AMP-activated protein kinase phosphorylates retinoblastoma protein to control mammalian brain development

Abstract

AMP-activated protein kinase (AMPK) is an evolutionarily conserved metabolic sensor that responds to alterations in cellular energy levels to maintain energy balance. While its role in metabolic homeostasis is well documented, its role in mammalian development is less clear. Here we demonstrate that mutant mice lacking the regulatory AMPK beta1 subunit have profound brain abnormalities. The beta1(-/-) mice show atrophy of the dentate gyrus and cerebellum, and severe loss of neurons, oligodendrocytes, and myelination throughout the central nervous system. These abnormalities stem from reduced AMPK activity, with ensuing cell cycle defects in neural stem and progenitor cells (NPCs). The beta1(-/-) NPC deficits result from hypophosphorylation of the retinoblastoma protein (Rb), which is directly phosphorylated by AMPK at Ser(804). The AMPK-Rb axis is utilized by both growth factors and energy restriction to increase NPC growth. Our results reveal that AMPK integrates growth factor signaling with cell cycle control to regulate brain development.

Fig. 1. AMPKβ1-deficient mice show reduced AMPK activity and manifest brain abnormalities

(A) Immunoblot analysis of wildtype (+/+) and β1−/− brain lysates using β1/β2 C-terminal antibodies or (B) total or phosphorylated AMPK (AMPK^Thr172) and ACC (ACC^Ser79) antibodies. (n = 4 independent experiments). (C) Densitometric analysis of bands in A and B. (D) Macroscopic view of wildtype and β1−/−brain at P14. (CB; dotted & lower panel) in β1−/− mice. (E, F) Coronal brain sections showing atrophy of dentate gyrus (E; arrow) and cerebellum (F; IGL inner granule cell layer). NeuN immunohistochemistry of (G) cortex and dentate gyrus (arrow) and (H) cerebellum. (I) Quantification of neuronal losses assessed by counting NeuN⁺ cells. * p < 0.001 (n = 5).

Fig. 2. AMPKβ1-deficient mice demonstrate both neuronal and glial CNS deficits, seizures and reduced GABA receptor phosphorylation

MAP2 immunohistochemistry showing dendrites (A), silver staining showing axonal tracts (B), APC immunohistochemistry showing oligodendrocytes (C) and MBP immunohistochemistry showing myelination (D) in the P14 brain. (E) Electron microscopic analysis of P14 β1−/− optic nerves. (F) Quantification of GFAP-positive cells in wildtype and β1−/− brains. (G) Immunohistochemistry of the forebrain of E18.5 wildtype and β1−/− embryos using GFAP antibody. LV, lateral ventricle; arrow heads indicate migrating GFAP+ astroglia. (H) Quantitation of GFAP+ migrating astroglia in E18.5 wildtype and β1−/− brains. (I) EEG showing three seizure episodes recorded for 30 min in P14 β1−/− mice. The trace of one episode is enlarged at the bottom for clarity. (J) Immunohistochemistry of wildtype and β1−/− P14 brain sections and (K) Western blot analysis of brain lysates using phospho-GABA_BR2^Ser783 antibody. (L) Quantification of signal intensities in (K). * p = 0.01; ** p = 0.002; # p = 0.006.

Fig. 3. Loss of β1 results in neural stem/progenitor cell developmental defects

Immunohistochemistry of the E14.5 forebrain with antibodies against Ki67 (red) (A), BrdU (green) and Ki67 (red) (B), phosphohistone H3 (red) (C) and cleaved Caspase3 (red) (D). TUNEL staining (red) in conjunction with immunohistochemistry using Sox2 antibodies (green) (E) or nestin antibodies (green) at E14.5 (F). Inset in (F) shows magnified view of TUNEL/Nestin double⁺ apoptotic neural precursors. (G) Immunohistochemistry using BrdU (green) and cleaved Caspase 3 (red) antibodies 1 hr after BrdU injection. (H) A cartoon illustrating the different layers of the embryonic forebrain. (I) Immunohistochemistry using BrdU (green) and Ki67 (red) antibodies 24 hr after BrdU injection and (J) BrdU (green) and cleaved Caspase 3 (red) antibodies 24 hr after BrdU injection. (K) Immunohistochemistry of E14.5 wild type and β1−/− brain using Tuj1 (green) and cleaved Caspase3 (red), (L) E18.5 brains using Olig2 (green) and cleaved Caspase3 (red) and (M) P7 brains using GFAP (green) and cleaved Caspase3 (red) antibodies to detect apoptotic cells. Inset in (K) shows colocalization of Tuj1 and cleaved Caspase3 in migrating neurons. (N)) Quantitative analysis of apoptotic neural cells. DAPI staining (blue) was used to highlight the nuclei. * p < 0.001.

