AKT Signaling Mediates IGF-I Survival Actions on Otic Neural Progenitors

Background Otic neurons and sensory cells derive from common progenitors whose transition into mature cells requires the coordination of cell survival, proliferation and differentiation programmes. Neurotrophic support and survival of post-mitotic otic neurons have been intensively studied, but the bases underlying the regulation of programmed cell death in immature proliferative otic neuroblasts remains poorly understood. The protein kinase AKT acts as a node, playing a critical role in controlling cell survival and cell cycle progression. AKT is activated by trophic factors, including insulin-like growth factor I (IGF-I), through the generation of the lipidic second messenger phosphatidylinositol 3-phosphate by phosphatidylinositol 3-kinase (PI3K). Here we have investigated the role of IGF-dependent activation of the PI3K-AKT pathway in maintenance of otic neuroblasts. Methodology/Principal Findings By using a combination of organotypic cultures of chicken (Gallus gallus) otic vesicles and acoustic-vestibular ganglia, Western blotting, immunohistochemistry and in situ hybridization, we show that IGF-I-activation of AKT protects neural progenitors from programmed cell death. IGF-I maintains otic neuroblasts in an undifferentiated and proliferative state, which is characterised by the upregulation of the forkhead box M1 (FoxM1) transcription factor. By contrast, our results indicate that post-mitotic p27Kip-positive neurons become IGF-I independent as they extend their neuronal processes. Neurons gradually reduce their expression of the Igf1r, while they increase that of the neurotrophin receptor, TrkC. Conclusions/Significance Proliferative otic neuroblasts are dependent on the activation of the PI3K-AKT pathway by IGF-I for survival during the otic neuronal progenitor phase of early inner ear development.


Introduction
The vertebrate inner ear is a complex sensory organ responsible for the senses of sound and balance. The inner ear derives from an ectodermal placode that invaginates to form the otic vesicle, or otocyst. This structure contains the information required to carry out an autonomous program of development resulting in generation of most of the cells that form the adult inner ear. Inner ear development requires the transition of otic progenitors through states of active cell proliferation, cell fate specification, synchronized cell cycle exit and differentiation to generate the distinctly fated cell populations within the highly ordered mosaic of the organ of Corti in mammals and the basilar papilla in birds [1], [2], [3]. Neurons of the acoustic and vestibular ganglia have a common origin in the otic placode. Neuronal progenitors are specified in the neurogenic otic epithelia, from where they migrate to form the acoustic-vestibular ganglion (AVG) [2]. The AVG later develops into the acoustic and vestibular ganglia that connect the sensory epithelia to the brain through the VIII cranial nerve [4]. Programmed cell death contributes to otic development and neurogenesis by controlling neuroblast cell number [5], [6], [7]. Neurodevelopmental programmed cell death occurs either in proliferating progenitors or upon loss of neurotrophins [8]. The postmitotic dependence of otic neurons on neurotrophins has been the subject of intense study [9], [10], [11], but the regulation of cell death during the expansion period of proliferative otic neuroblasts remains poorly understood.
IGF-I deficiency is associated with severe cochlear defects and human sensorineural deafness (http://www.ncbi.nlm.nih.gov/ omim/608747), [12], [13], [14], [15]. In the chicken embryo, IGF-I is required for the early morphogenesis and survival of otic vesicle epithelial cells [16], [17], as well as for AVG neurogenesis [18]. Binding of IGF-I to high affinity receptors, IGF1R, activates two main downstream signalling pathways, namely the RAF-MEK-ERK phosphorylation cascade and the phosphatidylinositol 3 kinase-AKT (PI3K-AKT) pathway [7], [19], [20]. Igf1 2/2 null mouse cochleae showed decreased activation of the PI3K-AKT pathway and an altered FoxM1/p27 kip ratio when compared to wild type mice [21]. In addition, the genomic analysis of the null mutant mouse for the zinc finger transcription factor Gata3, a fundamental player in auditory system development [22], has shown alterations in the expression of genes linked to IGFsignalling, including AKT and p27 Kip [23]. Still, a detailed study of the role of AKT during inner ear development has not yet been performed.
The serine-threonine protein kinase AKT (http://www.ncbi. nlm.nih.gov/omim/164730) is activated by binding of phosphatidylinositol 3-phosphate, which promotes its recruitment to the plasma membrane and sequential phosphorylation at threonine (Thr308) and serine (Ser473) residues by the phosphoinositide dependent kinases [24]. AKT can then phosphorylate and inactivate downstream proteins implicated in programmed cell death [24], [25] and cell cycle regulation, such as the cyclindependent kinase inhibitor p27 Kip [26], [27], [28]. The study of knockout mice for the Akt1 and Akt2 genes has shown that they play a fundamental role in cell growth and survival [29], [30], [31]. Furthermore, Akt1 levels regulate survival, proliferation and self-renewal of neural progenitors in response to extrinsic signals [31].
Here we show that AKT activation is fundamental for the IGF-I-dependent survival of the expanding population of otic neuroblasts.

