Pharmacogenomics study in a Taiwan methadone maintenance cohort

Abstract

Pharmacogenomics is research to study the drug treatment responses in subgroups of patients according to their genetic variants or genetic expression information. Methadone maintenance treatment, which is usually prescribed for patients with heroin dependence, was launched in Taiwan by the government in 2006. In this study, 366 patients who had taken methadone continually in the previous 7 days were examined. Data from administration of the Treatment Outcomes Profile (TOP), Severity of Dependence Scale (SDS), Clinical Opioid Withdrawal Scale (COWS), and Treatment Emergent Symptoms Scale (TESS) were obtained from patients' report records. Genes encoding the liver cytochrome P-450 (CYP) enzymes that are involved with the metabolism of methadone (CYP2B6, 3A4 and 2C19) were selected and genotyped in this cohort. We found that the SNPs on CYP2B6 were associated with plasma S-methadone concentration; SNPs on CYP3A4 were associated with withdrawal symptoms and side effects; and SNPs on CYP2C19 were associated with methadone dose. SNPs in the genes encoding the morphine phase II metabolic enzyme, UGT2B7, were associated with withdrawal symptom scores. In pharmacodynamic genes, the SNPs on OPRM1 were associated with insomnia and change in libido side effects. We conclude that SNP markers may be useful for future methadone dosage adjustment and to reduce adverse reactions.

1. Introduction

Individual steady state plasma drug concentrations may vary from person to person following the drug's administration. Research reports suggest that the steady state plasma drug concentrations may change from 5 to 10 times in a male population with the same body weight of around 70 kg [1]. If gender, age, and pregnancy status are taken into consideration, the steady state plasma drug concentrations may vary from 50 to100 times. Based on this result, using therapeutic drug monitoring to place patients in subgroups in order to examine treatment responses is impossible. Pharmacogenomics is trying to use information about individual's genomic make-up to subgroup patients and match them with a more optimal medication choice. This is a new field of pharmacology, and single nucleotide polymorphism (SNP) is often used as a genomic marker in subgrouping patients for their treatment responses.

A successful pharmacogenomics clinical trial requires a few set criteria. The most important is to ensure the participating patients' compliance through both observation by research nurses and the measurement of the steady state drug concentrations of each patient. A well-designed clinical trial is another key to achieving a successful pharmacogenomics study.

The Taiwanese government started a methadone maintenance treatment (MMT) program in 2006. Since then, more than 90 hospitals have provided MMT, and over 11,000 heroin-dependent patients have participated in this program. The purpose of this program is to reduce heroin abuse, the spread of infectious diseases, such as HIV and HCV infection, and the drug-related crime rate. As individual methadone doses range from 5 to 180 mg/day, the drug-drug interaction may cause overdose and might induce lethal complications. The National Health Research Institutes (NHRI), therefore, launched the study of pharmacogenomics in a MMT cohort.

2. Methods

2.1 Subjects

This study was designed as a cross-sectional methadone clinical trial and was approved by the Institutional Review Boards of the National Health Research Institutes (Miaoli County, Taiwan) and all seven participating hospitals (Tao-Yuan Mental Hospital, En-Chu-Kong Hospital, Far-Eastern Memorial Hospital, Taipei City Hospital Song-De and Yang-Ming Branches, China Medical University Hospital, and Wei Gong Memorial Hospital). A written informed consent was obtained from each participant. The project also has been registered with the U.S. National Institutes of Health Clinical Trial database (http://www.clinicaltrial.gov/ct/show/NCT01059747). We recruited 366 subjects with heroin dependence who were undergoing MMT as outpatients. The inclusion criteria included an age of 18 years or above, being under MMT for at least 3 months with regular attendance for the past 7 days, and a methadone dosage adjustment of no more than 10 mg in the past 7 days. Exclusion criteria included co-morbidity with physical or other mental disorders requiring immediate treatment as well as pregnancy.

2.2 Clinical assessments

Demographics, clinical characteristics, and methadone treatment courses, including the dose and treatment duration, and treatment adherence over the previous week were obtained from patients' report records (case report form; CRF). Information regarding the possible co-administration of other medications in the previous week was obtained either from medical records or the subjects' reports [2]. Several interviewer-administered assessments, including the Treatment Outcomes Profile (TOP) [3], which measures the amount and frequency of alcohol and other illicit substance used in the past 28 days, the Severity of Dependence Scale (SDS), and the Clinical Opioid Withdrawal Scale (COWS), which measures the severity of 11 opioid withdrawal symptoms [4], were conducted before methadone was administered.

