Methods Results Discussion References

Reagents and Materials. Irinotecan hydrochloride trihydrate (CPT-11) and SN-38, were supplied by Pharmacia and Upjohn, Inc. Camptothecin and inhibitors were obtained from Sigma or Aldrich (St. Louis, MO). Transplant quality human liver tissue was obtained through the International Institute for the Advancement of Medicine (Exton, PA). Microsomes were prepared by standard methods (16) and stored in 0.25 M sucrose at 70°C. Protein was determined by the Bicinchoninic Acid (BCA) assay (Pierce Chemical, Rockford, IL) and standardized relative to bovine serum albumin.

Kinetics and Inhibitors of CPT-11 Bioactivation. Incubations. Kinetic parameters were determined using CPT-11 that was at equilibrium in pH 7.4 phosphate buffer. Under these conditions the lactone/hydroxy acid anion ratio for CPT-11 was fixed at 87/13 (4). Kinetic studies were done in 2 ml Eppendorf Safe-lock polypropylene tubes (Brinkman, Westbury, NY) tubes using human hepatic microsomes and CPT-11, incubated in a final volume of 0.25 mL sodium phosphate (0.1 M, pH 7.4) at 37°C, with 80 rpm mixing. CPT-11 solutions were pre-incubated for 5 min and incubations were initiated by the addition of pre-warmed microsomes. Reactions were quenched and analyzed for SN-38 content by HPLC as described under sample preparation. All concentrations refer to total (lactone plus hydroxy acid anion) forms of CPT-11 and SN-38. Incubation times (5-60 min) and protein concentrations (0.125-1.8 mg/ml) were tested to identify conditions that resulted in quantifiable and linear formation of SN-38.

Determination of Michaelis-Menten kinetic parameters. Incubations were done in duplicate using microsomal protein (1 mg/ml) and CPT-11 (0.25-200 µM, 17 concentrations), incubated for 15 min. A 50 µM CPT-11 control in phosphate buffer (without microsomes) and a boiled microsome control were run with each experiment. The hydrolysis of CPT-11 in phosphate buffer, in the absence of microsomal protein, and in the presence of boiled microsomal protein, was negligible over the incubation time course; however, the CPT-11 stock solution contained a small amount of SN-38 (approximately 0.04%). Levels of the SN-38 impurity at other substrate concentrations were calculated by linear interpolation of a calibration line passing through the experimentally-determined 50 µM blank data point and the origin. Total SN-38 formation in the presence of microsomes was corrected for the SN-38 impurity at each substrate concentration. Substrate depletion in kinetic studies was maximally 10% at 0.25 µM CPT-11.

Determination of the effect of potential inhibitors on SN-38 formation. Incubations were done in triplicate using pooled human liver microsomes, as described in Incubations, at a CPT-11 concentration of 10 µM. Microsomes (25 µl) were mixed with potential inhibitors (0-25 µl) and phosphate buffer (100-125 µl) and pre-incubated for 5 min. Inhibitor concentrations were  21 µM and were chosen to represent clinically attainable concentrations, with the exception of loperamide, where the 25 µM test concentration was approximately 1000-fold higher than clinically relevant plasma levels. Incubations were initiated by the addition of prewarmed CPT-11 solution (100 µl). In the final incubation, DMSO content was 0.0016% and MeOH content was 0% or 1%. Reactions were quenched and analyzed as described under sample preparation. SN-38 formation was measured as ng/ml/15 min. Inhibition was expressed as per cent of control.

