Developmental Biology 338 (2010) 215–225

Contents lists available at ScienceDirect

Developmental Biology

j ourna l homepage: www.e lsev ie r.com/deve lopmenta lb io logy

Initiation of neuronal differentiation requires PI3-kinase/TOR signalling in thevertebrate neural tube

Katherine J. Fishwick, Roman A. Li, Pamela Halley, Peiyi Deng, Kate G. Storey ⁎Neural Development Group, Division of Cell & Developmental Biology, College of Life Sciences, University of Dundee, Dow Street, Dundee, DD1 5EH, Scotland, UK

⁎ Corresponding author.E-mail address: k.g.storey@dundee.ac.uk (K.G. Store

0012-1606/$ – see front matter © 2009 Elsevier Inc. Adoi:10.1016/j.ydbio.2009.12.001

a b s t r a c t

a r t i c l e i n f o

Article history:Received for publication 19 June 2009Revised 11 November 2009Accepted 1 December 2009Available online 11 December 2009

Keywords:Vertebrate neurogenesisChickMousePI3-kinaseTORRapamycinNeuronal differentiationCell cycleTuberous sclerosis

Regulated neuron production within the vertebrate nervous system relies on input from multiple signallingpathways. Work in the Drosophila retina has demonstrated that PI3-kinase and downstream TOR signallingregulate the timing of photoreceptor differentiation; however, the function of such signals during vertebrateneurogenesis is not well understood. Here we show that mutant mice lacking PKB activity downstream ofPDK1, the master kinase of the PI3-kinase pathway, exhibit deficient neuron production. We furtherdemonstrate expression of PI3-kinase signalling components and active PKB and TOR signalling in the chickspinal cord, an early site of neurogenesis. Neuron production was also attenuated in the chick neural tubefollowing exposure to small molecule inhibitors of PI3-kinase (LY294002) or TOR (Rapamycin) activity.Furthermore, Rapamycin repressed expression of early neuronal differentiation genes, such as Ngn2, but didnot inhibit expression of Sox1B genes characteristic of proliferating neural progenitors. In addition, somecells expressing an early neuronal marker were mis-localised at the ventricular surface in the presence ofRapamycin and remained aberrantly within the cell cycle. These findings suggest that TOR signalling isnecessary to initiate neuronal differentiation and that it may facilitate coordination of cell cycle anddifferentiation programmes. In contrast, stimulating PI3-kinase signalling did not increase neuronproduction, suggesting that such activity is simply permissive for vertebrate neurogenesis.

© 2009 Elsevier Inc. All rights reserved.

Introduction

Generation of the complex vertebrate nervous system takes placeover an extended period of time and relies on the regulatedproduction of neurons. PI3-kinase signalling controls the timing ofneuronal differentiation in the simpler fly retina (Bateman andMcNeill, 2004), but the role of this pathway during vertebrateneurogenesis is less clear. Insulin and insulin-like growth factors(IGFs) are ligands that promote PI3-kinase signalling via IGFRs andthese have been shown to work, along with other factors, tostimulate vertebrate neural progenitor cell proliferation in vivo andin vitro (reviewed in D'Ercole et al., 2002). Importantly, this results inan increase in neuron numbers, so although IGF/PI3-kinase signallingcan shorten the neural progenitor cell cycle and maintain these cellswithin the cycle (Hodge et al., 2004), this does not block eventual cellcycle exit and neuronal differentiation. PI3-kinase signalling couldtherefore simply expand the neural progenitor pool and therebygenerate more neurons or it might additionally play a role inpromoting neuronal differentiation. Interestingly, excess IGF signal-ling differentially affects distinct regions of the developing brain, forexample, many more neurons are generated in response to this factor

y).

ll rights reserved.

in cortical layer I than in layer VI (Hodge et al., 2005). IGF alsoinduces proliferation or differentiation of mouse striatal primaryneural stem cells depending on passage number and co-factors suchas BDNF (Arsenijevic and Weiss, 1998; reviewed in Bateman andMcNeill, 2006) and can promote survival, proliferation and alsodifferentiation of neuroblasts in the otic placode (Camarero et al.,2003). Different levels of PI3-kinase signalling can also elicit distinctcell behaviours, for example, in olfactory bulb stem cells reducedbasal level signalling decrease neuron and astrocyte numbers butdo not alter cell proliferation or survival, while more completeinhibition reduces both proliferation and neuron and astrocytenumbers and increased incidence of cell death (Otaegi et al., 2006).These data indicate that PI3-kinase signalling promotes proliferationor differentiation of neuroepithelia depending on signalling contextand suggest that this may reflect the level at which this pathwayis active.

It has been demonstrated that PI3-kinase acts in the neuroepithe-lium via PKB (Akt) downstream signalling, for example mice lackingPKBα and PKBγ have smaller brains and fewer cells (Easton et al.,2005; Tschopp et al., 2005; reviewed in Ye and D'Ercole, 2006). Afurther interconnecting, downstream pathway is the TOR complex 1(TORC-1) pathway, which regulates cell growth and translation(Fingar and Blenis, 2004; Gingras et al., 2007). In the fly retina,TOR/S6 kinase activity has also been shown to regulate the timing ofneuronal differentiation and so serves here to coordinate tissue

216 K.J. Fishwick et al. / Developmental Biology 338 (2010) 215–225

growth with developmental decisions (Bateman and McNeill, 2004;McNeill et al., 2008). A number of transgenic mouse models provideevidence for a TOR activity requirement during vertebrate neuraldevelopment: mice with reduced TOR levels fail to undergotelencephalic (forebrain) expansion (Hentges et al., 2001; Hentgeset al., 1999) while rats lacking TSC2, a negative regulator of TORsignalling, have an expanded forebrain (Rennebeck et al., 1998). Theseobservations indicate that TOR signals mediate proliferation in thisbrain region, however, recent work has also demonstrated arequirement for the TOR pathway for neuronal differentiation intelencephalon-derived progenitors in vitro (Han et al., 2008).Elevated TOR signalling underlies the genetic disease tuberoussclerosis caused by mutation of TSC1 or 2 genes and characterisedby tuber-like growths that destroy tissue architecture in neural andother tissues. Tubers contain enlarged “balloon-like” cells but alsodifferentiated cells and may form as a result of aberrant neuronalmigration during development (Mizuguchi and Takashima, 2001;reviewed in Crino et al., 2006). This phenotype is consistent withmultiple roles for TOR signalling during neurogenesis.

The developing hindbrain and spinal cord are the earliest sites ofvertebrate neurogenesis and the timing and regulation of neuronbirth in these regions has been studied in detail (Jessell, 2000; Priceand Briscoe, 2004; Sechrist and Bronner Fraser, 1991). Here wedemonstrate a requirement for PI3-kinase signalling for neuronproduction in this region of the neural tube in both mouse and chickembryos. Inhibition of TOR signalling with Rapamycin (a specificTOR/TORC1 inhibitor; reviewed in Bain et al., 2007; Hidalgo andRowinsky, 2000) attenuates neurogenesis in chick neuroepitheliumexplants in vivo and in vitro by inhibiting expression of genes thatinitiate neurogenesis. Exposure to Rapamycin can also disruptcoordination of cell cycle exit and progression of the neuronaldifferentiation programme and interfere with neuroepithelial orga-nisation. Stimulating PI3-kinase signalling, however, does notincrease neuron production in explanted neural tube, indicatingthat this signalling pathway does not drive neurogenesis in thiscontext.