Chicken embryos
Chicken embryos were obtained from fertilized eggs purchased from a local farm (Granja Santa Isabel, Córdoba, Spain), they were incubated in a humidified atmosphere at 37.8uC and the embryos were staged according to Hamburger and Hamilton criteria (HH; [32]). All experiments were performed following the recommendations of the European Communities Council Directive (86/609/EEC) and they were approved by the Ethics Committee of the CSIC.

Embryo and tissue preparation for in situ hybridization and immunofluorescence
Whole embryos or tissues were dissected in phosphate-buffered saline (PBS) and fixed overnight in 4% paraformaldehyde in PBS at 4uC. Subsequently, embryos were cryoprotected overnight in 15% sucrose/PBS at 4uC and then embedded at 37uC in 15% sucrose/10% gelatine in PBS. Gelatine-embedded tissues were frozen in isopentane at 280uC and then sectioned (20 mm) at 225uC in a cryostat (Cryocut 1900; Leica Microsystems, Deerfield, IL). The sections obtained were used for in situ hybridization or immunofluorescent staining.

Igfr1 cloning and in situ hybridization
In situ hybridization with digoxigenin-labelled antisense RNA probes (1 mg/ml) was performed essentially as described previously with some minor modifications [33]. Three HH19 embryos were tested in parallel in at least two independent experiments and no specific signal was obtained with the control sense probes (data not shown). The ckIgf1r gene was cloned by PCR (Igf1r forward 59-CAGTTCTCTCCTTGCCATCC-39; Igf1r reverse 59-CAGCAT-CCAACTCCTCTTCC-39) and ligated into a pGEM-T plasmid (Promega, Madison, WI). The probes for ckTrkC were kindly provided by Dr. Martin-Zanca (CSIC-USAL, Salamanca, Spain, [34]). Antisense single-stranded RNA probes for ckTrkC (BSU361/ T7) and ckIgf1r (ApaI/Sp6) were prepared by in vitro transcription.

Immunofluorescence
The sources, dilution, and cell specificities of the antibodies used for immunofluorescent staining are shown in Table 1. Samples were washed and permeabilized in 1% or 0.05% PBS/Triton-X-100 (PBS-T) (whole-mount otic vesicles or AVG and frozen sections, respectively). An additional permeabilization step of 30 min in PBS-T at 37uC, was carried out in the explant wholemount immunostaining. Non-specific binding sites were blocked for 1 h in PBS-T with 3% (wt/vol) BSA (Sigma-Aldrich) and 5% (vol/vol) normal goat or donkey serum. Samples were exposed to the primary antibodies overnight at 4uC, diluted in PBS-Tween20 (0.05%). The fluorescence conjugated secondary antibodies (antimouse Alexa488, anti-rabbit Alexa488, anti-mouse Alexa546, anti-rabbit Alexa546, anti-rabbit Alexa647 or anti-goat Alexa660; Molecular Probes, Eugene, OR) were incubated with the samples for 3 h at room temperature (RT) at a dilution of 1:400 in PBS Tween 20. For dual-fluorescence immunolabelling, samples were incubated with a mixture of fluorescent-conjugated secondary antibodies. The otic vesicles were mounted in Prolong Gold with DAPI (Invitrogen, Carlsbad, CA) and fluorescence staining was visualized by confocal microscopy (Leica TCS SP2, Wetzlar, Germany). At least six otic vesicles per condition were assayed from three independent experiments. For immunohistofluorescent staining of frozen sections, the sections were thawed and the protocol indicated above was followed. The sections were then mounted in Prolong Gold with DAPI (Invitrogen) and they were visualized under a fluorescence microscope (Nikon 90i, Tokyo, Japan).
Levels of SOX2, G4 and p27 Kip immunostaining were quantified using Adobe Photoshop CS4 software (Adobe Systems Inc., CA, USA) in compiled confocal microscopy projections. The colour channels of the signals of interest were converted into grey scale images. Then both the area and the intensity of the signal were measured, and normalised to the 0S condition, which was given an arbitrary value of 100. The data are presented as the mean 6 SEM and the statistical significance was estimated with the Student's t-test.
AVG areas (neural processes and neural somas) were measured using the ImageJ software (Wayne Rasband, National Institutes of Health, Bethesda, MD) and the results are presented as the mean 6 SEM of the areas. Statistical significance was estimated by Student's t-test.