Methadone-related adverse events were assessed by research nurses using the Treatment Emergent Symptoms Scale (TESS) [5]. The TESS consists of 43 treatment emergent symptoms, and only 31 were complained from the patients in this study. Only symptoms that occurred after the initiation of MMT were counted as adverse events related to methadone. The severity of each symptom was rated on a 3-point Likert scale ranging from mild, moderate, to severe. In the present study, only adverse events with a frequency of 15% or above were included in the analyses.

2.3.1 Serum and urine drug test

Levels of glutamate oxaloacetate transaminase (GOT, reference range: <38 U/L), glutamic pyruvic transaminase (GPT, reference range: <41 U/L) and gamma-glutamyl transpeptidase (γ-GT, reference range: 8-61 U/L) from serum samples of patients were measured at the Taipei Institute of Pathology (Taipei, Taiwan).

Urine specimens were collected prior to the administration of methadone on the study day. The morphine screen test was performed via a kinetic interaction of microparticles (KIMS) on an Integra 800 device (Roche Diagnostics, Basel, Switzerland). In our present analyses and previous reports [2,6,7], the urine morphine test was used as a surrogate measurement for the methadone treatment outcome.

2.4 Analyses of methadone and its metabolites in the plasma

Whole blood samples of 12 ml were taken at around 24 ± 2 hrs after the last methadone dose, when the plasma concentration of methadone is likely to be at its lowest. Plasma concentrations of racemic methadone and its metabolite, 2-ethylidene-1,5-dimethyl-3,3-diphenylpyrrolidine (EDDP) were measured using high-performance liquid chromatography (HPLC) using the settings described in our previous study [8].

2.5 SNP selection and genotyping

Selection of SNPs was based on literature reports [2,6,9-13], a minor allele frequency of greater than 0.1 on the HapMap in Chinese ethnic groups (http://hapmap.ncbi.nlm.nih.gov/index.html.en), and possible functionalities predicted by the bioinformatics tool FastSNP [14].

Genomic DNA was extracted from the buffy coat of 6 ml whole blood lymphocyte pellets using the Puregene^® Blood kit C (QIAGEN Sciences; Maryland, USA). Genotypes of the selected 17 SNPs were identified using matrix-assisted laser desorption/ ionization-time of flight mass spectrometry (MALDI-TOF MS) [15]. This genotyping method has been utilized in a broad variety of clinical applications due to the accuracy of SNP detection, sufficient sensitivity to score SNPs from small amounts of template, flexibility of the procedure, and cost-effectiveness [16].

2.6 Statistics

All statistical analyses were conducted using SAS software, Version 9.1 (SAS Institute, Inc., Cary, NC). The Wilcoxon rank-sum test for continuous data and chi-square test or Fisher's exact test for categorical data were applied in clinical character, TESS, and COWS comparisons between patients testing positive for urine morphine and patients testing negative. Association analyses between the selected SNPs and treatment outcomes, withdrawal symptoms, and adverse events were calculated using a generalized linear model (GLM) procedure. False discovery rate (FDR) test of the MULTTEST procedure was employed to correct for multiple hypotheses tests of p-values [17]. The Hardy-Weinberg equilibrium test and haplotype block association analysis were performed using HAPLOVIEW version 4.1 [18]. A statistical significance was designed with p-value less than 0.05.

3. Results

3.1 General clinical characteristics

Half of these 366 MMT patients tested positive for urine morphine (Table 1). The average age of these patients was 38.2 ± 7.7 years and almost all of the participants smoked cigarettes (99.5%). Males were predominant (n = 297, 81.1%) in this cohort. The average starting dose of methadone was 32 ± 11 mg/day and the dosage at the time of testing was 54.7 ± 28 mg/day. The ratios of plasma R- and S-methadone concentration/methadone dose were higher in the patients who tested negative for urine morphine than those who tested positive. The HCV incidence in these patients was 95%. There were no significant differences between those who tested positive for urine morphine and those who tested negative in regard to BMI, HIV, and HCV tests.

Table 1. General demograph of 366 methadone maintenance treatment patients.

