The BH3 mimetic ABT-737 induces cancer cell senescence

Abstract

ABT-737, a small molecule cell-permeable Bcl-2 antagonist that acts by mimicking BH3 proteins, induces apoptotic cell death in multiple cancer types. However, when incubated with this agent many solid tumor cell lines do not undergo apoptosis. The current study reveals a novel mechanism whereby ABT-737 when added to apoptosis-resistant cancer cells has profound biologic effects. In PV-10 cells, a renal cell carcinoma that does not die after ABT-737 treatment, this agent induces a two-fold change in the transcription of nearly 430 genes. Many of these induced mRNA changes are in secreted proteins, IL-6, IL-8, and IL-11 and chemokines CXCL2 and CXCL5, or genes associated with an "inflammatory" phenotype. Strikingly, these gene changes are highly similar to those changes previously identified in cellular senescence. Brief exposure of apoptosis-resistant renal, lung and prostate cancer cell lines to ABT-737, although not capable of inducing cell death, causes the induction of senescence-associated β-galactosidase and inhibition of cell growth consistent with the induction of cellular senescence. Evidence indicates that the induction of senescence occurs as a result of reactive oxygen species elevation followed by low-level activation of the caspase cascade, insufficient to induce apoptosis, but sufficient to lead to minor DNA damage and increases in p53, p21, IL-6 and 8 proteins. By overexpression of a dominant-negative p53 protein, we show that ABT-737-induced cellular senescence is p53-dependent. Thus, in multiple cancer types in which ABT-737 is incapable of causing cell death, ABT-737 may have additional cellular activities that make its use as an anticancer agent highly attractive.

Figure 1

Gene transcription changes after ABT-737 treatment. A, microarray analysis of gene expression induced by ABT-737. PV-10 cells were treated in triplicate with DMSO or 10 µmol/L ABT-737 for 24 hours and gene microarray changes documented. B, gene changes associated with senescence. C, qT-PCR analysis of IL-6 and IL-8 transcripts. PV-10 and 22Rv1 cells were treated in triplicate with DMSO, ABT-737 (10 µmol/L) or enantiomer (En., 10 µmol/L) for 24 hours (mean ± SD, n = 3). D, the secretion of IL-6 and IL-8 in PV-10 cells treated with DMSO, ABT-737 or enantiomer for 24 hours was determined by ELISA (n = 6, mean ± SD).

Figure 2

Transcriptional regulation of IL-6 increased by ABT-737 is mediated by transcription factors C/EBPβ and NF-κB. A, qT-PCR analysis of C/EBPβ mRNA expression. PV-10 cells were treated with DMSO or 10 µmol/L ABT-737 (ABT) for 24 hours (triplicate, ±SD). B, C/EBPβ protein expression is determined by Western blotting. C, EMSA on nuclear lysates from PV-10 cells with an NF-κB radiolabeled probe. Competition with wild type (100-fold excess) was used to show the specificity of this assay. Nuclear extracts from PV-10 cells treated with TNFα (3 nmol/L) for 30 minutes were used as positive control. D, luciferase reporter assay of IL-6. PV-10 cells were cotransfected with a luciferase construct, a wild-type (Wt-luc), NF-κB (−ΔNF-κB-luc), or C/EBPβ (−ΔC/EBPβ-luc) deletion mutant, and 0.01 µg of pEF-Renilla-luc. IL-6 luciferase activity (mean ± SD, n = 3) was determined by dual luciferase assay as described in Materials and Methods. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

Figure 3

ABT-737 induces senescent-like growth arrest in human cancer cells. A, cell growth curves were determined by MTT assay. PV-10 cells were treated with DMSO or ABT-737 (ABT; 10 µmol/L) for 24 hours only, the cell washed and then maintained in fresh growth media for up to 6 days (triplicate, ±SD). B, BrdU incorporation. The percentage of BrdU-positive cells was determined 5 days after ABT-737 exposure. C, clonogenic assay. Cells were treated with the indicated doses of ABT-737 for 5 days. The colony formation was visualized by crystal violet staining. The bar graph indicates the percentage inhibition of colony formation from control (DMSO). D and E, SA-β-gal staining. At 5 days after exposure to ABT-737, SA-β-gal–stained cells were visualized under phase contrast microscopy (D). The bar graph indicates percentage of SA-β-gal–positive cells (E) in each cell type.

