Neural stem cells depend on insulin-like growth factor I (IGF-I) for differentiation. We analysed how activation and inhibition of the PI 3-kinase–Akt signalling affects the number and differentiation of mouse olfactory bulb stem cells (OBSCs). Stimulation of the pathway with insulin and/or IGF-I, led to an increase in Akt phosphorylated on residues Ser473 and Thr308 (P-AktSer473 and P-AktThr308, respectively) in proliferating OBSCs, and in differentiating cells. Conversely, P-AktSer473 levels decreased by 50% in the OB of embryonic day 16.5-18.5 IGF-I knockout mouse embryos. Overexpression of PTEN, a negative regulator of the PI 3-kinase pathway, caused a reduction in the basal levels of P-AktSer473 and P-AktThr308 and a minor reduction in IGF-I-stimulated P-AktSer473. Although PTEN overexpression decreased the proportion of neurons and astrocytes in the absence of insulin/IGF-I, it did not alter the proliferation or survival of OBSCs. Accordingly, overexpression of a catalytically inactive PTEN mutant promoted OBSCs differentiation. Inhibition of PI 3-kinase by LY294002 produced strong and moderate reductions in IGF-I-stimulated P-AktSer473 and P-AktThr308, respectively. Consequently, LY294002 reduced the proliferation of OBSCs and the number of neurons and astrocytes, and also augmented cell death. These findings indicate that OBSC differentiation is more sensitive to lower basal levels of P-Akt than proliferation or death. By regulating P-Akt levels in opposite ways, IGF-I and PTEN contribute to the fine control of neurogenesis in the olfactory bulb.
The number of neurons and glia that form during the development of the CNS is controlled by the coordination of extracellular factors, cell to cell interactions and intrinsic mechanisms that regulate neuroepithelial cell proliferation, differentiation and survival (Edlund and Jessell, 1999). Manipulating neural stem cells (NSC) is an experimental approach to determine the effects of extrinsic factors and of signalling pathways during development (Rajan et al., 2003; Sinor and Lillien, 2004). The differentiation of murine embryonic olfactory bulb stem cells (OBSCs) requires insulin-like growth factor I (IGF-I) both in culture and in vivo and, in addition, IGF-I and (pro)insulin promote OBSC proliferation (Vicario-Abejón et al., 2003). IGF-I has also been reported to regulate the proliferation and differentiation of neuroepithelial cells in other regions of the brain (Brooker et al., 2000; Arsenijevic et al., 2001; Hodge et al., 2004; Hsieh et al., 2004). The respective membrane receptors through which (pro)insulin and IGF-I transmit their signals – the insulin and IGF-I receptor – are widely expressed in the CNS and are found in the olfactory bulb (OB) and OBSCs (De Pablo and de la Rosa, 1995; Varela-Nieto et al., 2003; Vicario-Abejón et al., 2003; Bondy and Cheng, 2004). The intracellular signalling pathways activated upon insulin/IGF-I receptor stimulation include the phosphoinositide 3-kinase (PI 3-kinase) pathway and its downstream target, the serine/threonine kinase Akt (Taniguchi et al., 2006). The PI 3-kinase is made up of a regulatory (p85) and a catalytic (p110α) subunit. The p110α-knockout mice die at embryonic day (E) 9.5-10 showing CNS defects, in part due to impaired cell division (Bi et al., 1999). Accordingly, enhanced PI 3-kinase activation in the retina augmented proliferation and interfered with cell death in such a way that excessive growth of this tissue occurred during early development (Pimentel et al., 2002). Stimulation of Akt by exogenous insulin also decreased apoptosis in the proliferating retina (Diaz et al., 1999). Three isoforms of Akt (Akt1, Akt2 and Akt3, also known as PKBα, β, and γ, respectively) have been found in mammals, and recent reports indicate that Akt1 and Akt3 influence brain size and the number of cells in the brain (Easton et al., 2005; Tschopp et al., 2005).
Akt is activated by the association of phosphorylated phosphoinositide (PI) generated in reactions catalysed by PI 3-kinase. For complete activation, Akt must be translocated to the plasma membrane where it is phosphorylated at Thr308 and Ser473 (Bayascas and Alessi, 2005; Song et al., 2005). Akt activity is negatively regulated by molecules that antagonise PI 3-kinase, such as the dual lipid and protein phosphatase PTEN (Cantley and Neel, 1999; Simpson and Parsons, 2001). There is evidence that PTEN regulates multiple steps in CNS development (for reviews, see Li et al., 2003a; Stiles et al., 2004). Using a nestin-promoter-driven Cre transgene, PTEN has been deleted in neural precursor cells, producing an enlargement of the brain, possibly due to the enhanced proliferation and survival of neuroepithelial cells (Groszer et al., 2001; Groszer et al., 2006). It is also thought that PTEN regulates neuronal positioning, migration and cell size (Backman et al., 2001; Kwon et al., 2001; Li et al., 2002; Marino et al., 2002; Li et al., 2003a; Li et al., 2003b), as well as neuronal dendritogenesis and Bergman glia differentiation (Lachyankar et al., 2000; Musatov et al., 2004; Yue et al., 2005; Jaworski et al., 2005). Thus, it appears that the regulation and function of PTEN depends on the cell type and the stage of development. Furthermore, the PI 3-kinase–PTEN–Akt signalling pathway appears to be implicated in multiple processes that remain to be fully analysed in defined neural systems. Here, we have examined whether this signalling pathway mediates the effects of IGF-I on OBSCs and on the OB in vivo. Our results indicate that this pathway plays a prominent role in regulating the number of OBSCs, as well as their differentiation into neurons and astrocytes.
