Glycogen synthase kinase 3alpha and 3beta mediate a glucose-sensitive antiapoptotic signaling pathway to stabilize Mcl-1

Abstract

Glucose uptake and utilization are growth factor-stimulated processes that are frequently upregulated in cancer cells and that correlate with enhanced cell survival. The mechanism of metabolic protection from apoptosis, however, has been unclear. Here we identify a novel signaling pathway initiated by glucose catabolism that inhibited apoptotic death of growth factor-deprived cells. We show that increased glucose metabolism protected cells against the proapoptotic Bcl-2 family protein Bim and attenuated degradation of the antiapoptotic Bcl-2 family protein Mcl-1. Maintenance of Mcl-1 was critical for this protection, as glucose metabolism failed to protect Mcl-1-deficient cells from apoptosis. Increased glucose metabolism stabilized Mcl-1 in both cell lines and primary lymphocytes via inhibitory phosphorylation of glycogen synthase kinase 3alpha and 3beta (GSK-3alpha/beta), which otherwise promoted Mcl-1 degradation. While a number of kinases can phosphorylate and inhibit GSK-3alpha/beta, we provide evidence that protein kinase C may be stimulated by glucose-induced alterations in diacylglycerol levels or distribution to phosphorylate GSK-3alpha/beta, maintain Mcl-1 levels, and inhibit cell death. These data provide a novel nutrient-sensitive mechanism linking glucose metabolism and Bcl-2 family proteins via GSK-3 that may promote survival of cells with high rates of glucose utilization, such as growth factor-stimulated or cancerous cells.

FIG. 1.

Glut1/HK1 expression increases glucose metabolism and delays commitment to cell death. (A) Glycolytic flux was measured in control (Neo) and Glut1/HK1 cells in the presence or absence of IL-3. (B) Tandem mass spectrometry-based analysis of organic acids in control and Glut1/HK1 cells cultured in the presence or absence of IL-3 for 6 h. (C) Clonogenicity of multiple individual control, Glut1/HK1, and Bcl-x_L cells cultured in the absence of IL-3 for the indicated times, followed by readdition of IL-3. (D) Control and Glut1/HK1 cells were IL-3 withdrawn and cultured in 10 mM glucose, 10 mM 2DOG plus 1 mM methylpyruvate, or 5 mM glucose plus 5 mM 2DOG, and viability was determined by PI exclusion. (E) Bax activation was analyzed by immunoprecipitation (I.P.) of active Bax from cells cultured in the presence or absence of IL-3 for 10 h followed by Western blotting for total Bax. (F) Cytochrome c release in control (Neo), Glut1/HK1, and Bcl-x_L cells after IL-3 withdrawal. Cells were cultured in the presence or absence of IL-3 for the indicated times, fractionated to isolate cytosolic and mitochondrial proteins, and analyzed by Western blotting for cytochrome c. Values are means and standard deviation (n = 3) (A, B, and D) or means and standard errors of the means (n = 5) (C). N, Neo control; GH, Glut1/HK1; X, Bcl-x_L.

FIG. 2.

Bim is necessary for death and is induced normally in Glut1/HK1 cells, yet Bim toxicity is inhibited. (A and B) Cells were transfected with control (GFP) or Bim shRNAi and analyzed in the presence or absence of IL-3 for Bim expression (A) and survival (B). (C and D) Bim RNA (C) and protein (D) induction after IL-3 withdrawal in control (Neo), Glut1/HK1, and Bcl-x_L cells. Cells were cultured in the presence or absence of IL-3 for 10 h. Bim RNA levels were determined by RNase protection assay, and protein levels were determined by Western blotting. (E) Control (Neo), Glut1/HK1, and Bcl-x_L cells were cultured in the presence or absence of IL-3 for the indicated times. Bim levels in mitochondrial fractions were determined by Western blotting. (F) Control (Neo) and Glut1/HK1 cells were cultured in the presence of or withdrawn from IL-3 for 10 h. Bim was immunoprecipitated, and the precipitates were analyzed by immunoblotting for Bim, Bcl-2, and Mcl-1. (G and H) Control (Neo) or Bim_L plasmid was transiently transfected into control (Neo), Glut1/HK1, and Bcl-x_L cells. Bim expression was determined by Western blotting (G), and cell viability was determined by PI exclusion (H). Values are means and standard deviations (n = 3) (B and H). N, Neo control; GH, Glut1/HK1; X, Bcl-x_L.

FIG. 3.

Maintenance of Mcl-1 protein level is critical for protection of Glut1/HK1 cells upon IL-3 withdrawal. (A and B) Control, Glut1/HK1, and Bcl-x_L cells were cultured in the presence or absence of IL-3 for 10 h, and mitochondria were isolated. (A and B) Representative gel (A) and quantitated averages and standard deviations of mitochondrial Mcl-1 levels using Bcl-2 as a normalization control (B). (C) Control (Neo), Glut1/HK1, and Bcl-x_L cells were withdrawn from IL-3, and the total cellular Mcl-1/actin ratio was determined over time. (D and E) Control and Glut1/HK1 cells were transiently transfected with control (GFP) or Mcl-1 shRNAi, and then Mcl-1 protein level was determined by immunoblotting (D) and cell viability was observed by PI exclusion (E). Values are means for cell survival from triplicate samples; error bars indicate standard deviations. N, Neo control; GH, Glut1/HK1; X, Bcl-x_L.

FIG. 4.