HPLC and analytical standard preparation. A Perkin-Elmer ISS-200 autoinjector and PE 410 pump were used (Perkin Elmer, Norwalk, CT). Chromatography was done on a Zorbax SB-C8 column (4.6 × 250 mm, 5 µm particle size, Mac Mod Analytical Inc., Chadds Ford, PA) with a Brownlee Newguard RP8 guard column using a mobile phase of 74:26, v:v triethylamine buffer:acetonitrile at a flow rate of 1 mL/min. Detection was done using a Waters Model 474 fluorescence detector (Millipore Corp., Milford, MA). Calibration standards (9-10 concentrations) were prepared in duplicate in phosphate buffer to cover a concentration range of approximately 1 to 4000 ng/ml SN-38. Quality control (QC) standards were prepared in duplicate from separately prepared stock solutions at nominal SN-38 concentrations of 23, 76, and 1900 ng/ml. The stability of stock and control solutions in MeOH and in phosphate buffer at 4°C (storage conditions) were proven by HPLC analysis after storage periods of 0, 5, 22, 28, and 41 days. SN-38 recovery from microsomes relative to phosphate buffer was determined to be complete by comparing triplicate incubations of microsome suspensions spiked with 4, 90, 300, and 1900 ng/ml SN-38 with triplicate QC samples prepared identically in phosphate buffer.

Sample preparation. A 250 µl aliquot of incubation sample, calibration standard or QC sample was mixed with 500 µl quench solution (acetonitrile/acetic acid/methanol (95.6/4.0/0.4, w/w/w) containing 24 ng/ml camptothecin (CPT) internal standard). The samples were heated in a water bath at 37°C for 20 min to convert SN-38 to the lactone form. Samples were mixed with 1.0 ml of 50 mM, pH 4.2 triethylamine/acetate buffer. Precipitated protein was separated by centrifugation at 14,000 rpm for 2 min. A 50 µl aliquot of supernatant was analyzed by HPLC. Retention times were sensitive to small changes in mobile phase composition. The detector was programmed to change wavelengths and detector gain. From 0-10 min the excitation and emission wavelengths were 372 and 425, respectively (gain 10, measures CPT-11). From 10-14 min the excitation and emission wavelengths were 372 and 535, respectively (gain 100, measures SN-38). From 14 min onward the excitation and emission wavelengths were 372 and 425, respectively (gain 100, measures CPT).

Data collection and analysis. Data were collected and processed using the UPACS (Upjohn) chromatography system, Version 5.2. Quantitation was done using peak height ratios of the analyte to internal standard. Calibration curves were determined by using linear regression through-the-origin best fit with a 1/concentration weighting factor. Unknown SN-38 concentrations in incubation solutions were determined by comparison of the peak height ratios with the calibration curve. Correction factors for the SN-38 impurity at each concentration/time datapoint were calculated by subtracting background SN-38 from the total SN-38 formed in the presence of microsomes. Unknown concentrations were determined as corrected (carboxylesterase-mediated) ng/ml SN-38 produced in 15 min and were converted to units of velocity (pmol/mg/min) using a formula weight of 410.4 g/Mol for SN-38 monohydrate.

Analytical validation. Standards of CPT-11 in MeOH or in phosphate buffer, prepared in polystyrene centrifuge tubes and stored at 4°C, were stable over a period of 41 days. Because of a limited supply of human microsomes, standards and QC samples prepared in phosphate buffer were used to quantify microsomal concentrations of SN-38. Recoveries of SN-38 from microsomes spiked in triplicate with SN-38 at concentrations of 4, 90, 300, and 1900 ng/ml were approximately 100% in both microsomes and phosphate buffer. Correlation coefficients (r²) for linear standard curves in kinetic assays were  0.998. QC data were typically within ± 15%. The assay lower limit of quantitation (LLOQ) was 1.1-2.4 ng/ml SN-38.

Curve fitting methods. K_M and V_max were determined by nonlinear least squares regression using Systat Version 5.2.1 for Macintosh (Version 7.0). SN-38 formation V versus [S] curves were fit for a two enzyme system with 1/V² weighting. -NA hydrolysis V versus [S] curves were best fit by a single enzyme Michaelis-Menten equation.

Determination of inhibitor potency. Data were analyzed using Microsoft Excel Version 4.0a (Microsoft Corp., Redmond, WA). Means and standard deviations of uncorrected SN-38 production in the presence and absence of inhibitors were compared with a Student's t test. Mean SN-38 production was corrected for mean SN-38 impurity measured in the presence of heat denatured (boiled) microsomes and SN-38 production was expressed as per cent of enzyme-catalyzed control. When MeOH was necessary as an aid to solution for inhibitors, an additional control containing an equal amount of MeOH was run. The carboxylesterase inhibitor BNPP and a phosphate buffer control (no protein) were used as positive control and blank, respectively.

