The Pharmacogenomics Journal (2014), 1–8 © 2014 Macmillan Publishers Limited All rights reserved 1470-269X/14 www.nature.com/tpj
ORIGINAL ARTICLE
ABCB1 polymorphisms predict imatinib response in chronic myeloid leukemia patients: a systematic review and meta-analysis Q Zheng1, H Wu2, Q Yu3, DH (Dennis) Kim4, JH Lipton4, S Angelini5, S Soverini6, D Vivona7, N Takahashi8 and J Cao1 Imatinib mesylate, a competitive tyrosine kinase inhibitor, is considered the first-line therapy drug for Ph+ chronic myeloid leukemia (CML). Three single-nucleotide polymorphisms (SNPs) in the ATP-binding cassette, subfamily B (MDR/TAP), member 1 gene (ABCB1/MDR1), c.1236C4T, c.2677G4T/A and c.3435C4T, have been shown to affect cellular transport/metabolism of imatinib. The associations between these SNPs and imatinib response in CML patients have been widely evaluated, but the results were inconsistent. To derive a conclusive assessment of the associations, we performed a meta-analysis by combining data from a total of 12 reports including 1826 patients. The results showed that the 2677G allele or 3435T allele predicted a worse response to imatinib in CML patients, whereas 1236CC genotype was associated with better response in CML patients from Asian region. In conclusion, this meta-analysis suggests that c.1236C4T, c.2677G4T/A and c.3435C4T can be served as predictive markers for the therapeutical use of imatinib in CML patients. The Pharmacogenomics Journal advance online publication, 23 September 2014; doi:10.1038/tpj.2014.54
INTRODUCTION Chronic myeloid leukemia (CML) is a clonal myeloproliferative disorder of pluripotent hematopoietic stem cells that acquires a Philadelphia (Ph) chromosome encoding the BCR–ABL oncogenic fusion protein with constitutive tyrosine kinase activity.1 Use of imatinib mesylate, which is a competitive BCR–ABL tyrosine kinase inhibitor, is the first-line treatment of Ph+ CML.2 Imatinib has dramatically improved the clinical outcome and prognosis for CML patients.3 Despite the marked improvement in CML treatment, however, resistance exists in many patients.4 Several mechanisms for imatinib resistance in CML have been proposed, including heterogeneity in absorption, distribution and metabolism of imatinib.5 The ATP-binding cassette, subfamily B (MDR/TAP), member 1 gene (ABCB1/MDR1) encodes the cell surface P-glycoprotein (Pgp) acting as an energy-dependent multidrug efflux pump.6 Imatinib is a substrate of Pgp and the intracellular concentrations of imatinib were shown to be significantly lower in Pgp-expressing cells.7,8 The genetic polymorphisms in ABCB1 gene have the potential to alter the expression and function of Pgp and therefore influence imatinib disposition.9 The c.1236C4T, c.2677G4T/A and c.3435C4T are the most frequent three single-nucleotide polymorphisms (SNPs) in the coding region of ABCB1 and in linkage disequilibrium.9 The c.1236C4T (in exon 12 position 412) and c.3435C4T (in exon 26 position 1145) are synonymous SNPs, whereas the c.2677G4T/A in exon 21 results in the substitution of alanine in position 893 by serine (2677T) or threonine (2677A),
respectively.10 Early report has found that the functional variation in ABCB1 gene could explain at least in part variable responses to imatinib.9 Silent mutations in Pgp have recently been found to alter the substrate specificity that might be affecting the timing of cotranslational folding and insertion of Pgp into the membrane.10 In addition, there is evidence that the synonymous SNP c.3435C4T resulted in decreased levels of mRNA expression.11 Together, these results implicate that these three SNPs, though their exact role is yet undefined, may result in altered expression and function of Pgp. Mounting evidence has demonstrated that these SNPs, c.1236C4T, c.2677G4T/A and c.3435C4T, were associated with imatinib response in CML patients. Dulucq et al.12 firstly reported that patients with 1236TT genotype or 2677G allele achieved better major molecular response (MMR) at 12 months. However, among a larger cohort, they found no significant difference in genotype frequencies between patients with or without MMR was observed for the three polymorphisms.13 In addition, several other studies have analyzed the relationship between c.1236C4T, c.2677G4T/A, c.3435C4T and imatinib response, but the results were inconsistent. For instance, enhanced MMR and complete molecular response at 12 months significantly associated with the 1236CC genotype was reported by Deenik et al.,14 and patients carrying 3435TT and 2677TT showed lower probabilities to obtain a MMR and complete molecular response. The 3435CT/TT genotype and 2677T allele were inversely associated with primary failure of imatinib treatment.15 Takahashi et al.16 found no significant associations between the three polymorphisms and
1 Clinical Research Center, The Second Affiliated Hospital, School of Medicine, Zhejiang University, Hangzhou, China; 2Department of Neurosurgery, The Second Affiliated Hospital, School of Medicine, Zhejiang University, Hangzhou, China; 3Department of Surgical Oncology, The First Affiliated Hospital, School of Medicine, Zhejiang University, Hangzhou, China; 4Chronic Myelogenous Leukemia Group, Princess Margaret Cancer Centre, University Health Network University of Toronto, Toronto, Ontario, Canada; 5Department of Pharmacy and Biotechnology, University of Bologna, Bologna, Italy; 6Department of Experimental, Diagnostic and Specialty Medicine, Institute of Hematology, University of Bologna, Bologna, Italy; 7Departmento de Análises Clínicas e Toxicológicas da Faculdade de Ciências Farmacêuticas da Universidade de São Paulo, São Paulo, Brazil and 8 Department of Hematology, Nephrology, and Rheumatology, Akita University Graduate School of Medicine, Akita, Japan. Correspondence: Dr J Cao, Clinical Research Center, The Second Affiliated Hospital, Zhejiang University School of Medicine, 88 Jiefang Road, Hangzhou 310009, China, E-mail: [email protected] Received 1 April 2014; revised 18 July 2014; accepted 13 August 2014
ABCB1 polymorphisms predict imatinib response in CML Q Zheng et al
2 imatinib response among CML patients. Each single study with small sample size had insufficient power to conclusively determine whether these SNPs might be useful in response prediction to imatinib therapy in CML patients. In this meta-analysis, we aim to summarize the available evidence for the associations between ABCB1 polymorphisms and imatinib response in CML patients, and provide the highest level of evidence for their clinical significance up to date. MATERIALS AND METHODS Literature search strategy We performed a comprehensive literature search from Pubmed, ISI web of knowledge, Embase and Cochrane databases up to January 2014 using the following terms: ‘chronic myeloid leukemia’, ‘CML’, ‘MDR1’, ‘ABCB1’ and ‘imatinib’. The titles and abstracts of articles were screened to exclude clearly irrelevant studies, and the full texts of remaining articles were retrieved and independently assessed by two authors to determine whether they contained information on the topic of interest. The reference lists from the main reports and reviews were reviewed for additional relevant works.
Selection criteria and data extraction Eligible studies had to meet the following criteria: (1) the study with independent data has been published in peer-reviewed journals; (2) the study explored the impact of SNPs c.1236C4T, c.2677G4T/A, and c.3435C4T in ABCB1 on the response to imatinib in CML patients; (3) the study supplied enough information of genotype to calculate odds ratio (OR) with its 95% confidence interval. Studies with overlapping data and/or insufficient genotype data were excluded. Two investigators independently reviewed and extracted data with any discrepancies resolved by consensus. The information extracted included the first author’s surname, year of publication, country of origin, number of patients, age, disease phase, therapy, follow-up period, response criteria and the genotype counts in responders and non-responders. The corresponding authors of studies were contacted for additional data when necessary.
Statistical methods We examined the effect of each polymorphism on imatinib response in CML patients using (1) allele contrast, (2) recessive and (3) dominant models.17 Pooled ORs (95% confidence intervals) were calculated using fixed-effects (Mantel–Haenszel method)18 or random-effects (DerSimonian–Laird method).19 We chose the fixed-effects model when homogeneous effects were assumed; otherwise, the random-effects model was more appropriate. Subgroup analyses were performed according to geographic region (Europe, Asia, America or Africa), therapeutic approach (imatinib only or imatinib/+other drugs) and response criteria (MMR or other non-uniform criteria). Heterogeneity of the combined studies was assessed with Cochran’s Q-statistic test20 and I2-test.2 Although the P-value of Cochran’s Q-statistic is below 0.05, heterogeneity was considered statistically significant. The I2-test provides a measure of the degree of heterogeneity in the results. Typically, values of 0–25% are considered to represent no heterogeneity, 25–50% to modest heterogeneity, 50–75% to large heterogeneity and 75–100% to extreme heterogeneity.21 Sensitivity analysis sequentially omitting individual studies was performed to assess the stability of the results. The Egger’s test and funnel plots were used to identify publication bias.22 All of the above analyses were performed using STATA version 12.0 (StataCorp, College Station, TX, USA).
RESULTS On the basis of the above criteria, 12 eligible studies12–16,23–29 were included in this systematic review and meta-analysis involving a total of 1826 cases (Figure 1). These pharmacogenetic studies investigated the influence of the SNPs (c.1236C4T, c.2677G4T/A and c.3435C4T) in ABCB1 gene on response to imatinib in CML patients. Raw genotype data supplied by the corresponding authors from five studies have not been published.15,16,23,26,27 Table 1 describes the main characteristics of all included studies. Among them, five studies were conducted The Pharmacogenomics Journal (2014), 1 – 8
Figure 1.
Flowchart of the study selection process.
in the region of Europe,12–15,27 four in Asia,16,24,25,29 one in South America,26 one in Africa28 and the last one was conducted both in Asia and North America, which was divided into two groups.23 Patients in nine studies were treated with imatinib only at a dose of 300–800 mg per day.15,16,23–29 Patients in two studies following the treatment with imatinib/+interferon,12 with imatinib+cytarabine (AraC),14 respectively, were included in the group of imatinib/+other drugs. The last one was divided into two groups as the data were available.13 The response criterion was defined by MMR in six studies,12–14,16,23,27 among which five were followed at 12 months,12–14,23,27 one was not known.16 The results in the remaining six studies15,24–26,28,29 were based on other nonuniform criteria, which were also summarized in Table 1. ABCB1 c.1236C4T polymorphism and imatinib response in CML Overall, a total of 1623 patients from 11 studies were included for this polymorphism in this meta-analysis.12–16,23–27,29 As shown in Table 2, we found a significant association between the polymorphism and imatinib response in the Asia region under both dominant and recessive models. Figure 2 illustrates subsidiary analysis of geographic region under the dominant model. The presence of 1236CC genotype significantly increased response to imatinib in Asian CML patients (OR = 0.59, P = 0.036). We did not observe significant heterogeneity between these studies (I2 = 0.0%; P = 0.518). No publication bias was detected in these studies (Supplementary Figure S1). ABCB1 c.2677G4T/A polymorphism and imatinib response in CML The meta-analysis of 10 studies13–16,23–28 involving 1579 patients revealed no obvious association between ABCB1 c.2677G4T/A polymorphism and imatinib response in overall populations (Table 3). In the subgroup analyses, the results show several significant associations. Notably, the odds of the G allele carrier state was lower in responders versus non-responders in the subgroup of other non-uniform criteria under the allele contrast model (OR = 1.42, P = 0.031; Figure 3). We obtained similar results in this group under recessive model (Table 3). Conclusively, patients carrying the 2677T/A allele achieved better response with non-uniform response criteria. When stratified by geographic region, there was only one study carried out in Africa. Again the odds of the 2677 homozygous variant carrier state was © 2014 Macmillan Publishers Limited
ABCB1 polymorphisms predict imatinib response in CML Q Zheng et al
3 Table 1.
Characteristics of studies included in this meta-analysis
Reference
Dulucq et al.12 Kim et al.