IGF-I stimulates the PI 3-kinase–Akt pathway in cultured OB neural stem cells and in the OB in vivo
We previously showed that IGF-I promotes the proliferation and differentiation of cultured OBSCs, and that neuronal and glial development is altered in the OB of Igf-I knockout mice (Pichel et al., 2003; Vicario-Abejón et al., 2003). To elucidate the biochemical pathways that may mediate the effects of IGF-I, we determined the levels of Akt phosphorylated on residues Ser473 and Thr308 (P-AktSer473 and P-AktThr308, respectively) and phosphorylated Erk1/2 (P-Erk1/2) in extracts from IGF-I-stimulated OBSCs cells (Fig. 1) and from the OB of E16.5-E18.5 Igf-I–/– and Igf-I+/+ mice (Fig. 2). Treatment of cultured cells with 100 ng/ml IGF-I produced a marked increase in the levels of P-AktSer473 and P-AktThr308 during OBSC proliferation and differentiation when compared with those found in control cultures. The increase in P-Akt levels was observed 5 minutes after addition of IGF-1 and lasted for at least 24 hours (data not shown). By contrast, Erk1/2 was highly phosphorylated in basal control conditions and phosphorylation was not stimulated further by IGF-I. Akt phosphorylation was IGF-I dependent because a 50% reduction was observed in P-AktSer473 levels in the OB obtained from E16.5-E18.5 Igf-I–/– mice compared with those of Igf-I+/+ mice (Fig. 2). The absence of IGF-I did not change the pattern of TdT-mediated dUTP nick-end labelling (TUNEL) staining in the E18.5 OB (data not shown).
PTEN overexpression impairs OBSC differentiation into neurons and astrocytes as well as basal levels of P-Akt
IGF-I appears to be an important factor in controlling P-Akt levels in the developing OB cells. To study the influence of the PI 3-kinase–P-Akt pathway in the proliferation and differentiation of neural stem cells, OBSCs were infected with a retroviral vector expressing PTEN-GFP (Fig. 3). On average, at least 80% of OBSCs were infected with the control GFP vector (Fig. 3B,D,F) and these cells proliferated in the same manner as the mock-infected cells (Fig. 3A,C,E). A similar high infection efficiency was obtained when cells were infected with the PTEN-GFP and the PTEN-C/S-GFP vector, and PTEN overexpression was confirmed by immunoblotting with an antibody against the HA tag or PTEN (Fig. 3G). The relative levels of PTEN-GFP and PTEN-C/S-GFP were increased 2.2-fold and 3.1-fold, respectively, relative to those of endogenous PTEN (Fig. 3G,H). Upon mitogen removal, the infected cells differentiated into neuronal class III β-tubulin (TuJ1)-positive neurons (Fig. 3I-K), glial fibrillary acidic protein (GFAP)-positive astrocytes (Fig. 3L-N), and oligodendrocyte marker O4 (O4)-positive oligodendrocytes (Fig. 3O-Q).
We analysed the effects of PTEN overexpression on OBSC proliferation and differentiation by immunolabelling proliferative cells with 5-bromo-2-deoxyuridine (BrdU), neurons with TuJ1, astrocytes with GFAP and oligodendrocytes with O4 (Fig. 4). Since the standard culture medium contained a high concentration of insulin (10 μg/ml) that might counteract the effects of PTEN overexpression, these analyses were performed in cells cultured with and without insulin/IGF-I. In a previous study (Vicario-Abejón et al., 2003) we had demonstrated that this concentration of insulin promoted differentiation to a similar extent as 100 ng/ml of IGF-I and, thus, we have used these concentrations of either factor interchangeably. The total number of GFP-positive cells and the percentage of BrdU-GFP double-labelled cells was similar in the PTEN-infected cultures and in the GFP-infected cultures, both with or without insulin (Fig. 4A). By contrast, PTEN overexpression produced a significant decrease in the percentages of TuJ1-positive (46%) and GFAP-positive (35%) cells, in insulin-free cultures compared with cultures exposed to insulin (Fig. 4B,C). The percentage of oligodendrocytes in the insulin/IGF-I-free culture conditions was increased (40%) by PTEN overexpression although the increase was not statistically significant (Fig. 4D). To determine whether the reduction in the number of neurons and astrocytes following PTEN overexpression was due to an increase in cell death, cells were stained by using the TUNEL method. In differentiating OBSC cultures, PTEN overexpression did not augment the percentage of TUNEL-positive cells, nor did it decrease the total number of cells growing in the presence of insulin or IGF-I, or in the absence of these factors (Fig. 4E and data not shown). These results suggest that a moderate increase in PTEN levels in OBCS impairs differentiation into neurons and astrocytes without significantly affecting cell survival.