Materials and methods

Experimental animals

Mice used in this study carried a knock-in mutation in the PHdomain of PDK1 as (PDK1-PH KI; described in McManus et al 2004).Chick embryos were obtained from fertilised chicken eggs RhodeIsland Red x High Sex (Winter Farm, Hertfordshire, UK) and incubatedat 38 °C for 29–36 h to yield embryos of HH8–10 (Hamburger andHamilton, 1951).

Immunocytochemistry

Standard methods were used for immunohistochemistry onwhole-mount embryos, sections or explants. Nuclei were counter-stained using DAPI or propidium iodide (Sigma). Primary antibodiesused: neurofilament-associated antigen (3A10) (DevelopmentalStudies Hybridoma Bank), phospho-PKB (Ser473), phospho-S6(Ser235/236), cleaved caspase 3 (Cell Signalling), HuC/D (MolecularProbes), phospho-H3 (Upstate), Tuj-1 (Babco/Covance) and BrdU(Sigma). Fluorescent-conjugated secondary antibodies used (1:1000):anti-rabbit Alexa488, anti-mouse Alexa647, anti-rabbit Alexa594 andanti-guinea-pig Alexa488 (Molecular Probes).

In situ hybridisation

In situ hybridisation was carried out according to a standardprotocol, after Wilkinson and Nieto (1993). Plasmids for analysisof key genes were kindly provided as follows: Ngn2 (A. Graham)

Pax6, Sox2, Sox3, Delta-1 (D. Henrique), cIGF2 (P. Rotwein) and cIGFR-1 (G.Allan).

Chick neural tube explant culture, in ovo grafts and drug delivery

Neural tube explants were taken as described in Diez del Corral etal. (2002). Explants were always taken as pairs from the sameembryo, cultured +/− reagents/inhibitors and processed in parallel.Culture media contained OptiMEM (Gibco), glutamine and antibio-tics to which was added 20 μM LY294002 (Calbiochem), 1 μMRapamycin (Calbiochem), 10 or 100 μg/ml insulin (Sigma). Smallmolecule inhibitors were dissolved in DMSO, which was also addedto control media. Western blotting was performed as described inLunn et al. (2007) with embryos cultured in media identical to thatused for explants for 2 h at 37 °C, 5%CO2 before lysis and Westernblotting using standard techniques. Formate derived AGX1-2 beadswere prepared for grafting by soaking for 2+ h in either DMSO or 0.5 mg/ml Rapamycin. Bead grafting was carried out at HH10 in ovo,via a small window in the eggshell, through which the vitellinemembrane was reflected to allow access to the embryo. In in ovoexperiments, the bead was grafted into the neural tube at the level ofthe most recently formed somites. Alternatively, beads were graftedvia the ventral surface into embryos prepared in EC culture(Chapman et al., 2001) and positioned beneath the paraxialmesoderm in contact with the neuroepithelium, just caudal to themost recently formed somites.

In vivo electroporation

Standard in ovo electroporation techniques (Itasaki et al., 1999)were used to transfect one side of the neural tube at HH8–9 with aplasmid containing S6K (pMT2 (CMV) HA-p70S6 Kinase) in either aconstitutively active (T412/E) or an S6K kinase-dead (T412/A)construct, as described in Balendran et al. (1999).

Quantitative RT-PCR

Pairs of neural tube explants from 10 embryos were cultured ineither 1 μMRapamycin or DMSO (0.185%) for 24 h. RNAwas extractedfrom explants using “Total RNA Isolation” kit (Macherey-Nagel)according to manufacturer's instructions. Purified RNA was used astemplate for cDNA synthesis using “ImProm-II Reverse TranscriptionSystem” (Promega) according to manufacturer's protocol. Briefly,random primers and RNA mix were thermally denatured at 70 °C for5 min and chilled on ice for 5 min. Reverse transcription mix(containing reaction buffer, reverse transcriptase, MgCl, dNTPs andribonuclease inhibitor) was added and primers were allowed toanneal at 25 °C for 5 min. First-strand synthesis was then carried outat 42 °C for 1 h and enzymes were inactivated at 70 °C for 15 min.Resulting cDNAwas amplified using Eppendorf realplex2Mastercyclerwith the following primers: βActin Fw CCAGCCATCTTTCTTGGGTA RvATGCCAGGGTACATTGTGGT, Histone3 Fw GAGATCCGTCGCTACCAGAARv AACAGACCCACCAGATACGC, Pax6 Fw GGCAGAAGATCGTGGAACTCRv TTTGATTGTCCAGCACTTGG, Sox2 Fw AAACAGCACCACGAGTTTCCRv ACCTCGGGAAGAAAGGAAGA. Obtained expression levels of genesof interest were normalised to βActin or Histone3 expression andaveraged and are presented in arbitrary units±SEM.

Cell counts and statistical analyses

All counts for cells labelled with key markers are presented as anaverage percentage of total cell number in individual neural tubeexplants or per transverse section of the neural tube for in vivoanalyses. Student T-test (paired or unpaired as indicated) was carriedout and SEM is presented for all statistical analyses, except wherestated otherwise (see figure legends).

217K.J. Fishwick et al. / Developmental Biology 338 (2010) 215–225

Results

Neuronal differentiation is defective in mice lacking PI3K signalling

Neuronal differentiation was analysed in mice possessing aknock-in inactivating mutation in the PH domain of PDK1(McManus et al., 2004). This prevents PDK1 binding to PIP3 andleads to failure to activate PKB and to phosphorylate PKBdownstream targets including, TSC2, GSK-3α/β, FOXO-1 and -3,WNK1 and PRAS-40. In addition, these animals have reduced S6K1activity in response to IGF1 and the mutant protein is unstable andis expressed at 10–20% of normal PDK1 levels (McManus et al.,2004). Unlike PDK1 null embryos, which die at early somite stages(Lawlor et al., 2002), the homozygous PDK1-PH KI mutationsurvives at least until E10.5 (with no embryos found after E11.5)and so provides an opportunity to examine the requirement for PKBsignalling during early neural development. Embryos at E9.5 wereanalysed in the first instance by assessing expression of aneurofilament-associated protein, which identifies neurons. Thisrevealed that these embryos have fewer neurons and axon tractsthan heterozygous littermates (n=2/2 PDK1-PH KI and n=2/2heterozygous embryos, Figs. 1A–B′). A few neurons are detected inthe hindbrain region posterior to the otic vesicle in homozygousmutant embryos but this is severely reduced compared with

Fig. 1. Neuronal differentiation is defective in mice lacking PKB signalling. Neurofilament detwhite box in B) PDK1-PHD KI homozygotes (n=2 embryos).Ngn2 expression in (C, C′ transvTS of D) (n=2 embryos). Activated caspase 3 (aCas-3) detection in (E) wild-type littermatesstaining of G. Arrowheads indicate an occasional dying cell in the neural tube of WT or mutanthe hindbrain/spinal cord examined in each of 2WT (10 aCas-3 cells in 28 sections) and 2 PDin the first pharyngeal arch of both PDK1-PHD KI homozygotes, but not in WT embryos (da

heterozygous embryos (Figs. 1A′, B′). To investigate this defectfurther, we next assessed expression of the proneural gene Ngn2, akey transcription factor expressed in neural progenitors thatpromotes neuron production, reviewed in Bertrand et al. (2002).In contrast with heterozygous littermates, Ngn2 transcripts inhomozygous mice were only detected in a few cells in the hindbrainand were absent in the spinal cord (n=2/2 PDK1-PH KI and n=2/2 heterozygous embryos, Figs. 1C–D′). PI3Kinase/PKB signalling iswell known to promote cell survival in many contexts; however, asin wild-type embryos (Fig. 1E), few cells positive for activatedcaspase 3 indicative of apoptotic cells were detected in the neuraltube of PDK1-PH KI homozygous mutant mice (n=2/2 PDK1-PH KIand n=2/2 wild-type embryos; Figs. 1F, G, G′). Many activatedcaspase-3-expressing cells were, however, found in the firstpharyngeal arch of homozygotes (Figs. 1G, G′) consistent with thepreviously reported malformation of this region in these mice(McManus et al., 2004). These findings suggest that cell deathcannot account for the neuronal differentiation defects apparent inPDK1-PH KI mice. However, these embryos die not long after thetime that neuronal differentiation commences during normaldevelopment and so although we observe defective neuronproduction we cannot rule out the possibility that this is due to ageneral developmental delay rather than a direct and specificrequirement for PI3-kinase signalling within the neuroepithelium.

