Abstract
Objective
To characterize the allelic and diplotype variability of TPMT and NUDT15 in pediatric patients with B-cell acute lymphoblastic leukemia from the central-southern region of Mexico.
Methods
Samples from 275 pediatric B-cell acute lymphoblastic leukemia patients were analyzed. Next-generation sequencing was used for TPMT and NUDT15 genotyping. Alleles and diplotypes were assessed according to the Clinical Pharmacogenetics Implementation Consortium guidelines. Their geographic distribution was compared across Mexican states and global populations. In-silico analyses were conducted to assess the structural and functional impact of TPMT variants not associated with star alleles.
Results
The wild-type *1 allele, associated with normal enzymatic activity, was predominant in both genes: TPMT (94.15%) and NUDT15 (90.45%). TPMT showed greater allelic diversity compared with previous studies in Mexican populations. Alleles conferring absent or indeterminate enzymatic activity in TPMT were distributed across six diplotypes (11.62%), with *3A allele (4.73%) and *1/*3A diplotype (9.45%) being the most frequent. Additionally, two unclassified TPMT variants, p.G126A and p.D137Y, were identified. For NUDT15, three non-wild-type diplotypes were observed (19.09%), with the *2 allele (6.74%) and *1*/2 diplotype (13.48%) being the most prevalent.
Conclusion
Approximately 28% of patients carried TPMT and/or NUDT15 variants associated with non-wild-type enzymatic activity, increasing the risk of mercaptopurine-induced myelotoxicity. Preemptive genotyping is essential to reduce toxicity, optimize treatment, and advance precision medicine in this population. Additionally, the two TPMT variants p.G126A and p.D137Y, currently not classified within Clinical Pharmacogenetics Implementation Consortium-defined star alleles, highlight the need for functional validation and potential clinical classification to improve pharmacogenetic interpretation in diverse populations.
Keywords: NUDT15 diplotypes, pharmacogenetics, TPMT variants, Mexican Mestizos
Introduction
B-cell acute lymphoblastic leukemia (B-ALL) is the most common pediatric cancer and the leading cause of cancer-related death among Mexican children [1,2]. Although advances in therapy have improved survival, treatment-related mortality remains a concern, often because of prolonged neutropenia and lymphopenia, caused by cytotoxic and immunosuppressive drugs [3]. Standard treatment comprises induction, consolidation, and maintenance phases. During the maintenance phase, 6-mercaptopurine (6-MP) is administered for 2 or 3 years after diagnosis [4]. Its antileukemic effect depends on metabolic conversion to thioguanine nucleotides (TGN), which are incorporated into DNA, inducing damage and apoptosis [5,6].
Although 6-MP has significantly improved survival in B-ALL, 15–28% of patients experience adverse reactions at standard doses. These include severe myelosuppression, primarily prolonged neutropenia, and, to a lesser extent, lymphopenia, as well as hepatotoxicity. Such toxicities often require dose reduction, treatment interruption, or discontinuation of chemotherapy, potentially compromising therapeutic efficacy [7,8].
These adverse reactions are attributable to interindividual variability in thiopurine metabolism, driven primarily by genetic variants in thiopurine S-methyltransferase (TPMT), a phase II metabolizing enzyme, and Nudix hydrolase 15 (NUDT15), which hydrolyzes cytotoxic thioguanine metabolites, preventing their incorporation into DNA. Over 40 TPMT alleles have been identified, with *3A and *3C accounting for most low or null enzyme activity phenotypes across populations [8]. The prevalence of TPMT deficiency varies by ethnicity, representing the principal cause of thiopurine intolerance in European and African populations [5,6].
TPMT activity is inherited as a codominant trait. Patients homozygous or compound heterozygous for nonfunctional TPMT alleles are at a high risk of severe, potentially life-threatening myelosuppression when treated with standard thiopurine doses. Heterozygous individuals exhibit elevated TGN levels and a greater risk of myelosuppression compared with those carrying two wild-type alleles [8].
Demographic and genome-wide association studies have identified NUDT15 as a key determinant of 6-MP tolerance. Functional studies have shown that low or null NUDT15 alleles increase the incorporation of TGN into DNA, enhancing cytotoxicity [5,9,10]. Patients carrying one loss-of-function NUDT15 allele (*2, *3, or *9) are classified as intermediate metabolizers. In contrast, individuals with two loss-of-function alleles are poor metabolizers and tolerate less than 10% of the standard 6-MP dose [5,6]. NUDT15 deficiency is common in Asians and Hispanics, infrequent in Europeans, and rare in Africans [5,9–11].
