Multidrug resistance-1 gene polymorphisms associated with treatment outcomes in de novo acute myeloid leukemia
Article first published online: 5 DEC 2005
Copyright © 2005 Wiley-Liss, Inc.
International Journal of Cancer
Volume 118, Issue 9, pages 2195–2201, 1 May 2006
How to Cite
Kim, D. H., Park, J. Y., Sohn, S. K., Lee, N. Y., Baek, J. H., Jeon, S. B., Kim, J. G., Suh, J. S., Do, Y. R. and Lee, K. B. (2006), Multidrug resistance-1 gene polymorphisms associated with treatment outcomes in de novo acute myeloid leukemia. Int. J. Cancer, 118: 2195–2201. doi: 10.1002/ijc.21666
- Issue published online: 21 FEB 2006
- Article first published online: 5 DEC 2005
- Manuscript Accepted: 4 OCT 2005
- Manuscript Received: 29 APR 2005
- Regional Technology Innovation Program of the Ministry of Commerce, Industry and Energy (MOCIE), The Republic of Korea. Grant Number: RTI04-01-01
- single nucleotide polymorphism;
- multidrug resistance-1 gene;
- acute myeloid leukemia
Multidrug resistance-1 (MDR-1) gene single nucleotide polymorphisms (SNPs) have been identified as associated with the treatment outcomes of acute myeloid leukemia (AML) in Caucasians; yet, similar evidence is lacking for Asian populations. A total of 101 AML patients were enrolled in the current study. Two MDR1 SNPs (C3435T and G2677T/A) were analyzed with PCR/RFLP assay. As regards C3435T polymorphism, C/C genotype was significantly correlated with lower functional P-glycoprotein (P-gp) activity in leukemic blasts (7.5%) compared with C/T (10.7%) or T/T genotype (19.9%, p = 0.029). In genotypic analyses, C/C at −3435 (p = 0.05) and G/G at −2677 (p = 0.04) were strongly associated with a higher probability of complete remission (CR). In addition, the 3-year event-free survival (EFS) was higher in G/G genotype at −2677 (60.6%) than nonG/G (21.9%; p = 0.0241), in C/C at −3435 was higher than nonC/C genotype (p = 0.0139), and was higher in GC haplotype homozygote (58.2%) than nonGC homozygote (22.6%; p = 0.0427). In a multivariate analysis, the group without GC haplotype showed worse EFS (p = 0.030), with unfavorable cytogenetic risk (p = 0.008). However, no differences were noted in overall survival according to the MDR1 SNPs (p = 0.491 for C3435T and p = 0.955 for G2677T/A). © 2005 Wiley-Liss, Inc.
In patients with acute myeloid leukemia (AML), drug resistance is mainly determined by the function of P-glycoprotein (P-gp) generated by the multidrug resistance (MDR) gene (mdr1). Proposed mechanisms for upregulating P-gp include altering the activity of the transcription factor,1 gene rearrangement,2 or hypomethylation of the mdr1 promoter region.3 Recently, systematic screens of the mdr1 gene have identified multiple single nucleotide polymorphisms (SNPs), some of which appear to be associated with an altered P-gp function and expression based on structural variations of the mdr1 gene.4 Among the already identified 29 kinds of MDR1 gene SNPs,5 the G2677T SNP at exon 21 was found to lead most frequently to a change in the amino acid sequence, thereby influencing the function of P-gp. In addition, linking with C2677T SNPs, the C3435T at exon 26 was also identified as associated with the function of P-gp.4, 6
To date, the MDR1 gene SNP has already been known to be associated with the treatment outcomes of AML in Caucasians.7 However, the mdr1 gene is not only known to have an influence on the drug efflux of leukemic blasts, but has also been related to the metabolism of various kinds of drugs, such as calcineurin inhibitors (i.e. cyclosporine, FK-506) or corticosteroids. As such, the impact of the MDR1 gene SNPs would seem to differ depending on the clinical situation. For example, in a conventional chemotherapy setting, the MDR1 gene SNPs may be correlated with the chemotherapeutic response in the leukemic blasts, while in a stem cell transplantation (SCT) setting, the toxicity or effects of intensive immunosuppression on the SCT outcomes may be influenced by the MDR1 gene SNPs through the pharmacokinetic effects.
