6-Mercaptopurine Dosage and Pharmacokinetics Influence The

Home , Bone marrow suppression, Mercaptopurine

Leukemia (2001) 15, 74–79  2001 Nature Publishing Group All rights reserved 0887-6924/01 $15.00 www.nature.com/leu 6-Mercaptopurine dosage and pharmacokinetics inﬂuence the degree of bone marrow toxicity following high-dose methotrexate in children with acute lymphoblastic leukemia K Schmiegelow and U Bretton-Meyer

The Laboratory for Pediatric Oncology, The Pediatric Clinic II, The Juliane Marie Centre, The National University Hospital, Rigshospitalet, Copenhagen, Denmark

Through inhibition of purine de novo synthesis and enhance- are the primary mediators of the cytotoxic effect of 6MP ment of 6-mercaptopurine (6MP) bioavailability high-dose through their incorporation into DNA.3 Another important methotrexate (HDM) may increase the incorporation into DNA of 6-thioguanine nucleotides (6TGN), the cytoxic metabolites of metabolic pathway is the methylation by the enzyme thiopur- 6MP. Thus, coadministration of 6MP could increase myelotoxi- ine methyltransferase (TPMT, E.C. 2.1.1.67) of 6MP and some city following HDM. Twenty-one children with standard risk of its metabolites such as 6-thioinosine 59-monophosphate.4 (SR) and 25 with intermediate risk (IR) acute lymphoblastic leu- As a result of a common genetic polymorphism, large interin- kemia (ALL) were studied. During consolidation therapy they dividual variations in TPMT activity exist, and this significantly received either three courses of HDM at 2 week intervals with- influences the degree of methylation and intracellular 6TGN out concurrent oral 6MP (SR-ALL) or four courses of HDM given 5,6 at 2 week intervals with 25 mg/m2 of oral 6MP daily (IR-ALL). accumulation. During the first year of maintenance with oral 6MP (75 High-dose methotrexate (HDM) is an important part of the mg/m2/day) and oral MTX (20 mg/m2/week) they all received five therapy given to children with ALL during consolidation ther- courses of HDM at 8 week intervals. In all cases, HDM consisted apy with or without concurrent oral 6MP and as re-inductions of 5000 mg of MTX/m2 given over 24 h with intraspinal MTX and during maintenance therapy with oral daily 6MP and weekly leucovorin rescue. Erythrocyte levels of 6TGN (E-6TGN) and methotrexate (MTX) as the backbone.7–9 HDM often causes methotrexate (E-MTX) were, on average, measured every significant bone marrow toxicity which carries a risk of infec- second week during maintenance therapy. When SR consoli- 10 dation (6MP: 0 mg), IR consolidation (6MP: 25 mg/m2), and tions and a need for transfusions. This myelosuppression SR/IR maintenance therapy (6MP: 75 mg/m2) were compared, may lead to treatment interruptions and a reduction of the white cell and absolute neutrophil count (ANC) nadir, lympho- dose intensity, which can affect the cure rate.11–13 MTX may cyte count nadir, thrombocyte count nadir, and hemoglobin increase the bioavailability of 6MP,14 and in addition, through nadir after HDM decreased significantly with increasing doses inhibition of de novo purine synthesis, enhance the avail- of oral 6MP. Three percent of the HDM courses given without 15 oral 6MP (SR consolidation) were followed by an ANC nadir ability of 6TGN and its incorporation into DNA. We specu- ,0.5 × 109/l compared to 50% of the HDM courses given during lated that this drug interaction could in part be responsible for SR/IR maintenance therapy. Similarly, only 13% of the HDM the bone marrow toxicity which follows HDM given during courses given as SR-ALL consolidation induced a thrombocyte 6MP/MTX maintenance therapy. In the present study, we count nadir ,100 × 109/l compared to 58% of the HDM courses demonstrate that the dose of concurrently given oral 6MP and given during maintenance therapy. The best-fit model to predict the individual variations in 6MP metabolism significantly the ANC nadir following HDM during maintenance therapy included the dose of 6MP prior to HDM (β =−0.017, P = 0.001), influence the degree of myelotoxicity following HDM. the average ANC level during maintenance therapy (β = 0.82, P = 0.004), and E-6TGN (β =−0.0029, P = 0.02). The best-fit model to predict the thrombocyte nadir following HDM during Patients and methods maintenance therapy included only mPLATE (β = 0.0057, P = 0.046). In conclusion, the study indicates that reductions of the Patients dose of concurrently given oral 6MP could be one way of reduc- ing the risk of significant myelotoxicity following HDM during maintenance therapy of childhood ALL. Leukemia (2001) 15, The Nordic NOPHO ALL-92 protocol began on 1 January 74–79. 