23
Deenik et al.
2008
French
90
4–46
CP; AP
2010
416
17–87
CP; AP; BC
46
NA
CP
Imatinib 400 mg; imatinib 400 mg+IFN Imatinib 400, 600 or 800 mg Imatinib 800 mg+AraC
Follow-up period (months)
MMR
12
MMR
12
MMR
12
MMR
NA
MMR
2010 2010
French
557
NA
NA
2010
Japan
67
20–81
CP
2011
Spain
65
19–75
CP
Imatinib 400 mg; imatinib 400 mg+IFN; Imatinib 400 mg+AraC; imatinib 600 mg Imatinib 400 mg; imatinib other doses Imatinib 400 mg
2011
China
52
18–76
CP
Imatinib 400 mg
⩾12
2012
China
48
18–76
CP; AP
Imatinib 300–800 mg
105
2012
Brazil
118
18–80
CP
Imatinib 400 mg
18
2013
Italy
189
18–75
CP
Imatinib 400 or 800 mg
12
2013 2013
Egypt Korea
96 82
44 17–79
CP; AP; BC CP
Imatinib 400–600 mg Imatinib 400 mg
18 6
15
24 25 26
Angelini et al.
27
Elghannam et al. Seong et al.29
28
Response criteria
12
13
Maffioli et al.
Vivona et al.
Number of Age Disease phase Therapy patients (years)
Canada; Korea Nederland
Takahashi et al.16
Chen et al.
Country
14
Dulucq et al.
Ni et al.
Year
18
Genotype
c.1236C4T; c.2677G4T/A; c.3435C4T c.1236C4T; c.2677G4T/A; c.3435C4T c.1236C4T; c.2677G4T/A; c.3435C4T c.1236C4T; c.2677G4T/A; c.3435C4T
c.1236C4T; c.2677G4T/A; c.3435C4T Response c.1236C4T; c.2677G4T/A; c.3435C4T CR c.1236C4T; c.2677G4T/A; c.3435C4T CCR; MMR c.1236C4T; c.2677G4T/A; c.3435C4T CCR c.1236C4T; c.2677G4T/A; c.3435C4T MMR c.1236C4T; c.2677G4T/A; c.3435C4T CR; MR c.2677G4T/A CCR c.1236C4T; c.3435C4T
Abbreviations: AP, accelerated phase; AraC, cytarabine; BC, blastic crisis; CCR, complete cytogenetic response; CP, chronic phase; CR, cytogenetic response; IFN, interferon; MMR, major molecular response; MR, molecular response; NA, not applicable.
significantly increased in responders versus non-responders under recessive model (OR = 6.14, P = 0.007; Table 3). Sensitivity analysis and publication bias analysis showed no significant results (Supplementary Figure S2). ABCB1 c.3435C4T polymorphism and imatinib response in CML We compiled 11 studies12–16,23–27,29 that analyzed the association between ABCB1 c.3435C4T polymorphism and imatinib response in CML, nine presented enough data to allow meta-analysis with data from a total of 1373 patients.13,14,16,23–27,29 In overall comparison, significant association was observed under the allele contrast model (OR = 0.85, P = 0.042; Table 4) or dominant model (OR = 0.73, P = 0.011; Figure 4). The overall comparison under allele contrast model indicated that the presence of 3435T allele predicted a worse response to imatinib in CML patients (OR = 0.85, P = 0.042; Table 4). In subgroup analyses, there was no significant result among different regions. However, the group treated with imatinib/+ other drugs under the allele contrast model (OR = 0.72, P = 0.044; Table 4) or recessive model (OR = 0.58, P = 0.035; Table 4) and the group evaluated by MMR under dominant model (OR = 0.72, P = 0.021; Figure 4) achieved significant results. Sensitivity analysis and publication bias analysis showed no significant results (Supplementary Figure S3). DISCUSSION Imatinib is the first generation of tyrosine kinase inhibitor, which exerts a potent inhibitory activity against Ph+ CML. Imatinib have proven efficacious in the treatment of CML; however, a proportion of patients do not respond and/or exhibit resistance.30 To overcome the deficiency of imatinib, nilotinib and dasatinib have been developed and demonstrated remarkable efficacy in the induction of cytogenetic and molecular responses in newly diagnosed CML patients.31,32 Therefore, biomarkers predicting response to therapy would be helpful for making treatment decisions, in view of a personalized medicine. The pharmacokinetics of imatinib including absorption, distribution, metabolism and excretion may partly explain the © 2014 Macmillan Publishers Limited
interpatient variability in imatinib response.6 Pgp, encoded by ABCB1 gene, is an energy-dependent multidrug efflux pump mediating the efflux of imatinib.7 Overexpression of Pgp confer resistance to imatinib in leukemia cell lines,33,34 and Pgpexpressing cells had lower imatinib intracellular levels.8 The ABC transporter ABCG2 (ATP-binding cassette, subfamily G (WHITE), member 2), also known as breast cancer resistance protein, is another drug efflux pump that mediate imatinib efflux and has been implicated in imatinib resistance.35,36 ABCB1 is more likely involved in imatinib transport at clinically relevant concentrations, compared with ABCG2.35 Besides, the factors resulting in imatinib resistance include the difference in population pharmacogenetics.6,37 The three most common SNPs c.1236C4T, c.2677G4T/A and c.3435C4T in ABCB1 have been showed to be associated with imatinib pharmacokinetic characteristics. The imatinib clearance has been demonstrated to relate to the ABCB1 genotype and was least reduced in thymidine homozygotes at the c.1236C4T, c.2677G4T/A and c.3435C4T loci.9 Another report confirmed that patients with the 3435CT and TT genotypes had lower imatinib clearance compared with patients with 3435CC genotype.38 In addition, polymorphisms in ABCG2 may influence concentration and clinical response of imatinib in both solid and hematologic malignancies. Human embryonic kidney cells transfected with ABCG2 c.141Q4K exhibited greater imatinib accumulation in vitro in comparison with cells expressing wild-type ABCG2, but the clinical relevance of the observation is limited.39 Patients heterozygous for ABCG2 c.421C4A showed a 22% reduction in imatinib clearance compared with wild type.40 The GG genotype in ABCG2 c.34G4A was significantly associated with poor response to imatinib in CML patients.41 Another study reported that the ABCG2 c.421C4A variant allele was related to a higher rate of MMR in CML patients using imatinib.29 In 2008, Dulucq et al.12 firstly identified that c.1236C4T, c.2677G4T/A and c.3435C4T variants in ABCB1 were associated with imatinib response. Patients with 1236TT genotype achieved better response, whereas 2677G allele was associated with a worse response. In recent years, a number of reports have focused on the associations between those SNPs and imatinib response, The Pharmacogenomics Journal (2014), 1 – 8
ABCB1 polymorphisms predict imatinib response in CML Q Zheng et al
4 Table 2.
Stratified analyses for polymorphism c.1236C4T and imatinib response in CML
Analyses
Allele contrast model
Dominant model
Recessive model
OR (95% CI)
P(Z)
I2 (%)
OR (95% CI)
P(Z)
I2 (%)
OR (95% CI)
P(Z)
I2 (%)
Overall
0.86 (0.66, 1.12)
0.263
63.4
0.87 (0.70, 1.09)
0.235
26.1
0.93 (0.74, 1.18)
0.558
50.8
Region Europe Asia America
0.96 (0.66, 1.42) 0.61 (0.37, 1.00) 1.19 (0.87, 1.64)
0.849 0.051 0.282
67.2 59.0 0.0
0.91 (0.68, 1.23) 0.59 (0.36, 0.97) 1.14 (0.70, 1.87)
0.555 0.036 0.589
45.6 0.0 0.0
1.05 (0.75, 1.47) 0.61 (0.40, 0.93) 1.43 (0.82, 2.48)
0.788 0.022 0.207
46.7 55.0 0.0
Therapy Imatinib Imatinib/+other drug
0.82 (0.63, 1.06) 1.02 (0.41, 2.55)
0.129 0.965
49.8 85.4
0.86 (0.66, 1.11) 0.91 (0.59, 1.43)
0.253 0.690
0.0 77.6
0.85 (0.65, 1.12) 1.24 (0.76, 2.03)
0.250 0.387
42.0 75.2
Response criteria MMR Other criteria
0.98 (0.74, 1.29) 0.62 (0.33, 1.17)
0.865 0.139
61.1 65.8
0.95 (0.74, 1.21) 0.58 (0.33, 1.02)
0.672 0.058
34.7 0.0
1.04 (0.80, 1.36) 0.61 (0.36, 1.02)
0.767 0.058
30.9 64.2
Abbreviations: CML, chronic myeloid leukemia; CI, confidence interval; MMR, major molecular response; OR, odds ratio; Z, z-score in normal distribution model.
