Suppression of acetylpolyamine oxidase by selected AP-1 members regulates DNp73 abundance: mechanistic insights for overcoming DNp73-mediated resistance to chemotherapeutic drugs

Abstract

Enhanced resistance to chemotherapy has been correlated with high levels of Delta-Np73 (DNp73), an anti-apoptotic protein of the p53 tumor-suppressor family which inhibits the pro-apoptotic members such as p53 and TAp73. Although genotoxic drugs have been shown to induce DNp73 degradation, lack of mechanistic understanding of this process precludes strategies to enhance the targeting of DNp73 and improve treatment outcomes. Antizyme (Az) is a mediator of ubiquitin-independent protein degradation regulated by the polyamine biosynthesis pathway. We show here that acetylpolyamine oxidase (PAOX), a catabolic enzyme of this pathway, upregulates DNp73 levels by suppressing its degradation via the Az pathway. Conversely, downregulation of PAOX activity by siRNA-mediated knockdown or chemical inhibition leads to DNp73 degradation in an Az-dependent manner. PAOX expression is suppressed by several genotoxic drugs, via selected members of the activator protein-1 (AP-1) transcription factors, namely c-Jun, JunB and FosB, which are required for stress-mediated DNp73 degradation. Finally, chemical- and siRNA-mediated inhibition of PAOX significantly reversed the resistant phenotype of DNp73-overexpressing cancer cells to genotoxic drugs. Together, these data define a critical mechanism for the regulation of DNp73 abundance, and reveal that inhibition of PAOX could widen the therapeutic index of cytotoxic drugs and overcome DNp73-mediated chemoresistance in tumors.

Figure 1

Polyamine catabolic enzyme paox regulates DNp73 stability. (a) Schematic showing the polyamines (putrescine, spermidine and spermine) biosynthesis pathway and the involvement of anabolic (ODC, SAMDC, SPDS and SPMS) and catabolic enzymes (PAOX and SSAT; upper panel) and the effects of polyamines on Az1 frameshifting process (lower panel). N-acetylspermine (NASp) is the intermediate in the catabolic conversion from spermine to spermidine. (b) Semi-quantitative RT-PCR analysis of the expression of enzymes in the polyamine biosynthesis pathway in p53 null H1299 cells exposed to doxorubicin (5μM) for the indicated times. (c) Immunoblot analysis of steady-state DNp73β levels in H1299 cells transfected with or without Flag-tagged PAOX for 24 h followed by exposure to doxorubicin for 8 h. Expression of exogenous PAOX and DNp73 was detected by anti-Flag and anti-p73 (GC15) antibodies, respectively. (d) Immunoblot and semi-quantitative RT-PCR (top and middle panels), and quantitative real-time PCR (lower bar charts) analysis of overexpressed DNp73β or endogenous cyclin D1 in H1299 cells was analyzed 48 h after siRNA-mediated silencing of paox. All values are means±S.D. of duplicate experiments, reflected by error bars. (e) H1299 cells inducibly expressing DNp73β were induced by doxycycline for 24 h before treatment with the PAOX inhibitor, MDL72527, at the indicated doses for 16 h and harvested for immunodetection. All experiments were repeated at least thrice and representative data are shown throughout. (f and g) Intracellular levels of spermine and putrescine were measured in H1299 cells (left bars) or HCT116 cells (right bars) without (Untted) or with cisplatin (25 μM CDDP) or doxorubicin (5 μM Doxo) treatment for 6 h by HPLC and the ratio (see Supplementary Figure 1D for raw data) is presented (f). The levels of the intermediate product in the conversion of spermine back to spermidine, NASp was also measured and depicted (g). All values are means±S.D. of triplicate experiments for g

Figure 2

Polyamine oxidase (paox) expression is selectively regulated by AP-1 family members. (a) Expression of processed antizyme, Az1(p), and the indicated AP-1 members were analyzed by immunoblotting, and mRNA expression of paox was monitored by semi-quantitative RT-PCR, after exposure of H1299 cells to different concentration of doxorubicin (doxo) or cisplatin (CDPP), for 16 h. (b) Expression of paox was determined by RT-PCR 24 h after transfection of increasing amounts of indicated AP-1 members in H1299 cells. (c) Effect of siRNA-mediated silencing of c-jun, fosB or junB on the expression of paox mRNA upon doxorubicin treatment at the indicated times was analyzed by semi-quantitative PCR in H1299 cells (upper panels). Knockdown efficiency after 48 h was assessed by immunoblotting (lower panels). (d) Schematic showing human paox promoter with two potential AP-1-binding sites (P1 and P2) predicted in silico (upper panel). The effect of overexpresssing c-Jun, JunB, FosB or the c-Jun/FosB and JunB/FosB dimers on the promoter activity of paox (left) or collagenase (right) was assessed by luciferase reporter assay in H1299 cells. RLU: relative luciferase units, which are ratios of luciferase and β-galactosidase activities. All values are means±S.D. of duplicate experiments. (e) The effect of ablating the predicted AP-1-binding sites on paox promoter activity was assessed using the wild-type and mutated (Mutant P1+P2) promoters (left panel) in the absence or presence of dominant-negative c-Jun (TAM67) or FosB (DNFosB) (right panel). All values are means±S.D. of duplicate experiments. (f) H1299 cells exposed to doxorubicin for the indicated time periods were used for chromatin immunoprecipitation (ChIP) using antibodies against the indicated AP-1 members (lower panel). Total DNA collected before immunoprecipitation was used as input control and binding specificity was demonstrated by the negative signal in anti-rabbit IgG control. Schematic shows sites of enrichment for AP-1 on the human paox promoter (upper panel)

