Therapy-resistant acute lymphoblastic leukemia (ALL) cells inactivate FOXO3 to escape apoptosis induction by TRAIL and Noxa

Abstract

Forkhead transcription factors (FOXO) are downstream targets of the phosphoinositol-3-kinase (PI3K) protein kinase B (PKB) signaling cascade and play a pivotal role in cell differentiation, cell cycle and apoptosis. We found that cells from prednisone-resistant T-acute lymphoblastic leukemia (T-ALL) patients showed cytoplasmic localization of FOXO3 in comparison to prednisone-sensitive patients suggesting its inactivation. To determine the impact of FOXO3, T-ALL cells were infected with a 4OH-tamoxifen-regulated, phosphorylation-independent FOXO3(A3)ERtm allele. After FOXO3-activation these cells undergo caspase-dependent apoptosis. FOXO3 induces the death ligand TRAIL and the BH3-only protein Noxa implicating extrinsic as well as intrinsic death signaling. Whereas dnFADD partially inhibited cell death, CrmA and dnBID efficiently rescued ALL cells after FOXO3 activation, suggesting a caspase-8 amplifying feedback loop downstream of FADD. Knockdown of TRAIL and Noxa reduced FOXO3-induced apoptosis, implicating that mitochondrial destabilization amplifies TRAIL-signaling. The-reconstitution of the cell cycle inhibitor p16INK4A, which sensitizes ALL cells to mitochondria-induced cell death, represses FOXO3 protein levels and reduces the dependency of these leukemia cells on PI3K-PKB signaling. This suggests that if p16INK4A is deleted during leukemia development, FOXO3 levels elevate and FOXO3 has to be inactivated by deregulation of the PI3K-PKB pathway to prevent FOXO3-induced cell death.

Figure 1. FOXO3 localizes to the cytoplasm in bone marrow cells from prednisone-resistant T-ALL pediatric patients

Bone marrow cells from pediatric T-ALL patients sensitive (PGR) or resistant (PPR) to prednisone were fixed before treatment in 4% formaldehyde. FOXO3 expression was visualized with a specific rabbit monoclonal antibody against FOXO3, followed by the addition of ALEXA488 (green) labeled anti-rabbit antisera. DAPI (blue) was applied to visualize the nuclei. Images were acquired by the videoconfocal system ViCo microscope Nikon Eclipse 80i (Nikon, Japan). The shown stainings are representative for three PGR and three PPR patients.

Figure 2. FOXO3 activation induces caspase-dependent apoptotic cell death in CEM cells

The vector pLIB-FOXO3(A3)ERtm-iresNeo was retrovirally infected into CEM cells. Expression of the endogenous FOXO3 and the fusion protein FOXO3(A3)ERtm in CEM/FOXO3 cells were assessed by immunoblot analysis. α-Tubulin served as loading control. Mock-transfected CEM/Ctr cells were used as control (A, left panel). CEM/FOXO3 cells were incubated with 4OHT (50 nM) for 24 hours to activate transgenic FOXO3. Exposure of phosphatidylserin was analyzed by AnnexinV staining and loss of mitochondrial activity was detected by CMXRos staining (B). CEM/Ctr and CEM/FOXO3 cells were treated with 4OHT (50 nM) and/or the caspase inhibitor qVD (10 µM) for 24 and 48 hours. Apoptosis was measured by PI-FACS analyses (C). Statistical difference between treatments was assessed by unpaired t-test (***P < 0.001).

Figure 3. FOXO3-induced apoptosis depends on extrinsic and intrinsic death signaling

CEM/FOXO3 cells were retrovirally infected with pLIB-BCL2-iresPuro, pLIB-CrmA-iresPuro, pLIB-dnFADD-iresPuro, pLIB-dnBID-iresPuro or pLIB-BclxL-iresPuro supernatants. Transgenic expression of BCL2, BclxL, CrmA, dnBID and dnFADD in these cells was verified by immunoblot analyses. Mock-infected cells were used as controls. α-Tubulin served as loading control (A). CEM/FOXO3-Ctr, CEM/FOXO3-BCL2, -BclxL, -CrmA, -dnBID, and -dnFADD cells were incubated with 4OHT (50 nM) for 18 hours. Cleavage of caspases-8, -9, and -3 was assessed by immunoblot analyses. α-Tubulin was used as loading control (B). Apoptosis induction was assessed by PI-FACS analyses (C). Statistical difference between 4OHT-treated controls and cell lines with ectopic expression of apoptosis inhibitors was calculated by unpaired t-test (***P < 0.001, **P < 0.005).

