Influence of Anticonvulsants on the Metabolism and Elimination of Irinotecan

ABSTRACT: The hepatic metabolism and biliary secretion of irinotecan (CPT-11, Camptosar) and metabolites is complex and involves cytochrome P450 isoenzymes, carboxylesterases, glucuronosyltransferase, and the ATP-dependent export pumps MRP-2 and MXR. Enzyme-inducing antiepileptic drugs (EIAEDs) such as phenytoin and carbamazepine are known to induce several of the metabolic pathways relevant to ininotecan’s elimination. The North American Brain Tumor Consortium phase I study is designed to determine the maximum tolerated dose and pharmacokinetics of irinotecan given every 3 weeks to patients who are receiving EIAEDs. The EIAEDs have altered both the pharmacokinetics and pharmacodynamics of irinotecan. Peak concentrations and the area under the plasma-time curves for both irinotecan and SN-38 were significantly decreased in patients receiving EIAEDs. The recommended phase II dose of irinotecan administered every 3 weeks is 750 mg/m² for patients who have been receiving stable doses of EIAEDs. [ONCOLOGY 16(Suppl 7):33-40, 2002]

Irinotecan (CPT-11,Camptosar) is a semisynthetic, water-soluble camptothecin derivative approvedfor the treatment of patients with colorectal cancer. Irinotecan is alsoundergoing phase I/II clinical testing and has shown promising activity againsta wide variety of tumor types, including recurrent malignant gliomas. A phase IItrial conducted at Duke University in patients with recurrent high-grade gliomareceiving weekly irinotecan doses of 125 mg/m² reported a 15% objective responserate.[1] Of note, the toxicity and plasma exposure to irinotecan and its potentmetabolite SN-38 appeared to be lower than that of a historical databasecomprising colorectal cancer patients.

Intravenously administered irinotecan (molecular weight [MW]: 587) ismetabolized in the liver by cytochrome P450 3A4/5 (CYP 3A4/5) enzymes to theless active oxidative metabolites APC—7-ethyl-10-[4-N-(5-aminopentanoicacid)-1-piperi-dino]-carbonyloxycamptothecin (MW: 619)—and NPC—7-ethyl-10[4-(piperidino)-1amino]carbonyl-oxycamptothecin (MW: 519).[2,3] Irinotecan is also bioactivatedin the liver by carboxylesterases (predominantly hCE2) to the potenttopoisomerase I inhibitor SN-38 (MW: 393)[4-6] (Figure1).

SN-38 is eliminated mainly through conjugation by hepatic uridine-diphosphoglucuronosyltransferase (UGT-1A1) to SN-38 glucuronide (SN-38G)[7] (Figure2). SN-38G hasonly 1/100 the antitumor activity of SN-38. UGT-1A1 is the same isoenzymeresponsible for glucuronidation of bilirubin.[8] Grade 4 toxicity(neutropenia/diarrhea) has been reported in patients with deficient UGT-1A1activity (Gilbert’s syndrome) after receiving irinotecan.[9] The relationshipof genetic polymorphism of UGT-1A1 and irinotecan toxicity was recentlyreported.[10] Patients who are either homozygous or heterozygous for theUGT-1A1*28 genotype are at risk for severe irinotecan-associated toxicity.

The canalicular multispecific organic anion transporter (MRP-2) reportedly isresponsible for the biliary transport of the carboxylate forms of irinotecan andSN-38, as well as the lactone and carboxylate forms of SN-38G.[11] MRP-2 ispresent in apical membranes of the liver, kidney, and intestine, and isexpressed in human lung, gastric, and colorectal cancer cells.[12] MRP-2functions as a conjugate export pump mediating the unidirectional transport ofbilirubin glucuronides, reduced folate and amphiphilic anions, particularlylipophilic substances conjugated with glutathione, glucuronide or sulfide.[13]In the rat, a single nucleotide deletion results in a frame shift, reduced mRNAlevels, and absence of the protein.[14] The phenotype found in humans with thedeletion is termed Dubin-Johnson syndrome, an autosomal recessive disordercharacterized by chronic hyperbilirubinemia.[15] Reports also suggest that SN-38may be a substrate for the mitoxantrone (Novantrone)-resistance half-transporter(MXR) protein.[16,17] The effect of MXR on hepatic disposition of SN-38 has yetto be determined.

