Hypoxanthine guanine phosphoribosyltransferase activity is related to 6-thioguanine nucleotide concentrations and thiopurine-induced leukopenia in the treatment of inflammatory bowel disease




Thiopurine drugs are widely used in the treatment of inflammatory bowel disease (IBD). The polymorphic enzyme thiopurine S-methyltransferase (TPMT) is of importance for thiopurine metabolism and adverse events occurrence. The role of other thiopurine-metabolizing enzymes is less well known. This study investigated the effects of TPMT and hypoxanthine guanine phosphoribosyltransferase (HPRT) activities on 6-thioguanine nucleotides (6-TGNs) concentrations and thiopurine-induced leukopenia in patients with IBD.


Clinical data and blood samples were collected from 120 IBD patients who were receiving azathioprine (AZA)/6-mercaptopurine (6-MP) therapy. Erythrocyte TPMT, HPRT activities and 6-TGNs concentrations were determined. HPRT activity and its correlation with TPMT activity, 6-TGNs level, and leukopenia were evaluated.


The HPRT activity of all patients ranged from 1.63–3.33 (2.31 ± 0.36) μmol/min per g Hb. HPRT activity was significantly higher in patients with leukopenia (27, 22.5%) than without (P < 0.001). A positive correlation between HPRT activity and 6-TGNs concentration was found in patients with leukopenia (r = 0.526, P = 0.005). Patients with HPRT activity > 2.70 μmol/min per g Hb could have an increased risk of developing leukopenia (odds ratio = 7.47, P < 0.001). No correlation was observed between TPMT activity and HPRT activity, 6-TGNs concentration, or leukopenia.


High levels of HPRT activity could be a predictor of leukopenia and unsafe 6-TGN concentrations in patients undergoing AZA/6-MP therapy. This could partly explain the therapeutic response or toxicity that could not be adequately explained by the polymorphisms of TPMT. (Inflamm Bowel Dis 2011;)

The thiopurine drugs azathioprine (AZA) and 6-mercaptopurine (6-MP) are well established in the treatment of inflammatory bowel disease (IBD) and have proven to be effective in both inducing and maintaining remission of Crohn's disease (CD) and ulcerative colitis (UC).1–3 AZA and 6-MP are the most commonly used immunomodulatory drugs in the treatment of IBD, with efficacy rates of 55%–70%.2, 4, 5 Unfortunately, the occurrence of side effects, such as bone marrow toxicity and hepatotoxicity, is a major problem in the use of these drugs. Long-term thiopurine therapy fails in approximately 50% patients who experience significant toxicity or inadequate response during treatment.6

After one oral dose, AZA is rapidly converted to 6-MP and an imidazole derivative, but up to 12% of the dose can be split to form the purine base hypoxanthine and thioimidazole.7 Three enzymes compete to metabolize 6-MP: thiopurine methyltransferase (TPMT), xanthine oxidase (XO), and hypoxanthine guanine phosphoribosyltransferase (HPRT). 6-MP activation, catalyzed by HPRT, initially forms the 6-MP nucleotides and eventually the active metabolites, the 6-thioguanine nucleotides (6-TGNs). The TGN metabolites act as purine antagonists and induce cytotoxicity and immunosuppression by inhibition of RNA, DNA, and protein synthesis. These cytotoxic properties are, at least partially, due to the direct incorporation of TGN into DNA.8 It has been suggested that the immunosuppressive effects of thiopurines are mediated by binding of TGN triphosphate instead of guanine triphosphate to the Rac1 protein. This binding could result in suppressed Rac1 activation and induction of apoptosis.9 Moreover, thiopurines have been shown to selectively inhibit inflammatory gene expression in activated T lymphocytes.10