Since PTEN negatively regulates the PI 3-kinase–Akt pathway by dephosphorylating phosphatidylinositol (3,4,5)-trisphosphate [PtdIns(3,4,5)P3] (Waite and Eng, 2002), we determined the consequence of PTEN overexpression on Akt phosphorylation in control cultures and cultures stimulated with 100 ng/ml IGF-I (Fig. 5A). Basal P-Akt levels were reduced by 58.4% in cells infected with the PTEN-GFP construct, although stimulation with IGF-I resulted in a similar Akt phosphorylation in both PTEN-GFP and control cells. PTEN thus appeared to regulate basal levels of P-AktSer473 and P-AktThr308, whereas overexpression of PTEN did not modify basal or IGF-I-stimulated levels of P-Erk1/2 (data not shown). To further assess whether overexpression of PTEN partially inhibits Akt phosphorylation by IGF-I, stimulation experiments were performed by using low concentrations of the growth factor on OBSC-differentiating cultures prepared from Igf-I knockout mice. Ten ng/ml IGF-I increased P-AktSer473 levels and this effect was partially abrogated (46% reduction, average data from two experiments) by overexpression of PTEN; only minimal reductions in P-AktSer473 levels occurred when higher concentrations (20-100 ng/ml) of IGF-I were used in Igf-I knockout cells that overexpressed PTEN (Fig. 5B).
Overexpression of a PTEN mutant (PTEN-C/S-GFP) promotes neuron and astrocyte differentiation
To corroborate the implication of PTEN in neuronal and astrocyte differentiation, a catalytically inactive mutant (PTEN-C/S-GFP) was overexpressed in Igf-I knockout OBSCs. This mutant previously showed markedly reduced in vitro lipid phosphatase activity (Lacalle et al., 2004). PTEN-C/S-GFP overexpression caused an ∼20% increase in the number of TuJ1-positive neurons and of GFAP-positive astrocytes in OBSC cultures differentiating in the presence of IGF-I (Fig. 6A,B). In OBSC cultures differentiating in the absence of IGF-I (and of insulin), however, overexpression of the catalytically inactive mutant PTEN produced an increase of 52% and a 42% (P<0.05) of TuJ1-positive cells (Fig. 6A) and GFAP-positive cells (Fig. 6B), respectively, further indicating a role of the phosphatase PTEN in neuron and astrocyte differentiation. PTEN-C/S-GFP overexpression produced no change in P-Akt levels in OBSC extracts (data not shown), as has been reported for Jurkat cells (Lacalle et al., 2004).
Effects of PI 3-kinase inhibition on OBSC number and differentiation
Since a moderate increase of PTEN levels in the cells caused a reduction in the basal levels of P-Akt and OBSC differentiation into neurons and astrocytes, we studied the consequences of further inhibiting the PI 3-kinase activity on these processes. Exposing cultures to the PI 3-kinase inhibitor LY294002 (30 μM) (Vanhaesebroeck et al., 2005) in the presence of IGF-I provoked a marked reduction in the levels of P-AktSer473 after 24 hours, during both OBSC proliferation (Fig. 7A) and differentiation (Fig. 8A). The P-AktThr308 levels were only moderately reduced by LY294002 (Fig. 7A, Fig. 8A) and, although P-Erk1/2 was not significantly affected in proliferative OBSCs (Fig. 7A), the levels of P-Erk1 (but not P-Erk2) were slightly diminished in LY294002-treated cells undergoing differentiation (Fig. 8A). We then examined the effects of downregulating PI 3-kinase activity on the number of proliferative OBSCs (Fig. 7) and on the production of OBSC-derived neurons and astrocytes (Fig. 8). LY294002 inhibition of Akt phosphorylation produced decreases of 56% (P<0.01) and 46% (P<0.001) in the proportion of BrdU-positive cells growing in the presence of IGF-I or its absence (BSA added instead of IGF-I), respectively, compared with cultures maintained with the vehicle (DMSO added) (Fig. 7B,E,F). The effect of LY294002 was partially reverted by IGF-I; the percentage of BrdU-positive-cells was 64% higher (P<0.01) than that of cells exposed to LY294002 in the presence of BSA (basal levels, Fig. 7B). Total cell number was not significantly affected by LY294002 in basal cultures but it was reduced by 27% (P<0.05) in the presence of IGF-I (Fig. 7C,E-H). Indeed, addition of LY294002 to the cultures caused a 47.6% (P<0.05) and a twofold increase (P<0.001) in the proportion of TUNEL-positive cells in the presence of BSA and IGF-I, respectively (Fig. 7D,G,H). IGF-I counteracted the effect of LY294002 on cell death because the percentage of TUNEL-positive cells in the presence of IGF-I was 50% (P<0.05) lower than with BSA. IGF-I thus partially protected cells from LY294002-induced cell death under conditions in which P-Akt levels were low.