Accordingly, the current study attempted to evaluate the association between the MDR1 gene SNPs and the treatment outcomes, including the results after SCT, in de novo AML patients of Korean ethnicity. In addition, the correlation of the MDR1 gene SNPs with the functional P-gp activity in the leukemic blasts was also investigated.
Material and methods
Patient characteristics and treatment
A total of 101 AML patients newly diagnosed at Kyungpook National University Hospital between December 1996 and September 2004 were identified, among whom 6 patients with M3 subtypes were excluded from the study, while 14 other patients were not included in the treatment program for reasons of age (n = 10), poor performance (n = 3), and treatment refusal (n = 1). Finally, 81 patients were enrolled into the current study.
The patient characteristics of 81 patients were as follows: the median age was 39.0 years (range 15–72 years) and the male to female ratio was 54:46 (n = 44:37). FAB classification then revealed the following: M0, 5 patients (6%); M1, 4 (5%); M2, 45 (55%); M4, 8 (10%); M5, 8 (10%); M6, 5 (6%); M7, 3 (4%); unclassified type, 3 (4%). The favorable, intermediate and unfavorable cytogenetic groups included 18 (23%), 48 (62%), and 12 patients (15%), respectively. The laboratory findings at presentation were as follows: peripheral white cell counts, 40.8 ± 6.3 × 109/l (mean ± standard error), percentage of peripheral and marrow blasts, (41.5 ± 3.2)% and (55.6 ± 2.3)%. The internal tandem duplication of FLT3 (FLT3/ITD) was observed in 12 patients (14.8%).
The patients were treated with standard induction chemotherapy (Idarubicin 12 mg/m2/day IV for 3 consecutive days and Cytarabine 100 mg/m2/day IV for 7 consecutive days). In the case of patients over 55 years, the dose of Idarubicin was modified to two thirds of the total dose. Among the 81 patients who received remission–induction chemotherapy, 5 patients (6%) died of early treatment-related mortality (sepsis, n = 3; bleeding n = 1; DIC, n = 1) before a follow-up marrow examination. The CR rate for the remaining 76 evaluable cases was 74% (n = 56).
Consolidation therapies were performed based on 2 more cycles of high-dose Cytarabine (3 g/m2/day IV twice a day on day 1, 3 and 5). Allogeneic or autologous transplantation was recommended based on donor availability, primary treatment failure, and cytogenetic risk. Those patients who were not enrolled on the SCT program were given 1 additional cycle of high-dose Cytarabine.
Overall, 51 patients were enrolled on the transplantation program, where the 44 allogeneic recipients were in a state of first CR (n = 29), third CR (n = 1), first relapse (n = 6), or primary refractoriness (n = 8), while the 7 autologous transplant recipients were all in a state of first CR. The transplant procedures were conducted as previously described.8, 9, 10 The conditioning regimens for the allogeneic recipients included busulfan/cytoxan (BUCY) containing myeloablative (n = 39) or fludarabin/busulfan-containing reduced intensity conditioning (n = 5), and for the autologous recipients, all patients were conditioned with a BUCY regimen (n = 7).
MDR1 genotyping and functional P-gp assays
Mononuclear cells (MNCs) were isolated from the patients' marrow samples at diagnosis, using Histopaque®-1077 (Sigma Diagnostics, St. Louis, MO). A functional P-gp assay was then performed as previously described.11, 12 In brief, instead of a rhodamine-123 efflux assay, a daunorubicin intracellular accumulation assay was adopted and the mean fluorescence distribution (MFI) percentage were measured using an FACSCalibur™ flow cytometer (Becton Dickinson, San Jose, CA) after CD45 gating. The functional P-gp activity (Daunorubicin accumulation percentage) was then calculated using the following formula: functional P-gp activity (%) = 100 × MFI60min/MFI60min/verapamil.
