Leukemia (1998) 12, 346–352  1998 Stockton Press All rights reserved 0887-6924/98 $12.00

Etoposide and pharmacology in patients who develop secondary acute myeloid MV Relling1,6, Y Yanishevski1, J Nemec, WE Evans1,5,6, JM Boyett3,5, FG Behm4,5 and C-H Pui2,4,5

Departments of 1Pharmaceutical Sciences, 2Hematology/Oncology, 3Biostatistics and 4Pathology, St Jude Children’s Research Hospital, Memphis, and Colleges of 5Medicine and 6Pharmacy, The University of Tennessee, Memphis, TN, USA

Etoposide, an effective agent for acute lymphoblastic leukemia ALL patients develop AML as a consequence of therapy that (ALL), can cause secondary (AML) in is curative in Ͼ70% of children so treated. a subset of patients. Our objectives were to determine whether Although several treatment regimen-related factors (eg patients who develop secondary AML displayed altered etopo- 18 22 side or other pharmacologic characteristics schedule, concurrent use of or cumulative compared to identically treated patients who did not develop dose23) have been hypothesized to be associated with an AML. Children with newly diagnosed ALL were treated accord- increased risk of epipodophyllotoxin-induced AML, there are ing to a protocol which included etoposide 300 mg/m2 given no known patient-specific risk factors by which one can ident- three times over 8 days during remission induction and once ify patients at higher risk for its development. Furthermore, no every 2–4 weeks during 120 weeks of continuation therapy. Characteristic 11q23 rearrangements were documented in studies have evaluated whether interindividual differences in seven of the eight patients with AML. Etoposide clearance, time the pharmacokinetics of etoposide, or of its reactive O- that etoposide concentrations exceeded 10 ␮M, etoposide or demethylated metabolities, or both, contribute to the risk for etoposide catechol area-under-the-plasma-concentration vs development of secondary AML. time curve (AUC), serum albumin, and average In our recently completed front-line protocol for treatment concentration did not differ significantly between the two of childhood ALL, SJCRH protocol Total XIII, eight children groups; methyltransferase (TPMT) activity tended to be lower in the eight children who did vs the 23 matched con- developed secondary AML. Children on this protocol had trol children who did not develop AML (P = 0.16). Further assessments of etoposide disposition and antimetabolite phar- regression analyses likewise indicated that lower TPMT activity macology performed at times specified by the treatment proto- tended to be associated with shorter onset of secondary AML col, affording us the opportunity to compare these assessments (P = 0.11); it also tended to be lower among the eight index in children who went on to develop secondary AML vs a cases compared to the entire unmatched cohort of 105 ident- matched cohort of identically treated children who did not ically treated children with ALL (P = 0.10). We observed no stat- istically significant differences in etoposide disposition and develop secondary AML. antimetabolite pharmacology between patients who did and did not develop secondary AML. Keywords: etoposide; secondary leukemia; pharmacokinetics; Patients and methods thiopurine methyltransferase; methotrexate; chromosome 11q23 Patients

Introduction A total of 154 children with newly diagnosed ALL were enrolled (after obtaining informed consent from the parent or The epipodophyllotoxins, etoposide and , are a guardian) on our front-line therapeutic protocol SJCRH Total class of antineoplastics for which pharmacodynamic relation- XIIIHR (Table 1).24,25 With a minimum of 38 months of follow- ships have been established: ie plasma concentration of par- up from diagnosis of ALL for all patients, eight children ent drug correlates with both anticancer effect1–3 and acute developed secondary AML, four of whom have been toxicity (eg myelosuppression).2,4–8 However, defining the described previously.22,25 In order to control for as many fac- optimal dosages and schedules of the epipodophyllotoxins has tors as possible that could affect drug disposition and to proven challenging, in part due to a 10-fold interpatient ensure equivalent drug treatment, patients who were treated variability in systemic clearance9–14 and because we do not on the same protocol and had not developed secondary AML understand the risk factors for their most devastating delayed were matched to secondary AML cases using the following toxicity: the development of secondary acute myeloid leuke- criteria: same immunophenotype category for the original ALL mia (AML) in as many as 10–20% of adult and pediatric (B progenitor vs T cell), identical number of asparaginase patients.15–19 The epipodophyllotoxins are often incorporated doses delivered during induction therapy, same sex, same into treatment regimens for the most common childhood number of days (±2 days) required to deliver remission induc- malignancy, acute lymphoblastic leukemia (ALL).20 However, tion therapy, and identical pre-induction methotrexate therapy some of the same epipodophyllotoxin-including dosage regi- (ie low-dose vs high-dose)26 except in one case, in which the mens that produce excellent antileukemic effect for ALL