Fig. 4. AMPKβ1 loss results in cell-autonomous NPC defects

(A) Photomicrographs of telencephalic neurospheres cultured for 48 hr. Neurospheres were assayed for growth (B) and self renewal (C). 1° & 2° NS = primary & secondary neurospheres. (D) Wildtype and β1−/−NPCs were incubated with CFSE dye and the fluorescence intensity of the cells was measured at 0 and 96 hr. The numbers in parenthesis are mean fluorescence intensities. (E) Propidium iodide staining of unfixed wildtype and β1−/− neurospheres and quantitative analysis (F) of apoptosis of β1+/+, β1+/− and β1−/− NPCs. (G) Quantification of immunohistochemical results obtained from in vitro differentiation of β1+/+, β1+/− and β1−/− neurospheres showing percentage of neurons (Tuj1), oligodendrocytes (O4) and astrocytes (GFAP). * p < 0.001. Data is representative of at least three independent experiments.

Fig. 5. Expression of β1 but not β2 rescues β1−/− NPC phenotypes

(A) Immunoblot analysis of NPC lysates using phospho-AMPK^Thr172, total ACC and phospho-ACC^Ser79 antibodies. AMPKβ1−/− NPCs were infected with lentivirus expressing GFP (control), β1, β2 or constitutively active (ca) AMPKα2, while wildtype NPCs were infected with dominant negative (dn) AMPKα2 and the growth rate (B), self-renewal capacity (C) and number of apoptotic cells (D) were monitored. * p< 0.005; # p <0.05. Data is representative of three independent experiments.

Fig. 6. AMPK phosphorylates Rb to regulate Rb-E2F interaction

(A) AMPK consensus sequence (top), AMPK phosphorylation site in ACC and potential site found in Rb (bottom). (B) Immunoblot analysis of β1−/− NPC lysates using pRb(Ser^800/804), pRb(Ser⁷⁸⁰), pan Rb, and tubulin antibodies. Data is representative of three independent experiments. (C, D) Nonradioactive kinase assay: CDK4/6 complex immunoprecipitated using Cyclin D1/D2 antibodies from wild type and β1−/− NPCs (C) and AMPK holoenzyme immunoprecipitated using AMPKα1/2 antibody from wildtype NPCs, were used to phosphorylate recombinant Rb protein (residues 701–928). Phosphorylation was monitored by immunoblot using phosphoRb^800/804 antibody. Data is representative of three independent experiments. (E, F) Immunoprecipitation of NPC lysates with pan Rb antibody followed by Western blot with E2F1 antibody and densitometry showed the increased amount of E2F1-Rb complexes in β1−/−NPCs. (G) Proliferation assay of wildtype and β1−/− NPCs expressing GFP or the indicated Rb proteins. Data is representative of two independent experiments. (H) Flow cytometric analysis of NPCs expressing GFP or Rb proteins displaying the percentage of cells in G2M phase. FACS data is available with Online Supplement (Fig. S12–13; S15–21). (I) Proliferation assay of wildtype NPCs expressing GFP, constitutively active (ca) AMPK, β1, or β2 subunits. Data is representative of three independent experiments. (J) Immunoblot analysis of wildtype NPCs expressing caAMPK using pAMPKα^Thr172 and pRb^Ser800/804 and pan AMPKα and Rb antibodies. (K) Immunoprecipitation assay showing the level of E2F1 bound to Rb in NPCs expressing GFP (control), caAMPK, or dnAMPK. * p< 0.005.

Figure 7. Energy restriction and Growth factor signaling utilize the AMPK-Rb axis to enhance NPC proliferation

Immunoblot (A) and densitometric analysis (B) of wildtype NPCs using indicated antibodies after 2h growth factor stimulation in the presence or absence of AMPK inhibitor Compound C (CC, 5 mM). Proliferation assay of wildtype NPCs (C) and β1−/− NPCs (D) cultured for 48 hr under glucose limiting conditions. (Cont = 25 mM glucose, which is amount in Neurobasal medium). Data is representative of three independent experiments. Immunoblot (E) and densitometric analysis (F–K) using phospho-specific AMPKα^Thr172, Rb^Ser800/804, ACCSer⁷⁹, and pan AMPKα, Rb, and ACC antibodies. (L) Immunoprecipitation assay using Rb antibody and Western blot with E2F1 antibody from NPCs grown in the indicated glucose concentration in the absence or presence of Compound C (5 mM). # p= 0.07 * p 0.005. AU = arbitrary units.