Regulation of Neuronal Survival by the Serine-Threonine Protein Kinase Akt

See allHide authors and affiliations

Science  31 Jan 1997:
Vol. 275, Issue 5300, pp. 661-665
DOI: 10.1126/science.275.5300.661


A signaling pathway was delineated by which insulin-like growth factor 1 (IGF-1) promotes the survival of cerebellar neurons. IGF-1 activation of phosphoinositide 3-kinase (PI3-K) triggered the activation of two protein kinases, the serine-threonine kinase Akt and the p70 ribosomal protein S6 kinase (p70S6K). Experiments with pharmacological inhibitors, as well as expression of wild-type and dominant-inhibitory forms of Akt, demonstrated that Akt but not p70S6K mediates PI3-K-dependent survival. These findings suggest that in the developing nervous system, Akt is a critical mediator of growth factor-induced neuronal survival.

The intracellular signaling pathways by which growth factors promote survival—in particular, survival of neurons of the central nervous system—are not well characterized. The survival of certain subsets of neurons of the peripheral nervous system can be promoted by the activation of a pathway that includes the guanosine triphosphate (GTP)- binding protein Ras and a series of protein kinases leading to mitogen-activated protein kinase (MAPK) (1, 2). In addition, a pathway that includes the lipid kinase PI3-K is important for the survival of several cell lines (3, 4), although the mechanisms by which PI3-K promotes survival are unclear. We investigated the contribution of two targets of PI3-K, the serine-threonine kinase Akt (57) and p70S6K (8), to the IGF-1-mediated survival of cerebellar neurons. We found that Akt has a critical role in promoting IGF-1-dependent survival.

For these studies, we used a well-characterized culture system of cerebellar neurons (9, 10). Large numbers of neurons of relatively homogeneous composition (consisting primarily of granule neurons) can be obtained, thus allowing biochemical analyses (9). Withdrawal of survival factors leads to the rapid and synchronous apoptosis of cerebellar neurons (9). About 50% of the cells were apoptotic within 1 day (Fig. 1), and almost all of the cells died within 3 to 4 days. The dying neurons showed characteristic features of apoptosis, including nuclear condensation and cleavage of chromatin into oligonucleosomal fragments (Fig. 1E) (9, 11). The apoptosis could be inhibited by defined trophic factors (9), including IGF-1 (Fig. 1C); insulin at superphysiological concentrations (Fig. 1D), which is believed to act through the IGF-1 receptor (10); and membrane-depolarizing concentrations of KCl, which lead to increased concentrations of intracellular calcium and may therefore simulate neuronal activity (9). The effects of IGF-1 and high concentrations of insulin on cerebellar neuron survival may reflect an in vivo function of IGF-1, because both IGF-1 and its receptor are expressed in the cerebellum, and transgenic mice overexpressing IGF-1 show increases in cell number in the brain (12).

Fig. 1.

Induction of apoptosis or survival of cerebellar neurons (10). (A) Phase-contrast photomicrograph of cerebellar neurons in complete medium [containing 25 mM KCl (high KCl) and 10% calf serum]. Cells were healthy, as indicated by large size and phase-brightness. (B) After 22 hours of deprivation (5 mM KCl, no serum), many neurons died. Death of cells was also confirmed by staining with trypan blue (11). (C and D) Death was inhibited by IGF-1 (25 ng/ml) (C) or insulin (10 μg/ml) (D). (E) Deprivation of cerebellar neurons induces chromatin cleavage. Starting at 4 to 8 hours after deprivation of trophic factors, extensive DNA laddering was present (2). U, untreated; positions of molecular size markers (in kilobases) are indicated on the left.

We first identified the signal transduction pathways that are activated in cerebellar neurons by IGF-1 or insulin. We examined activation of the Ras-MAPK pathway, because this pathway has been implicated in the control of nerve growth factor (NGF)-induced survival (1, 2). Insulin and IGF-1 failed to activate MAPK, although brain-derived neurotrophic factor (BDNF) efficiently activated MAPK (Fig. 2, A and B) (13). These results suggest that activation of the Ras-MAPK pathway is not critical for IGF-1-promoted or insulin-promoted survival of cerebellar neurons, and they also corroborate reports that insulin fails to activate MAPK in certain cell types, including chick forebrain neurons (14).

