Enzyme:NUDT15

NUDT15 (nudix hydrolase 15, historically also called MTH2) is a cytosolic hydrolase enzyme whose genetic polymorphism is, alongside TPMT, one of the two central pharmacogenomic determinants of thiopurine safety. It is encoded by the NUDT15 gene on chromosome 13q14.2 and belongs to the nudix hydrolase superfamily, a group of enzymes that hydrolyze nucleoside diphosphates linked to another moiety x. Like TPMT, NUDT15 has no role in the metabolism of any medicine outside the thiopurine class, and like TPMT, its loss-of-function variants convert a standard thiopurine dose into a potentially lethal one. The reason NUDT15 warrants its own page rather than a footnote on the TPMT page is a matter of human population genetics: NUDT15 and TPMT loss-of-function alleles have nearly complementary ancestry distributions, and testing for one without the other systematically under-protects a large fraction of the world's patients.

The history here is recent and unusually well dated. In 2014, a research team centred on the Asan Medical Center in Seoul, studying a Korean inflammatory-bowel-disease cohort, ran a genome-wide association study against the phenotype of thiopurine-induced early leukopenia. The signal that emerged was not at TPMT, the gene every clinician would have predicted, but at a then-obscure gene called NUDT15: a single common missense variant, rs116855232 (Arg139Cys), was strongly associated with thiopurine-induced leukopenia, and it explained a large share of the myelosuppression that TPMT genotyping had been failing to predict in East Asian patients. Two years later, Moriyama and colleagues at St. Jude Children's Research Hospital in Memphis worked out the mechanism: NUDT15 polymorphisms alter thiopurine metabolism by changing how much active thioguanine nucleotide reaches DNA, and the loss-of-function variants produce measurably greater hematopoietic toxicity. A parallel 2015 study had already extended the clinical association to children with acute lymphoblastic leukaemia, and by 2019 the Clinical Pharmacogenetics Implementation Consortium had folded NUDT15 into its thiopurine-dosing guideline as a co-equal partner to TPMT.

Tissue distribution

NUDT15 is a small cytosolic enzyme expressed broadly across human tissues. As with TPMT, the tissue that matters clinically is the hematopoietic compartment of the bone marrow, where the consequences of impaired thiopurine-metabolite clearance show up as the neutropenia, leukopenia, and broader myelosuppression that limit thiopurine therapy. Clinical assessment of NUDT15 status is done by genotyping rather than by an enzyme-activity assay; there is no routine red-cell phenotype test for NUDT15 equivalent to the long-established red-cell TPMT activity assay.

Function and substrate spectrum

NUDT15 is a nudix hydrolase that dephosphorylates the triphosphate forms of the active thioguanine nucleotides, principally thio-deoxyguanosine triphosphate (the thiopurine analogue of dGTP) and the corresponding thioguanosine triphosphate. By converting these active triphosphate metabolites back to monophosphates, NUDT15 limits the quantity of thioguanine nucleotide that is available for incorporation into DNA and RNA. Incorporation of thioguanine into DNA is one of the principal mechanisms of thiopurine cytotoxicity, so NUDT15 functions as a guardian enzyme: it removes the most dangerous metabolic species before they can do damage.

This produces the same inverted logic seen on the TPMT page. When NUDT15 activity is reduced, more thioguanine triphosphate survives to be incorporated into DNA, DNA-thioguanine levels climb, and hematopoietic toxicity follows. The clinical danger of NUDT15 deficiency, like that of TPMT deficiency, comes from the loss of a protective metabolic step rather than from the failure to activate a prodrug.

The clinically relevant substrate set is identical to that of TPMT, because the two enzymes act at different points on the metabolism of the same three thiopurine medicines. Because the list is complete rather than curated, this table is presented in full rather than as a collapsible near-complete excerpt.

Medicine	Therapeutic class	NUDT15 relevance	Clinical notes
Azathioprine	Immunosuppressant (IBD, transplant, autoimmune disease)	major	Prodrug of mercaptopurine; the mercaptopurine pharmacology applies after conversion. NUDT15-deficiency myelosuppression risk is the same as for mercaptopurine. FDA labeling references NUDT15 (and TPMT) testing.
Mercaptopurine	Antineoplastic (ALL maintenance); immunosuppressant (IBD)	major	6-MP. NUDT15 poor metabolizers receiving standard-dose mercaptopurine develop severe myelosuppression. FDA labeling references NUDT15 and TPMT testing.
Thioguanine	Antineoplastic (AML, ALL); occasionally IBD	major	6-TG. Because thioguanine enters the active-nucleotide pool more directly, NUDT15 status is, if anything, particularly relevant for this medicine.

NUDT15 acts on the downstream thionucleotide triphosphate metabolites of these medicines, not on the parent compounds directly. The three thiopurines above are the entire clinically relevant substrate set; NUDT15 has no established role in the metabolism of any other medicine.

Phenotype categories

NUDT15 phenotype is assigned from the two NUDT15 alleles, and the CPIC convention recognises three metabolizer phenotypes:

• Normal metabolizer (NM): two normal-function alleles (typically \*1/\*1).
• Intermediate metabolizer (IM): one normal-function and one no-function (or decreased-function) allele.
• Poor metabolizer (PM): two no-function or decreased-function alleles.

There is no ultra-rapid NUDT15 category; gain-of-function alleles are not established for this enzyme.