ection in (A, A′, white box in A) PDK1-PHD KI heterozygotes (n=2 embryos) and (B, B′,erse section (TS) of C) heterozygotes (n=2 embryos) and in PDK1-PH KI embryos (D, D′(n=2 embryos) and PDK1-PHD KI homozygotes (F and G) (n=2 embryos), (G′) DAPIt embryos (images are representative of sections through the neural tube at the level ofK1-PHD KI homozygotes (4 aCas-3 cells in 23 sections), massive cell death was detectedta not shown).

218 K.J. Fishwick et al. / Developmental Biology 338 (2010) 215–225

PI3-kinase/TOR signalling is specifically required in chickneuroepithelium for neuron production

To examine the requirement for PI3-kinase signalling withinthe neuroepithelium, we next exploited the early chick embryo,which is amenable to explant culture and in vivo manipulations.First, we confirmed that a PI3-kinase stimulating growth factorIGF2 and an IGF receptor were expressed in the chick embryo asthe neural tube forms at Hamburger and Hamilton stages HH8–10.IGF2 is expressed in future telencephalic and diencephalic regionsof the nervous system and in medial somites and at a low levelwithin the developing hindbrain and spinal cord (Figs. 2A–A‴) andIGFR1 is detected at low level throughout the central nervoussystem and somites and strongly in the notochord at these stages(Figs. 2B–B‴), consistent with previous analysis in whole-mount(Allan et al., 2003). PI3-kinase signalling is also active in the chickneural tube, as indicated by immunocytochemical detection ofphosphorylated PKB (on Serine 473) and phosphorylated S6ribosomal protein, a downstream target of PI3K/mTOR signalling(Figs. 2C, D) and see summary pathway diagram (Fig. 2E). To testthe requirement for this pathway for neuronal differentiation,bilateral pairs of neural tube explants were dissected from stageHH8–9 and cultured for 24 h in the presence or absence of thePI3-kinase inhibitor LY294002 (Vlahos et al., 1994). In contrast tothe PDK1-PHD knock-in mutants, this drug should affect allsignalling activity downstream of PI3-kinase. To confirmLY294002 activity and to determine the concentration range foruse in embryos, the impact of this drug on PKB phosphorylationwas assessed by Western blot following treatment of wholeembryos (Fig. 2F). To confirm LY294002 action in explantedtissues, explants were treated for 2 h with 100 μM LY294002;pPKB was clearly detected in the cytoplasm of cells in controlvehicle (DMSO) only treated explants, but this was completelyabolished in the presence of LY294002 (Figs. 2G, H). However, forlonger incubations, the lowest effective concentration (20 μM) wasused to reduce the incidence of cell death (see below). Neural tubeexplants cultured in the presence or absence of 20 μM LY294002for 24 h were assessed for expression of the pan-neuronal markerHuC/D (Wakamatsu and Weston, 1997). This analysis indicatesthat fewer HuC/D-positive cells are present when PI3-kinasesignalling is reduced, n=4 explant pairs (Figs. 2I, J). To excludethe possibility that an increase in apoptotic cells accounts for thereduction in neuron numbers (assessed below), HuC/D-positivecells were quantified as a percentage of the total number of cellsin each explant (Fig. 2K). This reduction in HuC/D expressing cellsin neural tube explants demonstrates a requirement for PI3Kinasesignalling within the neuroepithelium for the normal productionof neurons.

PI3-kinase/PKB signals regulate a large number of downstreampathways, including TOR signalling (Fingar and Blenis, 2004). Neuraltube explants were therefore next exposed to Rapamycin. To identifya Rapamycin concentration that inhibits TOR signalling in chicktissue, we assessed phosphorylated S6 levels in whole embryos andin neural tube explants following exposure to this drug (Figs. 3A–C).We also determined that, in comparison with the effects ofLY294002, n=10 explant pairs (Figs. 3D–E′), treatment withRapamycin induced less apoptosis in neural tube explants (Figs.3F–G′) and no statistical difference in numbers of cells expressingactivated caspase 3 was detected between DMSO and Rapamycinexplant pairs, n=10 explant pairs (Supplementary Fig. 1). We thenassessed the effects of Rapamycin on numbers of HuC/D-positivecells, determined as above. This revealed a reduction in neuronnumbers, n=13 explant pairs (Figs. 3H–J). Together these studiessuggest that TOR activity, a downstream consequence of PI3Ksignalling, is required within the neuroepithelium to promoteneuron production.

TOR inhibition attenuates neurogenesis and increases mitoticindex in vivo

To examine the effects of TOR inhibition in an in vivo context,beads soaked in Rapamycin or vehicle control (DMSO) were graftedinto the developing neural tube in whole chick embryos at the levelof the last formed somite. We first established that Rapamycinbeads reduced pS6 at the 24-h time point (Figs. 4A, B); however, at36 h pS6 levels were normal again (data not shown), indicatingthat this experimental procedure provides effective drug delivery atleast up to 24 h. We also confirmed that Rapamycin did not elicitan increase in cell death in this in vivo assay as indicated bycaspase 3 activity at 24 h, n=3 DMSO and n=2 Rapamycin beadimplanted embryos (Figs. 4C, D). The number of HuC/D-positivecells was then quantified, and this revealed fewer neurons local tothe Rapamycin bead, n=9 DMSO and n=14 Rapamycin beadimplanted embryos (Figs. 4E–G) and at a distance of 1000 μmcaudal to the bead neuron numbers, were still significantly reduced(data not shown). The specificity of Rapamycin action in thiscontext was also confirmed by a rescue experiment in which wemis-expressed a constitutively active form of the TOR target S6K ora control kinase-dead S6K (Balendran et al., 1999) in cells on oneside of the neural tube, followed by exposure to a Rapamycinsoaked bead for 18 h. These experiments revealed that expressionof active S6K, but not kinase-dead S6K, could maintain pS6 levelson Rapamycin exposure (Supplementary Figs. 2A–F) and rescue thedecrease in HuC/D-positive cells induced by this small molecule(Supplementary Figs. 2G–I). These data confirm that Rapamycin isacting via its inhibition of TOR to attenuate neuronal differentiation.Overall, these findings indicate that inhibition of TOR signallinginterferes with neuron production in an in vivo as well as in vitrocontext in the chick embryo.