Figure 2. Forest plot of ABCB1 c.1236C4T polymorphism and imatinib response stratified by geographic region under the dominant model. Data from Dulucq et al.13 were divided into two groups, including patients treated with imatinib only and patients treated with imatinib and other drugs. Data from Kim et al.23 were divided into two groups, including patients from Asia and from America. OR41 indicates a better response to imatinib in CML patients carrying 1236CT and 1236TT genotypes. The vertical dash line represents OR = 0 and the vertical bold line represents OR = 1. The length of horizontal line for each study represents the 95% CI, the middle spot represents the value of OR and the area of the gray square box represents the proportion of weight. The length of diamond represents the combined 95% CI and the centra represents the combined OR. The overlap between the bold vertical line and each horizontal line or diamond represents that the result is not statistically significant. CI, confidence interval; CML, chronic myeloid leukemia; OR, odds ratio.
with inconsistent results. Recently, Zu et al.42 reported 1236TT/CT genotype exposures as having higher risk of imatinib resistance in Asian patients. In the current meta-analysis, we included much more cases with subgroup analyses in more detail among three genetic models, which may lead to a more reliable conclusion. Eleven studies analyzed the synonymous SNP c.1236C4T that is located in ABCB1 exon 12.12–16,23–27,29 One report found an The Pharmacogenomics Journal (2014), 1 – 8
association between the 1236TT genotype and increased imatinib response rate.12 However, three studies14,24,25 found the opposite result and the other seven studies13,15,16,23,26,27,29 found no association between c.1236C4T and imatinib response. This systematic review has shown that c.1236C4T was associated with imatinib response in CML patients in the region of Asia, which is consistent with a recent meta-analysis.42 Patients carrying the CC © 2014 Macmillan Publishers Limited
ABCB1 polymorphisms predict imatinib response in CML Q Zheng et al
5 Table 3.
Stratified analyses for polymorphism c.2677G4T/A and imatinib response in CML
Analyses
Allele contrast model
Dominant model
Recessive model
OR (95% CI)
P(Z)
I2 (%)
OR (95% CI)
P(Z)
I2 (%)
OR (95% CI)
P(Z)
I2 (%)
Overall
1.01 (0.87, 1.17)
0.874
40.9
1.05 (0.84, 1.31)
0.686
12.8%
0.97 (0.76, 1.25)
0.840
48.4
Region Europe Asia America Africa
0.94 0.98 1.09 1.65
1.15) 1.35) 1.50) 2.95)
0.538 0.894 0.618 0.090
54.2 50.1 0.0 NA
1.01 1.05 1.17 1.04
1.36) 1.83) 1.89) 2.41)
0.972 0.873 0.530 0.936
50.0 30.3 0.0 NA
0.81 (0.57, 1.15) 0.92 (0.56, 1.52) 1.04 (0.58, 1.88) 6.14 (1.65, 22.94)
0.242 0.749 0.888 0.007
26.8 56.4 0.0 NA
Therapy Imatinib Imatinib/+other drug
1.09 (0.92, 1.28) 0.77 (0.56, 1.06)
0.316 0.107
27.9 65.2
1.11 (0.86, 1.42) 0.86 (0.53, 1.39)
0.434 0.531
1.1 68.2
1.13 (0.85, 1.50) 0.56 (0.32, 0.97)
0.411 0.040
43.6 0.0
Response criteria MMR Other criteria
0.92 (0.78, 1.09) 1.42 (1.03, 1.96)
0.337 0.031
34.7 17.1
0.99 (0.77, 1.28) 1.25 (0.79, 1.95)
0.935 0.339
0.0 37.4
0.81 (0.61, 1.07) 2.30 (1.23, 4.31)
0.130 0.009
31.3 20.9
(0.76, (0.71, (0.79, (0.92,
(0.75, (0.60, (0.72, (0.45,
Abbreviations: CML, chronic myeloid leukemia; CI, confidence interval; MMR, major molecular response; NA, not applicable; OR, odds ratio; Z, z-score in normal distribution model.
Figure 3. Forest plot of ABCB1 c.2677G4T/A polymorphism and imatinib response stratified by response criteria under the allele contrast model. Data from Dulucq et al.13 were divided into two groups, including patients treated with imatinib only and patients treated with imatinib and other drugs. Data from Kim et al.23 were divided into two groups, including patients from Asia and from America. OR41 indicates a better response to imatinib in CML patients carrying 2677T/A allele. The vertical dash line represents OR = 0, and the vertical bold line represents OR = 1. The length of horizontal line for each study represents the 95% CI, the middle spot represents the value of OR and the area of the gray square box represents the proportion of weight. The length of diamond represents the combined 95% CI and the centra represents the combined OR. The overlap between the bold vertical line and each horizontal line or diamond represents that the result is not statistically significant. CI, confidence interval; CML, chronic myeloid leukemia; OR, odds ratio.
genotype of c.1236C4T were more likely to respond to imatinib as compared to the variant genotypes. However, the conclusion only pertained to population from Asia and could not be generalized to cohorts of other regions. The diversity may come from different frequencies of mutant alleles in different ethnic groups, and diverse gene–environment interactions leading from © 2014 Macmillan Publishers Limited
different life styles and environmental factors among the different populations.43,44 From the 11 reports that studied c.2677G4T/A with imatinib response,12–16,23–28 one study found an association between 2677G allele and increased imatinib response,13 whereas four found the opposite result.12,15,25,28 However, the other six reports The Pharmacogenomics Journal (2014), 1 – 8
ABCB1 polymorphisms predict imatinib response in CML Q Zheng et al
6 Table 4.
Stratified analyses for polymorphism c.3435C4T and imatinib response in CML
Analyses
Allele contrast model
Dominant model
Recessive model
OR (95% CI)
P(Z)
I2 (%)
OR (95% CI)
P(Z)
I2 (%)
OR (95% CI)
P(Z)
I2 (%)
Overall
0.85 (0.73, 0.99)
0.042
36.5
0.73 (0.57, 0.93)
0.011
36.3
0.85 (0.64, 1.11)
0.232
37.2
Region Europe Asia America
0.85 (0.69, 1.04) 0.73 (0.49, 1.07) 0.96 (0.70, 1.33)
0.106 0.104 0.803
25.5 66.3 0.0
0.73 (0.53, 1.01) 0.68 (0.39, 1.20) 0.76 (0.48, 1.23)
0.059 0.184 0.267
11.7 62.2 63.8
0.90 (0.64, 1.27) 0.64 (0.30, 1.35) 0.83 (0.46, 1.49)
0.551 0.241 0.536
56.7 36.0 54.3
Therapy Imatinib Imatinib/+other drug
0.90 (0.75, 1.07) 0.72 (0.53, 0.99)
0.236 0.044
33.0 58.7
0.77 (0.59, 1.02) 0.59 (0.35, 1.00)
0.065 0.052
38.4 56.0
0.99 (0.71, 1.37) 0.58 (0.34, 0.96)
0.939 0.035
33.2 0.0
Response criteria MMR Other criteria
0.89 (0.75, 1.06) 0.71 (0.50, 1.01)
0.188 0.057
2.0 61.6
0.72 (0.55, 0.95) 0.76 (0.47, 1.24)
0.021 0.273
14.1 63.7
0.95 (0.70, 1.28) 0.43 (0.21, 0.88)
0.730 0.021
32.0 9.2
Abbreviations: CML, chronic myeloid leukemia; CI, confidence interval; MMR, major molecular response; OR, odds ratio; Z, z-score in normal distribution model.
Figure 4. Forest plot of ABCB1 c.3435C4T polymorphism and imatinib response stratified by response criteria under the dominant model. Data from Dulucq et al.13 were divided into two groups, including patients treated with imatinib only and patients treated with imatinib and other drugs. Data from Kim et al.23 were divided into two groups, including patients from Asia and from America. OR41 indicates a better response to imatinib in CML patients carrying 3435CT and 3435TT genotypes. The vertical dash line represents OR = 0 and the vertical bold line represents OR = 1. The length of horizontal line for each study represents the 95% CI, the middle spot represents the value of OR and the area of the gray square box represents the proportion of weight. The length of diamond represents the combined 95% CI and the centra represents the combined OR. The overlap between the bold vertical line and each horizontal line or diamond represents that the result is not statistically significant. CI, confidence interval; CML, chronic myeloid leukemia; OR, odds ratio.
that analyzed this polymorphism did not find any association.14,16,23,24,26,27 In this meta-analysis, we observed that individuals with 2677T/A allele achieved better response with non-uniform response criteria. It is possible that even more discriminatory results would have been obtained if a comprehensive set of criteria were set including CR and MR. In conclusion, ABCB1 c.2677G4T/A does not seem to be a good marker of imatinib response. In the 11 studies that analyzed synonymous SNP in exon 26 (c.3435C4T),12–16,23–27,29 six studies found no association The Pharmacogenomics Journal (2014), 1 – 8
between c.3435C4T and imatinib response.12,13,16,23,27,29 In three studies, the authors reported a worse response to imatinib of 3435T allele,14,24,26 whereas the other two reported the opposite result.15,25 The current meta-analysis supported that 3435T allele predicted a worse response to imatinib in CML patients. Many factors can lead to inaccurate, inconsistent or misleading results in various studies. First, the sample size in eight studies was less than 100 cases.12,14–16,24,25,28,29 A study with low statistical power, which results from relatively small sample size, has a reduced chance of detecting a true effect.45 Second, the clinical © 2014 Macmillan Publishers Limited
ABCB1 polymorphisms predict imatinib response in CML Q Zheng et al
7 characteristics, such as patient ethnicity, previous therapy and disease phase, may lead to relatively high heterogeneity in each study. Patients included in five studies had different disease phase,12,13,23,25,28 and in four studies had received other drugs in addition to imatinib.12–14,16 Considering previous therapy, patients in five studies were newly diagnosed,13–15,27,29 in two studies were previously diagnosed,23,26 and other studies didn’t declare.12,16,24,25,28 Third, bias may come from patients selection process, evaluation process and censoring data. Furthermore, ABCB1 c.1236C4T, c.2677G4T/A and c.3435C4T form the most common haplotypes, and that contain the mutated alleles are associated with major structural modifications on Pgp.46 There is report that found significant association between ABCB1 haplotypes and Pgp activity in CML patients.47 In our review of the literature, we identified three studies in which the haplotypes were associated with imatinib response. One study has reported that the 1236C/2677G/3435C haplotype was statistically linked to less frequent MMR,12 whereas another found that 1236CT/ 2677GT/3435CT haplotype was positively associated with the MMR to imatinib in CML patients.26 In addition, the ABCB1 haplotype 1236T/2677G/3435C was more frequently found in patients primarily resistant to imatinib.15 Yet, there are some limitations in this meta-analysis. First, the relatively small sample size results in a lack of uniformly strong statistical power. Second, there exists clinical heterogeneity among the CML patients in this meta-analysis. The differences include clinical parameters such as disease duration, disease phase and use of concomitant medication. Finally, several different outcome measures were used to assess the outcomes from imatinib therapy. In conclusion, this systematic review and meta-analysis demonstrated that the probability of successful treatment with imatinib is influenced, at least in part, by c.1236C4T, c.2677G4T/ A and c.3435C4T variants in ABCB1 gene possessed by the CML patient. The presence of 1236CC genotype (in Asian population under both dominant and recessive models), 2677T/A allele (in patients evaluated by non-uniform response criteria under both allele contrast and recessive models) or 3435C allele (under both allele contrast and dominant models in overall comparison) marks a better response to imatinib in CML patients. The pharmacokinetics for imatinib in CML is not fully known, future detailed studies are needed to elucidate the molecular basis. CONFLICT OF INTEREST Dr Naoto Takahashi reports grants, personal fees from Novartis, Bristol-Myers Squibb and Pfizer, outside the submitted work. The remaining authors declare no conflict of interest.
ACKNOWLEDGMENTS We would like to thank Dr Francisco Cervantes from Hematology Department, Hospital Clínic, IDIBAPS, University of Barcelona, Barcelona, Spain for providing us their raw genotype data. Part of the work was supported by grant FAPESP, Brazil (#09/54184-0) to Dr Douglas Vivona.