It has been reported that patients who achieved very increased concentrations of the active 6-TGNs with standard thiopurine doses could have a high likelihood of profound myelosuppression.11–13 The observed interindividual differences in metabolite concentration, therapeutic response, or toxicity were partly explained by the variable formation of active metabolites and genetic polymorphisms of the crucial enzymes in thiopurine metabolism. The enzyme TPMT, which catalyzes the S-methylation of thiopurine drugs to inactive metabolites, can indirectly influence 6-TGN concentrations by shunting 6-MP metabolism away from 6-TGNs. Previous studies showed that TPMT deficiency was associated with severe, sometimes even fatal hematopoietic toxicity. Subjects with inherited TPMT deficiency presented higher levels of the 6-TGNs and had increased risk of myelotoxicity.14, 15 The gene encoding TPMT is located on chromosome 6 (6p22.3) and contains 10 exons. Allelic variants of the TPMT gene responsible for changes in the enzyme activity have been characterized. To date, two wildtype alleles (TPMT*1 and *1 S) and 20 mutant alleles (TPMT*2, *3 A, *3B, *3C, *3D, *4, *5, *6, *7, *8, *9, *10, *11, *12, *13, *14, *15, *16, *17, *18) related to TPMT deficiency have been described.16, 17 AZA dose selection based on pharmacogenetic testing of TPMT and metabolite monitoring (MM) may offer a safety and efficacy advantage over traditional dosing strategies, as TPMT testing is more beneficial for initial response to treatment and MM is more beneficial for sustained response to treatment.18 According to these outcomes, a pharmacogenetic-guided thiopurine therapy based on TPMT genetic polymorphism was established in the treatment of IBD.19 However, not all adverse events or metabolite patterns can be explained by genetic variations in TPMT. A study by Colombel et al20 showed that only 27% of patients with CD and myelosuppression had mutant alleles of the TPMT gene associated with enzyme deficiency, and myelosuppression was more often caused by other factors. In a prospective study, the authors could not confirm whether the choice of AZA/6-MP dose based on TPMT activity prevented myelotoxicity in patients with IBD. Some patients with wildtype TPMT alleles developed mercaptopurine-related adverse events, for reasons that are not fully understood.21

The interindividual variation in concentration of thiopurine active metabolite and therapeutic outcomes could be a result of polymorphisms in all thiopurine-metabolizing enzymes. The relevance of enzymes other than TPMT to the clinical effects of these drugs has not been extensively evaluated.

HPRT is a purine salvage enzyme that catalyzes the conversion of hypoxanthine and guanine into inosine monophosphate and guanosine monophosphate, respectively. In humans, HPRT deficiency is associated with two clinical disorders: Lesch-Nyhan (LN) syndrome (MIM# 300322) caused by a complete deficiency,22 and a milder disorder, Kelley-Seegmiller syndrome (MIM# 300323), characterized by severe gout and uric acid nephrolithiasis.23 The former disorder is associated with hyperuricemia, mental retardation, choreoathetosis, and compulsive self-mutilation. In thiopurines metabolism, HPRT catalyzes the conversion of 6-MP to its active TGNs metabolites. The individual variability of HPRT might be a potential reason for the different metabolite concentration and therapeutic outcomes in IBD patients undergoing thiopurine treatment.

To our knowledge, there are no data on the relation between HPRT activity, the production of 6-TGN, and incidence of leukopenia in IBD patients. Therefore, the aims of this study were to investigate 1) the HPRT activity in IBD patients undergoing stable AZA/6-MP therapy; 2) whether the HPRT activity is correlated with thiopurine metabolism and toxicity; and 3) the association of TPMT polymorphism with leukopenia.



All consecutive patients with a diagnosis of IBD who received AZA/6-MP treatment at the Gastroenterology Outpatient Clinic of the First Affiliated Hospital of Sun Yat-sen University were included in this study.

Diagnoses of CD and UC were according to the criteria of Lennard-Jones,24 based on clinical, endoscopic, histopathological, and radiological findings. Location of disease was made according to the criteria of the Montreal Classification.25

No patients had used any sulf-purine drugs including AZA, 6-thioguanine (6-TG), or 6-MP before. Inclusion criteria included steroid-dependent disease: unable to reduce corticosteroids below the equivalent of prednisolone 15 mg/day (or budesonide below 3 mg/day) within 3 months of starting corticosteroids or relapse within 3 months of stopping corticosteroids; frequent relapses: >3 relapses in 1 year or >2 relapses in 6 months; remission maintenance; and postoperative prophylaxis.

Exclusion criteria included blood transfusion or administration of cyclosporine or methotrexate (MTX) within the last 3 months; treatments potentially interfering with AZA metabolism, including allopurinol and diuretics; insufficient function of heart, liver, or kidney; active infection; and pregnancy.

Drug dose was started with 1 mg/kg daily for AZA (Imuran, GlaxoSmithKline, Sweden) or 0.5 mg/kg daily for 6-MP (Purinethol, GlaxoSmithKline) in the first week, and then increased to 2 mg/kg daily for AZA and 1 mg/kg daily for 6-MP without alteration in the following weeks.

Two mL venous blood samples (EDTA anticoagulation) were obtained prior to treatment for HPRT and TPMT detection. The same volume of venous blood was drawn at week 8 after administration of the stable dose or at the time of leukopenia occurrence for erythrocyte 6-TGNs concentration determination.

