Thiopurine and Fluoropyrimidine Pharmacogenetics

Thiopurine and fluoropyrimidine pharmacogenetics encompasses the pre-treatment molecular screening of germline variants in three catabolic and dephosphorylating enzymes—thiopurine S-methyltransferase (TPMT), Nudix hydrolase 15 (NUDT15), and dihydropyrimidine dehydrogenase (DPYD)—that regulate the intracellular inactivation of purine and pyrimidine antimetabolite drugs. Because both thiopurines (azathioprine, 6-mercaptopurine, and 6-thioguanine) and fluoropyrimidines (5-fluorouracil, capecitabine, and tegafur) possess narrow therapeutic indices and rely on a single dominant enzymatic pathway to clear >80% of the administered dose or dephosphorylate the active nucleotide pool, inherited deficiency of these enzymes diverts drug flux into cytotoxic anabolic nucleotides.

Without preemptive dose adjustment, patients who are heterozygous Intermediate Metabolizers (IMs) or homozygous Poor Metabolizers (PMs) for TPMT, NUDT15, or DPYD develop life-threatening or fatal adverse drug reactions—including grade 4 neutropenia, pancytopenia, hemorrhagic mucositis, and secretory diarrhea—within days to weeks of initiating standard doses. Pre-treatment genotyping and CPIC/DPWG-guided dose reductions allow clinicians to lower systemic exposure to match that of wild-type Normal Metabolizers while preserving antileukemic, antitumor, or immunosuppressive efficacy.

Anabolic Bioactivation to 6-Thioguanine Nucleotides (6-TGN)

The thiopurine antimetabolites—azathioprine (an imidazole prodrug cleaved non-enzymatically and by glutathione S-transferases into 6-MP), 6-mercaptopurine (6-MP), and 6-thioguanine (6-TG)—are cornerstone therapies for pediatric and adult acute lymphoblastic leukemia (ALL), inflammatory bowel disease (Crohn's disease and ulcerative colitis), autoimmune hepatitis, and organ transplant rejection prophylaxis.

All three thiopurines are inactive prodrugs that undergo intracellular salvage-pathway bioactivation initiated by hypoxanthine-guanine phosphoribosyltransferase (HGPRT):

1. HGPRT converts 6-MP into 6-thioinosine monophosphate (6-TIMP), which inosine monophosphate dehydrogenase (IMPDH) and GMP synthetase oxidize and aminate into 6-thioguanosine monophosphate (6-TGMP) (whereas 6-TG is converted directly by HGPRT into 6-TGMP).
2. Kinases sequentially phosphorylate 6-TGMP into 6-TGDP and 6-thioguanosine triphosphate (6-TGTP)—collectively termed 6-thioguanine nucleotides (6-TGN).
3. 6-TGTP (and its deoxy-derivative 6-dTGTP) exerts cytotoxicity through two mechanisms: (i) incorporation into nascent DNA strands during S-phase replication, which triggers futile mismatch-repair cycles and double-strand DNA breaks in proliferating leukocytes, and (ii) binding to the small GTPase Rac1 in activated T lymphocytes in place of GTP, blocking Vav guanosine nucleotide exchange factor activity and inducing mitochondrial apoptosis.

TPMT Inactivation Pathway and Proteasomal Misfolding (*\*2*, *\*3A*, *\*3C*)

In hematopoietic tissues (which lack xanthine oxidase), two enzymatic checkpoints prevent excessive 6-TGN accumulation. The first is thiopurine S-methyltransferase (TPMT), encoded by the TPMT gene on chromosome 6p22.3. Using S-adenosyl-L-methionine (SAM) as a methyl donor, cytosolic TPMT catalyzes the S-methylation of 6-MP into inactive 6-methylmercaptopurine (6-MeMP) and methylates 6-TIMP into 6-methylthioinosine monophosphate (6-MeTIMP, an inhibitor of de novo purine synthesis).

Over 40 TPMT star alleles have been identified, but three non-synonymous alleles account for >90% to 95% of all reduced- or no-function cases worldwide:

• *TPMT\*2* (c.238G>C, p.Ala80Pro)
• *TPMT\*3A*: A cis-haplotype combining two missense mutations—*\*3B* (c.460G>A, p.Ala154Thr in exon 7) and *\*3C* (c.719A>G, p.Tyr240Cys in exon 10). *TPMT\*3A* is the predominant loss-of-function allele in European and Latino populations (allele frequency ~3–5%).
• *TPMT\*3C* (c.719A>G, p.Tyr240Cys alone): The most common loss-of-function allele in African (~4–7%) and East/South Asian (~1.5–2%) populations.

Structurally, the Ala80Pro, Ala154Thr, and Tyr240Cys substitutions disrupt the hydrophobic core of the TPMT Rossmann-fold domain, inducing conformational misfolding. Misfolded TPMT proteins are recognized by molecular chaperones Hsp70 and Hsp90, polyubiquitinated, and rapidly degraded by the 26S ubiquitin–proteasome system (reducing protein half-life from ~18 hours to <45 minutes).