Figure 4. TRAIL and Noxa are critical mediators of FOXO3-induced apoptosis in T-ALL cells

Total RNA was prepared from CEM/FOXO3 cells after incubation with 4OHT (50 nM) for 0 and 6 hours. The mRNA levels of TRAIL, Bim and Noxa were measured by quantitative RT-PCR (A). Induction of TRAIL, Bim and Noxa after incubation of CEM/FOXO3 cells with 4OHT (50 nM) for the times indicated was assessed by immunoblot analyses. Equal protein loading was confirmed by α-Tubulin (B). For knockdown experiments CEM/FOXO3 cells were infected with retroviruses coding for short-hairpin RNAs against TRAIL, Bim or Noxa. Knockdown efficiency was controlled by RT-PCR after incubation with 50 nM 4OHT for 4 hours (C). Apoptosis was measured by PI-FACS analyses in bulk-selected CEM/FOXO3-shCtr, -shTrail, -shBim or -shNoxa cells (D). Unpaired t-test was used to assess the statistical difference between 4OHT-treated controls and cell lines expressing shRNAs (***P < 0.001, **P < 0.005).

Figure 5. p16^INK4A regulates FOXO3 steady state expression and thereby apoptosis sensitivity

CEM/p16 cells were treated for 24 hours with 250 ng/ml doxy. The G1-arrest was measured by flow cytometry after PI-staining (A). p16^INK4A and FOXO3 levels were assessed by immunoblot analysis after incubation of CEM/Ctr and CEM/p16 cells with doxy (250 ng/ml) for 24 hours. Equal protein loading was confirmed by α-Tubulin detection (B). CEM/Ctr, CEM/p16 and CEM/p16-ECFP-FOXO3wt cells were incubated with doxy for 24 hours and/or Ly294002 (40 µM) for another 48 hours. The expression of ECFP-FOXO3wt was assessed by live cell fluorescence microscopy. Bar is 50 µm (left panel). Apoptosis was assessed by PI-FACS analyses (C). CEM/p16 cells were retrovirally infected with pLIB-FOXO3(A3)ERtm-iresNeo supernatants. The expression of the fusion protein FOXO3(A3)ERtm and the endogenous FOXO3 were verified by immunoblot analyses. α-Tubulin was used as loading control (D, left panel). CEM/FOXO3-Ctr and CEM/p16-FOXO3(A3)ER cells were treated with doxy for 24 hours and/or 4OHT (50 nM) for another 36 hours (D, right panel). Apoptosis levels were assessed by PI-FACS analysis, for statistical analysis unpaired t-test was used (***P < 0.001, **P < 0.005).

Figure 6. Model for FOXO3-induced apoptosis in T-ALL

Our data suggest that the FOXO3 targets TRAIL and the BH3-only protein Noxa are critical for cell death induction. As dnFADD only partially inhibited FOXO3-induced cell death, an additional apoptosis signal might be triggered via TRADD/RIP/TRAF2 and NF-κB, leading then to cleavage of caspase-8 and activation of tBID independent of FADD. FOXO3-induced expression of Noxa on the other hand may partially sequester anti-apoptotic BCL2 proteins such as MCL1, BclxL and BCL2 and thereby sensitize mitochondria to tBID. The concerted impact of both apoptotic stimuli at mitochondria thereby will cause activation of BAX/BAK, loss of outer mitochondrial membrane integrity, Cytochrome c release and caspase-9 and -3 activation. Cleavage of caspase-8 by small amounts of caspase-9 may further amplify this death circuit eventually leading to Cytochrome c release and apoptosis.