Additive effects of TPMT and NUDT15 on thiopurine toxicity in children with acute lymphoblastic leukemia across multiethnic populations

Abstract

Background

Thiopurines such as mercaptopurine (MP) are widely used to treat acute lymphoblastic leukemia (ALL). Thiopurine-S-methyltransferase (TPMT) and Nudix hydrolase 15 (NUDT15) inactivate thiopurines, and no-function variants are associated with drug-induced myelosuppression. Dose adjustment of MP is strongly recommended in patients with intermediate or complete loss of activity of TPMT and NUDT15. However, the extent of dosage reduction recommended for patients with intermediate activity in both enzymes is currently not clear.

Methods

MP dosages during maintenance were collected from 1768 patients with ALL in Singapore, Guatemala, India, and North America. Patients were genotyped for TPMT and NUDT15, and actionable variants defined by the Clinical Pharmacogenetics Implementation Consortium were used to classify patients as TPMT and NUDT15 normal metabolizers (TPMT/NUDT15 NM), TPMT or NUDT15 intermediate metabolizers (TPMT IM or NUDT15 IM), or TPMT and NUDT15 compound intermediate metabolizers (TPMT/NUDT15 IM/IM). In parallel, we evaluated MP toxicity, metabolism, and dose adjustment using a Tpmt/Nudt15 combined heterozygous mouse model (Tpmt^+/−/Nudt15^+/−).

Results

Twenty-two patients (1.2%) were TPMT/NUDT15 IM/IM in the cohort, with the majority self-reported as Hispanics (68.2%, 15/22). TPMT/NUDT15 IM/IM patients tolerated a median daily MP dose of 25.7 mg/m² (interquartile range = 19.0-31.1 mg/m²), significantly lower than TPMT IM and NUDT15 IM dosage (P < .001). Similarly, Tpmt^+/−/Nudt15^+/− mice displayed excessive hematopoietic toxicity and accumulated more metabolite (DNA-TG) than wild-type or single heterozygous mice, which was effectively mitigated by a genotype-guided dose titration of MP.

Conclusion

We recommend more substantial dose reductions to individualize MP therapy and mitigate toxicity in TPMT/NUDT15 IM/IM patients.

Thiopurines, including mercaptopurine (MP), are cytotoxic drugs used for the maintenance therapy of acute lymphoblastic leukemia (ALL), for which a prolonged drug exposure is required to effectively prevent relapse (1,2). However, these antimetabolite drugs are characterized by a narrow therapeutic index, with dose-limiting toxicity (eg, myelosuppression), which necessitate frequent dose adjustments and close monitoring of the hematopoietic function to prevent life-threatening complications (3-5).

As prodrugs, thiopurines undergo extensive intracellular metabolic activation into thioguanine nucleotides (6). These synthetic analogs compete with endogenous nucleotides for integration into the DNA double strand (as DNA-TG), which triggers futile mismatch repair, followed by DNA strand breaks, and eventually apoptosis (7). Thiopurine-S-methyltransferase (TPMT) and Nudix hydrolase 15 (NUDT15) are 2 key enzymes involved in the inactivation of thiopurines. Genetic variations in these 2 genes explain approximately 45% of the interpatient variability in thiopurine-induced myelosuppression (8-17). TPMT variants are found predominantly in White patients (18), whereas NUDT15 variants are prevalent in individuals of Hispanic or Asian origin (8-11).

Guidelines from the Clinical Pharmacogenetics Implementation Consortium (CPIC) recommend dose adjustments of thiopurines to mitigate drug toxicity in TPMT or NUDT15 intermediate metabolizers (IM) and poor metabolizers (PM), defined as individuals carrying 1 or 2 no-function alleles, respectively, in either TPMT or NUDT15 genes (19). These recommendations are classified at the highest level as “strong” recommendations, meaning they are supported by high-quality evidence. However, it is not clear whether patients who have no-function alleles in both genes, that is, TPMT and NUDT15 compound IM (TPMT/NUDT15 IM/IM), should receive even lower doses than those who are IM for only one of those genes. In lieu of adequate evidence, currently, they are recommended to receive the same starting dose as patients who are IM for TPMT alone or NUDT15 alone (ie, a 30%-80% decrease in the initial dose of MP) (19).