REFERENCES 1 Groffen J, Stephenson JR, Heisterkamp N, de Klein A, Bartram CR, Grosveld G. Philadelphia chromosomal breakpoints are clustered within a limited region, bcr, on chromosome 22. Cell 1984; 36: 93–99. 2 Bhamidipati PK, Kantarjian H, Cortes J, Cornelison AM, Jabbour E. Management of imatinib-resistant patients with chronic myeloid leukemia. Ther Adv Hematol 2013; 4: 103–117. 3 Etienne G, Dulucq S, Nicolini FE, Morisset S, Fort MP, Schmitt A et al. Achieving deeper molecular response is associated with a better clinical outcome in chronic myeloid leukemia patients on imatinib front-line therapy. Haematologica 2014; 99: 458–464.
© 2014 Macmillan Publishers Limited
4 Rosti G, Castagnetti F, Gugliotta G, Palandri F, Baccarani M. Second-generation BCR-ABL inhibitors for frontline treatment of chronic myeloid leukemia in chronic phase. Crit Rev Oncol Hematol 2012; 82: 159–170. 5 Szakacs G, Varadi A, Ozvegy-Laczka C, Sarkadi B. The role of ABC transporters in drug absorption, distribution, metabolism, excretion and toxicity (ADME-Tox). Drug Discov Today 2008; 13: 379–393. 6 Diamond JM, Melo JV. Mechanisms of resistance to BCR-ABL kinase inhibitors. Leuk Lymphoma 2011; 52(Suppl 1):12–22. 7 Hegedus T, Orfi L, Seprodi A, Varadi A, Sarkadi B, Keri G. Interaction of tyrosine kinase inhibitors with the human multidrug transporter proteins, MDR1 and MRP1. Biochim Biophys Acta 2002; 1587: 318–325. 8 Widmer N, Colombo S, Buclin T, Decosterd LA. Functional consequence of MDR1 expression on imatinib intracellular concentrations. Blood 2003; 102: 1142. 9 Gurney H, Wong M, Balleine RL, Rivory LP, McLachlan AJ, Hoskins JM et al. Imatinib disposition and ABCB1 (MDR1, P-glycoprotein) genotype. Clin Pharmacol Ther 2007; 82: 33–40. 10 Kimchi-Sarfaty C, Oh JM, Kim IW, Sauna ZE, Calcagno AM, Ambudkar SV et al. A "silent" polymorphism in the MDR1 gene changes substrate specificity. Science 2007; 315: 525–528. 11 Wang D, Johnson AD, Papp AC, Kroetz DL, Sadee W. Multidrug resistance polypeptide 1 (MDR1, ABCB1) variant 3435C4T affects mRNA stability. Pharmacogenet Genomics 2005; 15: 693–704. 12 Dulucq S, Bouchet S, Turcq B, Lippert E, Etienne G, Reiffers J et al. Multidrug resistance gene (MDR1) polymorphisms are associated with major molecular responses to standard-dose imatinib in chronic myeloid leukemia. Blood 2008; 112: 2024–2027. 13 Dulucq S, Preudhomme C, Guilhot F, Mahon FX. Response Is there really a relationship between Multidrug Resistance Gene (MDR1) polymorphisms and major molecular response to imatinib in chronic myeloid leukemia?. Blood 2010; 116: 6145–6146. 14 Deenik W, van der Holt B, Janssen JJ, Chu IW, Valk PJ, Ossenkoppele GJ et al. Polymorphisms in the multidrug resistance gene MDR1 (ABCB1) predict for molecular resistance in patients with newly diagnosed chronic myeloid leukemia receiving high-dose imatinib. Blood 2010; 116: 6144–6145. 15 Maffioli M, Camos M, Gaya A, Hernandez-Boluda JC, Alvarez-Larran A, Domingo A et al. Correlation between genetic polymorphisms of the hOCT1 and MDR1 genes and the response to imatinib in patients newly diagnosed with chronic-phase chronic myeloid leukemia. Leuk Res 2011; 35: 1014–1019. 16 Takahashi N, Miura M, Scott SA, Kagaya H, Kameoka Y, Tagawa H et al. Influence of CYP3A5 and drug transporter polymorphisms on imatinib trough concentration and clinical response among patients with chronic phase chronic myeloid leukemia. J Hum Genet 2010; 55: 731–737. 17 Lu PH, Yang J, Li C, Wei MX, Shen W, Shi LP et al. Association between mitogenactivated protein kinase kinase kinase 1 rs889312 polymorphism and breast cancer risk: evidence from 59,977 subjects. Breast Cancer Res Treat 2011; 126: 663–670. 18 Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959; 22: 719–748. 19 DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986; 7: 177–188. 20 Cochran WG. The combination of estimates from different experiments. Biometrics 1954; 10: 101–129. 21 Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003; 327: 557–560. 22 Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997; 315: 629–634. 23 Kim DH, Kong JH, Byeun JY, Jung CW, Xu W, Liu X et al. The IFNG (IFN-gamma) genotype predicts cytogenetic and molecular response to imatinib therapy in chronic myeloid leukemia. Clin Cancer Res 2010; 16: 5339–5350. 24 Ni LN, Li JY, Miao KR, Qiao C, Zhang SJ, Qiu HR et al. Multidrug resistance gene (MDR1) polymorphisms correlate with imatinib response in chronic myeloid leukemia. Med Oncol 2011; 28: 265–269. 25 Chen WW, Meng FY, Zhong JS, Yin CX, Wang ZX. Effects of hOCT1 and ABCB1 gene on the efficacy of imatinib mesylate in chronic myelocytic leukemia. Zhonghua Yi Xue Za Zhi 2012; 92: 1405–1408. 26 Vivona D, Bueno CT, Lima LT, Hirata RD, Hirata MH, Luchessi AD et al. ABCB1 haplotype is associated with major molecular response in chronic myeloid leukemia patients treated with standard-dose of imatinib. Blood Cells Mol Dis 2012; 48: 132–136. 27 Angelini S, Soverini S, Ravegnini G, Barnett M, Turrini E, Thornquist M et al. Association between imatinib transporters and metabolizing enzymes genotype and response in newly diagnosed chronic myeloid leukemia patients receiving imatinib therapy. Haematologica 2013; 98: 193–200. 28 Elghannam DM, Ibrahim L, Ebrahim MA, Azmy E, Hakem H. Association of MDR1 gene polymorphism (G2677T) with imatinib response in Egyptian chronic myeloid leukemia patients. Hematology 2013; 19: 123–128.
The Pharmacogenomics Journal (2014), 1 – 8
ABCB1 polymorphisms predict imatinib response in CML Q Zheng et al
8 29 Seong SJ, Lim M, Sohn SK, Moon JH, Oh SJ, Kim BS et al. Influence of enzyme and transporter polymorphisms on trough imatinib concentration and clinical response in chronic myeloid leukemia patients. Ann Oncol 2013; 24: 756–760. 30 Milojkovic D, Apperley J. Mechanisms of resistance to imatinib and secondgeneration tyrosine inhibitors in chronic myeloid leukemia. Clin Cancer Res 2009; 15: 7519–7527. 31 Saglio G, Kim DW, Issaragrisil S, le Coutre P, Etienne G, Lobo C et al. Nilotinib versus imatinib for newly diagnosed chronic myeloid leukemia. N Engl J Med 2010; 362: 2251–2259. 32 Kantarjian H, Shah NP, Hochhaus A, Cortes J, Shah S, Ayala M et al. Dasatinib versus imatinib in newly diagnosed chronic-phase chronic myeloid leukemia. N Engl J Med 2010; 362: 2260–2270. 33 Mahon FX, Belloc F, Lagarde V, Chollet C, Moreau-Gaudry F, Reiffers J et al. MDR1 gene overexpression confers resistance to imatinib mesylate in leukemia cell line models. Blood 2003; 101: 2368–2373. 34 Assef Y, Rubio F, Colo G, del Monaco S, Costas MA, Kotsias BA. Imatinib resistance in multidrug-resistant K562 human leukemic cells. Leuk Res 2009; 33: 710–716. 35 Eadie LN, Hughes TP, White DL. Interaction of the efflux transporters ABCB1 and ABCG2 with imatinib, nilotinib, and dasatinib. Clin Pharmacol Ther 2014; 95: 294–306. 36 Burger H, van Tol H, Boersma AW, Brok M, Wiemer EA, Stoter G et al. Imatinib mesylate (STI571) is a substrate for the breast cancer resistance protein (BCRP)/ ABCG2 drug pump. Blood 2004; 104: 2940–2942. 37 Cortes JE, Egorin MJ, Guilhot F, Molimard M, Mahon FX. Pharmacokinetic/pharmacodynamic correlation and blood-level testing in imatinib therapy for chronic myeloid leukemia. Leukemia 2009; 23: 1537–1544. 38 Yamakawa Y, Hamada A, Nakashima R, Yuki M, Hirayama C, Kawaguchi T et al. Association of genetic polymorphisms in the influx transporter SLCO1B3 and the
39
40
41
42
43 44 45
46
47
efflux transporter ABCB1 with imatinib pharmacokinetics in patients with chronic myeloid leukemia. Ther Drug Monit 2011; 33: 244–250. Gardner ER, Burger H, van Schaik RH, van Oosterom AT, de Bruijn EA, Guetens G et al. Association of enzyme and transporter genotypes with the pharmacokinetics of imatinib. Clin Pharmacol Ther 2006; 80: 192–201. Petain A, Kattygnarath D, Azard J, Chatelut E, Delbaldo C, Geoerger B et al. Population pharmacokinetics and pharmacogenetics of imatinib in children and adults. Clin Cancer Res 2008; 14: 7102–7109. Kim DH, Sriharsha L, Xu W, Kamel-Reid S, Liu X, Siminovitch K et al. Clinical relevance of a pharmacogenetic approach using multiple candidate genes to predict response and resistance to imatinib therapy in chronic myeloid leukemia. Clin Cancer Res 2009; 15: 4750–4758. Zu B, Li Y, Wang X, He D, Huang Z, Feng W. MDR1 gene polymorphisms and imatinib response in chronic myeloid leukemia: a meta-analysis. Pharmacogenomics 2014; 15: 667–677. Hirschhorn JN, Lohmueller K, Byrne E, Hirschhorn K. A comprehensive review of genetic association studies. Genet Med 2002; 4: 45–61. Dick DM. Gene-environment interaction in psychological traits and disorders. Annu Rev Clin Psychol 2011; 7: 383–409. Button KS, Ioannidis JP, Mokrysz C, Nosek BA, Flint J, Robinson ES et al. Power failure: why small sample size undermines the reliability of neuroscience. Nat Rev Neurosci 2013; 14: 365–376. Fung KL, Gottesman MM. A synonymous polymorphism in a common MDR1 (ABCB1) haplotype shapes protein function. Biochim Biophys Acta 2009; 1794: 860–871. Vivona D, Lima LT, Rodrigues AC, Bueno CT, Alcantara GK, Barros LS et al. ABCB1 haplotypes are associated with P-gp activity and affect a major molecular response in chronic myeloid leukemia patients treated with a standard dose of imatinib. Oncol Lett 2014; 7: 1313–1319.