Figure 3

Selectivity in AP-1-mediated regulation of paox correlates with selectivity in AP-1-dependent DNp73 degradation. (a) Effect of co-expressing DNp73β with the indicated AP-1 members on the steady-state levels of DNp73 were assessed by immunoblotting 24 h after transfection in H1299 cells. Expression of AP-1 members was analyzed by immunoblotting and semi-quantitative PCR. Egfp was co-transfected as a control for transfection efficiency. (b) Half-life of transfected DNp73β after siRNA-mediated knockdown of FosB and JunB were assessed by the addition of 25 μg/ml cycloheximide for the indicated times before immunoblotting. The adjacent graphs show the quantification of the remaining DNp73 (arbitrary values). (c) H1299 cells inducibly expressing DNp73β were transfected with control scramble, c-jun-, fosB- or junB-specific siRNA for 24 h before induction of DNp73β expression for a further 24 h by doxycycline addition. Cells were then treated with doxorubicin for 4 h and harvested for immunodetection. (d) c-jun^−/− mouse embryonic fibroblasts (MEFs) were transfected with DNp73β and the AP-1 members for 36 h before analysis (right panel). Status of endogenous c-Jun in wild-type and knockout MEFs is shown by immunoblotting (left panel)

Figure 4

AP-1-mediated DNp73 depletion occurs via the PAOX-antizyme pathway. (a) Effect of overexpressing AP-1 members on the expression of Az1 mRNA and processed Az, [Az1(p)], was assessed by semi-quantitative RT-PCR and immunodetection, respectively, 24 h after transfection in H1299 cells. (b and c) Effect of the AP-1 members on transfected DNp73β in the presence of the naturally occurring antizyme inhibitor, AZI (b) or Flag-tagged PAOX (c) was determined by immunoblot analysis, 24 h after transfection in H1299 cells. Lower panels show quantification of DNp73 from the respective lanes from the blots above (b). (d and e) Effect of siRNA-mediated silencing of paox in the presence of AZI (d) or Az1 siRNA (e) on co-transfected DNp73β levels was determined by immunoblotting 24 h and 48 h after transfection in H1299 cells, respectively. Knockdown efficiency was determined by semi-quantitative RT-PCR

Figure 5

PAOX inhibition reverses chemoresistance by inducing apoptosis. (a and b) SHSY5Y-DNp73β or SHSY5Y-pcDNA cells were treated with the indicated doses of cisplatin (CDDP) (a) in the absence or presence of 200 μM of MDL72527 for 24 h, and the proportion of cell death was determined by Annexin V/propidium iodide (PI) staining and flow cytometry. Raw data are shown as density plots (left) and percentage of Annexin V⁺/PI⁺ cells are represented in a graphical format (right). The same experiment as a was repeated with HCT116-pcDNA and HCT-116-DNp73β (b). All values are means±S.D. of duplicate experiments. (c and d) SHSY5Y-DNp73β or SHSY5Y-pcDNA cells were treated with the indicated doses of cisplatin (CDDP) in the absence or presence of 200 μM of MDL72527 for 24 h and caspase-3 activity in cell lysates were measured using a colorimetric protease assay. The values were calculated by subtracting background signal (with no substrate) from duplicate experiment (c). The same experiment as c was repeated in HCT116-pcDNA and HCT116-DNp73β cells (d). All values are means±S.D. of duplicate experiments. (e) Schematic illustrates the proposed mechanism by which genotoxic stresses can induce the degradation of anti-apoptotic DNp73 by activating c-Jun, FosB and JunB. Upon stress, these three AP-1 family members are induced and inhibit the expression of the polyamine catabolic enzyme paox, resulting in the eventual increase in the expression of Az1(p), which leads to the degradation of DNp73, and consequently, chemosensitizes to genotoxic drugs