Protein kinase B (AKT) is a major downstream effector of survival and growth signaling cascades. A large body of evidence, spanning multiple disciplines, implicates dysregulation of AKT in an array of human diseases and includes diseases of the brain (Franke 2008). Indeed, dysregulation of AKT plays an important role in the development and/or progression of Huntington’s(Humbert et al. 2002; Gines et al. 2003; Colin et al. 2005), Parkinson’s (Xiromerisiou et al. 2008; Timmons et al. 2009), Alzheimer’s (Dickey et al. 2008; Lee et al. 2009), schizophrenia (Kalkman 2006; Thiselton et al. 2008), and bipolar depression (Yu et al. 2010), to name a few. Moreover, evidence suggests psychotropic drugs alter the activity of the AKT signaling pathway, which may partially explain the therapeutic benefit of these compounds (Beaulieu et al. 2009). A complete understanding of how endogenous AKT activity is regulated in the brain, and characterizing the biochemical mechanisms involved, will help in the development of novel CNS drugs able to alter AKT activity for the treatment of neurological diseases.
Pleckstrin homology and leucine rich repeat protein phosphatases (PHLPPs) are a recently identified group of proteins which inhibit several kinases including AKT (Gao et al. 2005), protein kinase C alpha (PKCα) (Gao et al. 2008), and MAPK (extracellular regulated kinase; ERK) (Shimizu et al. 2003). Two isoforms of PHLPP, PHLPP1 and PHLPP2, have been identified (Brognard et al. 2007). Further, PHLPP1 exists as two splice variants, PHLPP1α (Gao et al. 2005) and PHLPP1β/SCN circadian oscillatory protein (SCOP) (Shimizu et al. 1999). While the majority of studies have probed PHLPP function in cancer cells, PHLPP proteins in the brain have been comparatively uncharacterized. However, one recent study in PC12 cells found the PHLPP1β/SCOP splice variant inhibits ERK by binding to K-ras, an upstream activator of ERK signaling, and prevents ERK’s phosphorylation (Shimizu et al. 2003). The leucine rich repeat domain in PHLPP1β/SCOP is required for this inhibition, and deletion of the leucine rich repeat domain prevents PHLPP1β/SCOP’s association with K-ras. In addition, subsequent studies found PHLPP1β/SCOP regulates ERK signaling in the mouse hippocampus, and altering hippocampal PHLPP1β/SCOP levels affects learning and memory (Shimizu et al. 2007). Alternatively, studies in cancer cells show PHLPP1α and PHLPP2 inhibit AKT/PKCα by selectively dephosphorylating their Ser473 (pAKT473)/Ser657 (pPKCα657) site, respectively (Gao et al. 2005; Brognard et al. 2007; Hirano et al. 2009). Phosphorylation at pPKCα657 stabilizes PKCα and prevents its degradation (Gysin and Imber 1997). Therefore, dephosphorylation by PHLPP1α or PHLPP2, in addition to decreasing phosphorylated levels, also decreases total levels of PKCα. Although further characterization of PHLPP protein expression and function is needed, these studies indicate that they can inhibit multiple kinases and may provide a unique target for modulating several signaling pathways.
Previously, we found hippocampal regional differences in the level of AKT and PKCα signaling across the lifespan. When compared to the hippocampal Ammon’s horn (area CA3) subregion, the hippocampal Ammon’s horn (area CA1) subregion showed reduced levels of pAKT473 and pPKCα657. Further, regional differences in nuclear AKT inversely correlated with differences in the level of nuclear PHLPP1α but not other well known AKT regulators including phosphatase and tensin homolog or protein phosphatase 2, subunit A (Jackson et al. 2009). These studies suggest PHLPP1α plays an important role in reducing AKT and PKCα signaling in the CA1 hippocampal subregion. However, it has yet to be directly demonstrated in any neuronal model system that PHLPP1α inhibits AKT or PKCα. Therefore, the role of PHLPP1α in mediating the regional differences in AKT activity, remain speculative. In addition, because PHLPP1 knockdown studies to date have all targeted PHLPP1 (i.e. both splice variants are simultaneously knocked down) no study has investigated how PHLPP1β/SCOP selectively contributes to AKT or PKCα regulation.