Programmed cell death, BrdU incorporation and immunodetection
Cell death was studied by Tdt-mediated dUTP nick-end labelling (TUNEL) staining of fragmented DNA using the kit Dead-End TM Fluorometric TUNEL System (Invitrogen) adapted to whole organ labelling [38]. The otic vesicles were mounted in Prolong Gold/DAPI and visualized by confocal microscopy. To study cell proliferation, otic vesicles were incubated with 5-Bromo-29-deoxyuridine (BrdU, 10 mg/ml, Sigma-Aldrich) during the last hour of culture. BrdU incorporation was immunodetected as described with additional steps to denature the DNA. Samples were incubated in 50% (vol/vol) formamide-SSC for 40 min at 65uC and in 2N HCl for 30 min at 37uC, followed by a 10 min wash in Tris 0.1 M (pH 8). At least six otic vesicles per condition were assayed from three independent experiments.
TUNEL-positive nuclei were quantified in compiled confocal microscopy projections with Image Analysis Software (Olympus) and the results are presented as the mean 6 SEM of the positive cells per total area. Values were normalised to the 0 h time point (otic vesicles fixed immediately after removal from the embryo). Areas from medial sections of the otic vesicle and the AVG were measured using the ImageJ software (Wayne Rasband, National Institutes of Health, Bethesda, MD) and the results are presented as the mean 6 SEM of the areas (mm 2 ). Statistical significance was estimated by Student's t-test or by ANOVA with Bonferroni and Tukey's multiple comparisons post-hoc tests, using SPSS for Windows 15.0 software (SPSS Inc., Chicago, IL).

Quantitative RT-PCR
Chicken embryos were dissected out and pooled to obtain RNA at different stages: HH17 (n = 45), HH18 (n = 30), HH19 (n = 25) and HH23 (n = 5). Four independent pools of RNA from each stage were isolated with TRIZOL (Invitrogen) following the manufacturer's instructions. The integrity and concentration of the RNA was assessed with an Agilent Bioanalyzer 2100 (Agilent Technologies) and cDNA was generated by reverse transcription (High Capacity cDNA Reverse Transcription Kit; Applied Biosystems). Real-Time PCR of each pool was performed in triplicate using specific oligonucletides for the available chicken AKT genes (Akt1 and Akt3; Quantitect Primer Assays, Geneglobe, Qiagen). Primers used were: Akt1 and Akt3 (QT00593411 and QT01484105, respectively; Geneglobe, Qiagen), using SYBR Green as the detection system. Eukaryotic 18S rRNA was used as the endogenous housekeeping gene (Hs99999901_s1, TaqMan, Applied Biosystems). PCR was performed on a 7900HT Real-Time PCR System and gene expression was estimated as 2 2DDCt .

Western blotting
Otic vesicles were made quiescent by incubating overnight in serum-free M199 culture medium; they were pre-treated with the inhibitors (LY and AKTi) for 1 h and then stimulated with IGF-I for 30 min. At the end of this period, the otic vesicles were homogenized in ice cold Laemmli Buffer containing 50 mM dithiotreitol, phosphatase inhibitor cocktail 2 and protease inhibitor cocktail (both 1:100, from Sigma-Aldrich) and stored immediately at 220uC. Before loading the samples, they were heated at 95uC for 5 min. Electrophoresis, protein transfer and immunodetection were carried out as described [20] using the antibodies shown in Table 1. Four independent experiments were performed using 15 HH18 otic vesicles per condition. Western blot films were scanned and the bands were quantified by densitometry with Adobe Photoshop CS4 software (Adobe Systems Inc.). The results are presented as the mean 6 SEM of the phosphorylated AKT (on Ser 473)/AKT ratio. The IGF-I condition was considered as the reference and given a value of 100. The statistical significance was estimated with the Student's t-test.