	Kinetics of -naphthyl acetate hydrolysis and inhibition by CPT-11

An assay for the carboxylesterase-mediated formation of -naphthol from -NA was developed based on the formation of an azo dye from the reaction of fast blue RR (a diazonium salt, Sigma, St. Louis, MO) with -naphthol (17). Enzyme kinetics were determined using modified literature methods (14, 15) on a THERMOmax Microplate Reader (Molecular Devices, Sunnyvale, CA) (18). Optical Density (O.D.) was measured at 450 nm. Initial velocities were recorded in units of mO.D./min and were corrected for background hydrolysis in buffer, determined simultaneously. CPT-11 (80-440 µM) was tested as an inhibitor of -NA hydrolysis. Control carboxylesterase-mediated hydrolysis was determined on the same plate, simultaneous with each inhibitor concentration, since preincubation effected the control kinetics. Results were expressed as per cent of control sample initial velocity (mO.D./min).

Reagent Preparation. BNPP and NaF were prepared in 0.1 M, pH 7.4 phosphate buffer. Stock solutions of -naphthol for standard curves were prepared in acetone. Final acetone concentration in the standard curve wells was 3.3%. Stock CPT-11 solutions were prepared at 360 µM in phosphate/MeOH or at 80, 140, and 440 µM in phosphate/dimethylsulfoxide (DMSO) at 90/10 w/w. The co-solvent concentration in each incubation well was 0.33%. Controls containing an equal amount of co-solvent were prepared. Stock -NA solutions were prepared in acetone/phosphate at 50/50 (w/w). The final acetone concentration of 1.7% in all wells does not effect enzyme activity (14). Fast Blue RR was prepared fresh in phosphate buffer at 0.25% v/v for each experiment and was filtered prior to use.

Preliminary Experiments. Low protein concentrations and short reaction times at 23°C and substrate concentrations of  40 µM were needed to avoid substrate depletion in excess of 15-20%.

Incubation Conditions. In a 96 well microplate, each well contained 260 µl buffer, 20 µl microsomes (0.075 mg/mL) or buffer, 10 µl fast blue RR (0.25 mg/ml), and 10 µl substrate or buffer. The total volume was 300 µl. Substrate was added last using a multi-channel pipette. Immediate automatic mixing was employed and incubations were conducted with continuous assessment of optical density within the thermostatted compartment of the plate reader. Preincubations with inhibitors were done inside the plate reader.

Kinetics of -Naphthyl Acetate Hydrolysis. Substrate concentrations of 0, 5, 10, 20, 40, 60, 80, 100, 200, 400, 600, and 800 µM (12 concentrations/lane) were incubated in quadruplicate on the same plate with a matching quadruplicate set of enzyme-free controls. Substrate depletion, calculated from the -naphthol standard curve, was maximally 17% at 2 min at 60 µM -NA for donor 20F, and was 13% or less in all other donors at -NA  40 µM. Data were fit to a single enzyme Michaelis Menten equation using substrate concentrations of 40-60 µM to 600-800 µM. V_max data were compared to V_max data for low and high K_M isoforms of the CPT-11 hydrolyzing carboxylesterase by linear regression comparing donors 10M, 24M, 17F, and 20F.

Inhibition by CPT-11. Microsomes from donor 17 were preincubated with CPT-11 (360 µM), or BNPP (360 µM), or NaF (500 µM) for 10 min before the addition of -NA (0-120 µM, 8 concentrations/plate lane) to start the reaction. The experiment was repeated with CPT-11 (80, 140, and 440 µM) preincubated for 1 h at 23°C over -NA concentrations of 0-800 µM.