1992.16 Patients were eligible for the present study and for Keywords: Bone marrow toxicity; child; acute lymphocytic leuke- statistical analyses, if they: (1) were diagnosed with standard mia; 6-mercaptopurine; methotrexate; thioguanine nucleotides risk (SR) or intermediate risk (IR) non-B cell ALL after 1 January 1992, at the University Hospital, Rigshospitalet, Copenhagen; (2) were treated according to the NOPHO ALL-92 program; Introduction (3) entered 6MP/MTX maintenance therapy before 1 January 1997, and (4) were in first remission when the study was Acute lymphoblastic leukemia (ALL) is the most common can- closed, 1 January 1999. Fifty patients fulfilled these criteria. cer in childhood. During maintenance therapy of ALL with Four patients were excluded because they: (1) received only oral 6-mercaptopurine (6MP) and methotrexate (MTX), 6-thio- two courses of HDM due to severe neurotoxicity (n = 1); guanine nucleotides (6TGN) and MTX polyglutamates (2) were treated elsewhere (n = 2); or (3) data were lacking (n accumulate in erythrocytes (E-6TGN and E-MTX), and E- = 1). The remaining 30 boys and 16 girls included 21 cases 6TGN and E-MTX have been related to the cure rate.1,2 6TGN of SR- and 25 cases of IR-ALL. The SR/IR risk classification depended on age and white blood cell count (WBC) (SR: age 2.0–9.9 years and WBC ,10 × 109/l; IR: age 1.0–1.9 years or age 10.0–14.9 years and/or WBC 10–49 × 109/l) and the Correspondence: K Schmiegelow, The Pediatric Clinic II, JMC-4064, × 9 The National University Hospital, Rigshospitalet, Blegdamsvej 9, DK- absence of high risk criteria (WBC 50 10 /l, T cell disease, 2100 Copenhagen, Denmark; Fax: +45 3545 4673 mediastinal or CNS or testicular or lymphomatous disease, Received 27 June 2000; accepted 13 September 2000 t(4;11), t(9;22), a day 14 bone marrow with more than 25% Bone marrow toxicity after high-dose methotrexate K Schmiegelow and U Bretton-Meyer 75 lymphoblasts, or a day 29 bone marrow with more than 5% given in age-related doses (8–12 mg). Starting 36 h from the lymphoblasts). beginning of the HDM infusion, leucovorin rescue was given at a dose of 15 mg/m2 i.v. every 6 h until the serum MTX was below 200 nmol/l. If serum MTX was .1000 nmol/l at 42 h, Induction therapy the leucovorin doses were increased according to the serum MTX concentrations. Blood counts were taken from 10 days Induction therapy consisted of prednisolone (60 mg/m2/day after the start of HDM and repeated at 1–3 day intervals until divided into three doses on days 1–36, then tapered), weekly they were within the above given target range for maintenance vincristine (VCR, 2.0 mg/m2 ×6), doxorubicin (40 mg/m2 days therapy. For each HDM, we registered the hemoglobin, plate- 1, 22 and 36), Erwinia asparaginase (30 000 IU/m2 days 37 to let counts, WBC, absolute neutrophil (ANC) and lymphocyte 46), and i.s. MTX (days 1, 8, 15 and 29).16 count nadirs, which were experienced 10 to 20 days from the start of HDM. Since changes in the dose of 6MP will change the intracellular 6TGN levels, the dose of 6MP prior to HDM Consolidation therapy was only included in the statistical analyses if they were unaltered for at least 14 days before HDM. From day 50, children with SR-ALL received three courses of HDM given at 2 week intervals without concurrent oral 6MP. Consolidation therapy for patients with IR-ALL consisted of: E-6TGN/MTX and TPMT (1) from day 50: daily oral 6MP (75 mg/m2 for 2 × 2 weeks), cyclophosphamide (1 g/m2 ×2), four series of low-dose cytara- During maintenance therapy, blood samples were taken for bine (each with i.v. doses of 75 mg/m2 on 4 consecutive days), 6MP/MTX metabolite determination together with every blood and i.s. MTX (×2); (2) from day 106: daily oral 6MP for 8 count. The median monthly number of blood samples for the weeks (25 mg/m2 given once in the evening) with four courses 46 children was 2.0 (75% range: 0.9–2.8). E-MTX analyses of HDM given at 2 week intervals, (3) from day 169: 4 weeks were done at least 48 h after the latest oral dose of MTX. All of re-induction with dexamethasone (10 mg/m2/day divided analyses of E-6TGN and E-MTX were performed at The Lab- into three doses for 3 weeks, then tapered), weekly VCR (2.0 oratory for Pediatric Oncology, The University Hospital, Rig- mg/m2/day ×4), weekly daunorubicin (30 mg/m2/day ×4), and shospitalet. E-6TGN and E-MTX were measured by an HPLC asparaginase four times (30 000 IU/m2 at 3–4 day intervals); method18 and a radioligand assay,19 respectively. One of the and (4) from day 197: daily oral thioguanine (60 mg/m2 for 2 46 patients was TPMT deficient.20 E-6TGN and E-MTX weeks), cyclophosphamide (1 g/m2 ×1), two series of low-dose measurements were included in the multivariate regression cytarabine (each with i.v. doses of 75 mg/m2 on 4 consecutive analyses if available within 14 days prior to the start of HDM. days), and i.s. MTX (×1). If more than one measurement was available during this period, the one closest to the start of HDM was chosen. Maintenance therapy

Maintenance therapy with starting 6MP doses of 75 Statistics mg/m2/day and MTX doses of 20 mg/m2/week given in the evening was initiated at week 14 (SR) or week 33 (IR). The For each patient, we calculated: (1) the mean neutrophil nadir doses of oral 6MP and oral MTX were to be targeted to a WBC and thrombocyte nadir following HDM during consolidation × 9 , × 9 of 1.5–3.5 10 /l and reduced to 50% at a WBC 1.5 10 /l therapy and maintenance therapy (mANC-nadirMT-HDM, , × 9 and interrupted at a WBC 1.0 10 /l and/or a thrombocyte mPLATE-nadirMT-HDM); and (2) the mean WBC, neutrophil and count ,100 × 109/l. During maintenance therapy blood thrombocyte counts during the last 38 weeks (IR) or 53 weeks counts were taken at least every 2 weeks. The median (SR) of maintenance therapy which did not include HDM monthly number of blood counts for the 46 children was 3.6 (mWBC, mANC, mPLATE). The mWBC, mANC and mPLATE (75% range: 2.8–5.3). As part of a randomized study, 10 were calculated as weighted means using as weight the inter- patients with SR-ALL and 11 patients with IR-ALL were (in val between the sample in question and the next blood sample addition to dose adjustments by toxicity) been stratified to as previously described.21 The Mann–Whitney U test, Wil- have their doses of oral 6MP and MTX adjusted by E-6TGN coxon’s test, and Spearman’s rank order correlation analysis and E-MTX.2,17 However, even for these patients a WBC ,1.5 were applied to compare the distributions of parameters × 109/l or a thrombocyte count ,100 × 109/l indicated dose between subgroups, in related samples, and the correlations = 22 reductions as outlined above. During the first year of mainte- between parameters (rS correlation coefficient). Since nance therapy, patients with SR- and IR-ALL received at 4 HDM was given right at the start of maintenance therapy for week intervals alternate pulses of either (1) VCR (2.0 mg/m2 IR-ALL (at week 33) without preceding oral 6MP therapy, this ×1) and prednisolone (60 mg/m2/day for 1 week) or (2) HDM, HDM course was excluded from the multivariate regression the first being given at week 18 (SR) or week 33 (IR). These analysis of the impact of oral 6MP on HDM toxicity. Stepwise, re-inductions were continued until five courses of HDM had forward, multivariate linear regression analyses were done to been given.16 No oral MTX was given for a week from the identify the independent variables that best predicted the start of HDM. degree of bone marrow toxicity following HDM during maintenance therapy. Parameters were included in the models at significance limits of 0.05. Prior to the regression analyses High-dose MTX natural logarithmic transformation of the ANC and the thrombocyte nadir were done to approximate normal distributions. Five thousand milligrams of MTX/m2 were given over 24 h. Data analyses were done with SPSS statistical software.23 Eighteen to 24 h after the beginning of the HDM, i.s. MTX was Two-sided P values ,0.05 were regarded as being significant.

