Possible Implication of Thiopurine S-Methyltransferase In

Home , Thiopurine

Leukemia (2001) 15, 1706–1712  2001 Nature Publishing Group All rights reserved 0887-6924/01 $15.00 www.nature.com/leu Possible implication of thiopurine S-methyltransferase in occurrence of infectious episodes during maintenance therapy for childhood lymphoblastic leukemia with mercaptopurine T Dervieux1,YMe´dard1, P Verpillat2, V Guigonis1, M Duval3, B Lescoeur3, S Suciu4, E Vilmer3 and E Jacqz-Aigrain1

1Service de Pharmacologie Pe´diatrique et Pharmacoge´ne´tique, Hoˆpital Robert Debre´, Paris; 2Departement de Biostatistique et Epide´miologie, Groupe Hospitalier Bichat, Paris; 3Service d’He´mato-Immunologie, Hoˆpital Robert Debre´, Paris, France; and 4European Organization for Research and Treatment of Cancer (EORTC), Brussels, Belgium

6-Mercaptopurine (6-MP) is metabolized by thiopurine S- TGN) incorporated into DNA.2,3 Alternatively, mercaptopur- methyltransferase (TPMT), an enzyme subject to genetic poly- ine is catabolized into thiouric acid by xanthine oxidase morphism. We investigated the relationships between the TPMT locus (TPMT activity and genotype) and the pharmaco- (formation of thiouric acid) and into inactive methylmercapto- logical response to 6-MP during maintenance therapy of 78 purine by thiopurine methyltransferase (TPMT; E.C: 2.1.1.67). children with acute lymphoblastic leukemia (ALL). For each TPMT is a cytosolic enzyme subject to genetic polymor- patient, 6-MP dosage, leukocyte counts and occurrence of phism.4,5 According to the trimodal distribution of TPMT infectious episodes were monitored on an 8 week basis. Higher activity in red cells, approximately 89% of Caucasians are 6-MP dosage was associated with higher TPMT activity homozygous for the high activity allele (wild type), 11% are (P = 0.03) and higher average leukocyte counts (P Ͻ 0.01). Eight patients (10%) carrying a TPMT mutant genotype (one homo- heterozygous and about one subject in 300 inherits TPMT 4 zygous and seven heterozygous) received lower 6-MP doses deficiency as an autosomal recessive trait. Red blood cells (average: 48 vs 65 mg/m2/day; P = 0.02) and had on average (RBC) are convenient cells to study the thiopurine methyl- lower leukocyte counts (2834 vs 3398 cells/mm3; P = 0.003) than transferase polymorphism and the intracellular metabolism of patients carrying the wild-type TPMT genotype. Higher occur- mercaptopurine. RBC TPMT activity has been demonstrated rence of infectious episodes graded 2 or 3 was correlated with Ͻ to correlate with the TPMT activity measured in various cells higher 6-MP dosage (P 0.01) but no difference was observed 6 7 8 9 between TPMT mutants and TPMT wild-type patients. Patients or organs (lymphocytes, lymphoblasts, (kidney, liver ), and who received 6-MP dosage above the group median RBC 6-thioguanine nucleotide concentrations are proposed (62 mg/m2/day) or having a TPMT activity above the group for the monitoring of 6-MP during maintenance of ALL. Indi- median (21.5 nmol/h/ml) had a higher percentage of 8 week per- viduals inheriting a TPMT deficiency or patients heterozygous iods with infectious episodes requiring treatment (34% vs 17% at the TPMT locus exhibit hematological intolerance of stan- and 33% vs 19%, respectively) than those with 6-MP dose or Ͻ dard doses of mercaptopurine by a mechanism involving the TPMT activity below the group median (P 0.01). In the last 25 10–13 patients enrolled in the study, steady-state erythrocyte thiogu- accumulation of thioguanine nucleotides. However, the anine nucleotide (TGN) concentrations were associated with human thiopurine methyltransferase gene is now well charac- lower leukocyte counts (P = 0.01) but not with a higher occur- terized, and simple RFLP PCR can easily detect mutations rence of infectious episodes. In contrast, higher steady-state responsible for the TPMT deficiency.14,15 On the other hand, erythrocyte methylmercaptopurine nucleotide (MeMPN) con- little is known about the potential effects of mercaptopurine centrations were associated with higher 6-MP dosage Ͻ among patients exhibiting high TPMT activities. Studies by (P 0.01) and higher occurrence of infectious episodes 11 (P Ͻ 0.001). In conclusion, during maintenance therapy of ALL, Lennard and coworkers have shown that patients with high children with higher TPMT activity receive a higher 6-MP dos- RBC TPMT activity accumulate lower active TGN concen- age and may have infectious episodes caused by metabolism trations; thus, these individuals have the potential to be under- of 6-MP into methylmercaptopurine nucleotides. Leukemia treated with standard dosages of MP and are therefore more (2001) 15, 1706–1712. subject to relapse. Keywords: lymphoblastic leukemia; childhood; 6-mercaptopurine; The present study was conducted during maintenance ther- thiopurine methyltransferase; pharmacogenetics apy of a group of children receiving methotrexate and 6-MP for the treatment of ALL. We investigated the relationship between TPMT pharmacogenetics and the pharmacological Introduction properties of mercaptopurine with management of treated patients by dose administration, leukocyte counts and The common treatment of acute lymphoblastic leukemia (ALL) infectious episode side-effects. is based on induction remission therapy followed by intensi- fication and maintenance therapy1 which consists of administration of daily 6-mercaptopurine (6-MP) with weekly Materials and methods methotrexate. In vivo, the thiopurine pro-drug 6-MP, is metabolized by Patients, treatment and study design three competing pathways. The anabolic pathway is metabolism of 6-MP by hypoxanthine guanine phosphorybosyl Children with ALL were enrolled at the time of diagnosis in transferase into thioinosine monophosphate that is sub- the European chemotherapy protocol 58881 approved by an sequently metabolized into 6-thioguanine nucleotides (6- institutional review committee for clinical trials. The ethic committee of Paris Bichat-Claude Bernard approved the INSERM protocol 95020 allowing additional blood sampling Correspondence: E Jacqz-Aigrain, Service de Pharmacologie Pe´diatrique et Pharmacogeńe´tique, Hoˆpital Robert Debre´, 48 boulevard for pharmacogenetic studies. The parents of our patients Se´rurier, 75019 Paris, France; Fax: (33) 1 40 03 47 59 signed written informed consents. After 27 weeks of Received 9 January 2001; accepted 25 June 2001 induction/consolidation therapy (described in detail TPMT in maintenance therapy for childhood ALL T Dervieux et al 1707 elsewhere16), maintenance therapy started. This consisted of G460A and A719G mutations were presumed to be combined daily 6-mercaptopurine (started at the standard dose of on the same alleles (TPMT*3A). 