	Results

Abstract Introduction Methods Results Discussion References

Hydrolysis Kinetics of CPT-11 in the Presence of Human Liver Microsomal Carboxylesterases

Incubation Conditions. In pilot studies, heat denaturation (boiled microsomes) and the organophosphate esterase inhibitor, BNPP reduced SN-38 production by human hepatic microsomes to zero, providing evidence for the involvement of esterase enzymes. An optimal protein concentration of 1 mg/ml and incubation time of 15 min were chosen from the linear regions of SN-38 production. These conditions were necessary to surpass the assay limit of quantitation at low substrate concentrations and to adequately differentiate enzyme-catalyzed SN-38 formation from the SN-38 impurity.

Kinetic Results for SN-38 Formation. HPLC chromatograms showing the effect of CPT-11 concentration on the formation of SN-38 by microsomes are shown in fig. 1. In microsomes from seven different liver donors, Eadie Hofstee plots were clearly biphasic over 0.25-200 µM CPT-11 (data not shown). In accord with the linear plots, V versus [S] data, modeled using nonlinear least squares regression analysis, were best fit by a two enzyme equation. K_M and V_max estimates are presented in table 1. A representative Eadie Hofstee plot for the pooled microsomes, comparing fitted with actual data is shown in fig. 2. The actual and theoretical curves were superimposed. The calculated K_M and V_max for each carboxylesterase isoform are individually represented in the same figure by the calculated linear plots for each isoform.

View larger version (26K): [in this window] [in a new window]   Fig. 1.   Superimposed HPLC chromatograms showing SN-38 formed by donor 10M microsomes (1 mg/mL protein, 15 min) at CPT-11 concentrations of 0.5 (lower trace), 5 (second trace), 25 (third trace), and 100 µM (upper trace).

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Eadie Hofstee plot generated following nonlinear curve fitting of V versus [S] data for pooled microsomes to a 2 enzyme Michaelis-Menten equation. Linear interpolations for each isoform, based on nonlinear KM and Vmax estimates, illustrate graphically the correct intercept and slope for the individual isoforms. KM and Vmax estimates are in table 1.

View larger version (25K): [in this window] [in a new window]   Fig. 3.   Histogram showing the calculated relative contributions of each esterase isoform to the total production of SN-38 at increasing concentrations of CPT-11. At clinically relevant hepatic concentrations of below 20 µM, the high affinity, low KM isoform is the dominant contributor to CPT-11 bioactivation.

Effects of Potential Inhibitors of CPT-11 Bioactivation

Results of four separate inhibition experiments are compared with their respective control values in table 2. Inhibition was expressed as per cent of control. Controls containing an equal amount of 1% MeOH were run when MeOH was used to dissolve inhibitors. MeOH decreased activity by 3.5, 6.6, and 18.2% in experiments 2, 3, and 4, respectively. Organic solvent effects on carboxylesterase activity have been documented previously (19).

As expected, BNPP at both 25 and 100 µM reduced CPT-11 hydrolysis to 1.3-1.6% of control values. Physostigmine, a known inhibitor of butyrylcholinesterases (E.C. 3.1.1.8, closely related to carboxylesterases) showed significant (77% decrease, (P < 0.05)) inhibition of CPT-11 hydrolysis, in accord with previous data (6, 11). Probenecid (25 µM) showed a significant (13%, P < 0.01) inhibition of CPT-11 hydrolysis, while 5-FU (25 µM) had no significant inhibitory effect (2%, P < 0.1). Sodium fluoride (NaF), a known carboxylesterase inhibitor (10, 20) inhibited hydrolysis significantly (P < 0.01) by 9% at 25 µM, 14% at 100 µM, and 35% at 481 µM.

Aggressive therapy with loperamide is used to treat CPT-11-induced delayed diarrhea (21). A small decrease (7.5%) in SN-38 formation owing to loperamide was observed; however, the difference was not significantly different from control values. Ethanol at 25 and 100 µM had no significant effect on CPT-11 hydrolysis. Two mercuribenzoate inhibitors of A-esterases (HMB and CMB) also had no significant effect. Caffeine, ASA, and coumarin all showed no significant effect on CPT-11 hydrolysis. Small, but statistically significant, decreases were observed with metoclopramide (9.1% decrease, P < 0.01) and acetaminophen (7% decrease, P < 0.05). Dexamethasone, cisplatin, and prochlorperazine were without significant effect with 3.4%, 5.6%, and 7.0% differences from control values, respectively.