The goals of this study were to (i) characterize the protein expression profile of all three PHLPP proteins in the hippocampus, and (ii) directly verify PHLPP1 proteins are regulators of AKT signaling in neurons. The work presented here is the first to show in a neuronal system, that inhibition of PHLPP1α increases AKT activity while inhibition of PHLPP1β/SCOP promotes AKT/PKCα activity. Thus, PHLPP1 splice variants are antagonistic regulators of critical signaling pathways. Further, by taking a broad approach and characterizing multiple facets of PHLPP regulation, we conclude PHLPP1α is a major contributor to AKT regulation in the adult rat brain, and specifically in neurons. Together our findings suggest PHLPP1α is a potential therapeutic target for future CNS drugs designed to regulate AKT signaling.
- Top of page
- Materials and methods
- Supporting Information
Pleckstrin homology and leucine rich repeat phosphatase proteins are novel regulators of multiple cell signaling pathways. Studies using non-neuronal culture models indicate PHLPP1 inhibits the kinases AKT and PKCα (Gao et al. 2005; Brognard et al. 2007; Hirano et al. 2009). However, it has not been directly shown PHLPP1 can inhibit AKT and PKCα in neurons. Multiple CNS diseases are linked to dysregulation of neuronal AKT. Validation that PHLPP proteins are in fact AKT regulators in the brain is an important advancement in our understanding of the underlying biochemical mechanisms controlling CNS AKT signaling.
The results of the current study provide unique evidence for a divergence in the function of PHLPP1α and PHLPP1β/SCOP splice variants, and suggest that PHLPP1α plays a major role in regulating AKT signaling in neurons. As PHLPP1β/SCOP possesses some of the same regulatory domains as PHLPP1α, it is thought PHLPP1β/SCOP may regulate AKT and PKCα, in addition to ERK (Shimizu et al. 2010). Here, we report, for the first time, significant functional differences in the role of PHLPP1 splicoforms for regulating the AKT and PKC signaling pathways in neurons. The difference was revealed by comparing the effects of knockdown of both PHLPP1α and PHLPP1β/SCOP, relative to knockdown of PHLPP1β/SCOP alone (Figs 7–10). Knockdown of both phosphatases results in the expected increase in pAKT473, consistent with previous studies in cancer cells (Gao et al. 2005). Quite the opposite was observed following selective knockdown of PHLPP1β/SCOP, which significantly reduced pAKT473, pAKT308, and pPKCα657 in hippocampal neurons. Similarly, the decrease in pPKCα657 and total PKCα levels observed following knockdown of PHLPP1β/SCOP was not observed following knockdown of PHLPP1α/β, consistent with the idea that PHLPP1α and PHLPP1β/SCOP have opposing influences on some signaling pathways. In contrast, PHLPP1α/β KO, like PHLPP1β/SCOP KO, reduced total ERK1 and increased pERK expression. Together, the results suggest that PHLPP1α exhibits greater control over the AKT signaling pathway and PHLPP1β/SCOP has effects opposite PHLPP1α on AKT and PKCα, while acting as the primarily regulator of the Ras/MAP/ERK signaling pathway. Finally, the decreased AKT/PKC phosphorylation and increased ERK phosphorylation was observed under steady state conditions in PHLPP1β/SCOP knockdown neurons (Figure S5) indicating that PHLPP1β/SCOP contributes to the basal activation of AKT, PKC, and ERK kinases.