Leukemia Bone marrow toxicity after high-dose methotrexate K Schmiegelow and U Bretton-Meyer 76 apy. The median duration of these treatment interruptions was 10 days (75% range: 3–18 days, maximum: 30 days). The duration of treatment interruptions was signiﬁcantly related to =− , both the ANC nadir (rS 0.55, P 0.001) and thrombocyte =− , nadir (rS 0.72, P 0.001). The average ANC nadir experienced by the individual patient following HDM given as consolidation therapy was for all 21 patients with SR-ALL (mean of three courses) higher , than the mANC nadirMT-HDM (mean of ﬁve courses) (P 0.001, Wilcoxon) with a mean difference of 1.7 × 109/l (95% CI: 1.3–2.1). Similarly, the average ANC nadir after consolidation HDM (mean of four courses) was higher for 24 of the

25 patients with IR-ALL compared to the mANC nadirMT-HDM (P , 0.001) with a mean difference of 0.8 × 109/l (95% CI: 0.4–1.2) (Figure 2a). The average thrombocyte count nadir experienced by individual patients following HDM given as Figure 1 The neutrophil nadir following high-dose methotrexate consolidation therapy was for 19 of 21 patients with SR-ALL in relation to the coadministration of oral 6-mercaptopurine. Two higher than the mPLATE-nadir (P , 0.001) with a mean patients had unexplained granulocytosis prior to and during three MT-HDM × 9 HDM courses. difference of 143 10 /l (95% CI: 98–188), and similarly higher for 21 of 25 patients with IR-ALL (P , 0.001) with a mean difference of 83 × 109/l (95% CI: 55–111) (Figure 2b). Results

The median WBC nadir after HDM was significantly lower Predictive factors for myelotoxicity after HDM during during maintenance therapy (median: 1.6 × 109/l; 75% range: maintenance therapy 0.9–2.9) compared to SR consolidation (median: 5.1 × 109/l; =− = 75% range: 3.3–6.9) and IR consolidation (median: 2.7 × In univariate analysis, both ANC nadir (rS 0.20, P 0.007) =− = 109/l; 75% range: 1.8–4.4) (Figure 1). Similarly, comparing SR and thrombocyte nadir (rS 0.13, P 0.07), was negatively consolidation, IR consolidation, and maintenance therapy, related to the dose of 6MP at the time of HDM. In contrast, ANC nadir, lymphocyte count nadir, thrombocyte count the average dose of 6MP for individual patients during mainte- = nadir, and hemoglobin nadir after HDM decreased signifi- nance therapy was positively related to mWBC (rS 0.54), to = = cantly with increasing doses of oral 6MP (Table 1). The mANC mANC (rS 0.48), and to average thrombocyte counts (rS 0.48). This reflected that during the last part of maintenance nadirMT-HDM and mPLATE nadirMT-HDM were significantly = , therapy the patients with the least myelotoxicity would be pre- related (rS 0.63, P 0.001). Only 3% of the HDM courses given as SR-ALL consolidation was followed by an ANC nadir scribed the highest dose of 6MP. The dose of 6MP prior to ,0.5 × 109/l compared to 9% given as IR-ALL consolidation, HDM was not correlated to the time spend on maintenance = = and 50% of the HDM courses given during maintenance ther- therapy (rS 0.03, P 0.22). apy. Similarly, only 13% of the HDM courses given as SR- The degree of neutropenia following HDM did not correlate ALL consolidation induced a thrombocyte count nadir ,100 significantly with the time spent on maintenance therapy =− = × 109/l compared to 15% of the HDM given as IR-ALL consoli- either for children with SR-ALL (rS 0.03, P 0.79) or for =− = dation, and 58% of the HDM courses given during mainte- those with IR-ALL (rS 0.16, P 0.10). nance therapy. Due to neutropenia and/or thrombocytopenia, We found no significant difference in the degree of neutro- 6MP/MTX maintenance therapy was discontinued following penia following HDM during maintenance therapy for SR- vs × 96 of the 230 HDM courses given during maintenance ther- IR-ALL (median (75% range) mANC nadirMT-HDM; SR: 0.56 109/l (0.22–1.3), IR: 0.74 × 109/l (0.32–1.5); P = 0.49). Simi- larly, there was no significant difference in the degree of Table 1 Bone marrow toxicity following high-dose methotrexate thrombocytopenia following HDM during maintenance ther-