Because TPMT and NUDT15 modulates thiopurine metabolism through non-overlapping mechanisms, we hypothesized that the loss of both enzymes results in additive effects on drug toxicity; therefore, TPMT/NUDT15 IM/IM patients should require further dose reduction compared with TPMT or NUDT15 single IM patients, that is, patients who are TPMT IM/NUDT15 NM or TPMT NM/NUDT15 IM. To this end, we evaluated the dose tolerance of MP in TPMT/NUDT15 IM/IM patients with ALL across diverse international cohorts. In addition, we established a Tpmt and Nudt15 compound heterozygous murine model to show that further dose reduction can mitigate the toxicity caused by the reduced metabolism of MP.

Methods

Patients and thiopurine therapy

Children with newly diagnosed ALL were identified retrospectively from multiple frontline clinical trials internationally: the Guatemala LLAG-0707 study (Unidad Nacional de Oncología Pediátrica, UNOP, N = 230) (20), Malaysia-Singapore (Ma-Spore) group’s MS2003 and MS2010 protocols (N = 80) (21), Indian Postgraduate Institute of Medical Education and Research (PGIMER) ALL clinical protocol (N = 172), Children’s Oncology Group (COG) clinical trial AALL03N1 (N = 642) (22,23), St Jude Children’s Research Hospital protocol TOTAL16 (N = 490) (24), and children treated per COG AALL0932 (25), AALL1131 (26), and AALL1231 (27) protocols at Children’s Hospital Los Angeles (CHLA, N = 154). All protocols included daily MP as part of maintenance therapy, with a planned dosage ranging from 50 to 75 mg/m². Dose adjustments were allowed on the basis of the degree of myelosuppression or the occurrence of infections per standard of care practices or as dictated by protocol. The clinical characteristics (sex and age at ALL diagnosis) were provided by participating sites. At the time of study enrollment, patients and/or parents self-reported their race/ethnicity as Asian, Black, Hispanic, White, or “other” if not included in these categories (eg, patients with multiple races/ethnicities). Each clinical trial was approved by their respective institutional review boards, and written informed consents (or assent when appropriate) for enrollment and banking of specimens were obtained from parents, guardians, and/or patients. This study was approved by the St Jude Children’s Research Hospital Institutional Review Board.

Determination of MP dose tolerance

For each patient, the tolerated dosage of MP (mg/m²) was calculated as the average of the dose administered during the maintenance therapy (Supplementary Table 1, available online). Drug exposure was considered stable after at least 9 weeks (or one complete cycle) of maintenance therapy with daily MP dosing. Missed dosages were not included in the calculation of the average tolerated dosage. The protocol dose was specific to each participating institution. In addition, for TOTAL16 patients, the dosage was preemptively adjusted on the basis of TPMT metabolizer status (24). The dosage of MP for CHLA patients was adjusted based on TPMT and NUDT15 metabolizer status, when available. For patients included in the TOTAL16 clinical trial (24) and those treated at CHLA and PGIMER, the dose intensity (DI) of MP was determined by dividing the total cumulative prescribed dosage per patient by the total cumulative protocol-specified dosage for maintenance therapy. The DI of patients from the AALL03N1 clinical trial was calculated over 6 months of therapy (11). For patients treated per Ma-Spore protocols and at UNOP, DI was inferred by the average tolerated dosage divided by the protocol dosage (9).

TPMT and NUDT15 genotyping and phenotype assignment

Germline DNA was extracted from peripheral blood and was used for TPMT/NUDT15 genotyping or sequencing, according to the institution’s practices (see Supplementary Methods, available online). Actionable pharmacogenomic variants of TPMT and NUDT15 were identified using the CPIC allele functionality table (19): TPMT *2, *3A, *3B, and *3C and NUDT15 *2, *3, and *9 were used to assign phenotypes as normal metabolizer (NM), intermediate metabolizer (IM), or poor metabolizer (PM) for TPMT or NUDT15. Patients with compound IM phenotype were reported as TPMT/NUDT15 IM/IM.

Animals

Tpmt-knockout mice (Tpmt^−/−) on a 129x1/SvJ background and Nudt15-knockout mice (Nudt15^−/−) on an FVB/NJ background were established previously (28,29). Using a speed congenics approach, the FVB/NJ strain was backcrossed with a 129x1/SvJ strain (Jackson Labs, stock number 000691, Bar Harbor, ME) for 9 consecutive generations to obtain more than 99% 129x1/SvJ background similarity. Wild-type, heterozygous for either Tpmt or Nudt15 (Tpmt^+/− or Nudt15^+/−, respectively), and compound heterozygous (Tpmt^+/−/Nudt15^+/−) mice in a 129x1/SvJ background were used for this study. All animal experiments were approved by the Institutional Animal Care and Use Committee of St Jude Children’s Research Hospital.

