Phosphatase and TENsin homolog (PTEN) is a protein/lipid phosphatase that is responsible for regulating a myriad of signal transduction events in the brain including those essential for central nervous system patterning. It is an ancient component of many highly conserved signaling pathways including most notably the Phosphatidyl Inositide 3-Kinase (PI3K)/mammalian Target Of Rapamycin Complex 1 (mTORC1) pathway for which PTEN serves as a major antagonist (Figure 1). This article will explore the roles of this enzyme based on findings made primarily using genetic loss-of-function studies within the developing and adult nervous system and disciplines beyond where appropriate. Recent reviews have addressed the general cell biological functions of PTEN as well as the regulatory mechanisms controlling its activity and expression (Song et al., 2012; Ortega-Molina and Serrano, 2013). Many of its roles in mature nervous system physiology and pathology are tackled by other articles within this Research Topic. Here, we explore diverse aspects of brain development influenced by PTEN including progenitor cell proliferation, cell fate determination, migration, polarity, axon-dendrite morphogenesis, cell soma size, and myelination (Figure 2). The observations discussed in this review reveal PTEN's key roles in cellular signaling that have the potential to be harnessed via cellular, pharmacologic and genetic strategies to improve disease outcomes (Yang et al., 2009; Ortega-Molina and Serrano, 2013; Maire et al., 2014).

The control of cell cycle timing and differentiation in the developing brain are essential elements for populating the nervous system with the appropriate numbers and types of cells. Multiple studies have demonstrated a key role for PTEN in stem cell maintenance and fate specification utilizing Cre recombinase mouse strains to conditionally knockout (cKO) PTEN expression in various populations of cells in the embryonic retina and brain, early post-natal, and adult brain regions including the cerebellum, cortex, hippocampus. We have included an accompanying table listing the Cre lines discussed in this review along with their primary reference and expression patterns (Table 1). Examination of regulated loss-of-function models has resulted in many unifying themes of PTEN function and observations that implicate extra-cellular, trans-cellular or cell-type specific roles for PTEN. While Cre recombinases offer a tremendous asset for understanding tissue and cell-type specific gene function, it is important to consider that recent work indicates that the timing of recombination may not correspond with the expected promoter-based expression pattern (Liang et al., 2012; Harno et al., 2013).

The appropriate localization and distribution of stem cell progeny make cell migration a key event in proper brain formation. Given the central role of PTEN in the transduction of many pathways that shape cellular responses to chemo-attractant or -repellent cues, it would be reasonable to expect that certain aspects of cell migration would be perturbed following the loss of PTEN. Many studies of the developing brain find little indication of massive alterations in migration, but lamination defects have been observed following Nestin-Cre deletion (Groszer et al., 2001). However, it is unclear if this observation reflects defects in migration or in progenitor proliferation.

Neuronal helix-loop-helix protein-1 (Nex1)^Cre (Table 1) deletion of PTEN only targets neurons that have left the cell cycle and become fated as neurons. Immunohistochemical staining for lamination markers such as Tbr1 (early born neurons) indicated a less compact cerebral cortex with some Tbr1⁺ cells near the marginal zone, and labeling of later born neurons using BrdU birthdating or Cux1 staining appeared to have normal placement, indicating no major defect in migration or positioning (Kazdoba et al., 2012). Somewhat surprisingly, these mice did display an increased expression of Reelin, a secreted regulator of migration, relative to controls. It is possible that the lack of effect on migration could be due to inefficient NEX1^Cre-mediated deletion of PTEN in these cells. Additional analyses of cortical structure using Cre recombinases that target earlier steps in neurogenesis (Emx1^Cre and hGFAP-Cre) do find significant alterations in cortical organization and hippocampal radial glial persistence, respectively (Lehtinen et al., 2011; Wen et al., 2013).

As described above, PTEN-null cells migrating from the SVZ/SEZ along the RMS have been observed to fall into two subsets of cells—those that prematurely differentiate within the SVZ/SEZ, and those that continue to migrate to the OB (Zhu et al., 2012). Live imaging studies of this second group of cells indicated that their migratory speed was not statistically distinguishable from wild-type cells. The identity of the factors that distinguish these two classes of migrating cells may hold a wealth of information regarding chemotactic strategies employed by various RMS-transiting neuroblasts.