The fact PHLPP1α and PHLPP1β/SCOP have divergent effects on AKT and PKC signaling, may help explain why over-expressing PHLPP1α in HEK293-FT cells results in a massive increase in PHLPP1β/SCOP levels. That is, expression of PHLPP1β/SCOP may be altered in order to adjust PHLPP1α actions on AKT activity. This balance in PHLPP1 splice variants is reflected by hippocampal differences in the level of PHLPP proteins in region CA1 and CA3. Specifically, PHLPP1α is significantly lower in the CA3 and PHLPP1β/SCOP tends to be lower in this region as well. However, many signaling pathways are affected by selective PHLPP1β/SCOP knockdown, thus it is not certain which pathway mitigates this cross-talk. Moreover, because we over-expressed PHLPP1α and knocked down PHLPP1α/β in HEK293-FT cells to verify the quality of the antibodies used in our work, we had the opportunity to determine if PHLPP1 regulates AKT, PKC, and ERK in HEK293-FT cells in a manner similar to neurons. Intriguingly, we did not observe the same role of PHLPP1 proteins in regulating AKT, PKC, and ERK signaling in HEK293-FT cells, compared to hippocampal neurons (Figures S6 and S7). Knockdown of PHLPP1 had no effect on the phosphorylated levels of AKT or ERK; however, a mild (although significant) increase in pPKCα657 levels was observed. In addition, over-expression of PHLPP1α robustly increased pERK levels and tended to increase pAKT473 levels. While we cannot explain the differential regulation of PHLPP1 proteins on cell signaling pathways in HEK393-FT cells, our findings are in agreement with work by Qiao and colleagues. These authors recently examined PHLPP signaling in multiple cell types. In some cells, including those of a breast cancer lineage, AKT activation is inhibited by over-expression of PHLPP proteins. However, over-expression of PHLPP proteins in many other cell types, including HEK293-T cells, did not alter AKT phosphorylation (Qiao et al. 2010). It is conceivable that HEK293 cells do not express crucial adapter proteins required for PHLPP1 regulation of AKT.
In addition to knockdown experiments, characterization of PHLPP proteins in the brain reveal several significant findings which underscore the potential importance of the PHLPP1α splice variant in regulation of neuronal AKT. First, developmental studies reveal the PHLPP1α isoform accumulates over development and expression in the adult is greatly increased relative to PHLPP1β/SCOP and PHLPP2. Interestingly, the abrupt increase in PHLPP1α expression is associated with a decrease in AKT activity during development. A recent study found overactivation of AKT during development disrupts normal neuronal growth and may contribute to mental illness (Kim et al. 2009). Our results suggest that PHLPP1α contributes to the stringent regulation of AKT activity, which is critical for normal CNS growth during maturation. It would be interesting to learn if changes in PHLPP1α activity accompany changes in AKT activity in models of CNS disease.
In addition to developmental studies, subcellular fractionation experiments using adult hippocampal tissue reveal an important aspect of PHLPP1α regulation. While PHLPP1β/SCOP and PHLPP2 predominantly localize to the cytoplasmic/plasma membrane fraction, PHLPP1α is highly abundant in the nucleus and nuclear membrane. The latter finding is supported by intense nuclear and perinuclear immunofluorescent staining of PHLPP1α in cultured hippocampal neurons. The importance of this observation relates to AKT’s well elucidated mechanisms of action. Specifically, activation of AKT by phosphorylation at Thr308 and Ser473 elicits a rapid and robust translocation of AKT into the nucleus. Once in the nuclear compartment AKT targets multiple substrates for phosphorylation, including FOXO3a, and thereby promotes survival signaling mechanisms (Zheng et al. 2002). The finding that PHLPP1α is the primary PHLPP protein within the nuclear compartment, suggests that PHLPP1α alone is positioned to turn off AKT once activated. Therefore, nuclear PHLPP1α, rather than the predominately cytoplasmic PHLPP1β/SCOP or PHLPP2, is poised to regulate nuclear AKT activity in neurons.
