Requirement for ribosomal protein S6 kinase 1 to mediate glycolysis and apoptosis resistance induced by Pten deficiency

Abstract

Pten inactivation promotes cell survival in leukemia cells by activating glycolytic metabolism. We found that targeting ribosomal protein S6 kinase 1 (S6K1) in Pten-deficient cells suppressed glycolysis and induced apoptosis. S6K1 knockdown decreased expression of HIF-1α, and HIF-1α was sufficient to restore glycolysis and survival of cells lacking S6K1. In the Pten^fl/fl Mx1-Cre⁺ mouse model of leukemia, S6K1 deletion delayed the development of leukemia. Thus, S6K1 is a critical mediator of glycolytic metabolism, cell survival, and leukemogenesis in Pten-deficient cells.

The Pten phosphatase is among the most commonly inactivated tumor suppressor proteins in human cancer. Spontaneous inactivating mutations in the Pten gene are found in cancers of the central nervous system, prostate, and endometrium at a frequency of 15–38% (1). In addition to mutations, Pten function can be reduced in cancer cells through epigenetic modifications, miRNA regulation, subcellular translocation, and posttranslational modification (2).

Pten expression levels determine the tissue spectrum and aggressiveness of neoplastic tumors. In hematopoietic cells, heterozygous mice with one functional allele of Pten develop a lymphoproliferative autoimmune disease (3), whereas complete deletion in hematopoietic cells triggers aggressive lymphoid and myeloid leukemias (4, 5). Pten deficiency contributes to the accumulation of tumor-initiating cells in cancers of hematopoietic, prostate, and brain tissues (4, 6, 7). Increased numbers of tumor-initiating cells indicate a need for targeted chemotherapeutic approaches to achieve long-term cancer remission in cancers associated with Pten inactivation.

Loss of Pten triggers the accumulation of the lipid products of the class 1A phosphatidylinositol-3 kinases (PI3K) and activation of the Akt/PKB protein kinases. Among the three mammalian isoforms of the Akt kinases, Akt1 is required for oncogenesis in mice that are heterozygous for a null allele of Pten (8). Activation of Akt induces glycolytic metabolism and renders cells hypersensitive to interruptions in glycolysis, suggesting that Akt metabolic control can be targeted to induce apoptosis in cancer cells (9, 10). Rapamycin, an inhibitor of the mammalian target of rapamycin complex 1 (mTORC1), can prevent Akt-induced glycolysis (11). This indicates that substrates of mTORC1 are likely mediators for Akt-induced glycolysis, but the array of mTORC1 substrates that mediate glycolysis in Pten-deficient cells is not known.

The ribosomal protein S6 kinase 1 (S6K1) is an attractive target downstream of mTORC1 for activation of glycolysis in Pten-deficient cells. mTORC1 phosphorylation activates the protein kinase activity of S6K1, which in turn regulates protein translation by phosphorylating proteins that regulate translation initiation (12–14). S6K1 also functions in hormonal control of circulating glucose through effects in insulin-responsive tissues—S6K1^−/− mice are glucose intolerant and exhibit increased blood glucose levels when fed a high fat diet (15). Because it can be inhibited using compounds selective for its ATP-binding pocket, S6K1 is a potential target for developing novel chemotherapeutics. We tested the potential for targeting S6K1 to reduce glycolytic metabolism and restore apoptosis in cellular and mouse models of Pten-deficient leukemogenesis.

Results

S6K1 Is Required to Maintain Glycolysis and Survival in Pten-Deficient Cells.