MP metabolism and toxicity in mice

Female and male mice between 6 and 10 weeks of age were administered oral MP (cat. #852678, Sigma Aldrich) in drinking water at different dosage levels: 5, 10, or 20 mg/kg per day or 25, 50, or 100 mg/mL, respectively, assuming an average drinking volume of 5 mL per day (30). These dosages were equivalent to 15, 30, and 60 mg/m² per day in children, respectively (31). Hematopoietic toxicity was evaluated by complete blood count on peripheral blood collected by retro-orbital puncture performed before treatment, and after 5, 10, 17, and 30 days of MP (n = 10 mice per group). Mice were followed up for up to 70 days and were euthanized when body weight loss exceeded 20% (or they became moribund). The incorporation of the thioguanine metabolite deoxythioguanosine (dTG) into DNA (as DNA-TG) was evaluated in a separate group of mice euthanized after 5 days of treatment (n = 6 mice per group) (see Supplementary Methods, available online) (32,33).

Statistical analyses

All statistical tests were 2-sided and chosen as appropriate according to data distribution. In patients, the tolerated dose of MP was compared between groups with the nonparametric Wilcoxon rank-sum test. P less than .05 was used as the cutoff for statistical significance. The χ² test was used to assess homogeneity between groups, when applicable. A multivariable general linear model was applied to assess the association between MP dosage and TPMT and NUDT15 metabolizer phenotypes while controlling for the patients’ clinical characteristics and institutions. A meta-analysis using a random effect model on the basis of estimates and their standard errors was conducted to evaluate the effect of metabolizer status on MP dose intensity across cohorts (34). Mice hematopoietic parameters and thiopurine metabolite concentrations in vivo were compared using the Wilcoxon rank-sum test. Mice survival curves were computed by Kaplan-Meier analysis and compared using the log-rank test. All analyses were performed with R software (version 4.1.3; http://www.r-project.org/), and data were plotted with GraphPad Prism (version 9; San Diego, CA).

Results

TPMT/NUDT15 IM/IM patients tolerate lower dosages of MP than TPMT or NUDT15 single IM patients

A total of 1768 children with ALL (628 female patients = 35.5%; mean [SD] age at diagnosis = 6.7 [4.5] years) were included in this study (Table 1). TPMT and NUDT15 metabolizer status was determined on the basis of the genotype of functionally characterized TPMT and NUDT15 variants, that is, TPMT *2, *3A, *3B, and *3C and NUDT15 *2, *3, and *9. Twenty-two patients (22/1768, 1.2%) were found to be TPMT/NUDT15 IM/IM. The self-reported race/ethnicity breakdown in the entire study cohort was as follows: 20.8% Asian (n = 368), 9.9% Black (n = 174), 32.8% Hispanic (n = 580), 3.3% other (n = 59), and 33.2% White (n = 587). Among the TPMT/NUDT15 IM/IM patients (n = 22), the self-reported race/ethnicity breakdown was 22.7% Asian (n = 5), 68.2% Hispanic (n = 15), and 9.1% White (n = 2) (Table 1). The distribution of TPMT and NUDT15 metabolizer status across self-identified race/ethnicity categories was significant (χ²P < .001).

Table 1.

Clinical characteristics and tolerated MP dosage according to TPMT and NUDT15 metabolizer status