Finally, our characterization of PHLPP proteins reveal among different CNS cell types, PHLPP1α is primarily confined to neurons (Fig. 3). This ostensibly trite finding, in fact, may be one of the more intriguing and important discoveries within this work. Searching for methods to selectively increase AKT activity in neurons but not the supporting glia has been a continuous yet unsuccessful enterprise. Increasing AKT globally is decidedly undesirable because of the increased risk for developing peripheral or CNS cancers. This concern was highlighted in a recent review suggesting the hypoactivity of the AKT signaling cascade contributes to the development of schizophrenia (Kalkman 2006). At the same time, decreased AKT activity might explain the reduced risk for developing cancer in the schizophrenic population (Barak et al. 2005; Levav et al. 2007). Therefore, the author predicts using pharmacological interventions to increase CNS AKT activity may be a good strategy to prevent or reduce mental illness but increases the risk of developing CNS tumors as a consequence to AKT mediated therapies (Kalkman 2006). Together our results show that (i) PHLPP1α is primarily present in neurons and (ii) a potent AKT inhibitor in neurons, it is possible that inhibition of PHLPP1α with small molecule inhibitors may be a useful strategy to increase neuronal AKT levels while side-stepping the associated risk of developing gliomas. In addition, the comparatively low levels of PHLPP1α in multiple organs outside the brain may also help to reduce other side effects potentially caused by increased AKT activity. As PHLPP research advances its likely selective agonists and antagonists will become available, allowing the hypothesis to be put to the test.
As interest in PHLPP proteins for treatment of human disease mounts, our results for a divergence in the function of PHLPP1α and PHLPP1β/SCOP provide insight into potential therapeutic strategies. In cancer therapy, where activated AKT needs to be reduced, enhancing PHLPP1α signaling while reducing PHLPP1β/SCOP signaling might be more effective. Alternatively, in neurodegenerative disease where increased AKT activity is desired, the opposite strategy could apply. Consistent with the idea that targeting PHLPP1 proteins may improve outcomes in neurodegenerative disease, recent work inversely correlates decreased PHLPP1α levels with increased pAKT473 levels and reduced cell death in the striatum in Huntington’s Disease mouse models (Saavedra et al. 2010). Moreover, decreased PHLPP1 levels in the putamen are observed in samples taken from Huntington’s Disease patients, suggesting the PHLPP1/AKT axis is important in human disease.
Our finding that different PHLPP1 splice variants serve different functions also dovetails well with recent advances in PHLPP research. The first published report of a PHLPP1 knockout animal was recently described (Masubuchi et al. 2010). In the context of this new animal model, our observations on PHLPP1 splice variants are particularly important because they highlight subtleties in the dual regulation and interaction between PHLPP1α and PHLPP1β/SCOP which may otherwise be overlooked. Complete PHLPP1 gene knockout may hide effects caused by loss of PHLPP1β/SCOP and preferentially reveal PHLPP1α’s influence on cell signaling. In our system, the effect of PHLPP1β/SCOP knockdown on AKT and PKCα was reversed and lost, respectively, when PHLPP1α was simultaneously knocked down; this finding indicates PHLPP1α’s effects on the AKT and PKCα pathways are dominant to PHLPP1β/SCOP’s when both our reduced; such is the case with the PHLPP1 KO mouse. Nevertheless, this animal model promises exciting new data and will help elucidate the roles PHLPP1 plays in the body, as well as, determine what consequences dysfunctional PHLPP1 signaling has on behavior.
In conclusion our results indicate PHLPP1α and PHLPP1β/SCOP do not serve the same function for regulation of AKT, PKC, or ERK signaling in hippocampal neurons. PHLPP1α inhibits AKT and PKC signaling in neurons. Alternatively, PHLPP1β/SCOP is important for the normal activation of AKT and PKC signaling but plays a dominant role in inhibiting the ERK pathway in neurons. The demonstration that PHLPP1α is highly abundant in neurons is a major advance in understanding potential strategies to increase neuronal AKT in diseased neurons. While work here points to PHLPP1α is a potent regulator of AKT in neurons, future studies will need to define PHLPP2’s contribution to AKT regulation in neurons. Although, PHLPP2 appears to be less abundant, no broad conclusions can be made as to the potential importance of PHLPP2 for AKT regulation. It is possible the phosphatase activity of PHLPP2 is differently regulated relative to PHLPP1α and could provide another level of cell signaling regulation. In agreement with this possibility, PHLPP1β/SCOP expression in neurons is far lower than PHLPP1α, and yet PHLPP1β/SCOP is a major regulator of ERK phosphorylation.