Pten inactivation induces Akt signaling, apoptosis resistance, and glycolytic metabolism in cancer cells. Loss of Pten is known to activate the protein kinase S6K1, but the role of S6K1 in regulating apoptosis resistance and glycolytic metabolism in carcinogenesis is not known. To determine the role of S6K1 in regulating apoptosis in Pten-deficient cells, we transduced IL-3–dependent hematopoietic progenitor FL5.12 cells with shRNA expression vectors targeting Pten (shPten) and/or S6K1 (shS6K1; Fig. S1A). S6K1 shRNA and siRNA reduced S6K1 expression without affecting the expression of the related kinase S6K2 (Fig. S1 B and 1C). Pten-deficient cells but not controls have elevated Akt phosphorylation and resist the induction of apoptosis when cultured 48 h in the absence of growth factor (Fig. 1A and Fig. S1A). S6K1 knockdown abrogated cell survival in Pten-deficient cells, indicating a requirement for S6K1 to transmit prosurvival signals downstream of Akt (Fig. 1A). S6K1 knockdown inhibited cell survival in Pten-deficient cells to a similar degree as the mTORC1 inhibitor rapamycin, indicating that S6K1 is a major mediator of survival signals downstream of mTORC1 (Fig. 1B).

Fig. 1.

S6K1 is necessary for survival and glycolysis in Pten-deficient cells. (A) S6K1 knockdown prevented cell survival of PTEN-knockdown cells. SD of triplicate viability measurements was <4% unless otherwise noted. (B) Similar to S6K1, 20 nM rapamycin reduced cell viability in shPten cells. (C) Decreased mitochondrial membrane potential in Pten/S6K1 double-knockdown cells. Viable cells (DAPI negative) were stained with the potentiometric dye TMRE after culture in absence of IL-3. TMRE staining in Bcl-xL–expressing cells is shown in both panels for reference. (D) S6K1 knockdown reduces glycolysis in Pten-deficient cells. The rate of cellular glycolysis was determined in a radiolabeled tracer assay after culture in the absence of IL-3 for 3 h.

Next we determined the impact of S6K1 knockdown on metabolic parameters associated with Akt-dependent survival. S6K1 knockdown correlated with decreased mitochondrial membrane potential compared with cells with Pten knockdown alone, suggesting that glucose-dependent cell survival may be compromised (Fig. 1C) (10). We therefore measured the cellular glycolytic rate using a radiolabeled tracer assay that correlates tightly with the flux of glucose through the glycolytic pathway (16). In cells cultured in the absence of growth factor, S6K1 knockdown reduced the rate of glycolysis in Pten-deficient cells (Fig. 1D). Consistent with reduced glycolysis, Pten/S6K1-deficient cells produced less lactate compared with Pten-deficient cells in the absence of growth factor (Fig. S2). Interestingly, S6K1 knockdown alone was sufficient to suppress glycolysis and accelerate cell death (Fig. 1 A and D). Metabolic decline preceded cell death because glycolysis and mitochondrial measurements were performed at 3 h post-IL-3 withdrawal, well before any decline in viability is evident (17).

Recent work has shown that programmed cell death can proceed through distinct pathways, each with its own molecular mediators (18). To determine whether loss of S6K1 activated the apoptosis pathway for cell death, we measured the effects of S6K1 knockdown on subcellular localization of Bax and cytochrome c in Pten-deficient cells. In viable cells, Bax is maintained in a cytosolic location, whereas in apoptotic cells Bax is associated with the mitochondrial outer membrane (19). When apoptosis was induced by culturing cells in the absence of growth factor, Pten knockdown significantly reduced Bax translocation from the cytosol to mitochondria (Fig. 2 A, lanes 7 and 8, and B). S6K1 knockdown restored Bax translocation in Pten-deficient cells, strongly indicating that S6K1 is required to prevent apoptosis induction in Pten-deficient cells. Bax translocation to the outer membrane is required for mitochondrial outer membrane permeability (MOMP) and the release of cytochrome c (20, 21). To determine if Bax translocation to mitochondria induced MOMP, we measured cytochrome c release to the cytosol in cells cultured in the absence of growth factor to induce cell death. S6K1 knockdown increased the fraction of cytochrome c in the cytosol in Pten-deficient cells, demonstrating that S6K1 inactivation induces an apoptotic form of programmed cell death in Pten-deficient cells (Fig. 2C).