• TPMT Intermediate Metabolizers (~10% of Europeans, e.g., *\*1/\*3A*) accumulate 2- to 3-fold higher erythrocyte 6-TGN levels and require a 30% to 70% reduction in starting dose.
• TPMT Poor Metabolizers (~1 in 300 Europeans, e.g., *\*3A/\*3A*) accumulate 10-fold higher erythrocyte 6-TGN concentrations and suffer universal, life-threatening pancytopenia on standard doses. For malignant disease (ALL), CPIC mandates reducing the starting thiopurine dose 10-fold (administering 10% of the standard dose three times weekly rather than daily); for non-malignant autoimmune conditions, thiopurines should be replaced with a non-thiopurine immunosuppressant.

NUDT15 Nucleotide Dephosphorylation and the *\*3* Allele in Asian and Indigenous Cohorts

Whereas TPMT loss-of-function alleles are uncommon in East Asians (~1.5–2%), East Asian, South Asian, and Indigenous/Hispanic American patients experience high rates of thiopurine-induced leukopenia. In 2014–2015, genome-wide association studies identified the second major thiopurine checkpoint: Nudix hydrolase 15 (NUDT15 on chromosome 13q14.2).

NUDT15 is a nucleotide diphosphatase that directly hydrolyzes the active cytotoxic triphosphates 6-TGTP and 6-dTGTP back into inactive monophosphates (6-TGMP and 6-dTGMP) plus inorganic pyrophosphate (PP_i), preventing their incorporation into DNA. The signature loss-of-function allele *NUDT15\*3* (c.415C>T, p.Arg139Cys, rs116855232) replaces an arginine residue at the base of the catalytic α-helix with cysteine, breaking an intramolecular salt bridge and causing thermal unfolding of the NUDT15 catalytic pocket. *NUDT15\*3* occurs at an allele frequency of ~7% to 10% in East Asians, South Asians, and Indigenous/Admixed Americans (versus <0.5% in Europeans and Africans). Because NUDT15 acts downstream of TPMT on the final 6-TGTP pool, NUDT15 Poor Metabolizers (*\*3/\*3*) suffer marked leukopenia despite having normal total erythrocyte 6-TGN counts (since standard clinical RBC 6-TGN assays do not distinguish 6-TGMP from 6-TGTP) and require the same 10-fold dose reduction or alternative therapy as TPMT PMs.

Dihydropyrimidine Dehydrogenase (DPD) Catabolism of 5-Fluorouracil

The fluoropyrimidines—intravenous 5-fluorouracil (5-FU) and its oral prodrugs capecitabine and tegafur—are backbone chemotherapeutic agents for colorectal, breast, pancreatic, gastric, and head-and-neck carcinomas. Inside tumor and rapidly dividing mucosal/marrow cells, only 1% to 3% of an administered 5-FU dose is anabolized into active cytotoxic nucleotides: 5-fluoro-2'-deoxyuridine 5'-monophosphate (FdUMP), which forms a covalent ternary complex with thymidylate synthase (TYMS) and 5,10-methylenetetrahydrofolate to block dTTP synthesis ("thymine-less death"), and FUTP, which misincorporates into RNA.

The remaining 80% to 85% of administered 5-FU is rapidly inactivated in the liver by the rate-limiting catabolic enzyme dihydropyrimidine dehydrogenase (DPD), encoded by the large DPYD gene (23 exons spanning 843 kb on chromosome 1p21.3). DPD is a homodimeric flavoprotein containing FAD, FMN, and four iron-sulfur ([4Fe-4S]) clusters that transfers electrons from NADPH to reduce the pyrimidine 5,6-double bond of 5-FU into inactive 5,6-dihydrofluorouracil (DHFU), which is subsequently cleaved into α-fluoro-β-alanine (FBAL) for renal excretion.

The Four Canonical Actionable DPYD Variants and Emergency Antidote Therapy

Approximately 3% to 8% of European patients are heterozygous carriers of a loss-of-function or decreased-function DPYD variant, and ~0.1% to 0.2% have complete DPD deficiency. Clinical guidelines from CPIC, DPWG, and the European Medicines Agency (EMA) mandate pre-treatment testing for four core DPYD variants (plus the African-enriched variant c.557A>G, p.Tyr186Cys):

1. *DPYD\*2A* (c.1905+1G>A, IVS14+1G>A, rs3918290; Activity Value = 0): Disrupts the invariant GT splice donor dinucleotide at the 5' end of intron 14, causing the spliceosome to skip the entire 165-bp exon 14 and producing a catalytically dead truncated protein.
2. *DPYD\*13* (c.1679T>G, p.Ile560Ser, rs55886062; Activity Value = 0): Destabilizes the FMN-binding domain, abolishing enzymatic activity.
3. c.2846A>T (p.Asp949Val, rs67376798; Activity Value = 0.5): Impairs electron transfer across the C-terminal [4Fe-4S] iron-sulfur clusters, reducing catalytic activity by ~50%.
4. HapB3 (c.1129-5923C>G deep intronic splice variant in intron 10, in linkage disequilibrium with synonymous c.1236G>A p.Glu412Glu, rs75017182 / rs56038477; Activity Value = 0.5): Creates a cryptic splice donor site that inserts a 44-bp intronic fragment containing a premature stop codon into a fraction of DPYD transcripts. HapB3 is the single most common decreased-function DPYD allele in Europeans (carrier frequency ~4.1–4.8%).

