NUDT15 Haplotypes and Diplotypes Predict Thiopurine-Induced Leukopenia and the Influence of Prolonged Exposure to Azathioprine on Hematologic Indices in Patients with Inflammatory Bowel Diseases

Background . NUDT15 gene polymorphisms have been identifed to predispose Asian patients with an infammatory bowel disease (IBD) to thiopurine-induced leukopenia. Tis study predicted the infuence of NUDT15 haplotypes and diplotypes on azathioprine (AZA)-induced leukopenia as well as the long-term infuence of AZA on hematologic parameters in IBD. Methods . 194 IBD patients were tested for NUDT15 genotypes. We collected clinical data of 80 patients with AZA treatment including adverse events, dosage, white blood cell (WBC) count, platelet (PLT) count, and mean corpuscular volume (MCV) after AZA initiation. Patients without adverse events and drug withdrawal were followed up for at least one year. Te relationship between NUDT15 haplotypes and diplotypes and leukopenia was analyzed. Results . Te haplotypes NUDT15 c.415C > Tand c.36_37insGGAGTC as well as the diplotypes NUDT15 ∗ 1/ ∗ 2, ∗ 3/ ∗ 3, and ∗ 3/ ∗ 5 were signifcantly associated with AZA-induced leukopenia. Only one patient with NUDT15 c.52G > A experienced leukopenia. NUDT15 ∗ 1/ ∗ 3 was not associated with leukopenia. After AZA initiation, the WBC count showed a downward trend in both wild types and mutants. Te mean of WBC count in the mutant group at 1st month after AZA initiation was lower than that in the wild-type group ( P � 0 . 006). Te MCV increased gradually in mutant cases ( P � 0 . 039), and the diferences were obvious at 6th and 12th months compared with the baseline ( P � 0 . 014 and P � 0 . 042, respectively). Te PLT count showed a decreasing trend in the mutant group, but there was no diference until 11months after initiating treatment ( P � 0 . 023). Te fnal dose of AZA in the NUDT15 mutant group was signifcantly lower than that in the wild-type group ( P � 0 . 006). Conclusion . NUDT15 polymorphisms may be an appropriate predictor of AZA abnormal hematologic indices in IBD patients. It is necessary for IBD patients to monitor hematological indices and optimize AZA therapy.


Introduction
Tiopurine are widely applied to maintain remission in patients with infammatory bowel disease (IBD), especially steroid-dependent and steroid-resistant cases [1][2][3]. In addition, thiopurine can prevent the development of antidrug antibodies in those receiving antitumor necrosis factor (TNF)-α antibody [4,5]. In patients with Crohn's disease (CD), it can also be used to prevent relapse after enterectomy [6]. Nevertheless, drug-induced adverse events often interfere with the application of thiopurine in clinical practice. Up to 15%-30% of IBD patients discontinue thiopurine therapy due to the intolerable adverse efects [7]. Myelosuppression, especially leukopenia, is the most common adverse event which might increase the risk of infection and mortality [8][9][10].
However, it remains unknown whether the diplotypes to predict thiopurine-induced leukopenia in IBD patients is superior to the haplotypes. In addition, there is still a lack of studies on the long-term efects of thiopurine on hematological indices in Chinese IBD patients. Terefore, we evaluated the associations between NUDT15 polymorphisms and azathioprine (AZA)-induced leukopenia by testing the haplotypes and diplotypes. Meanwhile, we also evaluated the infuence of prolonged exposure to AZA on hematologic parameters in Chinese patients with IBD by NUDT15 polymorphisms.  [20] and (2) complete clinical data. Exclusion criteria were as follows: (1) recent blood transfusion; (2) administration of other immunosuppressants such as ciclosporin or methotrexate (MTX); (3) treatments potentially interfering with AZA metabolism, including allopurinol, febuxostat, and angiotensin-converting enzyme inhibitor (ACEI); (4) severe cardiac, hepatic, and/or renal insufciency; (5) active infection; and (6) pregnancy or lactation.

Materials and Methods
Te initial dose of AZA was 0.5-1.0 mg/kg daily [20]. Patients treated with AZA underwent routine blood tests once a week during the frst month, then twice a month for two months, followed by every month, as well as liver function tested every month. AZA dose was increased to 2 mg/kg daily as the target maintenance dosage if patients had no adverse efects [20]. HWE analysis of the research subjects was carried out by comparing the detected distribution of allele frequencies with the theoretical distribution estimated based on the SNP allelic frequencies. P > 0.05 (chi-squared statistics) was considered to indicate equilibrium.