The dosing stakes mirror those for TPMT. CPIC recommends standard thiopurine dosing for normal metabolizers; a reduced starting dose for intermediate metabolizers; and, for poor metabolizers, a drastically reduced dose (for azathioprine and mercaptopurine, on the order of 10% of the standard dose) or an alternative non-thiopurine agent. A NUDT15 poor metabolizer given a standard thiopurine dose faces the same risk of profound, potentially fatal myelosuppression as a TPMT poor metabolizer.

Major variants

• \*1, the reference normal-function allele.
• \*2, a decreased-to-no-function haplotype that includes the Arg139Cys substitution along with a second variant in cis.
• \*3 (rs116855232, Arg139Cys), the single most important NUDT15 loss-of-function allele worldwide. It is the variant identified in the original 2014 association study, and it carries the large majority of the clinical NUDT15 loss-of-function burden.
• \*4, \*5, \*6, \*9 and other rarer alleles, several of them characterized in the 2017 work that extended NUDT15 allele discovery across diverse ancestral populations.

The defining feature of NUDT15 population genetics is its ancestry distribution. The \*3 allele reaches an allele frequency of roughly 10% in East Asian populations and is also common in Hispanic and Latino populations and in South Asian populations, while it is rare in European-ancestry populations (on the order of 0.2%) and rare in African-ancestry populations. This is very nearly the mirror image of the TPMT loss-of-function distribution, and the consequences of that mirror are the subject of the next section.

Inhibitors and inducers

NUDT15 is not an established drug-interaction enzyme. There is no clinically characterized NUDT15 inhibitor or inducer that produces a meaningful interaction in the way the aminosalicylates inhibit TPMT or rifampin induces the cytochromes P450. NUDT15 status is therefore, in practical terms, a fixed genetic property of the patient rather than something modulated by co-prescription.

Relationship to TPMT

NUDT15 and TPMT are the two genes of the modern thiopurine-pharmacogenomics story, and the clinical case for testing both rests on a single population-genetics fact: their loss-of-function alleles have nearly complementary ancestry distributions.

TPMT loss-of-function is comparatively common in European-ancestry and African-ancestry populations and uncommon in East Asian populations. NUDT15 loss-of-function is comparatively common in East Asian, Hispanic and Latino, and South Asian populations and uncommon in European-ancestry and African-ancestry populations. Neither gene, tested alone, protects patients across the full range of human ancestry.

The clinical history made this concrete in a way worth stating plainly. For decades, TPMT was the thiopurine-safety test, and it worked reasonably well in the European-ancestry populations in which it had been developed and validated. In East Asian patients, however, clinicians repeatedly observed severe thiopurine myelosuppression in individuals whose TPMT genotype and TPMT activity were entirely normal. The 2014 discovery of NUDT15 resolved that long-standing puzzle: those patients were NUDT15 poor or intermediate metabolizers, and a TPMT-only testing strategy was structurally incapable of detecting them. A TPMT-only testing program is therefore not merely incomplete but inequitable, because the patients it fails to protect are concentrated in particular ancestral populations. Current best practice, and the CPIC guideline, is to test TPMT and NUDT15 together before starting a thiopurine. See Enzyme:TPMT for the TPMT account in full.

Clinical implications, summary

• Pre-prescription testing. Test NUDT15 together with TPMT before initiating any thiopurine. FDA labeling for azathioprine and mercaptopurine references NUDT15 status, and a TPMT-only strategy under-protects patients of East Asian, Hispanic, Latino, and South Asian ancestry in particular.
• Known PM phenotype: drastically reduce the thiopurine dose (on the order of 10% of standard) or select an alternative non-thiopurine agent. Standard dosing in a poor metabolizer risks fatal myelosuppression.
• Known IM phenotype: start at a reduced dose with blood-count monitoring and titration, per the CPIC algorithm.
• Combined TPMT and NUDT15 interpretation: the CPIC guideline provides a joint phenotype framework. A patient who is an intermediate metabolizer for both genes carries more risk than a patient who is intermediate for one and normal for the other, and the guideline addresses these combinations explicitly.

Authoritative resources

Because the clinically relevant NUDT15 substrate set is limited to the three thiopurine medicines listed in full above, there is no large external substrate table for NUDT15 (the canonical Flockhart cytochrome P450 interaction table does not cover NUDT15, which is not a cytochrome P450 and not a classical drug-metabolizing enzyme at all). The authoritative maintained resources for NUDT15 genotype, allele-function assignment, and thiopurine dosing are:

• The CPIC Guideline for Thiopurine Dosing Based on TPMT and NUDT15 Genotypes, which carries the consensus allele-function table and the joint phenotype-dosing algorithm.
• PharmGKB, the pharmacogenomics knowledge base hosted at Stanford University; the NUDT15 gene page indexes the allele catalogue and the underlying primary literature. Available at https://www.pharmgkb.org/.
• The Pharmacogene Variation Consortium (PharmVar) maintains the definitive NUDT15 star-allele nomenclature.

See also

• Enzyme:TPMT, the parallel thiopurine-safety gene with a nearly complementary ancestry distribution.
• Phenotype:NUDT15 poor metabolizer
• Phenotype:NUDT15 intermediate metabolizer
• Phenotype:NUDT15 normal metabolizer
• Enzyme:CYP2D6, Enzyme:CYP3A4, Enzyme:CYP2C19, Enzyme:CYP2C9, Enzyme:UGT1A1 (other drug-metabolizing enzymes covered in the wiki)
• Azathioprine, Mercaptopurine, Thioguanine (the entire clinical NUDT15 substrate set)
• Category:Drug-metabolizing enzymes