Clinical data including sex, age, age at diagnosis, and site of disease, type of IBD, weight, dose of AZA/6-MP, indication for AZA/6MP therapy, concomitant therapy (5-aminosalicylates, infliximab, or other drugs), and toxicity data including full blood counts and liver function tests were recorded. Patient information was collected by a clinician.

Control visits were performed every 2 weeks for the first month, every month for the following 2 months, and then every 3 months. During these control visits patients had complete blood count measurements and liver function tests, and were clinically reviewed while adverse effects were recorded. Hematotoxicity was observed as leukopenia. This was defined as a leukocyte count (WBC) less than 3.5 × 109/L. Each decrease of WBC should be continuously observed in 2 days and recovered in the next 1 or 2 weeks after AZA/6-MP withdrawal. Hepatotoxicity was defined as an increase in transaminase at least two times higher than the normal value. Pancreatitis was diagnosed when compatible symptoms (abdominal pain) were present and serum amylase was increased two times above the upper normal limit. Flu-like symptoms included febris, headache, courbature, and arthralgia all over the body while gastrointestinal intolerance was defined as hypogeusia, nausea, and vomiting.

Equipment and Materials

The high-performance liquid chromatography (HPLC) system was composed of a Waters (Milford, MA) M510 Pump, a Waters 717 Autosampler, a Waters 486 UV Detector, and a Waters M32 Chromatography Workstation.

Mini NucleoSpin blood kit was purchased from MN (Germany). Taq DNA polymerases, Mwo I, and Acc I were all purchased from Takara Biotechnology (Dalian, China). Methanol of HPLC grade was purchased from Fisherbrand (Thermo Fisher, UK). Tris was purchased from Bio Basic (Ontario, Canada). Allopurinol, S-adenosyl-l-methionine (SAM), 6-MP, 6-methylmercaptopurine (6-MMP), 6-TG, 5-phosphoribosyl-1-pyrophosphate (PRPP), hypoxanthine and alkaline phosphates (ALP) from bovine calf intestine were all purchased from Sigma (St. Louis, MO). Magnesium chloride (MgCL2), triethylamine, and perchloric acid were all purchased from Tedia (Guangzhou, China). All other reagents were of analytical grade. Human plasma was obtained from healthy volunteers who did not take any medications for 1 month at Guangzhou Blood Center (Guangzhou, China). Ultrapure water was obtained from a Milli-Q Plus water purification system (Millipore, Bedford, MA).

Enzyme and Metabolite Assays

Venous blood samples were collected in EDTA anticoagulation tubes. Briefly, 2 mL of whole blood were centrifuged at 800 g for 10 minutes at 4°C to isolate red cells. After washing the pellet twice with 2 mL of normal saline and centrifuged at 800 g for 10 minutes, cells were gently resuspended in 2 mL of normal saline and the hematocrit was determined. After this step, the red cells were lysed with cold distilled water (4 mL for 1 mL of solution) and then centrifuged at 13,000g for 10 minutes at 4°C. The supernatant was kept at −80°C until analysis.

HPRT activity was determined according to a previously reported HPLC method with minor modifications. The method could detect 0.3% of normal HPRT activity by using erythrocyte lysates.26 The HPRT-mediated catalytic reaction occurred under neutral buffered condition, taking hypoxanthine as substrate and PRPP as phosphate radical donor. Chromatographic separation was achieved on a C-18 column (Hypersil BDS C-18, I.D. 4.6 × 150 mm, 5 μm, Elite HPLC, China) at room temperature. The mobile phase consisting of water–methanol–triethylamine (95.9:4:0.1, v/v/v) was adjusted to pH 3.0 with phosphoric acid and pumped at a flowrate of 1.0 mL/min. The detection wavelength was 248 nm. HPRT activity was expressed by the formation of inosine in the enzymatic reaction. The unit of HPRT activity was μmol/min per g Hb.

The erythrocytic TPMT activity was determined by an HPLC assay as described previously.27 One unit of TPMT activity represented the formation of 1 nmol of 6-MMP from 6-MP per hour of incubation. TPMT activity was normalized per milliliter of packed erythrocytes (pRBC).

6-TGNs concentration in erythrocyte lysates was determined by using a previously described HPLC method28 which hydrolyzed (100°C for 45 minutes) 6-TGNs in the separated supernatant to release the 6-TG by perchloric acid. The chromatographic separation of the free 6-TG after the sample preparation procedure was consistent with the analysis of HPRT activity. The detection wavelength was set at 345 nm for 6-TG detection. The concentration of 6-TGNs was normalized to pmol/8 × 108 RBC.

