Association of TPMT and NUDT15 gene polymorphisms with azathioprine-induced leukopenia: A case–control study in Eastern India

Abstract

BACKGROUND:

Azathioprine (AZA) is a widely used immunosuppressant drug. Leukopenia is a serious adverse effect of the drug which often necessitates dose reduction or drug withdrawal. Predictors of leukopenia include genetic and nongenetic factors. Genetic polymorphism of AZA-metabolizing enzyme, thiopurine S-methyltransferase (TPMT) is well established. There is inconclusive evidence about the role of Nudix hydrolase (NUDT15) gene polymorphism. This case–control study assessed the association of genetic polymorphisms of NUDT15 and TPMT with leukopenia induced by AZA.

MATERIALS AND METHODS:

Cases were patients on AZA who developed leukopenia (white blood cell count <4000/μl) within 1 year of treatment initiation that necessitated dose reduction or drug withdrawal. Age and gender-matched patients without leukopenia within 1 year of treatment with AZA served as controls. TPMT (3 loci: c238G to C, c460G to A, c719A to G) and NUDT15 (c 415C to T, rs116855232) genotyping were done using TPMT strip assay and polymerase chain reaction–restriction fragment length polymorphism, respectively. Genotype frequencies were noted, and the odds ratio was calculated to determine the association between genotypes and leukopenia.

RESULTS:

Twenty-nine subjects (15 cases and 14 controls) were enrolled. Statistically significant differences were not observed in the TPMT genotype (*1/*1 and *1/*3C) (P = 0.23) between cases and controls. NUDT15 genotypes (*1/*1 and *1/*3) (P = 0.65) also showed no statistically significant difference between cases and controls.

CONCLUSION:

The above genotypes do not appear to be associated with AZA-induced leukopenia in an eastern Indian population.

Introduction

Azathioprine (AZA) is used for treating autoimmune diseases, lymphomas, and inflammatory bowel disease. It is also used in patients after organ transplantation.[1] AZA is a prodrug and gets metabolized to 6-mercaptopurine (6-MP) by the hepatic enzyme glutathione-S-transferase. Thiopurine-S-methyltransferase (TPMT) catalyzes the conversion of 6-MP to 6-methyl MP. Variability in AZA response may be attributed to several factors notable among which is polymorphism of genes encoding for its metabolism. Four mutant alleles (*2, *3A, *3B, and *3C) of the TPMT gene are known to account for 80%–95% of decreased enzyme activity. The risk of developing severe leukopenia is high in individuals having mutant alleles when treated with standard doses of thiopurines. Therefore, genotyping for the major TPMT variant alleles may be a valuable tool in optimizing AZA therapy. However, some patients with normal TPMT activity still demonstrate substantial toxicity, suggesting that additional factors, possibly other genetic variants, might contribute to inter-patient variability in thiopurine metabolism. In Asian populations, the TPMT mutant variants are relatively less common compared to Caucasians and thus may have limited significance in guiding thiopurine dose regimens in such populations.[2] Therefore, the role of other enzymes like the Nudix hydrolase family was investigated. The NUDT15 gene is located on chromosome 13 and encodes for the Nudix hydrolase enzyme superfamily which catalyzes the hydrolysis of nucleoside diphosphates which results in oxidative damage and induces base mispairing during DNA replication.[3,4,5] Mutations of the NUDT15 gene result in poor metabolism of thiopurines and may increase the possibility of AZA-induced leukopenia. The incidence of NUDT15 genetic polymorphism is higher in Asian populations than in Caucasians, thus justifying that NUDT15 genotyping may be useful for predicting thiopurine toxicity in our populations.[4]

The role of NUDT15 and TPMT genetic polymorphisms with AZA is not well studied in the eastern Indian population. Hence, our study objectives were to evaluate whether there is any association between TPMT (238G>C, 460G>A, and 719 A>G) and NUDT15 (415 C>T) single-nucleotide polymorphisms with AZA-induced leukopenia.

Materials and Methods

This study was initiated after clearance from the Institutional Ethics Committee. It was registered with the Clinical Trials Registry of India (CTRI/2019/07/020354). Before enrollment into the study, written informed consent was taken from all study participants. The study participants included adult patients with an autoimmune disease treated with AZA as an immunosuppressant for induction or maintenance therapy. Such patients were screened and recruited from the Rheumatology, Medicine, and Nephrology OPD of our hospital. Genotyping and other study-related activities were done in the Pharmacology Department of the institute.