In DPYD Intermediate Metabolizers (Activity Score = 1.0 or 1.5), 5-FU clearance drops by 40% to 60%; reducing the initial fluoropyrimidine dose by 50% (followed by cautious pharmacokinetic or clinical titration) reduces grade ≥3 toxicity from >70% down to the ~30% baseline seen in wild-type patients (*1/*1, AS = 2.0). In DPYD Poor Metabolizers (AS = 0 or 0.5), fluoropyrimidines are contraindicated. If a DPD-deficient patient inadvertently receives 5-FU or capecitabine, emergency administration of the oral nucleoside prodrug antidote uridine triacetate (10 g orally every 6 hours for 20 doses within 96 hours) delivers high intracellular concentrations of uridine triphosphate (UTP), which competitively displaces FUTP from RNA incorporation and rescues bone marrow and gastrointestinal crypt stem cells (improving survival from ~16% to >96%).

Why does 6-thioguanine (6-TG) still cause myelosuppression in TPMT-deficient and NUDT15-deficient patients?

Although 6-thioguanine bypasses the initial HGPRT/IMPDH conversion steps of 6-mercaptopurine, 6-TG itself is a direct substrate of TPMT (which methylates 6-TG to inactive 2-amino-6-methylthiopurine) and its active triphosphate metabolite 6-TGTP is the direct substrate of NUDT15. Therefore, patients with either TPMT or NUDT15 deficiency accumulate toxic 6-TGTP levels from 6-thioguanine and require identical pre-treatment genotype-guided dose reductions.

What should be done if a patient inherits heterozygosity in both TPMT and NUDT15 (e.g., *TPMT\*1/\*3A* plus *NUDT15\*1/\*3*)?

Because TPMT acts upstream (methylating 6-MP and 6-TIMP) and NUDT15 acts downstream (dephosphorylating 6-TGTP), inheriting one loss-of-function allele in both genes produces an additive bottleneck on thiopurine inactivation. Compound digenic heterozygotes exhibit thiopurine intolerance comparable to homozygous single-gene Poor Metabolizers and require a deeper starting dose reduction (50% to 80% reduction, or 20–50% of standard dose) with close complete blood count monitoring.

How does uridine triacetate rescue patients from life-threatening 5-fluorouracil toxicity?

Uridine triacetate is an orally bioavailable acetylated prodrug of uridine that achieves 4- to 6-fold higher plasma uridine bioavailability than equimolar oral uridine. Esterases cleave uridine triacetate into free uridine, which is phosphorylated into UTP. High intracellular UTP outcompetes toxic FUTP for incorporation into cellular RNA in normal bone marrow progenitors and gastrointestinal mucosal cells, halting lethal cell death when initiated within 96 hours of fluoropyrimidine administration.

1. Relling, M. V., Schwab, M., Whirl-Carrillo, M., Suarez-Kurtz, G., Pui, C. H., Stein, C. M., et al. (2019). "Clinical Pharmacogenetics Implementation Consortium guideline for thiopurine dosing based on TPMT and NUDT15 genotypes: 2018 update." Clinical Pharmacology & Therapeutics, 105(5), 1095–1105. https://doi.org/10.1002/cpt.1304
2. Moriyama, T., Nishii, R., Perez-Andreu, V., Yang, W., Klussmann, F. A., Zhao, X., et al. (2016). "NUDT15 polymorphisms alter thiopurine metabolism and hematopoietic toxicity." Nature Genetics, 48(4), 367–373. https://doi.org/10.1038/ng.3508
3. Amstutz, U., Henricks, L. M., Offer, S. M., Barbarino, J., Schellens, J. H. M., Swen, J. J., et al. (2018). "Clinical Pharmacogenetics Implementation Consortium (CPIC) guideline for dihydropyrimidine dehydrogenase genotype and fluoropyrimidine dosing: 2017 update." Clinical Pharmacology & Therapeutics, 103(2), 210–216. https://doi.org/10.1002/cpt.911
4. Henricks, L. M., Lunenburg, C. A. T. C., de Man, F. M., Meulendijks, D., Frederix, G. W. J., Kienhuis, E., et al. (2018). "*DPYD* genotype-guided dose individualisation of fluoropyrimidine therapy in patients with cancer: a prospective safety analysis." The Lancet Oncology, 19(11), 1459–1467. https://doi.org/10.1016/S1470-2045(18)30686-730686-7)