Fig. 2.

S6K1 knockdown limits glycolysis to induce apoptosis. (A) S6K1 knockdown restores Bax translocation in Pten-deficient cells upon IL-3 withdrawal. Cytosolic (C) and mitochondrial (M) fractions were probed for Bax after culture ± IL-3 for 15 h. Cytochrome c oxidase subunit IV (Cox IV) was used as a marker for the mitochondrial fraction. (B) Bax abundance in cytosolic and mitochondrial fractions from A was quantified as a ratio. Pten-deficient cells maintained a relatively high cytosol:mitochondria ratio of Bax after withdrawal of growth factor (IL-3), but S6K1 knockdown counteracted this effect. (C) Cytochrome c is released in S6K1-deficient cells upon IL-3 withdrawal. Cytosolic and mitochondrial fractions were probed for cytochrome c after culture in the presence or absence of IL-3 for 18 h. (D) Glycolysis upon S6K1 knockdown is upstream of a cellular commitment to apoptosis. The rate of glycolysis was measured in cells expressing exogenous Bcl-xL to prevent the induction of apoptosis after culture in the absence of growth factor.

Decreased glycolysis in shPten/shS6K1 cells cultured in the absence of growth factor occurs as cells progress toward irreversible commitment to apoptosis. To rule out the possibility that S6K1 regulation of glycolysis is a consequence of cellular commitment to apoptosis, we measured glycolysis in knockdown cells overexpressing Bcl-xL. Bcl-xL expression prevented growth factor withdrawal-induced death (Fig. S3), permitting analysis of metabolic control in the absence of the effects of apoptosis commitment. In the absence of IL-3, S6K1 knockdown reduced glycolysis in Pten-deficient cells, demonstrating that S6K1 metabolic control is upstream of apoptosis commitment (Fig. 2D). Together, the data show that loss of S6K1 compromises the metabolic hallmarks of Akt-dependent survival, resulting in the induction of apoptosis.

S6K1 Induces Glycolysis Through HIF-1α.

Several studies have reported increased expression of hypoxia-inducible factor (HIF-1α) in Pten-deficient cancer cell lines (22, 23). Because HIF-1α can regulate the transcription of glycolytic enzymes, we analyzed HIF-1α expression in Pten/S6K1-deficient cells. Consistent with previous reports, we observed a small increase in HIF-1α expression in shPten cells cultured under normoxic conditions, compared with vector control cells (Fig. 3A). S6K1 knockdown in Pten-deficient cells suppressed HIF-1α expression under both normoxic and hypoxic (1% O₂) conditions, suggesting that limited HIF-1α expression could be responsible for decreased glycolysis. We confirmed this finding in PTEN-deficient PC3 prostate cancer cells. S6K1 siRNA suppressed HIF-1α expression compared with control siRNA-nucleofected cells (Fig. 3B). We treated PC3 cells with dimethyloxallyl glycine (DMOG), a prolyl hydroxylase inhibitor that prevents the oxygen-dependent degradation of HIF-1α. Treatment with DMOG triggered rapid accumulation of HIF-1α in all cells, but the amount of HIF-1α stabilized by DMOG was significantly reduced in S6K1 knockdown PC3 cells (Fig. 3B). Reduced accumulation of HIF-1α in response to an inhibitor of oxygen-dependent degradation indicates that S6K1 regulates synthesis of HIF-1α. No change in HIF-1α mRNA levels was observed upon S6K1 knockdown in PTEN-deficient cells (Fig. S4), consistent with a role for S6K1 in regulating HIF-1α production at the level of translation.

Fig. 3.