apy for SR- vs IR-ALL (median (75% range) mPLATE nadirMT- × 9 × 9 = Median nadir SR IR Maintenance HDM; SR: 95 10 /l (33–165), IR: 108 10 /l (39–191); P (75% range) consolidation consolidationa therapyb 0.52). The mANC nadirMT-HDM and the mPLATE nadirMT-HDM 10–20 days no 6MP 6MP: 6MP: = = were related to mANC (rS 0.49, P 0.001) and to mPLATE from start of HDM 25 mg/m2 75 mg/m2 = = (rS 0.47, P 0.01), respectively. Thus, the degree of neutropenia and thrombocytopenia-induced HDM reﬂected the White cell count 5.1 (3.3–6.9) 2.7 (1.8–4.4) 1.6 (0.9–2.9) Neutrophil count 1.9 (1.0–4.0) 1.2 (0.6–2.4) 0.5 (0.1–1.6) relative bone marrow suppression in relation to the patient’s Lymphocyte count 1.9 (1.1–2.9) 0.9 (0.5–1.5) 0.7 (0.4–1.2) average blood counts. Thrombocyte count 240 (94–382) 203 (74–303) 80 (13–211) With multivariate linear regression analysis, we tested the Hemoglobin 6.6 (5.4–7.3) 6.8 (6.0–7.5) 5.8 (5.0–6.9) impact on ANC nadir (or thrombocyte nadir) following HDM during maintenance therapy of gender (0 = girls vs 1 = boys), HDM, high-dose methotrexate; 6MP, 6-mercaptopurine; SR/IR, age at diagnosis, risk group (0 = SR, 1 = IR), number of days standard/intermediate risk leukemia. on maintenance therapy, mANC (or mPLATE), dose of 6MP, a + Prior to HDM 6MP, patients with IR-ALL received as consolidation E-6TGN and E-MTX levels prior to HDM, serum MTX concen- oral 6MP (75 mg/m2/day for 6 weeks), cyclophosphamide (1 g/m2 × 2), four series of low-dose cytarabine (75 mg/m2 × 16 in total), tration at the end of the MTX infusion, and 42 h serum MTX and i.s. MTX (×2). concentration. The regression analyses were performed in two bMTX/6MP maintenance therapy with HDM reinductions was similar ways: (1) only a single data set was used for each child using for SR- and IR-ALL. the mean neutrophil and thrombocyte nadirs following HDM,

Leukemia Bone marrow toxicity after high-dose methotrexate K Schmiegelow and U Bretton-Meyer 77

Figure 2 (a) Average ANC nadirs following HDM during consolidation and maintenance therapy for 21 patients with SR-ALL (----) and 25 patients with IR-ALL (—). For each patient mean values were calculated for all the HDM courses given during the specific therapy phase. (b) Average thrombocyte nadirs following HDM during consolidation and maintenance therapy for 21 patients with SR-ALL (----) and 25 patients with IR-ALL (—–). For each patient mean values were calculated for all the HDM courses given during the specific therapy phase. and the mean of the individual parameters to be tested in the tine monitoring of serum MTX concentrations is an integrated regression analyses; or (2) each of the HDM courses were part of HDM therapy. However, apart from the myelotoxicity regarded as independent. In the first alternative, only mANC in patients who have delayed MTX clearance, most cases of and the mPLATE, respectively, were found to be related to the clinically significant bone marrow toxicity leading to treat- average degree of neutropenia and thrombocytopenia follow- ment withdrawal have been unpredictable.26 ing HDM. In the second alternative, the best-fit model to pre- HDM and low-dose oral 6MP can interact in several ways. dict ANC nadir following HDM during maintenance therapy MTX is an inhibitor of the enzyme dihydrofolate reductase,27 included in the following order: dose of 6MP prior to HDM (β which converts folates to their active reduced (tetra- =−0.017, P = 0.001), average ANC level during maintenance hydrofolate) form. The subsequent accumulation of dihydro- therapy (β = 0.82, P = 0.004), and E-6TGN (β =−0.0029, P folate can inhibit de novo purine synthesis.15,28 In addition, = 0.02). The age of the patient (β = 0.19, P = 0.09) and the MTX can inhibit purine de novo synthesis by virtue of the risk group (β =−0.19, P = 0.08) were of borderline signifi- polyglutamate metabolites which directly inhibit aminoimida- cance. Thus, those who received the highest doses of 6MP, zole carboxamide ribonucleoside monophosphate.28 Finally, those with the lowest average ANC during maintenance ther- low-dose MTX as well as HDM can inhibit xanthine oxidase apy, and those with the highest E-6TGN levels prior to HDM and enhance the bioavailability of 6MP.14,29 These interac- had the most severe degree of