ABCB1 Expression in Acute Myeloid Leukemia (AML): A Possible Predictive Value for Treatment Resistance?

Introduction: Over the last years, there has been a tremendous increase in understanding acute myeloid leukemia (AML) biology and a great effort has been taken in order to improve AML chemotherapy strategies. However, the growing knowledge of leukemia associated molecular mechanisms just started to translate into improved outcome. With regard to conventional chemotherapy multidrug resistance (MDR) is a persisting problem and the impact of ABCB1 (MDR1) expression is still controversially discussed.

Methods: In this study we evaluated the ABCB1 expression using qRT-PCR and gene expression profiling (Affymetrix U133plus2.0 arrays) in 250 diagnostic AML samples derived from patients enrolled on a prospective treatment trial of the German-Austrian AML Study Group (AMLSG 07-04 trial; NCT00151242), in which patients were treated with an intensive anthracycline/cytarabine-based induction therapy. Findings were also evaluated in 154 TCGA AML cases receiving a 7+3 induction treatment (data available at http://cancergenome.nih.gov/) and put into perspective with previous reports. Furthermore, we investigated ABCB1 expression associated gene signatures and examined epigenetic regulation mechanisms by COBRA and methyl-CpG immunoprecipitation sequencing (MCIp-seq) in selected cases.

Results: Our global analysis showed that patients who obtained a complete response (CR) following double induction therapy had lower ABCB1 mRNA levels compared to patients with refractory disease (RD) (p=0.07). Regarding cytogenetic AML subtypes, ABCB1 mRNA levels varied among the different cytogenetic groups with the complex karyotype group showing the highest ABCB1 and the inv(16) group the lowest ABCB1 expression levels. A comparison of CR versus RD cases within the cytogenetically determined prognostic groups showed that in the intermediate [CN-AML, t(11q23), and other intermediate risk cytogenetic aberrations (others^inter)] and poor risk groups (complex karyotype and others^high), RD patients presented with significantly higher ABCB1 mRNA levels (p=0.02). Similarly, patients with favorable risk cytogenetics [t(8;21) and inv(16)], who achieved a CR, presented with lower ABCB1 levels compared to the ones, who were refractory. Patients with the lowest ABCB1 expression quartile (ABCB1^low) showed significantly longer event-free survival (EFS) times than patients in the highest quartile cohort (ABCB1^high) (median EFS 322 vs 105 days; p=0.02), while no differences were observed with regard to overall survival. In accordance, there was a significant enrichment of RD cases in the ABCB1^high patient group (p=0.03). Next, in order to better understand the regulation of ABCB1 in AML, we specifically evaluated the DNA methylation level of a previously identified GC box important for ABCB1 expression regulation in CML and we performed global analyses of the entire ABCB1 5' region. While both analyses did not reveal significant differences, further investigation of an ABCB1 associated gene pattern showed a correlation with CD34 and KIT expression (p<0.001). This suggests that like in CML, ABCB1 might be regulated by WNT, and in line, normal CD34⁺ hematopoietic stem cells also showed high ABCB1 expression levels.

Conclusions: In summary, our data provide further evidence for a potential impact of ABCB1 deregulation on the response to AML chemotherapy, especially in more stem cell like leukemia cohorts as well as cytogenetically high risk AML. While we are currently further investigating the involvement of the Wnt/β-catenin pathway in the regulation of ABCB1 transcription in AML, further integration of molecular findings are warranted to better decipher the underlying drug resistance mechanisms. Ultimately, these analyses will improve patient management by adding valuable predictive biomarkers.