have AML case was an unusual patient who received no window unfortunately been associated with an increased risk of sec- therapy. At least two and up to four matches were found for ondary AML.18,21 Thus, it is important to determine why some each AML case. If there were more than two matches for the discrete variables, proximity of age (younger or older than 9 years of age and usually within 2 years of each other) was used as the criterion for the selection of the matches. Thus, except for the one child who received no pre-induction Correspondence: MV Relling, St Jude Children’s Research Hospital, 332 North Lauderdale, Memphis, Tennessee 38105, USA; Fax: 901 methotrexate, all cases and controls received identical chemo- 525 6869 therapy. All controls have been followed for at least 38 Received 22 August 1997; accepted 14 November 1997 months (as of October 1997), which is longer than the longest Pharmacologic risk factors for secondary AML MV Relling et al 347 Table 1 Treatment regimen for study XIIIHR weeks 15, 29, 53, 69, 81, 94 and 120, or at any time when relapse was suspected. When the diagnosis of AML was sus- Remission induction/Consolidation pected, the samples were analyzed by cytogenetics and by 2 × 2 Weeks 0–4 Methotrexate, either 1 g/m 24 h or 30 mg/m q Southern blot for MLL rearrangements27 or by reverse tran- 6h× 6 Prednisone 40 mg/m2 daily, scriptase polymerase chain reaction for common translocation 25 mg/m2 weekly × 2 doses, 1.5 mg/m2 transcripts involving MLL (eg t(9;11) and t(11;19)). weekly, asparaginase 10 000 U/m2 i.m. 3 times per week × 9 doses Weeks 5–6 Etoposide 300 mg/m2 + 300 mg/m2 i.v. Pharmacologic studies days 22, 25, 29 Weeks 7–8 Methotrexate 2 g/m2 days 44 and 51; daily oral 75 mg/m2 Etoposide and its catechol were measured in duplicate using minor modifications of a previously described HPLC assay28 Continuation with electrochemical detection. Using 200 ␮l of plasma per Week 1 Etoposide 300 mg/m2 + 2 replicate, the limits of detection for etoposide and etoposide 300 mg/m i.v. ␮ Week 2 Daily oral mercaptopurine catechol in plasma were approximately 0.001 M and 75 mg/m2 + methotrexate 40 mg/m2 i.v./i.m. 0.0017 ␮M, respectively. A first-order two-compartment struc- Week 3 Methotrexate 40 mg/m2 i.v./i.m. + cytarabine tural pharmacokinetic model was used to simultaneously fit 300 mg/m2 i.v. the parent drug and etoposide catechol data (Figure 1), using 2 + Week 4 Daily prednisone 40 mg/m vincristine a Bayesian estimation algorithm implemented in Adapt II.29 1.5 mg/m2 i.v. + asparaginase 10 000 U/m2 i.m. Week 5 Etoposide 300 mg/m2 + cyclophosphamide The pharmacokinetic parameters and coefficients of variation 300 mg/m2 i.v. used in the prior mean vector and the initial estimates for eto- Week 6 Daily oral mercaptopurine poside (Vc, Ke-Total, Kcp, Kpc) were identical to those we have 75 mg/m2 + methotrexate 2 g/m2 i.v. previously reported;28 in addition, we assumed that the inter- Week 7 Etoposide 300 mg/m2 + cytarabine 300 mg/m2 i.v. compartmental distribution rate constants and central volume 2 + Week 8 Daily prednisone 40 mg/m vincristine of distribution for the etoposide catechol were identical to 1.5 mg/m2 i.v. + asparaginase 10 000 U/m2 i.m. those for etoposide (preliminary experiments indicate compa- rable for both etoposide catechol and Continuation therapy lasted for 120 weeks; reinduction/ consolidation was repeated at weeks 32–38 with etoposide/ parent drug), and we used initial estimates and prior means cytarabine on day 22 only; high-dose methotrexate was omitted for the elimination rate constant for the etoposide catechol −1 after 1 year of continuation therapy. (Kecat) of 0.002 h and rate constant for conversion of etopo- −1 side to etoposide catechol (KvpPcat) of 1.25 h , incorporating extremely large coefficients of variation, since no prior data latency period between ALL and AML diagnosis in the were available. In addition to the primary etoposide pharma- indexed cases. In addition to the eight children who cokinetic parameters, we also estimated: AUC for both etopo- developed secondary AML, a total of 105 children who did side catechol and parent drug from 0 to 24 h using standard not develop secondary AML received all nine doses of aspara- noncompartmental techniques with simulated concentrations ginase during remission induction, completed remission every 10 min, etoposide clearance as Vc·Ketotal or as induction within 10 days of the scheduled time period, and dose/AUC, and time that plasma concentrations exceeded a had erythrocyte thiopurine methyltransferase (TPMT) activity mutagenic concentration30 of 10 ␮M (t Ͼ 10 ␮M). Erythrocyte measured at day 44 of remission induction; these 105 patients TPMT activity was measured as previously described.31 constituted the entire cohort from which the matched controls Plasma methotrexate concentrations were measured using a were selected and served as an additional control group for fluorescence polarization immunoassay (TDx, Abbott Labora- comparison of TPMT activity (see below). tories, Abbott Park, IL, USA); the average of the concentrations measured 44 h (at which time leucovorin rescue began) fol- lowing the first five courses of MTX (over the first 6 months Samples of therapy) was used in the analyses.

