Measurements of 6-thioguanine nucleotide levels with TPMT and NUDT15 genotyping in patients with Crohn’s disease

The association between the 6-thioguanine nucleotide (6-TGN) level and clinical remission in Crohn’s disease (CD) remains controversial. Thiopurine-induced leukopenia is a life-threatening complication of CD in Asians that was recently shown to strongly correlate with NUDT15 genetic variants. This study aimed to determine the relationship between thiopurine metabolite levels and therapeutic response, and to investigate the association of NUDT15, TPMT, and thiopurine metabolites with leukopenia in patients with CD. We enrolled 165 adult patients with CD undergoing thiopurine treatment. Clinical evaluation and laboratory examinations were carried out every 2–3 months. We measured thiopurine metabolites levels and genotyped NUDT15 and TPMT. During the median 12-month observational period, 95 (67.9%) patients exhibited clinical response and 45 (32.1%) did not respond to the treatment. The median 6-TGN level was significantly higher in responders than in non-responders (P < 0.001). The odds ratio of patients with a 6-TGN level ≥230 pmol/8 × 108 red blood cells for showing a clinical response was 4.63 (95% CI 1.62–11.9). NUDT15 variant types were strongly associated with developing leukopenia. Patients with NUDT15 homozygous variant genotype developed severe early leukopenia with an average reduction of 88.2% (range, 84–94%) from the baseline white blood cell count at 4 weeks. Our findings support the role of therapeutic drug monitoring in thiopurine maintenance treatment to optimize thiopurine therapy, especially, for non-responding CD patients. Thiopurine treatment should not be recommended to patients with NUDT15 homozygous variant genotype due to severe early leukopenia.


Introduction
Thiopurines, such as azathioprine (AZA) and 6-mercaptopurine , are drugs used in the induction and maintenance therapy of Crohn's disease (CD) [1,2]. AZA and 6-MP have no intrinsic activity, so both drugs must undergo extensive metabolic transformations, resulting PLOS ONE | https://doi.org/10.1371/journal.pone.0188925 December 5, 2017 1 / 13 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 in 6-thioguanine nucleotides (6-TGN) and 6-methylmercaptopurine nucleotides , which are considered to be the primary active metabolites of these drugs [3]. While 6-TGN and 6-MMPN are responsible for therapeutic efficacy, they exert undesirable myelosuppressive effects, and the 6-MMPN level is also associated with hepatotoxicity of 6-MP and AZA [4,5]. Thiopurine S-methyltransferase (TPMT) is the key enzyme in the metabolic pathway of thiopurine compounds. Previous studies have shown that lower TPMT activity is associated with higher 6-TGN and a higher likelihood of clinical response to thiopurine therapy, but also higher risk for leukopenia [6][7][8][9]. However, only about a quarter of inflammatory bowel disease (IBD) patients with thiopurine-induced leukopenia carry a TPMT mutation [10][11][12]. Moreover, while the frequency of TPMT mutations is lower in Asians (1-3%) than in Caucasians (~10%), the occurrence of thiopurine-induced leukopenia is considerably higher in Asians [13][14][15][16]. In addition to the previously known TPMT mutation, NUDT15 (nucleoside diphosphate-linked moiety X-type motif 15) genetic variants have been recently identified to be strongly associated with thiopurine-induced leukopenia in Asians [17][18][19]. Yang et al. reported that the NUDT15 allele encoding p.Arg139Cys was found in 89.4% of early leukopenia cases but in only 6.8% of controls, suggesting that the presence of the NUDT15 allele had a sensitivity of 89.4% and specificity of 93.2% for early leukopenia in Koreans [17].
The aim of this study was to assess the association between thiopurine metabolite levels and therapeutic response and to evaluate the association of NUDT15, TPMT, and metabolites of thiopurines with leukopenia in patients with Crohn's disease (CD). In addition, we evaluated factors that predict 6-TGN concentration and clinical response.

Patients and data collection
This retrospective study included 165 patients with CD older than 20 years of age at Samsung Medical Center, South Korea, from January 2014 to March 2015. Patients received a consistent dose of AZA or 6-MP for at least 3 months prior to measurement of thiopurine metabolite levels. Patients with leukopenia or elevated liver enzyme levels before treatment with thiopurines and pregnant women were excluded. Additionally, patients with 6-thioguanine metabolite levels that had been measured within 12 h of taking AZA or 6-MP and patients taking anti-TNF drugs were excluded. We measured the patients' levels of thiopurine metabolites, including 6-TGN and 6-MMPN, at baseline. Initially, patients received AZA at 0.5-1.0 mg/kg (6-MP 0.25-0.5 mg/kg), and this dose was increased by 0.5 mg/kg (6-MP 0.25 mg/kg) every 1-2 weeks until the target dose was reached. Clinical evaluation and laboratory examinations were performed every 2-3 months according to the clinical condition of the patient during the median 12-month observational period. Clinical and demographic data, including age, sex, weight, diagnosis, AZA or 6-MP dosage in mg/kg of body weight per day, concomitant drug treatments, and toxicity, were collected at every visit. The following laboratory data were assessed: counts of leukocytes, neutrophils, lymphocytes, and platelets; levels of alanine aminotransferase, aspartate aminotransferase, C-reactive protein (CRP); and erythrocyte sedimentation rate (ESR). In addition, blood samples were collected for determination of 6-TGN and 6-MMPN metabolite levels and for TPMT genotyping. This study was approved by the Institutional Review Board of Samsung Medical Center and was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all patients.