Results
In vivo expression of Igf1r in the developing AVG During otic neurogenesis there is a sequence of otic neural maturation, which is reflected by the expression of molecular markers characteristic of the transition from epithelial to ganglionic neuroblasts, then to immature and, finally, to mature neurons ( Figure 1A). Islet-1 is a LIM/homeodomain transcriptional regulator that is expressed by early otic neuroblasts [39], [40], and TuJ-1 is a b-III-tubulin that labels fibers in immature ganglionic neuroblasts. The axonal glycoprotein G4 [18], [20] [41] and the TrkC neurotrophin receptor [42] are characteristic of mature otic neurons Neuroblasts are characterized by their capacity to proliferate, whilst neurons are post-mitotic and express cell cycle inhibitors such as p27 Kip [43].
To identify the neural otic populations modulated by IGF-I, we studied Igf1r expression in the developing HH19 otic vesicle, and its cellular co-localization with neural differentiation markers (Figure 1 B). Igf1r was ubiquitously expressed in the developing inner ear at stage HH19, although its expression was notably stronger in the otic epithelium adjacent to the AVG and in the ganglion neuroblasts (Figure 1 B, a and e, asterisks and arrows, respectively). Within the AVG, Igf1r expression was intense in neuroblasts strongly labeled for Islet-1 (Figure 1 B, b and c, arrowheads). In the otic epithelium, Igf1r and Islet-1 were expressed strongly in the neurogenic zone (Figure 1 B, c9, arrowheads). In striking contrast, the ganglion population most distal to the neurogenic epithelium exhibited a weak Igf1r expression, concomitantly with an intense p27 Kip label, G4-positive neuronal processes, TuJ-1 and TrkC expression (Figure 1 B, d, f, g and h, arrows). These data suggest that undifferentiated otic neuroblasts show high levels of IGF1R, but, as development proceeds, otic neurons down-regulate IGF1R expression. Spatial distribution of neural markers in cultured otic vesicle explants Otic vesicles can be explanted from the embryo and their ex vivo development can be followed in a defined culture medium to study the molecular cues that instruct the cellular diversity found in vivo [7]. The AVG also develops ex vivo, and thus this constitutes an excellent model to study otic neurogenesis [38], [37] (Figure 2 A). To further explore the IGF-I actions in the different neural differentiation stages, we used the ex vivo model to compare the neural markers G4, Islet-1, TuJ-1 and the post-mitotic marker p27 Kip in the untreated condition (cultured in the absence of any additives; 0S) and in IGF-I-treated otic vesicles (Figure 2 B). In both conditions, the AVG showed areas populated by cells at distinct neural maturation stages. Thus, the most distal and ventral aspect of the AVG showed weaker Islet-1 and stronger TuJ-1/G4 and p27 Kip staining than the proximal and dorsal aspects (Figure 2 B, a-d, arrows; c, double-headed arrow). Accordingly, immature G4 and p27 Kip -negative neuroblasts were confined to the more proximal and dorsal part of the AVG, adjacent to the otic epithelium (Figure 2 B, a-d, arrowheads). Treatment with IGF-I for 20 h resulted in an increased size of the AVG (Figure 2 B, compare a, c with b, d) and the G4-negative cell population was confined to the proximal-dorsal part of the AVG (Figure 2 B  Otic vesicles were isolated from HH18 chicken embryos and cultured in serum-free medium without additives (0S; a-e and a9-e9) or supplemented with IGF-I (10 nM; g-k and g9-k9). Immunostaining for incorporated BrdU was performed (green), and light microscopy images are shown for comparison (a-k). The arrowhead points to the accumulation of BrdU-positive cells in the neurogenic region in the otic epithelium (c9), whilst the arrows indicate the areas of low BrdU incorporation in the AVG (c9, j9, k9). Panels j0 and k0 correspond to the boxed areas of the ventral part of the AVG in j9 and k9, respectively, where BrdU incorporation is very low. Representative images of six otic vesicles per condition and from at least three independent experiments are shown, and they were obtained from compiled confocal microscopy projections of otic vesicles. Orientation, A, anterior; D, dorsal. Scale bars, 150 mm.  arrowhead). Indeed, in the AVG cultured with IGF-I there was a significant reduction in the proportion of G4 positive labeling with respect to the total area when compared with the 0S condition (Figure 2 B, quantification in e). Accordingly, the proportion of postmitotic p27 Kip -cells was also significantly higher in the IGF-Itreated AVG than in control cultures (Figure 2 B, compare c with d, arrowheads; quantification in f). Thus, IGF-I-treated cultures show a wider p27 Kip /G4-negative immature neuroblast population than control cultures.
Glial cells for the AVG are derived from the neural crest and can be detected by immunolocalization of the transcription factor SOX10 [37], [44]. In HH19 embryos, the AVG shows SOX10positive cells associated to Islet-1 positive neuroblasts.In contrast, cultures of otic vesicles develop a ganglion that shows no detectable contribution of glial cells in any of the culture conditions tested, 0S, IGF-I treatment or incubation with LY ( Figure S1).

Spatiotemporal patterns of cell proliferation in cultured otic vesicles
To study the state of proliferation in the IGF-I-cultures, otic vesicles were cultured for different periods of time (1, 4, 8, 20 or 30 h) in the presence of BrdU, in both 0S and IGF-I conditions (Figure 3 A). During the first 4 h, BrdU uptake by otic vesicles was similar in the IGF-I and 0S conditions (Figure 3 A, compare a9 with g9, and b9with h9). After 8 h in culture, BrdU incorporation in the 0S condition was observed in the AVG and in the ventromedial aspect of the otic vesicle (Figure 3 A, c9, arrow and arrowhead, respectively; Figure S2 A). BrdU incorporation was negligible after 20 and 30 h (Figure 3 A, d9 and e9). Interestingly, after 20 h in culture in presence of IGF-I, BrdU positive cells were abundant in the AVG, except in its most ventral aspect (Figure 3 A, j9, arrow; detail in j0). This trait was more evident after 30 h in culture, where negligible IGF-I-induced BrdU incorporation was observed in the ventral aspect of the AVG (Figure 3 A, k9, arrow; detail in k0). The areas of both the otic vesicle and the AVG were measured in both 0S and IGF-I conditions for the different time-points (Figure 3 B). There were no significant differences in the size of the otic vesicles maintained in the presence or absence of IGF-I up to 4 h in culture. By contrast, from 8 h onwards, the otic vesicles cultured in the presence of IGF-I were significantly larger than those of the 0S condition. The AVG followed a similar trend showing a small but significant increase in size after 8 h in the presence of IGF-I (Figure 3 B). These data indicate that IGF-I promotes a widespread and sustained entry in to S-phase of the cell cycle in cultured otic vesicles.