As neuronal differentiation involves cell cycle exit, the effects ofRapamycin on proliferationwere examined. Embryoswere graftedwithDMSO or Rapamycin beads as above and the number of mitotic cellsquantified using an antibody against phospho-H3 (pH3). A smallincrease in mitotic cells was found in the presence of Rapamycin,n=8DMSO andn=9Rapamycin bead implanted embryos (Figs. 4H–J).We next collected cells from neural tubes exposed to Rapamycin orDMSO beads after 24 h and analysed these for cell cycle phasedistribution using flow cytometry. These data indicate a slight trendtowardsdecreasednumbers of cells inG1and increasednumbersof cellsin G2/M within this short time frame (Supplementary Fig. 3). Overall,these data suggest that exposure to Rapamycin in this assay maintainscells in the cycle, consistent with decreased neuron number.

TOR signalling is required at the initiation of the neurogenesis cascade

We next sought to establish at what point in the neurogenesisprogramme TOR signalling is first required. Rapamycin beads weregrafted as above, at the level of themost recently formed somite or justcaudal to this in contact with the newly formed neuroepithelium, aposition that approximately coincideswith the onset of neurogenesis inthe extending body axis (Diez del Corral et al., 2002). The Notch ligand,Delta1, is expressed by newly born neurons in the neural tube and after16 h fewer Delta-1-positive cells were observed close to Rapamycinbeads, n=0/3 DMSO and n=12/14 Rapamycin bead implantedembryos (Figs. 5A, A′). One of the earliest expressed genes within thecascade of transcription factors leading to neuronal differentiation isNeurogenin-2 (Ngn2) (Bertrandet al., 2002). After 16h,Ngn2 transcriptsare also lower in the vicinity of Rapamycin beads, n=0/3 DMSO andn=4/6 Rapamycin bead implanted embryos (Figs. 5B, B′).

Pax6 is a direct cross-regulatory target of Ngn2 (Scardigli et al.,2003) (Scardigli et al., 2001) that primes neural progenitors forneuronal differentiation (Bel-Vialar et al., 2007) and consistent withthis Rapamycin beads also inhibit Pax6 expression, n=0/6 DMSO and

Fig. 2. PI3-kinase signalling is active in chick neural tube and attenuation of this pathway inhibits neuronal differentiation. In situ hybridisation for IGF2 (A, A′–A‴) and IGFR1 (B, B′–B‴)in the stage 9/10 chick embryo. Black bars indicate level of transverse sections. Immunocytochemical detection of phospho (p) PKB (Ser473) (C) and pS6 (Ser235/236) (D) inHH10 chick neural tube; Summary diagram of PI3-kinase pathway and some downstream targets (E); Western blot to detect pPKB in HH10 chick embryos following culture in arange of LY294002 concentrations for 2 h (F); schematic of HH8 chick embryo neural tube (NT) explants indicated by pink boxes and HH8 NT explant pair from same embryolabelled with pPKB (Ser473) cultured in DMSO (G) or 100 μM LY294002 (H) for 2 h; HH8 NT explant pair cultured in DMSO (I, I′) or 20 μM LY294002 (J, J′) for 24 h labelled withHuC/D and stained with DAPI; average percentage of HuC-positive cells in DMSO or LY294002-treated NT explant pairs (n=4 explant pairs) (K) ⁎p=b0.05). Scale bars, inC=75 μm, in F=100 μm.

219K.J. Fishwick et al. / Developmental Biology 338 (2010) 215–225

n=4/7 Rapamycin bead implanted embryos (Figs. 5C, C′). As bothNgn2 and Pax6 are characteristic of neural progenitors, we nextinvestigated whether TOR signalling regulates the general neural

progenitor cell state, as indicated by expression of pan-neuralprogenitor markers Sox3 and Sox2. Rapamycin beads do not reduceSox3 or Sox2 transcripts within the early spinal cord (transcript

Fig. 3. Calibration of Rapamycin inhibition of TOR signalling in chick embryo, comparison of cell death following inhibition of PI3-kinase and TOR signalling and inhibition of TORattenuates neuronal differentiation. Western blot for pS6 (Ser235/236) in HH10 chick embryos cultured in Rapamycin for 2 h (A); HH8 NT explant pair labelled with pS6 cultured inDMSO (B) or 1 μMRapamycin (C) for 24 h; HH8 NT explant pairs cultured for 24 h in DMSO (D, D′) or 20 μMLY294002 (E, E′) (representative of 10 explant pairs) and DMSO (F, F′) or1 μM Rapamycin (G, G′) assessed for expression of activated caspase3 (Cas3) and stained with DAPI (representative of 3 explant pairs). HH8 NT explant pairs cultured for 24 h inDMSO (H, H′) or 1 μM Rapamycin (I, I′) assessed for expression of HuC/D and stained with DAPI. Average percentage HuC/D-positive cells in 13 DMSO/Rapamycin-treated NTexplant pairs ⁎p=b0.05 (J). Scale bar in B applies to B–I′=100 μm.

220 K.J. Fishwick et al. / Developmental Biology 338 (2010) 215–225

reduction: Sox3, n=1/3 DMSO and n=0/9 Rapamycin beadimplanted embryos; Sox2, n=1/11 DMSO and n=3/15 Rapamycinbead implanted embryos; Figs. 5D, D′ and E, E′). In a slightly later assay,beads graftedmore rostrally into the newly formedneural tube atHH10also elicited a reduction in Ngn2 and Pax6, while Sox2 expression wasconsistently unaffected (transcript reduction: Ngn2 n=1/11 DMSOand n=11/16 Rapamycin; Pax-6 n=0/8 DMSO and n=9/12Rapamycin; Sox-2 n=0/7 DMSO and 0/9 Rapamycin bead implantedembryos) (data not shown). Similarly, neural tube explants (excised asabove) cultured in the presence of Rapamycin showed reduced Ngn2expression relative to their DMSO-treated counterparts, n=9/11

explant pairs (Figs. 5F, F′) but showed no difference in Sox2 levels inall cases, n=8 explant pairs (Figs. 5G, G′). Gene expression levels inexplants were further determined by quantitative PCR. This revealed asignificant reduction of Pax6 transcripts in Rapamycin-treated explants,but little effect on Sox2 expression (normalised to actin or Histone-3levels, Fig. 5H), in agreement with in situ hybridisation analysis. AsNgn2 is expressed at the start of the neurogenesis cascade and at highlevels promotes expression of Delta-1 (Bertrand et al., 2002), thesefindings suggest that TOR-mediated signalling is not required forexpression of these pan-neural progenitor markers but contributes tothe initiation of the neuronal differentiation programme.

Fig. 4. Rapamycin inhibits neuronal differentiation and increases mitotic index in vivo. Chick neural tube at HH10 at the level of a DMSO (A) or Rapamycin soaked bead (B) (sectionsrepresentative of 3 DMSO and 4 Rapamycin bead-grafted embryos, dotted lines indicate bead position) assessed for expression of pS6. Detection of activated caspase 3 and DAPIstaining in neural tube sections in the region of DMSO (C) or Rapamycin (D) beads; sections representative of 3 DMSO (12 sections, 3 aCasp3 cells) and 2 Rapamycin-treated embryos(8 sections, 0 aCasp3 cells). Detection of HuC/D near DMSO (E) or Rapamycin (F) beads in the neural tube. Average percentage of HuC+ cells in neural tube sections counted at thelevel of DMSO (11 sections, 9 embryos) or Rapamycin beads (23 sections, 14 embryos), ⁎⁎p=b0.001 Mann–Whitney U test (G). Detection of pH3-positive cells near DMSO (H) orRapamycin (I) beads in the neural tube. Average percentage of pH3+ cells in neural tube sections counted at the level of DMSO (8 embryos, 8 sections) or Rapamycin beads (9embryos, 14 sections) and 1 mm from the caudal to the bead DMSO (4 embryos, 5 sections) and Rapamycin (6 embryos, 10 sections) ⁎p=b0.05 Student's T-test (J). Scale bars in A(applies to A, B and E–H)=100 μm and in C, D=50 μm.