Fig. 2.

Activation of signaling pathways in cerebellar neurons. (A and B) Lack of activation of MAPK by insulin (13). Cells were left untreated in complete medium containing 25 mM KCl and calf serum (10%) (U), or were placed in medium containing insulin, IGF-1, no stimulant (−), or BDNF for the indicated times. In (A), activation of MAPK is accompanied by phosphorylation that reduces its mobility upon SDS-PAGE (38). Untreated cells (U) contained a basal amount of MAPK activation; BDNF, but not insulin, caused increased activation. In (B), similar results were seen when MAPK activity was measured with an immune-complex kinase assay with MBP as substrate (10 min treatment). (C) Induction of association between IRS-1 and p85 by insulin and IGF-1. Cells were left untreated or were placed in medium containing no stimulant (−), insulin, or IGF-1. At the indicated times, cells were lysed, IRS-1 was immunoprecipitated, and associated p85 was detected by immunoblotting (13). Untreated cells contained a basal amount of p85 associated with IRS-1, which was lost when cells were starved (−) but increased when cells were stimulated with insulin or IGF-1. (D) Activation of PI3-K lipid kinase activity (16). IGF-1 [and insulin (11)] rapidly activated PI3-K, leading to increased production of the reaction products phosphatidylinositol-3-phosphate (PIP) and phosphatidylinositol-3,4,5-trisphosphate (PIP3). (E and F) Induction of p70S6K phosphorylation and activation (13). Cells were left unstimulated, or were placed in media containing insulin, no stimulant (−), or IGF-1, in the absence or presence of wortmannin (Wort, 100 nM). At the indicated times, cells were lysed, and lysates were used in (E) for immunoblotting with antibody C2 to p70S6K (31) or in (F) for immune-complex kinase assays with ribosomal protein S6 as substrate. Activation of p70S6K is accompanied by p70S6K phosphorylation, which reduces its mobility upon SDS-PAGE (8). Insulin (and IGF-1) induced phosphorylation of p70S6K, which was blocked by wortmannin, and also activated p70S6K activity. For (A), (C), and (E), the migration positions of prestained molecular size markers (in kilodaltons) are indicated.

Activation of PI3-K is required for growth factor-induced survival of the PC12 neuronal cell line (3). By several criteria, we established that IGF-1 and insulin efficiently activate PI3-K in cerebellar neurons. First, IGF-1 and insulin induced binding of PI3-K to the receptor-associated protein IRS-1 (insulin receptor substrate 1) (Fig. 2C) (13), an interaction that mediates activation of PI3-K (15). Second, IGF-1 and insulin caused increased lipid kinase activity of PI3-K (Fig. 2D) (16). Third, IGF-1 and insulin activated p70S6K (Fig. 2, E and F) (13). As seen in many cell lines (8), phosphorylation of p70S6K was blocked by the PI3-K inhibitor wortmannin (Fig. 2E) (11), suggesting that IGF-1-induced or insulin-induced p70S6K activation in cerebellar neurons is dependent on PI3-K.

Consistent with a role for PI3-K in cell survival, we found that the PI3-K inhibitor LY294002 (3, 4) inhibited insulin-dependent survival of cerebellar neurons (Fig. 3) (17). LY294002 had little effect on cells grown in 25 mM KCl plus serum (Fig. 3) [or 25 mM KCl plus insulin (17)]. This suggests that the inhibition of insulin-dependent survival by LY294002 did not represent nonspecific toxicity, and also suggests that additional survival pathways may be activated by stimuli such as KCl.

Fig. 3.

Inhibition of insulin-promoted survival by the PI3-K inhibitor LY294002. Cells were placed in medium containing insulin (10 μg/ml), 25 mM KCl plus serum (10%), or no survival factor (−), in the absence or presence of LY294002 (LY, 10 μM) (17). After 2 or 4 days, cells were fixed, and the number of healthy cells was determined by staining with the DNA dye propidium iodide and counting nonapoptotic nuclei. Survival in the presence of both serum and 25 mM KCl (complete media) is defined as 100%. LY294002 inhibited the promotion of survival by insulin. Data are from three experiments; error bars indicate SEM.