Effect of CPT-11 on -Naphthyl Acetate, a High Turnover Substrate of Human Liver Carboxylesterase

Fig. 4 shows initial velocity versus substrate concentration plots over 40-800 µM for microsomes from five donors. Data were fit by nonlinear regression to a single enzyme Michaelis Menten equation giving corrected correlation coefficient (R²) values of 0.94-0.98. K_M estimates for -NA hydrolysis were 115 ± 20.8 µM (range 90.6-146 µM, CV = 18%). V_max was estimated at 1.64 ± 0.45 µMol/mg/min (range 1.1-2.6 µMol/mg/min, CV = 27.3%). Linear regression (forced through the origin) of V_max for -NA hydrolysis versus V_max for low and high K_M isoforms of the CPT-11 hydrolase afforded correlation coefficients of 0.94 with the CPT-11 high K_M isoform and a lower correlation coefficient of 0.82 for the CPT-11 low K_M isoform in five donors. The narrow range of V_max for the CPT-11 low K_M isoform and the limited number of donors (N = 4, range 2.41-2.61), may be responsible for the lower correlation coefficient. A larger sample size would be needed to conclude that CPT-11 and -NA carboxylesterase activities are correlated.

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Effect of -naphthyl acetate concentration (40-800 µM) on the initial velocity of human liver microsome catalyzed, carboxylesterase-mediated formation of -naphthol. Data shown are mean ± SD for simultaneous quadruplicate determinations on a 96 well microplate. Results for each well are corrected for any measurable background hydrolysis, determined simultaneously in enzyme-free phosphate buffer. Mean data in mO. D./min were converted using a standard curve to units of pmol/mg/min for calculation of KM and Vmax.

At -NA concentrations of 0 to 100 µM, hydrolysis was almost completely inhibited by positive control carboxylesterase inhibitors, BNPP (360 µM) and NaF (500 µM), whereas CPT-11 preincubated for 10 min at 360 µM inhibited -NA hydrolysis by only 9 and 20% relative to control values, at concentrations of -NA of 50 and 100 µM (fig. 5). In a second experiment, CPT-11 preincubated for 1 h at 80, 140, or 440 µM inhibited -NA hydrolysis to 69-89% (mean 80 ± 6.6%, N = 9) of control values, at concentrations of -NA of 200, 400, and 800 µM (data not shown). It is clear from these data that inhibition of -NA hydrolysis by CPT-11 was maximal at approximately 80% of control values at the lowest CPT-11 concentration tested (80 µM), and therefore this concentration was well above the K_i for CPT-11-sensitive carboxylesterases. Thus, CPT-11-sensitive carboxylesterases may only account for 20% of the turnover of the model substrate, -NA in vitro.

View larger version (36K): [in this window] [in a new window]   Fig. 5.   Effect of CPT-11 on the carboxylesterase-mediated hydrolysis of -naphthyl acetate. Control experiments with a MeOH cosolvent were done for direct comparison with CPT-11 incubations that required methanol as an aid to solution. The control incubation for NaF and BNPP treatments was free of cosolvent. In a separate experiment (not shown), CPT-11 inhibited -naphthyl acetate hydrolysis by only 20% (80 ± 6.6% of control, N = 9). The remaining 80% of control activity may be attributed to CPT-11-insensitive carboxylesterase isoforms.