S6K1 regulates HIF-1α. (A) Endogenous HIF-1α is reduced in shPten/shS6K1 cells under normoxic conditions (N). shPten/shS6K1 also exhibited reduced HIF-1α stabilization under hypoxic (1% O₂) conditions (H). (B) S6K1 knockdown decreased HIF-1α expression in PC3 prostate cancer cells. DMOG treatment stabilized HIF-1α expression, but there was decreased HIF-1α accumulation in S6K1-depleted PC3 cells upon treatment with DMOG. (C) shPten cells were nucleofected with control, siHIF-1α, or siS6K1. siHIF-1α decreased glycolysis to a similar extent as observed with siS6K1. (D) Comparable decrease in mRNA expression of the glycolytic enzyme PGK-1 upon knockdown of S6K1 or HIF-1α. Data are the mean fold change in PGK-1 expression from three independent experiments ± SE. (E) HIF-1α is required to sustain cell survival in shPten cells. Cells were cultured in the absence of growth factor to measure apoptosis resistance.

Although there is strong evidence that increased HIF-1α correlates with Pten inactivation, it is not clear if HIF-1α is required for increased glycolysis and survival in Pten-deficient cells. To test if reduced HIF-1α abundance may account for decreased glycolysis in S6K1 knockdown cells, we compared the effects of HIF-1α and S6K1 knockdown in Pten-deficient cells. Depletion of HIF-1α reduced glycolysis to a level comparable to S6K1 knockdown (Fig. 3C). Moreover, down-regulation of HIF-1α and S6K1 together did not induce an additive decrease in glycolysis compared with knockdown of either HIF-1α or S6K1 alone, suggesting that the two proteins function in the same pathway (Fig. 3C).

The phosphoglycerate kinase 1 (PGK-1) enzyme is a glycolytic enzyme whose expression is responsive to HIF-1α. Expression of PGK-1 in cells upon knockdown of S6K1 or HIF-1α was reduced to a similar extent, indicating that S6K1 regulates the expression of HIF-1α target genes in glycolysis (Fig. 3D). Similar to glycolysis, knockdown of S6K1 or HIF-1α triggered reduced apoptosis resistance, and there was little additive activity of combined S6K1/HIF-1α knockdown in a viability time course (Fig. 3E). Thus, HIF-1α is necessary to promote glycolysis and apoptosis resistance in Pten knockdown cells.

Previous work demonstrated that limited glycolysis prevents survival of cells with inactivated Pten or constitutively active Akt (9, 10). To test whether decreased glycolysis contributes to limited survival in S6K1 knockdown cells, we restored glycolysis by expressing HIF-1α PP402,564AA, a mutant HIF-1α protein that lacks the proline hydroxylation sites in HIF-1α and thereby evades proteasomal degradation (24). Expression of the protein product was confirmed by intracellular flow cytometry (Fig. S5). We measured glycolysis in cells expressing the HIF-1α mutant and observed increased glycolysis relative to control cells, as expected (Fig. 4A). Some variability was observed in the rate of glycolysis in cell lines expressing mutant HIF-1α, suggesting additional layers of glycolysis regulation by a complex signaling network in response to Pten and/or S6K1 inactivation.

Fig. 4.

HIF-1α is sufficient to rescue glycolysis and survival in the absence of S6K1. (A) Expression of HIF-1α PP402,564AA (+HIF-1α) increased glycolysis over control (dashed line) independent of S6K1. (B) Increased viability in cells where glycolysis is supported by HIF-1α PP402,564AA independent of S6K1 expression.

Having increased glycolysis by expressing mutant HIF-1α, experiments could now address whether elevated glycolysis correlated with viability in Pten/S6K1 knockdown cells. Survival in Pten/S6K1 double-knockdown cells expressing mutant HIF-1α was comparable to cells deficient in Pten alone (Fig. 4B), which suggests that increased glycolysis contributes to the prosurvival effects of S6K1 in Pten-deficient cells. Although expression of the HIF-1α mutant also increased viability in control cells, the level of cell survival mediated by HIF-1α itself did not approach that of Pten knockdown cells, suggesting that increased glycolysis mediated by S6K1 and HIF-1α contributes to apoptosis resistance, but increased glycolysis alone is not sufficient for cell survival. Pten-deficient cells also activate S6K1-independent prosurvival activities, such as Akt inactivation of FOXO transcription factors. S6K1 activation of glycolysis likely cooperates with parallel prosurvival mechanisms to mediate apoptosis resistance in Pten-deficient cells.