We considered the possibility that the effects of PI3-K on cell survival might be mediated by the protein kinase Akt (also known as PKB-α or RAC-α) (18, 19), which is activated by a number of growth factors, including insulin, through a PI3-K-dependent mechanism (57, 20). Akt is a widely expressed cytoplasmic serine-threonine kinase, and its aberrant expression has been implicated in tumorigenesis (18, 19, 21). Akt contains at its NH2-terminus a domain termed the pleckstrin homology (PH) domain, which may regulate the activation of Akt by binding D3-phosphorylated phosphoinositides that are the products of PI3-K (6, 22). Phosphorylation of Akt also influences its activation (5, 20, 23), and the PH domain may influence the activation of Akt by promoting its dimerization (24). The functions of Akt are mostly unknown, with the exception of the identification of glycogen synthase kinase-3 (GSK-3) as a substrate; phosphorylation of GSK-3 by Akt is believed to regulate glycogen synthesis (20). To determine whether Akt might mediate survival, we stimulated cerebellar neurons with IGF-1 or insulin and assayed the activity of Akt (25). Both insulin and IGF-1 activated Akt (Fig. 4A). Activation of Akt was blocked by wortmannin (Fig. 4A) and by LY294002 (11), which suggested that this activation was dependent on PI3-K. KCl was found not to activate Akt or p70S6K (11), consistent with the possibility that KCl activates other survival pathways.

Fig. 4.

Activity and expression of Akt proteins. (A) PI3-K-dependent activation of Akt in cerebellar neurons (25). Cells were left untreated (U) or were treated for the indicated times with no stimulant (−), insulin, or IGF-1, in the absence or presence of wortmannin (W). Cell lysates were immunoprecipitated with an antibody to Akt, and immune-complex kinase assays were performed with either histone H2B or recombinant GSK-3 as substrate. With H2B as substrate, activation of Akt in cells treated with insulin for 15 min (n = 3) was 3.3 ± 0.44 times that in unstimulated cells. (B) Expression of transfected Akt in cerebellar neurons (26). Neurons were cotransfected with HA-Akt and CMV-β-Gal, and 2 days later they were fixed and immunostained with an antibody to β-Gal (Texas Red-coupled secondary antibody, red) and 12CA5 antibody to HA (Cy2-coupled secondary antibody, green). Transfected cells efficiently expressed both β-Gal (left) and HA-tagged Akt (center); a double exposure is shown on the right.

To determine the importance of Akt for insulin-dependent survival, we transfected expression vectors encoding wild-type Akt (HA-Akt) or two mutant forms of Akt, a catalytically inactive mutant [HA(K179M)] and a mutant encoding the PH domain (HA-PH) (26), into cerebellar neurons. Transfected neurons were identified by cotransfecting an expression vector for β-galactosidase (β-Gal) and immunostaining cells for β-Gal expression (Fig. 4B). To assess the effects of Akt, we scored transfected cells in a blinded manner as healthy or apoptotic by nuclear morphology. Apoptotic cerebellar neurons showed pronounced nuclear condensation, which was visualized with the DNA dye bisbenzimide (Hoechst 33258) (2); these nuclei also stained positively for DNA degradation in the TUNEL assay (11, 27).

We first tested whether expression of the mutant forms of Akt would interfere with survival; both mutants have been found to have dominant-inhibitory activity toward wild-type Akt kinase activity (28). Neurons transfected with HA-Akt had a normal, noncondensed nuclear morphology (Fig. 5A, in insulin). In contrast, cells transfected with either HA-Akt(K179M) or HA-PH showed increased apoptosis, as evidenced by nuclear condensation and disintegration of processes and the cell body (Fig. 5, A and B). The extent of apoptosis in cells transfected with HA-Akt(K179M) or HA-PH and then grown with insulin as the sole survival factor (∼60%) was as large as that in vector control transfectants in the presence of no survival factor (Fig. 5C) (11). Cells transfected with HA-Akt(K179M) or HA-PH showed less apoptosis when grown in the presence of both KCl and insulin [or KCl and serum (29)], consistent with the possibility that KCl, serum, or combinations of these factors may activate survival pathways in addition to the Akt pathway (29). Taken together, these results suggest that the promotion of survival by insulin requires Akt.