The latest Clinical Pharmacogenetics Implementation Consortium (CPIC) update provides thiopurine dosing recommendations based on TPMT and NUDT15 genotypes to minimize thiopurine-related myelotoxicity [6]. Additionally, the National Comprehensive Cancer Network guidelines for B-ALL recommend genotyping for nonfunctional TPMT and NUDT15 alleles before initiating thiopurine therapy or in patients with unexpected high toxicity [12]. However, limited access to pharmacogenetic testing across many public healthcare settings in Mexico hinders the implementation of genotype-guided 6-MP therapy.
This study characterizes allele and diplotype frequencies of TPMT and NUDT15 in 275 pediatric B-ALL patients from the south-central region of Mexico, providing population-specific evidence to support the implementation of personalized thiopurine dosing strategies in the healthcare system of Mexico.
Materials and methods
Population
This retrospective analysis included 275 pediatric patients with de-novo B-ALL born in nine different states of Mexico. All patients were diagnosed between January 2018 and December 2024 by pediatric onco-hematologists or oncologists at participating institutions, based on bone marrow aspiration and immunophenotyping. The mean age at diagnosis was 7.9 ± 4.8 years (range: 0–17 years). Before treatment initiation, 3–5 ml of bone marrow aspirate was collected into EDTA tubes at each institution and delivered to the Genomic Diagnostics Laboratory at the National Institute of Genomic Medicine (INMEGEN). The study was approved by the scientific review boards of Comisión Nacional de Investigación Científica del Instituto Mexicano del Seguro Social (number: R-2020-785-022). Parents provided informed consent in accordance with the Declaration of Helsinki and authorized the use of samples and clinical data.
Nucleic acid extraction and quantification
DNA was extracted using the Maxwell 16 RSC Instrument (Promega Corporation, Madison, Wisconsin, USA). DNA concentration and quality were assessed using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, Massachusetts, USA), and a Qubit 4 fluorometer with the Qubit 1X double-stranded DNA High Sensitivity Assay Kit (Thermo Fisher Scientific, Waltham, Massachusetts, USA).
TPMT and NUDT15 sequencing
TPMT and NUDT15 were sequenced using a customized next-generation sequencing panel synthesized by ArcherDX (ArcherDX, Inc., Boulder, Colorado, USA), targeting all exons and exon-intron boundaries. Library preparation was carried out according to the manufacturer’s protocol (see Supplementary material, Supplemental digital content 1, https://links.lww.com/FPC/B534 for details). Sequencing was performed on the NextSeq 2000 instrument (Illumina, Inc., San Diego, California, USA). Alignment to the GRCh37.p13/hg19 and variant calling were conducted using Archer Suite Analysis v7.3.2. The bioinformatic algorithm detects single-nucleotide variants, small insertions and deletions, and copy number variants. For 99% of target regions, the read depth exceeded 1000×. For heterozygous variants, the allelic variant fraction ranged from 40 to 55% and exceeded 98% for homozygous variants.
Allele assignment and metabolic phenotype
Allele assignment, diplotype identification, and metabolic classification for both genes were performed according to the CPIC guidelines [6]. Additional details on *1/*2 and *3/*6 NUDT15 diplotype discrimination are included in Supplementary material, Supplemental digital content 1, https://links.lww.com/FPC/B534. Diplotype distribution in B-ALL patients was compared with that reported for other populations, including only studies with available diplotype data. Chi-square or Fisher’s exact tests were applied to evaluate differences in diplotype distribution, depending on expected cell counts. Population details and references are provided in Supplementary Tables S1 and S2, Supplemental digital content 1, https://links.lww.com/FPC/B534. Additionally, diplotype distribution among B-ALL was compared by state of birth in Mexico.
In-silico evaluation of variants in TPMT not assigned to a star allele
An in-silico evaluation was conducted to assess the potential structural and functional impact of TPMT variants not assigned to a star allele. For structural analysis, the crystal structure of TPMT (PDB ID: 2BZG) was used. Variant effects on protein stability and flexibility were evaluated using DynaMut [13]. Structural representations were generated in PyMOL to visualize modifications in hydrogen bonding, hydrophobic interactions, and overall structural integrity. A multiple sequence alignment was performed using Clustal Omega to assess evolutionary conservation. In-silico predictions were further evaluated using the Franklin by Genoox platform (https://franklin.genoox.com), which integrates multiple computational tools like MutationTaster, DANN, BayesDel, REVEL, PolyPhen, and SpliceAI for variant interpretation.