50 mg/m2 orally) and weekly methotrexate (20 mg/m2 orally). The total duration of the planned maintenance treatment was 72 weeks (nine 8 week periods). Mercaptopurine dosage was Erythrocyte thiopurine nucleotides concentration adapted to maintain leukocyte count within the range of 2000 to 3000 cells/mm3. 6-MP dose was increased by 25% if two Red blood cell preparation and liquid chromatographic con- consecutive leukocyte counts at 2 weeks interval were above ditions were similar to the procedure described previously.19 4000 cells/mm3, and reduced by 25% if the leukocyte count Intra- and inter-day coefficients of variation were less than reached values below 1500 cells/mm3. All patients received 10% for both 6-thioguanine (6-TGN) and methylmercapto- trimethoprime–sulfamethoxazole for prophylaxis against purine (Me6MPN) nucleotides, respectively. Results are Pneumocystis carinii. expressed as pmol/8 × 108 red blood cells. For each patient, the cumulative 6-MP dose (expressed in mg) administered orally for 8 consecutive weeks (ie 8 week period) was determined sequentially during the entire mainte- Analysis of the study nance period and concomitantly with the occurrence of infections and leukocyte counts. Leukocyte counts were determ- For each sequential period of 8 weeks, the 6-MP dosage ined each week or every 2 weeks. The infection episodes were (mg/m2/day) administered to each patient was calculated by graded according to the following World Health Organization dividing the cumulative dose received (in mg) during the per- Toxicity Grading: grade 0 (absence of infection), grade 1 iod by the body surface area determined at the start of each (minor infection requiring no treatment), grade 2 (moderate period and by the number of days (56 days). infection requiring treatment), grade 3 (major infection requir- Similarly, the mean white blood cell count within each 8 ing treatment and hospitalization) or grade 4(major infection week period was calculated using all measurements collected. with hypotension). If more than one infectious episode Subsequently, we calculated the average 6-MP dosage occurred within an 8 week period, the highest grade was (mg/m2/day) and average leukocyte counts recovered during assigned to the period. If an infectious episode occurred at the the entire maintenance therapy using the values calculated end an 8 week period and continued at the beginning of the during the time periods (N) the patient was followed. The per- following 8 week period, the infectious episode was only centage of 8 week periods per patient having a mean white assigned to the first 8 week period. In all cases the toxicity blood cell counts within a range of 2000 to 3000 cells/mm3 grading was determined by the Pediatric Oncology Center. was also calculated. Because the grading system used could TPMT activity and thiopurine nucleotide concentrations were be dependent upon the perceived need of the physician for measured during follow-up visits after at least 4consecutive antibiotics and or hospitalization, for each patient we determ- weeks of unchanged 6-MP dosage (steady state). RBC TPMT ined the percentage of 8 week periods without occurrence of activities were excluded from the analysis if the patients had infectious episode by dividing the number of periods where received a blood transfusion in the 6 weeks preceding no infections occurred (grade 0) over the total number of per- blood sampling. iods the patient was followed. Also, the percentage of 8 week periods with infectious episodes graded 2 or more was calculated. RBC TPMT activity Statistical analysis was performed with StatView and SAS software (SAS Institute, Cary, NC, USA). For each patient, the Blood sample preparation and incubation conditions were average 6-MP dose, average leukocyte counts, percentage of similar to those previously described by Jacqz-Aigrain et al17 periods with maintenance of leukocyte count within the range and Me´dard et al.18 Briefly, the amount of methyl 6-MP for- of 2000 to 3000 cells/mm3 (%), percentage of 8 week periods med during incubation of erythrocyte lysates in the presence with infectious episodes graded 2 or 3 or percentage of per- of 6-MP and S-adenosylmethionine (as the methyl donor) was iods without infections were calculated as described above measured by HPLC. Results were expressed as nmol of methyl and used as individual values in all subsequent analyses. 6-MP formed per hour and per milliliter of packed red blood Averages were compared by the non-parametric Mann– cells. The intra-assay and inter-assay coefficients of variation Whitney test or Wilcoxon test and frequencies by the Fisher over the concentration range were less than 10%. A quality exact test. The correlations between quantitative variables control lysate was included in each assay run. Mean within were evaluated with Spearman rank tests. Linear multivariate day TPMT activity was 13.0 nmol/h/ml (coefficient of vari- regression models were applied to adjust for potential cofoun- ation: 2.6%, n = 5). Mean between days TPMT activity was ders. All significance levels reported were two-sided and set 13.6 nmol/h/ml (coefficient of variation: 5.2%, n = 8). at P Ͻ 0.05.