Spatiotemporal pattern of cell death in otic vesicle explant cultures
To study the survival of otic neural progenitors, we performed TUNEL labelling in otic vesicles, cultured for different periods of time (1,4,8,20 or 30 h), in both 0S and IGF-I conditions (Figure 4 A). TUNEL-positive cells were counted at the different time-points (Figure 4 B). Otic vesicles cultured in control 0S conditions showed apoptosis that could be rescued in the presence of either IGF-I or FBS to the culture medium (Figure S2 B). In the 0S otic vesicles, TUNEL staining increased rapidly during the first 8 h in culture, and afterwards it remained elevated and widespread (Figure 4 A, a9-e9). Although in the presence of IGF-I there was also a timedependent increase in TUNEL staining, it was significantly reduced compared with the 0S condition (Figure 4 A, g9-k9). In addition, the distribution of TUNEL-positive cells was different. In the IGF-I-treated explants, up to 20 h of culture, TUNEL-positive cells were restricted to the ventromedial neurogenic zone of the otic vesicle (Figure 4 A, j9 and k9, arrows) and to the ventral part of the AVG (Figure 4 A, arrowheads in j9 and k9; details in j0 and k0). Accordingly, this area showed negligible BrdU incorporation even in the presence of IGF-I (see Figure 3 A, j9, k9, j0 and k0). These data indicate that IGF-I protects otic progenitors in culture from apoptosis and that as development proceeds the most mature neural cells present high TUNEL-staining, even in the presence of IGF-I.

IGF-I protects otic progenitors from caspase-3dependent programmed cell death
To determine whether IGF-I-abolished programmed cell death was caspase-dependent, we studied active caspase-3 expression levels ( Figure 5). Explants in the 0S condition, showed areas of intense apoptotic cell death where TUNEL-labelled apoptotic nuclei were surrounded by cytoplasm containing active caspase-3 ( Figure 5, a9; detail in d). The addition of exogenous IGF-I caused a striking reduction in both TUNEL and active caspase-3 labelling, although some stained regions could still be seen, in accord with previously shown data (Figure 5, b9; see Figure 4). Figure 5. IGF-I protects otic progenitors from caspase-3dependent apoptosis. Otic vesicles were isolated from HH18 chicken embryos and cultured for 20 h in serum-free culture medium without additives (0S; a and a9), in the presence of IGF-I (10 nM; b and b9) or the pan-caspase inhibitor BOC (50 mM; c and c9). Cell death was visualized by TUNEL staining (green), and immunostaining for activated-caspase-3 (red) was performed. Light microscopy images are also shown for comparison (a-c). BOC treatment caused a reduction in the AVG size (c, arrowhead) and the otic epithelium appeared thickened (c, arrow). The boxed area in a9 is shown at a higher magnification in d, which shows the TUNEL-positive nuclei surrounded by activated caspase-3. Representative images of six otic vesicles per condition and from at least three independent experiments are shown, and they were obtained from compiled confocal microscopy projections of otic vesicles. A, anterior; D, dorsal. Scale bar, 150 mm. doi:10.1371/journal.pone.0030790.g005  (Figure 5, c9), increased otic vesicle size, and interestingly, caused an abnormal thickening in the otic epithelium that showed a rounded morphology, and also caused a reduction in the AVG size ( Figure 5, compare a with c, arrow and arrowheads, respectively). These data suggest that apoptosis is required for morphogenesis and neurogenesis.