Results
TPMT alleles and diplotypes
All samples were successfully genotyped for TPMT. The identified alleles are summarized in Supplementary Table S3, Supplemental digital content 1, https://links.lww.com/FPC/B534. In two patients, exon 5 variants not currently assigned to any CPIC-defined star alleles were detected: NM_000367.5(TPMT):c.377G>C, p.(G126A) (rs773335632), and NM_000367.5(TPMT):c.409G>T, p.(D137Y) (rs376217758).
Supplementary Table S4, Supplemental digital content 1, https://links.lww.com/FPC/B534, shows the distribution of TPMT diplotypes and their corresponding predicted metabolizer phenotypes.
Figure 1a illustrates the distribution of *1/*3A, *1/*3C, and *1/*2 TPMT diplotypes across global populations and B-ALL patients. Figure 1b depicts the TPMT diplotype distribution among B-ALL patients stratified by state of birth.
Fig. 1.
TPMT diplotype distribution. Panel (a): The *1/*3A, *1/*3C, and *1/*2 diplotype distribution among populations from different countries. Data obtained from studies listed in Supplementary Table S1; numbers following country names correspond to the different publications. Panel (b): TPMT diplotype distribution among pediatric B-ALL patients, stratified by state of birth in Mexico. 1/X refers to diplotypes containing p.(G126A) or p.(D137Y) variants. NUDT15 diplotype distribution among populations. Panel (c): Diplotype distribution of *1/*2, *1/*3, and*1/*4 diplotypes among populations from different countries. Data were retrieved from Supplementary Table S2. Panel (d): NUDT15 diplotype distribution by state of birth in B-ALL patients. The numbers following the country name indicate data from different publications. African A, African American; B-ALL, B-cell acute lymphoblastic leukemia; LATAM: Latin American; N. Am, North American; NB, newborns; SLE, systemic lupus erythematosus; RA, rheumatoid arthritis.
In-silico evaluation of p.(G126A) and p.(D137Y) TPMT variants
The modeled structures carrying the p.(G126A) and p.(D137Y) variants are shown in Figs. 2a and b, respectively, with mutated residues highlighted in red. The G126 residue interacts closely with I112, S107, and S108 via hydrogen bonds and van der Waals forces. DynaMut analysis predicted a mildly destabilizing effect on protein stability (ΔΔG: –0.713 kcal/mol) for p.(G126A), which was attributed to increased molecular flexibility, as indicated by a positive vibrational entropy change (ΔΔSVib ENCoM: 0.148 kcal·mol−1·K−1).
Fig. 2.
Structural representation of the TPMT variants. Panel (a): p.(G126A) variant and Panel (b): p.(D137Y) variant. Panel (c): MSA of the TPMT protein across species. Color key: in both structural models, the WT residues are shown in red on the orange protein surface. The yellow molecule represents the ligand bound to TPMT in the crystal structure (PDB ID: 2BZG). Light green highlights the positions of residues 126 or 137 in both WT and mutant structures. Yellow dots indicate ionic interactions, red dashed lines represent hydrogen bonds, green dots denote hydrophobic contacts, and gray dots indicate weak interactions. MSA, multiple sequence alignment; WT, wild type.
The p.(D137) residue (Fig. 2b) interacts with I119, S118, C117, P123, R124, and I99 through hydrogen bonds, van der Waals forces, hydrogen bond-like interactions, and hydrophobic contacts. The bulkier tyrosine residue at position 137 alters interactions with C117 and S118 by eliminating ionic interactions and creating new hydrophobic contacts with I99 and R124. DynaMut predicted a stabilizing effect (ΔΔG: 1.070 kcal/mol) and reduced molecular flexibility, associated with a decrease in vibrational entropy energy (ΔΔSVib ENCoM: –0.447 kcal·mol−1·K−1).
Protein sequence alignment of TPMT showed that residues G126 and D137 are conserved (Fig. 2c). In-silico algorithms indicated a potential functional impact for p.(G126A) (Combined annotation dependent depletion score: 24.0; Deleterious annotation of genetic variants using neural networks (DANN): deleterious = 0.99; MutationTaster: deleterious = 1), while other algorithms predicted moderate-to-low pathogenicity (Rare exome variant ensemble learner (REVEL) = 0.506; PolyPhen = 0.481; SpliceAI = 0.0700; Pangolin = 0.0100). The c.377G nucleotide is evolutionarily conserved (phyloP = 5.04; PhastCons = 1), and the MutationTaster algorithm predicted a possible splicing alteration because of increased donor activity at position gDNA15366 (wt: 0.64/mu: 0.82).