Supplementary Information accompanies the paper on the The Pharmacogenomics Journal website (http://www.nature.com/tpj)
The Pharmacogenomics Journal (2014), 1 – 8
© 2014 Macmillan Publishers Limited
ORIGINAL ARTICLE
ABCB1 polymorphisms predict imatinib response in chronic myeloid leukemia patients: a systematic review and meta-analysis Q Zheng1, H Wu2, Q Yu3, DH (Dennis) Kim4, JH Lipton4, S Angelini5, S Soverini6, D Vivona7, N Takahashi8 and J Cao1 Imatinib mesylate, a competitive tyrosine kinase inhibitor, is considered the first-line therapy drug for Ph+ chronic myeloid leukemia (CML). Three single-nucleotide polymorphisms (SNPs) in the ATP-binding cassette, subfamily B (MDR/TAP), member 1 gene (ABCB1/MDR1), c.1236C4T, c.2677G4T/A and c.3435C4T, have been shown to affect cellular transport/metabolism of imatinib. The associations between these SNPs and imatinib response in CML patients have been widely evaluated, but the results were inconsistent. To derive a conclusive assessment of the associations, we performed a meta-analysis by combining data from a total of 12 reports including 1826 patients. The results showed that the 2677G allele or 3435T allele predicted a worse response to imatinib in CML patients, whereas 1236CC genotype was associated with better response in CML patients from Asian region. In conclusion, this meta-analysis suggests that c.1236C4T, c.2677G4T/A and c.3435C4T can be served as predictive markers for the therapeutical use of imatinib in CML patients. The Pharmacogenomics Journal advance online publication, 23 September 2014; doi:10.1038/tpj.2014.54
INTRODUCTION Chronic myeloid leukemia (CML) is a clonal myeloproliferative disorder of pluripotent hematopoietic stem cells that acquires a Philadelphia (Ph) chromosome encoding the BCR–ABL oncogenic fusion protein with constitutive tyrosine kinase activity.1 Use of imatinib mesylate, which is a competitive BCR–ABL tyrosine kinase inhibitor, is the first-line treatment of Ph+ CML.2 Imatinib has dramatically improved the clinical outcome and prognosis for CML patients.3 Despite the marked improvement in CML treatment, however, resistance exists in many patients.4 Several mechanisms for imatinib resistance in CML have been proposed, including heterogeneity in absorption, distribution and metabolism of imatinib.5 The ATP-binding cassette, subfamily B (MDR/TAP), member 1 gene (ABCB1/MDR1) encodes the cell surface P-glycoprotein (Pgp) acting as an energy-dependent multidrug efflux pump.6 Imatinib is a substrate of Pgp and the intracellular concentrations of imatinib were shown to be significantly lower in Pgp-expressing cells.7,8 The genetic polymorphisms in ABCB1 gene have the potential to alter the expression and function of Pgp and therefore influence imatinib disposition.9 The c.1236C4T, c.2677G4T/A and c.3435C4T are the most frequent three single-nucleotide polymorphisms (SNPs) in the coding region of ABCB1 and in linkage disequilibrium.9 The c.1236C4T (in exon 12 position 412) and c.3435C4T (in exon 26 position 1145) are synonymous SNPs, whereas the c.2677G4T/A in exon 21 results in the substitution of alanine in position 893 by serine (2677T) or threonine (2677A),
respectively.10 Early report has found that the functional variation in ABCB1 gene could explain at least in part variable responses to imatinib.9 Silent mutations in Pgp have recently been found to alter the substrate specificity that might be affecting the timing of cotranslational folding and insertion of Pgp into the membrane.10 In addition, there is evidence that the synonymous SNP c.3435C4T resulted in decreased levels of mRNA expression.11 Together, these results implicate that these three SNPs, though their exact role is yet undefined, may result in altered expression and function of Pgp. Mounting evidence has demonstrated that these SNPs, c.1236C4T, c.2677G4T/A and c.3435C4T, were associated with imatinib response in CML patients. Dulucq et al.12 firstly reported that patients with 1236TT genotype or 2677G allele achieved better major molecular response (MMR) at 12 months. However, among a larger cohort, they found no significant difference in genotype frequencies between patients with or without MMR was observed for the three polymorphisms.13 In addition, several other studies have analyzed the relationship between c.1236C4T, c.2677G4T/A, c.3435C4T and imatinib response, but the results were inconsistent. For instance, enhanced MMR and complete molecular response at 12 months significantly associated with the 1236CC genotype was reported by Deenik et al.,14 and patients carrying 3435TT and 2677TT showed lower probabilities to obtain a MMR and complete molecular response. The 3435CT/TT genotype and 2677T allele were inversely associated with primary failure of imatinib treatment.15 Takahashi et al.16 found no significant associations between the three polymorphisms and
1 Clinical Research Center, The Second Affiliated Hospital, School of Medicine, Zhejiang University, Hangzhou, China; 2Department of Neurosurgery, The Second Affiliated Hospital, School of Medicine, Zhejiang University, Hangzhou, China; 3Department of Surgical Oncology, The First Affiliated Hospital, School of Medicine, Zhejiang University, Hangzhou, China; 4Chronic Myelogenous Leukemia Group, Princess Margaret Cancer Centre, University Health Network University of Toronto, Toronto, Ontario, Canada; 5Department of Pharmacy and Biotechnology, University of Bologna, Bologna, Italy; 6Department of Experimental, Diagnostic and Specialty Medicine, Institute of Hematology, University of Bologna, Bologna, Italy; 7Departmento de Análises Clínicas e Toxicológicas da Faculdade de Ciências Farmacêuticas da Universidade de São Paulo, São Paulo, Brazil and 8 Department of Hematology, Nephrology, and Rheumatology, Akita University Graduate School of Medicine, Akita, Japan. Correspondence: Dr J Cao, Clinical Research Center, The Second Affiliated Hospital, Zhejiang University School of Medicine, 88 Jiefang Road, Hangzhou 310009, China, E-mail: [email protected] Received 1 April 2014; revised 18 July 2014; accepted 13 August 2014
ABCB1 polymorphisms predict imatinib response in CML Q Zheng et al
2 imatinib response among CML patients. Each single study with small sample size had insufficient power to conclusively determine whether these SNPs might be useful in response prediction to imatinib therapy in CML patients. In this meta-analysis, we aim to summarize the available evidence for the associations between ABCB1 polymorphisms and imatinib response in CML patients, and provide the highest level of evidence for their clinical significance up to date. MATERIALS AND METHODS Literature search strategy We performed a comprehensive literature search from Pubmed, ISI web of knowledge, Embase and Cochrane databases up to January 2014 using the following terms: ‘chronic myeloid leukemia’, ‘CML’, ‘MDR1’, ‘ABCB1’ and ‘imatinib’. The titles and abstracts of articles were screened to exclude clearly irrelevant studies, and the full texts of remaining articles were retrieved and independently assessed by two authors to determine whether they contained information on the topic of interest. The reference lists from the main reports and reviews were reviewed for additional relevant works.
Selection criteria and data extraction Eligible studies had to meet the following criteria: (1) the study with independent data has been published in peer-reviewed journals; (2) the study explored the impact of SNPs c.1236C4T, c.2677G4T/A, and c.3435C4T in ABCB1 on the response to imatinib in CML patients; (3) the study supplied enough information of genotype to calculate odds ratio (OR) with its 95% confidence interval. Studies with overlapping data and/or insufficient genotype data were excluded. Two investigators independently reviewed and extracted data with any discrepancies resolved by consensus. The information extracted included the first author’s surname, year of publication, country of origin, number of patients, age, disease phase, therapy, follow-up period, response criteria and the genotype counts in responders and non-responders. The corresponding authors of studies were contacted for additional data when necessary.
Statistical methods We examined the effect of each polymorphism on imatinib response in CML patients using (1) allele contrast, (2) recessive and (3) dominant models.17 Pooled ORs (95% confidence intervals) were calculated using fixed-effects (Mantel–Haenszel method)18 or random-effects (DerSimonian–Laird method).19 We chose the fixed-effects model when homogeneous effects were assumed; otherwise, the random-effects model was more appropriate. Subgroup analyses were performed according to geographic region (Europe, Asia, America or Africa), therapeutic approach (imatinib only or imatinib/+other drugs) and response criteria (MMR or other non-uniform criteria). Heterogeneity of the combined studies was assessed with Cochran’s Q-statistic test20 and I2-test.2 Although the P-value of Cochran’s Q-statistic is below 0.05, heterogeneity was considered statistically significant. The I2-test provides a measure of the degree of heterogeneity in the results. Typically, values of 0–25% are considered to represent no heterogeneity, 25–50% to modest heterogeneity, 50–75% to large heterogeneity and 75–100% to extreme heterogeneity.21 Sensitivity analysis sequentially omitting individual studies was performed to assess the stability of the results. The Egger’s test and funnel plots were used to identify publication bias.22 All of the above analyses were performed using STATA version 12.0 (StataCorp, College Station, TX, USA).
RESULTS On the basis of the above criteria, 12 eligible studies12–16,23–29 were included in this systematic review and meta-analysis involving a total of 1826 cases (Figure 1). These pharmacogenetic studies investigated the influence of the SNPs (c.1236C4T, c.2677G4T/A and c.3435C4T) in ABCB1 gene on response to imatinib in CML patients. Raw genotype data supplied by the corresponding authors from five studies have not been published.15,16,23,26,27 Table 1 describes the main characteristics of all included studies. Among them, five studies were conducted The Pharmacogenomics Journal (2014), 1 – 8
Figure 1.
Flowchart of the study selection process.
in the region of Europe,12–15,27 four in Asia,16,24,25,29 one in South America,26 one in Africa28 and the last one was conducted both in Asia and North America, which was divided into two groups.23 Patients in nine studies were treated with imatinib only at a dose of 300–800 mg per day.15,16,23–29 Patients in two studies following the treatment with imatinib/+interferon,12 with imatinib+cytarabine (AraC),14 respectively, were included in the group of imatinib/+other drugs. The last one was divided into two groups as the data were available.13 The response criterion was defined by MMR in six studies,12–14,16,23,27 among which five were followed at 12 months,12–14,23,27 one was not known.16 The results in the remaining six studies15,24–26,28,29 were based on other nonuniform criteria, which were also summarized in Table 1. ABCB1 c.1236C4T polymorphism and imatinib response in CML Overall, a total of 1623 patients from 11 studies were included for this polymorphism in this meta-analysis.12–16,23–27,29 As shown in Table 2, we found a significant association between the polymorphism and imatinib response in the Asia region under both dominant and recessive models. Figure 2 illustrates subsidiary analysis of geographic region under the dominant model. The presence of 1236CC genotype significantly increased response to imatinib in Asian CML patients (OR = 0.59, P = 0.036). We did not observe significant heterogeneity between these studies (I2 = 0.0%; P = 0.518). No publication bias was detected in these studies (Supplementary Figure S1). ABCB1 c.2677G4T/A polymorphism and imatinib response in CML The meta-analysis of 10 studies13–16,23–28 involving 1579 patients revealed no obvious association between ABCB1 c.2677G4T/A polymorphism and imatinib response in overall populations (Table 3). In the subgroup analyses, the results show several significant associations. Notably, the odds of the G allele carrier state was lower in responders versus non-responders in the subgroup of other non-uniform criteria under the allele contrast model (OR = 1.42, P = 0.031; Figure 3). We obtained similar results in this group under recessive model (Table 3). Conclusively, patients carrying the 2677T/A allele achieved better response with non-uniform response criteria. When stratified by geographic region, there was only one study carried out in Africa. Again the odds of the 2677 homozygous variant carrier state was © 2014 Macmillan Publishers Limited
ABCB1 polymorphisms predict imatinib response in CML Q Zheng et al
3 Table 1.
Characteristics of studies included in this meta-analysis
Reference
Dulucq et al.12 Kim et al.