- Top of page
- Materials and methods
- Supporting Information
Figure S1. Validation of hippocampal cell culture. Isolated hippocampal neurons were grown on glass cover slides for 10 days, and subsequently stained for cell markers indicated above each panel. All images represent overlays using cell type specific antibodies (GREEN) and nuclei (BLUE). No oligodendrocytes or microglia were detected, however, a small percentage (~ 5–7%) of cells were astrocytes. Hippocampal neurons represent the majority of cell types present in 10 DIV cultures. The bottom two figures show two random fields overlaying astrocytes (RED) and neurons (GREEN) for comparison. More neuronal staining is evident compared to astrocyte staining.
Figure S2. Cayman PHLPP1 antibody detects PHLPP1α. Blot showing that the PHLPP1 antibody, (Cayman Chemicals) used in all immunofluorescence experiments, also detects PHLPP1α by western analysis. The Cayman antibody detects PHLPP1α in HEK293-FT cells over-expressing human PHLPP1α but not in cells over-expressing GFP only. GAPDH (loading control) did not differ between PHLPP1α or GFP over-expressing cells.
Figure S3. Enhanced image of astrocyte & neurons. From white box in Fig. 3 Panel (i). Image shows an astrocyte in green devoid of RED PHLPP1 staining. Astrocytic processes extend around several large neuronal nuclei that show perinuclear PHLPP1 (RED) staining.
Figure S4. PHLPP1 brain immunofluorescence, control and CA1 228-ms exposure. (a–c) Control images showing little to no staining in brains incubated with secondary only. (a) red channel 228-ms exposure and (b) green channel. (d–f) The original exposure time (228-ms exposure) of hippocampal sub-region CA1 in the adult rat brain showing intense PHLPP1 (RED) staining.
Figure S5. Neuronal PHLPP1β knockdown alters basal activation of AKT, PKC, and ERK kinases. Hippocampal neurons were grown for 10 days in vitro (10 DIV). At DIV 0 and 7 neurons were transduced with either a non-targeting (NT) shRNA or selective PHLPP1β targeting shRNA. At DIV 10, cells were serum starved for 2 h, washed twice in ice cold PBS, and homogenized in RIPA buffer containing protease and phosphatase inhibitors. Activation of AKT, PKC, and ERK were analyzed by western analysis. (a) Representative blots (n = 3) show without IGF-1 stimulation after a 2-h starvation period, AKT and PKC phosphorylation is reduced and ERK activation is increased in PHLPP1β knockdown neurons. (b–i) Densitometry of proteins analyzed by western blot (n = 5 for NT-Controls; n = 4 for PHLPP1β knockdown neurons). The results indicate PHLPP1β knockdown alters the basal activation of AKT, PKC, and ERK kinases. Data are significant at p < 0.05 (*), p < 0.01 (**), p < 0.001 (***).
Figure S6. Effect of PHLPP1 knockdown to alter AKT, PKC, and ERK activation in HEK293-FT cells. HEK293-FT cells were transduced with lentivirus delivering either a non-targeting (NT) shRNA or human specific PHLPP1 shRNA. (a) Western blot showing changes in phosphorylated/total levels of AKT, PKC, and ERK kinases after knockdown. (b and c) Densitometry of western blots in panel (a) (n = 4) showing knockdown of PHLPP1 proteins induced a significant increase (p = 0.02211) in pPKCα657 levels but not in PKCα total levels. Only PKCα phosphorylation was altered by PHLPP1 knockdown, no significant change in AKT or ERK phosphorylation was observed.
Figure S7. Effect of PHLPP1α over-expression to alter AKT, PKC, and ERK activation in HEK293-FT cells. HEK293-FT cells were transiently transfected with a plasmid containing either GFP only (i.e. control cells) or the human cDNA for PHLPP1α. (a) Western blot showing changes in phosphorylated/total levels of AKT, PKC, and ERK kinases after over-expression of PHLPP1α. (b) Densitometry of western blots in panel (a) (n = 3) showing pAKT473 levels had a tendency (p = 0.0776) to be higher in PHLPP1α over-expressing HEK293-FT cells. (c) pERK phosphorylation was significantly increased in PHLPP1α over-expressing cells (p = 0.0003).
As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer-reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.
Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.