On day 29 of remission induction therapy, serum albumin was measured, and blood in heparinized tubes was obtained pre- Statistics and at 1, 1.5, 2, 6 and 24 h following etoposide 300 mg/m2 i.v. over 2 h for etoposide measurements. This was the third A matched case–control study design was used to screen for of a planned total of 48 identical doses of etoposide that were possible pharmacologic factors that differed between children given over 2.5 years of ALL therapy (Table 1).22 The plasma who did and did not develop secondary AML. The values of was stored at −70°C until the time of analysis. On day 44 of parameters from the matched controls were compared to the remission induction therapy, lysates from hep- observed parameter values for the patients with secondary arinized blood samples were used to measure TPMT activity, AML using an exact Wilcoxon test32,33 stratified for matching. just prior to the first consolidation doses of 6-mercaptopurine. An exact conditional logistic regression analysis,34 stratified Plasma for methotrexate concentration was measured 44 h for matching, was used to assess whether age, etoposide phar- following the start of each course of high-dose MTX. macologic parameters, albumin, TPMT, or average 44 h MTX concentration were predictive for the development of second- ary AML. Among those with AML, correlations among phar- Diagnosis of AML macologic parameters and time to onset of AML were assessed using Spearman’s rank tests. A Cox proportional hazards Children had aspirates performed at diagnosis regression model was used in a forward stepwise fashion to of ALL, on day 43 of remission induction therapy, and at assess pharmacologic factors for possible correlation with time Pharmacologic risk factors for secondary AML MV Relling et al 348

Figure 1 Structural pharmacokinetic model used to estimate parameters for etoposide and etoposide catechol disposition, with plasma con- centration of etoposide (upper line) and etoposide catechol (lower line) vs time for a representative patient.

to onset of secondary AML. Based on results of the case– Table 2 Characteristics at diagnosis of ALL patients who control analysis, one of the pharmacologic variables (ie TPMT developed secondary AML and matched controls activity) was compared in the entire cohort from which the controls were selected and the eight AML cases, using a Wil- Secondary AML No AML coxon test. n 823 Mean age in years 9.2 8.9 (range) (3.1–17.3) (2.3–17.9) Results White/Blacka (%) 6/2 17/6 (75/25) (74/26) Eight patients developed secondary AML at 12–30 months T/B lineage ALL (%) 1/7 4/19 (median 24 months) from diagnosis of ALL; the myeloid blasts (13/87) (17/83) were characterized by 11q23 translocations in seven cases. Mean (range) initial 32 800 18 200 leukocyte count (1800–192 000) (1800–162 000) Characteristics of the eight patients who developed secondary (cells/␮l) AML and those of the control patients are summarized in Male/Female (%) 4/4 13/10 Table 2. The only two discrete variables that were not (50/50) (57/43) matched between the AML group and the controls were race Window therapy: 4/3/1 12/11/0 and pre-induction methotrexate therapy; however, these two high-dose/low-dose/no (50/38/12) (52/48/0) variables are unlikely to affect the pharmacologic parameters methotrexate (%) of interest, because there was no significant difference in eto- aNone of the patients were Hispanic. With the exception of one case poside clearance, AUC, t Ͼ 10 ␮M, etoposide catechol AUC, with AML having two white and two black control matches, races fraction unbound, albumin, erythrocyte TPMT activity, and of index cases and their controls were identical. average MTX concentrations in the low-dose vs the high-dose group or in blacks vs whites (data not shown). There was no attempt to match initial leukocyte count in the two groups, diagnosis), we explored whether the time of onset of AML was but they did not differ in this respect (P = 0.70). related to TPMT activity or to other pharmacologic variables, Pharmacologic parameters (Table 3) were compared hypothesizing that early onset of AML might be more likely between the eight patients with secondary AML and their related to measured pharmacologic variables than later onset. matched