TPMT genotypes Results

Total genomic DNA was isolated from peripheral blood leu- Patient treatment and pharmacological parameters kocytes using a conventional chloroform/phenol extraction procedure. The mutations in exon 5 (G238C), exon 7 (G460A) Between August 1990 and February 1998, 78 children (47 and exon 10 (A719G) were detected by RFLP-PCR assays, as boys and 31 girls) aged 1 to 17 years (5.5 ( 3.1; average ± s.d.) reported previously.15 The homozygote TPMT wild-type and weighing from 6 to 66 kg (19.4 ± 1.04; average ± s.d.) at allele, was called TPMT*1, alleles presenting the G238C the time of diagnosis were included. All patients had received mutation were called TPMT*2, alleles presenting the A719G at least 5 months of maintenance therapy and 55 patients mutation were called TPMT*3C and the detection of both (70%) had completed the treatment at the time of the analysis.

Leukemia TPMT in maintenance therapy for childhood ALL T Dervieux et al 1708 The median duration of the maintenance period for the whole Median 6-TGN concentration was 333 pmol/8.0 × 108 cells group of patients was 16.8 months (range 5.6–20.5). (range 173–1334; n = 25) and median MeMPN concentration A total of 6308 week periods (median of nine per patient) was 5749 pmol/80 × 108 cells (range 20–19682; n = 25). were recorded, 3353 leukocyte counts were measured (median of 45 per patient), and 198 RBC TPMT activities were determined (median of two per patient). For each patient, the Relationship between mercaptopurine dosage, TPMT average values for 6-MP dosage, the weekly methotrexate, activity/genotype and leukocyte counts TPMT activity, leukocyte count and percentage of 8 week periods with successful maintenance of leukocyte count within In a univariate analysis, higher average mercaptopurine dose the range of 2000 to 3000 cells/mm3 were calculated. The correlated with higher average leukocyte counts (r = 0.52; infectious episodes occurred in a total of 306 8 week periods P Ͻ 0.0001), higher TPMT activity (r = 0.24; P = 0.03) and in which 47% (n = 143) were graded 1, 43% (n = 132) were higher TPMT activity correlated with higher average leukocyte graded 2, and 10% (n = 31) were graded 3. No grade 4infec- counts (r = 0.31; P = 0.006) (Figure 1). In a multivariate linear tious episodes occurred. The percentage of 8 week periods regression analysis for leukocyte counts with 6-MP and TPMT without occurrence of infectious episodes or with infectious activity as covariates, only 6-MP dosage remained signifi- episodes graded у2 (requiring treatment) was calculated for cant (P Ͻ 0.0001). each patient. Results are summarized in Table 1. Patients carrying a TPMT mutant genotype (heterozygous TPMT genotypes were determined in 75/78 patients. The and homozygous mutant) received a lower mercaptopurine RBC TPMT activity in the three patients with no DNA avail- dose (P = 0.02) and had lower average leukocyte counts than able ranged from 25.3 to 27.9 nmol/h/ml and they were con- patients having the wild-type TPMT genotype (P = 0.02 and sidered homozygous wild type. Under these conditions, 70 P = 0.003, respectively) (Figure 2). patients (89.7%) were homozygous wild type, seven patients Children who received mercaptopurine doses above the (9%) were heterozygous (TPMT*1/*2: n = 2, TPMT*1/*3A: group median (62 mg/m2/day) presented significantly lower n = 4, TPMT*1/*3C: n = 1) and one patient (1.2%) was homo- percentage periods with successful maintenance of leukocyte zygous mutant (TPMT*3C/*3C). The RBC TPMT activity of counts in the range 2000 to 3000 cells/mm3 than patients who patients carrying point mutations at the TPMT locus was sig- received mercaptopurine doses below the group median nificantly lower than that of patients having a wild-type TPMT (average ± s.d.: 47.9 v± 18.7% vs 65.4 ± 17.6%; P Ͻ 0.001). genotype (average ± s.d.: 10.7 ± 5.0 vs 22.0 ± 5.6; Similarly, children having TPMT activity above the group P Ͻ 0.0001). The TPMT activity measured in the patient hav- median (21.5 nmol/h/ml) had significantly lower percentage ing the homozygous mutant genotype (TPMT*3C/*3C) was periods with successful maintenance of leukocyte counts in 3.2 nmol/h/ml. the range of 2000 to 3000 cells/mm3 than patients with TPMT A total of 41 RBC thiopurine nucleotide concentrations activity below the group median (average ± s.d.: 50.2 ± 19.4% were determined in the last 25 patients included in the trial vs 63.0 ± 18.6%; P = 0.002). (average of 1.6 measurement per patient). The 6-MP dose administered during the 4weeks prior to blood sampling was not different from the average 6-MP dose administered during the entire maintenance period. Therefore, 6-TGN and Relationship between mercaptopurine dose, TPMT MeMPN concentrations were considered at steady state. activity/genotype and infectious episodes

The percentage of 8 week periods with infectious episodes Table 1 Pharmacological parameters in the 78 children during graded у2 did not vary significantly over the maintenance continuation therapy of ALL period (P = 0.21; Figure 3). Methotrexate dose administered was not associated with the percentage of 8 week periods with Median Range infectious episodes graded у2(P = 0.23), or with the percentage of 8 week periods with no infectious episodes (P = 0.80). Duration of the maintenance 16.8 5.6–20.5 When comparing patients carrying the TPMT mutant geno- therapy (months) type and patients having the wild-type TPMT genotype, no Average 6-mercaptopurine 62 26–123 2 significant difference in the percentage of periods with no dosage (mg/m /day) ± ± Average methotrexate dosage 19 8–21 infectious episodes (average s.d.: 52.1 23.2% vs (mg/m2/week) 50.3 ± 24.6%; P = 0.78), or in the percentage of periods with Average RBC TPMT activity 21.5 3.3–41.2 infectious episodes graded у2 (average ± s.d.: 22.0 ± 21.9% (nmol/h/ml) vs 26.0 ± 20.0%; P = 0.56) was observed. The patients carry- Average leukocyte counts 3278 2194–5255 ing the TPMT homozygous mutant genotype had 41.3% per- (cells/mm3) у Percentage of periods with 57 0–100 iods with infectious episodes graded 2 and 29% periods maintenance of leukocyte count with no infectious episodes. within the range 2000–3000/mm3 Patients having received 6-MP dosage above group median (%) (62 mg/m2/day) had a significantly lower percentage of 8 week Percentage of periods without 50 0–100 periods with no infections (P = 0.03), and conversely, a two- infectious episodes (%) fold higher percentage of 8 week periods with infectious epi- Percentage of periods with 24 0–71 у = infectious episodes graded 2 or sodes graded 2(P 0.0005) than those having received 6- 3(%) MP dosage below the group median. Similarly, patients having TPMT activity above the group median (21.5 nmol/h/ml) The average value for each parameter was determined in each had a significantly lower percentage of 8 week periods with patient as described in Materials and methods. Results for the 78 no infections (P = 0.003) and a 1.7-fold higher occurrence of children are presented as median and range. 8 week periods with infectious episodes graded у2