Variable	TPMT/NUDT15 NM (N = 1512)			TPMT IM (N = 131)			NUDT15 IM (N = 103)			TPMT/NUDT15 IM/IM (N = 22)
	N (%)	MP (IQR)a in mg/m2	P b	N (%)	MP (IQR)a in mg/m2	P b	N (%)	MP (IQR)a in mg/m2	P b	N (%)	MP (IQR)a in mg/m2	P b
Sex
Male	973 (64.4%)	60.1 (43.4-70.6)	.02	81 (61.8%)	50.4 (34.9-56.1)	.16	74 (71.8%)	43.7 (31.3-55.0)	.70	12 (54.5%)	30.5 (15.6-35.4)	.51
Female	539 (35.6%)	58.0 (45.6-69.9)		50 (38.2%)	42.5 (31.2-54.9)		29 (28.2%)	42.4 (37.0-56.0)		10 (45.5%)	24.2 (20.0-29.2)
Age at diagnosis (years)
<1	50 (3.3%)	59.8 (53.0-72.1)	.40	3 (2.3%)	55.0 (35.5-65.6)	.85	3 (2.9%)	47.0 (43.0-51.3)	.0004	0 (0%)	N/A	.44
1-10	1149 (76.0%)	59.7 (47.9-70.0)		96 (73.3%)	49.1 (33.7-55.6)		78 (75.7%)	46.2 (38.4-55.9)		15 (68.2%)	26.8 (21.0-33.1)
>10	313 (20.7%)	59.3 (46.1-71.6)		32 (24.4%)	43.3 (28.5-56.1)		22 (21.4%)	27.1 (18.9-37.7)		7 (31.8%)	23.7 (18.2-30.4)
Race/ethnicity
Asian	297 (19.6%)	55.0 (49.3-60.0)	<.0001	10 (7.6%)	53.5 (49.3-55.0)	.53	56 (54.4%)	44.0 (37.2-55.1)	.63	5 (22.7%)	31.2 (26.8-35.0)	.33
Black	157 (10.4%)	66.8 (54.9-73.3)		18 (13.7%)	39.7 (26.5-53.4)		0 (0%)	N/A		0 (0%)	N/A
Hispanic	473 (31.3%)	58.0 (38.3-71.6)		50 (38.2%)	48.1 (35.0-58.6)		42 (40.8%)	42.2 (28.9-54.5)		15 (68.2%)	24.7 (19.0-30.9)
Otherc	51 (3.4%)	61.1 (47.6-69.6)		6 (4.6%)	43.9 (36.3-53.8)		2 (1.9%)	53.2 (47.1-59.3)		0 (0%)	N/A
White	535 (35.3%)	63.6 (51.5-71.3)		47 (35.9%)	49.1 (28.7-55.8)		3 (2.9%)	41.6 (30.1-45.0)		2 (9.1%)	18.6 (16.1-21.2)

^aThe median of tolerated MP dosage and interquartile ranges (IQRs) were calculated for each category of patients.

^b P values were computed with a Kruskal-Wallis rank-sum test.

^cPatients with multiple self-identified races/ethnicities. IM = intermediate metabolizer; MP = mercaptopurine; N/A = not applicable; NM = normal metabolizer; NUDT15 = Nudix hydrolase 15; TPMT = thiopurine-S-methyl transferase.

The standard starting MP dosage for maintenance therapy ranged between 50 and 75 mg/m² and was adjusted clinically to a target white blood cell and/or absolute neutrophil count, sometimes with consideration of platelet count, TPMT metabolizer status alone (for patients from TOTAL16), or TPMT and NUDT15 metabolizer status (for patients from CHLA, if available) (Supplementary Table 1, available online). MP tolerated dose was the highest in the patients with normal activity of both TPMT and NUDT15 (median MP dosage of 59.7 mg/m², interquartile range [IQR] = 47.8-70.5 mg/m²), followed by patients IM for either TPMT (49.0 mg/m², 32.6-55.8 mg/m²; P < .001) or NUDT15 (43.5 mg/m², 31.5-55.0 mg/m²; P < .001) (Figure 1). TPMT/NUDT15 IM/IM patients were more sensitive to MP as they tolerated a median dosage of only 25.7 mg/m² (IQR = 19.0-31.1 mg/m²), which was significantly lower than TPMT/NUDT15 NM, TPMT IM, or NUDT15 IM patients (all Ps < .001). The difference in MP tolerated dose between IM/IM and TPMT IM or NUDT15 IM patients remained significant in a multivariate model that considered institution and race as co-factors (respective P < .001). There was no significant difference in tolerated dosages by race/ethnicity within the TPMT/NUDT15 IM/IM patients (P = .33, Table 1), although there was an over-representation of Hispanic patients having this compound phenotype.

Figure 1.

Association of TPMT and NUDT15 metabolizer status with MP toxicity. TPMT and NUDT15 metabolizer status of 1768 patients with acute lymphoblastic leukemia was determined on the basis of the genotype of functionally characterized TPMT and NUDT15 variants, that is, TPMT *2, *3B, and *3C and NUDT15 *2, *3, and *9. Data represent the median, minimum, and maximum values and interquartile ranges of individual MP dosages. ***P < .001, according to the Wilcoxon rank-sum test conducted between patients who were single intermediate metabolizer for TPMT or NUDT15 and patients compound intermediate metabolizers. MP = mercaptopurine; NUDT15 = Nudix hydrolase 15; TPMT = thiopurine-S-methyltransferase.

With regard to TPMT and NUDT15 diplotypes, the majority of IM/IM patients carried the TPMT *1/*3A and NUDT15 *1/*2 (or *3/*6) combination (N = 9, MP dosage = 24.7 mg/m², IQR = 23.1-36.5), and 7 patients with TPMT *1/*3A and NUDT15 *1/*3 tolerated a median dosage of 23.7 mg/m² (IQR = 19.0-30.4 mg/m²) (Supplementary Table 2, available online). One patient with TPMT *1/*3C allele combined to NUDT15 *1/*2 (or *3/*6) tolerated only 6.6 mg/m².