Glucose uptake attenuates Mcl-1 ubiquitination and degradation after growth factor withdrawal. (A, B, and C) Control and Glut1/HK1 cells were withdrawn from IL-3 for 8 h and then treated with cycloheximide (CHX) to observe Mcl-1 degradation. (A) Total cellular Mcl-1 as a fraction of starting levels prior to CHX treatment. (B and C) Representative gels (B) and quantitated averages and standard deviations from three independent experiments (C) of mitochondrial Mcl-1 levels, using Bcl-2 as a normalization control, before (−) and after (+) CHX treatment. (D) HA-Mcl-1-expressing control and Glut1/HK1 cells were transiently transfected with FLAG-ubiquitin (FLAG-Ub) and withdrawn from IL-3 for 10 h. The proteasome inhibitor MG-132 (20 μM) was added in the last hour of culture. Lysates were immunoprecipitated with anti-HA and probed with anti-FLAG, anti-HA, or anti-heavy chain immunoglobulin. N, Neo control; GH, Glut1/HK1.

FIG. 5.

GSK-3 phosphorylation is increased in Glut1- and HK1-expressing cells and regulates both Mcl-1 degradation and cell death. (A) Mcl-1-expressing control, Glut1, or Glut1/HK1 cells were analyzed for phospho-S159 and total Mcl-1. Ratios are shown. (B) Multiple FL5.12 clones expressing Glut1 and/or HK1 were analyzed for phospho-Akt, phospho-GSK-3, and total GSK-3β by immunoblotting. (C) Control and Glut1/HK1 FL5.12, 32D, and BAF3 cells were cultured in IL-3 or withdrawn from IL-3 for 6 hours, and phospho-GSK-3 and total GSK-3β were determined by immunoblotting. Ratios are shown. (D and E) Cells were treated with GSK-3 inhibitor (10 μM SB216763), followed by immunoblotting for Mcl-1 and actin (ratios are shown) (D) and determination of cell viability upon IL-3 withdrawal (E). (F and G) FL5.12 cells were transiently transfected with shRNAi for GFP or GSK-3β. Cells were analyzed by immunoblotting for GSK-3β, Mcl-1, and actin in the presence or absence of IL-3 (ratios are shown) (F), and cell survival upon IL-3 withdrawal was determined by PI exclusion (G). Values are means from triplicate samples; error bars indicate standard deviations. M, hMcl-1; N, Neo control; G, Glut1; H, HK1; GH, Glut1/HK1.

FIG. 6.

Glut1 expression increases GSK-3 phosphorylation, Mcl-1 level, and survival in primary hematopoietic cells. (A) Bone marrow cells were cultured in IL-3, infected with retrovirus to express Glut1 and truncated hNGFR as a marker on day 4, and stained by immunofluorescence for phospho-GSK-3 (red), hNGFR (green), and DAPI on day 7. (B to E) Glut1 protein levels were analyzed by immunoblotting (B), glucose uptake was measured (C), and phospho-GSK-3 levels were determined by immunoblotting (D) and quantified (E) in purified T lymphocytes from nontransgenic (Non) and Glut1-transgenic (Glut1) littermates freshly after purification or after 1 day of culture in the absence of stimulation (neglect). (F) Mcl-1 levels were determined by immunoblotting of freshly isolated, neglected, or GSK-3 inhibitor (10 μM SB216763 [SB])-treated nontransgenic T cells. Ratios are shown. (G) Mcl-1/actin ratio quantitated from nontransgenic and Glut1-transgenic T cells after isolation (fresh), neglect, or culture in GSK-3 inhibitor (SB). (H) Purified T cells from nontransgenic and Glut1-transgenic littermate mice were cultured without stimulation (neglect), and cell survival determined after 2 days. Values shown are the means from triplicate samples; error bars indicate standard deviations.

FIG. 7.

PKC activity is necessary for GSK-3 phosphorylation and may be stimulated by altered diacylglycerol distribution in Glut1/HK1 cells. (A) Control and Glut1/HK1 FL5.12 cells were cultured in IL-3 in the presence of various kinase inhibitors (PKC, 10 μM Ro31-8220 [Ro]; PI3K, 10 μM LY294002 [LY]; PKA, 10 μM H89; CamK II, 10 μM KN93; mTOR, 25 nM rapamycin [Rap]; PKC, 10 μM GF 109203X; PKC, 2 μM Go6976; PKC, 10 μM Go 6983; PKC, 20 μM Rottlerin [Rott]) for 1 h, and phospho-GSK-3 levels were determined by immunoblotting. (B and C) Control and Glut1/HK1 cells transiently transfected to express PKD-CRD-GFP or PKD-CDR-GFP P155/287G were imaged by confocal microscopy (B), and cells with membrane-targeted GFP were blindly quantified (means and standard deviations from three independent experiments are shown) (C). (D) Glut1/HK1 cells were withdrawn from IL-3 for 3 hours in vehicle alone (Ctrl), PKC inhibitor (2 μM Ro31-8220 [Ro]), or PKC inhibitor plus GSK-3 inhibitor (10 μM SB216763 [SB]), and Mcl-1 and actin levels were detected by immunoblotting. Ratios are shown. (E) Control and Glut1/HK1 cells were withdrawn from IL-3 and cultured in the absence or presence of PKC (10 μM Ro31-8220 [Ro]) or GSK-3 (10 μM SB216763 [SB]) inhibitor, and cell survival was observed over time. Ratios are shown. N, Neo control; GH, Glut1/HK1.