Leukemia TPMT in maintenance therapy for childhood ALL T Dervieux et al 1709

Figure 2 Comparison of average mercaptopurine dosage administered (a), average leukocyte count (b), between TPMT mutant patients and patients having a wild-type genotype. Bars represent the average ± s.e.m. P values are given (Mann–Whitney test).

Figure 1 Relationship between average 6-MP dosage and average P = 0.007). However, in a multivariate linear regression analy- leukocyte counts (a), average 6-MP dosage and TPMT activity (b); sis including mercaptopurine dosage, TPMT activity and leu- average leukocyte counts and TPMT activity (c). Each point corre- kocyte counts, only mercaptopurine dosage was associated sponds to the average value of repeated measurements determined with the percentage of periods with infectious episodes graded during the maintenance in individual patients (ie average mercaptopu- у2(P = 0.03). rine dosage, average white blood cell count). (᭹) corresponds to patients carrying mutations at the TPMT locus. Spearman correlation coefﬁcients (r) with P value are given for each panel. The arrow points to the patient homozygous mutant. Steady-state erythrocyte thiopurine nucleotide concentrations

(P = 0.002) than those having TPMT activity below the group In the last 25 patients included in the trial, 6-TGN and median. Results are summarized in Figure 4. MeMPN concentrations were measured in RBC. 6-TGN con- In a univariate analysis, higher occurrence of infections epi- centrations measured in three patients carrying TPMT sodes graded у2 correlated with higher mercaptopurine doses mutations (two were heterozygous and one homozygous) administered (r = 0.40; P = 0.001), higher white blood cell were 3.3-fold higher (median: 1031 pmol/8 × 108 cells, range counts (r = 0.36; P = 0.01) and higher TPMT activity (r = 0.30; 918–1334) than those measured in the 22 patients having a

Leukemia TPMT in maintenance therapy for childhood ALL T Dervieux et al 1710

Figure 3 Percentage of 8 week periods with infectious episodes graded у2 (2 or 3) over the maintenance periods in the 78 patients. Bars represent the average with conﬁdence interval at 95% at each 8 week period.

wild-type genotype (median: 307 pmol/8 × 108 cells, range 173–559) (P Ͻ 0.001). In contrast, the 22 patients having a wild-type TPMT genotype had significantly higher MeMPN concentrations (median: 6137 pmol/8 × 108 cells, range 637– 19 682) than the three patients carrying TPMT mutations (median: 307 pmol/8 × 108 cells, range 20–336). Higher TPMT activity correlated significantly with lower 6- TGN concentrations (r =−0.48; P = 0.01) but not with higher MeMPN concentrations. 6-TGN concentrations were not associated with 6-MP dosage, percentage of periods with infectious episodes graded 2 or 3, or percentage of periods without infectious episodes. However, higher steady-state 6- TGN concentrations correlated significantly with lower average leucocyte counts (r =−0.56; P = 0.003) (Figure 5a). Steady-state MeMPN concentrations correlated with higher 6- MP dosage (r = 0.64; P = 0.0004), higher percentage of periods with infectious episodes graded у2(r = 0.50; P = 0.01) Figure 4 Comparison of percentage of 8 week periods with no and lower percentage of periods without infectious episodes у =− = infectious episodes (a) or with infectious episodes graded 2 (b) in (r 0.56; P 0.003) (Figure 5b). Finally, in a multivariate lin- patients who received an average 6-MP dose below or above the ear regression analysis including 6-MP dosage and MeMPN median group (62 mg/m2/day) or, had erythrocyte TPMT activity concentrations, higher MeMPN concentrations remained below or above the median group (21.5 nmol/h/ml). Bars represent associated with the occurrence of infectious episodes the average with confidence interval at 95%. P values are given. The (P Ͻ 0.01). below median groups include the eight patients having a TPMT mutant genotype (one homozygous mutant and seven heterozygous) plus 31 patients having a wild-type TPMT genotype; the above median groups include 39 patients having a TPMT wild-type genotype. Discussion