DNA Extraction and Genotyping

DNA was isolated using the Mini NucleoSpin blood kit. Allele-specific polymerase chain reaction (PCR) and PCR-restriction fragment length polymorphism (RFLP) were used to determine the frequency of TPMT mutant alleles (TPMT*2, TPMT*3A, TPMT*3B, and TPMT*3C) in IBD patients, as described previously.27


Descriptive statistics was calculated using SPSS for Windows 16.0 package (SPSS, Chicago, IL). A one-sample Kolmogorov–Smirnov test was used to evaluate the normal distribution of enzyme activities and metabolite concentrations. Quantitative variables were expressed as median and range, or as mean ± standard deviation when normally distributed. The chi-square test was used to evaluate differences in percentages, odds ratios (OR) with 95% confidence intervals (CI). A parametric Student's t-test or nonparametric Mann–Whitney U-test was adopted to evaluate the differences between two independent groups. Analysis of covariance (ANCOVA) was used when factors such as age, gender, or combination of AZA were not matched between two groups. A bivariate correlate was used to analyze the correlation between enzyme activities and metabolite concentrations. Receiver operating characteristic (ROC) curves were obtained to plot the sensitivity and specificity for various HPRT activities to predict the development of leukopenia. P less than 0.05 was considered statistically significant.

Ethical Considerations

The study was approved by the local Ethics Committee of Sun Yat-Sen University. Informed consent were obtained from the patients before inclusion. It was also registered on the International Standard Randomized Controlled Trial Number Register (ISRCTN) with a trial number of ISRCTN58287360.


Patient Characteristics

A total of 120 patients with IBD were included in this study; 78 were men; 116 of them were treated with AZA while four of them were treated with 6-MP. Thirty-one (25.8%) had CD and 89 (74.2%) had UC. The mean age was 32 ± 13 years. The selected baseline demographics and disease characteristics of the patients are shown in Table 1. All patients completed the 1-year follow-up visit after taking AZA/6-MP.

Table 1. Baseline Characteristics of Patients at Time of Inclusion
CharacteristicNo. of Patients (N = 120)%
Age (years)  
 Ranger3 ∼ 74 

Adverse effects were observed in 43 patients (35.8%) treated with AZA/6-MP. The most frequent was leukopenia (27, 22.5%). Other adverse effects included gastrointestinal intolerance (12, 10%), flu-like symptoms (5, 4.2%), alopecia (6, 5%), and hepatotoxicity (1, 0.8%). No pancreatitis observed. The frequency of each particular adverse effect and corresponding medical decision are summarized in Table 2.

Table 2. Summary of AEs and Corresponding Medical Decision in AZA/6-MP Therapy
Adverse EffectsNo. of Patients (N=120)Medical Decision
Drug WithdrawalOther Treatment
All43 (35.8%)35 (81.4%) 
Leukopenia27 (22.5%)27 (100%) 
Hepatotoxicity1 (0.8%)0Hepatinica
Flu-like symptoms5 (4.2%)4 (80%) 
Gastrointestinal intolerance12 (10.0%)6 (50%)Digestant
Alopecia6 (5%)2 (33.3%) 

6-TGN Concentration and Leukopenia

The 6-TGN concentrations of 27 patients who developed leukopenia were determined at the time of hematological toxicity. Other samples were collected at week 8 after administration of the stable dose: 2 mg/kg daily for AZA and 1.0 mg/kg daily for 6-MP.

In all, 77 patients reached the stable dose (2 mg/kg daily for AZA and 1.0 mg/kg daily for 6-MP) and did not experience any adverse effects during the entire 1-year follow-up visit. The median 6-TGN concentrations of these patients was 269.48 (77.55–782.66) pmol/8 × 108 RBC. The patients who suffered leukopenia had a median 6-TGN concentration of 397.34 (225.72–1913.53) pmol/8 × 108 RBC. There was a significant difference between these two groups (P < 0.001) (Table 3).

Table 3. Correlation of TPMT, HPRT Activity, and 6-TGNs Concentration with Adverse Effects
Activity (U/mL RBC)aActivity (μmol/min per g Hb)aConcentration (pmol/ 8×108 RBC)b
  • a

    Mean ± SD.

  • b

    Median (range).

  • c

    Compared with Without AEs.