Outcome definition
Primary outcomes included correlations between thiopurine metabolite levels and CD activity index (CDAI) scores and a calculated therapeutic 6-TGN cutoff level. CDAI score was calculated at every visit. Therapeutic response and non-response were determined according to these scores. Therapeutic response was defined as patients with remission, which was determined by a CDAI score <150 without a step-up therapy of anti-TNF agents during the followup period. Therapeutic non-response was defined as patients with active disease, which was determined by a CDAI score !150 with or without a step-up therapy of anti-TNF agents during the follow-up period. Additionally, we defined the therapeutic response using CRP cutoff point 0.6 mg/dl instead of CDAI score. A 6-TGN concentration of 235-450 pmol/8 × 10 8 red blood cells (RBC) was considered the optimal therapeutic range. Poor compliance was defined as a 6-TGN level below 100 pmol/8 × 10 8 RBC in the absence of a metabolite profile suggesting hypermethylation of thiopurines to MMPN (6-MMPN/6-TGN >11) [20]. We also evaluated predictors of 6-TGN levels and clinical response.

Thiopurine metabolite assays
Blood samples for determination of thiopurine metabolite levels were obtained at least 12 h after thiopurine treatment. 6-TGN and 6-MMPN RBC levels were measured on an API4000 tandem mass spectrometer (Applied Biosystems, Foster City, CA, USA) equipped with an Agilent Technologies Series 1200 HPLC system (Agilent Technologies, Santa Clara, CA, USA) [21]. Intra-and inter-assay coefficients of variation for thiopurine metabolite measurements were below 10%. The lower detection limits for quantification were 0.1 mol/l for 6-TGN and 0.5 mol/L for 6-MMPN.

TPMT and NUDT15 genotyping
After peripheral blood samples were obtained, genomic DNA was extracted from the peripheral blood leukocytes of patients using a Wizard Genomic DNA Purification Kit (Promega, Madison, WI, USA) according to the manufacturer's instructions. We genotyped four SNPs in the TPMT gene (rs1142345 for Ã  compared using Student's t-test or Mann-Whitney U-test, while differences between categorical variables were analyzed using Chi-squared or Fisher's exact test. To evaluate the precise association between 6-TGN levels and clinical remission, we conducted univariable and multivariable analyses after excluding non-compliant patients. Univariable and multivariable regression analyses were performed to evaluate factors that predict 6-TGN levels. The variables analyzed included age, sex, BMI, 5-aminosalicylic acid (5-ASA) use, steroid use, anti-TNF use, presence of leukopenia, thiopurine dosage, duration of thiopurine treatment, and TPMT genotype. The predictive value of the 6-TGN level for clinical response and adverse event was evaluated using receiver operating characteristic (ROC) curves. The area under the curve (AUC), Youden's index, sensitivity, and specificity were calculated. Youden's index is the difference between the true positive rate (sensitivity, %) and the false positive rate (1 -specificity, %). Finding the point on the ROC curve that maximizes Youden's index provides an optimal cutoff value [25]. Multivariable logistic regression analyses were performed to evaluate factors that predict clinical response and leukopenia. P < 0.05 was considered statistically significant. Statistical analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC, USA).

Patient characteristics
One hundred forty patients with CD, aged 20 years or older, were included in the analysis. The mean (SD) age of the patients was 33.6 (8.8) years. Patients were treated with thiopurine (mean dose, AZA: 1.2 mg/kg/day, 6-MP: 0.6 mg/kg/day) for a period ranging from 7 to 96 months (median = 34 months). During the median 12 months (ranging from 5 to 22 months) of the observational period, 95 (67.9%) patients showed clinical remission based on CDAI score, and 98 (70%) patients showed clinical remission based on CRP level. Patient characteristics are shown in Table 1.
During the study period, several adverse drug reactions were observed. The most common side effect was lymphopenia (37.9%). Leukopenia, neutropenia, and lymphopenia were more common in responders than in non-responders. Incidence rates of other toxic events did not differ between both groups. There were five (5.3%) responders with a variant TPMT genotype ( Ã 1/ Ã 3), while there were no non-responders with variant genotypes.