6-MP is an antileukemic agent administered during the main- between the thiopurine methyltransferase locus and response tenance phase of childhood ALL.1 The objective is to maintain to mercaptopurine treatment. In our study, TPMT activities the complete remission obtained after several courses of inten- were highly variable and, as expected, patients with higher sive chemotherapy for approximately 6 months (induction TPMT activity received higher doses of 6-MP. According to phase). The inter-individual variation in the antileukemic the therapeutic protocol, the mercaptopurine dose was effect of mercaptopurine is in part related to variations of the increased to achieve the antileukemic effect. Unexpectedly, thiopurine methyltransferase activity.11,12,20 It is generally we observed a positive relationship between higher TPMT accepted that patients with high TPMT activity require an activity, higher mercaptopurine doses and higher leukocyte increase of their mercaptopurine dosage to overcome mercap- counts. The inﬂuence of TPMT genotype was further investi- topurine catabolism and obtain an antileukemic effect,12,21 gated. Children carrying a TPMT mutant genotype received a and that patients with low TPMT activity or a TPMT mutant lower mercaptopurine dose and had lower leukocyte counts genotype require dosage reduction to avoid severe hematolog- than patients with a wild-type genotype. In addition, patients ical toxicity due to the extensive formation of cytotoxic with a TPMT mutant genotype had higher steady-state red thioguanine nucleotides.12,22 blood cell thioguanine nucleotide concentrations, which were We prospectively followed 78 children during the mainte- associated with lower leukocyte counts. This is in accordance nance therapy of ALL and investigated the relationship with previous reports, demonstrating greater leukopenia sec-

Leukemia TPMT in maintenance therapy for childhood ALL T Dervieux et al 1711 TPMT genotype as demonstrated previously.12 Here again, in a multivariate analysis including 6-MP dosage, TPMT activity, and leukocyte counts, only 6-MP remained significantly associated with the occurrence of infectious episodes. Therefore, 6-MP appears as the risk factor of infections during maintenance therapy in patients with ALL. As previously stated, 6-MP is metabolized into 6-thioguanine nucleotides (6-TGN) and alternatively catabolized into inactive methylmercaptopurine by thiopurine methyltransferase. Besides mercaptopurine, TPMT also methylates thiopurine nucleotides into methylthiopurine nucleotides (ie methylmercaptopurine nucleotides). In vitro studies have shown that methylmercaptopurine nucleotides contribute to the antipro- liferative properties of 6-MP,24 probably by inhibiting phos- phoribosylamidotransferase, the first enzymatic step in purine de novo synthesis.25,26 Clinical data have shown that TPMT- deficient patients can tolerate higher thioguanine nucleotide concentrations than patients with high TPMT, suggesting a deleterious effect of 6-methylmercaptopurine nucleotides.10,20 In addition, lymphocytes having high TPMT activity have greater inhibition of mitogen stimulation after mercaptopurine treatment than lymphocytes with low TPMT activity.27 There- fore, the occurrence of infections associated with the formation of methylmercaptopurine nucleotides may be related to the inhibition of de novo purine synthesis, already reported as the mechanism of action of immunosuppressive agents such as mycophenolate.28 Although TGN concentrations were not related to the occurrence of infectious episodes, patient having a TPMT mutant genotype may also be at high risk of infectious episodes, as the treatment induces neutropenia and myelotoxicity. This is supported by the fact that our patient with a TPMT homozygous mutant genotype had high TGN concentrations, no detectable MeMPN concentrations, and high occurrence of infectious episodes. The analysis of the type of infection was not planned in the present protocol but we speculate that infectious episodes occurring during maintenance therapy may be related to high concentrations of methylmercaptopurine nucleotides in patients having high TPMT activity or to neutropenia induced by 6-thioguanine Figure 5 Relationships between (a) steady-state erythrocyte thiog- nucleotides in patients having low TPMT activity. uanine nucleotides and average leukocyte counts, (b) steady-state In conclusion, our results suggest that during the mainte- erythrocyte methylmercaptopurine nucleotide concentrations and percentage of periods with infectious episodes graded 2 or 3. Spear- nance therapy of ALL, children with high TPMT activity man correlation coefficients (r) with P values are given for each panel. receiving high doses of 6-MP may be at risk of infections, Lines correspond to the linear regression. The arrow points to the probably through metabolism of 6-MP into methylmercapto- patient TPMT homozygous mutant. purine nucleotides.