PI 3-kinase inhibition by LY294002 also affected OBSC differentiation. Exposure to LY294002 produced a 58% (non significant) and 51% (P<0.01) reduction in the number of TuJ1-positive cells in the presence of BSA and IGF-I, respectively (Fig. 8B,E,F). After exposure to LY294002, the number of TuJ1-positive cells was approximately fivefold greater (P<0.001) in the presence of IGF-I than with BSA, further evidence that IGF-I partially protects the cells from the effects of LY294002 when P-Akt levels were reduced. Similar responses were observed in the number of astrocytes in differentiating OBSC cultures (Fig. 8C,G,H). LY294002 produced a 40% (P<0.001) and 43% (P<0.001) decrease in the percentage of GFAP-positive cells in the presence of BSA and IGF-I, respectively. However, the percentage of IGF-I-treated GFAP-positive cells was twofold (P<0.001) in the presence of IGF-I than compared with control cells treated with BSA. Astrocytes cultured with IGF-I together with LY294002 had less processes than cells growing in IGF-I together with DMSO (Fig. 8G,H). In the absence of IGF-I, the percentage of TUNEL-positive cells was similar in cultures treated with or without LY294002, whereas in the presence of IGF-I, LY294002 produced in a 27% increase (P<0.05) in these cells (Fig. 8D,I,J). In accordance with the earlier data, the proportion of TUNEL-positive cells induced by LY294002 in the presence of IGF-I was nevertheless significantly lower (P<0.01) than in BSA. Overall, these results demonstrate an important role of the PI 3-kinase–Akt pathway during the differentiation of OBSCs, concurring with a minor effect on cell survival.
Neuronal differentiation and allocation of neurons to cell layers in the OB requires IGF-I in vivo. Similarly, proliferation and differentiation of OBSCs in culture depends on (pro)insulin and IGF-I (Vicario-Abejón et al., 2003). We have studied the biochemical pathways that are activated by IGF-I in the OBSCs, and the role of the PI 3-kinase–Akt pathway in regulating OBSC number and differentiation. We show that in the absence of insulin or IGF-I, the basal levels of phosphorylated Akt in cultured OBSCs are very low and that IGF-I greatly stimulated Akt phosphorylation, at both the regulatory Ser473 residue and the catalytic Thr308 residue. More importantly in physiological terms, IGF-I is needed to maintain normal levels of P-Akt in the OB in vivo. By contrast, high basal levels of P-Erk1/2 were detected and further stimulation by IGF-I did not alter the phosphorylation of this kinase. Moderate overexpression of PTEN in the OBSC diminished the basal phosphorylation of Akt as well as that stimulated by IGF-I (only by 10 ng/ml). Moreover, PTEN overexpression impaired neuronal and astrocyte differentiation without significantly affecting cell proliferation or cell death. Accordingly, overexpression of a catalytically inactive mutant PTEN promoted OBSC differentiation. Pharmacological inhibition of PI 3-kinase by LY294002 blocked the effects of IGF-I on Akt phosphorylation. Likewise, the proliferation and survival of OBSC was impaired, and neuron and astrocyte differentiation was diminished. Thus, whereas a reduction in basal P-Akt affects OBSC differentiation, inhibition of IGF-I-induced PI 3-kinase–P-Akt activation blocks proliferation and survival.
The marked effect of IGF-1 on Akt phosphorylation and the failure to activate Erk1/2 in the OBSCs is in contrast to the stimulation of the Erk1/2 pathway by IGF-I in adult hippocampal stem cells (HPSC). However, in contrast to the high basal P-Erk1/2 levels found in embryonic OBSCs, adult HPSC display low levels of P-Erk1/2 (Aberg et al., 2003). This difference may partially account for the relatively higher rate of proliferation of embryonic NSC when compared with adult NSC (Gritti et al., 2002; Aberg et al., 2003; Vicario-Abejón et al., 2003; Yusta-Boyo et al., 2004). Therefore, we focused our attention on the PI 3-kinase–Akt pathway.