controls. TPMT activity tended to be lower (P = 0.16) Figure 2 depicts the relationship of onset of AML vs various in the patients who developed secondary AML. A logistic pharmacologic variables. Whether the analysis of the time to regression analysis, stratified by the matching chosen for the onset of AML was stratified using the eight strata based on controls, revealed no statistically significant predictors of matching controls for every AML case, or those strata were secondary AML. ignored and replaced by a post hoc stratification based on Because the pharmacologic variables evaluated were all age, race, and immunophenotype of ALL, no factors were sig- assessed early in ALL treatment (less than 6 months from nificantly related to shorter onset of development of AML, Pharmacologic risk factors for secondary AML MV Relling et al 349 Table 3 Pharmacologic parameters in patients who did and did not develop AML

Patients who developed Controlsa Stratified Wilcoxon P Conditional logistic AML value regression P value n 823 Etoposide clearance (ml/min/m2) 45.9 [40.0–53.4] 42.0 [37.1–49.3] 0.51 0.82 (24.5–61.1) (29.2–71.4) Etoposide AUC (␮M·h) 178 [155–208] 194 [167–222] 0.55 0.94 (134–356) (113–278) Etoposide catechol AUC (␮M·h) 1.15 [0.63–2.86] 1.65 [0.77–2.03] 0.74 0.82 (0.19–3.82) (0.055–3.44) % Unbound in plasma 6.6 [5.3–7.9] 7.2 [5.8–10.6] 0.37 0.83 (4.6–28.5) (4.2–26.5) Time etoposide Ͼ10 ␮M (h) 5.8 [5.2–6.7] 5.9 [5.2–6.5] 0.75 0.57 (4.4–8.9) (3.9–8.3) Day 29 albumin (g/dl) 3.4 [3.0–3.5] 3.4 [3.3–3.6] 0.26 0.38 (2.3–3.8) (2.3–4.0) Erythrocyte TPMT activity (U/ml packed 14.2 [11.2–19.6] 18.0 [13.0–21.7] 0.16 0.19 erythrocytes) (8.8–23.1) (8.8–33.9) Average methotrexate 44 h 0.16 [0.13–0.23] 0.12 [0.11–0.17] 0.18 0.55 concentration (␮M) (0.09–0.44) (0.08–0.69) aEach secondary AML case had two to four matched controls (see Methods for matching characteristics); the median [25th–75th percen- tiles], and (range) of values for each parameter are indicated. with P values of 0.11 and 0.065, respectively, for lower TPMT onset to AML as the dependent variable, there was also a activity, and P = 0.17 and P = 0.15, respectively, for higher suggestion that MTX plasma concentration tended to be average MTX concentrations. higher in those with shorter onset AML. Although etoposide pharmacokinetic data were not evalu- Most reports of very high frequencies of epipodophyllo- ated in all patients enrolled on the protocol, erythrocyte TPMT toxin-associated AML have involved clinical protocols in activity was assessed in most patients at day 44 of remission which epipodophyllotoxins immediately followed fairly pro- induction therapy. In order to further investigate whether longed courses of antimetabolite therapy (ie daily 6-mercapto- TPMT activity is related to the risk of secondary AML using purine and weekly methotrexate).15,17,18 Antimetabolite ther- all data available, we compared this parameter in the eight apy and folate deficiency have been shown to contribute to children who developed secondary AML and in the 105 iden- mutagenesis and carcinogenesis.35,36 In addition, complex tically treated children for whom erythrocyte TPMT activity interactions of mercaptopurine and methotrexate, which had been assessed. Median TPMT activity was 14.2 in the sec- might affect S-adenosyl methionine concentrations and the ondary AML patients vs 17.0 units/ml packed red blood cells state of DNA methylation,37 could predispose to nonhomolo- in the controls (P = 0.10 by Wilcoxon test). gous recombination following etoposide. One of the most important determinants of toxicity from exposure to active thi- oguanine nucleotides formed following 6-mercaptopurine Discussion therapy is the activity of erythrocyte thiopurine methyltransfer- ase.38 These considerations were the underpinnings for our This is the first report of pharmacologic characteristics of assessments of TPMT and methotrexate plasma concentrations patients who have developed epipodophyllotoxin-related as possible risk factors for etoposide-induced AML. Thiopurine AML. Here, we were able to compare etoposide and its methyltransferase activity tended to be lower in patients who metabolite pharmacokinetics, methotrexate plasma concen- developed AML vs those who did not