11,12,20 ondary to the formation of thioguanine nucleotides, and Acknowledgements with the concept that prospective identification of TPMT mutations may help to decrease excessive hematological tox- We thank the patients and their parents for participating in icity in individuals with low TPMT activity.22 this study, and the nurses for taking care of the patients. We Factors associated with white blood cell counts were further also thank Caroline Delnevo for secretarial assistance. This investigated in a multivariate analysis including TPMT activity work was supported by grants from AFSSAPS (Agence França- and 6-MP. Only 6-MP remained significant, showing that, 6- ise de Sećurite´ Sanitaire des Produits de Sante´) and INSERM MP dose remains the major determinant of the leukocyte (Institut National de la Sante´ et de la Recherche Me´dicale). counts during maintenance therapy. This suggests that, despite high 6-MP dosage, patients with high leukocyte counts may not have received 6-MP doses high enough to maintain the white blood cell counts in the range of 2000–3000/mm3.A References similar result was recently reported by Schiemegelow and Bretton-Meyer23 during continuation therapy of ALL. 1 Pui CH, Evans WE. Acute lymphoblastic leukemia. N Engl J Med In the present study, higher doses of mercaptopurine and 1998; 339: 605–615. 2 Elion GB. The purine path to chemotherapy. Science 1989; 24: higher TPMT activity were also associated with a higher 441–447. occurrence of infectious episodes. However, there was no 3 Warren DJ, Andersen A, Slordal L. Quantitation of 6-thioguanine association between the occurrence of infectious episodes and residues in peripheral blood leukocyte DNA obtained from