23
Deenik et al.
2008
French
90
4–46
CP; AP
2010
416
17–87
CP; AP; BC
46
NA
CP
Imatinib 400 mg; imatinib 400 mg+IFN Imatinib 400, 600 or 800 mg Imatinib 800 mg+AraC
Follow-up period (months)
MMR
12
MMR
12
MMR
12
MMR
NA
MMR
2010 2010
French
557
NA
NA
2010
Japan
67
20–81
CP
2011
Spain
65
19–75
CP
Imatinib 400 mg; imatinib 400 mg+IFN; Imatinib 400 mg+AraC; imatinib 600 mg Imatinib 400 mg; imatinib other doses Imatinib 400 mg
2011
China
52
18–76
CP
Imatinib 400 mg
⩾12
2012
China
48
18–76
CP; AP
Imatinib 300–800 mg
105
2012
Brazil
118
18–80
CP
Imatinib 400 mg
18
2013
Italy
189
18–75
CP
Imatinib 400 or 800 mg
12
2013 2013
Egypt Korea
96 82
44 17–79
CP; AP; BC CP
Imatinib 400–600 mg Imatinib 400 mg
18 6
15
24 25 26
Angelini et al.
27
Elghannam et al. Seong et al.29
28
Response criteria
12
13
Maffioli et al.
Vivona et al.
Number of Age Disease phase Therapy patients (years)
Canada; Korea Nederland
Takahashi et al.16
Chen et al.
Country
14
Dulucq et al.
Ni et al.
Year
18
Genotype
c.1236C4T; c.2677G4T/A; c.3435C4T c.1236C4T; c.2677G4T/A; c.3435C4T c.1236C4T; c.2677G4T/A; c.3435C4T c.1236C4T; c.2677G4T/A; c.3435C4T
c.1236C4T; c.2677G4T/A; c.3435C4T Response c.1236C4T; c.2677G4T/A; c.3435C4T CR c.1236C4T; c.2677G4T/A; c.3435C4T CCR; MMR c.1236C4T; c.2677G4T/A; c.3435C4T CCR c.1236C4T; c.2677G4T/A; c.3435C4T MMR c.1236C4T; c.2677G4T/A; c.3435C4T CR; MR c.2677G4T/A CCR c.1236C4T; c.3435C4T
Abbreviations: AP, accelerated phase; AraC, cytarabine; BC, blastic crisis; CCR, complete cytogenetic response; CP, chronic phase; CR, cytogenetic response; IFN, interferon; MMR, major molecular response; MR, molecular response; NA, not applicable.
significantly increased in responders versus non-responders under recessive model (OR = 6.14, P = 0.007; Table 3). Sensitivity analysis and publication bias analysis showed no significant results (Supplementary Figure S2). ABCB1 c.3435C4T polymorphism and imatinib response in CML We compiled 11 studies12–16,23–27,29 that analyzed the association between ABCB1 c.3435C4T polymorphism and imatinib response in CML, nine presented enough data to allow meta-analysis with data from a total of 1373 patients.13,14,16,23–27,29 In overall comparison, significant association was observed under the allele contrast model (OR = 0.85, P = 0.042; Table 4) or dominant model (OR = 0.73, P = 0.011; Figure 4). The overall comparison under allele contrast model indicated that the presence of 3435T allele predicted a worse response to imatinib in CML patients (OR = 0.85, P = 0.042; Table 4). In subgroup analyses, there was no significant result among different regions. However, the group treated with imatinib/+ other drugs under the allele contrast model (OR = 0.72, P = 0.044; Table 4) or recessive model (OR = 0.58, P = 0.035; Table 4) and the group evaluated by MMR under dominant model (OR = 0.72, P = 0.021; Figure 4) achieved significant results. Sensitivity analysis and publication bias analysis showed no significant results (Supplementary Figure S3). DISCUSSION Imatinib is the first generation of tyrosine kinase inhibitor, which exerts a potent inhibitory activity against Ph+ CML. Imatinib have proven efficacious in the treatment of CML; however, a proportion of patients do not respond and/or exhibit resistance.30 To overcome the deficiency of imatinib, nilotinib and dasatinib have been developed and demonstrated remarkable efficacy in the induction of cytogenetic and molecular responses in newly diagnosed CML patients.31,32 Therefore, biomarkers predicting response to therapy would be helpful for making treatment decisions, in view of a personalized medicine. The pharmacokinetics of imatinib including absorption, distribution, metabolism and excretion may partly explain the © 2014 Macmillan Publishers Limited
interpatient variability in imatinib response.6 Pgp, encoded by ABCB1 gene, is an energy-dependent multidrug efflux pump mediating the efflux of imatinib.7 Overexpression of Pgp confer resistance to imatinib in leukemia cell lines,33,34 and Pgpexpressing cells had lower imatinib intracellular levels.8 The ABC transporter ABCG2 (ATP-binding cassette, subfamily G (WHITE), member 2), also known as breast cancer resistance protein, is another drug efflux pump that mediate imatinib efflux and has been implicated in imatinib resistance.35,36 ABCB1 is more likely involved in imatinib transport at clinically relevant concentrations, compared with ABCG2.35 Besides, the factors resulting in imatinib resistance include the difference in population pharmacogenetics.6,37 The three most common SNPs c.1236C4T, c.2677G4T/A and c.3435C4T in ABCB1 have been showed to be associated with imatinib pharmacokinetic characteristics. The imatinib clearance has been demonstrated to relate to the ABCB1 genotype and was least reduced in thymidine homozygotes at the c.1236C4T, c.2677G4T/A and c.3435C4T loci.9 Another report confirmed that patients with the 3435CT and TT genotypes had lower imatinib clearance compared with patients with 3435CC genotype.38 In addition, polymorphisms in ABCG2 may influence concentration and clinical response of imatinib in both solid and hematologic malignancies. Human embryonic kidney cells transfected with ABCG2 c.141Q4K exhibited greater imatinib accumulation in vitro in comparison with cells expressing wild-type ABCG2, but the clinical relevance of the observation is limited.39 Patients heterozygous for ABCG2 c.421C4A showed a 22% reduction in imatinib clearance compared with wild type.40 The GG genotype in ABCG2 c.34G4A was significantly associated with poor response to imatinib in CML patients.41 Another study reported that the ABCG2 c.421C4A variant allele was related to a higher rate of MMR in CML patients using imatinib.29 In 2008, Dulucq et al.12 firstly identified that c.1236C4T, c.2677G4T/A and c.3435C4T variants in ABCB1 were associated with imatinib response. Patients with 1236TT genotype achieved better response, whereas 2677G allele was associated with a worse response. In recent years, a number of reports have focused on the associations between those SNPs and imatinib response, The Pharmacogenomics Journal (2014), 1 – 8
ABCB1 polymorphisms predict imatinib response in CML Q Zheng et al
4 Table 2.
Stratified analyses for polymorphism c.1236C4T and imatinib response in CML
Analyses
Allele contrast model
Dominant model
Recessive model
OR (95% CI)
P(Z)
I2 (%)
OR (95% CI)
P(Z)
I2 (%)
OR (95% CI)
P(Z)
I2 (%)
Overall
0.86 (0.66, 1.12)
0.263
63.4
0.87 (0.70, 1.09)
0.235
26.1
0.93 (0.74, 1.18)
0.558
50.8
Region Europe Asia America
0.96 (0.66, 1.42) 0.61 (0.37, 1.00) 1.19 (0.87, 1.64)
0.849 0.051 0.282
67.2 59.0 0.0
0.91 (0.68, 1.23) 0.59 (0.36, 0.97) 1.14 (0.70, 1.87)
0.555 0.036 0.589
45.6 0.0 0.0
1.05 (0.75, 1.47) 0.61 (0.40, 0.93) 1.43 (0.82, 2.48)
0.788 0.022 0.207
46.7 55.0 0.0
Therapy Imatinib Imatinib/+other drug
0.82 (0.63, 1.06) 1.02 (0.41, 2.55)
0.129 0.965
49.8 85.4
0.86 (0.66, 1.11) 0.91 (0.59, 1.43)
0.253 0.690
0.0 77.6
0.85 (0.65, 1.12) 1.24 (0.76, 2.03)
0.250 0.387
42.0 75.2
Response criteria MMR Other criteria
0.98 (0.74, 1.29) 0.62 (0.33, 1.17)
0.865 0.139
61.1 65.8
0.95 (0.74, 1.21) 0.58 (0.33, 1.02)
0.672 0.058
34.7 0.0
1.04 (0.80, 1.36) 0.61 (0.36, 1.02)
0.767 0.058
30.9 64.2
Abbreviations: CML, chronic myeloid leukemia; CI, confidence interval; MMR, major molecular response; OR, odds ratio; Z, z-score in normal distribution model.