in a univariate analysis trations, and thiopurine methyltransferase activity in a group (P = 0.16 compared to matched controls and P = 0.10 com- of children who developed AML to an identically treated con- pared to all available controls), in a conditional logistic trol group of children who did not develop AML. Although regression analysis (P = 0.19), and in a multivariate Cox pro- we observed no statistically significant differences at the tra- portional hazards model with time to onset of AML as the ditional 0.05 level in these pharmacologic parameters dependent variable (P = 0.11). The fact that P values of less between the two groups of patients, there was a trend for than the traditional 0.05 were not observed is not surprising, lower thiopurine methyltransferase activity in the children given the small number of secondary AML cases. Lower TPMT who went on to develop AML. This trend was also observed activity would result in greater exposure to active thiopurine when TPMT was compared in the eight patients who nucleotide metabolities, and would be consistent with prior developed AML and the larger unselected control group of antimetabolite therapy enhancing etoposide’s leukemogenic 105 children who did not develop AML. Because our assess- effects through a number of possible mechanisms. ments of etoposide and antimetabolite pharmacology were Our choice of pharmacologic factors for study warrants made early on in the course of these children’s ALL therapy, some comment. The secondary AML associated with epipodo- we reasoned that these pharmacologic assessments might be phyllotoxins, characterized by short onset, MLL gene most informative for prediction of early-onset secondary AML, rearrangements, and monoblastic phenotype,19 seems to be assuming that early onset corresponds with the early occur- related to the ability of these agents to interfere with the nor- rence of an etoposide-related leukemogenic effect. There was, mal activity of topoisomerase II. Most cases have been in fact, a trend indicating lower thiopurine methyltransferase reported following administration of etoposide and teniposide, activity associated with shorter onset of AML; in analyses with both of which interact very potently with topoisomerase II.39 Pharmacologic risk factors for secondary AML MV Relling et al 350

Figure 2 Plots of onset of secondary AML vs various pharmacologic parameters; lines are simply the least squares best fit. The Spearman = = rank correlation coefficients (rs) and associated P values are as follows: TPMT (thiopurine methyltransferase) activity, rs 0.73, P 0.04; etoposide = = ␮ =− = = clearance, rs 0.85, P 0.007; time that plasma etoposide concentrations exceeded 10 M, rs 0.85, P 0.007; serum albumin, rs 0.76, = =− = = = P 0.028; percent unbound etoposide in plasma, rs 0.18, P 0.67; average methotrexate 44 h plasma concentration, rs 0.38, P 0.35.

Similarly characterized secondary AML cases have been not tend to be different in children who did vs those who did reported less frequently following the use of less potent topo- not develop AML suggests that additional factors not directly isomerase II-active agents (eg ).16 Moreover, the related to etoposide exposure may be involved in protecting balanced translocations27 and concentration-dependent patients from development of etoposide-induced AML. recombinogenic effects30 of etoposide are plausible com- There are several caveats that must be acknowledged in pensatory changes in response to inhibition of topoisomerase interpreting these findings. One consideration is that etopo- II’s religation of double-stranded DNA breaks. Therefore, it is side pharmacokinetics were measured on only one occasion, possible that secondary AML might be more likely in patients with the third dose of etoposide, at the end of remission induc- who achieve high systemic concentrations of etoposide, or in tion therapy. This is problematic in that it is not known to those whose plasma concentrations of etoposide remained for what degree this measurement of etoposide pharmacokinetics a longer time above a concentration with high recombino- would be correlated with subsequent measurements through- genic effects in vitro,10␮M.30 Although increased etoposide out the next 2.5 