Early studies of conditional PTEN loss in the brain detected migration defects in the granule cells of the cerebellum (Backman et al., 2001; Kwon et al., 2001). Interestingly, while deletion of the Akt-activating kinase, Phosphoinositide-dependent kinase-1 (Pdk1), reverses some phenotypes observed in GFAP^Cre/PTEN cKO mice (see Somal Hypertrophy below), it does not rescue cerebellar granule cell migration defects (Chalhoub et al., 2009). Experiments using L7/PCP2-cre to target PTEN deletion in Purkinje cells of the cerebellum did not detect a major alteration in cell placement, but ablation of PTEN by En2-Cre resulted in a significant alteration in organization involving essentially all constituent cells of the cerebellum (Marino et al., 2002). Later work using hGFAP-Cre and viral delivery of Cre suggested that the migration defect observed for granule cells is the result of premature differentiation of Bergmann glia rather than a cell autonomous effect (Yue et al., 2005). Taken together, these results highlight the intrinsic diversity related to cell migration cues, cellular responses and the multivalent ways in which PTEN may shape them.

PTEN plays a role in acquisition of polarity by augmenting signal transduction from extracellular cues. Much of the work exploring the role of PTEN in polarity and axon outgrowth has focused on its phosphatase activity during membrane receptor signaling. Experiments specifically exploring the ability of PTEN to alter axon specification place it as a key element of axogenesis. Over-expression of PTEN results in a failure of axon formation and neurite outgrowth (Shi et al., 2003; Jiang et al., 2005). Conversely, RNA interference (RNAi) blockade of PTEN expression leads to the formation of multiple axon projections (Jiang et al., 2005). Caenorhabditis elegans PTEN ortholog daf-18 mutant neurons fail to properly polarize and extend axons (Adler et al., 2006), demonstrating the high degree of conservation across species. Additionally, in vivo loss-of-function studies using Neural Specific Enolase (NSE)-Cre to generate cKO PTEN mice indicate exuberant axon projections in the dentate gyrus (Kwon et al., 2006a). It will be important in future studies to assess whether these axonal defects are due to axon mis-targeting or altered neuronal polarity, but it is clear that axonal development is significantly altered following PTEN ablation in many types of neurons. While these studies make a compelling case for PTEN requirement during axon formation, these results do not delineate a specific role for PTEN in the early asymmetry events in axon specification or a role in regulation of axon outgrowth. Clarity for this morphologic aspect of PTEN function will await further investigation and new tools to probe the earliest events in neuronal polarization.

Beyond affecting early neurite specification, PTEN loss has been associated with alterations in both axonal and dendritic structure. The nuclear pool of PTEN may contribute to this process as well through its interaction with APC/CDH1. This complex has been shown to regulate axon formation and outgrowth by targeting transcriptional components for degradation. As described earlier, PTEN could be contributing to the timing of axon specification/outgrowth by increasing the association of APC and CDH1 (Lasorella et al., 2006; Stegmuller et al., 2006). The related complex of APC/CDC20 has been implicated in axonal outgrowth (De La Torre-Ubieta and Bonni, 2011), but whether PTEN associates with this complex as well remains to be determined. PTEN/daf-18 has also been shown in multiple species to be crucial for shaping of neuronal morphology by regulating the transcription factor FoxO/daf-16 (Christensen et al., 2011) and by controlling the PI3K-Akt signaling pathway.

Axon projections were more exuberant in both dopamine transport promoter (DAT^Cre) (Diaz-Ruiz et al., 2009; Domanskyi et al., 2011; Inoue et al., 2013) and Nse-Cre (Kwon et al., 2006a) deleted PTEN mouse lines. Adult newborn neurons respond to PTEN depletion much like embryonic cells, as retroviral siRNA in vivo increased axon diameter and bouton size (Luikart et al., 2011). However, midbrain dopaminergic neurons targeted for PTEN deletion using DAT^Cre did not display an increase in terminal bouton size (Diaz-Ruiz et al., 2009; Inoue et al., 2013). In contrast, PTEN deletion using Ca²⁺/Calmodulin-dependent protein kinase 2 promoter-driven (CaMKIIα)-Cre in mature neurons of the forebrain resulted in altered morphologies of neuronal processes (Sperow et al., 2012).