Total12012.87 ± 4.16 2.31 ± 0.36 296.95 (77.55 ∼ 1913.53) 
Adverse effects (AEs)4312.0 ± 5.00.1132.45 ± 0.410.004369.85 (77.55 ∼ 1913.53)<0.001
 Leukopenia2711.87 ± 5.360.3252.56 ± 0.40<0.001397.34 (225.72 ∼ 1913.53)<0.001
 Hepatotoxicity118.04 1.97 322.50 
 Flu-like symptoms512.53 ± 2.790.8182.07 ± 0.120.233288.44 (239.98 ∼ 782.66)0.363
 Gastrointestinal intolerance129.57 ± 3.860.0062.31 ± 0.410.216345.35 (181.25 ∼ 1913.53)0.053
 Alopecia612.62 ± 2.970.8522.65 ± 0.460.081394.16 (77.55 ∼ 766.07)0.079
 Without AEs7713.30 ± 3.81 2.24 ± 0.31 269.48 (77.55 ∼ 782.66) 

TPMT Genotype and Phenotype

Four alleles of the TPMT gene, TPMT*2, TPMT*3A, TPMT*3B, and TPMT*3C, were evaluated in 120 IBD patients. Only one subject carrying a mutant TPMT allele was identified, while TPMT*2, TPMT*3A, and TPMT*3B alleles were not detected. Thus, the total frequency of mutant alleles in these patients was 0.8%. The patient who suffered leukopenia during the treatment was heterozygous for TPMT*3C and her TPMT activity was 3.4 U/mL RBC.

A 7.4-fold interindividual variation was observed in the TPMT activity ranging from 3.4 to 25.02 (12.87 ± 4.16) U/mL RBC. The distribution of TPMT activity in IBD patients was normal-skewed (Z = 0.602, P = 0.862) and no deficient individual was found in this study. TPMT activities in different characteristics are shown in Table 4.

Table 4. TPMT, HPRT Activity, and 6-TGNs Concentrations in Different Characteristics
Activity (μmol/min per g Hb)aActivity (U/mL RBC)aConcentration (pmol/ 8×108 RBC)b
  • a

    Mean ± SD.

  • b

    Median (range).

All1202.31 ± 0.36 12.87 ± 4.16 296.95 (77.55 ∼ 1913.53) 
 Male782.30 ± 0.330.50713.07 ± 4.250.488291.82 (108.44 ∼ 782.66)0.813
 Female422.35 ± 0.4012.51 ± 4.01303.30 (77.55 ∼ 1913.53)
 CD892.31 ± 0.350.86512.84 ± 4.120.885327.44 ± 158.62 (77.55 ∼ 976.47)0.430
 UC312.32 ± 0.36 12.97 ± 4.32 372.38 ± 305.57 (108.44 ∼ 1913.53) 
 With 5-aminosalicylates282.41 ± 0.410.11612.11 ± 4.570.275341.83 (113.16 ∼ 1913.53)0.014
 Without 5-aminosalicylates922.29 ± 0.3313.10 ± 4.01280.37 (77.55 ∼ 782.66)
 With infliximab112.25 ± 0.400.58412.12 ± 4.140.535269.25 (162.60 ∼ 475.76)0.989
 Without infliximab1092.32 ± 0.3512.95 ± 4.18297.35 (77.55 ∼ 1913.53)

There was no significant difference in the mean TPMT activity between patients with leukopenia and without (11.87 ± 5.36 versus 13.3 ± 3.81 U/mL RBC, P = 0.119). No relationship between TPMT activity and 6-TGN concentration (r = −0.043, P = 0.051) was found in all 120 IBD patients. However, approximately 48% of the variance in 6-TGN concentration could be explained by TPMT activity in patients with leukopenia (r = −0.481, P = 0.011).

HPRT Activity and Its Correlation with TPMT, 6-TGN Concentration, and Leukopenia

HPRT activity of 120 IBD patients ranged from 1.63–3.33 (2.31 ± 0.36) μmol/min per g Hb. The distribution of HPRT activity in IBD patients was normal-skewed (Z = 0.682, P = 0.741). No deficient individual was found in this study (Fig. 1). No sex difference in HPRT activity was observed in the present study, and the HPRT activity showed no difference between the UC patients and CD patients (Table 4).

Figure 1.

The distribution of the HPRT activity in 120 IBD patients.

There was no correlation between HPRT activity and TPMT activity either in all 120 IBD patients (r = −0.207, P = 0.023) or in patients with leukopenia (r = −0.365, P = 0.061). A significant relationship between HPRT activity and 6-TGN concentration (r = 0.211, P = 0.021) was found in all 120 IBD patients (Fig. 2A). Moreover, there was a positive correlation between HPRT activity and 6-TGN concentration in patients with leukopenia (r = 0.526, P = 0.005) (Fig. 2B).

Figure 2.