years of therapy. We do know that, on the exposure was correlated with onset of AML among those who whole, the systemic clearance of a model substrate for hepatic developed AML (Figure 2), the fact that these parameters did cytochrome P450 metabolism (antipyrine) measured at the Pharmacologic risk factors for secondary AML MV Relling et al 351 end of remission induction therapy does not change signifi- References cantly throughout the course of 2.5 years of continuation ther- apy (unpublished data), and that etoposide is metabolized by 1 Clark PI, Slevin ML. The clinical pharmacology of etoposide and 40 one of the same enzymes (CYP3A4) metabolizing anti- teniposide. Clin Pharmacokinet 1987; 12: 223–252. pyrine.41 However, there are individual patients who 2 Clark PI, Slevin ML, Joel SP, Osborne RJ, Talbot DI, Johnson PW, occasionally exhibit changes in etoposide clearance, and Reznek R, Masud T, Gregory W, Wrigley PF. A randomized trial whether such changes occur more or less frequently among of two etoposide schedules in small-cell lung : the influence children who develop AML than among those who do not of pharmacokinetics on efficacy and toxicity. J Clin Oncol 1994; 12: 1427–1435. cannot presently be known without making measurements 3 Rodman JH, Abromowitch M, Sinkule JA, Rivera GK, Evans WE. with every single course of etoposide. Nonetheless, there are Clinical pharmacodynamics of continuous infusion teniposide: data to indicate that intrapatient variability in etoposide phar- systemic exposure as a determinant of response in a Phase I trial. macokinetics4 and in CYP3A442 is less than interpatient varia- J Clin Oncol 1987; 5: 1007–1014. bility, suggesting that the data here from a single time-point 4 Ratain MJ, Mick R, Schilsky RL, Vogelzang NJ, Berezin F. Pharma- may have some relevance to the total exposure to etoposide cologically based dosing of etoposide: a means of safely increasing dose intensity. J Clin Oncol 1991; 9: 1480–1486. for any given patient. Another major caveat in interpreting our 5 Karlsson MO, Port RE, Ratain MJ, Sheiner LB. A population model data is that, because of the relatively small number of second- for the leukopenic effect of etoposide. Clin Pharmacol Ther 1995; ary AML cases (eight), we have very limited power to discern 57: 325–334. differences in pharmacologic parameters between children 6 Rodman JH, Murry DJ, Madden T, Santana VM. Altered etoposide who do and do not develop secondary AML. Because of our pharmacokinetics and time to engraftment in pediatric patients limited power, we emphasize that the finding that TPMT undergoing autologous bone marrow transplantation. J Clin Oncol 1994; 12: 2390–2397. activity tended to be lower among those with AML may be 7 Arbuck SG, Douglass HO, Crom WR, Goodwin P, Silk Y, Cooper important and worthy of further study, while the lack of sig- C, Evans WE. Etoposide pharmacokinetics in patients with normal nificant associations with other pharmacologic parameters and abnormal organ function. J Clin Oncol 1986; 4: 1690–1695. must not be interpreted as definitive. 8 Evans WE, Rodman JH, Relling MV, Petros WP, Stewart CF, Pui A remarkable feature of the etoposide pharmacokinetic data C-H, Rivera GK. Differences in teniposide disposition and pharma- described here is the very high systemic etoposide clearance codynamics in patients with newly diagnosed versus relapsed acute lymphocytic leukemia. J Clin Oncol 1992; 260: 71–77. 2 ¨ in the entire group of patients, with a median of 43 ml/min/m . 9 Boos J, Real E, Schulze-Westhoff P, Wolff J, Euting T, Jurgens H. This clearance is higher than that reported in seven previous Investigation of the variability of etoposide pharmacokinetics in studies in children1,6,9,12,28 and 14 previous studies in children. Int J Clin Pharmacol Ther Toxicol 1992; 30: 495–497. adults.1,2,7,11,13 One possible explanation for the very high 10 D’Incalci M, Farina P, Sessa C, Mangioni C, Conter V, Masera G, clearance we observed is that the previous 4 week course of Rocchetti M, Pisoni MB, Piazza E, Beer M, Cavalli F. Pharmacoki- prednisone may have induced CYP3A443 metabolism of eto- netics of VP16-213 given by different administration methods. Cancer Chemother Pharmacol 1982; 7: 141–145. 