Because the protocols used by each institution stipulated a different starting dosage of MP for maintenance and provided different criteria for dose adjustment, we also used the DI to compare the response with MP in TPMT/NUDT15 IM/IM patients across institutions (Figure 2, A). Patients TPMT/NUDT15 IM/IM included in the AALL03N1 clinical trial received a median DI of 41.5% (IQR = 31.7%-48.6%) of a standard 75 mg/m² protocol dosage (Table 2). TPMT/NUDT15 IM/IM patients from UNOP in Guatemala tolerated 47.8% (IQR = 38.0%-61.2%) of a standard 50 mg/m² dosage. TPMT/NUDT15 IM/IM patients treated per Ma-Spore protocols tolerated 35.7% (IQR = 28.2%-49.0%) of a risk-stratified 50-75 mg/m² standard dosage. The 2 IM/IM patients from TOTAL16 received a DI of 11.5% and 18.4%, respectively. Similarly, the only patient from CHLA who was TPMT/NUDT15 IM/IM tolerated 40.9% of the protocol dose. Finally, the 2 patients IM/IM from PGIMER received 53.5% and 66.7% of a standard 60 mg/m² dosage. The meta-analysis comparing the DI in TPMT/NUDT15 IM/IM patients versus single IM patients showed a lower DI in compound IM patients (P = 4.5 × 10⁻⁶; effect size of -21.9%; 95% confidence interval = -31.1, -12.5), with no heterogeneity observed between the cohorts (P_{heterogeneity} = 0.84) (Figure 2, B), suggesting that the association was independent of institution-specific practices in MP dosage adjustment.

Figure 2.

Adjustment of MP dosage in TPMT/NUDT15 IM/IM patients across cohorts. A) The dose intensity (%) of MP was used to compare the dosage adjustment in compound intermediate metabolizer patients (TPMT/NUDT15 IM/IM) across cohorts. Data represent the median, minimum, and maximum values and interquartile ranges of individual dose intensity. B) A meta-analysis was conducted to evaluate the homogeneity of clinical practices in adapting the MP dosage in TPMT/NUDT15 IM/IM patients versus patients who were single IM (TPMT IM or NUDT15 IM). The estimates and standard errors of independent regression analyses from each cohort were pooled using a random effect model. *P < .05, according to the Wilcoxon rank-sum test conducted between single IM patients and TPMT/NUDT15 IM/M patients. CI = confidence interval; IM = intermediate metabolizer; MP = mercaptopurine; NM = normal metabolizer; NUDT15 = Nudix hydrolase 15; TPMT = thiopurine-S-methyltransferase; UNOP = Unidad Nacional de Oncología Pediátrica.

Table 2.

MP dose intensity per TPMT/NUDT15 metabolizer status across cohorts

Institution and protocol	TPMT/NUDT15 NM		TPMT IM		NUDT15 IM		TPMT/NUDT15 IM/IM
	N	DI (IQR)a %	N	DI (IQR)a %	N	DI (IQR)a %	N	DI (IQR)a %
COG AALL03N1 (N = 642)	572	87.9 (71.7-96.7)	38	70.2 (55.4-91.9)	28	63.2 (46.1-86.2)	4	41.5 (31.7-48.6)
CHLA ALL clinical protocols (N = 154)	133	88.8 (72.9-98.4)	10	96.5 (69.9-102.5)	10	71.0 (51.2-88.5)	1	40.9
Ma-Spore protocols (N = 80)	61	89.9 (66.2-105.4)	0	N/A	16	67.7 (45.0-80.4)	3	35.7 (28.2-49.0)
PGIMER ALL clinical protocol (N = 172)	138	85.9 (78.0-91.3)	8	82.2 (72.0-85.9)	24	69.5 (67.1-83.4)	2	60.1 (56.8-63.4)
SJCRH TOTAL16 (N = 490)	424	67.6 (52.2-79.7)	57	47.7 (29.9-59.3)	7	47.3 (32.8-59.1)	2	15.0 (13.3-16.7)
UNOP LLAG-0707 study (N = 230)	184	66.7 (48.2-95.8)	18	70.1 (49.9-87.0)	18	61.7 (27.2-98.0)	10	47.8 (38.0-61.2)

^aThe median dose intensity of mercaptopurine and interquartile ranges (IQRs) were calculated for each category of patients. CHLA = Children’s Hospital Los Angeles; COG = Children’s Oncology Group; DI = dose intensity; IM = intermediate metabolizer; MP = mercaptopurine; N/A = not applicable; NM = normal metabolizer; NUDT15 = Nudix hydrolase 15; PGIMER = Postgraduate Institute of Medical Education and Research; SJCRH = St Jude Children’s Research Hospital protocol; TPMT = thiopurine-S-methyl transferase; UNOP = Unidad Nacional de Oncología Pediátrica.