Although the activated PI 3-kinase–Akt pathway plays a major role in preventing cell death in many cell types (Brazil et al., 2004; Song et al., 2005), our results show that this pathway also increases OBSC proliferation. Moreover, both the prevention of cell death and the increase in proliferation, lead to an increase in the number of stem cells in this system. The PI 3-kinase–Akt pathway has also been shown to influence self-renewal in cortical stem and/or precursor cells overexpressing Akt-1 (Sinor and Lillien, 2004) as well as in retinal precursors overexpressing p65, an activated form of the regulatory subunit of PI 3-kinase (Pimentel et al., 2002). Indeed, defects in neural proliferation have been observed in mice carrying a deletion in the p110α subunit of PI 3-kinase (Bi et al., 1999). We found that the PI 3-kinase–P-Akt pathway is involved in OBSC differentiation into neurons and astrocytes, with a minimal influence on cell death. The reduction in P-Akt levels in the OB of E16.5-E18.5 IGF-I knockout mice, concomitant with impaired neuronal and glial differentiation further supports a role of Akt in neuronal differentiation within the OB in vivo (Pichel et al., 2003; Vicario-Abejón et al., 2003). Akt overexpression was recently shown to promote the generation of NG2-positive oligodendrocyte precursors but not the generation of neuron or astrocyte precursors (Sinor and Lillien, 2004). However, by inhibiting PI 3-kinase activity and Akt phosphorylation, we found that the PI 3-kinase–Akt pathway did regulate neuron and astrocyte differentiation. Further, our findings are in agreement with those of Easton et al. (Easton et al., 2005) and Tschopp et al. (Tschopp et al., 2005), who recently reported a reduction in brain size and in the number of brain cells in Akt1 and Akt3 knockout mice. This emphasises the important role of Akt activation in establishing the final composition of the nervous system.
LY294002 produced a strong decrease in the levels of IGF-I-stimulated P-AktSer473, as well as a significant increase in cell death during OBSC proliferation and inhibition of OBSC differentiation. Inhibition of the PI 3-kinase pathway in the presence of IGF-I, however, did not increase the number of TUNEL-positive cells nor reduce the number of BrdU-positive cells, neurons or astrocytes beyond the control (BSA) levels. This could be due to residual Akt kinase activity because LY294002 did not completely block phosphorylation at Thr308 (our results) (Brodbeck et al., 1999; Sarbassov et al., 2005). Alternatively, other biochemical pathways stimulated by IGF-I or other growth factors produced by the OBSCs may be involved and they may affect the processes observed. High concentrations of LY294002 also slightly reduced the levels of P-Erk1, as we observed in differentiating cells, suggesting a degree of cross-talk between these pathways in OBSCs. Preliminary experiments in which Erk1/2 is blocked by the MAP kinase inhibitor UO126 indicate a contribution of this pathway to regulate proliferation of OBSCs. It is also known that there are downstream targets of PI 3-kinase, apart from Akt, which can influence cell cycle kinetics and cell migration (García et al., 2006).
IGF-I is a very potent stimulus for the OBSCs as the effects of overexpressing the phosphatase PTEN were completely blocked by 100 ng/ml IGF-I or μM concentrations of insulin. In fact, whereas PTEN overexpression reduced basal P-Akt levels, it only affected the levels of IGF-I-stimulated P-Akt at low concentration of IGF-I (10 ng/ml). These findings are in accordance with data from other systems (Myers et al., 1998; Lacalle et al., 2004; Seminario et al., 2004), and support the concept that the main role of PTEN is in regulating the basal levels of PtdIns(3,4,5)P3 and thus, of phosphorylated Akt. The reduction in basal P-Akt appears to result in the specific effects of high PTEN expression, that is the decrease in OBSC differentiation into neurons and astrocytes in the absence of insulin or IGF-I. Accordingly, PTEN overexpression in PC12 cells inhibits neuronal differentiation; a reduction in nerve growth factor (NGF) stimulated levels of both P-Akt and P-Erk1/2 was also reported (Musatov et al., 2004). By contrast, PTEN overexpression in neuroblastoma cells was seen to decrease the stimulatory effect of IGF-I on Akt phosphorylation without affecting the Erk1/2 pathway (van Golen et al., 2001). Thus PTEN appears to influence neuronal and astrocyte differentiation, as supported by our studies with a catalytically inactive PTEN mutant and in a recent study indicating that PTEN is involved in the differentiation of cerebellar Bergmann glia cells (Yue et al., 2005). The decrease in the number of neurons and astrocytes in OBSCs overexpressing PTEN (growing in the absence of insulin or IGF-I) was concomitant with a tendency to augment the number of oligodendrocytes. Hence, this cell lineage may well depend on additional signalling pathways for differentiation, a phenomenon that ought to be explored. Indeed, we have previously shown that IGF-I and BDNF collaborate to promote oligodendrocyte differentiation (Vicario-Abejón et al., 2003). We would have expected to observe an increase in P-Akt levels in cultures that overexpress PTEN-C/S-GFP concomitant with the increase in OBSC differentiation. However, no putative rise in P-Akt levels was observed in a single time-point studied, according to the effects of the mutant PTEN-C/S-GFP seen in linfoid cells (Lacalle et al., 2004). Whether PTEN-C/S-GFP overexpression can produce a transient change in Akt phosphorylation that is undetectable by the methods used here remains to be explored.