40 ¨ poside, a finding consistent with our previous observations 11 Liliemark EK, Liliemark J, Pettersson B, Gruber A, Bjorkholm M, of enhanced P450 metabolism following remission Peterson C. In vivo accumulation of etoposide in peripheral leu- induction.44 kemic cells in patients treated for acute myeloblastic leukemia; In summary, although our data should be interpreted with relation to plasma concentrations and protein binding. Leuk Lym- caution due to their retrospective nature and the small number phoma 1993; 10: 323–328. of patients, they represent the only pharmacokinetic data 12 Lowis SP, Pearson ADJ, Newell DR, Cole M. Etoposide pharmaco- kinetics in children: the development and prospective validation available in identically treated patients who did and did not of a dosing equation. Cancer Res 1993; 53: 4881–4889. ¨ ¨ go on to develop epipodophyllotoxin-associated AML, and 13 Pfluger K-H, Hahn M, Holz J-B, Schmidt L, Kohl P, Fritsch H-W, may provide insights into variables that deserve further study. Jungclas H, Havemann K. Pharmacokinetics of etoposide: corre- Our results suggest that low thiopurine methyltransferase lation of pharmacokinetic parameters with clinical conditions. activity may increase the risk of epipodophyllotoxin- Cancer Chemother Pharmacol 1993; 31: 350–356. associated AML among ALL patients. Ongoing studies in pre- 14 Sinkule JA, Hutson P, Hayes FA, Etcubanas E, Evans WE. Pharma- cokinetics of etoposide (VP-16) in children and adolescents with clinical models in our laboratory are aimed at defining less refractory solid tumors. Cancer Res 1984; 44: 3109–3113. recombinogenic etoposide-containing regimens without com- 15 Sugita K, Furukawa T, Tsuchida M, Okawa Y, Nakazawa S, Akat- promising their efficacy. suka J, Ohira M, Nishimura K. High frequency of etoposide-related secondary leukemia in children with non-Hodgkin’s . Am J Pediatr Hematol Oncol 1993; 15: 99–104. 16 Pedersen-Bjergaard J, Philip P, Larsen SO, Andersson M, Dau- gaard G, Ersboll J, Hansen SW, Hou-Jensen K, Nielsen D, Sigs- gaard TC, Specht L, Osterlind K. Therapy-related myelodysplasia Acknowledgements and acute myeloid leukemia. Cytogenetic characteristics of 115 consecutive cases and risk in seven cohorts of patients treated intensively for malignant diseases in the Copenhagen series. Leu- This work was supported by NIH CA51001, CA20180, kemia 1993; 7: 1975–1986. CA36401, Cancer Center CORE grant CA21765, by a Center 17 Winick NJ, McKenna RW, Shuster JJ, Schneider NR, Borowitz MJ, Bowman WP, Jacaruso D, Kamen BA, Buchanan GR. Secondary of Excellence grant from the State of Tennessee, a research acute myeloid leukemia in children with acute lymphoblastic leu- award from Astra USA, and American Lebanese Syrian Asso- kemia treated with etoposide. J Clin Oncol 1993; 11: 209–217. ciated Charities (ALSAC). We thank research nurses Ms Sheri 18 Pui C-H, Ribeiro R, Hancock ML, Rivera GK, Evans WE, Raimondi Ring, Lisa Walters and Terri Kuehner; Dr Kenneth Caldwell, S, Head DR, Behm FG, Mahmoud MH, Sandlund JT, Crist W. Ms Nancy Kornegay, Kim Juneau and Mr Michael Hancock Acute myeloid leukemia in children treated with epipodophyllo- for assistance with data analysis; Dr Jean Cai and Mr Ken Cox toxins for acute lymphocytic leukemia. New Engl J Med 1991; 325: 1682–1687. for expert chemical analysis; members of the clinical staff for 19 Smith MA, Rubinstein L, Ungerleider RS. Therapy-related acute patient care and laboratory analyses; and the patients and myeloid leukemia following treatment with epipodophyllotoxins: their families for participating in clinical research studies. estimating the risks. Med Pediatr Oncol 1995; 23: 86–98. Pharmacologic risk factors for secondary AML MV Relling et al 352 20 Pui C-H, Crist WM. Biology and treatment of acute lymphoblastic properties, -erythrocyte correlation and presence of isozymes. leukemia. J Clin Oncol 1994; 124: 491–503. Pharmacogenetics 1992; 2: 148–159. 21 Rivera GK, Pui C-H, Crist WM. The epipodophyllotoxins: both 32 Mehta C, Patel N. In: StatXact 3 for Windows: Statistical Software sides of the coin (letter; comment). J Clin Oncol 1993; 11: for Exact Nonparametric Inference. Cytel Software Corp: Cam- 1624–1625. bridge, MA, 1995. 