221K.J. Fishwick et al. / Developmental Biology 338 (2010) 215–225

Rapamycin treatment can dissociate cell cycle and neuronaldifferentiation programmes

Proliferating and differentiating cells are located in distinctpositions across the medial–lateral axis of the neural tube. Asneuroepithelial cells progress through the cell cycle they undergointerkinetic nuclear migration (Frade, 2002) with cells transitingmitosis at the ventricular surface and the nuclei of their daughtercells reaching the G1/S checkpoint at the lateral surface where theyeither re-enter or exit the cell cycle. Post-mitotic cells thereforeaccumulate at the lateral edge of the neural tube, forming the mantlelayer. Viewed from the dorsal surface in whole embryos, this spatialpattern is clearly visible with mitotic cells at the midline and cellsexpressing neuronal differentiation markers at the lateral edges ofthe neural tube. Strikingly, examination of embryos grafted withRapamycin beads viewed in whole-mount and in sections revealedcells at the midline expressed the early neuronal marker βIII-tubulin(Tuj-1) more frequently than controls (n=14 DMSO and n=17Rapamycin bead implanted embryos; see Figs. 6A–E). TuJ-1 isnormally expressed in cells as they commence neuronal differenti-ation in this tissue, entering G1 and moving away from theventricular surface (Hammerle et al., 2002). We therefore nexttested whether these ectopic Tuj-1 cells are simply mis-localisedpost-mitotic cells or if they continue within the cell cycle. MidlineTuj-1-positive cells were assessed for BrdU incorporation following1 h treatment, which should label cells in S-phase/early G2, but is notlong enough to label cells that have progressed to G1. Surprisingly,some midline TuJ-1-positive cells were found to incorporate BrdU(Figs. 6F–H). This indicates that aberrantly located Tuj-1-positivecells cycle at the ventricular surface despite expression of this earlyneuronal differentiation gene. These observations suggest that PI3K/TOR signalling helps to coordinate the neuronal differentiationprogramme with cell cycle regulatory mechanisms.

Stimulating PI3K signalling does not promote neuronal differentiation

We next sought to establish whether neuron production could beincreased by stimulation of this pathway. Neural tube explants

(excised as above) were cultured in media supplemented with insulinto stimulate PI3-kinase signalling. To confirm that the concentrationsof insulin used were sufficient to promote signalling through thispathway, whole embryos were cultured on filters in either controlmedia or increasing concentrations of insulin. Both 10 and 100 μg/mlinsulin caused a large increase in the levels of phosphorylated PKB(Fig. 7A). However, when explants grown in either concentration ofinsulin were compared to their control-treated halves, there was nodifference in the percentage of HuC-positive cells, n=5 explant pairs(Figs. 7B–F). This indicates that neuronal differentiation cannot beincreased by stimulating PI3-kinase activity and suggests that neuronproduction within the vertebrate neural tube requires, but is notdriven by signalling through the PI3-K/TOR pathway.

Discussion

Here we identify a requirement for PI3-kinase/TOR signallingduring neurogenesis in the developing mouse and chick neural tube;mice lacking PDK1/PKB activity and chick neural tube exposed to PI3-kinase or TOR inhibitors have fewer neurons. Consistent with this,treatment with Rapamycin maintains cells in the cell cycle andinhibits expression of early neuronal differentiation genes. However,Rapamycin exposure generates a complex phenotype that alsodisrupts neuroepithelial organisation; some cells expressing aneuronal marker are mis-localised at the ventricular surface andcontinue to cycle, indicating that differentiation and cell cycleprogrammes can become uncoordinated. Stimulation of PI3-kinasesignalling, however, does not increase neuronal differentiation,indicating that such activity, although necessary, is not sufficient todrive this process.

A central finding of this study is that signalling via PI3-kinase andTOR is required for neuronal differentiation in the vertebrate neuraltube; we show that constitutive reduction of PDK1/PKB signallinginhibits Ngn2 onset and neuron production in the developing mouseembryo and that blocking TOR signalling, using Rapamycin, directlywithin chick neuroepithelium reduces Ngn2 expression (as well asother neurogenesis genes, Pax6 and Delta1in the neural tube) andneuron production. The specificity of Rapamycin as an inhibitor of

Fig. 5. Inhibition of TOR signalling inhibits early neuronal differentiation genes. Expression patterns of key neurogenesis pathway genes following grafting of DMSO or Rapam soaked beads to the chick pre-neural tube, scored for reductionof gene expression; analysis of Delta1 (A–A′) (DMSO beads n=0/3, Rapamycin 12/14); Neurogenin 2 (Ngn2) (B, B′) (DMSO beads n=0/3, Rapamycin 4/6); Pax6 (C, C′) SO beads n=0/6, Rapamycin 4/7); Sox3 (D, D′) (DMSO beadsn=1/3, Rapamycin 9/9 unaffected) and Sox2 (E, E′) (DMSO beads n=0/3, Rapamycin 5/7, unaffected); also see Supplementary Fig. 1 for similar results following graftin beads into the recently closed neural tube. Neural tube explantpairs cultured directly with DMSO or Rapamycin scored for reduction of gene expression assessed for Ngn2 (F, F′) n=9/11 explant pairs exhibit clear reduction in Ngn2 in R mycin-treated explant and for Sox2, n=8 explant pairs with nochange in transcript levels (G, G'). Scale bar=100 μm. Gene expression levels assessed by qPCR in 10 pairs of explants. Values for Pax6 and Sox2 transcripts were normalis gainst Histone3 expression (H) or βActin (data not shown) andaveraged. Error bars are SEM and asterisks indicate pb0.01 (paired T-test).

222K.J.Fishw

icket

al./Developm

entalBiology

338(2010)

215–225

ycin(DMg ofapaed a

Fig. 6. TOR Inhibition elicits aberrant onset of neuronal differentiation at the ventricular surface of the neural tube. Neural tubes grafted with a bead soaked in either DMSO (A) orRapamycin (B) at HH10, analysed after 24 h for Tuj-1 expression in whole-mount embryo just caudal to bead and in transverse section for DMSO (C) and Rapamycin (D). Dotted lineindicates midline, red arrowheads in (B) cells ectopically expressing Tuj-1 at the midline; quantification of Tuj-1 cells located at the midline observed per section, (E) (DMSO, 5/662Tuj-1+ cells in 13 embryos at the midline; Rapamycin, 47/765 Tuj1+ cells in 17 embryos at the midline); detection of BrdU incorporation and Tuj-1 expression in bead-graftedembryos after 24 h exposure to DMSO/Rapamycin and BrdU treatment for the final hour (F) (4/9 midline TuJ-1/BrdU double-labelled cells were observed in 9 embryos, none wereobserved in 6 DMSO embryos data not shown); red arrowheads indicate Tuj-1/BrdU double-labelled cells at the ventricular surface in (G) and in transverse section (from differentembryo), (H). Scale bar=50 μm.