For the MDR1 genotyping, the genomic DNA was extracted from patients' peripheral blood using a Wizard genomic DNA purification kit (Promega, Madison, WI). Two MDR1 gene polymorphisms were detected based on a polymerase chain reaction (PCR), using primers amplifying a short fragment of DNA containing the polymorphic sites. The PCR primers for the G2677T polymorphism in exon 21 were 5′-TGC AGG CTA TAG GTT CCA GG-3′ and 5′-TTT AGT TTG ACT CAC CTT CCC G-3′, generating a 224 bp fragment, while those for the G2677A polymorphism were 5′-AGA GCA TAG TAA GCA GTA GG-3′ and 5′-GTT TTG CAG GCT ATA GGT TC-3′, generating a 304 bp fragment. Plus, the PCR primers for the C3435T polymorphism in exon 26 were 5′-GCT GCT TGA TGG CAA AGA AA-3′ and 5′-ATT AGG CAG TGA CTC GAT GAT GA-3′, generating a 208 bp fragment. The PCR reactions were performed in a 20 μl reaction volume containing 100 ng of genomic DNA, 10 pmol of each primer, 0.2 mM of each deoxynucleotide triphosphate, 1 × a PCR buffer [50 mM KCl and 10 mM Tris-HCl (pH 8.3)], 1.5 mM MgCl2, and 1 unit of Taq polymerase (Takara Shuzo Co., Otsu, Shiga, Japan). The PCR program for exon 21 consisted of 35 cycles at 94°C for 30 sec, 58°C for 30 sec, 72°C for 30 sec, and a final elongation step at 72°C for 10 min. The PCR program for exon 26 consisted of 35 cycles at 94°C for 30 sec, 54°C for 30 sec, 72°C for 30 sec, and a final elongation step at 72°C for 10 min. The PCR products were checked on a 1.5% agarose gel, photographed using Polaroid film, and then subjected to an RFLP analysis.
To distinguish the SNPs, the restriction enzyme BanI (New England BioLabs, Beverly, MA) was used for G2677T, BseYI (New England BioLabs, Beverly, MA) for G2677A, and DpnII (New England Biolabs, Beverly, MA) for C3435T. Each 5 μl of the PCR products was digested overnight with 5 units of BanI (G2677T in exon 21), BseYI (G2677A in exon 21), and DpnII (exon 26) at 37°C. The digestion products were separated on a 6% acrylamide gel. Selected PCR-amplified DNA samples were then examined by DNA sequencing to confirm the genotyping results. The current study was approved by the institutional research board of Kyungpook National University Hospital and each patient gave written informed consent.
Cytogenetic risk group
Cytogenetic studies on the marrow samples at presentation were performed using standard G-banding with trypsin-Wright's staining after unstimulated short-term (24 hr) cultures. The criteria used to describe a cytogenetic clone and the karyotype followed the recommendations of the International System for Human Cytogenetic Nomenclature.13 At least 20 BM metaphase cells were analyzed in the patients designated as having a normal karyotype.
The cytogenetic risk groups were classified according to the MRC10 criteria, as described previously14; Favorable risk: AML associated with t(8;21), t(15;17) or inv(16), unfavorable risk: the presence of a complex karyotype, −5, del(5q), −7 or abnormalities of 3q and intermediate risk: the remaining group of patients, including those with 11q23 abnormalities, +8, +21, +22, del(9q), del(7q) or other miscellaneous structural or numerical defects not encompassed by the favorable or unfavorable risk groups.
Complete remission (CR) was defined as the presence of no more than 5% blast cells in the bone marrow aspirates and a trephine bone marrow biopsy and have an absolute neutrophil count of more than 1 × 109/l and platelets of >100 × 109/l without evidence of extramedullary leukemia. Relapse was defined as marrow infiltration by more than 5% blast cells in previously normal bone marrow or evidence of extramedullary leukemia.