	Overall N=366	Urine Morphine Positive n=185	Urine Morphine Negative n=178	P-value
	Mean ± SD	Mean ± SD	Mean ± SD
Starting dose of Methadone (mg/day)	32.04 ± 11.15	31.64 ± 10.07	32.44 ± 12.22	0.77a
Current dose of Methadone (mg/day)	54.67 ± 28.12	54.53 ± 26.07	55.32 ± 30.13	0.88a
Age (years)	38.17 ± 7.72	38.37 ± 7.96	37.87 ± 7.46	0.57a
Male	297 (81.2%)	152 (82.2%)	142 (79.8%)	0.56b
BMI (kg/m-1)	23.58 ± 3.52	23.62 ± 3.57	23.59 ± 3.49	0.89a
R-Methadone/methadone dose ratio	3.86 ± 2.32	3.7 ± 2.71	4.03 ± 1.82	0.001a
S-Methadone/methadone dose ratio	2.77 ± 1.57	2.58 ± 1.45	2.98 ± 1.66	0.012a
R-EDDP/methadone dose ratio	0.31 ± 0.5	0.26 ± 0.33	0.33 ± 0.54	0.26a
S-EDDP/methadone dose ratio	0.33 ± 0.49	0.31 ± 0.38	0.33 ± 0.58	0.95a
Human immunodeficiency virus (HIV)	86 (24.0%)	50 (27.3%)	36 (20.8%)	0.15b
Hepatitis C virus (HCV)	334 (94.9%)	173 (96.1%)	158 (93.5%)	0.27b

Values are shown as mean ± SD or N (%).

Bold P-value: P<0.05.

^aWilcoxon rank-sum test.

^bChi-Square test.

The average withdrawal symptom score rated by COWS was 1.49 ± 1.86 (Table 2). There was a statistically significant higher incidence of bone or joint aches symptom scores in patients testing positive for urine morphine (13% and 0.14 average score) than for patients who tested negative (6.2% and 0.08 average score). The top 15 (by percentage) side effects rated by TESS are listed in Table 3. The side effect of impaired mentation showed a significantly higher percentage of incidence in the patients who tested positive for urine morphine than those who tested negative (22.2% vs. 20.8%).

Table 2. Withdrawal Symptoms of Methadone Maintenance Subjects (N=366).

	Overall N=366			Urine Morphine Positive N=185			Urine Morphine Negative N=178			P-valuea	P-valueb,c
	n	%	Mean ± SD	n	%	Mean ± SD	n	%	Mean ± SD
Sum of COWS	366		1.49 ± 1.86	185		1.51 ± 1.93	178		1.44 ± 1.78	0.85
Heart Rate	365		77.55 ± 11.86	184		76.83 ± 11.29	178		78.20 ± 12.24	0.37
Sweating	41	11.20%	0.13 ± 0.40	26	14.10%	0.17 ± 0.48	14	7.90%	0.08 ± 0.30	0.06	0.06b
Restlessness	24	6.60%	0.07 ± 0.25	15	8.10%	0.08 ± 0.27	8	4.50%	0.04 ± 0.21	0.16	0.16b
Pupil size	69	18.90%	0.21 ± 0.46	34	18.50%	0.20 ± 0.44	35	19.70%	0.22 ± 0.48	0.74	0.77b
Bone or Joint aches	36	9.80%	0.11 ± 0.36	24	13.00%	0.14 ± 0.38	11	6.20%	0.08 ± 0.33	0.033	0.028b
Runny nose or tearing	31	8.50%	0.11 ± 0.39	14	7.60%	0.10 ± 0.36	17	9.60%	0.13 ± 0.43	0.49	0.50b
GI Upset	16	4.40%	0.07 ± 0.33	9	4.90%	0.08 ± 0.39	7	3.90%	0.05 ± 0.27	0.65	0.67b
Tremor	39	10.70%	0.15 ± 0.46	20	10.80%	0.13 ± 0.40	19	10.70%	0.17 ± 0.53	0.91	0.97b
Yawning	10	2.70%	0.04 ± 0.23	6	3.20%	0.04 ± 0.25	3	1.70%	0.02 ± 0.18	0.34	0.50c
Anxiety or Irritability	39	10.70%	0.12 ± 0.35	19	10.30%	0.12 ± 0.37	19	10.70%	0.11 ± 0.31	0.94	0.90b
Gooseflesh skin	8	2.20%	0.07 ± 0.44	4	2.20%	0.06 ± 0.44	4	2.30%	0.07 ± 0.45	0.96	1.00c

Values are shown as mean ± SD or n, %.

Bold P-value: P<0.05.

^aWilcoxon rank-sum test.

^bChi-Square test.

^cFisher Exact test.

Table 3. The top 15 Side Effects of Methadone Maintenance Subjects (N=366).