Work in Xenopus spinal neurons using pharmacologic inhibition of PTEN by Bisperoxo (1, 10-phenanthroline) oxovanadate, morpholino knock-down, or catalytically-inactive PTEN over-expression enhances axon growth when encountering target muscle tissue (Li and Peng, 2012). This suggests a role for PTEN in target recognition or down-regulation of axon growth signaling. Furthermore, spinal neuron growth cone chemotaxis has been shown to utilize PTEN selectively in response to chemorepulsive cues, demonstrating a potential context-dependent difference in PTEN modulated responses (Henle et al., 2013). Loss of the spinal motor neuron 1 gene (SMN1) leads to spinal motor atrophy and the axon outgrowth and growth cone defects associated with the loss of the SMN1 gene expression can be rescued by siRNA knockdown of PTEN (Ning et al., 2010). As the SMN1 protein plays a role in both small nuclear ribonucleic particle regulation and axonal RNA transport, this result may indicate that PTEN contributes to the regulation of SMN1-regulated translation for axonal RNA through regulation of mTOR signaling, or that the local translation of PTEN itself may be increased following the loss of SMN1, or both. During axon growth, PTEN may be the target of mRNA stability/translational regulation via the microRNA cluster 17-92 member miR19 (Figure 1) (Zhang et al., 2013), as described above in the regulation of neural stem cells.

Experiments directed at understanding the ubiquitin proteasome system in neurons led to the finding that the ubiquitin ligase NEDD4 is capable of targeting PTEN for degradation (Figure 1) (Drinjakovic et al., 2010). This work also demonstrated that PTEN degradation is crucial for the regulation of axon branching. Recently, adult dorsal root ganglion axons were also shown to utilize NEDD4 to regulate PTEN during axon growth (Christie et al., 2012). This work highlights the significance of localized regulation of protein expression and signaling responsiveness of cellular compartments such as growth cones.

Similarly, dendrites are also affected by loss of PTEN. Most studies of PTEN deletion in the nervous system that detected changes in somatic size (see below) also report primary dendrites of enlarged caliber. A systematic study of the PI3K-Akt-mTOR signaling cascade in dendrite development found that shRNA reduction of PTEN resulted in increased branching consistent with increased activity in this transduction cascade (Jaworski et al., 2005).

The importance of the mTORC1 signaling pathway and its regulation by PTEN are apparent in the cellular hypertrophy that results from PTEN loss (Backman et al., 2001). This effect is reversed in neurons by inhibitors of mTOR kinase activity (Kwon et al., 2003; Ljungberg et al., 2009; Zhou et al., 2009). Similar hypertrophy is also observed in astrocytes (Fraser et al., 2004), but there is some indication that this phenotype may be sensitive to the timing of PTEN loss. Embryonic deletion in post-mitotic neurons using NEX1^Cre (Kazdoba et al., 2012) as well as DAT^Cre mice (Diaz-Ruiz et al., 2009; Domanskyi et al., 2011; Inoue et al., 2013) resulted in somatic hypertrophy, but deletion in mature, post-natal forebrain neurons using CaMKIIα-Cre did not have this effect (Sperow et al., 2012). Paralleling this result, virally-mediated shRNA knock-down of PTEN in adult born dentate gyrus neurons resulted in somatic hypertrophy, but this effect is delayed relative to neonatally injected animals (Luikart et al., 2011). Epistasis experiments pursuing the components of the pathway responsible for hypertrophy have shown that loss of the kinase Pdk1 leads to reversal of hypertrophy in GFAP^Cre; PTEN^fl/fl mice (Chalhoub et al., 2009). Loss of S6K1, a key downstream effector of mTORC1 linking it to protein translation regulation, did not have the same effect (Chalhoub et al., 2006). The disruption of the regulated transport of PTEN in neurons by the motor protein Myosin V has also been shown to result in somatic hypertrophy (Figure 1) (Van Diepen et al., 2009). In this case, a direct interaction between these two proteins was shown to exist and to be regulated by either casein kinase 2 or glycogen synthase kinase 3 beta (GSK3β). These two studies demonstrate that a number of PTEN regulatory pathways can contribute to hypertrophy. An exact delineation of the mechanisms capable of influencing somatic hypertrophy await further studies to dissect the individual contributions of each PTEN regulator.