In terms of the survival analysis, event-free survival (EFS) was defined as the interval from the date of treatment to the date of a confirmed relapse, persistence of the disease or death from any cause. Overall survival (OS) was defined as the length of time from the date of diagnosis to the date of death from any cause. As regards the survival analysis with censoring patients at the time point of transplantation, EFS and OS were estimated from the date of treatment to the time point of censoring, i.e. transplantation date, and denominated as EFS-r and OS-r to clarify the terminology.
The results were analyzed according to information available as of December 2004. The medical records of the patients, including the age, white blood cell (WBC) counts, platelet counts, peripheral and marrow blast percentage and level of serum-lactate dehydrogenase (LDH), were reviewed. The patients were then divided into 2 groups according to each clinical parameter (Age: < vs ≥50, WBC < vs ≥50 × 109/l, marrow blast percentage < vs ≥50% and LDH < vs ≥1,000 IU/l). The univariate associations between CR and the individual clinical parameters were analyzed using a χ2 test, Fisher's exact test or Mann-Whitney's U test.
The Hardy-Weinberg equilibrium for the MDR1 SNPs was tested using a χ2 analysis, while the haplotypes were determined based on a Bayesian algorithm using the Phase program15 (available at http://www.stat.washington.edu/stephens/phase.html), as previously studied.16
The frequencies of the genotypes or haplotypes were compared using a χ2test or Fisher's test in terms of the clinical characteristics or achievement of CR. A multiple logistic regression analysis was performed to identify independent predictive factors related to the achievement of CR, using such variables as the WBC count, age, serum LDH level, cytogenetic risk, FLT3/ITD status and MDR1 SNPs, and calculating the Odds ratio (OR) and 95% confidence interval (CI) for the relative risks. The functional P-gp activity according to the genotype was compared using an ANOVA test.
The EFS and OS estimates were calculated using the method of Kaplan and Meier, and the differences between the EFS or OS rates were compared using a log-rank test. Multivariate survival analyses were performed using Cox's proportional hazard model to define the prognostic factors for EFS and OS using a step-up procedure until the p-value for the likelihood ratio test was >0.05. The following variables were included for the analyses: the MDR1 SNPs, age, WBC count, serum LDH level, cytogenetics, FLT3/ITD status and use of stem cell transplantation for EFS and OS. The hazard ratio (HR) and 95% CI were also estimated. In multivariate analyses, all data were available except for 3 cases lacking cytogenetic data, which was substituted as “intermediate” cytogenetic risk.
A cut off p-value of 0.05 was adopted for all statistical analyses. The statistical data were obtained using an SPSS software package (SPSS 11.5, Chicago, IL).
Frequency of MDR1 gene polymorphisms and functional P-gp assay
The allele frequency of G-, T- and A-allele for G2677T/A was 48.1, 37.7 and 14.2%, while those of C- and T-allele for C3435T was 63.6 and 36.4%, respectively. The frequencies of the MDR1 genotypes at exon 21 and 26 are summarized in Table I. According to the MDR1 SNPs, no significant differences were found in the marrow blast cells, peripheral WBC count, serum LDH level or cytogenetic risk. Among the 6 haplotypes, which were identified by a linkage disequilibrium analysis, 3 types were predominant (GC, TT and AC) based on the following frequencies: GC, 45.7%; TT, 33.3%; AC, 13.6%; TC, 4.3%; GT, 2.5% and AT, 0.6%.
