The JNK- and AKT/GSK3β- signaling pathways converge to regulate Puma induction and neuronal apoptosis induced by trophic factor deprivation

Abstract

The AKT, GSK3 and JNK family kinases have been implicated in neuronal apoptosis associated with neuronal development and several neurodegenerative conditions. However, the mechanisms by which these kinase pathways regulate apoptosis remain unclear. In this study we have investigated the role of these kinases in neuronal cell death using an established model of trophic factor deprivation induced apoptosis in cerebellar granule neurons. BCL-2 family proteins are known to be central regulators of apoptosis and we have determined that the pro-apoptotic family member Puma is transcriptionally up-regulated in trophic factor deprived neurons and that Puma induction is required for apoptosis in vitro and in vivo. Importantly, we demonstrate that Puma induction is dependent on both JNK activation and AKT inactivation. AKT is known to regulate a number of downstream pathways, however we have determined that PI3K-AKT inactivation induces Puma expression through a GSK3β-dependent mechanism. Finally we demonstrate that the JNK and AKT/GSK3β pathways converge to regulate FoxO3a-mediated transcriptional activation of Puma. In summary we have identified a novel and critical link between the AKT, GSK3β and JNK kinases and the regulation of Puma induction and suggest that this may be pivotal to the regulation of neuronal apoptosis in neurodegenerative conditions.

Figure 1. Puma expression is induced by potassium withdrawal in cerebellar granule neurons.

After 7 days in culture CGNs were either maintained in media containing 25 mM potassium (K25) or switched to low potassium medium containing 5 mM potassium (K5). A, RNA was harvested after 4, 6 and 8 hours and Puma, Bim and Hrk expression was analyzed by qRT-PCR. Expression was normalized to ribosomal S12 levels and is reported as fold increase over untreated controls (n = 4, p<0.05). B, Protein extracts were collected from CGNs derived from Puma+/+ and Puma−/− littermates or Bim+/+ and Bim−/− littermates 8 hours after potassium withdrawal and Puma and Bim protein levels were analyzed by western blot. Actin was included as a loading control.

Figure 2. Puma is essential for potassium-withdrawal induced apoptosis in CGNs.

CGNS derived Puma+/+ and Puma−/− littermates were maintained in high potassium medium (K25) or switched to low potassium medium (K5). A, The fraction of apoptotic cells was determined at 24 h by assessing nuclear morphology following Hoechst staining of Puma+/+ vs Puma−/− CGNs (n = 7, *p<0.01). B, Representative images of Hoechst stained wild type and Puma-deficient neurons 24 hours following potassium withdrawal. C, CGNs were stained with Mitotracker Red to assess mitochondrial membrane potential. The fraction of Mitotracker Red labeled neurons was determined 20 hours after potassium withdrawal and compared between genotypes (n = 4, *p<0.05). D, Cell lysates were collected 20 hours after potassium withdrawal and assayed for caspase-3-like activity. Caspase activity is reported as relative fluorescence units of cleaved caspase substrate and was compared between genotypes (n = 4,* p<0.05). E, CGNs derived from Bim+/+ and Bim−/− littermates were maintained in high potassium medium or switched to low potassium medium and the fraction of apoptotic cells was determined at 24 h (n = 5).

Figure 3. Puma contributes to CGN apoptosis during postnatal development in vivo.

Sagittal sections from the cerebellum of postnatal day 7 Puma+/+ and Puma−/− mice were TUNEL stained to detect apoptotic cells. A, Representative images showing Tunel labeling (brown DAB+ cells) of cerebellar sections from Puma+/+ and Puma−/− mice. B, The mean number of TUNEL positive cells in the internal granule layer (IGL) per field was determined for each animal and the data is presented as the mean +/− SEM for Puma+/+ and Puma−/− mouse pups (n = 4 mice of each genotype, *p<0.05).

Figure 4. JNK is required for Puma induction and potassium withdrawal induced neuronal apoptosis.

After 7 days in culture CGNs were either maintained in high potassium medium or switched to low potassium medium with or without 10 µM SP600125 (SP). A, Puma mRNA levels were assessed by qRT-PCR 6 hours after potassium withdrawal and are reported as fold increase over K25 controls (n = 7, *p<0.05). B, Protein extracts were collected 8 hours after potassium withdrawal and Phospho-ATF2, Phospho-c-Jun, ATF3 and Puma protein levels were assessed by western blot. Calnexin was included as a loading control. C, The fraction of apoptotic neurons was determined 24 hours after potassium withdrawal by examining nuclear morphology in Hoechst stained cells (n = 5, *p<0.05).

Figure 5. Activation of the AKT pathway inhibits potassium withdrawal induced Puma expression and neuronal apoptosis.