An important corollary to PTEN's effect on neuronal process outgrowth is the observation of myelination defects following PTEN loss (Fraser et al., 2008) and the hypermyelination by oligodendrocytes and Schwann cells lacking PTEN following deletion by oligodendrocyte transcription factor 2 (Olig2)^Cre (Harrington et al., 2010; Maire et al., 2014). 2', 3'-cyclic nucleotide 3' phosphodiesterase (CNP1)^Cre or Proteolipid protein 1 (Plp1) ERT2-Cre (Goebbels et al., 2010). Interestingly, rapamycin treatment of animals exhibiting altered myelination resulted in amelioration of the phenotype (Goebbels et al., 2010). This points to the lipid phosphatase activity of PTEN, and its antagonism of the PI3 kinase/mTOR pathway. A recent report has implicated PTEN as a key component of the polarized growth and wrapping of the myelin sheath at its inner tongue via regulation of the maturation state of the myelinating cell (Snaidero et al., 2014). Conditional deletion of PTEN using the Plp1-ERT2-Cre line resulted in a larger inner tongue and more cytoplasmic channels relative to controls, a sign of less mature myelinating sheets. Cytoplasmic-rich edges of the sheath in longitudinal sections that were absent in controls either 23 or 60 days post-natally were also seen (Goebbels et al., 2010; Snaidero et al., 2014). Deletion of PTEN at postnatal day 100 led to the reappearance of cytoplasmic myelin transport channels and increased myelination, relative to controls (Snaidero et al., 2014). These data indicate that it is possible to reinitiate the myelination program, even in mature animals, by increasing PI(3,4,5)P3 levels through loss of PTEN activity. This establishes PTEN both as a key regulator of embryonic development, and as a potential therapeutic target for myelination disorders and demyelinating diseases.

The profound effect on the inner tongue may be related to the localization of PTEN within the myelin sheet through interaction with scaffolding proteins such as Discs large homology 1 (Dlg1) or PAR-3. Reduction of Dlg1 has been shown to increase myelination and decrease levels of PTEN, suggesting a reciprocal stabilizing relationship between these two proteins (Cotter et al., 2010). However, neither Olig2^Cre nor CNP1^Cre PTEN cKO mice exhibit alterations in Dlg1 levels (Maire et al., 2014), (Goebbels et al., 2010), but there may be small, but significant, changes in signaling in these mice. Alternatively, the observed hypermyelination may reflect an alternate PTEN partner since myelination is also coordinated by a complex of the polarity protein PAR-3 and the neurotrophin receptor p75NTR (Chan et al., 2006). PTEN binds PAR-3 (Von Stein et al., 2005; Wu et al., 2007; Feng et al., 2008), and p75NTR receptor stimulation can increase PTEN activity (Song et al., 2010), thus it is possible that together as a tripartite complex these proteins could regulate axon ensheathment. However, much like Dlg1, localization of PAR-3 was not altered in CNP1^Cre; PTEN^flox/flox mice (Goebbels et al., 2012). While the identity of the molecules targeting PTEN within myelinating cells may not be resolved, the important role for PTEN/PI3K/mTOR signaling is clear.

Since its discovery almost 20 years ago, PTEN has remained a centerpiece of inquiry into signal transduction for a vast array of disciplines including tumor biology, metabolism, and neuroscience. Work over the last decade and a half has significantly strengthened our understanding of the major neurodevelopmental roles for PTEN. Going forward, a clearer understanding of the catalytic and non-catalytic functions of PTEN during brain development will be crucial, particularly in terms of transcriptional regulation. Similarly, the phosphatase activity linked to protein versus lipid dephosphorylation will expand our compendium of targets regulated by PTEN during nervous system patterning. Systematically perturbing PTEN interacting partners during brain development will allow us to begin to form a holistic view of how PTEN loss leads to the brain phenotypes observed.

It will also be important to develop “next-generation” tools with which to interrogate PTEN function in the developing brain. These might include chemical genetic variants of PTEN for rapid and specific elimination of phosphatase activity either in cell lines or mouse models, and inducible transgenic lines that permit cell-type specific rescue either with a doxycycline-regulated system or Cre-induced expression. To define functionality and understand the consequence of PTEN interacting with binding partners, point mutant variants should be developed that remove binding sites for particular partner proteins. Alternatively, systems can be devised that permit pharmacologic regulation of PTEN association with partner molecules. Forms of PTEN that are inducibly targeted to the sub-cellular locales known to be key sites of action for PTEN would also provide a mechanism to discern the various roles of PTEN within cellular compartments. Combinations of these and other future technologies will permit us to further delineate the PTEN-dependent mechanisms responsible for sculpting the developing and adult brain. Furthermore, the continued synergy with other disciplines that have already contributed significantly to our understanding of PTEN will continue to shape the field and provide new insights into PTEN biology enhancing our awareness of the role of PTEN in the brain.