(A) Correlation of HPRT activity and 6-TGN concentration in all patients (n = 120, r = 0.211, P = 0021). (B) Correlation of HPRT activity and 6-TGN concentration in patients with leukopenia (n = 27, r = 0.526, P = 0.005).

A significant difference of HPRT activity was observed among patients with or without leukopenia (2.56 ± 0.40 versus 2.24 ± 0.31 μmol/min per g Hb, P < 0.001). The mean HPRT activity was also higher in patients with overall adverse effects than those without adverse effects (2.45 ± 0.41 versus 2.24 ± 0.31 μmol/min per g Hb, P = 0.004) (Table 3).

The area under the ROC curve for highest HPRT activity was 0.737 (95% CI 0.623–0.850, P < 0.001). A specificity of 100% was seen at the cutoff level of 3.05 μmol/min per g Hb but at the expense of a sensitivity of 11.1%. Patients with an HPRT activity exceeding 2.70 μmol/min per g Hb could have an increased risk of developing leukopenia (OR = 7.47, 95% CI 2.68–20.79, P < 0.001). This level could be used for predicting leukopenia to AZA/6-MP with a specificity of 90.3% and a sensitivity of 44%.

Effects of Coadministration on TPMT, HPRT Activity, and 6-TGN Concentration

The combination of other drugs did not show any influence on TPMT and HPRT activity (Table 4). But the patients who were coadministered 5-ASA had significantly higher concentrations of 6-TGNs (341.83 versus 280.37 pmol/8 × 108 RBC, P = 0.014). There were 28 patients coadministered 5-ASA and 92 patients without. 5-ASA was not matched between these two groups by chi-square test. After controlling for the presence of 5-ASA and converted 6-TGN concentrations into Gaussian distribution with the equation Y = LOG10(X), two-way analysis of variance (ANOVA) showed that the median 6-TGN level was still significantly higher in patients with leukopenia than those without leukopenia (F = 12.563, P = 0.001). Moreover, there was also a significant difference in 6-TGN concentration between patients with overall adverse effects than those without adverse effects (F = 10.417, P = 0.002) when controlled for the presence of 5-ASA.


In this study the HPRT and TPMT activities were related to thiopurine active metabolite concentrations and clinical characteristics in order to elucidate the significance of these enzymes in thiopurine metabolism.

The thiopurine drugs are used widely to treat malignancies, rheumatic diseases, dermatologic conditions, IBD, and solid organ transplant rejection. As these drugs have a relatively narrow therapeutic index, it is important to recognize the risk factors that may lead to toxicity, especially leukopenia.

Previous studies have demonstrated that TPMT-deficient patients were at high risk for severe adverse effects. In IBD patients, this may occurred as early as the initial weeks of instituting the therapy.29, 30 TPMT testing is now recommended for patients with clinical or laboratory evidence of severe bone marrow toxicity, particularly myelosuppression.31 But the distribution of TPMT mutant alleles differs significantly among ethnic populations. TPMT*3 A is the most common mutant allele in Caucasian populations, followed by TPMT*2 and TPMT*3C accounting for the vast majority (>95%) of mutant alleles.32, 33 However, in Asian and African populations TPMT*3C is the most frequent mutant allele.34, 35 In this study, only one patient with TPMT*3C was found. Although this patient developed leukopenia during the treatment, the other 26 patients who suffered leukopenia all had the wildtype TPMT genotype. This indicated that the TPMT genotype might only be useful for predicting leukopenia in patients with TPMT homozygote and heterozygote. The significance of predicting myelosuppression by testing TPMT genotype in Chinese IBD patients was limited, as the frequency of TPMT*3C was rare.

In this study, no correlation was observed among TPMT activity and HPRT activity, 6-TGNs concentration, and leukopenia. These findings are supported by some20, 21 but not all36, 37 studies. The absence of correlations might have been effected by the coadministration of 5-ASA in 28 patients, as 13 (46%) of them developed leukopenia. However, neither TPMT activity nor HPRT activity differed between patients with and without concomitant 5-ASA. Furthermore, our study suggests that TPMT activity explains only approximately 48% of the variation in 6-TGN concentrations of the patients who developed leukopenia. Examination of further factors influencing thiopurine metabolism is still demanded.

Previous studies also documented that the influence of metabolism enzymes other than TPMT could affect the therapeutic response and toxicity of thiopurine drugs. A case report on XO showed a negative correlation between XO activity and the concentration of 6-TGNs.38 There are also many studies about the correlation between inosine triphosphate pyrophosphatase (ITPA), inosine-5-monophosphate dehydrogenase (IMPDH), and the efficacy or toxicity of AZA treatment.39, 40 However, data on HPRT activity in IBD patients taking thiopurine drugs are still deficient.