S6K1 Deficiency Delays Leukemogenesis Induced by Pten Deletion.

Decreased glycolytic metabolism and apoptosis resistance upon loss of S6K1 would be predicted to impair oncogenic transformation in Pten-deficient cells. We crossed Pten^fl/fl mice expressing the Cre recombinase from the IFN-inducible Mx1-Cre transgene with S6K1^+/+ and S6K1^−/− mice. S6K1^−/− mice are viable but small in size, with no apparent alterations in hematopoiesis (25). Deletion of Pten in hematopoietic cells leads to myeloproliferative disease (MPD), acute lymphoblastic leukemia (ALL), and/or acute myeloid leukemia (AML) (4, 5). After injection with pIpC to induce Pten deletion, mice were monitored for development of leukemia. Pten^fl/fl Mx1-Cre⁺ S6K1^+/+ mice developed a fatal combination of MPD and T-cell ALL (T-ALL) with a mean survival of 35 d. Pten^fl/fl Mx1-Cre⁺ S6K1^−/− developed fatal disease with slower kinetics, increasing average lifespan to 46 d, an improvement of 32% (Fig. 5A). The end points of the disease in Pten^fl/fl S6K1^+/+ and Pten^fl/fl S6K1^−/− mice did not differ significantly in the magnitude of splenomegaly or thymic enlargement (Fig. S6). We also found no significant difference in the frequency of cell subpopulations (CD4⁺, CD8⁺, TCRβ⁺, B220⁺, Mac1⁺, and GR-1⁺) when assessed by flow cytometry (Fig. S6). To determine the role of S6K1 in regulating glycolysis in leukemogenesis, we analyzed the expression of PGK-1 in bone marrow progenitor cells that lack the expression of defined differentiation markers (lineage-negative, or Lin⁻, cells). Lin⁻ cells have been shown to contain leukemia-initiating cells in Pten^fl/fl mice (4). These cells isolated from Pten^fl/fl S6K1^−/− mice showed >90% reduction in PGK-1 expression compared with Pten^fl/fl S6K1^+/+ mice, suggesting that S6K1 loss may affect disease initiation and development by impairing glycolysis in this population (Fig. 5B). Together, the data reveal that S6K1 is required for oncogenic glycolytic metabolism and apoptosis resistance in PTEN-deficient neoplasia.

Fig. 5.

S6K1 loss delays Pten-deficient leukemia. (A) Survival was compared in Pten^fl/fl S6K1^+/+ (n = 24) and Pten^fl/fl S6K1^−/− (n = 14) mice after pIpC injection. Mean survival for Pten^fl/fl S6K1^+/+ mice was 35 d and 46 d for Pten^fl/fl S6K1^−/− mice. P value calculated by log-rank test. (B) PGK1 mRNA expression was analyzed in pooled Lin⁻ bone marrow cells from Pten^fl/fl S6K1^+/+ mice (n = 5) and Pten^fl/fl S6K1^−/− mice (n = 5) using qRT-PCR. Target gene expression in Pten^fl/fl S6K1^+/+ BM cells was set to 1.

Discussion

The findings shown here identify S6K1 as a critical kinase that activates glycolysis to support cell survival and transformation in Pten-deficient cells by controlling the production of HIF-1α. Pten-deficient cells accumulate increased levels of HIF-1α, which requires mTORC1 signaling (22, 23). In response to increased mTORC1 signaling, HIF-1α translation is increased via mechanisms that include increased phosphorylation of the inhibitor of cap-dependent initiation 4EBP1 (26). Our studies reveal a similar role for S6K1 in regulating HIF-1α levels in Pten-deficient cells. Furthermore, we show that HIF-1α is required to induce glycolysis to sustain cell survival in Pten knockdown cells. Importantly, HIF-1α reexpression was sufficient to restore glycolysis and apoptosis resistance in cells that lack S6K1, which strongly supports the regulation of glycolysis as a critical function of S6K1.