neutropenia following HDM. tions between 6MP and MTX were indicated by clinical stud- Similarly, we tested the impact on thrombocyte count nadir ies more than 30 years ago when 6MP/MTX combination following HDM. The best-fit model to predict thrombocyte maintenance therapy for childhood ALL was shown to be nadir following HDM during maintenance therapy included superior to monotherapy with either MTX or 6MP.30,31 In only mPLATE (β = 0.0057, P = 0.046). Thus, those with the recent years both in vitro and in vivo studies have supported lowest average thrombocyte level during maintenance therapy that the interaction between 6MP and MTX is of clinical sig- also had the most severe thrombocytopenia following HDM. nificance.2,15,20,32,33 However, this is the first study which, in detail, explores the influence of 6MP dose and pharmacokinetics on the myelotoxicity of HDM. Discussion Theoretically, treatment intensity prior to administration of HDM could, to some extent, have influenced the degree of HDM has been given to children with ALL since the 1960s to myelotoxicity. However, several findings make it unlikely that reduce the risk of systemic as well as extramedullary relapse. this explains our results: (1) there was no significant difference A number of clinical studies during the last 10 to 15 years has in the degree of neutropenia and thrombocytopenia experi- increased our pharmacokinetic and pharmacodynamic under- enced by the SR and IR group during maintenance therapy in standing of low- and high-dose MTX.24,25 However, most of spite of the difference in their consolidation therapy (in fact these studies have focused on MTX pharmacokinetics only, myelosuppression was slightly more pronounced for SR whereas the possible pharmacodynamic interaction between patients); (2) the degree of myelosuppresision was more pro- MTX and other drugs has received far less attention. nounced for SR patients on maintenance therapy compared Patients with delayed clearance have long been known to to IR patients on consolidation who have previously received be at increased risk for myelotoxicity, and for this reason rou- combinations of cyclophosphamide, low-dose ARA-C and

Leukemia Bone marrow toxicity after high-dose methotrexate K Schmiegelow and U Bretton-Meyer 78 6MP; (3) the degree of neutropenia after HDM did not pro- 2 Schmiegelow K, Schrøder H, Gustafsson G, Kristinsson J, gress during maintenance therapy, which indicates that repeti- Glomstein A, Salmi T, Wranne L. Risk of relapse in childhood tive HDM courses in themselves do not progressively reduce acute lymphoblastic leukemia is related to RBC methotrexate and mercaptopurine metabolites during maintenance chemotherapy. the bone marrow reserve; and (4) the significant influence of Nordic Society for Pediatric Hematology and Oncology. J Clin the dose of 6MP during maintenance therapy on the degree Oncol 1995; 13: 345–351. of myelotoxicity following HDM supports that the coadminis- 3 Waters TR, Swann PF. Cytotoxic mechanism of 6-thioguanine: tration of 6MP during IR consolidation and during SR/IR hMutS-alfa, the human mismatch binding heterodimer, binds to maintenance therapy is a major determinant of the risk of DNA containing S6-methylthioguanine. Biochemistry 1997; 36: HDM-induced myelotoxicity. 2501–2506. The correlation between the dose of 6MP and ANC nadir 4 Lennard L. The clinical pharmacology of 6-mercaptopurine. Eur J Clin Pharmacol 1992; 43: 329–339. even with the E-6TGN levels included in the linear regression 5 Lennard L, Lilleyman JS. Variable mercaptopurine metabolism and model indicates (1) that other 6MP metabolites such as the 6- treatment outcome in childhood lymphoblastic leukemia methyl-thioinosine 59-monophosphate could have influenced (published erratum appears in J Clin Oncol 1990; 8: 567). J Clin the degree of myelotoxicity, since these are strong inhibitors Oncol 1989; 7: 1816–1823. of purine de novo synthesis, and/or (2) that E-6TGN levels do 6 Weinshilboum RM, Otterness DM, Szumlanski CL. Methylation not sufficiently reflect DNA-6TGN levels. In support hereof, pharmacogenetics; catechol O-methyltransferase, thiopurine we recently demonstrated that