	Overall N=366			Urine Morphine Positive N=185			Urine Morphine Negative N=178			P-valuea	P-valueb
	n	%	Mean ± SD	n	%	Mean ± SD	n	%	Mean ± SD
Constipation	248	67.80%	1.96 ± 0.84	125	67.60%	2.01 ± 0.85	121	68.00%	1.92 ± 0.83	0.4	0.93
Sedation	172	47.00%	1.51 ± 0.70	94	50.80%	1.51 ± 0.68	77	43.30%	1.49 ± 0.72	0.73	0.15
Change in Libido	111	30.30%	1.72 ± 0.79	63	34.10%	1.70 ± 0.75	47	26.40%	1.77 ± 0.84	0.76	0.11
Dry Mouth	101	27.60%	1.55 ± 0.70	59	31.90%	1.61 ± 0.74	41	23.00%	1.49 ± 0.64	0.49	0.06
Excessive Sweating	71	19.40%	1.86 ± 0.85	38	20.50%	1.95 ± 0.87	32	18.00%	1.75 ± 0.84	0.34	0.54
Insomnia	67	18.30%	1.93 ± 0.78	32	17.30%	2.03 ± 0.78	34	19.10%	1.85 ± 0.78	0.36	0.66
Impaired Mentation	79	21.60%	1.61 ± 0.72	41	22.20%	1.78 ± 0.76	37	20.80%	1.43 ± 0.65	0.032	0.75
Fatigue	65	17.80%	1.62 ± 0.78	32	17.30%	1.78 ± 0.79	32	18.00%	1.47 ± 0.76	0.08	0.87
Difficulty with Urination	52	14.20%	1.38 ± 0.60	28	15.10%	1.50 ± 0.69	23	12.90%	1.26 ± 0.45	0.24	0.54
Increase in Appetite	46	12.60%	1.48 ± 0.59	25	13.50%	1.52 ± 0.51	21	11.80%	1.43 ± 0.68	0.37	0.62
Decrease in Appetite	44	12.00%	1.55 ± 0.70	26	14.10%	1.50 ± 0.71	18	10.10%	1.61 ± 0.70	0.54	0.25
Weight Gain	38	10.40%	1.68 ± 0.74	18	9.70%	1.67 ± 0.59	20	11.20%	1.70 ± 0.86	0.86	0.64
Weakness	33	9.00%	1.36 ± 0.70	18	9.70%	1.56 ± 0.86	14	7.90%	1.14 ± 0.36	0.17	0.53
Malaise	27	7.40%	1.48 ± 0.64	13	7.00%	1.62 ± 0.65	13	7.30%	1.38 ± 0.65	0.31	0.92
Tachycardia/Palpitations	25	6.80%	1.44 ± 0.71	14	7.60%	1.43 ± 0.65	10	5.60%	1.50 ± 0.85	1.00	0.45

Values are shown as mean ± SD or n, %.

Bold P-value: P<0.05.

^aWilcoxon rank-sum test.

^bChi-Square test.

3.2 Association analyses of Pharmacokinetic genes and methadone dose

Methadone is mainly metabolized in the liver through specific isoforms of the cytochrome P-450 (CYP) enzyme system [19]. These isoforms include CYP2B6, CYP2C19, and CYP3A4, which have metabolic preferences toward the two different methadone enantiomers; the R-form is preferably metabolized by CYP2C19 and the S-form is preferably metabolized by CYP2B6 [19]. In our previous studies, the SNPs on CYP2B6 were significantly associated with plasma S-methadone concentration, the ratio of S-methadone/methadone dose, and the S-methadone clearance [6]. We also found that the CYP2B6 has a higher level of activity in the HCV-positive patients [20]. SNPs in CYP3A4 were significantly associated with COWS, TESS, and betel nut use [12]. The higher the withdrawal symptoms scores, the higher the side effect symptom score, but the lower the betel nut use. The gene dose of CYP2C19 showed significant associations with methadone dose, R-methadone/methadone dose, and R-EDDP/methadone dose [13]. Gene dose is associated with TESS scores in patients testing positive for urine morphine. The extensive metabolizers had a higher side effect score than did poor metabolizers.

In gene-gene interaction association analyses, the CYP2C19 gene may not interact with either CYP2B6 or CYP2C19. However, the allele types in either CYP2B6 or CYP3A4 may subgroup the methadone dose into six different dosage ranges and increase the level of significant associations between the gene dose of CYP2C19 and methadone dose. In three-gene interaction association analyses, the CYP2C19 did not interact with either CYP2B6 or CYP3A4. The allelic combination on CYP2B6 and CYP3A4 further subgrouped the methadone dose into 12 different dose ranges and increased the level of significant associations between CYP2C19 and methadone dose from a p-value of 0.018 to less than 10^-4. This CYP2C19 interaction was also significantly associated with methadone tolerance when it was defined as the dose difference between the current dose and the starting dose of methadone.