S6K1 deficiency functioned similarly to rapamycin in Pten-deficient leukemia, extending the lifespan of mice following deletion of Pten. One difference between targeting S6K1 and rapamycin treatment is that rapamycin suppressed leukemogenesis when coadministered with pIpC in Pten^fl/fl Mx1-Cre⁺ mice, whereas the S6K1^−/− delayed disease development (4). It is possible that rapamycin suppressed leukemia because of an ability to suppress proliferation through the regulation of the 4EBP proteins (27). However, the antiproliferative effects of rapamycin-induced dephosphorylation of 4EBP-1 may be detrimental to cytotoxic chemotherapy, as some studies have shown that it is preferable to maintain cell cycle progression while treating with cytotoxic agents to maximize induction of apoptosis (28). Several inhibitors are described that inactivate S6K1 downstream of Akt (29, 30). Whether direct inactivation of S6K1-induced glycolysis or combined inhibition of metabolism and cell cycle progression by targeting mTORC1 is preferable in cancer therapy will require direct comparison in future studies.

Materials and Methods

Cell Culture and Viral Transductions.

IL-3–dependent FL5.12 cells were cultured as described (31). PC3 cells were cultured in DMEM (1×) supplemented with 10% FBS (HyClone), penicillin, and streptomycin. Bcl-xL–expressing FL5.12 cells were previously described (10, 32).

Constructs.

PCR reactions were used to prepare Pten shRNA targeting sequences fused to the human U6 promoter; the resulting cassette was cloned into the pKD-GFP vector (33). Targeting sequences: shPten: AAAAAAGGAGTATCTTGTACTCACCCTAACCTCGAGCTTAGGGTGAATACAAAATACTCCGGTGTTTCGTCCTTTCCACAA; Scrambled: AAAAAAGTCCTGCCTCGTAATAGCCGTACACCTCGAGCTGTACGGCTATTACGAGGCAGGACGGTGTTTCGTCCTTTCCACAA. The S6K1 knockdown construct was from Open Biosystems (TRCN0000022905). Bcl-xL and HIF-1α PP402,564AA constructs were expressed in the MIT retroviral vector. Retrovirus was prepared in 293 cells, and lentiviral vectors were produced by the Viral Vector Core at the Translational Core Laboratories, Cincinnati Children's Hospital Research Foundation in Cincinnati.

Immunoblots.

Immunoblots were performed with the following antibodies: Akt pS473, S6K1 pT389, Akt, S6K1, and cytochrome c from Cell Signaling Technology; Pten, Actin, Bax, and FOXO3a from Santa Cruz Biotechnology. HIF-1α blots were performed as described (34). Contrast, brightness, and levels adjustments were performed in Canvas and Photoshop software (ACDSee; Adobe). Average pixel density ± SD for densitometry was determined using Quantity One software (BioRad).

Flow Cytometry.

For viability assays, FL5.12 cells were washed three times, then plated in complete medium, lacking only IL-3, at a density of 2E5 cells per mL. Triplicate samples of each culture were resuspended in PBS containing 2 μg/mL propidium iodide and immediately analyzed on a flow cytometer. To assess mitochondrial membrane potential, cells were incubated for 30 min at 37 °C with 20 nM tetramethylrhodamine ethyl ester (TMRE) and 5 μg/mL DAPI with or without 5 μM chlorocarbonylcyanide phenylhydrazone (CCCP). TMRE fluorescence was measured using a FACSAria, with gating to remove DAPI⁺ dead cells. The HIF-1α intracellular stain used PE-conjugated anti-HIF-1α or isotype control Ab (R&D Systems) after fixing in 4% paraformaldehyde and permeabilizing in 0.1% saponin.

Nucleofection.