|Genotypes||Age (yr)||Cytogenetics (%)||p-value||Functional MDR assay (%, mean ± SE) (n = 69)||p-value|
|G/G (n = 21, 25.9%)||38.0||7 (33)||12 (57)||2 (10)||0.619||10.04 ± 3.20||0.410|
|G/T (n = 28, 34.6%)||40.5||5 (20)||17 (68)||3 (12)||19.19 ± 4.30|
|G/A (n = 8, 9.9%)||36.5||2 (25)||5 (63)||1 (12)||4.97 ± 2.28|
|T/T (n = 11, 13.6%)||36.0||3 (28)||4 (36)||4 (36)||13.49 ± 5.44|
|T/A (n = 11, 13.6%)||40.0||0 (0)||9 (82)||2 (12)||9.94 ± 4.45|
|A/A (n = 2, 2.5%)||25.5||1 (50)||1 (50)||0 (0)||7.30 ± 7.30|
|C/C (n = 32, 39.5%)||37.0||9 (28)||19 (59)||4 (13)||8.03 ± 2.19||0.037|
|C/T (n = 39, 48.1%)||43.0||8 (22)||25 (70)||3 (8)||12.39 ± 2.56|
|T/T (n = 10, 12.3%)||35.0||1 (10)||4 (40)||5 (50)||21.85 ± 6.33|
MDR1 gene polymorphisms and CR rates
In terms of the MDR1 genotypes, the CR rate for the patients with 2 G-alleles at exon 21 (G2677T/A) was 91%, while the CR rate for those with one G-allele or without G-allele was 58 and 67%, respectively (p = 0.04, Table II). The CR rate for the patients with a C/C homozygote at exon 26 (C3435T) was 84%, while the rate for those with a C/T heterozygote or T/T homozygote was 59 and 60%, respectively (p = 0.05, Table II). In addition, the CR rate was significantly different in favor of C/C homozygote at exon 26 (C3435T) (p = 0.029, p = 0.043 by Fisher's exact test) and of G/G homozygote at exon 21 (G2677T/A) (p = 0.021, p = 0.025 by Fisher's exact test) in the current study.
|Genotype||CR (%)||Refractory/TRM||p-value||3-year EFS||3-year OS|
|G/G (n = 21, 26%)||19 (91)||1/1||0.04||60.6 ± 11.6||0.07521||34.5 ± 13.9||0.9210|
|G/TA (n = 36, 44%)||21 (58)||11/4||18.6 ± 8.6||44.0 ± 9.6|
|Others (n = 24, 30%)||16 (67)||8/0||26.3 ± 10.9||44.4 ± 12.7|
|C/C (n = 32, 40%)||27 (84)||3/2||0.05||42.6 ± 10.5||0.01392||32.5 ± 12.5||0.4434|
|C/T (n = 39, 48%)||23 (59)||13/3||22.5 ± 9.3||45.6 ± 9.1|
|T/T (n = 10, 12%)||6 (60)||4/0||14.3 ± 13.2||17.1 ± 15.6|
|2 GC haplotypes (n = 19, 24%)||17 (90)||1/1||0.09||58.2 ± 12.1||0.1073||36.5 ± 14.6||0.9674|
|1 GC haplotype (n = 35, 43%)||22 (63)||9/4||21.8 ± 9.7||35.4 ± 12.7|
|0 GC haplotype (n = 27, 33%)||17 (63)||10/0||22.5 ± 9.6||36.8 ± 11.4|
|2 GC haplotypes (n = 19, 24%)||17 (90)||1/1||0.03||58.2 ± 12.1||0.0427||36.5 ± 14.6||0.9317|
|0–1 GC haplotype (n = 62, 76%)||39 (63)||19/4||22.6 ± 6.9||35.8 ± 9.1|
In a haplotypic analysis, the patients with GC haplotype homozygote at exon 21/26 (G2677T/A and C3435T) showed a higher CR rate (90%) compared with those of GC haplotype heterozygote (63%) or without GC haplotype homozygote (63%, p = 0.09; Table II). This difference was statistically significant when analyzed between GC haplotype homozygote and nonGC haplotype homozygote (p = 0.03).