PTEN knockout mice do not display any changes in neuron or astrocyte numbers (Groszer et al., 2001). We found that following PTEN overexpression, neuronal and glia differentiation was only inhibited in the absence of insulin. Thus, the effects of PTEN deletion on differentiation might be masked by the presence of insulin in the culture medium (Groszer et al., 2001) that maintained high levels of P-Akt in the knockout or wild-type cells. Indeed, PI 3-kinase–Akt has recently been proposed to participate in the control of neuronal dendritogenesis and arborisation, modulated by upstream Ras and balanced by PTEN (Jaworski et al., 2005; Kumar et al., 2005). In these studies, mTOR appears to be the downstream target of PI 3-kinase–Akt and it has been proposed that mTOR is a convergence point for multiple signalling pathways that regulate neuronal development.
Under our experimental conditions, the moderate increase in the levels of PTEN (2.2-fold) did not affect the proliferation of OBSCs in the presence of fibroblast growth factor 2 (FGF-2) and epidermal growth factor (EGF). These mitogens activate the Erk1/2 pathway (data not shown) (Kessaris et al., 2004) and could then counteract the putative effects of PTEN phosphatase activity on proliferation. Accordingly, we cannot rule out a role for PTEN in maintaining the homeostasis of cell proliferation in the OB in vivo. Indeed, deletion of PTEN provokes changes in proliferation in other systems, producing more proliferative cells in the ventricular zone of the telencephalon at E14.5 (Groszer et al., 2001; Groszer et al., 2006) and subventricular zone (SVZ) neurospheres (Li et al., 2002). The loss of PTEN also increases astrocyte proliferation (Fraser et al., 2004) and it reduces the number of proliferative cells in the E15.5 cerebellum (Marino et al., 2002) and in the SVZ (Li et al., 2002). These discordant results probably reflect the different influences of PTEN on neural proliferation, or the cell-type- and developmental-stage-dependent nature of its effects. This same conclusion might apply to the role of PTEN in cell death. Fewer TUNEL-positive cells have been reported in the cortical ventricular zone (Groszer et al., 2001) and cerebellum (Marino et al., 2002) of PTEN knockout mice, and in the SVZ of heterozygous PTEN mice (Li et al., 2002). However a GFAP-Cre controlled deletion of PTEN did not affect cell death in the hippocampus and cerebellum (Backman et al., 2001). The influence of PTEN on cell death may vary in the presence of a toxic insult, because hippocampal neurons overexpressing PTEN were more sensitive to cell death triggered by excitatory amino acids, and PTEN+/– and PTEN–/– neurons were more resistant to cell death than wild-type neurons (Gary and Mattson, 2002; Li et al., 2002; Yue et al., 2005). Overexpression of PTEN, nevertheless, did not induce cell death in OBSCs in the absence of a toxic insult.
Together, our findings suggest an important role for the PI 3-kinase–Akt pathway in controlling neural stem/precursor cell number, and their differentiation into neurons and astrocytes. They also indicate that these processes are differentially regulated by distinct levels of P-Akt. In this respect, neuronal and astrocyte differentiation is more sensitive to reduced P-Akt activity than neural stem cell proliferation or cell death. Through the fine modulation of P-Akt levels, in collaboration with additional signalling pathways, IGF-I and PTEN regulate cell development in the OB.
Materials and Methods
Neural stem cell cultures
Tissue culture reagents were purchased from Gibco-Life Technologies (Carlsbad, CA) and Sigma (St Louis, MO), whereas EGF, FGF-2 and brain-derived neurotrophic factor (BDNF) were obtained from PeproTech (Rocky Hill, NJ). Human insulin was a kind gift of Eli Lilly (Indianapolis, IN) or was purchased from Sigma. Neural stem cells (NSC) were prepared from the olfactory bulb (OB) of E14.5 embryos obtained from CD1 mice or from Igf-I knockout mice (Liu et al., 1993; Vicario-Abejón et al., 2003); the day on which a vaginal plug was detected being considered E0.5. Genotypes were determined by PCR using two primer sets specific either for the Igf-I gene (sense primer: 5′-GTCTAACACCAGCCCATTCTGATT-3′, antisense primer: 5′-GACTCGATTTCACCCACTCGATCG-3′) yielding a 250 bp product, or for the neomycin gene (sense primer: 5′-GCTTGGGTGGAGAGGCTATTC-3′, antisense primer: 5′-CAAGGTGAGATGACAGGAGAT-3′) yielding a 284 bp product. PCR conditions were: denaturing at 94°C for 30 seconds, annealing at 58°C for 45 seconds, and extension at 72°C for 45 seconds for 35 cycles. The product was detected by agarose gel electrophoresis. All animal care was in accordance with European Union guidelines.