223K.J. Fishwick et al. / Developmental Biology 338 (2010) 215–225

TOR activity has been well documented (Hidalgo and Rowinsky,2000); Rapamycin binds intracellularly to members of the immuno-philin family of FK506 binding proteins (FKBPs), biochemical andgenetic approaches have demonstrated that the most important ofthese is FKBP12 and that Rapamycin forms a complex with FKBP12that specifically inhibits mTOR. Furthermore, the lack of impact ofRapamycin on a large panel of other kinases has been established(Bain et al., 2007) and the single target specificity of Rapamycin isfurther born out by its ability to phenocopy the mTOR mutant, Flat-top (Hentges et al., 2001). We show here that Rapamycin treatmentreduces pS6 levels (a readout of the activity of the TOR target S6K1) inchick neural tube explant pairs and acts in a dose-dependent manneras indicated byWestern blot analysis of chick embryo tissue. Althoughformally the effects of Rapamycin might not be 100% TOR specific,these data show that a major consequence of Rapamycin treatment inthe chick embryo is inhibition of TOR activity. Furthermore, we wereable to rescue neuron production in the presence of Rapamycin bymis-expressing an activated form of S6K1 in the neuroepithelium.This indicates that Rapamycin acts via the TOR pathway to repressneuron production in this context.

Expression of Ngn2 promotes neuronal differentiation andconstitutive deletion of this gene delays neurogenesis (Bertrand etal., 2002; Ma et al., 1999). Proneural genes such as Ngns induceexpression of Delta-1 in cells as they become neurons and Ngn2 isthe earliest expressed proneural gene in the vertebrate neural tube.The reduction in neuron number observed in these assays istherefore most likely due to lack of proneural gene expression andso is caused by a defect at the top of the neurogenesis cascade. Ourfinding that markers of the neural progenitor state Sox2 and Sox3 areunaffected by Rapamycin further suggests a specific requirement forthe neuronal differentiation programme. Interestingly, signalling viaPKB has also been shown to enhance the transcriptional activity ofrelated bHLH proneural genes in complex with co-activators p300 orCBP in P19 cells (Vojtek et al., 2003), suggesting multiple inputs toneuronal differentiation at the level of proneural genes downstreamof PI3-kinase signalling.

A link between TOR signalling and neuronal differentiation wasfirst demonstrated by analysis of photoreceptor differentiation inthe fly retina (Bateman and McNeill, 2004). These researchersshowed that the timing of neuronal differentiation was controlledby TOR signalling and suggested that this might reflect TORregulation of translation and protein synthesis via its phosphoryla-tion of S6K, which in turn phosphorylates the ribosomal protein S6and 4E binding protein (see Fig. 2E), an inhibitor of translation

initiation factor 4E (Bateman and McNeill, 2004). One possibility isthat expression of a proneural gene required for this differentiationrelies on a key protein dependent on S6K regulated translation orsynthesis. However, it is also possible that a different mechanismoperates in the vertebrate nervous system; in the fly, manipulationof TOR signalling can delay and accelerate neuronal differentiationsuggesting a simple gene regulatory mechanism, but our dataindicate that stimulating PI3-kinase signalling does not increaseneuron number. This contrasts with a recent study of telencephalicprecursors in vitro, which increase neuronal differentiation inresponse to insulin (Han et al., 2008). However, this was assessedafter 4 days and so may reflect an initial expansion of theprogenitor pool and subsequent differentiation. Our data over ashorter time frame suggests that stimulating PI3K signalling doesnot drive neuron production in the vertebrate CNS but may simplybe permissive for differentiation, perhaps setting a critical level ofcell growth that is required before engagement of the differentia-tion programme. Consistent with this proposal, the retinoidsignalling pathway that directly regulates neuronal differentiationby acting on proneural gene promoters (Ribes et al., 2008) requiresan intact PI3-kinase pathway to induce neuronal differentiation in aneuroblastoma cell line and can also directly stimulate PI3-kinasesignalling (Lopez-Carballo et al., 2002).

The effects of Rapamycin on the cell cycle were consistent with thereduction in neuron numbers; the slight increase in mitotic indexsuggesting that more cells remain within the cell cycle. Recentfindings in insect S2 cells show that Rapamycin can accelerate S-G2-Mtransition, increasing the number of pH3-positive cells over a short(9 h) period (Wu et al., 2007). While this phenomenon couldcontribute to the increased mitotic index observed in our study, thiswould not explain the reduction of neuron numbers, which werescored as a percentage of total cell number. Our findings contrast withmany reports of Rapamycin exposure leading to G1 arrest; anattribute that has lead to use of Rapamycin as an anti-cancer drug(Hidalgo and Rowinsky, 2000). This discrepancy may reflect differ-ences in cell behaviour following exposure to Rapamycin for longerperiods (2–3 days) observed in cell lines maintained in vitro.Although Rapamycin can affect Cyclin D1 stability as quickly as30 min (Hashemolhosseini et al., 1998; Muise-Helmericks et al.,1998), it may be that more rounds of cell division are required beforechanges in cell cycle profile are apparent at a population level. Thismay be particularly so after a short exposure to Rapamycin in vivo,where neuroepithelial cells are heterogeneous with respect to cellcycle phase and cell division mode (Frade, 2002; Wilcock et al., 2007).

Fig. 7. Stimulation of PI3-kinase signalling does not increase neuron production.Western blot for pPKB in HH10 chick embryos cultured with insulin for 2 h (A). HH8chick NT explant pair cultured in control media (B, C) or 100 μg/ml insulin (D, E) for24 h assessed for expression of HuC/D or stained with DAPI. Average percentage ofHuC+ cells in 5 control or insulin-treated (100 μg/ml) NT explant pairs (F). Error barsindicate SD.

224 K.J. Fishwick et al. / Developmental Biology 338 (2010) 215–225

There is also growing evidence that different cell types respond indistinct ways to Rapamycin. This includes inhibition of trophoblastbut not ICM proliferation in wild-typemice (Murakami et al., 2004) aswell as heterogeneity within the central nervous system with respectto neuronal differentiation (Hodge et al., 2005). This indicates thatdifferent cell types have distinct sensitivities to Rapamycin or requireonly low-level TOR signalling for their growth or differentiation. Thisheterogeneity also extends to cell survival signalling, as we observeonly localised cell death in specific subpopulations in mice lackingPDK1/PKB signalling. Overall, this suggests a complex and patternedrequirement for PI3-kinase and TOR signalling in the developingembryo.

It is striking that TOR signalling is also required for myogenesis;the differentiation of myoblasts into myotubes, which similarlyinvolves cell cycle exit (Conejo et al., 2001; Coolican et al., 1997;Erbay and Chen, 2001). Myoblast–myotube differentiation isstrikingly characterised by a tenfold increase in mTOR protein(Erbay and Chen, 2001), which may reflect the requirement fortranslation and synthesis of new proteins that mediate thisfundamental differentiation step. In this sense, during bothneurogenesis and myogenesis, TOR signalling levels may functionas a checkpoint for differentiation progression. Perhaps consistentwith this, we find that Rapamycin does not simply inhibit neuronaldifferentiation but can also result in loss of coordination of cell cycleand differentiation programmes, as we find that Rapamycin doesnot just inhibit neuronal differentiation, but can also disruptcoordination of cell cycle and differentiation programmes; somecells expressing the early neuronal marker Tuj-1 continue aber-rantly within the cell cycle at the ventricular surface. Thisphenotype may be generated depending on cell cycle phase butaffects too few cells to account for the overall reduction in neuronnumbers. At later stages motor neuron progenitors undergoingterminal division do express a neuronal differentiation marker;however, the stages we examined cover a period of interneuronbirth and these cells do not normally express such markers at themidline (Sechrist and Bronner Fraser, 1991). Furthermore, theincorporation of BrdU in Tuj-1 expressing midline cells suggeststhat interkinetic nuclear migration movements are disrupted in thepresence of Rapamycin and that these cells are now cycling on thespot. This defect may relate to neuronal migration defects observedin tuberous sclerosis and future studies will determine whether thisindicates a link between TOR signalling and the mechanics ofnuclear migration movements.

