Role of aldo-keto reductases and other doxorubicin pharmacokinetic genes in doxorubicin resistance, DNA binding, and subcellular localization

Abstract

Background: Since proteins involved in chemotherapy drug pharmacokinetics and pharmacodynamics have a strong impact on the uptake, metabolism, and efflux of such drugs, they likely play critical roles in resistance to chemotherapy drugs in cancer patients.

Methods: To investigate this hypothesis, we conducted a whole genome microarray study to identify difference in the expression of genes between isogenic doxorubicin-sensitive and doxorubicin-resistant MCF-7 breast tumour cells. We then assessed the degree of over-representation of doxorubicin pharmacokinetic and pharmacodynamic genes in the dataset of doxorubicin resistance genes.

Results: Of 27,958 Entrez genes on the array, 7.4 per cent or 2,063 genes were differentially expressed by ≥ 2-fold between wildtype and doxorubicin-resistant cells. The false discovery rate was set at 0.01 and the minimum p value for significance for any gene within the "hit list" was 0.01. Seventeen and 43 per cent of doxorubicin pharmacokinetic genes were over-represented in the hit list, depending upon whether the gene name was identical or within the same gene family, respectively. The most over-represented genes were within the 1C and 1B families of aldo-keto reductases (AKRs), which convert doxorubicin to doxorubicinol. Other genes convert doxorubicin to other metabolites or affect the influx, efflux, or cytotoxicity of the drug. In further support of the role of AKRs in doxorubicin resistance, we observed that, in comparison to doxorubicin, doxorubincol exhibited dramatically reduced cytotoxicity, reduced DNA-binding activity, and strong localization to extra nuclear lysosomes. Pharmacologic inhibition of the above AKRs in doxorubicin-resistant cells increased cellular doxorubicin levels, restored doxorubicin cytotoxicity and re-established doxorubicin localization to the nucleus. The properties of doxorubicinol were unaffected.

Conclusions: These findings demonstrate the utility of using curated pharmacokinetic and pharmacodynamic knowledge bases to identify highly relevant genes associated with doxorubicin resistance. The induction of one or more of these genes was found to be correlated with changes in the drug's properties, while inhibiting one specific class of these genes (the AKRs) increased cellular doxorubicin content and restored drug DNA binding, cytotoxicity, and subcellular localization.

Figure 1

Significance Analysis of Microarrays (SAM) graph for MCF-7_DOX2-12and MCF-7_CC112cells. Red or green points denote genes with significant over- or under-expression in MCF-7_DOX2-12 cells compared to MCF-7_CC12 cells, while black points represent genes that do not show altered expression.

Figure 2

PharmGKB doxorubicin pharmacodynamic and pharmacokinetic pathways. Doxorubicin metabolic pathways were retrieved from the PharmGKB website (http://www.pharmgkb.org/drug/PA449412#tabview=tab4&subtab=33), combined, and used for further analysis. All diagrams are copyright PharmGKB. Used with permission from PharmGKB and Stanford University. (A) pharmacodynamics (B) Pharmacokinetics.

Figure 3

Gene and protein expression levels of doxorubicin to doxorubicinol metabolizing enzymes. (A) Relative changes in expression between MCF-7_CC12 and MCF-7_DOX2-12 cells were assessed by RTqPCR. Each sample was normalized first to RPS28 for loading and then to the average expression of MCF-7_CC12 cells to determine fold change. * p ≤ 0.05, ** p ≤ 0.01. (B) Western blot analysis of AKR1C3 expression in MCF-7_DOX2-12 and MCF-7_CC12 whole cell lysates.

Figure 4

Doxorubicin and doxorubicinol cytotoxicity in MCF-7_CC12and MCF-7_DOX2-12cells upon treatment with or without 5β-cholanic acid. MCF-7_CC12 and MCF-7_DOX2-12 cells were assessed for their sensitivity to doxorubicin and doxorubicinol in the presence or absence of the AKR1B10, AKR1C2, and AKR1C3 inhibitor 5β-cholanic acid using a clonogenic assay.

Figure 5

Intracellular localization of doxorubicin and doxorubicinol in MCF-7_CC12and MCF-7_DOX2-12cells upon treatment with or without 5β-cholanic acid. Cells were treated for 24 h with doxorubicin or doxorubicinol (red) in the presence or absence of the AKR1C2, AKR1C3, and AKR1B10 inhibitor 5β-cholanic acid. After treatment for 24 h, the nuclear dye DRAQ5 (blue) was added to the media as a counterstain for 15 minutes, and then coverslips were mounted and imaged by confocal fluorescence microscopy.

Figure 6

Intracellular levels of doxorubicin or doxorubicinol in the presence or absence of 5β-cholanic acid or cyclosporine A, measured by HPLC. MCF-7_CC12 and MCF-_7DOX2-12 cells were treated for 24 h with 0.5uM doxorubicin or doxorubicinol, 200uM 5β-cholanic acid, and/or 5uM cyclosporine A. After treatment, cellular and media extracts were prepared and assessed for doxorubicin or doxorubicinol content by HPLC: (A) Treatment with doxorubicinol and 5β-cholanic acid: 1 = no treatment; 2 = 0.5uM doxorubicinol; 3 = 0.5uM doxorubicinol and 200uM 5β-cholanic acid, (B) Treatment with doxorubicin, 5β-cholanic acid, and/or cyclosporine A: 1 = no treatment; 2 = 0.5uM doxorubicin; 3 = 0.5uM doxorubicin + 200uM 5β-cholanic acid; 4 = 0.5uM doxorubicin + 5uM cyclosporine A; 5 = 0.5uM doxorubicin + 200uM 5β-cholanic acid + 5uM cyclosporine A. The symbol # represents differences between MCF-7_CC12 and MCF-7_DOX2-12 cells (treatment 2) with a p value ≤ 0.01. The symbols * and ** represent significant differences between treatments at p values of ≤ 0.05 and ≤ 0.01, respectively.

Figure 7

Relative DNA binding affinity of doxorubicin and doxorubicinol measured by a FID assay. Relative DNA binding affinity of doxorubicin and doxorubicinol was measured by adding consecutive aliquots of either doxorubicin or doxorubicinol to ethidium bromide-saturated salmon sperm DNA. Addition of the doxorubicin or doxorubicinol displaced the ethidium bromide and resulted in a progressive net decrease in fluorescence at excitation 545 nm and emission 595 nm.