Multidrug-Related Protein 1 (MRP1) Polymorphisms rs129081, rs212090, and rs212091 Predict Survival In Acute Myeloid Leukemia

ABCB1 (=MDR1, multidrug resistance protein 1) single nucleotide polymorphisms (SNPs) were shown to have a significant impact on therapy outcome in patients with acute myeloid leukemia (AML). Furthermore, an independent significant impact on treatment response and patient survival of SNPs in the genes for ABCC4 (MRP4), ABCC5 (MRP5) and ABCC11 (MRP8) related SNPs has also been demonstrated. In contrast, therapeutic strategies trying to modulate the anthracycline efflux of these transporters have failed in most clinical trials so far. Recently, higher dosages of daunorubicin used during induction chemotherapy have been associated with a better outcome in certain subgroups of AML patients. Hence, in times of individual diagnostic genetic analyses available as point-of-care diagnostics, the goal of this study was to further investigate whether SNPs in ABC-transporter genes, which are responsible for anthracycline efflux, have an independent impact on treatment outcome.

DNA samples were obtained from bone marrow aspirates of 160 Caucasian patients with newly diagnosed AML as part of the prospective AML2003 trial (NCT00180102). The cohort solely consisted of patients with a normal karyotype, based on conventional G-banding, minimizing false results in case of gain or loss of chromosomal material. All patients received double induction chemotherapy with daunorubicin and cytarabine. After DNA extraction, quantitative real time PCR was performed, using a total of 49 SNP assays investigating SNPs of seven different ABC genes. The identification of the corresponding SNPs was performed in an in silico analysis using the NIH dbSNP database and HapMap while statistical univariate and multivariate analyses were performed using SPSS.

We detected three ABCC1 (MRP1) SNPs: rs129081 (CACCCC[C/G]ACTCCA), rs212090 (TTACTG[A/T]TCCCAC), and rs212091 (ACCTTA[A/G]AGAACA) with a significant influence on disease-free survival (DFS) or overall survival (OS), respectively. Patients carrying the homozygous rs129081 GG-SNP had a significant longer 5-year OS and 5-year DFS compared to the homozygous wildtype CC and heterozygous CG patients (OS: 68% [GG] vs. 40% [CC] vs. 64%, [CG], p=.035; DFS: 64% vs. 35% vs. 50%, p=.01). SNP rs212090 revealed a statistically significant difference in DFS when comparing homozygous alleles TT and AA (wildtype), 40% vs. 68%, p=.021. SNP rs212091 showed a significant difference concerning OS, with homozygous SNP GG leading to worse OS (0% vs. wildtype AA 64% vs. heterozygous AG 59%, p=.006). Again, there was a significant difference in DFS between both homozygous alleles AA (wildtype) and GG (55% vs. 0%, p=.018). Furthermore, there were no significant differences of standard clinical and laboratory baseline characteristics, FLT3-ITD mutation, or NPM1-mutation status, or chemotherapeutic toxicities. In order to exclude false positive findings of SNPs conferred as a result of leukemic transformation, we obtained saliva germline DNA from patients in complete remission who were treated by chemoconsolidation and performed a confirmatory analysis with the investigated SNPs, including rs129081, rs212090, and rs212091. Here, all SNPs were shown to be expressed in germline DNA in remission and bone marrow samples at diagnosis alike. The multivariate models for rs129081, rs212090 (TT), rs212091(AG), and rs212091(AA) revealed significances of p=.024, p=.029, p=.042, and p=.017 respectively for DFS but not for OS (except for rs212091[AA]). After adjustment for a false discovery rate of 5% still a trend towards the association of the SNPs and DFS could be seen. Therefore, more research is necessary to strengthen this evidence.

In this study we found a significant influence of rs129081, rs212090, and rs212091 SNPs (ABCC1, MRP1) on survival in AML in univariate analyses. Interestingly, these polymorphisms were not associated with other AML specific characteristics at diagnosis and were shown to be expressed in germline DNA and AML DNA alike. Hence, we suggest a prognostic effect of these SNPs which might be responsible for differential anthracycline susceptibility.

No relevant conflicts of interest to declare.