The inclusion criteria for “Cases” were adult patients (18–60 years) of all genders with a diagnosis of any autoimmune disease on AZA and had leukopenia (total white blood cell [WBC] count <4000/μl) within 1 year of treatment initiation with AZA that necessitated a reduction of dose or drug withdrawal, while “Controls” were patients who were age and gender matched, taking a comparable dose of the drug for the same indications as cases but without any incidence of leukopenia up to 1 year of treatment. All subjects had to provide written informed consent before enrollment.

Exclusion criteria were (a) subjects who were on other immunosuppressants (other than corticosteroid) or on drugs known to cause leukopenia and (b) subjects who had developed leukopenia after 1 year of AZA initiation.

A formal sample size calculation could not be done due to financial constraints and logistic issues, so we proposed to include 15 cases and 15 controls during the study period.

Reagents

All reagents were from Sigma-Aldrich (St. Louis, MO, USA) except QIAamp DNA Blood Mini Kit (Qiagen, Hilden, Germany), AmpliTaq Gold 360 Master Mix, 360 GC Enhancer (Applied Biosystems, Foster, CA, USA), FastDigest Taal restriction enzyme, FastDigest Green Buffer, nuclease-free water (Thermo Fisher Scientific, Waltham, MA, USA), and PGX-TPMT StripAssay™ (ViennaLab Diagnostics GmbH, Vienna, Austria).[6]

Extraction of DNA

Extraction of genomic DNA was done from peripheral blood using QIAamp DNA Blood Mini Kit as per the manufacturer’s instructions. DNA purity was estimated by a nanodrop spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) based on the ratio of the absorbance at 260 and 280 nm (A_260/280). A ratio of 1.8 was considered for purity assessment of the DNA.

TPMT Strip Assay for TPMT (238G>C, 460G>A, and 719 A>G) polymorphism

TPMT Strip Assay was performed according to the manufacturer’s instructions (PGX-TPMT StripAssay™ Kit, ViennaLab Diagnostics GmbH, Vienna, Austria). The assay is designed to evaluate 3 polymorphic loci in the TPMT gene: 238G>C, 460G>A, and 719 A>G.[7] Using this strip assay, we could detect *1, *2, *3A, *3B, and *3C TPMT alleles and the resulting homozygous or heterozygous genotypes (e.g., *1*1, *1*2, and *1*3A) based on the mutations stated before. The assay has two steps after DNA isolation: polymerase chain reaction (PCR) amplification using biotinylated primers followed by hybridization of amplification products to a test strip that contains allele-specific oligonucleotide probes. The genotyping results were based on the color development in the test strips.[6,7]

PCR and restriction fragment length polymorphism protocol for NUDT15 (415 C>T) polymorphism.[8]

The PCR for NUDT15: For each 25 μl reaction:

AmpliTaq Gold 360 (Applied Biosystems, Foster, CA, USA) Master Mix=12.5 μl; GC Enhancer=1.0 µl; each primer PCP-0023 (5’-CCCAAATAAACACCCTTTGTTTTCTGT-3’) and PCP-0024 (5’-CTTTGTATCCCACCAGAT GGTTC-3’)=1.0 µM; Genomic DNA=20 ng; PCR-grade water=7.5 μl.

Step-down PCR protocol for amplification was done using the Veriti 96-Well Thermal Cycler (Applied Biosystems, Foster, CA, USA) as follows:

• Denaturation (initial) – 95°C for 10 min

• Denaturation (subsequent) – 95°C for 30s

• Annealing – 30 s with 1°C decrease per cycle, from 66°C to 56°C for 10 cycles

• Then 30 cycles of annealing at 60°C for 30s, extension at 72°C for 1 min, and final extension at 72°C for 10 min.

The PCR products were electrophoresed in a 2% agarose gel and analyzed in G-BOX gel doc (Syngene, Cambridge, United Kingdom) using Gene Tools (version 4.01.04) software.

Each NUDT15 c.415C>T genotyping restriction digestion reaction contained 1.7 μl unpurified PCR product, 0.2 μl FastDigest Taal restriction enzyme, 0.3 μl 10X FastDigest Green Buffer, and nuclease-free water added up to 5 μl. The digestion mix was incubated at 65°C (for TaaI digestion) for 15 min on a thermal block (Applied Biosystems, Foster, CA, USA) and separated alongside 2.0 μl of undigested PCR product using a 2% agarose gel in TBE(1X) buffer by electrophoresis at 100 V for 45 min.