Thiopurine metabolite levels
Based on CDAI score, 47 (49.5%) patients in the responder group and 8 (17.8%) patients in the non-responder group had 6-TGN concentrations within the optimal range (235-450 pmol/8 × 10 8 RBC) ( Table 2). Based on CRP level, 51 (52%) patients in the responder group and 4 (9.8%) patients in the non-responder group had optimal 6-TGN concentrations. The median 6-TGN concentration among responders was significantly higher than that of nonresponders. The median 6-MMPN level of responders was also significantly higher than that of non-responders, particularly based on the CRP criteria. The non-responder group had a higher proportion of non-compliant patients (6-TGN levels < 100 pmol/8×10 8 RBC) than the responder group. Patients that developed leukopenia had higher levels of 6-TGN than those without leukopenia. The median 6-TGN levels of patients with a variant TPMT genotype were significantly higher than those of patients with wild-type TPMT. In contrast, there was no difference in the median 6-TGN level between patients harboring variant NUDT15 and those carrying wild-type NUDT15. The 5-ASA users exhibited higher levels of 6-TGN and lower ratios of 6-MMPN/6-TGN than 5-ASA non-users (P < 0.001).

NUDT15 and TPMT genotypes and leukopenia
Out of 165 patients, 34 (20.6%) had NUDT15 variant alleles. NUDT15 heterozygous variants were found in 28 patients (17%) and NUDT15 homozygous variants were found in 6 patients (3.6%). The allele frequencies of wild type NUDT15 was 87.2% and the frequency of variant alleles was 12.8%. Table 4 shows the risk of leukopenia according to NUDT15 and TPMT