the risk of secondary myelodys- methyltransferase and histamine N-methyltransferase. Annu Rev Pharmacol Toxicol 1999; 39: 19–52. plasia or myeloid leukemia following 6MP/MTX maintenance 7 Frankel LS, Wang YM, Shuster J, Nitschke R, Doering EJ, Pullen J. therapy of childhood ALL was related not only to E-6TGN lev- High-dose methotrexate as part of remission maintenance therapy els but also to the intracellular levels of methylated 6MP for childhood acute lymphocytic leukemia: a Pediatric Oncology metabolites, which could reflect an increased incorporation Group pilot study. J Clin Oncol 1983; 1: 804–809. of DNA-6TGN due to inhibition of de novo purine synthesis 8 Abromowitch M, Ochs J, Pui CH, Fairclough D, Murphy SB, Riv- mediated by these methylated metabolites.17 era GK. Efficacy of high-dose methotrexate in childhood acute Seemingly contradictory findings are: (1) the negative corre- lymphocytic leukemia: analysis by contemporary risk classi- fications. Blood 1988; 71: 866–869. lation between the individual dose of 6MP prior to HDM dur- 9 Pui CH. Childhood leukemias. New Engl J Med 1995; 332: ing maintenance therapy and the ANC and thrombocyte 1618–1630. nadirs, and (2) the positive correlation between the average 10 Rask C, Albertioni F, Bentzen SM, Schrøder H, Peterson C. Clinical dose of 6MP during maintenance therapy and mANC and and pharmacokinetic risk factors for high-dose methotrexate- mPLATE. However, whereas the latter reflect the tolerance to induced toxicity in children with acute lymphoblastic leukemia – standard doses of 6MP, which for a large part are based on a logistic regression analysis. Acta Oncol 1998; 37: 277–284. inter- and intraindividual variations in 6MP pharmacokinetics, 11 Peeters M, Koren G, Jakubovicz D, Zipursky A. Physician com- pliance and relapse rates of acute lymphoblastic leukemia in chil- the former reflects the impact of the dose of 6MP on HDM- dren. Clin Pharmacol Ther 1988; 43: 228–232. induced bone marrow toxicity when the differences in the 12 Schmiegelow K. Prognostic significance of methotrexate and 6- average ANC and thrombocyte count levels are also taken into mercaptopurine dosage during maintenance chemotherapy for account (as demonstrated by the multivariate regression childhood acute lymphoblastic leukemia (published erratum analyses). Thus, the impact of a certain dose of 6MP prior to appears in Pediatr Hematol Oncol 1992; 9: 198). Pediatr Hematol HDM will be stronger for a patient with a poor tolerance to Oncol 1991; 8: 301–312. 6MP than for a patient with a high average neutrophil level 13 Relling MV, Hancock ML, Boyett JM, Pui CH, Evans WE. Prognos- tic importance of 6-mercaptopurine dose intensity in acute lym- at standard doses of 6MP. phoblastic leukemia. Blood 1999; 93: 2817–2823. Treatment interruptions following HDM should, if possible, 14 Balis FM, Holcenberg JS, Zimm S, Tubergen DG, Collins JM, Mur- be avoided since at least theoretically these periods of treat- phy RF, Gilchrist GS, Hammond D, Poplack DG. The effect of ment withdrawals could increase the risk of treatment fail- methotrexate on the bioavailability of oral 6-mercaptopurine. Clin ure.12,13 But how clinically significant myelosuppression fol- Pharmacol Ther 1987; 41: 384–387. lowing HDM can be avoided is not clear. The data of the 15 Bo¨kkerink JPM, Bakker MAH, Hulscher TW, DeAbreu RA, present study indicate that reductions of the dose of oral 6MP Schretlen EDAM. Purine de novo synthesis as the basis of syner- gism of methotrexate and 6-mercaptopurine in human malignant prior to the administration of HDM rather than reductions of lymphoblasts of different lineages. Biochem Pharmacol 1988; 37: the dose of HDM could be one approach, although this needs 2321–2327. to be confirmed in clinical trials. 