Statistical analysis

The demography, disease, treatment profile, and genotype frequencies of the study participants were summarized by appropriate statistical measures, i.e., frequency and counts for categorical variables, mean and standard deviation for numerical data with normal distribution, and median with interquartile range for numeric data with nonnormal distribution. For the association of the different genotypes with the outcome, the odds ratio (OR) with its 95% confidence interval was calculated. Values of P < 0.05 were considered for statistical significance. For the computation of OR, if any cell value was 0, we used the Haldane-Anscombe correction.[9]

Results

During the study period spanning from April 2019 to June 2020, 29 subjects were enrolled. The baseline characteristics of the study population are depicted in Table 1. The cases were age and gender matched with the controls. There was a female preponderance in both groups. All subjects were unrelated, and they stated to be residents of West Bengal for at least the last two generations. As per selection criteria, the cases had leukopenia while the controls had normal leukocyte counts. The indication for AZA was similar, and the dose of the drug was also comparable.

Table 1.

Characteristics of study population and azathioprine intake details at enrollment

Characteristics	Case (n=15)	Control (n=14)	P
Age (years) mean±SD	32.47±11.22	26.64±7.49	0.11
Gender (%)
Male	13.33	7.14	0.99
Female	86.67	92.86
Indication for AZA (n)
SLE	10	11	0.57
MCTD	3	3
ANCA vasculitis	1
Scleroderma	1
AZA dose (mg)
Minimum	50	50	NS
Maximum	100	100
Leukocyte count (cells/mm)
Minimum	1200	4100	<0.0001
Maximum	2800	8900

SLE=Systemic lupus erythematosus, MCTD=Mixed connective tissue disorder, ANCA=Antineutrophil cytoplasmic antibody, NS=Not significant, SD=Standard deviation, AZA=Azathioprine

Thiopurine methyltransferase genotype frequencies

Two genotypes, i.e., wild type *1/*1 (c. 474T, rs2842934) and the heterozygous mutant *1/*3C genotype (c.719A>G, rs1142345), were observed in the cases, whereas only wild-type genotype was detected in the control population. There were no subjects with *2, *3A, or *3B alleles. No statistically significant differences (P = 0.23) were observed in the genotype frequencies (*1/*1 and *1/*3C) between cases and controls [Table 2 and Figure 1].

Table 2.

Thiopurine S-methyltransferase and Nudix hydrolase genotype frequencies and association of genotypes with leukopenia in the study population

Genotype	Cases (n=15), n (%)	Controls (n=14), n (%)	P	OR (95% CI)
TPMT *1/*1	12 (80)	14 (100)	0.22	0.12 (0.006–2.6)
TPMT *1/*3C	3 (20)	0
NUDT15 *1/*1	11 (73.3)	12 (85.7)	0.65	0.46 (0.07–3)
NUDT15 *1/*3	4 (26.7)	2 (14.3)

Values are expressed as n (%), P value computed using Fisher’s exact test, calculated with Haldane-Anscombe correction as some cells had zero value. TPMT=Thiopurine S-methyltransferase, NUDT15=Nudix hydrolase, OR=Odds ratio, CI=Confidence interval

Figure 1.

TPMT Strip Assay for TPMT (238G>C, 460G>A, and 719 A>G) polymorphism with wild type and mutant indicated in the inset

Restriction fragment length polymorphism genotyping

After digestion with restriction enzymes, wild-type NUDT15 samples demonstrated the 191 bp PCR product while those which were heterozygous for NUDT15 gene locus had additional bands of 69 bp and 122 bp size [Figure 2].

Figure 2.

Polymerase chain reaction-restriction fragment length polymorphism genotyping: NUDT15 genotyping. Lanes 1, 3, 5: Undigested sample (i.e., no restriction enzyme); lane 2: Heterozygous sample; lanes 4 and 6 homozygous wild-type sample, showing no digestion; lane M: 100 bp DNA ladder

Nudix hydrolase genotype frequencies

The wild genotype (*1*1) was present in 73.3% and 85.7% of cases and controls, respectively. The heterozygous mutant genotype (*1*3) was demonstrated in 26.7% of cases and 14.3% of controls. However, this difference had no statistical significance [P = 0.65, Table 2]. Two subjects were both TPMT heterozygous mutant (*1*3C) and NUDT15 heterozygous mutant (*1*3). Hardy–Weinberg equilibrium testing was done which showed no deviations. The OR of AZA-induced leukopenia with the different genotypes is shown in Table 2.