For p.(D137Y), the potential functional impact was supported by five algorithms (Combined annotation dependent depletion score: 29.6; DANN: deleterious = 1; BayesDel: deleterious, supporting = 0.17; varity: deleterious = 0.89; and MutationTaster: deleterious = 1). In contrast, moderate levels of pathogenicity were predicted by other algorithms (REVEL = 0.520; PolyPhen = 0.587; SpliceAI = 0.100). The c.409G nucleotide is also evolutionarily conserved (phyloP = 8.89; PhastCons = 1), and MutationTaster predicted splicing alteration with increased acceptor activity at gDNA15404 (wt: 0.41/mu: 0.91) and creation of new acceptor sites at positions gDNA15402, 15405, and 15406.
According to the American College of Medical Genetics and Genomics guidelines, both variants were classified as of uncertain significance [14]. The p.(G126A) variant is already listed as variants of uncertain significance in the ClinVar database (ID: 3460503). Since no functional studies are available, the metabolic consequences of both variants remain undetermined.
NUDT15 alleles and diplotypes
Eight samples were excluded because of insufficient DNA, leaving 267 samples for analysis. The identified alleles are presented in Supplementary Table S5, Supplemental digital content 1, https://links.lww.com/FPC/B534. Diplotype distribution and the corresponding metabolic types are shown in Supplementary Table S6, Supplemental digital content 1, https://links.lww.com/FPC/B534.
Comparison of the NUDT15 *1/*3, *1/*2, and *1/*4 diplotype distributions across populations is shown in Fig. 1c, which displays the frequencies in the selected populations. Other diplotypes present in these populations are not shown in Fig. 1c, but are available in Supplementary Table S2, Supplemental digital content 1, https://links.lww.com/FPC/B534. Figure 1d presents the NUDT15 diplotype distribution among B-ALL patients by state of birth in Mexico.
Individual TPMT and NUDT15 diplotype combinations
A total of 267 individuals were genotyped for both genes. TPMT and NUDT15 diplotype combinations and their potential consequences on enzymatic activity are summarized in Table 1. Among the 267 patients, 71.91% (192/267) were normal metabolizers for both genes, 25.09% (67/267) were heterozygous for a non-wild-type allele in either NUDT15 or TPMT, and 3% (8/267) carried a non-wild-type allele in both genes.
Table 1.
Combinations of TPMT and NUDT15 diplotypes in the B-cell acute lymphoblastic leukemia individuals
| Enzymatic activity of TPMT/NUDT15 | TPMT | NUDT15 | Count | Total count, n = 267 (%) |
|---|---|---|---|---|
| WT/WT | *1*/1 | *1*/1 | 192 | 192 (71.91) |
| Non-WT/WT | *1/*3A | *1*/1 | 20 | 24 (8.99) |
| *1/*3C | *1*/1 | 2 | ||
| *1/*12 | *1*/1 | 1 | ||
| *1/G126A | *1*/1 | 1 | ||
| WT/non-WT | *1*/1 | *1/*2 | 29 | 43 (16.10) |
| *1*/1 | *1/*3 | 3 | ||
| *1*/1 | *1/*4 | 11 | ||
| Non-WT/non-WT | *1/*3A | *1/*2 | 6 | 8 (3.00) |
| *1/D137Y | *1/*4 | 1 | ||
| *1/*2 | *1/*2 | 1 |
WT, wild type.
Discussion
This study assesses TPMT and NUDT15 allelic and diplotype variability in B-ALL patients primarily from the central-southern region of Mexico. A customized next-generation sequencing based panel targeting coding regions and exon-intron boundaries enabled the detection of both established star alleles and unclassified variants. For both genes, the *1 allele, associated with normal enzymatic activity, was the most frequent, occurring at a frequency of 94.15% for TPMT and 90.45% for NUDT15.
Sequencing revealed greater allelic diversity in TPMT than previously reported [15–17]. Six non-wild-type alleles were identified, including the less common *2 and *12 alleles, as well as the p.G126A and p.D137Y variants. The *3A allele was the most frequent (4.76%), while the combined frequency of the remaining non-wild-type alleles was only 1%.