Finally, our data suggest that cells experiencing elevated TORsignalling in tuberous sclerosis neural tumours are poised todifferentiate. Indeed, tuber cells are reported to express severalgenes characteristic of neural progenitors (Yamanouchi et al., 1997).This may account for the contrary phenotype of tuberous sclerosistumours, which, while displaying aberrant cell growth, have atendency to contain differentiated cells.

Acknowledgments

We thank Dario Alessi, Edward McManus and Barry Collins forproviding embryos carrying a loss of function mutation in the PH-domain of PDK1 and for plasmids containing activated or kinase-deadforms of S6K and Kees Weijer for critical reading of the manuscript.KJF and RL studentships were funded by the MRC, and PH and KGSwere funded by an MRC programme grant G0600234.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.ydbio.2009.12.001.

References

Allan, G.J., Zannoni, A., McKinnell, I., Otto, W.R., Holzenberger, M., Flint, D.J., Patel, K.,2003. Major components of the insulin-like growth factor axis are expressed earlyin chicken embryogenesis, with IGF binding protein (IGFBP)-5 expression subjectto regulation by Sonic Hedgehog. Anat. Embryol. (Berl.) 207, 73–84.

Arsenijevic, Y., Weiss, S., 1998. Insulin-like growth factor-I is a differentiation factor forpostmitotic CNS stem cell-derived neuronal precursors: distinct actions from thoseof brain-derived neurotrophic factor. J. Neurosci 18, 2118–2128.

Bain, J., Plater, L., Elliott, M., Shpiro, N., Hastie, J., McLauchlan, H., Klevernic, I., Arthur, S.,Alessi, D., Cohen, P., 2007. The selectivity of protein kinase inhibitors; a furtherupdate. Biochem. J. 408, 97–315.

Balendran, A., Currie, R., Armstrong, C.G., Avruch, J., Alessi, D.R., 1999. Evidence that 3-phosphoinositide-dependent protein kinase-1 mediates phosphorylation of p70 S6kinase in vivo at Thr-412 as well as Thr-252. J. Biol. Chem 274, 37400–37406.

Bateman, J.M., McNeill, H., 2004. Temporal control of differentiation by the insulinreceptor/Tor pathway in Drosophila. Cel 119, 87–96.

225K.J. Fishwick et al. / Developmental Biology 338 (2010) 215–225

Bateman, J.M., McNeill, H., 2006. Insulin/IGF signalling in neurogenesis. Cell. Mol. LifeSci. 63, 1701–1705.

Bel-Vialar, S., Medevielle, F., Pituello, F., 2007. The on/off of Pax6 controls the tempo ofneuronal differentiation in the developing spinal cord. Dev. Biol. 305, 659–673.

Bertrand, N., Castro, D.S., Guillemot, F., 2002. Proneural genes and the specification ofneural cell types. Nat. Rev. Neurosci. 3, 517–530.

Camarero, G., Leon, Y., Gorospe, I., De Pablo, F., Alsina, B., Giraldez, F., Varela-Nieto, I.,2003. Insulin-like growth factor 1 is required for survival of transit-amplifyingneuroblasts and differentiation of otic neurons. Dev. Biol. 262, 242–253.

Chapman, S.C., Collignon, J., Schoenwolf, G.C., Lumsden, A., 2001. Improved method forchick whole-embryo culture using a filter paper carrier. Dev. Dyn. 220, 284–289.

Conejo, R., Valverde, A.M., Benito, M., Lorenzo, M., 2001. Insulin produces myogenesis inC2C12myoblasts by induction of NF-kappaB and downregulation of AP-1 activities.J. Cell. Physiol. 186, 82–94.

Coolican, S.A., Samuel, D.S., Ewton, D.Z., McWade, F.J., Florini, J.R., 1997. The mitogenicand myogenic actions of insulin-like growth factors utilize distinct signalingpathways. J. Biol. Chem. 272, 6653–6662.

Crino, P.B., Nathanson, K.L., Henske, E.P., 2006. The tuberous sclerosis complex. N. Engl.J. Med. 355, 1345–1356.

D'Ercole, A.J., Ye, P., O'Kusky, J.R., 2002. Mutant mouse models of insulin-like growthfactor actions in the central nervous system. Neuropeptides 36, 209–220.

Diez del Corral, R., Breitkreuz, D.N., Storey, K.G., 2002. Onset of neuronal differentiationis regulated by paraxial mesoderm and requires attenuation of FGF signalling.Development 129, 1681–1691.

Easton, R.M., Cho, H., Roovers, K., Shineman, D.W., Mizrahi, M., Forman, M.S., Lee, V.M.,Szabolcs, M., de Jong, R., Oltersdorf, T., Ludwig, T., Efstratiadis, A., Birnbaum, M.J.,2005. Role for Akt3/protein kinase Bgamma in attainment of normal brain size.Mol. Cell. Biol. 25, 1869–1878.

Erbay, E., Chen, J., 2001. The mammalian target of rapamycin regulates C2C12myogenesis via a kinase-independent mechanism. J. Biol. Chem. 276, 36079–36082.

Fingar, D.C., Blenis, J., 2004. Target of rapamycin (TOR): an integrator of nutrient andgrowth factor signals and coordinator of cell growth and cell cycle progression.Oncogene 23, 3151–3171.

Frade, J.M., 2002. Interkinetic nuclear movement in the vertebrate neuroepithelium:encounters with an old acquaintance. Prog. Brain Res. 136, 67–71.

Gingras, A., Raught, B., Sonenberg, N., 2007. Regulation of translation initiation byFRAP/mTOR. Genes Dev. 15, 807–826.

Hamburger, H., Hamilton, H.L., 1951. A series of normal stages in the development ofthe chick embryo. J. Exp. Morphol. 88, 49–92.

Hammerle, B., Vera-Samper, E., Speicher, S., Arencibia, R., Martinez, S., Tejedor, F.J.,2002. Mnb/Dyrk1A is transiently expressed and asymmetrically segregated inneural progenitor cells at the transition to neurogenic divisions. Dev. Biol. 246,259–273.

Han, J., Wang, B., Xiao, Z., Gao, Y., Zhao, Y., Zhang, J., Chen, B., Wang, X., Dai, J., 2008.Mammalian target of rapamycin (mTOR) is involved in the neuronal differentiationof neural progenitors induced by insulin. Mol. Cell. Neurosci. 39, 118–124.

Hashemolhosseini, S., Nagamine, Y., Morley, S.J., Desrivieres, S., Mercep, L., Ferrari, S.,1998. Rapamycin inhibition of the G1 to S transition is mediated by effects on cyclinD1 mRNA and protein stability. J. Biol. Chem. 273, 14424–14429.

Hentges, K., Sirry, B., Gingeras, A.C., Sarbassov, D., Sonenberg, N., Sabatini, D.M., Peterson,A., 2001. FRAP/mTOR is required for proliferation and patterning during embryonicdevelopment in the mouse. Proc. Natl. Acad. Sci. U. S. A. 98, 13796–13801.

Hentges, K., Thompson, K., Peterson, A., 1999. The flat-top gene is required for theexpansion and regionalization of the telencephalic primordium. Development 126,1601–1609.

Hidalgo, M., Rowinsky, E.K., 2000. The rapamycin-sensitive signal transductionpathway as a target for cancer therapy. Oncogene 19, 6680–6686.