Preclinical modeling of genotype-guided dose adjustment for TPMT/NUDT15 IM/IM patients

To assess the cumulative effects of TPMT and NUDT15 decreased function on thiopurine tolerability and metabolism, wild-type and genetically modified mice were treated daily with 20 mg/kg of MP, equivalent to a standard dose of 60 mg/m² in humans (Figure 3, A). Tpmt^+/−/Nudt15^+/− mice experienced earlier toxic death with a median survival of 18 days, significantly shorter than Nudt15^+/− (23 days; P < .001), Tpmt^+/− (29 days; P < .001) and wild-type animals (35 days; P < .001; Figure 3, B). The monitoring of the hematopoietic parameters reflected a similar pattern of toxicity, with Tpmt^+/−/Nudt15^+/− mice showing the most profound depletion of neutrophils (0.2 × 1000 cells/µL, IQR = 0.1-0.3 × 1000 cells/µL) compared with wild-type (0.8 × 1000 cells/µL, 0.7-1.4 × 1000 cells/µL; P < .001), Tpmt^+/− (1.0 × 1000 cells/µL, 0.6-1.0 × 1000 cells/µl; P < .001) and Nudt15^+/− (0.4 × 1000 cells/µl, 0.3-0.5 × 1000 cells/µL; P = .002) (Figure 3, C). Because thiopurine hematopoietic toxicity is related to the incorporation of thioguanine triphosphate (TGTP) into the DNA as DNA-TG, we quantified the final metabolite in mouse bone marrow. After 5 days of MP treatment, Tpmt^+/−/Nudt15^+/− mice accumulated more DNA-TG (median of 481.0 fmol/μg DNA, IQR = 446.0-572.0 fmol/μg DNA) than wild-type (179.0 fmol/μg DNA, 164.0-230.0 fmol/μg DNA; P = .002), Tpmt^+/− (151.0 fmol/μg DNA, 99.4-202.0 fmol/μg DNA; P = .002), and Nudt15^+/− (376.0 fmol/μg DNA, 304.0-396.0 fmol/μg DNA; P = .18), confirming the increased conversion of MP into cytotoxic metabolites when TPMT and NUDT15 activities are both compromised (Figure 3, D).

Figure 3.

Mercaptopurine toxicity and metabolism in Tpmt/Nudt15 compound heterozygous mice. A) Tpmt/Nudt15 genotype-modified mice were treated with MP in drinking water. Hematopoietic toxicity was evaluated after blood collection after 5, 10, 17, and 30 days of treatment with MP administered in drinking water (N = 10) (left panel). Concentration of DNA-thioguanine (DNA-TG) in bone marrow was measured after 5 days of MP on a separated group of mice (N = 5) (right panel). B) Mice treated with 20 mg/kg of MP were followed up and were euthanized when body weight loss exceeded >20%, or they became moribund. C) Neutrophils were measured on day 10 of treatment with MP 20 mg/kg (median and interquartile range [IQR]). D) DNA-TG accumulation in bone marrow cells after 5 days of MP 20 mg/kg (median and IQR). Panel A was generated on Biorender.com. LC-MS/MS = liquid chromatography coupled to tandem mass spectrometry; MP = mercaptopurine; Nudt15 = Nudix hydrolase 15; Tpmt = thiopurine-S-methyltransferase. *P < .05, **P < .01, ***P < .001, according to the Wilcoxon rank-sum test conducted between single heterozygous mice (Tpmt^+/- or Nudt15^+/-) and compound heterozygous mice (Tpmt^+/-/Nudt15^+/-).