Pretreatment of rats with phenobarbital, an inducer of UGT-1A1 and CYP 3A4,was shown to enhance formation of SN-38G and diminish the area under theconcentration-time curve (AUC) for both SN-38 and irinotecan.[18] However,valproic acid increased the AUC of SN-38. Additionally, phenobarbital combinedwith irinotecan successfully treated a patient with Gilbert’s syndrome.[19]Recent studies have demonstrated that patients receiving enzyme-inducingantiepileptic drugs (EIAEDs) obtain lower plasma concentrations ofchemotherapeutic agents when administered at conventional doses.[20,21] Failureto achieve adequate plasma drug levels may account in part for lack ofchemotherapy efficacy demonstrated in previous brain tumor trials.

The trial reported herein (NABTC-9801) is being conducted to determine thedose-limiting toxicities, maximum tolerated dose, and antitumor activity ofirinotecan administered every 3 weeks to patients with progressive or recurrentmalignant glioma receiving EIAEDs (phase I portion). In the phase II portion ofthe study, the antitumor activity of irinotecan including response rate, time totumor progression, and 6-month progression-free survival (primary end point)will be determined. The influence of anticonvulsants on the metabolism andelimination of irinotecan and its metabolites is also to be characterized, andthis is the main focus of this report.

Patients and Methods

Patient Eligibility

Eligible patients are consenting adults (minimum age, 18 years; Karnofskyperformance status ³ 60%) with histologically documented malignant glioma andradiographically (CT or MRI scan) measurable recurrent disease after radiationtherapy. In the phase I portion, patients could not have received more than twoprevious chemotherapy regimens. For the phase II portion, no more than oneprevious chemotherapy regimen was allowed, either as adjuvant therapy or forrecurrent disease. Patients taking corticosteroids must have been on stable ordecreasing doses for 72 hours prior to study entry, with no dose escalationsover the entry dose. An interval before enrollment of at least 2 weeks fromprevious resection or 4 weeks from previous radiation therapy or chemotherapywas required. Nitrosourea, suramin, or mitomycin (Mutamycin) could not have beenadministered for at least 6 weeks, and no previous topoisomerase I inhibitortreatment was allowed. Other criteria included adequate hematologic (absoluteneutrophil count > 1,500/µL, platelets ³100,000/µL), renal (SCr ≤ 1.5 mg/dL),and hepatic functions (bilirubin £ 1.5 mg/dL).

Phase IStratification and Dose Escalation

Patients were stratified into two groups according to their pretreatmentanticonvulsant medications. Patients not receiving EIAEDs with or withoutsteroids (group A) received irinotecan at the current maximum tolerated dose of350 mg/m² IV over 90 minutes repeated every 3 weeks (Table1). No doseescalations were permitted. Pharmacokinetic, safety, and efficacy data werecollected from group A patients. Patients receiving EIAEDs with or withoutsteroids (group B) received irinotecan at a starting dose of 350 mg/m².Irinotecan dose was escalated in 50-mg/m² increments in cohorts of threepatients at each dose level until the maximum tolerated dose was established.

Dose-LimitingToxicity and Maximum Tolerated Dose in Group B

Toxicities were graded according to the National Cancer Institute (NCI)Common Toxicity Criteria scale. Dose-limiting toxicity was defined as any of thefollowing events occurring during the first treatment course: (1) grade 4neutropenia lasting ³ 5 days, neutropenic fever or infection, or grade 4thrombocytopenia; (2) delayed diarrhea ³ grade 3 despite loperamide support, ornausea or vomiting ³ grade 3 despite 5HT₃ coverage; or (3) nonhematologictoxicity ³ grade 3 attributable to irinotecan. The maximum tolerated dose wasdefined as the dose level at which zero out of three or one out of six patientsexperienced a dose-limiting toxicity, with two out of three or two out of sixpatients at the next higher dose level encountering a dose-limiting toxicity.

PharmacokineticSample Collection and Analysis

Heparinized blood samples (7 mL) were drawn via venipuncture or through anindwelling IV heparin lock. One milliliter of blood was withdrawn for theheparin lock and discarded prior to sample collection at the following times:prior to drug administration (baseline), 45 minutes into the infusion, and atthe end of the infusion, and then 15, 30, 60, and 90 minutes and 2, 3, 4, 6, 8,10, and 24 hours after the end of infusion on day 1 of the first treatmentcycle. Blood samples were immediately centrifuged and plasma was removed andfrozen (£ 20_oC) for subsequent high-performance liquid chromatography (HPLC)analysis for total concentrations of irinotecan and SN-38, as well as SN-38G andAPC concentrations. A validated, sensitive, and specific isocratic HPLC methodwas used.[22] Samples were stored frozen for no more than 1 year. The long-termstability of total irinotecan and SN-38 in plasma has been documented to be atleast 2 years when samples were stored at £ -20_oC.