Data Collection and Follow-Up.
Te following data were collected: age, sex, weight, type of IBD, Montreal classifcation, dosage and duration of AZA, concomitant medications, NUDT15 genotypes, white blood cell (WBC) count, platelet (PLT) count, mean corpuscular volume (MCV), and adverse events. Te primary endpoint was leukopenia. Te secondary endpoint was drug withdrawal for other adverse events or inefectiveness. Patients without adverse events and drug withdrawal were followed up for at least one year. Te tolerance dose of AZA was calculated from the fnal dose of the patients who could continue AZA for more than six months [18].
Te Ethics Committee of the First Afliated Hospital of Nanjing Medical University approved this study (2020-SR-132). Written informed consent was obtained from every patient included in the study. Te study conforms to the Declaration of Helsinki.

Statistical Analysis.
Student's t-test or one-way analysis of variance (ANOVA) was used for normally distributed variables. Te Mann-Whitney U-test was utilized for nonnormally distributed variables. Probability (P) values were adjusted by Bonferroni correction when conducting the pairwise comparison. Diferences between categorical variables were analyzed using chi-squared or Fisher's exact test. Logistic regression analyses were performed to identify the associations of leukopenia with candidate variant and the relevant factors. All statistical analyses were carried out using SPSS version 22.0 (SPSS Inc., Chicago, IL, USA) and GraphPad Prism 8 (GraphPad software, La Jolla, CA). P values less than 0.05 were considered statistically signifcant.

Clinical Characteristics of IBD Patients.
A total of 194 patients with IBD were tested for NUDT15 genotypes. 80 of 194 received AZA treatment. Te baseline characteristics of 80 patients treated with AZA are summarized in Table 1.
Tere was no diference in other clinical characteristics between the two groups ( Table 1).

Te Adverse Events and NUDT15
Polymorphisms of Patients with IBD. Te incidence of adverse events in patients with NUDT15 mutants was signifcantly higher than that in NUDT15 wild types (68.2% vs. 27.6%, P � 0.002). Leukopenia accounted for 68% of all adverse events. Te diference in the incidence of leukopenia between patients with NUDT15 wild types and NUDT15 mutants was signifcant (15.5% vs. 54.5%, P � 4 × 10 − 4 ). In addition, AZA fnal dosage of NUDT15 mutant cases was lower than the dosage of NUDT15 wild-type cases (P � 0.013).
Furthermore, we analyzed the incidence of diferent grades and phases of leukopenia among all NUDT15 genotypes (Table 3). Patients with NUDT15 c.415C > T heterozygotes were more likely to experience grade 2 leukopenia early (P � 0.002 and P � 8 × 10 − 4 , respectively). Te only one patient carrying NUDT15 c.415C > T homozygote developed grade 3 leukopenia early, while NUDT15 c.36_37insGGAGTC was associated with grade 1 and late leukopenia (P � 0.003 and P � 0.016, respectively). Te only one NUDT15 c.52G > A sufered grade 2 and early leukopenia.

Univariate and Multivariate Analyses of AZA-Induced
Leukopenia. Univariate analysis showed that NUDT15 c.415C > T and c.36_37insGGAGTC signifcantly contributed to AZA-induced leukopenia. In contrast, sex, age, type of IBD, AZA fnal dosage, AZA duration, and concomitant therapy were not associated with the risk of leukopenia (Table 5). Further multivariate logistic regression analysis confrmed that NUDT15 c.415C > T was the only independent factor increasing the risk of developing AZAinduced leukopenia (Table 5).