MP dose titration in TPMT/NUDT15 compound heterozygous animals attenuates toxicity

Given the impact of TPMT and NUDT15 on MP catabolism, we postulated that MP toxicity in IM/IM mice could be mitigated by further dose reduction compared with TPMT IM or NUDT15 IM. In a progressive reduction of MP dosages, from 50% of the full dose (half dose, 10 mg/kg) to 25% of the full dose (quarter dose, 5 mg/kg), we observed a reduction in MP toxicity in mice, which corresponded proportionally to the deficiency in only one or both genes (Figure 4). Reducing the dose to 50% of the full dose did not improve the survival or mitigate hematopoietic toxicity in Tpmt^+/−/Nudt15^+/− mice (Supplementary Figure 1, A, available online). On the other hand, the tolerance to MP was comparable in Tpmt^+/−/Nudt15^+/− mice treated with a quarter dose, single heterozygous Tpmt^+/− and Nudt15^+/− mice treated with a half dose, and wild-type animals receiving the full dose (median survival of 40.5, 52, 37.5, and 35 days, respectively; P = .23) (Figure 4, A and Supplementary Figure 1, B, available online). In addition, neutrophil counts in Tpmt^+/−/Nudt15^+/− (0.9 × 1000 cells/µL, IQR = 0.7-1.1 × 1000 cells/µL) reached levels equivalent to those in wild-type mice treated with the full dose of MP (0.8 × 1000 cells/µL, 0.7-1.4 × 1000 cells/µL; P = .90) or Tpmt^+/− mice treated with a half dose (1.2 × 1000 cells/µL, 0.9-1.5 × 1000 cells/µL; P = .12), although marginally higher than the levels observed in Nudt15^+/− animals treated with a half dose (0.7 × 1000 cells/µL, 0.6-0.8 × 1000 cells/µL; P = .08) (Figure 4, B, Supplementary Figure 1, C and D, available online).

Figure 4.

Pharmacogenetics-guided dose adjustment of mercaptopurine in mice. MP toxicity was evaluated in Tpmt^+/-/Nudt15^+/- after a 75% dose reduction (5 mg/kg) and was compared with one of the single heterozygous mice (Tpmt^+/- or Nudt15^+/-) treated with a half dose (10 mg/kg) and wild-type mice treated with a full dose (20 mg/kg). A) Mice overall survival was monitored for 70 days, and they were euthanized when body weight loss exceeded >20% or they became moribund. B) Neutrophils were measured on day 10 of treatment (median and interquartile range [IQR]). C) DNA-TG concentrations were measured after 5 days of treatment to evaluate MP metabolism and toxicity (median and IQR). MP = mercaptopurine; Nudt15 = Nudix hydrolase 15; Tpmt = thiopurine-S-methyltransferase. *P < .05, **P < .01, according to the Wilcoxon rank-sum test conducted between single heterozygous mice (Tpmt^+/- or Nudt15^+/-) and compound heterozygous mice (Tpmt^+/-/Nudt15^+/-).

MP dosage adjustment in mice led to similar concentrations of DNA-TG detected in the bone marrow of Tpmt^+/−, Nudt15^+/−, and wild-type mice (220.1 fmol/μg DNA, IQR = 190.0-236.7 fmol/μg DNA; 222.4 fmol/μg DNA, 220.8-333.5 fmol/μg DNA; and 179.4 fmol/μg DNA, 163.7-229.9 fmol/μg DNA, respectively) (Figure 4, C and Supplementary Figure 2, available online). However, the metabolite concentration was significantly reduced in Tpmt^+/−/Nudt15^+/− animals treated with a quarter dosage (121.6 fmol/μg DNA [110.6-132.0 fmol/μg DNA], P = .03, .02, and .009 compared with wild-type, Tpmt^+/−, and Nudt15^+/−, respectively).

Discussion

Pharmacogenetics-guided dose adjustment of MP has become a more commonly used practice in ALL therapy since the established genetic associations between TPMT and NUDT15 variants and thiopurine-induced myelosuppression (8-12,16,19,35-39). Current guidelines on thiopurine use strongly recommend reducing the starting dose for patients with actionable no-function variants in TPMT or NUDT15 (19). However, the extent of dosage decrease recommended for TPMT/NUDT15 IM/IM patients has not been clear, because only a limited number of studies have reported on MP tolerance in these patients (11,40). Our comprehensive study, encompassing diverse cohorts of patients with ALL, demonstrated that TPMT/NUDT15 IM/IM patients face a higher risk of thiopurine-induced toxicity compared with TPMT or NUDT15 single IM and would benefit from an additional dose reduction to mitigate toxicity.

Both TPMT and NUDT15 are involved in thiopurine inactivation: TPMT directly methylates MP and its first metabolite thioinosine monophosphate to produce methylmercaptopurine nucleotides (MMPNs), responsible for thiopurine-related liver toxicity (41). Three variants, namely TPMT *2, *3B, and *3C, account for more than 90% of the TPMT intermediate and poor metabolizer phenotypes characterized by higher thioguanine nucleotides but lower MMPNs compared with patients with a normal metabolizer phenotype (18,42,43). NUDT15, on the other hand, catabolizes the dephosphorylation of active metabolite TGTP into inactive thioguanine monophosphate, thereby reducing the cytotoxic effect of thiopurines but exposing patients to increased hematopoietic toxicity when carrying no-function variants such as NUDT15 *2 or NUDT15 *3 (9,44).