Irinotecan, APC, SN-38, and SN-38G plasma concentrations were analyzed bynoncompartmental methods. The time intervals relative to the start of theirinotecan infusion and the actual sample time were used to calculate time topeak concentration (t_max) and the area under the plasma concentration-timecurves. Peak concentrations (Cp_max) were determined by inspecting each patient’splasma concentration-time curve. Elimination rate constants were estimated bylinear regression of the last two data points on the terminal log linear portionof the concentration-time curve. Terminal half-lives (t_1/2) were calculated bydividing 0.693 by the elimination rate constants.

The AUC was calculated using the linear trapezoidal rule up to the lastmeasurable data point (for AUC_0-24), then extrapolated to infinity (AUC). Thesystemic clearance for irinotecan was determined by dividing the dose (inmilligrams free base of irinotecan per meter squared) by the AUC. A metabolicratio, estimated as the ratio of AUC_SN-38 or AUC_SN-38 +AUC_SN-38G to AUC_CPT-11,was used as a measure of the relative extent of the conversion of irinotecan toSN-38. The relative extent of metabolism of irinotecan to APC was estimated asAUC_APC/AUC_CPT-11, and the relative extent of glucuronidation of SN-38 as theratio of AUC_SN-38 to AUC_SN-38G.

Statistical Considerations

Pharmacokinetic parameters for all compounds are reported as mean values ±standard deviations. The relationships between the administered irinotecan dose(mg/m²) and AUCs were analyzed by Spearman’s correlation coefficient andlinear-regression analysis. Differences in pharmacokinetic parameters betweenthe non-EIAED and the EIAED groups were evaluated using an unpaired two-sample ttest. Probability values of less than .05 were considered statisticallysignificant.

Results

A total of 102 patients have been enrolled in this ongoing study, of whom 73had the histologic diagnosis of glioblastoma multiforme and 29 had anaplasticastrocytoma. Thirty-four patients were assigned to group A (non-EIAEDs); thetrial remains open to group A patients with anaplastic astrocytoma. Among the 68group B (EIAEDs) patients, 48 participated in the phase I portion and 20 in thephase II evaluation. The phase I dose-limiting toxicity was grade 3 delayeddiarrhea at the 800-mg/m² dose level. The recommended phase II dose for patientsreceiving EIAEDs was 750 mg/m².

Seventy-two patients were accrued for pharmacokinetic analyses in the firsttreatment course. Pharmacokinetic parameters were characterized for 56 patients:22 group A patients treated with 350 mg/m² of irinotecan and 34 group B patientstreated with irinotecan doses ranging from 350 to 800 mg/m². (Sixteen patientswere nonevaluable for pharmacokinetic analysis due to incomplete samplecollections.) A plasma concentration-time curve generated from two group Bpatients treated at the 350 mg/m² dose level, each receiving dexamethasone andphenytoin, compared with two representative group A patients at the 350 mg/m²dose level, receiving only dexamethasone, is displayed in Figure 3. Systemiclevels of irinotecan, APC, SN38-G, and SN-38 were all lower in group B than ingroup A patients.

Pharmacokinetic parameters of irinotecan and its metabolites are summarizedin Table 2. For comparison purposes, Table 2 also includes pharmacokineticparameters from another trial that included six patients with non-centralnervous system (CNS) malignancies who received 340 mg/m² irinotecan in the sameevery-3-week schedule.[23] Both trials used the same analytical procedure, butthe sampling schedule of the current study extended only to 24 hours comparedwith 48 hours in the comparator trial. Differences in sampling times may accountfor the minor differences in half-lives, clearance, and AUC values. Overall,kinetic parameters in patients receiving no anticonvulsants or steroids in ourtrial are similar to those of non-CNS malignancy patients.

Figure 4 shows the individual and overall mean irinotecan clearance values.Irinotecan clearance in the EIAED group (29.3 ± 7.11 L/h/m², n = 34) wasincreased 1.6-fold compared with that of the non-EIAED group (18.2 L/h/m², n =22). Antiepileptic agents the patients were receiving and their influence onirinotecan clearance are shown in Table 3. The non-EIAED drugs(Table 2) had asmall but statistically significant effect on irinotecan clearance (20.1 ± 5.77L/h/m²) as compared with clearance values (14.2 ± 4.12 L/h/m²) in patients notreceiving steroids or antiepileptic agents. In four patients receiving valproicacid in combination with a non-EIAED, irinotecan clearance was not decreased,nor were the AUCs for SN-38 and SN-38G increased (data not shown).