A further analysis was also conducted to identify any other predictive factors for the achievement of CR in addition to the MDR1 SNPs, resulting in the following results: a WBC count of more than 50 × 109/l (p = 0.011, p = 0.019 by Fisher's exact test) and an unfavorable cytogenetic risk (p = 0.042, p = 0.064 by Fisher's exact test) were found to be unfavorable risk factor for the achievement of CR, but not the serum LDH level or age at presentation were. In a multiple logistic regression analysis, the patients with a GC haplotype homozygote showed a higher probability of achieving CR (p = 0.050, OR 0.193, 95% CI 0.037–1.002), and a high peripheral WBC count was also associated with a lower probability of CR (p = 0.015, OR 4.505, 95% CI 1.343–15.105, Table III).
|Risk factor||HR [95% CI]||p-value|
|MDR1 gene SNPs||GC haplotype homozygote||1.0||0.050|
|NonGC haplotype homozygote||5.183 [0.998–26.917]|
|WBC counts||Less than 50 × 109/l||1.0||0.015|
|More than 50 × 109/l||4.505 [1.343–15.105]|
|MDR1 gene SNPs||GC haplotype homozygote||1.0||0.030|
|NonGC haplotype homozygote||2.455 [1.088–5.539]|
|FLT3/ITD||Absence of FLT3/ITD||1.0||0.001|
|Presence of FLT3/ITD||3.521 [1.672–7.407]|
MDR1 gene polymorphisms and survival
With a median follow-up duration of 13.6 months (range 0.5–102 months), the 3-year EFS and OS rates were estimated as (30.8 ± 6.6)% and (40.9 ± 6.7)%, respectively. According to the MDR1 genotypes, EFS was significantly different in favor of the patients with 2 G-alleles at exon 21 (G2677T/A; 60.6% ± 11.6%) than those with 1 G-allele (18.6% ± 8.6%) or without G-allele (26.3% ± 10.9%, p = 0.0752, Table II and Fig. 1a; p = 0.0241 when comparing between). This difference was statistically significant when analyzed between the patients groups with and without G/G homozygote, Fig. 1b). The 3-year EFS rate for the patients with a C/C homozygote at exon 26 (C3435T) was calculated as (42.6 ± 10.5)%, while the rate for those with a C/T heterozygote or T/T homozygote was (22.5 ± 9.3)% and (14.3 ± 13.2)%, respectively (p = 0.0139, Table II and Fig. 1c). However, in terms of OS, there were no differences among the patients according to each genotype.
In the haplotypic analysis, the patients with a GC haplotype homozygote at exon 21/26 showed a higher EFS rate (58.2% ± 12.1%) compared with those without a GC haplotype homozygote (22.6% ± 6.9%, p = 0.0427, Table II, Figs. 2a and 2b). However, in terms of OS, no difference was observed among the patients according to each genotype (p = 0.491 for C3435T and p = 0.955 for G2677T/A) or haplotype (p = 0.979), as well.
In a multivariate survival analysis using Cox's proportional hazard model, the patients without a GC haplotype homozygote exhibited a lower EFS than those with a GC haplotype (p = 0.030, HR 2.455 [1.088–5.539]), together with the patients with an unfavorable cytogenetic risk (p = 0.008, HR 2.388 [1.252–4.555], Table III). However, in terms of OS, cytogenetic risk and FLT3/ITD status were found to be independent prognostic factors in the current study.
MDR1 gene polymorphisms and survival before transplantation
To exclude the effect of transplantation procedure on the survival, survival analysis censoring at the time point of transplantation was performed. The 2-year EFS-r and OS-r rates were estimated as (36.4 ± 9.9)% and (60.2 ± 10.6)%, respectively.
With respects to the G2677T/A genotype at exon 21, the 2-year EFS-r rate for the patients with and without a G/G homozygote was (68.4 ± 14.8)% and (20.1 ± 11.5)%, respectively (p = 0.018, Fig. 2a), while the 2-year OS-r rate was not different (p = 0.522). Meanwhile, for the C3435T genotype at exon 26, the 2-year EFS-t rate for the patients with C/C, C/T or T/T genotype was estimated as (59.7 ± 14.6)%, (15.6 ± 13.2)% and (0 ± 0)%, respectively (p = 0.001, Fig. 2b), while the 3-year OS-r rate was not different (p = 0.214). In the haplotypic analysis, 2 alleles of GC haplotype at exon 21/26 showed a higher 2-year EFS-r rate (64.7% ± 16.5%) compared with those with 1 allele of or without GC haplotype (41.9% ± 18.2% and 0% ± 0%, p = 0.028, Fig. 2c). However, in terms of OS, no difference was observed (p = 0.510), as well.