For the first time, we evaluated the clinical effects of concurrent alterations in both genes, finding that TPMT/NUDT15 IM/IM patients were significantly more sensitive to thiopurines than patients with IM for TPMT alone or NUDT15 alone. Importantly, these observations were consistent across diverse cohorts with different clinical practices, highlighting the substantial impact of the combined loss of expression of TPMT and NUDT15 on MP toxicity, irrespective of racial and ethnic backgrounds.

In line with previous preclinical models of MP dose adjustment in Tpmt-deficient or Nudt15-deficient mice (28,29), we systematically titrated MP dosage in Tpmt^+/−/Nudt15^+/− mice, in which a 75% dosage reduction effectively mitigated hematopoietic toxicity. We reason that this was achieved by normalizing leukocytes’ exposure to DNA-TG, the active drug metabolite directly responsible for thiopurine cytotoxicity. However, the DNA-TG level was significantly lower in Tpmt^+/−/Nudt15^+/− mice after dosage reduction compared with wild-type mice receiving the full dosage, even though the degree of neutropenia was comparable across genotypes. This raises questions regarding the precise relationship between DNA-TG and clinical myelosuppression and how (and whether) to use DNA-TG as an endpoint for therapeutic drug monitoring during thiopurine therapy.

Nonetheless, these results underscore the importance of pharmacogenetic-guided thiopurine dose adjustment in preventing toxicity without compromising efficacy for the TPMT/NUDT15 IM/IM patients (36,45). Although this study focused on the additive effects of TPMT and NUDT15 on MP tolerance/toxicity, it is of critical importance to understand the impact of the genotype guided dose reduction on anti-leukemic efficacy, which should be carefully studied in future studies and clinical trials. Finally, although genetic variation in TPMT and NUDT15 strongly influence the risk of thiopurine-related myelosuppression, MP tolerance also depends on other factors, including leukemia sensitivity, adherence, co-medication administration, comorbidities, and other genetic determinants (22,24,46-50).

We provide evidence that patients with TPMT/NUDT15 IM/IM phenotype are at increased risk of thiopurine-induced toxicity compared with patients with a TPMT IM or NUDT15 IM phenotype alone, and additional dose reduction should be considered when initiating thiopurine in this population to mitigate toxicity.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Author contributions

Maud Maillard, PharmD, PhD (Data curation; Formal analysis; Investigation; Methodology; Writing—original draft; Writing—review & editing), Deepa Bhojwani, MD (Conceptualization; Resources; Writing—review & editing), Federico G. Antillon Klussmann, MD (Resources; Writing—review & editing), Prateek Bhatia, MD (Resources; Writing—review & editing), Amita Trehan, MD (Resources; Writing—review & editing), Seth E. Karol, MD (Resources; Writing—review & editing), Ching-Hon Pui, MD (Resources; Writing—review & editing), Hiroto Inaba, MD (Resources; Writing—review & editing), Sima Jeha, MD (Resources; Writing—review & editing), Allen E.J. Yeoh, MD (Resources; Writing—review & editing), Cyrine E. Haidar, PharmD (Conceptualization; Resources; Writing—review & editing), Joshua L.Y. Suang, Mr (Resources; Writing—review & editing), Mark Leggas, PhD (Resources; Writing—review & editing), Kristine R. Crews, PharmD (Resources; Writing—review & editing), Mary V. Relling, PharmD (Resources; Writing—review & editing), Jenny Q. Nguyen, PharmD (Resources; Writing—review & editing), Shawn H.R. Lee, MD, PhD (Data curation; Writing—review & editing), Divyabharathi Chepyala, PhD (Investigation; Writing—review & editing), Keito Hoshitsuki, PharmD (Writing—original draft; Writing—review & editing), Wenjian Yang, PhD (Data curation; Formal analysis; Methodology; Writing—original draft; Writing—review & editing), Rina Nishii, PhD (Investigation; Methodology; Writing—review & editing), Meenu Singh, MD, PhD (Resources; Writing—review & editing), Jun J. Yang, PhD (Conceptualization; Funding acquisition; Supervision; Writing—original draft; Writing—review & editing).

Funding

This research was supported by the National Institutes of Health (NIH, grant #R35GM141947).

Conflicts of interest

MVR and St Jude Children’s Research Hospital received an investigator-initiated research funding from Servier.