Discussion
In this study of CD patients on thiopurine maintenance therapy, we showed that patients with higher levels of 6-TGN have a significantly higher chance of being in remission. We also found that the optimal cutoff value of 6-TGN was 230.2 pmol/8 × 10 8 RBC. Patients with 6-TGN levels !230 pmol/8 × 10 8 RBC had a 4.6 times greater chance of clinical response than those with 6-TGN levels <230 pmol/8 × 10 8 RBC. In addition, NUDT15 variant types were strongly associated with developing leukopenia. The association between 6-TGN and clinical remission remains a matter of debate in IBD. Several studies have suggested that 6-TGN metabolite levels cannot be used to predict remission. Gonzalez et al. evaluated 113 patients in a prospective multicenter study and reported  that neither TPMT activity nor 6-TGN concentration predicted treatment outcome [26]. Gilissen et al. found that median 6-TGN levels did not significantly differ between exacerbation and remission groups [3]. In contrast to these findings, several other studies have reported that the monitoring of 6-TGN concentrations may be helpful for managing AZA or 6-MP therapy in IBD patients, as it may identify the optimal AZA/6-MP dose to maximize efficacy while minimizing the risk of toxicity [4,21,27]. A recent meta-analysis demonstrated that patients with 6-TGN levels above a predefined cutoff were three times more likely to experience remission, with a pooled OR of 3.15 (95% CI, 2.41-4.11) [28]. In our study, we found that the optimal 6-TGN level was 230 pmol/8 × 10 8 RBC; patients with 6-TGN levels above this cutoff were 4.6 times more likely to exhibit clinical responses.
TPMT genotyping prior to thiopurine treatment in patients with IBD has been shown to substantially reduce the risk of adverse effects without compromising therapeutic efficacy [29-31]. In Asian populations, including Koreans, the frequencies of TPMT mutations are lower than those reported from Western countries [13][14][15][16]. Among a total of 900 Korean patients genotyped for TPMT, 30 patients (3.3%) had known TPMT variant alleles [22]. In our study, 5 out of 140 patients were genotyped as Ã 1/ Ã 3 (3.6%), and we also found a correlation between TPMT genotype and 6-TGN metabolite levels. A recent genome-wide association study identified a missense variant in the NUDT15 gene that is strongly associated with thiopurine-related myelosuppression in patients with IBD [17]. The NUDT15 genotype frequencies in our study (wild type, 79.4%; variant type, 20.6%) were similar to those recently reported in studies on Koreans and Japanese, and NUDT15 variant genotype is much more common than TPMT mutation [17,32]. Several studies have reported an association between NUDT15 and thiopurine-induced leukopenia [33][34][35]. Our data also showed that leukopenia was more prevalent in patients with variant NUDT15 than in patients with wild-type NUDT15. In addition, 6 patients who could not tolerate thiopurine therapy because of severe neutropenic events harbored the homozygous NUDT15 variant genotype. These findings are consistent with recent reports and imply pretreatment determination of the NUDT15 genotype is necessary to identify IBD patients susceptible to thiopurine-induced leukopenia [17][18][19]. We recommend that treatment with thiopurine should be avoided in patients with NUDT15 homozygous variants, and in cases with heterozygous variants, slow dosage increase rather than the usual schedule, with careful monitoring of adverse effects, is highly desirable.
Thiopurine-induced leukopenia has been primarily explained by TPMT variants that induce the intracellular 6-TGN accumulations [36]. Recent studies found NUDT15 mutation as another critical determinant of thiopurine-induced leukopenia. NUDT15 inactivates thiopurine metabolites by converting thio-guanosine triphosphate (TGTP) to TGMP, thus preventing their incorporation into DNA and negatively affecting desired cytotoxic effects of thiopurines [37]. Theoretically, patients with defective NUDT15 alleles can have excessive levels of thiopurine metabolites and their toxicity. In our study, NUDT15 variants-related leukopenia was not accompanied by an elevation of 6-TGN levels, it may tend to affect the level of TGTP and TdGTP rather than total 6-TGN [38]. Two recent studies in Asia also reported that there was no significant difference in 6-TGN level among the NUDT15 genotypes [18,35]. Therefore, patients with the NUDT15 variants may call for a new therapeutic monitoring method to detect the active metabolites TGTP rather than 6-TGN to explain the mechanism of NUDT15-related leukopenia.
The mean AZA weight-based dosage in this study was 1.2 ± 0.5 mgÁkg -1 Ád -1 , which is much lower than the recommended dose according to most guidelines. A daily AZA dose of 2-3 mg/ kg for individuals with wild-type TPMT is recommended by the Clinical Pharmacogenetics Implementation Consortium guidelines [39], and 1.5 to 2.5 mgÁkg -1 Ád -1 is recommended by the European Crohn's and Colitis Organization [40]. Several studies have reported that lower doses of thiopurines exert sufficient clinical efficacy and result in therapeutic concentrations of 6-TGN in Japanese IBD patients [41][42][43]. The reason for the lower dose of AZA in Asian population can be explained by the frequency of the NUDT15 variants allele. The frequency of the wild-type allele of NUDT15 was higher in African and European populations (>99.0%) than in Asian populations (82.8-93.0%) [24]. The frequency (12.8%) of NUDT15 variant alleles in our study was comparable with that reported for Asian populations in previous studies. Therefore, our findings along with several previous studies suggest that appropriate thiopurine doses for Asian populations may be lower than those recommended by current guidelines [44].
5-ASA is generally used in the induction and maintenance of remission of ulcerative colitis. The studies evaluating the effect of 5-ASA as induction and maintenance treatment in Crohn's disease are controversial and the Cochrane database stated that 5-ASA was not effective in the treatment of Crohn's disease [45]. In this study, 5-ASA co-medication along with azathioprine was associated with improved clinical response. The efficacy of 5-ASA in combination with azathioprine may be due to the additive effect of 5-ASA. Previous studies revealed that thiopurine therapy in combination with 5-ASA led to increased 6-TGN levels in 82% to 100% of study patients, which suggests increased therapeutic efficacy [46,47]. Another study provided the evidence that 5-ASA are potent noncompetitive inhibitors of TPMT enzyme activity by strong affinity for TPMT, which results in increased 6-TGN accumulation and consequently a boost to the immunosuppressive effect [48].
This study had some limitations. First, it was a single-center, retrospective study and thus, carries certain limitations compared with a prospective design, such as lacking strictly controlled medical visit schedules and missing clinical or laboratory variables. Therefore, there is a need for additional prospective studies that follow patients on thiopurine treatment from the start of therapy and assess the efficacy and toxicity at several time points through monitoring of thiopurine metabolite levels. A previous prospective study on individualized dosing based on 6-TGN concentrations showed negative results; however, the statistical power was low owing to inadequate enrollment and early withdrawals [49]. A prospective study with adequate sample size and followup period to compare the clinical efficacy of dose adjustments based on thiopurine metabolite levels with standard dose adjustments based on white blood cell counts would also be informative.
In conclusion, our results show that patients with higher 6-TGN levels exhibit a significantly higher chance of being in remission. The optimal threshold level of 6-TGN for therapeutic response was 230 pmol/8\10 8 RBC. The clinical response rate was 88.3% among patients in our study who met this criterion. Combination therapy with 5-ASA was associated with higher 6-TGN levels and subsequent clinical responses. Patients with NUDT15 variant genotypes were more likely to develop leukopenia. Therefore, these results support the role of therapeutic drug monitoring in thiopurine maintenance treatment to optimize thiopurine therapy especially for non-responding CD patients, and predetermination of NUDT15 genotype is helpful minimize adverse events.
Supporting information S1