Establishment of a novel predictive model for thiopurine-induced leucopenia based on the polymorphisms in the DNA-thioguanine nucleotide metabolite pathway in Chinese patients with inflammatory bowel disease

Thiopurine-induced leucopenia (TIL) affects more than 20% of Asians despite dose optimization via NUDT15 and TPMT genotyping. Elevated levels of DNA-thioguanine nucleotide (DNA-TG) have been implicated in the development of TIL. This study aimed to identify genetic polymorphisms influencing TIL through the DNA-TG metabolic pathway and construct a predictive model in Chinese inflammatory bowel disease patients. A prospective cohort of inflammatory bowel disease patients receiving thiopurines from December 2018 to December 2019 was analyzed. A total of 294 single-nucleotide polymorphisms across 29 genes were screened using the MassARRAY system. Candidate single-nucleotide polymorphisms associated with DNA-TG exposure or TIL (P < .1) were selected for validation. Independent risk factors were identified through multivariate logistic regression and incorporated into a predictive nomogram. Thiopurine metabolites, including DNA-TG and 6TGN, were quantified. DNA-TG, but not 6TGN, exposure was significantly associated with TIL (P = 8.3 × 10-7, P = .41). Sex, along with variants in NUDT15 (rs116855232 c.415C > T), TPMT (rs9465102), MOCOS (rs73430958), and RRM1 (rs1735053), were identified as independent predictors of TIL. The resulting nomogram demonstrated good discriminative performance (area under the curve = 0.72, 95% confidence interval: 0.65-0.78). Individuals stratified into low-, medium-, and high-risk groups based on total point thresholds (25.0 and 196.8), showed significant differences in TIL incidence in both NUDT15 variants and nonvariants (P = 1.8×10-5, P = .045). Our findings indicate that in addition to NUDT15, novel genetic variants in TPMT, MOCOS, and RRM1 are potential predictive markers of TIL. The new nomogram enables more accurate identification of individuals at high risk of TIL, allowing for the proactive adjustment of thiopurine therapy. SIGNIFICANCE STATEMENT: Compared with prior studies limited to NUDT15 and TPMT variants, it was the first pharmacogenomic study to investigate genetic polymorphisms affecting thiopurine-induced leucopenia (TIL) and DNA-thioguanine nucleotide metabolism, constructing a clinically actionable TIL prediction model. We identified novel associations of TPMT, MOCOS, and RRM1 variants with TIL risk, offering key insights for precision dosing in Asian populations. The validated nomogram integrates these biomarkers to stratify patients into distinct risk groups, facilitating tailored thiopurine therapy.