	Discussion

Abstract Introduction Methods Results Discussion References

CPT-11 Bioactivation

The Enzymology of Carboxylesterases. To develop a clinically-relevant perspective on the bioactivation of CPT-11, and to understand the implications of carboxylesterase inhibition on antineoplastic activity, it is necessary to understand the basic chemistry of the enzyme class. The mechanism of hydrolysis by serine hydrolases such as the carboxylesterases is well understood from kinetic studies on substrates and inhibitors (19). However, the molecular biology of carboxylesterase enzymes is only now being elucidated, in part because of technical difficulties inherent in subtle differences in sequence, structure, and substrate and inhibitor specificity (22, 23). Two highly similar human liver microsomal carboxylesterases were recently sequenced and expressed by Kroetz et al. (22), and these authors conclude, based on substrate diversity, that additional carboxylesterases must exist. Ketterman et al. have purified and characterized mid and low Pi human hepatic carboxylesterases and have proposed, based on kinetic comparison of different donors, that polymorphism occurs within each purified carboxylesterase (24, 25). Therefore, it is likely that more than one carboxylesterase isoform may contribute to CPT-11 bioactivation.

CPT-11 Hydrolysis and Antitumor Activity. Although CPT-11 has weak antineoplastic activity of its own, a 10⁴-fold increase in cytotoxicity in vitro is realized when SN-38 is released by carboxylesterases in rat serum (26). The conversion of CPT-11 to SN-38 has been studied in a wide variety of tissues, cell lines, and purified enzyme preparations in vitro (5-8, 11, 27). The sensitivity of proliferating tissues or cell lines to the cytotoxic effects of CPT-11 may be related their carboxylesterase levels (27). Some studies indicate that tumor levels of carboxylesterases are reduced relative to peritumoral normal tissue (28).

A New CPT-11 Hydrolase. Our observation of biphasic CPT-11 kinetics in human liver microsomes suggests at least two carboxylesterase isoforms are present in our human microsomal preparation. We have observed the low K_M isoform by studying much lower substrate concentrations (250 nM-200 µM) relative to the concentrations used by others (7, 8). Satoh et al. studied a purified enzyme at 25-100 µM CPT-11 giving K_M = 169 µM and V_max = 169 pmol/mg/min by Lineweaver-Burk analysis of the linear portion of the plot. Rivory et al. studied the mid Pi purified enzyme at 8-150 µM CPT-11, observing a K_M of 52.9 µM by nonlinear regression analysis. Our data fit a two enzyme MM equation extremely well, and we speculate that the characterization of two or more carboxylesterase isoforms by Kroetz et al. (22) and Ketterman et al. (24, 25) may be relevant to the kinetic parameters determined in this study.

Lactone/hydroxy Acid Anion Equilibria in Relation to CPT-11 Bioactivation. The lactone form of CPT-11 is stable under acidic, but not neutral, conditions with a half life in pH 7.4 phosphate buffered saline of 25.6 min. At equilibrium, the lactone/hydroxy acid anion ratio is known to be 13/87 (4). Haaz et al. have recently shown that there is an approximately 2-fold greater hydrolysis of the lactone form of CPT-11 relative to the ring-opened hydroxy acid anion form (35). Equilibrium conditions for our kinetic studies were chosen for several reasons: First, the lactone concentration would not be changing rapidly and enzyme activity would be maximal during the incubation at pH 7.4. Second, because the lactone forms of CPT-11 and SN-38 are preferentially bound to and stabilized by serum albumin, the free fraction in plasma may be dominated by the hydroxy acid anion form. Third, the hydroxy acid anion form may hypothetically be more available for extraction by the liver, as shown by the short half life of SN-38 hydroxy acid anion when given intravenously to animals as the soluble sodium salt (36).

Effect of Potential Inhibitors of CPT-11 Bioactivation

Loperamide (24 µM) reduced SN-38 formation from 10 µM CPT-11 by approximately 7%. This difference from control was not statistically significant. Rivory et al. (8), using a purified enzyme, have shown that high concentrations of loperamide (200 µM) inhibits p-nitrophenylacetate hydrolysis, lowering V_max from 60.0 to 41.9 nMol/min and almost doubling the K_M (240 to 400 µM). Similarly, 50 µM loperamide reduced SN-38 production by approximately 50% at high CPT-11 concentrations (8). We estimate, based on single dose literature data, that when loperamide is given at 2 mg every 2 hr, that loperamide steady state plasma concentrations are approximately 34 nM in humans (37). We speculate that hepatic concentrations at this dose intensity will be much lower than the 24 µM incubation concentrations used in our study. On the basis of these data, we conclude that, even if administered at high doses at the same time as CPT-11, the likelihood of a clinically significant effect of loperamide on SN-38 production is negligible.