Figure 2. Forest plot of ABCB1 c.1236C4T polymorphism and imatinib response stratified by geographic region under the dominant model. Data from Dulucq et al.13 were divided into two groups, including patients treated with imatinib only and patients treated with imatinib and other drugs. Data from Kim et al.23 were divided into two groups, including patients from Asia and from America. OR41 indicates a better response to imatinib in CML patients carrying 1236CT and 1236TT genotypes. The vertical dash line represents OR = 0 and the vertical bold line represents OR = 1. The length of horizontal line for each study represents the 95% CI, the middle spot represents the value of OR and the area of the gray square box represents the proportion of weight. The length of diamond represents the combined 95% CI and the centra represents the combined OR. The overlap between the bold vertical line and each horizontal line or diamond represents that the result is not statistically significant. CI, confidence interval; CML, chronic myeloid leukemia; OR, odds ratio.
with inconsistent results. Recently, Zu et al.42 reported 1236TT/CT genotype exposures as having higher risk of imatinib resistance in Asian patients. In the current meta-analysis, we included much more cases with subgroup analyses in more detail among three genetic models, which may lead to a more reliable conclusion. Eleven studies analyzed the synonymous SNP c.1236C4T that is located in ABCB1 exon 12.12–16,23–27,29 One report found an The Pharmacogenomics Journal (2014), 1 – 8
association between the 1236TT genotype and increased imatinib response rate.12 However, three studies14,24,25 found the opposite result and the other seven studies13,15,16,23,26,27,29 found no association between c.1236C4T and imatinib response. This systematic review has shown that c.1236C4T was associated with imatinib response in CML patients in the region of Asia, which is consistent with a recent meta-analysis.42 Patients carrying the CC © 2014 Macmillan Publishers Limited
ABCB1 polymorphisms predict imatinib response in CML Q Zheng et al
5 Table 3.
Stratified analyses for polymorphism c.2677G4T/A and imatinib response in CML
Analyses
Allele contrast model
Dominant model
Recessive model
OR (95% CI)
P(Z)
I2 (%)
OR (95% CI)
P(Z)
I2 (%)
OR (95% CI)
P(Z)
I2 (%)
Overall
1.01 (0.87, 1.17)
0.874
40.9
1.05 (0.84, 1.31)
0.686
12.8%
0.97 (0.76, 1.25)
0.840
48.4
Region Europe Asia America Africa
0.94 0.98 1.09 1.65
1.15) 1.35) 1.50) 2.95)
0.538 0.894 0.618 0.090
54.2 50.1 0.0 NA
1.01 1.05 1.17 1.04
1.36) 1.83) 1.89) 2.41)
0.972 0.873 0.530 0.936
50.0 30.3 0.0 NA
0.81 (0.57, 1.15) 0.92 (0.56, 1.52) 1.04 (0.58, 1.88) 6.14 (1.65, 22.94)
0.242 0.749 0.888 0.007
26.8 56.4 0.0 NA
Therapy Imatinib Imatinib/+other drug
1.09 (0.92, 1.28) 0.77 (0.56, 1.06)
0.316 0.107
27.9 65.2
1.11 (0.86, 1.42) 0.86 (0.53, 1.39)
0.434 0.531
1.1 68.2
1.13 (0.85, 1.50) 0.56 (0.32, 0.97)
0.411 0.040
43.6 0.0
Response criteria MMR Other criteria
0.92 (0.78, 1.09) 1.42 (1.03, 1.96)
0.337 0.031
34.7 17.1
0.99 (0.77, 1.28) 1.25 (0.79, 1.95)
0.935 0.339
0.0 37.4
0.81 (0.61, 1.07) 2.30 (1.23, 4.31)
0.130 0.009
31.3 20.9
(0.76, (0.71, (0.79, (0.92,
(0.75, (0.60, (0.72, (0.45,
Abbreviations: CML, chronic myeloid leukemia; CI, confidence interval; MMR, major molecular response; NA, not applicable; OR, odds ratio; Z, z-score in normal distribution model.
Figure 3. Forest plot of ABCB1 c.2677G4T/A polymorphism and imatinib response stratified by response criteria under the allele contrast model. Data from Dulucq et al.13 were divided into two groups, including patients treated with imatinib only and patients treated with imatinib and other drugs. Data from Kim et al.23 were divided into two groups, including patients from Asia and from America. OR41 indicates a better response to imatinib in CML patients carrying 2677T/A allele. The vertical dash line represents OR = 0, and the vertical bold line represents OR = 1. The length of horizontal line for each study represents the 95% CI, the middle spot represents the value of OR and the area of the gray square box represents the proportion of weight. The length of diamond represents the combined 95% CI and the centra represents the combined OR. The overlap between the bold vertical line and each horizontal line or diamond represents that the result is not statistically significant. CI, confidence interval; CML, chronic myeloid leukemia; OR, odds ratio.
genotype of c.1236C4T were more likely to respond to imatinib as compared to the variant genotypes. However, the conclusion only pertained to population from Asia and could not be generalized to cohorts of other regions. The diversity may come from different frequencies of mutant alleles in different ethnic groups, and diverse gene–environment interactions leading from © 2014 Macmillan Publishers Limited
different life styles and environmental factors among the different populations.43,44 From the 11 reports that studied c.2677G4T/A with imatinib response,12–16,23–28 one study found an association between 2677G allele and increased imatinib response,13 whereas four found the opposite result.12,15,25,28 However, the other six reports The Pharmacogenomics Journal (2014), 1 – 8
ABCB1 polymorphisms predict imatinib response in CML Q Zheng et al
6 Table 4.
Stratified analyses for polymorphism c.3435C4T and imatinib response in CML
Analyses
Allele contrast model
Dominant model
Recessive model
OR (95% CI)
P(Z)
I2 (%)
OR (95% CI)
P(Z)
I2 (%)
OR (95% CI)
P(Z)
I2 (%)
Overall
0.85 (0.73, 0.99)
0.042
36.5
0.73 (0.57, 0.93)
0.011
36.3
0.85 (0.64, 1.11)
0.232
37.2
Region Europe Asia America
0.85 (0.69, 1.04) 0.73 (0.49, 1.07) 0.96 (0.70, 1.33)
0.106 0.104 0.803
25.5 66.3 0.0
0.73 (0.53, 1.01) 0.68 (0.39, 1.20) 0.76 (0.48, 1.23)
0.059 0.184 0.267
11.7 62.2 63.8
0.90 (0.64, 1.27) 0.64 (0.30, 1.35) 0.83 (0.46, 1.49)
0.551 0.241 0.536
56.7 36.0 54.3
Therapy Imatinib Imatinib/+other drug
0.90 (0.75, 1.07) 0.72 (0.53, 0.99)
0.236 0.044
33.0 58.7
0.77 (0.59, 1.02) 0.59 (0.35, 1.00)
0.065 0.052
38.4 56.0
0.99 (0.71, 1.37) 0.58 (0.34, 0.96)
0.939 0.035
33.2 0.0
Response criteria MMR Other criteria
0.89 (0.75, 1.06) 0.71 (0.50, 1.01)
0.188 0.057
2.0 61.6
0.72 (0.55, 0.95) 0.76 (0.47, 1.24)
0.021 0.273
14.1 63.7
0.95 (0.70, 1.28) 0.43 (0.21, 0.88)
0.730 0.021
32.0 9.2
Abbreviations: CML, chronic myeloid leukemia; CI, confidence interval; MMR, major molecular response; OR, odds ratio; Z, z-score in normal distribution model.
Figure 4. Forest plot of ABCB1 c.3435C4T polymorphism and imatinib response stratified by response criteria under the dominant model. Data from Dulucq et al.13 were divided into two groups, including patients treated with imatinib only and patients treated with imatinib and other drugs. Data from Kim et al.23 were divided into two groups, including patients from Asia and from America. OR41 indicates a better response to imatinib in CML patients carrying 3435CT and 3435TT genotypes. The vertical dash line represents OR = 0 and the vertical bold line represents OR = 1. The length of horizontal line for each study represents the 95% CI, the middle spot represents the value of OR and the area of the gray square box represents the proportion of weight. The length of diamond represents the combined 95% CI and the centra represents the combined OR. The overlap between the bold vertical line and each horizontal line or diamond represents that the result is not statistically significant. CI, confidence interval; CML, chronic myeloid leukemia; OR, odds ratio.
that analyzed this polymorphism did not find any association.14,16,23,24,26,27 In this meta-analysis, we observed that individuals with 2677T/A allele achieved better response with non-uniform response criteria. It is possible that even more discriminatory results would have been obtained if a comprehensive set of criteria were set including CR and MR. In conclusion, ABCB1 c.2677G4T/A does not seem to be a good marker of imatinib response. In the 11 studies that analyzed synonymous SNP in exon 26 (c.3435C4T),12–16,23–27,29 six studies found no association The Pharmacogenomics Journal (2014), 1 – 8
between c.3435C4T and imatinib response.12,13,16,23,27,29 In three studies, the authors reported a worse response to imatinib of 3435T allele,14,24,26 whereas the other two reported the opposite result.15,25 The current meta-analysis supported that 3435T allele predicted a worse response to imatinib in CML patients. Many factors can lead to inaccurate, inconsistent or misleading results in various studies. First, the sample size in eight studies was less than 100 cases.12,14–16,24,25,28,29 A study with low statistical power, which results from relatively small sample size, has a reduced chance of detecting a true effect.45 Second, the clinical © 2014 Macmillan Publishers Limited
ABCB1 polymorphisms predict imatinib response in CML Q Zheng et al
7 characteristics, such as patient ethnicity, previous therapy and disease phase, may lead to relatively high heterogeneity in each study. Patients included in five studies had different disease phase,12,13,23,25,28 and in four studies had received other drugs in addition to imatinib.12–14,16 Considering previous therapy, patients in five studies were newly diagnosed,13–15,27,29 in two studies were previously diagnosed,23,26 and other studies didn’t declare.12,16,24,25,28 Third, bias may come from patients selection process, evaluation process and censoring data. Furthermore, ABCB1 c.1236C4T, c.2677G4T/A and c.3435C4T form the most common haplotypes, and that contain the mutated alleles are associated with major structural modifications on Pgp.46 There is report that found significant association between ABCB1 haplotypes and Pgp activity in CML patients.47 In our review of the literature, we identified three studies in which the haplotypes were associated with imatinib response. One study has reported that the 1236C/2677G/3435C haplotype was statistically linked to less frequent MMR,12 whereas another found that 1236CT/ 2677GT/3435CT haplotype was positively associated with the MMR to imatinib in CML patients.26 In addition, the ABCB1 haplotype 1236T/2677G/3435C was more frequently found in patients primarily resistant to imatinib.15 Yet, there are some limitations in this meta-analysis. First, the relatively small sample size results in a lack of uniformly strong statistical power. Second, there exists clinical heterogeneity among the CML patients in this meta-analysis. The differences include clinical parameters such as disease duration, disease phase and use of concomitant medication. Finally, several different outcome measures were used to assess the outcomes from imatinib therapy. In conclusion, this systematic review and meta-analysis demonstrated that the probability of successful treatment with imatinib is influenced, at least in part, by c.1236C4T, c.2677G4T/ A and c.3435C4T variants in ABCB1 gene possessed by the CML patient. The presence of 1236CC genotype (in Asian population under both dominant and recessive models), 2677T/A allele (in patients evaluated by non-uniform response criteria under both allele contrast and recessive models) or 3435C allele (under both allele contrast and dominant models in overall comparison) marks a better response to imatinib in CML patients. The pharmacokinetics for imatinib in CML is not fully known, future detailed studies are needed to elucidate the molecular basis. CONFLICT OF INTEREST Dr Naoto Takahashi reports grants, personal fees from Novartis, Bristol-Myers Squibb and Pfizer, outside the submitted work. The remaining authors declare no conflict of interest.
ACKNOWLEDGMENTS We would like to thank Dr Francisco Cervantes from Hematology Department, Hospital Clínic, IDIBAPS, University of Barcelona, Barcelona, Spain for providing us their raw genotype data. Part of the work was supported by grant FAPESP, Brazil (#09/54184-0) to Dr Douglas Vivona.