The EIAEDs being taken by group B patients were empirically categorized intofour groups (phenytoin, carbamazepine, phenobarbital, and combination EIAEDs)for comparison. The majority of patients were receiving phenytoin with orwithout dexamethasone. Excluding the combination EIAED group, irinotecanclearance values were highest among patients receiving phenytoin orcarbamazepine (30.1 and 28.1 L/h/m², respectively), and were lowest among thosereceiving phenobarbital (23.6 L/h/m²).

Figure 5 shows the relationship between irinotecan dose and AUC over thedosage range of 350 to 800 mg/m² in group B patients. There was a moderate togood relationship between irinotecan dose and irinotecan AUC (r² = 0.61), only afair relationship between irinotecan dose and APC AUC (r² = 0.35), and norelationship between irinotecan dose and SN-38 AUC (r² = 0.04).

Table 4 summarizes the relative metabolic ratios in this trial compared withthose observed in the Duke University glioma trial,[1] in which irinotecan 125mg/m² was administered weekly, and in the Mayo Clinic phase I trial,[23] inwhich irinotecan was administered every 3 weeks and dose escalated from 240 to340 mg/m² in patients with non-CNS malignancies. APC was not measured in theDuke University and Mayo Clinic trials. The low ratio of SN-38 to irinotecan inthe EIAED vs the non-EIAED group in our trial and the Duke University trialsuggests a decreased conversion of irinotecan to SN-38, consistent with the1.9-fold decrease in irinotecan AUC. The comparatively low ratio of SN-38 toSN-38G in the EIAED group indicates an increase in glucuronidation of SN-38.Furthermore, the higher ratio of APC to irinotecan in the EIAED compared withthe non-EIAED group suggests induction of CYP 3A4 activity, promoting metabolismof irinotecan to its predominant metabolite APC, and also potentially NPC, whichwas not measured in the current trial.

Summary

The enzyme-inducing antiepileptic drugs alter both the pharmacokinetics andpharmacodynamics of irinotecan. The recommended phase II dose of irinotecan whenadministered every 3 weeks is 750 mg/m² for patients who have been receivingstable (³ 2 weeks) doses of EIAEDs. Compared to findings in the non-EIAEDpatient group and historic controls,[23] the peak concentration and AUC ofirinotecan and SN-38 were significantly decreased in EIAED-treated patients.Although not statistically significant, a quantitative decrease in peakconcentrations and AUCs for APC and SN-38G was also observed.

Overall, irinotecan clearance was 1.6-fold higher in the EIAED group than inthe non-EIAED group. A moderate-to-fair relationship over the dosage range of350 to 800 mg/m² was demonstrated between irinotecan dose and the AUCs foririnotecan and APC, but no relationship was shown between irinotecan dose andthe AUCs for SN-38 or SN-38G. The reportedly non-EIAEDs, especially gabapentin,appear to have a small but statistically significant effect on irinotecanclearance. The relative influence of individual anticonvulsant medications onirinotecan pharmacokinetics in glioma patients is being evaluated in a largerdatabase derived from four NCI-sponsored trials (Duke University, The NorthAmerican Brain Tumor Consortium [NABTC], The New Approaches to Brain TumorTherapy [NABTT] CNS Consortium, and The North Central Cancer Treatment Group [NCCTG]).

References:

1. Friedman HS, Petros WP, Friedman AH, et al: Irinotecan therapy in adultswith recurrent or progressive malignant glioma. J Clin Oncol 17:1516-1525, 1999.

2. Haaz M-C, Riché C, Rivory LP, et al: Biosynthesis of an aminopiperidinometabolite of irinotecan (7-ethyl-10-[-4-(1-piperidone)-1 piperidino]carbonyloxycamptothecine) by human hepatic microsomes. Drug Metab Dispos26:769-774, 1988.

3. Santos A, Zanetta S, Cresteil T, et al: Metabolism of irinotecan (CPT-11)by CYP 3A4 and CYP3A5 in humans. Clin Cancer Res 6:2012-2020, 2000.

4. Slatter JG, Sue P, Sams JP, et al: Bioactivation of the anticancer agentCPT-11 to SN-38 by human hepatic microsomal carboxylesterases and the in vitroassessment of potential drug interactions. Drug Metab Dispos 25:1157-1164, 1997.