Leukemia TPMT in maintenance therapy for childhood ALL T Dervieux et al 1712 patients receiving 6-mercaptopurine-based maintenance therapy. Y, Dastugue N, Fournier M, Souillet G, Manel AM, Robert A, Cancer Res 1995; 55: 1670–1674. Nelken B, Millot F, Lutz P, Rialland X, Mechinaud F, Boutard P, 4Weinshilboum RM, Sladek SL. Mercaptopurine pharmacogenetics: Behar C, Chantraine JM, Plouvier E, Laureys G, Brock P, Uytteb- monogenic inheritance of erythrocyte thiopurine methyltransfer- roeck A, Margueritte G, Plantaz D, Norton L, Francotte N, Gysel- ase activity. Am J Hum Genet 1980; 32: 651–662. inck J, Waterkeyn C, Solbu G, Philippe N, Otten J. Long-term 5 McLeod H, Krynetski E, Relling MV, Evans WE. Genetic polymor- results of three randomized trials (58831, 58832, 58881) in child- phism of thiopurine methyltransferase and its clinical relevance hood acute lymphoblastic leukemia: a CLCG-EORTC report. Chil- for childhood acute lymphoblastic leukemia. Leukemia 2000; 14: dren Leukemia Cooperative Group. Leukemia 2000; 14: 2257– 567–572. 2266. 6 Van Loon JA, Weinshilboum RM. Thiopurine methyltransferase 17 Jacqz-Aigrain E, Bessa E, Medard Y, Mircheva Y, Vilmer E. Thiopu- biochemical genetics: human lymphocyte activity. Biochem Genet rine methyltransferase activity in a French population: h.p.l.c. 1982; 20: 637–658. assay conditions and effects of drugs and inhibitors. Br J Clin Phar- 7 McLeod HL, Relling MV, Liu Q, Pui C-H, Evans WE. Polymorphic macol 1994; 38: 1–8. thiopurine methyltransferase in erythrocytes is indicative of 18 Medard Y, Nafa S, Jacqz-Aigrain E. Thiopurine methyltransferase activity in leukemic blasts from children with acute lymphoblastic activity: new high-performance liquid chromatographic assay con- leukemia. Blood 1995; 85: 1897–1902. ditions. J Chromatogr B Biomed Sci Appl 1997; 700: 275–277. 8 Ferroni MA, Marchi G, Sansone E, Romeo P, Giulianotti PC, Pie- 19 Dervieux T, Boulieu R. Simultaneous determination of 6-thioguan- trabissa A, Mosca F, Pacifici GM. Variability in the rate of 6-mer- ine and methyl 6-mercaptopurine nucleotides of azathioprine in captopurine methylation in the erythrocytes, liver and kidney in red blood cells by HPLC. Clin Chem 1998; 44: 551–555. an Italian population. Eur J Clin Pharmacol 1996; 51: 23–29. 20 Relling MV, Hancock ML, Boyett JM, Pui C-H, Evans WE. Prognos- 9 Szumlanski CL, Honchel R, Scott MC, Weinshilboum RM. Human tic importance of 6-mercaptopurine dose intensity in acute lym- liver thiopurine methyltransferase pharmacogenetics: biochemical phoblastic leukemia. Blood 1999; 93: 2817–2823. properties, liver-erythrocyte correlation and presence of isozymes. 