Hodge, R.D., D'Ercole, A.J., O'Kusky, J.R., 2004. Insulin-like growth factor-I acceleratesthe cell cycle by decreasing G1 phase length and increases cell cycle reentry in theembryonic cerebral cortex. J. Neurosci. 24, 10201–10210.

Hodge, R.D., D'Ercole, A.J., O'Kusky, J.R., 2005. Increased expression of insulin-likegrowth factor-I (IGF-I) during embryonic development produces neocorticalovergrowth with differentially greater effects on specific cytoarchitectonic areasand cortical layers. Brain Res. Dev. Brain Res. 154, 227–237.

Itasaki, N., Bel-Vialar, S., Krumlauf, R., 1999. ‘Shocking’ developments in chickembryology: electroporation and in ovo gene expression. Nat. Cell Biol. 1, E203–207.

Jessell, T.M., 2000. Neuronal specification in the spinal cord: inductive signals andtranscriptional codes. Nat. Rev. Genet. 1, 20–29.

Lawlor, M.A., Mora, A., Ashby, P.R., Williams, M.R., Murray-Tait, V., Malone, L., Prescott,A.R., Lucocq, J.M., Alessi, D.R., 2002. Essential role of PDK1 in regulating cell size anddevelopment in mice. EMBO J. 21, 3728–3738.

Lopez-Carballo, G., Moreno, L., Masia, S., Perez, P., Barettino, D., 2002. Activation of thephosphatidylinositol 3-kinase/Akt signaling pathway by retinoic acid is requiredfor neural differentiation of SH-SY5Y human neuroblastoma cells. J. Biol. Chem.277, 25297–25304.

Lunn, J.S., Fishwick, K.J., Halley, P.A., Storey, K.G., 2007. A spatial and temporal map ofFGF/Erk1/2 activity and response repertoires in the early chick embryo. Dev. Biol.302, 536–552.

Ma, Q., Fode, C., Guillemot, F., Anderson, D.J., 1999. Neurogenin1 and neurogenin2control two distinct waves of neurogenesis in developing dorsal root ganglia. GenesDev. 13, 1717–1728.

McManus, E.J., Collins, B.J., Ashby, P.R., Prescott, A.R., Murray-Tait, V., Armit, L.J., Arthur,J.S., Alessi, D.R., 2004. The in vivo role of PtdIns(3,4,5)P(3) binding to PDK1 PHdomain defined by knockin mutation. Embo J. 23, 2071–2082.

McNeill, H., Craig, G.M., Bateman, J.M., 2008. Regulation of neurogenesis and epidermalgrowth factor receptor signaling by the insulin receptor/target of rapamycinpathway in Drosophila. Genetics 179, 843–853.

Mizuguchi, M., Takashima, S., 2001. Neuropathology of tuberous sclerosis. Brain Dev.23, 508–515.

Muise-Helmericks, R.C., Grimes, H.L., Bellacosa, A., Malstrom, S.E., Tsichlis, P.N., Rosen,N., 1998. Cyclin D expression is controlled post-transcriptionally via a phospha-tidylinositol 3-kinase/Akt-dependent pathway. J. Biol. Chem. 273, 29864–29872.

Murakami, M., Ichisaka, T., Maeda, M., Oshiro, N., Hara, K., Edenhofer, F., Kiyama, H.,Yonezawa, K., Yamanaka, S., 2004. mTOR is essential for growth and proliferation inearly mouse embryos and embryonic stem cells. Mol. Cell. Biol. 24, 6710–6718.

Otaegi, G., Yusta-Boyo, M.J., Vergano-Vera, E., Mendez-Gomez, H.R., Carrera, A.C., Abad,J.L., Gonzalez, M., de la Rosa, E.J., Vicario-Abejon, C., de Pablo, F., 2006. Modulationof the PI 3-kinase-Akt signalling pathway by IGF-I and PTEN regulates thedifferentiation of neural stem/precursor cells. J. Cell. Sci. 119, 2739–2748.

Price, S.R., Briscoe, J., 2004. The generation and diversification of spinal motor neurons:signals and responses. Mech. Dev. 121, 1103–1115.

Rennebeck, G., Kleymenova, E.V., Anderson, R., Yeung, R.S., Artzt, K., Walker, C.L., 1998.Loss of function of the tuberous sclerosis 2 tumor suppressor gene results inembryonic lethality characterized by disrupted neuroepithelial growth anddevelopment. Proc. Natl. Acad. Sci. U. S. A. 95, 15629–15634.

Ribes, V., Stutzmann, F., Bianchetti, L., Guillemot, F., Dolle, P., Le Roux, I., 2008.Combinatorial signalling controls Neurogenin2 expression at the onset of spinalneurogenesis. Dev. Biol. 321, 470–481.

Scardigli, R., Schuurmans, C., Gradwohl, G., Guillemot, F., 2001. Crossregulationbetween Neurogenin2 and pathways specifying neuronal identity in the spinalcord. Neuron 31, 203–217.

Scardigli, R., Baumer, N., Gruss, P., Guillemot, F., Le Roux, I., 2003. Direct andconcentration-dependent regulation of the proneural gene Neurogenin2 by Pax6.Development 130, 3269–3281.

Sechrist, J., Bronner Fraser, M., 1991. Birth and differentiation of reticular neurons in thechick hindbrain: ontogeny of the first neuronal population. Neuron 7, 947–963.

Tschopp, O., Yang, Z.Z., Brodbeck, D., Dummler, B.A., Hemmings-Mieszczak, M.,Watanabe, T., Michaelis, T., Frahm, J., Hemmings, B.A., 2005. Essential role ofprotein kinase B gamma (PKB gamma/Akt3) in postnatal brain development butnot in glucose homeostasis. Development 132, 2943–2954.

Vlahos, C.J., Matter, W.F., Hui, K.Y., Brown, R.F., 1994. A specific inhibitor ofphosphatidylinositol 3-kinase, 2-(4-morpholinyl)-8-phenyl-4H-1-benzopyran-4-one (LY294002). J. Biol. Chem. 269, 5241–5248.

Vojtek, A.B., Taylor, J., DeRuiter, S.L., Yu, J.Y., Figueroa, C., Kwok, R.P., Turner, D.L., 2003. Aktregulates basic helix-loop-helix transcription factor-coactivator complex formationand activity during neuronal differentiation. Mol. Cell. Biol. 23, 4417–4427.

Wakamatsu, Y., Weston, J.A., 1997. Sequential expression and role of Hu RNA-bindingproteins during neurogenesis. Development 124, 3449–3460.

Wilcock, A.C., Swedlow, J.R., Storey, K.G., 2007. Mitotic spindle orientation distinguishesstem cell and terminal modes of neuron production in the early spinal cord.Development 134, 1943–1954.

Wilkinson, D.G., Nieto, M.A., 1993. Detection of messenger RNA by in situ hybridizationto tissue sections and whole mounts. Methods Enzymol. 225, 361–373.

Wu, M.Y., Cully, M., Andersen, D., Leevers, S.J., 2007. Insulin delays the progression ofDrosophila cells through G2/M by activating the dTOR/dRaptor complex. EMBO J.26, 371–379.

Yamanouchi, H., Jay, V., Rutka, J.T., Takashima, S., Becker, L.E., 1997. Evidence ofabnormal differentiation in giant cells of tuberous sclerosis. Pediatr. Neurol. 17,49–53.

Ye, P., D'Ercole, A.J., 2006. Insulin-like growth factor actions during development ofneural stem cells and progenitors in the central nervous system. J. Neurosci. Res. 83,1–6.