The role of the PDK1/PKB kinases in regulating neuronal survival and differentiation: characterization of the PDK1 K465E knock-in mice

Abstract

Neuronal cell death programmes are counteracted by survival signals during development in order to maintain the tissue homeostasis. Neuronal differentiation is a mechanism generating functionally integrated neuronal cells from their progenitors. These processes appear to be mediated via activation of the Ras/Raf/MAPK and the PI3K/PDK1/PKB signaling pathways and are associated with a selective increase in protein translation. Protein kinase B (PKB/Akt) is a serine/threonine protein kinase which is claimed to be the critical transducer for several extracellular signals provided by different neurotransmitters, growth factors and hormones that promote phosphoinositide 3-kinase (PI3K) activation. PI3K is a lipid kinase characterized by its ability to phosphorylate the 3-OH group in the inositol ring of phospholipids at the inner side of the plasma membrane to generate phosphatidylinositol-3,4,5-trisphosphate (PtdIns(3,4,5)P3 or PIP3), which is a potent second messenger. PKB regulation by its activator PDK1 precisely relies on a PtdIns(3,4,5)P3 binding domain, named Pleckstrin Homology domain (PH-domain). Both PDK1 and PKB are protein kinases of the AGC family containing PH-domains which mediate their recruitment to the plasma membrane, where PKB is activated by phosphorylation at two regulatory residues, namely Thr308 at the T-loop by PDK1, and Ser473 at the hydrophobic motif by mTORC2. In fact, PDK1 was shown to be a master kinase also playing an essential role in the activation of a number of AGC family members by phosphorylating their T-loops by means of a PH domain-independent mechanism. Activated PKB modulates the function of numerous substrates involved in the regulation of cell metabolism, survival, proliferation and growth, which deregulation has consequences in pathologies such as diabetes, cancer and neurodegeneration. The crystal structure high resolution of the PDK1 PH domain revealed that the positivelycharged lateral chain of Lysine at position 465 within the PH domain crystal establishes fundamental interactions with the negatively-charged phosphate groups of PIP3. Targeted mutagenesis of Lysine 465 to the negatively-charged aminoacid Glutamic acid abolished binding of PDK1 to PIP3 by disrupting the phosphoinositide binding pocket. Therefore, it was thought that this mutation could be instrumental in ablating this part of the PDK1 signaling pathway. In order to analyze the role of the PDK1-PIP3 interaction in vivo, PDK1K465E/K465E knock-in mice were generated which physiologically express from the endogenous locus a mutant form of PDK1 incapable of phosphoinositide binding. This knock-in mice model was shown to be a good tool to analyse the contributory role of the PKB signaling pathway to glucose metabolism. The PDK1K465E/K465E mice were shown to be viable but smaller, with a modest reduction in PKB activity compared with the wild type littermates, and prone to diabetes. The importance of this pathway in tumourogenesis has been highlighted by introducing the PDK1 PH domain knock-in mutation into cancer-prone PTEN+/- mice, which resulted in the delayed tumour onset, suggesting that even moderate reduction of PKB activity can significantly delay tumour initiation and development. This makes the PDK1K465E/K465E mutant mouse model an excellent tool to explore the contribution of PKB to different human pathologies and to identify downstream substrates that could provide targets for therapeutic intervention. In particular, I aimed to use this genetic model to unravel the role of PKB on different aspects of brain development and function. Stereological analysis of embryonic brain sections showed that the PDK1K465E/K465E mice displayed reduced brain size due to a reduction in neuronal cell size rather than cell number, since the number of cortical and hippocampal neuronal populations between PDK1K465E/K465E and PDK1+/+ mice was not significantly different, whereas the volume of the mutant neuronal soma was approximately 80% of the volume of the wild type neuronal soma. Stimulation of cortical neurons with BDNF induced a robust phosphorylation of Trk receptors followed by the phosphorylation of PKB at Thr308 in the PDK1+/+ cells, which is blunted in the PDK1K465E/K465E neurons, whereas PKB phosphorylation at the mTORC2 site (Ser473) proceeded normally in both type of cells. The moderate reduction of PKB activation was not rate-limiting for the phosphorylation of those PKB substrates governing neuronal survival and apoptosis such as FOXO and GSK3. Then, it was questioned whether such mutation could affect survival responses in primary neuronal cultures. The findings from this study illustrate that the integrity of the PDK1 PH domain is not essential to support the survival of different embryonic neuronal populations analyzed. Cell viability is compromised after trophic factor deprivation, whilst BDNF treatment rescues cells from death to the same extent in both PDK1+/+ and PDK1K465E/K465E neurons. In contrast, the moderate reduction of PKB activity in the PDK1K465E/K465E neurons markedly reduced phosphorylation of the PRAS40 and TSC2 substrates, leading to decreased mTORC1/S6K activation and also reduced BRSK protein synthesis. The PDK1K465E/K465E neurons in culture showed reduced neurite outgrowth, delayed polarization and deficient axonogenesis. To establish the possible causal relation between the PKB pathway defects and axon formation, the impact of specific pharmacological treatments with PKB and mTORC1 inhibitors on neuronal differentiation were assessed, which provided strong evidence that the differentiation defects were due to reduced PKB activity and inefficient activation of the mTORC1 signaling. Moreover, the overexpression of BRSK isoforms rescued the axonogenesis defects of the PDK1K465E/K465E hippocampal cells. Altogether, these findings illustrate how the binding of PDK1 to PIP3 creates a PKB signaling threshold which is sufficient to support survival, but not differentiation of neuronal cells. In this regard, there is increasing evidence that PI3K/PDK1 dependent, PKB independent pathways might be responsible for the control of essential cellular processes, for example cell survival, which rely on other members of the AGC family activated by PDK1. These other PDK1-regulated members of the AGC family include SGK1, S6K and RSK. The activation of these kinases is not dependent on PDK1 binding to PIP3 and therefore they should be normally phosphorylated in the PDK1K465E/K465E knock-in mice neurons. However, I observed decreased phosphorylation of the SGK substrate NDRG1. This study clearly states for the first time, that NDRG1 is regulated by PKB, at least in neurons. Activation of S6K was found also incomplete in the PDK1K465E/K465E neurons due to reduced mTORC1 PKBdependent activation, which could be overcome by nutrients. In fact, the only PDK1 substrate analyzed that appears to not to be affected by the PDK1 K465E mutation is RSK, which serves as a control of the specificity of this knock-in mutation. In summary, the data allow to conclude that full activation of PKB is not essential in controlling neuronal survival. In marked contrast, reduced PKB-mediated, mTORC1- dependent, BRSK expression resulting from lack of PDK1-phosphoinositide binding prevents neuronal differentiation.