After removing the brain from the skull, the OB was dissected and mechanically disaggregated. The resulting cells were resuspended in DMEMF12/N2 medium, made from Dulbecco's modified Eagle's medium (DMEM) nutrient mixture F12 (F12) containing insulin, apotransferrin, putrescine, progesterone, sodium selenite (N2), and then plated on uncoated tissue culture dishes at a density of 3.5×104 cells/cm2. FGF-2 and EGF (20 ng/ml each) were added daily to expand and passage the proliferating precursor cell population (Vicario-Abejón et al., 2003).
Infection of NSC with retroviral vectors
Infection was performed using a Moloney murine leukaemia virus-based retroviral vector (Mo-MLV). The retroviral construct pRV-IRES-GFP encodes an improved variant of the Aequorea Victoria green fluorescent protein (GFP), and it was obtained by cloning a SacII-NotI fragment of the pLZR-IRES-GFP vector into a pRV plasmid. This GFP vector was used as a control in infection experiments. A construct encoding the complete PTEN cDNA (1.2 kb) fused to a cDNA of the hemaglutinine (HA) N-terminus was cloned into the pRV-IRES-GFP vector, and in the resulting vector pRV–HA-PTEN-IRES-GFP (PTEN-GFP) PTEN expression was under LTR viral control. The PTEN catalytically inactive mutant (PTEN-C/S-GFP), described and characterised recently (Lacalle et al., 2004), was also cloned into the pRV-IRES-GFP vector. This mutated form has no HA-tag but is also recognised by the PTEN antibody used to detect the wild-type PTEN.
To infect CD1- or Igf-I–/–-derived NSC, proliferating cells (3.5×106) were resuspended in 500 μl of viral supernatant (3-5×106 IU/ml) and 500 μl DMEMF12/N2 in the presence of 6 μg/ml of polybrene, 20 ng/ml EGF and 20 ng/ml FGF-2. The optimum ratio of the number of cells to viral supernatant was established in pilot experiments. The cell suspension was plated in multiwell plates and centrifuged at 2000 g, 37°C for 1 hour. The cells were then collected, washed and cultured at a density of 15,000-20,000 cells/cm2 and incubated in the presence of EGF and FGF-2 (37°C, 5% CO2). The cells were passaged 3 days later and plated for 2-3 days in four-well chamber slides in proliferation (5000-6000 cells/cm2, with EGF and FGF-2) or differentiation conditions (100,000 cells/cm2 without mitogens) to study the effects of PTEN-GFP and PTEN-C/S-GFP overexpression in NSC. Proliferating cells were incubated with 5 μM 5′-bromo-2-deoxyuridine (BrdU, Boehringer-Mannheim) for 20-22 hours prior to fixation with 4% paraformaldehyde. Some cultures were mock-infected to test whether the protocol produced changes in the patterns of NSC proliferation and differentiation. To calculate the efficiency of infection, cells were harvested and analysed by flow cytometry. The three retroviral constructs had similar infection titles tested in NIH3 3T3 cells and produced comparable percentages of infected NSC.
Incubation of NSC with LY294002
To study the effect of inhibiting PI 3-kinase on neural stem cells, cultures were incubated with LY294002 (Calbiochem, La Jolla, CA). After passage, NSC cells were plated in proliferation conditions (DMEMF12/N2 without insulin containing 20 ng/ml EGF and 20 ng/ml FGF-2) or differentiation conditions (without mitogens). To these cultures, 100 ng/ml IGF-I or, as a control, 0.001% BSA were added for 24 hours, and 5 or 30 μM LY294002 (dissolved in DMSO) or 1-1.5% DMSO were also added for an additional 24 hours. The cells were then fixed for immunostaining and TUNEL assays, or lysed for western blot analysis.