In clinical practice, RBC is routinely used for TPMT and metabolite measurements in the management of patients on thiopurine therapy. The use of RBC is based on the fact that TPMT activity in RBC reflects that in lymphocytes and other tissues such as kidney, hepatic tissues, and leukemic blasts.41–44 Therefore, these RBC assays were also performed in our study. As the 6-TGN concentration was also detected in RBC, we therefore used RBC as the compartment for HPRT measurements.

Distribution of HPRT activity in IBD patients of this study was similar to that of the healthy Chinese population in Taiwan (2.22 ± 0.34 μmol/min per g Hb)45 and a little higher than that of the healthy Japanese population (2.01 ± 0.19 μmol/min per g Hb).46

We only detected HPRT activity at baseline and all the participants were at an active stage of disease. It was impossible to evaluate the correlation between disease activity and HPRT activity, as all the patients were at the same stage. In the studies of Peters and Veerkamp47 and Lennard,48 HPRT activity increased during long-term thiopurine treatment. However, a recent study by van Asseldonk et al49 showed that HPRT activity in erythrocytes decreased following the initiation of 6-TG therapy. There was discordance among these studies. We believe that the influence of disease activity on HPRT still needs to be elucidated.

The gene encoding HPRT (HPRT1) is located on the long arm of the X-chromosome (Xq26.1). The genetic mutational spectrum study on HPRT found that mutations distribute over the whole gene. To date, more than 302 mutations have been reported for HPRT with different clinical manifestations. These mutations include deletions, insertions, duplications, abnormal splicing, and point mutation at different sites on the coding region.45 All of these mutations on the HPRT gene could influence enzyme activity.46 This might be the genetic basis of the individual difference in HPRT activity. A previous study on HPRT1 risk genotypes showed that the frequency of C_11680155_10, C_11680164_10, C_27862676_10, and C_8940525_10 (rs1468266) variations was not significantly different in patients with adverse effects (included leukopenia) in IBD patients. This could be explained by the low frequency (less than 1% in the patients) of the investigated polymorphisms.50

There were four patients who had two kinds of adverse effects simultaneously. Two of them had leukopenia accompanied by gastrointestinal intolerance. One had alopecia accompanied by leukopenia. One had alopecia accompanied by gastrointestinal intolerance. Two patients had three kinds of adverse effects simultaneously. One of them had leukopenia, flu-like symptoms, accompanied by gastrointestinal intolerance, while one of them had leukopenia, alopecia, accompanied by gastrointestinal intolerance. There was an increase of HPRT activity in patients who experienced more than one kind of adverse effect (2.59 ± 0.49 μmol/min per g Hb) than those who experienced only one (2.43 ± 0.39 μmol/min per g Hb), although these differences did not reach statistical significance (P = 0.363).

In the present study, a positive correlation between HPRT activity and 6-TGN concentration was found in patients with leukopenia (r = 0.526, P = 0.005). This indicated that patients with high HPRT activity could have high 6-TGN levels and a higher potential for adverse effects. The significant difference in HPRT activity between patients with and without leukopenia confirmed this hypothesis. A high level of HPRT activity could be a predictor of adverse effects and high TGN concentrations above a safe threshold in IBD patients undergoing AZA/6-MP therapy. By using a stipulated cutoff value of 2.70 μmol/min per g Hb, we found that patients with HPRT activity above this cutoff level had an increased risk of leukopenia (OR = 7.47, 95% CI 2.68–20.79, P < 0.001)). Furthermore, the ROC analysis showed that HPRT activity above this cutoff level has a specificity of 90.3% and a sensitivity of 44%. This observation indicated that detection of HPRT activity before thiopurine administration might protect IBD patients from AZA/6-MP-induced toxicity and help dose individualization.

The HPRT activity of the patient who developed hepatotoxicity was lower than the mean value of other IBD patients. This finding suggested that the hepatotoxicity of AZA was not directly related to HPRT activity, but it was impossible to perform a statistical analysis to confirm this hypothesis, as there was only one case.