21 Lennard L, Lilleyman JS. Are children with lymphoblastic leu- Pharmacogenetics 1992; 2: 148–159. kaemia given enough 6-mercaptopurine? Lancet 1987; 2:785– 10 Evans WE, Horner M, Chu YQ, Kalwinsky D, Roberts WM. Altered 787. mercaptopurine metabolism, toxic effects, and dosage require- 22 McLeod H, Coulthard SA, Thomas AE, Pritchard S, King DJ, Rich- ment in a thiopurine methyltransferase-deficient child with acute ards SM, Eden OB, Hall AG, Gibson BE. Analysis of thiopurine lymphocytic leukemia. J Pediatr 1991; 119: 985–989. methyltransferase variant alleles in childhood actue lymphoblastic 11 Lennard L, Lilleyman JS, Van Loon J, Weinshilboum RM. Genetic leukaemia. Br J Haematol 1999; 105: 696–700. variation in response to 6-mercaptopurine for childhood acute 23 Schmiegelow K, Bretton-Meyer U. 6-mercaptopurine dosage and lymphoblastic leukaemia. Lancet 1990; 336: 225–229. pharmacokinetics influence the degree of bone marrow toxicity 12 Relling MV, Hancock ML, Rivera GK, Sandlund JT, Ribeiro RC, following high-dose methotrexate in children with acute lym- Krynetski EY, Pui C-H, Evans WE. Mercaptopurine therapy intoler- phoblastic leukemia. Leukemia 2001; 15: 74–79. ance and heterozygosity at the thiopurine S-methyltransferase 24Dervieux T, Blanco JG, Krynetski EY, Vanin E, Roussel MF, Relling gene locus. J Natl Cancer Inst 1999; 91: 2001–2008. MV. Differing contribution of thiopurine methyltransferase to mer- 13 Evans WE, Hon YY, Bomgaars L, Coutre S, Holdsworth M, Janco captopurine versus thioguanine effects in human leukemic cells. R, Kalwinsky D, Keller F, Khatib Z, Margolin J, Murray J, Quinn Cancer Res 2001; 61: 5810–5816. J, Ravindranath Y, Ritchey K, Roberts W, Rogers ZR, Schiff D, 25 Bokkerink JP, Stet EH, De Abreu RA, Damen FJ, Hulscher TW, Steuber C, Tucci F, Kornegay N, Krynetski EY, Relling MV. Prepon- Bakker MA, van Baal JA. 6-Mercaptopurine: cytotoxicity and bio- derance of thiopurine S-methyltransferase deficiency and heteroz- chemical pharmacology in human malignant T-lymphoblasts. ygosity among patients intolerant to mercaptopurine or azathiop- Biochem Pharmacol 1993; 45: 1455–1463. rine. J Clin Oncol 2001; 19: 2293–2301. 26 Vogt MH, Stet EH, De Abreu RA, Bokkerink JP, Lambooy LH, 14Krynetski EY, Fessing MY, Yates CR, Sun D, Schuetz JD, Evans Trijbels FJ. The importance of methylthio-IMP for methylmercapto- WE. Promoter and intronic sequences of the human thiopurine purine ribonucleoside (Me-MPR) cytotoxicity in Molt F4human S- methyltransferase (TPMT) gene isolated from a human PAC1 malignant T-lymphoblasts. Biochim Biophys Acta 1993; 1181: genomic library. Pharm Res 1997; 14: 1672–1678. 189–194. 15 Yates CR, Krynetski EY, Loennechen T, Fessing MY, Tai HL, Pui 27 Van Loon JA, Weinshilboum RM. Human lymphocyte thiopurine C-H, Relling MV, Evans WE. Molecular diagnosis of thiopurine S- methyltransferase pharmacogenetics: effect of phenotype on 6- methyltransferase deficiency: genetic basis for azathioprine and mercaptopurine-induced inhibition of mitogen stimulation. J Phar- mercaptopurine intolerance. Ann Intern Med 1997; 126: 608– macol Exp Ther 1987; 242: 21–26. 614. 28 Ransom JT. Mechanism of action of mycophenolate mofetil. Ther 16 Vilmer E, Suciu S, Ferster A, Bertrand Y, Cave H, Thyss A, Benoit Drug Monit 1995; 17: 681–684.

Leukemia