Sodium fluoride (NaF), a known carboxylesterase inhibitor (10, 20), inhibited hydrolysis by 9% at 25 µM. This is a relatively high NaF concentration. Fluoride, which is formed from the breakdown of the 5-FU metabolite, fluoro--alanine (FBAL) (38), could potentially decrease SN-38 AUC during co-therapy with 5-FU and CPT-11 (39). In a separate study, at 5-FU doses ranging between 8 and 30 mg/kg, free fluoride was observed in plasma in about one-fourth of patient samples that were examined (median 5 µM, maximum 33 µM) (38). The low fluoride concentrations observed clinically and the extent of in vitro inhibition observed at 25 µM NaF in the present study indicate that 5-FU-derived fluoride is not likely to result in clinically significant decreases in metabolism of CPT-11 to SN-38. A recent clinical report shows that co-administration of 5-FU does not significantly effect the conversion of CPT-11 to SN-38 (40).

Relative to the potent inhibitors, BNPP and physostigmine, which are not expected to be given with CPT-11, inhibition of SN-38 formation by 5-10% in vitro by high concentrations of inhibitor is probably not clinically relevant. Furthermore, any subtle changes would be irrelevant considering the high interpatient variability normally observed in CPT-11 pharmacokinetics (41).

Co-administration of any known potent carboxylesterase inhibitor with CPT-11 is probably rare. However, if a known therapeutic carboxylesterase inhibitor, such as physostigmine or echothiophate, can be safely withheld, staggering co-therapy with CPT-11 would be the prudent and cost effective way to ensure adequate production of SN-38 in patients needing both medications.

-Naphthyl Acetate Hydrolysis: Effect of CPT-11 on a High Turnover Substrate of Human Liver Carboxylesterases

The kinetics of -NA hydrolysis were determined. V _max data were about 50% lower than those determined by Huang et al. (3.89 ± 0.20 µMol/mg/min versus 1.64 µMol/mg/min) (14). In this study V_max varied from 1.1-2.6 µMol/mg/min. Data from four donors demonstrated a correlation between -NA V_max and the V_max of the high K_M component of CPT-11 hydrolysis. The kinetic difference between ester and carbamate substrates is apparent in the 10⁶-fold higher V_max for -NA hydrolysis relative to CPT-11 hydrolysis.

In one liver sample, CPT-11 inhibited the carboxylesterase-mediated hydrolysis of -NA by only 20%. Compared with BNPP and NaF, CPT-11 was therefore a relatively weak inhibitor of -NA hydrolysis. We did not examine low enough CPT-11 concentrations to establish a K_i (presumably this would equal the K_M). Since 80% of the hydrolysis of -NA continued unabated in the presence of high concentrations of CPT-11, we speculate that nonselective substrates of carboxylesterases will not be appreciably inhibited by CPT-11 in vivo, owing to the contributions of isoforms not affected by CPT-11.

Conclusions

We have characterized a new low K_M human carboxylesterase that we propose to be the major isoform responsible for the bioactivation of CPT-11 in vivo. The CPT-11-sensitive microsomal carboxylesterases characterized in this paper are inhibited by potent carboxylesterase inhibitors, but are inhibited to only an insignificant extent by high concentrations of some selected agents that are commonly coadministered with CPT-11. CPT-11 inhibits the carboxylesterase-mediated turnover of the model high turnover ester substrate -NA, but only by 20% even at the highest CPT-11 concentrations tested. We speculate that, since there is ample remaining activity related to carboxylesterase isoforms that are not sensitive to CPT-11, CPT-11 should not significantly inhibit the hydrolysis of high turnover, nonselective ester substrates of carboxylesterases.