Activation of the PI3K-AKT pathway is required for IGF-Imediated survival of otic neuroblasts
Akt gene expression was studied in the inner ear at selected stages of otic development by quantitative RT-PCR. Transcripts encoding for Akt1 and Akt3 were expressed in comparable levels from HH17 to HH19 (Figure 6 A). AKT activation by phosphorylation (pAKT), was analysed in chicken embryos at the stages HH17 to HH19 (Figure 6 B). At HH17, pAKT was detected strongly at the closing edges of the otic pore and in mitotic cells of the otic epithelium (Figure 6 B, a, arrows and arrowheads, respectively). At stage HH18, pAKT was detected in the epithelial and the migrating neuroblasts (Figure 6 B, b arrow and arrowhead respectively). At HH19, pAKT was shown in the otic epithelium, and it was abundant in the neurogenic region and in the AVG (Figure 6 B, c and c9, arrow and arrowhead, respectively). The effects of IGF-I on survival and apoptosis involve signalling through the PI3K-AKT pathway in most cell contexts. Specific drugs such as the PI3K inhibitor LY294002 (LY) and the AKT inhibitor AKTi VIII (AKTi) can be used to study the consequences of PI3K-AKT pathway impairment (Figure 6 C). Upon the addition of IGF-I to otic vesicles in culture, the levels of pAKT increased, and this effect was abolished by pre-treatment with both inhibitors LY and AKTi (Figure 6 D). Neither inhibitor caused significant changes in the pERK/ERK ratio determined in parallel experiments (data not shown). Interestingly, basal phosphorylation of AKT was detected at threonine 308 (Thr 308) but not at the serine 473 (Ser 473) residue, and while AKTi had no effect on the former, LY reduced both phosphorylations (Figure 6  D). Overnight AKT inhibition also triggered a decrease in FoxM1 levels, a transcriptional activator of the G2-M-specific gene cluster that promotes p27 Kip degradation [45] and is modulated by IGF-I in the mouse cochlea [7] (Figure 6 D).
To study the effects of AKT inhibition in cultured otic vesicles, explants were cultured for 20 h in the presence or absence of IGF-I, LY or both. In accord with previously shown data, IGF-I treatment caused an increase in the Islet-1-positive and TuJ-1-negative neuroblast population, when compared to the 0S condition (Figure 6 E, compare a and b dashed areas). LY treatment caused a decrease in the Islet-1-positive neuroblasts, whereas mature TuJ-1-positive neuroblasts survived, although they did not show neurite extension (Figure 6 E, c, arrow). Exposure to IGF-I partially recovered the neuroblast population after LY treatment (Figure 6 E,  d, dashed area). IGF-I-treated otic vesicles showed a wide expression of pAKT that was abundant in ganglionic neuroblasts. Exposure to LY dramatically decreased the pAKT levels in both the otic epithelium and the AVG (data not shown).

Epithelial and ganglionic neuroblast populations are reduced upon AKT-inhibition
In order to further study the consequences of PI3K-AKT inhibition in otic neuroblast populations, we explored SOX2 expression (Figure 7 A). The transcription factor SOX2 is essential for the sustained self-renewal of undifferentiated embryonic cells and it is expressed in the otic proneural domain, which contains the epithelial neuroblasts [46], [47]. Whilst the SOX2-positive population was significantly increased by IGF-I treatment (Figure 7 A, compare a with b, arrowheads; quantification in e), there was a dramatic reduction in this population in the presence of either inhibitor (Figure 7 A, compare a with c and d). TUNEL staining and immunodetection of phospho-histone-3 (PH3) were performed to study cell death and cell proliferation respectively upon PI3K-AKT-inhibition (Figure 7 B, a-d). IGF-I increased the number of PH3-positive cells mitotic cells, whereas treatment with LY or AKTi did not significantly alter it (data not shown). However, PI3K-AKT inhibition triggered a widespread TUNEL staining in the otic epithelium and AVG, whilst IGF-I reduced cell death (Figure 7 B, compare a with b and c and d; quantification in e). Moreover, IGF-I restored low levels of TUNEL staining even after pretreatment with either of the inhibitors (Figure 7 B, quantification in e). The increase in TUNEL-positive cells was more pronounced upon PI3K inhibition (Figure 7 B, c, quantification in e). Both inhibitors affected otic vesicle size, LYtreated and AKTi-treated otic vesicle sizes were 63% and 73% of those of controls, respectively. Whereas the reduction observed in the AVG size were of 55% and 65% after treatment with LY and AKTi, respectively, with respect to 0S otic vesicles (data not shown).
To further explore AKT actions in neural cells, AVG were explanted and cultured for 20 h in the presence of either LY or AKTi. Upon treatments, a reduction in Islet-1-positive neuroblast population was observed (Figure 7 C, compare a with b and c). The Islet-1-positive area of the AVG (neural somas area) was significantly reduced with either inhibitor (Figure 7 C, compare a with b and c; quantification in d). Neurite extension (neural processes area) was also significantly affected upon PI3K and AKT inhibition (Figure 7 C, compare a with b and c, arrowheads; quantification in d).