Six TPMT diplotypes associated with intermediate or indeterminate enzymatic activity were observed in 11.62% of our B-ALL patients, a frequency comparable to those reported in Mexican B-ALL patients (8.8%) and healthy controls (9.5%) [15]. No significant differences were found in the distribution of non-wild-type TPMT alleles or diplotypes between B-ALL patients and healthy individuals, supporting the lack of association with B-ALL susceptibility [6].
Among B-ALL patients, three out of six non-wild-type TPMT diplotypes (10.53%) were classified as intermediate metabolizers. The *1/*3A diplotype was the most frequent (9.45%), consistent with frequencies reported in Mexican patients with rheumatoid arthritis and systemic lupus erythematosus (combined frequency 8.7%) [16], and in Mexican newborns (10.56%) [17,18]. This aligns with data from South American, North American, and European populations, where *1/*3A frequencies range from 4.5 to 13%. No homozygous or compound heterozygous individuals for non-wild-type TPMT alleles were identified in this cohort.
The *1/*3A diplotype was also the most frequent across all birth states, with no statistically significant variation in distribution (8.3–18.2%). A study of individuals from Mexico City reported frequencies of *1/*3A (6%) and *1/*3C (1.6%), consistent with those observed in the Mexico City B-ALL patient subgroup (7.1 and 2.3%, respectively) [15]. The State of Mexico and Mexico City showed greater TPMT diplotype diversity, with four and three distinct diplotypes identified, respectively. This may reflect greater internal migration, especially into Mexico City, the nation’s capital.
As shown in Fig. 1a, the combined frequency of *1/*3A, *1/*3C, and *1/*2 TPMT diplotypes varies geographically, ranging from 1.4 to 14.8%. Individual diplotypes, particularly *1/*3C and *1/*3A, also exhibit regional differences. The *1/*3A diplotype predominates in Latin American and European populations, while the *1/*3C diplotype is most common in African and Asian populations.
The TPMT *1/*12 diplotype has been reported at very low frequencies in African Americans (0.0104%) and Hispanics (0.0165%). Its metabolic implications remain unclear and warrant further investigation. Although the *1/*3B diplotype was not observed in our cohort, it has been reported at low frequencies in the Mexican population [15].
The TPMT variants p.(G126A) and p.(D137Y) are not currently assigned to any CPIC-defined TPMT star allele. Both variants have been detected in healthy Mexican individuals. The p.(G126A) variant may be of Native American origin, as it was observed exclusively in Indigenous Mexican individuals (allele count: 29/176 801; Allele frequency (AF) = 0.000164) [19] and at a lower frequency in admixed American populations (AF = 0.00005038, gnomAD database).
In contrast, the p.(D137Y) variant shows broader distribution. It has been detected in Mexican individuals of European ancestry (allele count: 17/87 311; AF = 0.000195) [19]. According to the gnomAD database, the highest frequency was observed in South Asian populations (AF = 0.0002746), followed by European (non-Finnish) (AF = 0.00004663) and admixed American (AF = 0.00003339) populations.
In-silico analyses suggest that both p.(G126A) and p.(D137Y) may disrupt splicing and exert a deleterious effect on the encoded TPMT enzyme. Both variants lie within a genomic constraint region (Z-score = 4.15; observed/expected ratio = 0.69), indicating intolerance to variation. DynaMut analysis predicted disruption of hydrogen bonds and ionic interactions, potentially affecting protein flexibility. However, these changes may not directly impair catalytic activity or ligand-binding capacity.
Matreyek et al. [20] assessed TPMT variants using the variant abundance by massive parallel sequencing assay; however, p.(G126A) and p.(D137Y) were excluded from the final analysis because of insufficient coverage. Despite this, their findings suggest that residue D137 is highly intolerant to variation, indicating that substitutions at this site may impair protein stability or expression. Position G126 was not assessed, precluding evaluation of its functional impact.
Clinical data on myelotoxicity are limited for patients carrying these variants, hindering genotype–phenotype correlation. The patient with the p.(G126A) variant was diagnosed with high-risk pre-B-ALL with (1;19)(TCF3::PBX1) and was treated with the St. Jude Total XV protocol using standard dosing [21]. The patient experienced treatment resistance with early central nervous system relapse and later died from gastric hemorrhage; however, no clear evidence of myelotoxicity was documented.
The patient with the p.(D137Y) variant was diagnosed with high-risk pre-B-ALL and received treatment under the same protocol. Gene fusion analysis was not performed. During the maintenance phase, signs of severe myelotoxicity were present, leading to the 6-MP dose reduction. This patient also carried the NUDT15 *1/*4 diplotype, classified as undetermined, which may have contributed to the myelotoxicity. Experimental studies are needed to clarify the functional impact of these TPMT variants. In the interim, close clinical monitoring of carriers is essential to mitigate the risk of thiopurine-related complications.