Discussion

This case–control study evaluated the genotype frequencies of TPMT *1/*1 (c. 474T, rs2842934) and the heterozygous mutant *1/*3C (c.719A>G, rs1142345) and NUDT *1/*1 and *1/*3 in a cohort of autoimmune disorder patients from eastern India who were on AZA immunosuppressive therapy. As per our inclusion criteria, cases (n = 15) had developed leukopenia (total WBC count <4000/ul) within 1 year of treatment initiation that necessitated a reduction of dose or drug withdrawal while controls (n = 14) were age and gender matched on AZA for more than 1 year without developing leukopenia. The cases and controls were unrelated individuals and were residents of the state of West Bengal for at least two generations. The dose of AZA had to be reduced, and in some, the drug had to be withdrawn.

We did not observe any significant statistical difference in the TPMT and the NUDT genotypes studied between cases and controls. Three subjects were both heterozygous mutants for TPMT *1 * 3C and NUDT *1 * 3.

Published literature suggests that TPMT *1 * 3C is the most common variant in Asians, but its frequency is relatively low (1%–3%).[10] A population-based study from Singapore reported that 2.7% (n = 11/401) had heterozygous genotype for the locus studied.[2] In contrast, 10.3% of our study population (n = 3/29) are heterozygous. This finding was, however, consistent with the guidelines of the Clinical Pharmacogenetics Implementation Consortium which recommends genotype-based dosing of thiopurines since 3%–14% of the population are heterozygous for the following alleles (*1/*2, *1/*3A, *1/*3B, *1/*3C, *1/*4).[7,11,12] No statistically significant differences were observed in the TPMT genotype (*1/*1 and *1/*3C) and allele frequencies (*1 and *3C) (P = 0.23 and P = 0.22, respectively) between cases and controls in our study. Significant statistical differences in genotype (P = 0.65) and allele frequencies (P = 0.67) of NUDT15 between cases and controls were also not observed.

In 2014, Yang et al. showed in Korean IBD patients that there was a significant association between NUDT15 and AZA-induced leukopenia.[4] In a Japanese IBD cohort early onset leukopenia was observed with patients having NUDT15 RC139C genotype.[5] Similar findings were also observed in a cohort of Chinese patients with rheumatic arthritis.[7] A meta-analysis reported that in the Asian population, NUDT15 R139C was significantly associated with the development of early onset leukopenia (OR = 15.53, 95% CI = 7.91–30.50, P < 0.001).[13] A subgroup analysis in a meta-analysis published in 2018 showed that for rs186364861 and rs554405994 of the NUDT15 gene, the diagnostic OR for early-onset leukopenia was statistically significant but not for late-onset leukopenia.[14] In contrast, our study found no association between NUDT15 and leukopenia with AZA (OR = 2.18, 95% CI = 0.33–14.36, P = 0.42).

Several studies have been published on the Chinese population in the recent past. A recently published whole-genome sequencing study showed that the NUDT15 R139C variant was found to be associated with AZA-induced leukopenia in the study population suffering from autoimmune diseases.[15] Another study in a Chinese population with systemic lupus erythematosus and Sjogren’s disease demonstrated a significant association with NUDT15 R139C while no association was found with TPMT.[10] Another Chinese study demonstrated that TPMT *3C and NUDT15 * 3 are important prognostic gene markers for AZA-induced toxicity in patients with rheumatic diseases.[7] The study findings corroborated with several studies on the Chinese population concerning TPMT, but differences were observed in NUDT15 genotypes. Although we observed a difference in the NUDT15139C genotype frequencies in cases versus controls, this difference was statistically insignificant.

The present study is limited because of the limited sample size. Moreover, this study cohort consisted of autoimmune disease affected patients; therefore, extrapolation of our study findings should be done cautiously for other disease cohorts where AZA is used in the treatment regimen. Nevertheless, to the best of our understanding, this is one of the few studies from India to assess the association of TPMT and NUDT15 genotype polymorphisms with AZA-induced leukopenia in subjects with autoimmune disease. The study results have provided insight into the common genetic polymorphisms and incidences of leukopenia with this pharmacotherapy in the patient cohort attending public hospitals for treatment in Kolkata.

Conclusion

Our study found no significant association between the TPMT heterozygous (*1/*3C) mutant and NUDT15 heterozygous mutant (*1*3). and the development of leukopenia with AZA which is consistent with previous studies. Therefore, more studies need to be conducted with a larger sample size to indicate whether TPMT and NUDT15 genotype polymorphisms are associated with leukopenia caused by AZA in patients with autoimmune disease.

Financial support and sponsorship

DST-FIST, Govt of India and DST, Govt of West Bengal.

Conflicts of interest

There are no conflicts of interest.