Discussion
The mechanisms that control the survival and proliferation of otic neural progenitors are still poorly understood. We show here that otic neural progenitors are protected from programmed cell death by IGF-I, which increases FoxM1 levels and p27Kipnegative neural population, thus maintaining neuroblasts in a proliferative and undifferentiated state. IGF-I promotes survival through the activation of IGF1R, leading to PI3K and AKT activation. Inhibition of AKT causes generalized apoptosis of neural progenitors. Post-mitotic p27 Kip -positive neurons become independent of IGF-I for survival and, concomitantly, neurotrophin receptors are up-regulated and otic neurons extend the fibers that connect them to the sensory epithelia. Double immunostaining was carried out for SOX2 (magenta) and for the neuroblast nuclear marker Islet-1 (green). The quantification of SOX2-positive population was performed as described in Materials and Methods). SOX2 is expressed in the otic epithelial neuroblast population (a, arrowhead) and it is significantly increased in presence of IGF-I (b, arrowhead, e) but it is markedly reduced in the presence of either AKTi or LY (c and d, arrowheads, e). Representative images of compiled confocal microscopy projections are shown from at least six otic vesicles per condition from at least three independent experiments. All otic vesicles are orientated in the same way: A, anterior; D, dorsal. Scale bar, 75 mm. The data are shown as the mean 6 SEM and the statistical significance between the different conditions was estimated by Student's t-test: *P,0.05, ***P,0.001, versus control (0S). (B) TUNEL labeling and PH3 expression upon PI3K-AKT inhibition. Otic vesicles were isolated from HH18 chicken embryos and they were cultured in the same conditions as in A. Apoptotic cell death was visualized by TUNEL (green) in cultured otic vesicles, and immunostaining for the mitotic marker phospho-Histone-3 was performed (PH3, magenta). Both TUNEL and PH3-positive cells were quantified (as described in Materials and Methods). PH3 levels did not vary with the different treatments, but TUNEL levels decreased markedly in the presence of IGF-I and increased dramatically with either inhibitor. Treatment with IGF-I restored low TUNEL-levels (e). The data are shown as the mean6SEM and the statistical significance between the different conditions was estimated by Student's t-test: *P,0.05, **P,0.01 and ***P,0.001 versus control (0S). (C) Acoustic-vestibular ganglia (AVG) explants were obtained from HH19 embryos and cultured in serum-free medium for 20 h with no additives (0S, a) or in the presence of LY294002 (LY, 25 mM, b) or AKTi VIII (AKTi, 50 mM, c). Immunostaining for TuJ-1 (magenta) and Islet-1 (green) were performed. Compared to 0S (a, arrowhead), the neuronal soma area of the LY and the AKTi-treated AVG is smaller (d). Areas were measured as described in Materials and Methods (d). The data are shown as the mean6SEM and the statistical significance between the different conditions was estimated by Student's t-test: *P,0.05, **P,0.01, versus control (0S). Scale bars, 75 mm. Representative images from compiled confocal microscopy projections of AVG are shown, from at least six otic vesicles or AVG explants per condition studied in at least three independent experiments. doi:10.1371/journal.pone.0030790.g007 Otic neuroblasts depend on IGF-I for survival and expansion Programmed cell death is part of the normal development of the nervous system [8], and it plays an important role during the earliest morphogenetic events of inner ear development [5], [6]. Apoptosis contributes to the regulation of cell number in the epithelial and ganglionar neuroblast populations, and it serves to remove aberrant cells. Total blockage of apoptosis with the caspase inhibitor Boc-D-FMK causes malformations in the otic vesicle, including the thickening of the neurogenic area in the otic epithelium, which suggests an impairment on neural progenitor migration from the otic epithelium. Accordingly, the size of the associated AVG appears reduced. This suggests that programmed cell death fulfils a complementary role in the epithelial neurogenic area, allowing the neuroblasts to detach from the epithelium. IGF-I reduces caspase-3 activation and TUNEL staining. However, the associated cell survival is spatiotemporally regulated so that it results in the persistence of specific spots of cell death. After IGF-I addition, apoptosis was negligible during the first 8 h in culture, suggesting that it protects undifferentiated otic progenitors. Apoptosis always increases after long periods in culture (24 h), and, even in presence of IGF-I, there is a localized region of cell death that corresponds to G4/p27 Kip positive post-mitotic neurons. Another region of cell death can be observed in the otic epithelium and corresponds with the neurogenic region. This area of cell death is also observed in vivo [6]. Otic neurons depend on brain-derived neurotrophic factor and/or neurotrophin-3, which are provided by the peripheral sensory system. It is likely that the absence of such neurotrophins causes neurotrophic cell death, a common mechanism to adjust neuronal numbers and to eliminate unhealthy or undesired cells [48].
Our results show that in the developing inner ear, IGF-Idependent neural survival is marked by the state of differentiation and the microenvironment, and it occurs within a time-window that precedes neurotrophin-dependence. IGF-I protects otic precursors so they continue to proliferate for longer periods. Proliferation in combination with a decreased rate of cell death results in an increase in the size of the otic vesicle. IGF-I increased the population of neuroblasts (Islet-1-positive), while reducing that of mature neurons (G4-and p27 Kip -positive) which become confined to the ventral aspect of the AVG. Accordingly, cell death occurs in post-mitotic neurons that down-regulate the expression of high affinity IGF-I receptors.