In the B-ALL cohort, three non-wild-type NUDT15 alleles were identified. The *2 allele was the most frequent (6.74%), followed by *4 (2.25%). The normal metabolizer diplotype (*1/*1) was present in 216 individuals (80.91%). Intermediate metabolizer diplotypes were found in 39 individuals (14.60%), with *1/*2 being the most common (13.48%). The *1/*4 diplotype, classified as indeterminate by CPIC, was identified in 12 individuals (4.49%) [6].
When stratified by state of birth, *1/*2 was the most frequent non-wild-type NUDT15 diplotype across all states, followed by *1/*4. Although the frequency of *1/*2 varied substantially (7.9–36.4%), these differences were not statistically significant, likely because of small sample sizes in each state.
The distribution of NUDT15 diplotypes has been less extensively studied than that of TPMT. To the best of our knowledge, reports on NUDT15 variants in Mexican populations have focused on individual alleles rather than diplotypes. For example, Texis et al. [22] reported frequencies of 6.5% in Mestizos, 4.7% in Mexicans from Los Angeles, California, and 10% in Native Mexicans for the (c.415C>T, p.R139C) variant, which is present in *2 and *3 alleles, without distinguishing between them. Additionally, *1/*4 diplotype prevalence may be underestimated in studies using nonsequencing genotyping methods that fail to detect the p.R139H variant. This is clinically relevant, as carriers of the *4 allele require thiopurine dose reductions because of reduced enzymatic activity [23].
Despite this limitation, substantial geographical variation in NUDT15 diplotype distribution is evident from the available data (Fig. 1c and Supplementary Table S2, Supplemental digital content 1, https://links.lww.com/FPC/B534). The Asian population showed the highest frequencies of non-wild-type diplotypes, which are recognized as the primary cause of thiopurine-induced myelotoxicity [10]. Among Latin American populations, non-wild-type diplotype frequencies ranged from 5.9 to 18.3%, with the highest frequency observed in our cohort.
The *1/*2 diplotype was the most prevalent non-wild-type diplotype in our B-ALL patients (13.5%), significantly exceeding the frequencies reported in Chinese Kirghiz and Dai populations (P < 0.000001). This elevated frequency may have resulted from selective pressure or genetic drift within the Mexican population.
The *1/*4 diplotype, present in Latino populations but absent from European and most Asian populations, suggests a likely Native American origin. Supporting this, allele *4 was found exclusively in Indigenous Mexicans (3.6%) [19]. In contrast, the *1/*3 diplotype was significantly less frequent in our population than in Asian populations (P < 1 × 10−10).
In our cohort, 71.91% of patients were predicted to have normal thiopurine metabolism based on both gene genotypes, while 25.09% are expected to have reduced enzymatic activity in at least one gene, with the intermediate metabolizer phenotype in NUDT15 being the most common. For intermediate metabolizers in either gene, a 30–80% reduction in the 6-MP starting dose is recommended when the standard dose is greater than or equal to 75 mg/m2/day. Subsequent doses should be titrated based on the severity of myelosuppression and disease-specific treatment protocols.
Murine studies have demonstrated that NUDT15± mice accumulate higher levels of TGN in bone marrow cells and exhibit lower neutrophil counts than TPMT± mice when both are treated with the same dose of 6-MP [3].
In our cohort, 3.0% of B-ALL patients were classified as intermediate metabolizers for both TPMT and NUDT15. As these enzymes function via independent pathways, concurrent reduced activity results in additive increases in thiopurine toxicity. Such individuals face a higher risk of adverse effects and may require substantial initial dose reductions. Across diverse ethnic groups and clinical settings, patients with intermediate activity in both TPMT and NUDT15 tolerated a median mercaptopurine dose of 25.7 mg/m2 (IQR: 19.0–31.1 mg/m2), significantly lower than that tolerated by patients with intermediate metabolism in only one gene [3].
Several limitations must be considered when interpreting the results of this study. The sequencing panel included exons and exon–intron boundaries, excluding variants in deep intronic or regulatory regions that could affect enzymatic activity. The statistical approach used to distinguish the *1/*2 or *3/*6 diplotypes may have produced a slight overestimation of the *1/*2 diplotype frequency. The geographical restriction of individuals to central-southern states of Mexico may limit the generalizability of the findings to the broader Mexican pediatric B-ALL population. A limited sample size per state also reduces statistical power to detect regional differences in diplotype distribution. Furthermore, because of insufficient clinical data, associations between diplotypes and thiopurine-induced myelotoxicity could not be evaluated. TPMT enzyme activity was not biochemically assessed in any patient.