CGNs were maintained in high potassium medium or switched to low potassium medium with or without 200 nM IGF-1. A, Protein extracts were collected 8 hours after potassium withdrawal and Phospho-AKT, Phospho-GSK3β and Puma protein levels were assessed by western blot. B, RNA was collected 6 hours after potassium withdrawal and Puma mRNA levels were quantified by qRT-PCR. Puma expression is reported as fold increase over K25 controls (n = 4,*p<0.05). C, Neurons were Hoechst stained 24 hours after potassium withdrawal in the presence or absence of IGF-1 (200 nM) and the fraction of apoptotic cells was determined by examining nuclear morphology (n = 4, *p<0.05). D, CGNs were infected with adenovirus expressing either GFP or a constitutively active form of AKT (CA-AKT) at 10 MOI and seven days later cells were switched to low potassium medium to induce apoptosis. RNA was collected 6 hours after potassium withdrawal and Puma transcript levels were quantified by qRT-PCR. Puma mRNA levels are reported as fold increase over K25 controls (n = 3, *p<0.05).

Figure 6. PI3K/AKT inhibition induces Puma expression and Puma-dependent apoptosis in CGNs.

A, CGNs maintained in high potassium medium (K25) were treated with the PI3K inhibitor LY294002 (30 µM) or vehicle and protein extracts were collected after 12 hours and Phosph-AKT, Phospho-GSK3β and Puma protein levels were assessed by western blot. B, CGNs were infected with recombinant adenovirus expressing either CA-AKT or GFP at 10 MOI and after 7 days neurons were treated with or without LY294002 (30 µM). RNA was collected at 8 hours and Puma mRNA levels were quantified by qRT-PCR (n = 4, *p<0.05). C, CGNs derived from Puma+/+ and Puma−/− littermates were treated with or without 30 µM LY294002 (LY) and the fraction of apoptotic neurons was quantified after 24 hours by Hoechst staining.

Figure 7. GSK3β is required for Puma induction in potassium withdrawal induced apoptosis.

CGNs were switched to low potassium medium in the presence or absence of the GSK3α/β inhibitor SB415286 (SB, 30 µM) or the GSK3β specific inhibitor AR-A01 4418 (AR, 50 µM). A, RNA was collected six hours after potassium withdrawal and analyzed for Puma mRNA expression using qRT-PCR (n = 4, *p<0.05). B, Protein extracts were collected 8 hours after potassium withdrawal and Puma protein levels were analyzed by western blot. C, The fraction of apoptotic neurons was quantified after 24 hours by Hoechst staining (n = 3, *p<0.05).

Figure 8. AKT inactivation induces Puma expression and neuronal apoptosis via a GSK3β-dependent mechanism.

CGNs maintained for 7 days in high potassium medium were treated with or without the PI3K inhibitor LY294002 (30 µM, LY) in the presence or absence of the GSK3β inhibitor SB415286 (30 µM, SB). A, RNA was collected 8 hours post treatment and Puma mRNA levels were determined by qRT-PCR (n = 3,* p<0.05). B, Protein extracts were collected 12 hours post-treatment and Puma protein levels were analyzed by western blot. C, The fraction of apoptotic neurons was quantified after 24 hours by Hoechst staining (n = 3, *p<0.05).

Figure 9. The JNK and AKT/GSK3β pathways signal independently during potassium withdrawal in CGNs.

A, CGNs were maintained in high potassium medium (K25) or subjected to potassium withdrawal (K5) in the presence or absence of the JNK inhibitor SP600125 (10 µM) or the GSK3β inhibitor SB415286 (30 µM). Protein extracts were collected at 2, 4 and 6 hours and Phospho-ATF2, Phospho-c-Jun and ATF3 levels were analyzed by western blot. B, CGNs were subjected to potassium withdrawal in the presence or absence of 200 nM IGF-1. Protein extracts were collected after 6 hours and were analyzed for Phospho-ATF2, Phospho-c-Jun and ATF3 protein levels by western blot. C, CGNs were subjected to potassium withdrawal in the presence or absence of 10 µM SP600125 (SP) and protein extracts were collected after 6 hours and analyzed by western blot for Phospho-AKT and Phospho-GSK3β levels.

Figure 10. FoxO3a regulates Puma induction in potassium deprived CGNs.

A, CGNs were maintained in high potassium (K25) or subjected to potassium withdrawal (K5). Protein extracts were collected at 8 hours and Phospho(Thr32)-FoxO3a and total FoxO3a levels were analyzed by western blot. B, CGNs were transduced with lentivirus (1 MOI) expressing shRNA directed against FoxO3a or a non-targeting control shRNA and FoxO3a protein levels were analyzed by western blot at 7 days post-infection. C, Puma mRNA levels were measured by qRT-PCR following 8 hours of potassium deprivation in CGNs transduced with FoxO3a-shRNA or non-targeting control shRNA (n = 3; *p<0.05). D, Phospho(Thr32)-FoxO3a and total FoxO3a protein was analyzed by western blot in CGNs maintained in high potassium (K25) or subjected to potassium deprivation for 8 hours in the presence or absence of either IGF-1 (200 nM), SP600125 (10 µM) or SB415286 (30 µM).