Using Nucleofector II (Amaxa Biosystems) and the G-016 program, 7–10E6 FL5.12 cells were nucleofected with 10 μg siRNA (Accell). At 24 h postnucleofection, 1E6 cells were starved for IL-3 for 3 h, following which glycolytic rates were measured. Cell viability and RNA isolation experiments were also performed 24 h postnucleofection. For PC3 cells, we used the same nucleofection kit with the program T-016. PC3 cells were nucleofected with 5–10 μg of S6K1 siRNA in complete media. At 24 h postnucleofection, media was changed to 0.1% FBS containing DMEM. After 18–20 h, cells were treated with 5 mM DMOG (Cayman Chemical Co.) for 90 min, following which cells were harvested for protein.

Quantitative Reverse-Transcription PCR.

RNA was isolated using Qiagen RNeasy mini kit coupled to DNase 1 digestion. A total of 1 μg RNA was reverse transcribed using Taqman Reverse Transcription reagents (Applied Biosystems). Quantitative PCR was performed using TaqMan Gene Expression Master Mix and HIF-1α,S6K1, PGK1, and S6K2 TaqMan probes (Applied Biosystems).

Mitochondrial Fractionation.

1E7 cells were incubated on ice, resuspended in 250 μM sucrose/20 mM MOPS containing protease and phosphatase inhibitors, sonicated, and cleared in a 500 × g spin. A 16,000 × g spin yielded a mitochondrial pellet and a cytosolic supernatant. Mitochondria were resuspended in sucrose/MOPS buffer, and Nonidet P-40 was added to both fractions to a final concentration of 1%.

Glycolysis Measurement.

Glycolysis was measured using 5-³H-glucose in a protocol adapted from Ashcroft et al. (35). 1E6 cells were cultured with 5 μCi of 5-³H-glucose and incubated for 1 h at 37 °C. 5-³H-glucose is converted exclusively by glycolysis to ³H-water during this incubation. After acid lysis, ³H-water was separated from 5-³H-glucose in a closed system consisting of an outer chamber filled with 1 mL of water and an inner chamber containing ³H-water generated by glycolysis. After 24–48 h at room temperature, ³H-water in the inner chamber equilibrated with the water in the outer chamber through evaporation and condensation. ³H-water in the outer chamber was then measured by scintillation counting, and standardized to controls containing pure ³H-water or pure 5-³H-glucose. Data are expressed as nanomoles of glucose converted per 1 × 10⁶ cells per hour.

Lactate Production.

4E6 cells were nucleofected with 5 μg siRNA. At 24 h postnucleofection, cells were cultured in the absence of IL-3 for 7 h, and lactate secretion into media was measured using a Lactate Assay Kit (Biovision Inc.) following the manufacturer's instructions.

Mice.

Pten^fl/fl mice on a mixed background (The Jackson Laboratory) were crossed with Mx1-Cre and S6K1^−/− mice (a gift from the laboratories of Sara Kozma and George Thomas, University of Cincinnati). Six- to 8-wk-old mice were i.p. injected twice with 12.5 μg of polyinosine-polycytidine (pIpC) (Invivogen) per gram of body weight every other day. Mice were euthanized upon observation of leukemia-associated symptoms. Experiments were conducted in accordance with the animal care policies of the Institutional Animal Care and Use Committee of the University of Cincinnati.

Isolation of Lin⁻ Bone Marrow Cells.

Six- to 8-wk-old mice were i.p. injected with 12.5 μg of pIpC per gram of mouse body weight for two consecutive days. Mice were euthanized 5 d after the first pIpC injection. Bone marrow was harvested from femurs and tibia, and cells were pooled from five mice of each genotype. Depletion of lineage-positive cells was performed using biotinylated antibodies specific for Ter119, CD3, GR1, Mac1, and B220 (559971; BD Biosciences), and streptavidin-coated magnetic beads (Miltenyi Biotec), using an AutoMACS cell sorter.

Statistical Analysis.

Statistical calculations were t tests performed using GraphPad Prism software.