Expression of ABCG2 (BCRP) is regulated by Nrf2 in cancer cells that confers side population and chemoresistance phenotype

Abstract

ATP-binding cassette, subfamily G, member 2 (ABCG2) is expressed in both normal and cancer cells and plays a crucial role in side population (SP) formation and efflux of xenobiotics and drugs. Nrf2, a redox-sensing transcription factor, on constitutive activation in non-small-cell lung cancer cells upregulates a wide spectrum of genes involved in redox balance, glutathione metabolism, and drug detoxification, which contribute to chemoresistance and tumorigenicity. This study examined the mechanism underlying Nrf2-dependent expression of ABCG2 and its role in the multidrug resistance phenotype. In silico analysis of the 5'-promoter flanking region of ABCG2 identified an antioxidant response element (ARE) at -431 to -420 bp. A detailed promoter analysis using luciferase reporter assays showed that ARE at -431 to -420 bp is critical for the Nrf2-mediated expression in lung cancer cells. Electrophoretic mobility shift assays and chromatin immunoprecipitation assays revealed that Nrf2 interacts with the ABCG2 ARE element at -431 to -420 bp in vitro and in vivo. Disruption of Nrf2 expression in lung and prostate cancer cells, by short hairpin RNA, attenuated the expression of ABCG2 transcript and protein, and dramatically reduced the SP fraction in Nrf2-depleted cancer cells. Moreover, depleted levels of ABCG2 in these Nrf2 knockdown cells sensitized them to mitoxantrone and topotecan, two chemotherapy drugs detoxified mainly by ABCG2. As expected, overexpression of Nrf2 cDNA in lung epithelial cells led to an increase in ABCG2 expression and a 2-fold higher SP fraction. Thus, Nrf2-mediated regulation of ABCG2 expression maintains the SP fraction and confers chemoresistance.

Figure 1. Nrf2-dependent ABCG2 expression in lung cancer cells

(A) Relative expression of ABCG2 in A549 and H460 cells constitutively expressing Nrf2shRNA analyzed by real-time RT-PCR. β-actin was used for normalization. Data are presented as fold change calculated using gene expression levels in control luciferase shRNA cells as baseline. (B) Immunoblot analysis of ABCG2, Nrf2, and GAPDH expression in A549 and H460 cells expressing control luciferase shRNA and Nrf2 shRNA.

Figure 2. Nrf2 regulates ABCG2 transcription through an ARE element in the ABCG2 promoter

(A) A putative ARE at -431 bp to -420 bp in the proximal promoter region of the human ABCG2 gene was identified by in silico analysis and promoter reporter assay. Left, schematic representation of the two constructs with and without putative ARE. The 5′ end of each of the constructs relative to the transcription start site (arrows) is indicated. These constructs were transfected into A549 control cells and A549 Nrf2shRNA cells, and the luciferase activity was measured. Luciferase activities were normalized to the Renilla luciferase activity of a co-transfected reporter vector. NQO1 basal promoter construct was included as a positive control for measuring Nrf2-dependent reporter expression. (B) ABCG2 promoter fragment containing the ARE core sequence was cloned upstream of heterologous promoter driving pTAL vector. Mutations in ARE core sequence (sequence shown on top) were introduced by site-directed mutagenesis. These constructs were transfected into A549 cells, and the luciferase activity was measured. ‘*’, significant when compared to wild type ARE. *, p-value< 0.05, analyzed by Student’s t test. (C) In vitro DNA binding activity of Nrf2 to ABCG2 ARE using EMSA assays. Nuclear proteins from Nrf2-depleted cells (lane 2) and control A549 cells (lane 3) were incubated with ³²P-labeled oligonucleotides harboring the ARE consensus sequence of ABCG2 gene. The resulting complexes were resolved by non-denaturing PAGE and analyzed. For competition assays, a 25 and 50 -fold excess of unlabeled oligonucleotides harboring the wild type ARE (lanes 4 and 5) of the ABCG2 gene was added during the pre-incubation period. A 25–50-fold excess of unlabeled oligonucleotides with mutated ARE sequence was added (lanes 6 and 7). Nonspecific unlabeled oligonucleotides that contain AP1 and NFκB binding sequences were used as a negative control (lanes 8-11). Nuclear extract from A549 control cells were used in both specific and non-specific competitions (lanes 4 to 11). The arrow indicates ARE binding complex; FP, free probe. (D) ChIP assays were performed with A549 cells stably expressing Nrf2shRNA and the control cells expressing luciferase shRNA.

Figure 3. Alteration of SP phenotype in Nrf2-depleted lung cancer cells

A549 and H460 cells were incubated with Hoechst 33342 dye and PI and were analyzed by flow cytometry. To demonstrate the specificity of assay, cells were incubated with Hoechst dye in the presence of Fumitremorgin C (FTC). The SP is shown as a percentage of the whole viable cell population. Reactions were done in triplicates and were repeated three times.

Figure 4. Reversal of chemo-resistance in Nrf2-depleted lung cancer cells

A549 cells and H460 cells were treated with mitoxantrone and topotecan for 2-4 days, and cell viability was analyzed by MTT assays. Results shown are median values obtained from three independent experiments with 8 samples for each dose. Results were analyzed by One-way ANOVA analysis, ‘*’, p-value<0.001 in all panels.

Figure 5. Attenuated expression and activity of ABCG2 in Nrf2 deficient prostate cancer cells

(A) Relative expression of ABCG2 and Nrf2 in Du145 parent cells, Du145-LucshRNA and Du145-Nrf2shRNA cells analyzed by real-time RT-PCR. Data are presented as fold change calculated using gene expression levels in control luciferase shRNA cells as 1. (B) Immunoblot analysis of Nrf2 and ABCG2, expression in Du145 cells. (C) Densitometric quantization of Nrf2 and ABCG2 relative protein expression in Du145 parent cells, Du145-LucshRNA and Du145-Nrf2shRNA cells (arbitrary units [A.U.]).

Figure 6. Nrf2 activation induces ABCG2 expression in non-tumorigenic lung epithelial cells as well as tumorigenic lung epithelial cells

(A) Up-regulation of ABCG2 expression in cells expressing Keap1 shRNA. Total RNA from NuLi cells expressing Keap1 shRNA and Luciferase shRNA was isolated and expression of Keap1 and ABCG2 was determined by real time RT-PCR. Downregulation of Keap1 expression led to an increase in Nrf2 dependent NQO1, GCLM and ABCG2 expression. ‘*’, p-value<0.05, significant when compared with LucshRNA group. (B) Induction of ABCG2 expression in response to treatment with multiple oxidative stress inducing agents. NuLi cells were treated with TBHQ (20nM), and CS condensate (100 μg/ml) for 24h and expression of NQO1, GCLM and ABCG2 were measured by real-time RT-PCR. ‘*’, p-value<0.05, significant when compared with vehicle-treated sample. (C) Ecotopic expression of Nrf2 in lung cancer cells upregulates Nrf2 dependent ABCG2 expression. Overexpression of Nrf2 cDNA in H23 lung cancer cells resulted in activation of Nrf2 target gene expression including ABCG2. ‘*’, p-value<0.05, significant when compared with empty vector group. (D) Enforced expression of Nrf2 cDNA in H23 cells increases the proportion of SP cells.