5. Humerickhouse R, Lohrbach K, Li L, et al: Characterization of CPT-11hydrolysis by human liver carboxylesterase isoforms hCE-1 and hCE-2. Cancer Res60:1189-1192, 2000.

6. Khanna R, Morton CL, Dank MK, et al: Proficient metabolism of irinotecanby a human intestinal carboxylesterase. Cancer Res 60:4725-4728, 2000.

7. Iyer L, King CD, Roy SK, et al: Genetic predisposition to the metabolismof irinotecan: Role of UGT-1A1 in the glucuronidation of its active metabolite(SN-38) in human liver microsomes. J Clin Invest 101:847-854, 1998.

8. Iyer L, Hall D, Das S, et al: Phenotype-genotype correlation of in vitroSN-38 and bilirubin glucuronidation in human liver tissue with UGT-1A1polymorphism. Clin Pharmacol Ther 65:576-582, 1999.

9. Waserman E, Myara A, Lopiec F, et al: Severe CPT-11 toxicity in patientswith Gilbert’s syndrome: Two case reports. Ann Oncol 8:1049-1051, 1997.

10. Ando Y, Saka H, Ando M, et al: Polymorphisms ofUDP-glucuronosyltransferase and irinotecan toxicity: A pharmacogenetic analysis.Cancer Res 60:6921-6926, 2000.

11. Chu W-Y, Kato Y, Veda K, et al: Biliary excretion mechanism of CPT-11 andits metabolites in humans: Involvement of primary active transporters. CancerRes 58:5137-5143, 1988.

12. Narasaki F, Oka M, Nakano R, et al: Human canalicular multispecificorganic anion transporter (cMOAT) is expressed in human lung, gastric andcolorectal cancer cells. Biochem Biophys Res Comm 240:606-611, 1997.

13. Keppler D, Konig J, Buchler M: The canalicular multidrug resistanceprotein; CMRP2, a novel conjugate expert pump in the apical membrane ofhepatocytes. Adv Enzyme Regul 37:321-333, 1997.

14. Paulusma CC, Bosma PJ, Zaman GJ, et al: Congenital jaundice in rats witha mutation in the multidrug resistance-associated protein gene. Science271:1126-1128, 1996.

15. Paulusma CC, Kool M, Bosma PJ, et al: A mutation in the human canalicularmultispecific organic anion transporter gene causes the Dubin-Johnson syndrome.Hepatology 25:1539-1542, 1997.

16. Brangi M, Litman T, Ciotti M, et al: Camptothecin resistance: Role of theATP-binding cassette (ABC), mitoxantrone-resistance half-transporter (MXR), andpotential glucuronidation in MXR-expressing cells. Cancer Res 59:5938-5946,1999.

17. Maliepaard M, vanGasteten MA, Tohgo A, et al: Circumvention of breastcancer resistance protein (BCRP) -mediated resistance to comptotecins in vitrousing nonsubstrate drugs or the BCRP inhibitor GF 120918. Clin Cancer Res7:935-941, 2001.

18. Gupta E, Wang X, Ramirez J, et al: Modulation of glucuronidation ofSN-38, the active metabolite of irinotecan by valproic acid and phenobarbital.Cancer Chemother Pharmacol 39:440-444, 1997.

19. Gupta E, Mick R, Ramirez J, et al: Pharmacokinetic and pharmacodynamicevaluation of the topoisomerase inhibitor irinotecan in cancer patients. J ClinOncol 15:1502-1510, 1997.

20. Grossman SA, Hochberg F, Fisher J, et al: Increased 9-aminocamptothecin(9-AC) dose requirements in patients on anticonvulsants. Proc Am Soc Clin Oncol16:1387, 1997.

21. Chang S, Kuhn J, Rizzo J, et al: Phase I study of paclitaxel in patientswith recurrent malignant glioma: A North American Brain Tumor Consortium Report.J Clin Oncol 16:2188-2194, 1988.

22. Saltz LB, Kanowitz J, Kemeny NE, et al: Phase I clinical andpharmacokinetic study of irinotecan, fluorouracil and leucovorin in patientswith advanced solid tumors. J Clin Oncol 14:2959-2967, 1996.

23. Pitot H, Goldberg RM, Reid JM, et al: Phase I dose-funding andpharmacokinetic trial of irinotecan hydrochloride (CPT-11) using aonce-every-three-week dosing schedule for patients with advanced solid tumormalignancy. Clin Cancer Res 6:2236-2244, 2000.