Figure 4

DNA damage response induced by ABT-737 for induction of senescence. A, induction of γ-H2AX and p-ATM by ABT-737. Cells were treated with 10 µmol/L ABT-737 for 24 hours and then maintained in fresh media for 1 or 2 more days prior to Western blot analysis. B, Western blot analysis of dose-dependent induction of γ-H2AX expression. C, Western blot analysis of ATM and γ-H2AX expression levels. ATM expression in PV-10 cells was decreased by lentiviral infection with short hairpin microRNA (shRNAmir) targeting the ATM (shRNAmir-ATM). A nonsilencing control (shRNAmir-control) was also employed. Cells were treated with DMSO (−) or 10 µmol/L ABT-737 (+) for 24 hours. D, SA-β-gal activity in PV-10 cells infected with nonsilencing shRNAmir or ATM shRNAmir. These cells were exposed to DMSO or ABT-737 for 24 hours, washed with PBS, and then maintained in fresh growth media for 5 days. After staining with β-gal, the percentage of SA-β-gal–positive cells was evaluated under phase contrast microscopy (left). Phase contrast microscopy of SA-β-gal–positive cells is shown (right).

Figure 5

Contribution of caspase-3 cleavages to the ABT-737-induced senescence. A, Western blot analysis of caspase cleavage. PV-10 cells were treated with 10 µmol/L of ABT-737 with or without 40 µmol/L of z-VAD-FMK for 24 hours. Whole-cell lysates were subjected to Western blot analysis and the cleavage fragments of caspases identified by arrows. B, Western blot detection of γ-H2AX expression after caspase-3 (casp-3) knockdown. After 24 hours treatment, extracts were Western blotted with multiple antibodies. C, Western blot analysis of ICAD cleavage. The cleavage fragments of ICAD are identified by arrows. D, SA-β-gal staining was done at day 5 after exposure of ABT-737 or DMSO. E, qT-PCR analysis of IL-6 and IL-8 mRNA in cells depleted without (shRNAmir-Control) or with (shRNAmir-Casp-3) casp-3 expression. Cells were exposed to ABT-737 for 24 hours and then qT-PCR analysis (the mean ± SD of triplicate experiments of these values) was done. The IL-6 and IL-8 mRNA levels were normalized to GAPDH. DFF45, DNA fragmentation factor 45.

Figure 6

ABT-737 treatment increases the generation of ROS. A, ROS measurement. ROS levels were assessed by dichlorofluorescein (DCF) measurement after ABT-737 treatment for 24 hours. The data represent the mean ± SD of 3 independent experiments. B, Western blot analysis of caspase-3 and ICAD cleavage. Cells were pretreated with N-acetyl cysteine (NAC; 5 mmol/L) for 2 hours followed by DMSO or ABT-737 treatment for 24 hours. C, SA-β-gal activity. Cells were exposed to DMSO or ABT-737 for 24 hours in presence or absence of NAC, washed with PBS, and then maintained in fresh growth media for 5 days. After staining with β-gal, the percentage of SA-β-gal–positive cells was evaluated under phase contrast microscopy (triplicate experiments, the mean ± SD of measured values is shown).

Figure 7

ABT-737 induces cell cycle arrest through activation of p53-p21. A, Western blot analysis of p21 and p53 levels. Cells were treated with DMSO (−) or ABT-737 (10 µmol/L) for 24 hours and cellular extracts subjected to Western blotting. B, qT-PCR analysis of p21 transcripts. Cells were treated with DMSO, ABT-737, or enantiomer (En.) for 24 hours. The p21 mRNA levels were normalized to GAPDH. C, inactivation of p53 blocks ABT-737–mediated senescence. The p53 function in 22Rv1 cells was inactivated by retroviral infection with pBabe-hygro dominant-negative p53 (DN p53) and selected with hygromycin (150 µg/mL) for 14 days. These cells were treated with DMSO or 10 µmol/L ABT-737 (24 hours pretreatment) prior to Western blot analysis for p21 and p53. D, SA-β-gal activity. PV-10 cells in the top were treated for 24 hours and senescent cells evaluated 5 days later. The percentage of SA-β-gal–positive cells is given in the right and the morphologic features are shown in the left. PV-10 cells were treated with 500 µmol/L hydrogen peroxide (H₂O₂; for 2 hours pretreatment) and evaluated for SA-β-gal activity at day 5. E, qT-PCR analysis of IL-6 and IL-8 mRNA in 22Rv1 cells expressing dominant-negative p53. 22Rv1 cells expressing empty vector (EV) or pBabe-hygro dominant-negative (DD) p53 (DN p53) were treated with DMSO (−) or 10 µmol/L ABT-737 (+) for 24 hours. F, representative schematic diagram summarizing the signaling pathway for cancer cell senescence induced by ABT-737.