Fig. 5.

Effects of expression of mutant or wild-type Akt on cerebellar neuron apoptosis. (A) Cells were transfected with the indicated expression vectors (along with CMV-β-Gal), and 1 day later they were placed in medium containing insulin (10 μg/ml) as the survival factor. After 16 hours, cells were fixed, immunostained, and Hoechst-stained. Transfected cells are red (Texas Red, anti-β-Gal), and nuclei are stained with the DNA dye bisbenzimide (Hoechst 33258, blue and white). HA-Akt transfectants have large, nonapoptotic nuclei (white arrows) and normal processes (red arrow). Cells transfected with HA-Akt(K179M) or HA-PH (11) were apoptotic, as indicated by nuclear condensation (white arrow) and disintegration of cell body and processes (red arrow). (B) Quantitation of induction of apoptosis by Akt mutants. Cells were transfected with the indicated expression vectors, and 1 day later they were placed in medium containing insulin alone or insulin plus 25 mM KCl. After 16 hours, cells were fixed and stained, and transfectants were scored as healthy or apoptotic. Data are from four experiments; symbols above the bars indicate significant differences between conditions with identical symbols (ANOVA with Bonferroni correction for four plasmid conditions; *, P = 0.0034; #, P = 0.0047; ·, P = 0.0017). (C) Promotion of survival by Akt. Cells were transfected with either CMV-6 (control vector) or HA-Akt, and 1 day later they were placed in medium containing 25 mM KCl plus serum (10%) or in deprivation media (5 mM KCl, no serum). After 22 hours, cells were processed as in (B). Data are from three experiments; asterisks indicate significant difference (ANOVA, P = 0.0013). For (B) and (C), the total number of transfectants scored is shown within each bar; error bars indicate SEM. In (B) and (C), the percent apoptosis among control (CMV-6) transfectants in KCl plus insulin or KCl plus serum reflected a combination of the basal level of cell death in culture (typically 10% or less in KCl plus insulin or KCl plus serum) and transfection toxicity.

We next tested whether exogenously expressed Akt is sufficient to enhance survival. Cerebellar neurons were transfected with HA-Akt or with control vector and were deprived of survival factors after 1 day. Expression of HA-Akt markedly reduced the amount of apoptosis (Fig. 5C); control transfectants showed ∼60% apoptosis after 1 day of deprivation, whereas HA-Akt transfectants showed only 35% apoptosis [P = 0.0013 by analysis of variance (ANOVA)] (29). The ability of HA-Akt to block apoptosis was not reduced in the presence of LY294002 (11), consistent with Akt acting downstream of PI3-K.

Because both PI3-K and Akt can promote activation of p70S6K (5, 23), p70S6K is a potential mediator of the survival effects of IGF-1. Originally identified as a ribosomal protein S6 kinase (30), p70S6K has since been shown to regulate progression from the G1 to the S phase of the cell cycle (31). We inhibited the activation of p70S6K with rapamycin, which blocks phosphorylation of p70S6K (31). Rapamycin had no effect on the promotion of survival by insulin or by serum plus KCl, at a range of concentrations that blocked the activation of p70S6K (11). This is consistent with the reported lack of requirement for p70S6K activity for PI3-K-mediated survival of PC12 cells (4).

Taken together, our findings reveal that a critical function of Akt is to mediate the effects of IGF-1 on neuronal survival. Akt may promote the survival of a range of cell types in response to various growth factors, particularly those that activate PI3-K. The observation that Akt promotes survival may partially explain the oncogenic potential of Akt (18, 21). The promotion of survival by Akt may also be relevant to situations of pathological neuronal cell death, such as hypoxic-ischemic injury, for which IGF-1 can be protective (32). Because Akt is believed to be activated at least in part by lipid products of PI3-K (6), Akt may prove a propitious target for small-molecule therapeutics that promote cell survival.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
View Abstract

Navigate This Article