MDR1 SNPs and functional P-glycoprotein activity
Data on the functional P-gp activity were available for 69 patients (85.2%) out of 81 patients. For these patients, the mean percentage of daunorubicin efflux was (10.8 ± 1.7)% (range 0–51.0%), and 39.1 and 18.9% of the patients exhibited a functional P-gp activity of more than 10 and 20%, respectively. With respect to the association of P-gp activity with CR rates, although statistically insignificant (p = 0.487), the median percentage of functional P-gp activity was (8.1 ± 2.1)% for the patients that achieved CR and (11.8 ± 3.8)% for those without CR.
As regard to the association of MDR1 SNPs with functional P-gp assay, a significant correlation was found between the MDR1 genotype and the MDR functional activity. Interestingly, a significant correlation was noted between C3435T genotype and the functional P-gp activity. The patients with a C/C homozygote exhibited a lower P-gp activity (7.5% ± 2.1%; mean ± standard error) than those with a C/T heterozygote or T/T homozygote (10.7% ± 2.3% and 19.9% ± 6.1%, p = 0.029; Fig. 3). But, as regard to the G2677T/A genotype, no difference was observed (p = 0.181).
When using a cutoff value of 20% for positive functional P-gp activity, only 2 out of the 27 patients (7%) with a C/C homozygote were positive, while 6/23 (21%) and 5/11 (46%) were positive with a C/T heterozygote or T/T homozygote, respectively (p = 0.020, 0.022 by Fisher's exact test).
The current study raises 2 important issues related to MDR1 gene SNPs in the treatment of AML. First, the MDR1 gene SNPs were found to be associated with an achievement of CR and EFS in AML patients. However, second, they were not found to be associated with OS.
Various factors are associated with the prognosis of de novo AML, including the cytogenetic risk,17 FLT3 mutation,18, 19 MDR or performance of SCT. Recently, interest has increased in the role of genetic background in the prognosis and treatment outcomes of AML. In particular, systematic screens of the mdr1 gene have identified multiple SNPs, some of which appear to be associated with an altered transporter function and expression,4 thereby affecting the metabolism and disposition of chemotherapeutic agents.
A previous study by Illmer et al.7 that investigated 3 MDR1 SNPs at exons 12, 21 and 26 (positions 1236, 2677 and 3435) revealed the finding that the genotype at exon 26 (position 3435) and haplotype were associated with OS and the probability of relapse in the Caucasian AML population (n = 405). Similar outcomes were also observed in the current study, where a C/C genotype at exon 26 (position 3435, Fig. 1c) and G/G genotype at exon 21 (position 2677, Fig. 1b) were associated with better EFS (Table I). A haplotypic analysis also confirmed that the patients with a GC haplotype at exon 21/26 (positions 2677/3435) had better EFS (Fig. 2), suggesting that MDR1 SNPs could be used to predict the treatment outcomes and prognosis in the treatment of AML, at least among the Korean population.
However, there were some discrepancies between the results from the current and previous study. Whereas a C/C-genotype at exon 26 (position 3435) was found to be associated with a better EFS in the current study, the genotype was associated with worse survival in the study by Illmer et al.7 As well, the GC haplotype (position 2677/3435) was associated with better survival in the current study, yet worse survival in Illmer et al.'s study.7 The following may account for the differences between the results of the 2 studies. The discrepancies could be explained by ethnic differences between Caucasians and Koreans as regards MDR1 processing, including transcriptional initiation and RNA maturation.20 For instance, in a study of small cell lung cancer in Koreans, a C/C genotype at exon 26, C/C genotype at exon 21 and GC haplotype were found to be associated with a better response to chemotherapy, suggesting that these genetic markers could be used as a wide marker to predict treatment failure and survival among the Korean population (Jae Yong Park, personal communication).