IGF-I acts through the PI3K-AKT pathway to promote the survival of neural progenitors
Here we show that the survival actions of IGF-I on otic neuroblasts are mediated by the activation of AKT via PI3K. AKT expression is regulated during early inner ear development. Active pAKT is present in the otic epithelium and in migrating and ganglion neuroblasts, as well as in otic neurons. Blockade of either PI3K or AKT activity causes a dramatic loss of both epithelial and ganglionic neuroblast populations. IGF-I partially rescues this phenotype possibly by activating alternate pathways, including those initiated by CRAF activation, which has been shown to promote survival in the otic vesicle [20]. It is worth noting that blocking PI3K causes both an increase in apoptosis and other morphological alterations which are not observed when AKT activity is inhibited. Members of the PI3K family regulate apoptosis but they also participate in other biological processes including autophagy [49], [50]. AKT inhibition increases apoptosis and decreases the number of SOX2-positive epithelial precursors and of Islet-1-positive/TuJ-1-negative immature neu-roblasts. SOX2 has been proposed as a marker of otic progenitors committed to a neural fate [46]. AKT does not affect the mature ganglionic population, but neurons did not extend neural processes, suggesting a role for AKT in differentiation. Indeed, AKT has been implicated in neurite outgrowth in other systems [51]. AKT has also been reported to directly modulate cell cycle, thus AKT phosphorylates and promotes the accumulation in the cytoplasm of Skp2, resulting in p27 ubiquitylation, degradation and cell cycle activation [28], [27].
There is previous evidence that IGF-I can promote neurogenesis [52], proliferation of neural progenitors [53], [54], [55], and proliferation of neural stem cells in culture [56]. Depending on the cell type and context, these actions of IGF-I are mediated by distinct signaling pathways [16], [20], [57], [58]. The basic mechanism underlying these actions is the capacity of IGF-I to promote G 1 /S cell cycle progression by regulating cyclin kinase activation via the activation of specific signaling pathways [55], [59]. IGF-I regulates the nuclear levels of cyclin-dependent kinase inhibitor p27 Kip in the rat cerebral cortex [55], and in the developing mouse cochlea [21]. FoxM1 transcription factor is an intermediate step in the inhibition of cell-cycle inhibitors p21 Cip and p27 Kip [60], [61]. Indeed, FoxM1 knock-out mice are embryonic lethal due to the lack of progenitor cell proliferation [62]. Here, we show that IGF-I acts via AKT/FoxM1 to promote long-term maintenance of neural otic precursors with the concomitant decrease of p27 Kip -positive cell numbers. As the expression of IGF-I high affinity receptors is reduced, otic neurons mature and express cyclin inhibitor p27 Kip .
In summary, we show here that the IGF-I/AKT signalling pathway is fundamental for survival of proliferative otic neuroblasts and the maintenance of the undifferentiated state. This suggests a crucial role of this pathway in establishing the final number of neurons and the timing at which neuron generation proceeds during otic development.

Supporting Information
Figure S1 Glial SOX10 expression in the inner ear. SOX10 (red) and Islet-1 (green) expression was studied by immunohistochemistry in HH19 chicken embryo sections (a, b) and in cultured otic vesicles (c, d). SOX10 was labelling neural crest glial cells within the AVG (a, b, arrowheads). In cultured OV there were no detectable SOX10-positive cells associated to the neuroblasts that express Islet-1 (c, d). Panels b and d correspond to the boxed areas in a and c, respectively. Representative microphotographs are shown from at least two embryos and six otic vesicles. Orientation: A, anterior; D, dorsal. Scale bars, 150 mm. (TIF) Figure S2 IGF-I promotes proliferation and survival in otic neuroblasts. (A) IGF-I promotes BrdU incorporation in epithelial neuroblasts. Otic vesicles were isolated from HH18 chicken embryos and cultured in serum-free medium without additives (0S; s) or supplemented with IGF-I (10 nM; b). Immunostaining for incorporated BrdU was performed (green). Representative images from confocal optic planes of BrdU incorporation in cultured otic vesicles are shown. The arrowheads point to the accumulation of BrdU-positive cells in the neurogenic region in the otic epithelium from the IGF-I condition. (B) IGF-I protects otic progenitors from programmed cell death induced by serum deprivation. Otic vesicles were isolated from HH18 chicken embryos and cultured for 4 (a-c) or 20 hours (d-f) in serum-free medium either without additives (0S; a, d) or supplemented either with IGF-I (10 nM; b, e) or FBS (2.5% v/v c, f). Cell death was visualized by TUNEL staining (green). Representative images of six otic vesicles per condition and from at least three independent experiments are shown, and they were obtained from compiled confocal microscopy projections of otic vesicles. Orientation, A, anterior; D, dorsal. Scale bars, 150 mm. (TIF)