6-TGNs are known to possess cytotoxic properties, and their accumulation may contribute to a more persistent neutropenia, increasing the likelihood of myelosuppression.51, 52 In the present study the venous blood used to detect the 6-TGN concentrations in patients without leukopenia was drawn at week 8 after administration of the stable dose based on the result of a pharmacokinetics study. It has been determined that erythrocyte 6-TGNs reached steady state after 8 weeks on stable doses in patients with IBD.53 The 6-TGNs levels in patients with leukopenia were significantly higher than that in the patients without leukopenia in this study, even controlled for the presence of 5-ASA. This finding was consistent with other reports.54, 55 But the median value of 6-TGN concentration in patients who developed leukopenia (397.34 pmol/8 × 108 RBC) was not identical to those reports. This is likely explained by differences in assay method.

In our study we detected 6-TGNs concentration with the method published by Dervieux and Boulieu28 while some studies56, 57 were performed with the method published by Lennard.58 A research by Shipkova et al59 compared the difference between these two widely used methods. It showed that the sample preparation procedure for the Lennard assay was relatively laborious, time-consuming, and uses the neurotoxic reagent phenylmercuric acetate (PMA) for the extraction. Direct comparison of both methods showed that 6-TGN concentrations were, on average, 2.6-fold higher in the Dervieux–Boulieu method over the concentration range tested, although the correlation (r = 0.99; P < 0.001) was good. Replacement of sulfuric acid by perchloric acid reduced this difference to 1.4-fold (r = 0.99; P < 0.001). The difference between 6-TGN concentrations measured by the two methods was attributable, at least in part, to differences in the extent of nucleotide hydrolysis. The hydrolysis time used in the Lennard method was not sufficient to achieve complete hydrolysis. Therefore, we chose the Dervieux–Boulieu method in our study. A commentary on this article by both Lennard and Shipkova demonstrated that quantification of TGN could differ among laboratories and method-specific therapeutic ranges were required for the interpretation of the results generated.60

The frequency of leukopenia (27, 22.5%) was higher than previously reported in the Western population in this study.61–63 The could be explained by the different definition of myelotoxicity (WBC lower than 3.5 × 109/L) and the patients were followed up for a longer time (1 year). We found that although more than two-thirds of the adverse effects developed early in the 12 weeks after AZA/6-MP initiation, myelotoxicity could develop at any time during the treatment. The higher percentage of myelotoxicity could also be a result of racial differences and indicates that the pharmacodynamics of thiopurines might be different between Chinese and other populations.

In our study, exposure to 5-ASA medications was associated with an increase risk of leukopenia. This was consistent with other reports.64, 65 But the precise mechanism of the interaction between 5-ASA and AZA/6-MP is unclear. In a study by de Boer et al66 in 26 IBD patients, 5-ASA could raise the concentration of 6-TGNs in patient erythrocytes. Another retrospective study done on 126 IBD patients reported the same conclusion.67 The interaction leads to higher 6-TGN levels, probably due to diminished TPMT activity. Previous studies have established that 5-ASA medications and their metabolites inhibit TPMT activity in vitro.68–71 However, two prior trials in vivo showed a trend toward increased TPMT activity with exposure to a 5-ASA medication, although the results did not show statistical significance.64, 72 In Hande et al's67 study, TPMT inhibition might not explain this effect of 5-ASA, as the exposure did not affect 6-MMP levels. But a recent study by de Graaf et al73 showed that individual 6-TGN metabolites were increased after addition of 5-ASA, while 6-MMP-levels and the 6-MMP/6-TGN ratios were decreased. Apparently, 5-ASA did not raise the TPMT and HPRT activity in the present study. Nonetheless, the fact that we did not observe the change of TPMT activity could be due to the inhibition by 5-ASA being only temporary and was washed away in the process of TPMT activity determination. The mechanism of 5-ASA's effect on thiopurines therapy should be studied further.

A clinical follow-up visit showed that 27 patients had leukopenia. After drug withdrawal and given symptomatic treatment, all 27 patients recovered with normal blood cell count in the next 1 or 2 weeks. Seventeen patients with myelotoxicity were reintroduced to AZA (≤1 mg/kg daily, n = 9) or 6-MP (≤0.5 mg/kg daily, n = 8) and this rechallenge was successful in 16 patients as the dosage was restricted. One patient with myelotoxicity was introduced to MTX, five were introduced to 5-ASA, and four were shifted to infliximab. One patient showed hepatotoxicity and by giving hepatinica without drug withdrawal, his transaminase level recovered in 1 month.

In conclusion, the results of this study showed that HPRT activity was related to thiopurine drugs metabolism and leukopenia in the treatment of IBD. This could partly explain the therapeutic response or toxicity which could not be adequately explained by the polymorphisms of TPMT after thiopurine administration. Consideration of individual HPRT activity in thiopurine therapy may help identify patients at risk of associated hematological toxicity and can serve as a guide for dose individualization.