REFERENCES 1 Groffen J, Stephenson JR, Heisterkamp N, de Klein A, Bartram CR, Grosveld G. Philadelphia chromosomal breakpoints are clustered within a limited region, bcr, on chromosome 22. Cell 1984; 36: 93–99. 2 Bhamidipati PK, Kantarjian H, Cortes J, Cornelison AM, Jabbour E. Management of imatinib-resistant patients with chronic myeloid leukemia. Ther Adv Hematol 2013; 4: 103–117. 3 Etienne G, Dulucq S, Nicolini FE, Morisset S, Fort MP, Schmitt A et al. Achieving deeper molecular response is associated with a better clinical outcome in chronic myeloid leukemia patients on imatinib front-line therapy. Haematologica 2014; 99: 458–464.
© 2014 Macmillan Publishers Limited
4 Rosti G, Castagnetti F, Gugliotta G, Palandri F, Baccarani M. Second-generation BCR-ABL inhibitors for frontline treatment of chronic myeloid leukemia in chronic phase. Crit Rev Oncol Hematol 2012; 82: 159–170. 5 Szakacs G, Varadi A, Ozvegy-Laczka C, Sarkadi B. The role of ABC transporters in drug absorption, distribution, metabolism, excretion and toxicity (ADME-Tox). Drug Discov Today 2008; 13: 379–393. 6 Diamond JM, Melo JV. Mechanisms of resistance to BCR-ABL kinase inhibitors. Leuk Lymphoma 2011; 52(Suppl 1):12–22. 7 Hegedus T, Orfi L, Seprodi A, Varadi A, Sarkadi B, Keri G. Interaction of tyrosine kinase inhibitors with the human multidrug transporter proteins, MDR1 and MRP1. Biochim Biophys Acta 2002; 1587: 318–325. 8 Widmer N, Colombo S, Buclin T, Decosterd LA. Functional consequence of MDR1 expression on imatinib intracellular concentrations. Blood 2003; 102: 1142. 9 Gurney H, Wong M, Balleine RL, Rivory LP, McLachlan AJ, Hoskins JM et al. Imatinib disposition and ABCB1 (MDR1, P-glycoprotein) genotype. Clin Pharmacol Ther 2007; 82: 33–40. 10 Kimchi-Sarfaty C, Oh JM, Kim IW, Sauna ZE, Calcagno AM, Ambudkar SV et al. A "silent" polymorphism in the MDR1 gene changes substrate specificity. Science 2007; 315: 525–528. 11 Wang D, Johnson AD, Papp AC, Kroetz DL, Sadee W. Multidrug resistance polypeptide 1 (MDR1, ABCB1) variant 3435C4T affects mRNA stability. Pharmacogenet Genomics 2005; 15: 693–704. 12 Dulucq S, Bouchet S, Turcq B, Lippert E, Etienne G, Reiffers J et al. Multidrug resistance gene (MDR1) polymorphisms are associated with major molecular responses to standard-dose imatinib in chronic myeloid leukemia. Blood 2008; 112: 2024–2027. 13 Dulucq S, Preudhomme C, Guilhot F, Mahon FX. Response Is there really a relationship between Multidrug Resistance Gene (MDR1) polymorphisms and major molecular response to imatinib in chronic myeloid leukemia?. Blood 2010; 116: 6145–6146. 14 Deenik W, van der Holt B, Janssen JJ, Chu IW, Valk PJ, Ossenkoppele GJ et al. Polymorphisms in the multidrug resistance gene MDR1 (ABCB1) predict for molecular resistance in patients with newly diagnosed chronic myeloid leukemia receiving high-dose imatinib. Blood 2010; 116: 6144–6145. 15 Maffioli M, Camos M, Gaya A, Hernandez-Boluda JC, Alvarez-Larran A, Domingo A et al. Correlation between genetic polymorphisms of the hOCT1 and MDR1 genes and the response to imatinib in patients newly diagnosed with chronic-phase chronic myeloid leukemia. Leuk Res 2011; 35: 1014–1019. 16 Takahashi N, Miura M, Scott SA, Kagaya H, Kameoka Y, Tagawa H et al. Influence of CYP3A5 and drug transporter polymorphisms on imatinib trough concentration and clinical response among patients with chronic phase chronic myeloid leukemia. J Hum Genet 2010; 55: 731–737. 17 Lu PH, Yang J, Li C, Wei MX, Shen W, Shi LP et al. Association between mitogenactivated protein kinase kinase kinase 1 rs889312 polymorphism and breast cancer risk: evidence from 59,977 subjects. Breast Cancer Res Treat 2011; 126: 663–670. 18 Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959; 22: 719–748. 19 DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986; 7: 177–188. 20 Cochran WG. The combination of estimates from different experiments. Biometrics 1954; 10: 101–129. 21 Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003; 327: 557–560. 22 Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997; 315: 629–634. 23 Kim DH, Kong JH, Byeun JY, Jung CW, Xu W, Liu X et al. The IFNG (IFN-gamma) genotype predicts cytogenetic and molecular response to imatinib therapy in chronic myeloid leukemia. Clin Cancer Res 2010; 16: 5339–5350. 24 Ni LN, Li JY, Miao KR, Qiao C, Zhang SJ, Qiu HR et al. Multidrug resistance gene (MDR1) polymorphisms correlate with imatinib response in chronic myeloid leukemia. Med Oncol 2011; 28: 265–269. 25 Chen WW, Meng FY, Zhong JS, Yin CX, Wang ZX. Effects of hOCT1 and ABCB1 gene on the efficacy of imatinib mesylate in chronic myelocytic leukemia. Zhonghua Yi Xue Za Zhi 2012; 92: 1405–1408. 26 Vivona D, Bueno CT, Lima LT, Hirata RD, Hirata MH, Luchessi AD et al. ABCB1 haplotype is associated with major molecular response in chronic myeloid leukemia patients treated with standard-dose of imatinib. Blood Cells Mol Dis 2012; 48: 132–136. 27 Angelini S, Soverini S, Ravegnini G, Barnett M, Turrini E, Thornquist M et al. Association between imatinib transporters and metabolizing enzymes genotype and response in newly diagnosed chronic myeloid leukemia patients receiving imatinib therapy. Haematologica 2013; 98: 193–200. 28 Elghannam DM, Ibrahim L, Ebrahim MA, Azmy E, Hakem H. Association of MDR1 gene polymorphism (G2677T) with imatinib response in Egyptian chronic myeloid leukemia patients. Hematology 2013; 19: 123–128.
The Pharmacogenomics Journal (2014), 1 – 8
ABCB1 polymorphisms predict imatinib response in CML Q Zheng et al
8 29 Seong SJ, Lim M, Sohn SK, Moon JH, Oh SJ, Kim BS et al. Influence of enzyme and transporter polymorphisms on trough imatinib concentration and clinical response in chronic myeloid leukemia patients. Ann Oncol 2013; 24: 756–760. 30 Milojkovic D, Apperley J. Mechanisms of resistance to imatinib and secondgeneration tyrosine inhibitors in chronic myeloid leukemia. Clin Cancer Res 2009; 15: 7519–7527. 31 Saglio G, Kim DW, Issaragrisil S, le Coutre P, Etienne G, Lobo C et al. Nilotinib versus imatinib for newly diagnosed chronic myeloid leukemia. N Engl J Med 2010; 362: 2251–2259. 32 Kantarjian H, Shah NP, Hochhaus A, Cortes J, Shah S, Ayala M et al. Dasatinib versus imatinib in newly diagnosed chronic-phase chronic myeloid leukemia. N Engl J Med 2010; 362: 2260–2270. 33 Mahon FX, Belloc F, Lagarde V, Chollet C, Moreau-Gaudry F, Reiffers J et al. MDR1 gene overexpression confers resistance to imatinib mesylate in leukemia cell line models. Blood 2003; 101: 2368–2373. 34 Assef Y, Rubio F, Colo G, del Monaco S, Costas MA, Kotsias BA. Imatinib resistance in multidrug-resistant K562 human leukemic cells. Leuk Res 2009; 33: 710–716. 35 Eadie LN, Hughes TP, White DL. Interaction of the efflux transporters ABCB1 and ABCG2 with imatinib, nilotinib, and dasatinib. Clin Pharmacol Ther 2014; 95: 294–306. 36 Burger H, van Tol H, Boersma AW, Brok M, Wiemer EA, Stoter G et al. Imatinib mesylate (STI571) is a substrate for the breast cancer resistance protein (BCRP)/ ABCG2 drug pump. Blood 2004; 104: 2940–2942. 37 Cortes JE, Egorin MJ, Guilhot F, Molimard M, Mahon FX. Pharmacokinetic/pharmacodynamic correlation and blood-level testing in imatinib therapy for chronic myeloid leukemia. Leukemia 2009; 23: 1537–1544. 38 Yamakawa Y, Hamada A, Nakashima R, Yuki M, Hirayama C, Kawaguchi T et al. Association of genetic polymorphisms in the influx transporter SLCO1B3 and the
39
40
41
42
43 44 45
46
47
efflux transporter ABCB1 with imatinib pharmacokinetics in patients with chronic myeloid leukemia. Ther Drug Monit 2011; 33: 244–250. Gardner ER, Burger H, van Schaik RH, van Oosterom AT, de Bruijn EA, Guetens G et al. Association of enzyme and transporter genotypes with the pharmacokinetics of imatinib. Clin Pharmacol Ther 2006; 80: 192–201. Petain A, Kattygnarath D, Azard J, Chatelut E, Delbaldo C, Geoerger B et al. Population pharmacokinetics and pharmacogenetics of imatinib in children and adults. Clin Cancer Res 2008; 14: 7102–7109. Kim DH, Sriharsha L, Xu W, Kamel-Reid S, Liu X, Siminovitch K et al. Clinical relevance of a pharmacogenetic approach using multiple candidate genes to predict response and resistance to imatinib therapy in chronic myeloid leukemia. Clin Cancer Res 2009; 15: 4750–4758. Zu B, Li Y, Wang X, He D, Huang Z, Feng W. MDR1 gene polymorphisms and imatinib response in chronic myeloid leukemia: a meta-analysis. Pharmacogenomics 2014; 15: 667–677. Hirschhorn JN, Lohmueller K, Byrne E, Hirschhorn K. A comprehensive review of genetic association studies. Genet Med 2002; 4: 45–61. Dick DM. Gene-environment interaction in psychological traits and disorders. Annu Rev Clin Psychol 2011; 7: 383–409. Button KS, Ioannidis JP, Mokrysz C, Nosek BA, Flint J, Robinson ES et al. Power failure: why small sample size undermines the reliability of neuroscience. Nat Rev Neurosci 2013; 14: 365–376. Fung KL, Gottesman MM. A synonymous polymorphism in a common MDR1 (ABCB1) haplotype shapes protein function. Biochim Biophys Acta 2009; 1794: 860–871. Vivona D, Lima LT, Rodrigues AC, Bueno CT, Alcantara GK, Barros LS et al. ABCB1 haplotypes are associated with P-gp activity and affect a major molecular response in chronic myeloid leukemia patients treated with a standard dose of imatinib. Oncol Lett 2014; 7: 1313–1319.
Supplementary Information accompanies the paper on the The Pharmacogenomics Journal website (http://www.nature.com/tpj)
The Pharmacogenomics Journal (2014), 1 – 8
© 2014 Macmillan Publishers Limited