Conclusion
In summary, this study evaluates for the first time NUDT15 and TPMT allelic and diplotype distribution in a B-ALL group of Mexican patients using massive parallel sequencing. Approximately 28% of B-ALL patients are predicted to have non-wild-type enzymatic activity in TPMT, NUDT15, or both, placing them at increased risk for 6-MP-induced myelotoxicity. Preemptive genotyping of TPMT and NUDT15 in Mexican patients with B-ALL is urgently needed to reduce treatment-related toxicity, optimize outcomes, and advance precision medicine approaches.
Acknowledgements
This work was supported by Consejo Nacional de Humanidades, Ciencias y Tecnologías (CONAHCYT) through the following grants: Pronaii Leucemia Infantil, Proyecto Nacional de Investigación e Incidencia (FORDECYT-PRONACES/303082/2019, FORDECYT-PRONACES/303019/2019, and FORDECYT-PRONACES/302994/2019). JGS received a CONAHCYT fellowship (CVU 1042482).
Conceptualization: JGS, CAV, CMG. Methodology: JGS, MJO, BEVT. Investigation: JGS, CAV, CMG, VJSM. Resources: JJY, ICP, DCA, JCNE, JFL, MMR, JPG, VHOV, MRBL, AGS, CAGD, AOV, MSC, MAGH, NCLS, LSCC, RPC, JMMA, FFE, KCS, LLFL, ECMC, VGO. Validation: all the authors. Visualization: JGS: Formal analysis: JGS, CAV. Writing original draft: JGS, CAV. Writing review & editing: All authors. Supervision: All authors. Project administration: CAV. Funding acquisition: CAV, JCNE, JMMA, RPC.
All relevant data supporting the findings of this study are included in this article and its supplementary materials.
Conflicts of interest
The authors declare no conflicts of interest.
Supplementary Material
Footnotes
Joaquin Garcia-Solorio and Carolina Molina-Garay contributed equally to the writing of this article.
Supplemental Digital Content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's website, www.pharmacogeneticsandgenomics.com.
References
- 1.Rivera-Luna R, Perez-Vera P, Galvan-Diaz C, Velasco-Hidalgo L, Olaya-Vargas A, Cardenas-Cardos R, et al. Triple-hit explanation for the worse prognosis of pediatric acute lymphoblastic leukemia among Mexican and Hispanic children. Front Oncol 2022; 12:1072811. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.de Smith AJ, Jiménez-Morales S, Mejía-Aranguré JM. The genetic risk of acute lymphoblastic leukemia and its implications for children of Latin American origin. Front Oncol 2023; 13:1299355. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Maillard M, Nishii R, Yang W, Hoshitsuki K, Chepyala D, Lee SHR, et al. Additive effects of TPMT and NUDT15 on thiopurine toxicity in children with acute lymphoblastic leukemia across multiethnic populations. J Natl Cancer Inst 2024; 116:702–710. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Terwilliger T, Abdul-Hay M. Acute lymphoblastic leukemia: a comprehensive review and 2017 update. Blood Cancer J 2017; 7:e577. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Karol SE, Yang JJ. Pharmacogenomics and ALL treatment: how to optimize therapy. Semin Hematol 2020; 57:130–136. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Relling MV, Schwab M, Whirl-Carrillo M, Suarez-Kurtz G, Pui C-H, Stein CM, et al. Clinical Pharmacogenetics Implementation Consortium Guideline for Thiopurine Dosing Based on TPMT and NUDT15 Genotypes: 2018 Update. Clin Pharmacol Ther 2019; 105:1095–1105. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Jantararoungtong T, Wiwattanakul S, Tiyasirichokchai R, Prommas S, Sukprasong R, Koomdee N, et al. TPMT*3C as a Predictor of 6-Mercaptopurine-Induced Myelotoxicity in Thai Children with Acute Lymphoblastic Leukemia. J Pers Med 2021; 11:783. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Goh LL, Lim CW, Leong KP, Ong KH. TPMT and NUDT15 testing for thiopurine therapy: a major tertiary hospital experience and lessons learned. Front Pharmacol 2022; 13:837164. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Correa-Jimenez O, Yunis JJ, Linares-Ballesteros A, Sarmiento-Urbina I. Susceptibility to thiopurine toxicity by TPMT and NUDT15 variants in Colombian children with acute lymphoblastic leukemia. Colomb Med (Cali) 2021; 52:e2074569. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Moriyama T, Nishii R, Perez-Andreu V, Yang W, Klussmann FA, Zhao X, et al. NUDT15 polymorphisms alter thiopurine metabolism and hematopoietic toxicity. Nat Genet 2016; 48:367–373. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Yang S-K, Hong M, Baek J, Choi H, Zhao W, Jung Y, et al. A common missense variant in NUDT15 confers susceptibility to thiopurine-induced leukopenia. Nat Genet 2014; 46:1017–1020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Shah B, Mattison RJ, Abboud R, Abdelmessieh P, Aldoss I, Burke PW, et al. Acute lymphoblastic leukemia, Version 2.2024, NCCN Clinical Practice Guidelines in Oncology. J Natl Compr Canc Netw 2024; 22:563–576. [DOI] [PubMed] [Google Scholar]
- 13.Rodrigues CH, Pires DE, Ascher DB. DynaMut: predicting the impact of mutations on protein conformation, flexibility and stability. Nucleic Acids Res 2018; 46:W350–W355. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Richards S, Aziz N, Bale S, Bick D, Das S, Gastier-Foster J, et al.; ACMG Laboratory Quality Assurance Committee. Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology. Genet Med 2015; 17:405–424. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Jiménez-Morales S, Ramírez-Florencio M, Mejía-Aranguré JM, Núñez-Enríquez JC, Bekker-Mendez C, Torres-Escalante JL, et al. Analysis of thiopurine S-methyltransferase deficient alleles in acute lymphoblastic leukemia patients in mexican patients. Arch Med Res 2016; 47:615–622. [DOI] [PubMed] [Google Scholar]
- 16.Ramirez-Florencio M, Jiménez-Morales S, Barbosa-Cobos RE, López-Cano DJ, Ramírez-Bello J. High frequency of mutant thiopurine S-methyltransferase genotypes in Mexican patients with systemic lupus erythematosus and rheumatoid arthritis. Clin Rheumatol 2018; 37:963–970. [DOI] [PubMed] [Google Scholar]
- 17.Candelaria M, Taja-Chayeb L, Vidal S, Gutierrez O, Dueñas-Gonzalez A. Importance of the polymorphisms of the thiopurine S-methyltransferase gene (TMPT) determination in Mexican Mestizo patients with acute lymphoblastic leukemia (ALL). Blood 2007; 110:4222–4222. [Google Scholar]
- 18.González-del Angel A, Bermúdez-López C, Alcántara-Ortigoza MA, Vela-Amieva M, Castillo-Cruz RA, Martínez V, Torres-Espíndola L. Thiopurine S -methyltransferase (TPMT) genetic polymorphisms in Mexican newborns: TPMT polymorphisms in newborns from Mexico City. J Clin Pharm Ther 2009; 34:703–708. [DOI] [PubMed] [Google Scholar]
- 19.Ziyatdinov A, Torres J, Alegre-Díaz J, Backman J, Mbatchou J, Turner M, et al.; Regeneron Genetics Center. Genotyping, sequencing and analysis of 140,000 adults from Mexico City. Nature 2023; 622:784–793. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Matreyek KA, Starita LM, Stephany JJ, Martin B, Chiasson MA, Gray VE, et al. Multiplex assessment of protein variant abundance by massively parallel sequencing. Nat Genet 2018; 50:874–882. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Pui C-H, Campana D, Pei D, Bowman WP, Sandlund JT, Kaste SC, et al. Treating childhood acute lymphoblastic leukemia without cranial irradiation. N Engl J Med 2009; 360:2730–2741. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Texis T, Guzmán-Cruz C, Rodríguez-Dorantes M, Sánchez-García S, Mino-León D, Gonzalez-Covarrubias V. Genotyping NUDT15*3 rs1166855232 reveals higher frequency of potential adverse effects of thiopurines in Natives and Mestizos from Mexico. Pharmacol Rep 2022; 74:257–262. [DOI] [PubMed] [Google Scholar]
- 23.Maillard M, Nguyen JQ, Yang W, Hoshitsuki K, Relling MV, Caudle KE, et al. Clinical Actionability of the NUDT15 *4 (p.R139H) Allele and Its Association With Hispanic Ethnicity. Clin Pharmacol Ther 2025; 117:724–731. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.