22 Pui C-H, Relling MV, Behm FG, Hancock ML, Boyett JM, Rai- 33 Lehmann EL. In: Nonparametrics: Statistical Methods based on mondi SC, Krance RA, Mahmoud HH, Ribeiro RC, Sandlund JT Ranks. Holden-Day: San Francisco, 1975. et al. L-asparaginase may potentiate the leukemogenic effect of 34 Breslow NE, Day NE. In: Statistical Methods in Cancer Research. the epipodophyllotoxins. Leukemia 1995; 9: 1680–1684. IARC Scientific Publications: Lyon, 1980. 23 Ratain MJ, Kaminer LS, Bitran JD, Larson RA, LeBeau MM, Skosey 35 Weisburger EK. Bioassay program for carcinogenic hazards of can- C, Purl S, Hoffman PC, Wade J, Vardiman JW, Kaly K, Rowley JD, cer chemotherapeutic agents. Cancer 1977; 40: 1935–1949. Golomb HM. Acute nonlymphocytic leukemia following etopo- 36 James SJ, Basknakian AG, Miller BJ. In vitro folate deficiency side and combination for advanced non- induces deoxynucleotide pool imbalance, , and small-cell carcinoma of the lung. Blood 1987; 70: 1412–1417. mutagenesis in Chinese hamster ovary cells. Cancer Res 1994; 54: 24 Pui C-H, Rivera GK, Evans WE, Sandlund JT, Hancock ML, Relling 5075–5080. MV, Boyett J, Ribeiro RC, Furman WL, Mahmoud HH, Hurwitz 37 Chiang PK, Gordon RK, Tal J, Zeng GC, Doctor BP, Pardhasaradhi C, Crist WM. Current treatment and results of St Jude studies for K, McCann PP. S-Adenosylmethionine and methylation. FASEB J ¨ childhood acute lymphoblastic leukemia. In: Buchner T, 1996; 10: 471–480. Schellong G, Hiddemann W, Urbanitz D, Ritter J (eds). Acute Leu- 38 Lennard L, Van Loon JA, Lilleyman JS, Weinshilboum RM. Thiopu- kemias IV. Prognostic Factors and Treatment Strategies. Hematol- rine pharmacogenetics in leukemia: correlation of erythrocyte thi- ogy and Blood Transfusion. Springer-Verlag: Berlin, 1994, pp opurine methyltransferase activity and 6-thioguanine nucleotide 307–313. concentrations. Clin Pharmacol Ther 1987; 41: 18–25. 25 Pui C-H, Relling MV, Rivera GK, Hancock ML, Raimondi SC, Hes- 39 Ross W, Rowe T, Glisson B, Yalowich J, Liu L. Role of topoisomer- lop HE, Santana VM, Ribeiro RC, Sandlund JT, Mahmoud HH, ase II in mediating epipodophyllotoxin-induced DNA cleavage. Evans WE, Krance RA. Epipodophyllotoxin-related acute myeloid Cancer Res 1984; 44: 5857–5860. leukemia: a study of 35 cases. Leukemia 1995; 9: 1990–1996. 40 Relling MV, Nemec J, Schuetz EG, Schuetz JD, Gonzalez FJ, Kor- 26 Synold TW, Relling MV, Boyett JM, Rivera GK, Sandlund JT, Mah- zekwa KR. O-Demethylation of epipodophyllotoxins is catalyzed moud H, Crist WM, Pui C-H Evans WE. Blast cell methotrexate- by human cytochrome P450 3A4. Mol Pharmacol 1994; 45: polyglutamate accumulation in vivo differs by lineage, ploidy, and 352–358. methotrexate dose in acute lymphoblastic leukemia. J Clin Oncol 41 Engel G, Hofmann U, Heidemann H, Cosme J, Eichelbaum M. 1994; 94: 1996–2001. Antipyrine as a probe for human oxidative : 27 Rubnitz JE, Behm FG, Downing JR. 11q23 rearrangements in acute identification of the cytochrome P450 enzymes catalyzing 4-hyd- leukemia. Leukemia 1996; 10: 74–82. roxyantipyrine, 3-hydroxymethylantipyrine, and norantipyrine for- 28 Relling MV, McLeod H, Bowman L, Santana VM. Etoposide phar- mation. Clin Pharmacol Ther 1996; 59: 613–623. macokinetics and pharmacodynamics after acute and chronic 42 Watkins PB. Noninvasive tests of CYP3A enzymes. Pharmaco- exposure to cisplatin. Clin Pharmacol Ther 1994; 56: 503–511. genetics 1994; 4: 171–184. 29 D’Argenio DZ, Schumitzky A. A program package for simulation 43 Schuetz EG, Schuetz JD, Strom SC, Thompson MT, Fisher RA, and parameter estimation in pharmacokinetics systems. Comput Molowa DT, Li D, Guzelian PS. Regulation of human liver cytoch- Programs Biomed 1979; 9: 115–134. romes P450 in family 3A in primary and continuous culture of 30 Chen C-L, Fuscoe JC, Liu Q, Relling MV. Etoposide causes illegit- human hepatocytes. Hepatology 1993; 18: 1254–1262. imate V(D)J recombination in human lymphoid leukemic cells. 44 Relling MV, Crom WR, Pieper JA, Cupit GC, Rivera GK, Evans Blood 1996; 89: 1701–1707. WE. Hepatic drug clearance in children with leukemia: changes in 31 Szumlanski CL, Honchel R, Scott MC, Weinshilboum RM. Human clearance of model substrates during remission-induction therapy. liver thiopurine methyltransferase pharmacogenetics: biochemical Clin Pharmacol Ther 1987; 41: 651–660.