The Pharmacogenomics Journal https://doi.org/10.1038/s41397-018-0016-6 ARTICLE Associations between genetic polymorphisms of membrane transporter genes and prognosis after chemotherapy: meta-analysis and finding from Seoul Breast Cancer Study (SEBCS) 1 1 1 1 2 3 4 Ji-Eun Kim ● Jaesung Choi ● JooYong Park ● Chulbum Park ● Se Mi Lee ● Seong Eun Park ● Nan Song ● 5 6 4,7 8 1,4,9 10 Seokang Chung ● Hyuna Sung ● Wonshik Han ● Jong Won Lee ● Sue K. Park ● Mi Kyung Kim ● 4,7 9,11 1,4,9,12 1,4,9 Dong-Young Noh ● Keun-Young Yoo ● Daehee Kang ● Ji-Yeob Choi Received: 7 June 2017 / Revised: 13 October 2017 / Accepted: 4 December 2017 © Macmillan Publishers Limited, part of Springer Nature 2018 Abstract Membrane transporters can be major determinants of the pharmacokinetic profiles of anticancer drugs. The associations between genetic variations of ATP-binding cassette (ABC) and solute carrier (SLC) genes and cancer survival were investigated through a meta-analysis and an association study in the Seoul Breast Cancer Study (SEBCS). Including the SEBCS, the meta-analysis was conducted among 38 studies of genetic variations of transporters on various cancer survivors. 1234567890();,: The population of SEBCS consisted of 1 338 breast cancer patients who had been treated with adjuvant chemotherapy. A total of 7 750 SNPs were selected from 453 ABC and/or SLC genes typed by an Affymetrix 6.0 chip. ABCB1 rs1045642 was associated with poor progression-free survival in a meta-analysis (HR = 1.33, 95% CI: 1.07–1.64). ABCB1, SLC8A1, and SLC12A8 were associated with breast cancer survival in SEBCS (Pgene < 0.05). ABCB1 rs1202172 was differentially associated with survival depending on the chemotherapy (Pinteraction = 0.035). Our finding provides suggestive associations of membrane transporters on cancer survival. Introduction chemotherapy has been demonstrated in breast cancer sur- vival, the heterogeneity of response among patients still Breast cancer is the second most common cancer in Korean exists [3–5]. Such inter-individual differences can be elu- women, and an increasing trend has been seen in the mor- cidated by genetic variations of pharmacogenes that are tality of breast cancer [1, 2]. Although the effect of adjuvant involved in drug absorption, distribution, metabolism and excretion (ADME) [6]. Application of genome-wide association studies Electronic supplementary material The online version of this article (GWAS) in pharmacogenomics has been increased since (https://doi.org/10.1038/s41397-018-0016-6) contains supplementary 2007. Pharmacogenomic information could be used for material, which is available to authorized users. * Ji-Yeob Choi Institute, Bethesda MD USA [email protected] 7 Department of Surgery, Seoul National University College of Medicine, Seoul, Korea 1 Department of Biomedical Sciences, Seoul National University Graduate School, Seoul, Korea 8 Department of Surgery, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Korea 2 College of Pharmacy Chonnam National University, Gwangju, Korea 9 Department of Preventive Medicine, Seoul National University College of Medicine, Seoul, Korea 3 College of Pharmacy, Duksung Women’s university, Seoul, Korea 10 Division of Cancer Epidemiology and Management, National 4 Cancer Research Institute, Seoul National University, Cancer Center, Goyang, Korea Seoul, Korea 11 The Armed Forces Capital Hospital, Seongnam, Korea 5 Division for New Health Technology Assessment, National Evidence-based Healthcare Collaborating Agency, Seoul, Korea 12 Institute of Environmental Medicine, Seoul National University Medical Research Center, Seoul, Korea 6 Division of Cancer Epidemiology and Genetics, National Cancer J-E Kim et al. making practical decisions on drug prescriptions for an 346 records identified 472 records identified through PubMed searching through EMBASE searching individual patient [7, 8]. However, GWAS in pharmaco- (Jan, 2000 – Dec, 2016) (Jan, 2000 – Dec, 2016) genomics has potential limitations on sample size, replica- tion of findings and understanding of drug response 572 records remained after duplicates removed mechanisms [7, 9]. Previously, a two-stage GWAS on 274 records removed disease-free survival (DFS) in breast cancer stratified by Not original article or meta-analysis (n=107) tumor subtypes in the Seoul Breast Cancer Study (SEBCS) Unrelated topic (n=160) Not accessible (n=7) was conducted. Although two loci close to the methe- 298 records screened on the nyltetrahyfrofolate synthetase (MTHFS) and H3 histone base of abstract and title 206 records excluded protein (H3K27Ac) detected from GWAS were associated No chemotherapy fi or antitumor agent (n=47) with DFS in response to treatment, it is dif cult to under- Not for human study (n=35) No transporter genes (n=54) stand the mechanisms of inter-individual differences in drug Abstract without full-text (n=70) Full-text articles assessed response from the GWAS in SEBCS [10]. for eligibility (n=92) In this study, we mainly focused on drug transporters 55 full-text articles excluded involved in drug absorption and elimination. There are two No qualified data (n=55) major subfamilies of transporters: ATP-binding cassette 37 studies met (ABC) and solute carrier (SLC). The ABC transporters are the inclusion criteria efflux pumps that are involved in the movement of intra- The current study (SEBCS) cellular substrates using ATP hydrolysis energy, often included 38 studies included against the concentration gradient. The SLC transporters are in the meta-analysis involved in the uptake of substrates depending on con- centration gradient [11–13]. Impaired delivery of drug can Fig. 1 Flow chart summarizing the literature search and selection process for inclusion of studies result in poor absorption and, increased elimination of drug and can affect drug sensitivity due to genetic alterations of transporter genes [14]. (d) no chemotherapy or antitumor agent (n = 47); (e) not for The aim of this study was to investigate the association human study (n = 35); (f) no transporter gene (n = 54); (g) between genetic polymorphisms of ABC and SLC trans- abstract without full text (n = 70); (h) no qualified data (n porters and breast cancer survival among patients who had = 55). A total of 38 studies including SEBCS were con- been treated with adjuvant chemotherapy. First, we con- sidered eligible for the meta-analysis. All the reviewed ducted a systematic review and meta-analysis to summarize articles in the present study were limited to English lan- previously conducted genetic association studies. Second, guage reports (Fig. 1). we investigated the association between genetic variations of transporter genes and breast cancer survival among Data extraction and outcomes for meta-analysis patients who had received adjuvant chemotherapy based on GWAS in SEBCS. The following data were extracted: author, year, ethnicity, gene, variant, type of cancer, regimen of chemotherapy, type of outcome and estimates (hazard ratio (HR) and 95% Methods confidence interval (95% CI)) by three reviewers (JEK, CBP, and JYC). If one study was conducted separately by Search strategies and selection criteria for meta- the type of regimen or ethnicity, the study was separated as analysis (a) and (b) [15–19]. The outcomes were classified as overall survival (OS), A systematic literature search was performed in PubMed progression-free survival (PFS), disease-free survival and EMBASE from January 2000 to December 2016 using (DFS), and time-to-progression (TTP). OS is defined as the the following terms: “transporter” (“ABC” OR “SLC”) time from randomization to death from any cause. PFS is AND “cancer [title]” AND (“pharmacogenetics” OR defined as time from randomization until disease progres- “pharmacogenomics” OR “SNP”) AND (“therapy” OR sion or death from any cause. DFS is defined as the time “treatment” OR “drug” OR “medicine”) AND (“survival” from randomization until recurrence of tumor or death from OR “response” OR “outcome” OR “prognosis”). any cause [20]. TTP is defined as the time from treatment After removing duplicates in the electronic databases, the start to disease progression [21]. following study exclusion criteria was applied for 572 stu- The results were extracted according to the following dies: (a) not original article or meta-analysis (n = 107); (b) genetic models: additive, dominant, and codominant unrelated topic (n = 160); (c) not accessible articles (n = 7); (homozygote and heterozygote). In the additive model, the Associations between genetic polymorphisms of membrane transporter genes and prognosis after. risk of disease depends on increasing the number of the in the previous study [23]. SNPs were mapped to genes minor alleles. In the dominant model, genotypes including considering 10 kb upstream and downstream of the coding any minor allele contribute equally to the phenotype. In the region. We excluded the SNPs with minor allele frequency codominant model, the heterozygote is between a major (MAF)<1% and a P-value <10−4 for deviation from allele homozygote and minor allele homozygote. Hardy–Weinberg Equilibrium (HWE). Study population of SEBCS Statistical analysis As a multicenter case-control study, the SEBCS included A meta-analysis of reviewed studies was conducted in 4 040 incident breast cancer patients who were recruited various cancer patients with chemotherapy. The associa- from Seoul National University Hospital (SNUH), ASAN tions of genetic variants with each outcome (PFS and OS) Medical Center (AMC) and the National Cancer