Immunoblotting of extracts from cultured cells and OBs
Immunoblotting was performed on IGF-I-stimulated cultures, LY294002-treated cultures, PTEN-GFP- and GFP-infected cultures, and on dissected mouse OBs. To analyse the effect of IGF-I on the phosphorylation of Akt and Erk, cell cultures were grown for 24-48 hours under conditions of proliferation or differentiation in the absence of insulin, and they were then treated with BSA (control) or 10 to 100 ng/ml IGF-I for 5 minutes to 24 hours. The cells cultured under these different conditions, as well as OB tissue dissected out from E16.5-E18.5 mouse embryos from an Igf-I knockout colony (Liu et al., 1993; Vicario-Abejón et al., 2003), were incubated in lysis buffer [50 mM Tris HCl pH 7.4, 300 mM NaCl, 1% Triton X-100, 1 mM EDTA, 1 mM Ortovanadate, 25 mM NaF, 4 mM sodium pyrophosphate and a complete mini EDTA free protease inhibitor tablet (Roche)] for 20 minutes at 4°C. The lysed cells were centrifuged at 20,000 g, 4°C for 15 minutes, and the supernatant collected and stored at –20°C for further analysis by immunoblotting. Protein extracts were fractionated by electrophoresis on 7.5% polyacrylamide gels and transferred to nitrocellulose membranes. The membranes were treated with 2 mg/ml BSA, 1% Tween 20, PBS for 1 hour at room temperature (RT), and then incubated overnight at 4°C with primary antibodies against: PTEN (goat polyclonal 1:1000, St Cruz Biotechnology, CA or rabbit polyclonal 1:2000, Cell Signalling, Beverley, MA), β-tubulin (mouse monoclonal 1:10,000, Sigma), HA (mouse monoclonal 1:200, St Cruz), P-AktSer437 (rabbit polyclonal 1:2000, Cell Signalling), P-AktThr308 (rabbit polyclonal 1:2000, Cell Signalling), Akt1/2 (goat polyclonal 1:2000, St Cruz), P-Erk1/2 (rabbit polyclonal 1:2000, Cell Signalling) or Erk1/2 (mouse monoclonal 1:1000; Zymed, San Francisco, CA). After washing, the membranes were incubated for 1 hour at RT with horseradish peroxidase (HRP)-conjugated secondary antibodies and then incubated with the enhanced chemilumiscence (ECL) reagent (Amersham).
Immunostaining of cultured cells
Cultured cells were fixed with 4% paraformaldehyde in 0.1 M phosphate buffer pH 7.4, for 25 minutes. Cells were incubated overnight at 4°C or RT with primary antibodies against: β-III-tubulin (TuJ1, rabbit polyclonal, 1:2000; Covance, Berkeley, CA), GFAP (rabbit polyclonal, 1:1000; Dako, Glostrup, Denmark), O4 (mouse monoclonal IgM, 1:8, kindly shared by A. Rodríguez Peña, Instituto de Investigaciones Biomédicas, Madrid), BrdU (mouse monoclonal, 1:1000; the Developmental Studies Hybridoma Bank) or GFP (mouse monoclonal 1:500; rabbit polyclonal 1:1000; Molecular Probes, Eugene, OR). The anti-BrdU antibody (G3G4) was generated by S. J. Kaufman and was obtained from the Developmental Studies of Hybridoma Bank maintained by the University of Iowa Dept. Biological Sciences, Iowa City, IA. Cells were then incubated with Texas Red and/or Alexa Fluor 488-conjugated secondary antibodies (1:200; Molecular Probes). Alternatively, they were incubated with a biotinylated secondary antibody (1:200), then with an avidin-biotin-HRP complex (Vectastain ABC kit, Vector, Burlingame, CA) and subsequently developed with 3,3′-diaminobenzidine (DAB, Sigma) and H2O2. Controls were performed to confirm primary and secondary antibody specificity. Some cultures were stained with 4',6-diamidino-2-phenylindole (DAPI, Vector Laboratories, Burlingame, CA).
Detection of apoptosis
Cultured cells were fixed for 25 minutes in 4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.3-7.4. Apoptotic cell death was determined by TdT-mediated dUTP nick-end labelling (TUNEL) using the Apoptosis Detection System (Promega, Madison, WI) as described previously (Diaz et al., 1999).
To determine the number of cells expressing a specific antigen, ten random fields per chamber were counted using a 40× objective and fluorescence filters or bright-field optics (Zeiss Axioplan microscope). The results are expressed as the number of cells stained for the antigen per ten fields. In cultures infected with the GFP and PTEN-GFP expression vectors, the proportions of TuJ1-positive, GFAP-positive or BrdU-positive cells co-labelled with anti-GFP were calculated with respect to the total number of GFP-expressing cells. The data obtained in PTEN-GFP- or in PTEN-C/S-GFP-infected cultures were expressed as a percentage of the data obtained in GFP infected cultures, which were considered to be 100%. The results were expressed as the average ± s.e.m. of 4-8 cultures from 2-5 experiments unless otherwise stated. Statistical analysis was performed using the Student's t test. The optical density of the specific immunoblot bands was measured by densitometry using Quantity One software. The levels of PTEN, P-Akt and P-Erk1/2 were normalized against levels of actin, Akt1/2 and Erk1/2, respectively.
We thank Antonio Bernad (CNB, CSIC) for helpful comments throughout the study and Santos Mañes (CNB, CSIC) for providing the PTEN-C/S-GFP construct. This work was funded by the grants from the MEC (Spain) BFU 2004-2352 to F.d.P., SAF2004-05798 to C.V.-A. and BMC 2003-07751 to E.J.d.l.R., and from the Fundación `la Caixa' NE03/72-02 to C.V.-A. G.O., M.J.Y.-B, E.V.-V. and H.R.M.-G. were supported by doctoral fellowships either from the MEC or the `Comunidad Autónoma de Madrid'.