In phase II morphine metabolic enzyme UDP glucuronosyltransferase 2 family, polypeptide B7 (UGT2B7), the SNPs in this coding region showed significant associations with the severity of withdrawal symptom scores rated by COWS [7].

3.3 Association analyses of a Pharmacodynamic gene OPRM1

The pharmacodynamic gene was reported first in the mu-opioid receptors (OPRM1). The non-synonymous OPRM1 SNP rs1799971 (exon 1; A118G polymorphism; Asn40Asp) showed a marginally significant association with methadone dose (adjusted GLM, p-value=0.027). The intron 1 region of SNPs demonstrated significant associations with the side effects of change in libido and insomnia [2]. It was also significantly associated with plasma nicotine metabolite cotinine [21]. In the recessive model of association analyses, the patients who had a higher plasma cotinine level had a lower insomnia symptom score. The pharmacokinetic genes may be correlated to physical dependence, whereas the pharmacodynamic genes are correlated to psychological dependence.

4. Discussion

Using this methadone maintenance cohort, we found that pharmacogenomics studies in heroin dependence are impossible to explain using a single gene association analysis. The metabolism of methadone involves more than one single cytochrome P-450 (CYP) enzyme; e.g., 2B6, 2C19 and 3A4. Each CYP isozyme may play different roles in their genetic variants. The patients in this cohort were exposed to multiple environmental influences; 95% patients tested positive for the HCV antibody. All patients reported smoking cigarettes. It is impossible to report these important environmental influences with a single genetic association analyses.

In the current study, the pharmacokinetic genes especially, the CYP2C19 x CYP2B6 x CYP3A4 interaction, may subgroup the patients into different methadone dosage groups with a p-value lower than that from the p-value of a pharmacodynamic OPRM1 single gene. This result indicated that the required methadone dosage is regulated by genes in combination with pharmacokinetic and pharmacodynamic influences. SNPs in the pharmacokinetic genes, such as CYP2B6, CYP2C19, and CYP3A4, may be indicators for the plasma methadone concentrations, methadone dose, and severity of withdrawal symptoms.

SNPs in the pharmacodynamic genes may influence the dosage prediction and the treatment side effects. SNPs in the pharmacodynamic genes, especially the OPRM1, were associated with methadone dose, and may be predictors for methadone side effects, such as change in libido and insomnia.

Pharmacogenomics studies promise the advent of personalized medicine. In association analyses of SNPs in single genes, we found that each single gene covers different numbers of exons and introns. Using single genetic variants, studies may provide further information for genetic regulation and its functional correlation with the responses to direct further biological study or to be replicated in the study of other ethnic groups.

The limitation of this cohort include gender imbalance (81% are male), 95% HCV positive, around 50% of patients tested positive for urine morphine , and it was a relatively small sample. Replication studies are needed in order to establish future treatment guidelines.

Additional methods of optimizing drug therapy, with respect to patients' genotype, warrant further investigation to ensure maximum efficacy with minimal adverse effects. This study considered the potential in a single gene of both pharmacokinetics genes (e.g., CYP2B6, 3A4, 2C19, UGT2B7 and PXR) and pharmacodynamic genes (e.g., OPRM1). Studying random combinations of these genes and then making association analyses without having the results of each single gene would not reveal the individual contributions from each single gene in the systemic methadone treatment response. Gene-gene interaction should be based upon single gene discoveries, not nebulous theory. As the pharmacogenomics studies are still in their early stage, studies of a number of SNP markers in each single gene, which may potentially represent the function of a gene, are essential. We therefore showed the SNP IDs, so that it would be easier for researchers who are interested in replicating the study with other ethnic groups.

Methadone pharmacogenomics study may provide a key to deciphering the mechanisms of opioid dependence. There is great potential for pharmacogenomics to be a clinical practice guideline for future individualized medicine. Our studies provide essential information in terms of bridging the basic biological functional study, through understanding the exon and intron roles in each gene, and the potential drug-targeting of novel treatment processes, through the association analyses of treatment responses.

Figure 1.

Methadone metabolic pathway.

Figure 2.

Pharmacogenomics study diagram for methadone maintenance treatment patients.