Infuence of Prolonged Exposure to AZA on Hematologic
Indices in Diferent Genotypes of NUDT15. We investigated the alterations of PLT, MCV, and WBC counts after initial AZA treatment (Figure 1). At the baseline, there was no signifcant diference in theses hematologic indices between NUDT15 mutant group and NUDT15 wild-type group. Te WBC count showed a downward trend in both wild types and mutants after AZA therapy, but there was no statistical diference. Te mean of WBC count in the mutant group at 1st month after AZA initiation was signifcantly lower than that in the wild-type group (P � 0.006) (Figure 1(a)). We found that MCV in mutant cases increased gradually (P � 0.039) (Figure 1(b)). Te diferences were obvious at 6th and 12th months after initial therapy compared with the baseline (P � 0.014 and P � 0.042, respectively). Te PLT   count decreased gradually in the mutant group, but the decreasing trend was not signifcant (P > 0.05). Te changing trend of PLT count was also not apparent in the wild-type group. Although there was no statistical diference in PLT count between the two groups at 1st, 3rd, 6th, and 9th months, the PLT count was lower in mutant cases than wildtype cases at 12th month (P � 0.023) (Figure 1(c)).
In this study, we verifed again that NUDT15 c.415C > T, c.52G > A, and c.36_37insGGAGTC were associated with thiopurine-induced leukopenia. Our results suggested that NUDT15 polymorphisms can predict AZA-induced leukopenia, and the monitoring of hematologic indices in IBD patients with AZA therapy should be long-term and careful. Logistic regression analysis confrmed that NUDT15 polymorphism was the only factor increasing the risk of developing AZA-induced leukopenia. Tere were no diferences in sex, AZA dosage, and corticosteroid usage in patients with or without leukopenia. Combined with previous results, we inferred that corticosteroid might have no protective efect on AZA-induced leukopenia in IBD patients [13,18,21], and the myelosuppression caused by NUDT15 polymorphisms may be dose independent [8]. Te association between sex with leukopenia is still controversial [13,18,23]. More studies are needed to prove whether there is a gender diference in AZA-induced leukopenia in the IBD population.
In this study, we evaluated the infuence of prolonged exposure to AZA on hematologic indices in Chinese IBD population by NUDT15 polymorphisms, including WBC count, MCV, and PLT count. WBC count showed a downward trend after beginning AZA therapy. Te mean of WBC count in the mutant group at 1st month after AZA initiation   was signifcantly lower than that in the wild-type group. Te results were consistent with our previous conclusion that patients with mutant diplotypes were prone to early leukopenia, which was similar to a previous study in Japanese IBD population [24]. In addition, we also found that the diferences in the WBC count between two groups at 9th and 12th month were slight. Tis phenomenon might be due to the fact that the patients who could not tolerate AZA already discontinued the therapy, while the WBC count of remaining patients who could tolerate AZA might be similar to that of patients with NUDT15 wild-types. Terefore, we should monitor the WBC count more closely, especially in the frst month. Several studies demonstrated that thiopurine-mediated inhibition of DNA synthesis in bone marrow precursor cells leads to megaloblastic erythropoiesis, which is characterized by an increase in the MCV value [25][26][27][28]. MCV elevation above 7 f is also reported to be an indicator of optimal 6-TGN levels [26]. Measurement of MCV may be a simple and inexpensive alternative to measurement of 6-TGN concentrations in patients treated with azathioprine or 6mercaptopurine [26]. We did not measure the concentrations of folate and vitamin B12, but the MCV of all patients were in the normal range before AZA initiation. We observed that the MCV increased gradually in mutant cases, which indicated that the best MCV value for predicting the optimal 6-TGN levels should be adjusted according to NUDT15 genotypes. It also suggested that the red blood cells of patients with mutations were more sensitive to AZA than those with wild-types. Tese conclusions were similar to a Japanese study reporting that the MCV was higher in the NUDT15 c.415C > T C/T group than in the C/C group [24]. It is controversial whether there exists a diference in the PLT count between the NUDT15 wild-type group and mutant group after AZA initiation [21,24]. Our study revealed that the PLT count decreased gradually in the mutant group but not statistically diferent. However, the PLT count in the mutant group was signifcantly lower at the 12th month after initiation of AZA. Tese results suggested that it is necessary for IBD patients to carefully monitor hematological indices for a long time and optimize the AZA dose.
Finally, we compared the fnal dose of AZA in all patients and the tolerance dose in patients with diferent diplotypes. Te fnal dose of AZA in the NUDT15 mutant group is lower than that in the wild-type group. Tere was no signifcance in the tolerance dose among these diplotypes. Patients with both * 3/ * 3 and * 3/ * 5 diplotypes were unable to tolerate AZA and discontinued AZA therapy at an early stage. Tere might be type 2 errors in the statistical analysis due to the small sample size. However, we cannot rule out the possibility that there was no diference in the tolerance dose between patients with * 1/ * 1 and patients, who did not develop leukopenia after six months of AZA therapy, with mutations. Tis result may support the view that the myelosuppression caused by NUDT15 polymorphisms may be dose independent [8] and suggest that other factors besides NUDT15 polymorphisms may contribute to AZA intolerance in Chinese IBD patients.
Tis study also has several limitations. First, it was a small sample study in a single center. Second, because of the various causes of anemia in IBD patients, we did not analyze the hemoglobin count of IBD patients. Tird, other genes that may afect AZA metabolism were not detected.  In conclusion, despite these limitations, NUDT15 polymorphism is indeed an appropriate predictor of AZA intolerance in IBD patients. Our study compared the differences in grades and phases of leukopenia caused by AZA among diferent diplotypes. We found that testing diplotypes to predict thiopurine-induced leukopenia in IBD patients may be superior to the haplotypes. In addition, this study had evaluated the efects of NUDT15 polymorphisms on the prediction of prolonged exposure to AZA on hematological indices of Chinese IBD population, which provides more suggestions for rational regimen of AZA therapy in patients with IBD.

Data Availability
Te data used to support the fndings of this study are available from the corresponding author on reasonable request.

Disclosure
Tis study has been presented as AGA Abstracts previously.