16 Gustafsson G, Kreuger A, Clausen N, Garwicz S, Kristinsson J, Lie SO, Moe PJ, Perkkio M, Yssing M, Saarinen PU. Intensified treatment of acute childhood lymphoblastic leukaemia has improved Acknowledgements prognosis, especially in non-high-risk patients: the Nordic experi- ence of 2648 patients diagnosed between 1981 and 1996. Nordic The commitment and skillful technical assistance of Jannie Society of Paediatric Haematology and Oncology (NOPHO). Acta Gregers, Kristine Nielsen, and Michael Timm are greatly Paediatr 1998; 87: 1151–1161. appreciated. The study has received financial support from 17 Thomsen JB, Schrøder H, Kristinsson J, Madsen B, Szumlanski C, The Carl and Ellen Hertz Foundation, The Danish Childrens Weinshilboum R, Andersen JB, Schmiegelow K. Possible carcino- Cancer Foundation, The Danish Cancer Society (grant Nos 91- genic effect of 6-mercaptopurine on bone marrow stem cells – 048, 92-017, 93-017, 95-100-28), The JPC Foundation, The relation to thiopurine metabolism. Cancer 1999; 86: 1080–1086. Lundbeck Foundation (38/99), and The Minister Erna Hamil- 18 Bruunshuus I, Schmiegelow K. Analysis of 6-mercaptopurine, 6- thioguanine nucleotides, and 6- thiouric acid in biological fluids ton Foundation. by high-performance liquid chromatography. Scand J Clin Lab Invest 1989; 49: 779–784. References 19 Kamen BA, Takach PL, Vatev R, Caston JD. A rapid, radiochem- ical-ligand binding assay for methotrexate. Anal Biochem 1976; 1 Lilleyman JS, Lennard L. Mercaptopurine metabolism and risk of 70: 54–63. relapse in childhood lymphoblastic leukemia. Lancet 1994; 343: 20 Andersen JB, Szumlanski C, Weinshilboum RM, Schmiegelow K. 1188–1190. Pharmacokinetics, dose adjustments, and 6-mercaptopurine/

Leukemia Bone marrow toxicity after high-dose methotrexate K Schmiegelow and U Bretton-Meyer 79 methotrexate drug interactions in two patients with thiopurine 28 Chabner BA, Allegra CJ, Curt GA, Clendeninn NJ, Baram J, Kou- methyltransferase deﬁciency. Acta Paediatr 1998; 87: 108–111. zumi S. Polyglutamation of methotrexate: is methotrexate a pro- 21 Schmiegelow K, Pulczynska MK. Maintenance chemotherapy for drug? J Clin Invest 1985; 76: 907–912. childhood acute lymphoblastic leukemia: should dosage be 29 Innocenti F, Danesi R, Di Paolo A, Loru B, Favre C, Nardi M, guided by white blood cell counts? Am J Pediatr Hematol Oncol Bocci G, Nardini D, Macchia P, Del Tacca M. Clinical and experi- 1990; 12: 462–467. mental pharmacokinetic interaction between 6-mercaptopurine 22 Siegel S, Castellan NJ. Non-parametric Statistics for the Behavioral and methotrexate. Cancer Chemother Pharmacol 1996; 37: Sciences. McGraw-Hill: Singapore, 1988. 409–414. 23 SPSS statistical software for Windows release 10.0.5. SPSS Inc. 30 Frei E, Karon M, Levin RH, Freireich EJ, Taylor RJ, Hananian J, 1999. Selawry O, Holland JF, Hoogstraten B, Wolman IJ, Abir E, Sawit- 24 Synold TW, Relling MV, Boyett JM, Rivera GK, Sandlund JT, Mah- sky A, Lee S, Mills SD, Burgert EOJ, Spurr CL, Patterson RB, moud H, Crist WM, Pui CH, Evans WE. Blast cell methotrexate- Ebaugh FG, James GW, Moon JH. The effectiveness of combi- polyglutamate accumulation in vivo differs by lineage, ploidy, and nations of antileukemic agents in inducing and maintaining methotrexate dose in acute lymphoblastic leukemia. J Clin Invest remission in children with acute leukemia. Blood 1965; 26: 1994; 94: 1996–2001. 642–656. 25 Evans WE, Relling MV, Rodman JH, Crom WR , Boyett JM, Pui 31 Lonsdale D, Gehan EA, Fernbach DJ, Sullivan MP, Lane DM, C-H. Conventional compared with individualized chemotherapy Ragab AH. Interrupted vs. continued maintenance therapy in for childhood acute lymphoblastic leukemia. New Engl J Med childhood acute leukemia. Cancer 1975; 36: 341–352. 1998; 338: 499–505. 32 Schmiegelow K, Schrøder H, Schmiegelow M. Methotrexate and 26 Balis FM, Holcenberg JS, Bleyer WA. Clinical pharmacokinetics 6-mercaptopurine maintenance chemotherapy for childhood of commonly used anticancer drugs. Clin Pharmacokinet 1983; 8: acute lymphoblastic leukemia: dose adjustments by white cell 202–232. counts or by pharmacokinetic parameters. Cancer Chemother 27 White JC, Goldman ID. Mechanism of action of methotrexate. Pharmacol 1994; 34: 209–215. Free intracellular methotrexate required to suppress dihydrofolate 33 Giverhaug T, Loennechen T, Aarbakke J. The interaction of 6-mer- reduction to tetrahydrofolate by Erlich ascites tumor cells in vitro. captopurine (6-MP) and methotrexate (MTX). Gen Pharmacol Mol Pharmacol 1976; 12: 711–719. 1999; 33: 341–346.

Leukemia