In the current study, the correlation between the C3435T MDR1 SNP and EFS is presented in a dose-dependent manner (Fig. 1c), and definite correlation was also found between the C3435T MDR1 SNP and the functional P-gp activity of the leukemic blasts (Fig. 3). Furthermore, a C/C genotype at exon 26 was significantly associated with a lower expression of P-gp and better EFS. As such, one interpretation of these findings is that a C/C genotype at exon 26 (3435) is a strong predictor of the achievement of CR based on a reduced P-gp expression, thereby increasing the intracellular accumulation of chemotherapeutic agents. Conversely, a reduced intracellular accumulation of chemotherapeutic agents attributable to the action of P-gp in the leukemic blasts could be related to chemoresistance and treatment failure in AML patients without a C/C genotype at exon 26. However, influence of MDR1 SNPs on CR rate may depend on the drugs used within the induction therapy, i.e. classical MDR-mediated agents such as daunorubicin versus other agents, which were transported to a less amount by P-gp, including idarubicin or mitoxantrone. Thus, cautious interpretation is needed to apply the present result to other clinical situation.
It is notable to focus on the different definitions of survival adopted in the current study, EFS. The EFS can reflect treatment failure, such as primary refractory disease or early relapse, while OS includes nonrelapse mortality (including opportunistic infections) along with the relapse/refractoriness of the disease. Thus, in the current study, EFS was found to be associated with the MDR1 SNPs, whereas the MDR1 SNPs did not influence OS, especially in an SCT setting, which was consistent with the results of previous investigations, where MDR1 blast expression was observed as predictive of the achievement of CR; yet, failed to predict influence on long-term survival.7, 21 However, cautious interpretation is needed to understand the present results because of small patient cohort.
The reason that the MDR1 SNPs were not associated with overall survival might be the pharmacokinetic effects of the MDR1 gene SNPs. The MDR1 gene SNPs might even influence the detoxification organ systems in patients under conventional chemotherapy. Thus, SNPs do not only have impact on chemosensitivity of AML blasts but also on treatment toxicity, and both aspects may influence survival. The pharmacokinetic effects of the MDR1 SNPs, altering the drug clearance, may influence the treatment outcomes, including transplantation results.22–25 However, a final conclusion on this issue requires a further study with a larger number of patients.
In conclusion, the present results revealed an association between the MDR1 SNPs and the treatment outcomes or EFS for AML patients, plus the MDR1 SNPs seemed to be correlated with the functional P-gp activity in the leukemic blasts. However, OS did not seem to be linked with the MDR1 SNPs. Nonetheless, further study is needed to reach a final conclusion on the association of the MDR1 gene SNPs with the treatment outcomes for AML patients with different ethnicities.
- 10Large-volume leukapheresis using femoral venous access for harvesting peripheral blood stem cells with the Fenwal CS 3000 Plus from normal healthy donors: predictors of CD34+ cell yield and collection efficiency. J Clin Apher 2003; 18: 10–15., , , , , , .
- 13MitelmanF, ed. An international system for human cytogenetic nomenclature, ISCN. Basel, Switzerland: S Karger, 1995.
- 16IL-10 promoter gene polymorphism associated with the occurrence of chronic GVHD and its clinical course during systemic immunosuppressive treatment for chronic GVHD after allogeneic peripheral blood stem cell transplantation. Transplantation 2005; 79: 1615–22., , , , , , , .
- 18The presence of a FLT3 internal tandem duplication in patients with acute myeloid leukemia (AML) adds important prognostic information to cytogenetic risk group and response to the first cycle of chemotherapy: analysis of 854 patients from the United Kingdom Medical Research Council AML 10 and 12 trials. Blood 2001; 98: 1752–9., , , , , , , , , , , .